Universidad de Granada

Facultad de Ciencias

Departamento de Física Aplicada Grupo de Física de Fluidos y Biocoloides

Caracterización Físico-química de Sistemas Coloidales Aplicados como Transportadores de

Fármacos

Manuel J. Santander Ortega Tesis Doctoral

Editor: Editorial de la Universidad de GranadaAutor: Manuel J. Santander OrtegaD.L.: GR. 2124-2008ISBN: 978-84-691-6449-5

Caracterización Físico-química de Sistemas Coloidales Aplicados como Transportadores

de Fármacos por

Manuel J. Santander Ortega Licenciado en Química

Directores de la Tesis Dr. D. Juan Luis Ortega Vinuesa Dra. Dña. Delfina Bastos González Prof. Titular de Física Aplicada Profa. Titular de Física Aplicada

Este trabajo de investigación se presenta para alcanzar el grado de

Doctor por la Universidad de Granada DOCTOR EUROPEUS

Granada, 2008

A mi Familia

Agradecimientos

En esta memoria está plasmado gran parte del trabajo que he llevado a cabo durante los últimos cinco años en el Departamento de Física de Fluidos y Biocoloides, perteneciente al Departamento de Física Aplicada de la Universidad de Granada. No obstante, antes de entrar en materia me gustaría dar mi más sincero agradecimiento a todos aquellos que me han apoyado y ayudado o que de alguna manera han contribuido a la realización de este trabajo.

En primer lugar quisiera empezar por mis directores de Tesis, Juan Luis Ortega Vinuesa y Delfi Bastos González. Os agradezco la confianza depositada en mí desde el principio, vuestro apoyo constante y dedicación. He de agradeceros la gran capacidad que habéis mostrado para trabajar en equipo y coordinarnos estando incluso a 2000 Km. de distancia, lo cual queda plasmado en esta memoria. A parte de lo meramente profesional os he de agradecer vuestra calidad humana, la cual nos ha permitido pasar muy buenos momentos durante estos cinco años. En fin, gracias por portaros conmigo más como maestros que como directores.

También quisiera agradecer a todo el grupo de Física de Fluidos y Biocoloides, tanto a los miembros activos como a los que lo fueron en su día, la oportunidad que me han brindado de conocer y participar en el mundo de la investigación, poniendo a mi disposición toda su experiencia. Especialmente a Pepe Callejas, J.M. Peula, gracias por vuestra ayuda.

Como se verá durante el desarrollo de esta memoria este trabajo se ha desarrollado en colaboración con dos departamentos de Tecnología Farmacéutica. Por lo que quisiera agradecer a la catedrática María José Alonso su dedicación durante los seis meses que duró mi estancia en Santiago de Compostela en el Grupo de Nanotecnologías Aplicadas al Diseño de Sistemas de Liberación de Fármacos. Esta estancia me abrió a su vez la posibilidad de optar a una beca Marie Curie predoctoral dentro de la red europea Galenos para realizar una estancia de un año en el grupo de Biopharmazie und Pharmazeutische Technologie perteneciente a la Universidad de Saarland, Alemania. En este momento he de agradecer la confianza depositada en mí por parte del catedrático Claus Michael Lehr, director del grupo, así como al Dr. Ulrich F. Schäfer, Uli, y a Herr Meiers por su apoyo y ayuda constantes, Vielen Dank!.

Como no podría ser de otra forma, en esta sección también deben aparecer mis compañeros becarios (o contratados), con los que he compartido penas y alegrías, y sin los cuales la realización de este trabajo habría sido mucho más dura. Les agradezco a todos ellos los momentos que hemos pasado juntos tanto en el laboratorio como fuera de él. Muchas gracias a Teresica quien me apadrinó en mis primeros meses en el 13 y con quien he pasado tan gratos momentos. A Roberto con quien he compartido tan buenas sobremesas aniquilándolo (pese a lo que él diga en su Tesis) y compartiendo sus teorías leónidas. Al Moro por ser el Moro, bienvenido otra vez al mundo exterior. A Sándalo, ese hombrecillo de tallo ramoso con hojas lampiñas originario de Persia y a Fernando Vereda, con quienes compartí tan enriquecedores momentos en el 16 y fuera de él cultivando mente, espíritu y las macetas de los estudiantes de Biología, y como no a Juanjo, Julia, Alberto, Miguel Angel, Maria, Arturo, Catalina, Juan Carlos, Fernando, Joaquín, Ceci, Jose Manuel y Pedro Gea (por los buenos momentos del 24), Cedric, Javi y a los nuevos, Miguel Alberto, Cesar, Miguel Wulf, Carlos, Amelia, Pablo, Carmen Rocío, Miriam y todos aquellos con los que he compartido estos cinco años.

Tampoco podría olvidar a mis compañeros de Santiago de Compostela. Gracias a Noémi por recibirme allá por septiembre de 2004 y enseñarme la diferencia entre sintetizar partículas y formularlas. Por supuesto a Nela y su pato con pelo, Ivana, Yolanda, Patri, Rafa, Puri, Ángela, María Alonso por su gran ayuda en el momento clave, María de la Fuente, Desi, Sascha, Noelia, Daya y Chachi y…..Aitana (¡bienvenida!), Fran, Pablo, Francisco y tantos otros que hicieron que me sintiese como en casa.

Cómo no, también he de acordarme de Ana, con quien tantos cafés compartí hablando sobre la idiosincrasia hispano-alemana, Steffania, katherina, Birgitt, Nico, Noha, Marc y el resto de becarios del grupo Biopharmazie und Pharmazeutische Technologie con los cuales pasé muy buenos momentos durante mi periplo alemán.

Ya fuera del ámbito profesional me gustaría dar las gracias a las personas que siempre han estado ahí, independientemente de la distancia y las circunstancias, a Raúl, Fernando, Jesús, Tito, Chorques (Duque de Ruf y vizconde de Canena) y Maria José, con quienes he compartido grandes momentos desde hace ya más de 10 añazos. Muchas gracias por estar ahí (o en Albuquerque) sin pedir nunca nada a cambio. Y especialmente a Vicky por darme todo su cariño y estar siempre a mi lado.

En este apartado tampoco podría faltar mi familia. Gracias a mis padres, Carmen y Manolo, por haberme dado todo y un poco más. A mis hermanos, por haber sido siempre un gran apoyo y referente. A Javi y Rocío, por ser como son y por supuesto a Maria quien sin saberlo me ayudó a discernir las cosas importantes de las que no lo son tanto en los malos momentos. Tampoco puedo olvidar a mis tíos Pepe y Ana quienes han sido para mi unos segundos padres. Gracias a todos.

Por último he de agradecer el apoyo financiero recibido por parte del Grupo de Física de Fluidos y Biocoloides de la Universidad de Granada así como de los proyectos MAT-2003-01257, P07-FQM-2496 y a la beca Marie Curie predoctoral perteneciente a la red Galenos enmarcada en el proyecto europeo “Towards an European PhD in Advanced Drug Delivery” MEST-CT-2004-404992, sin los cuales no habría sido posible desarrollar este trabajo.

Manuel J. Santander Ortega

30 de Agosto, 2008

Índice

I. Introducción……………………………………………………..………..1

II. Objetivos y Esquema……………………………………………....…….15

III. Breve Resumen y Discusión de los Resultados…………………………19

IV. Conclusiones……………….……………………………………………31

V. Bloque I: Caracterización de nanopartículas aplicadas como sistemas de liberación controlada de fármacos.

Paper I: Stability and Physico-chemical Characteristics of PLGA, PLGA:poloxamer and PLGA: poloxamine Blend Nanoparticles: a Comparative Study.......................................43

Paper II: Colloidal Stability of Pluronic F68® Coated PLGA Nanoparticles a Variety of Stabilization Mechanisms……………………………………………………………...…61

Paper III: Electrophoretic Mobility and Colloidal Stability of PLGA Particles Coated with IgG…………………………………………………………………………………..81

Paper IV: Nanoparticles made from Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers. Part 1.……………………………………………………...101

Paper V: Characterization of Core-Shell Lipid-Chitosan and Lipid-Poloxamer nanocapsules Insulin-Loaded PLGA Nanoparticles for Oral Administration: an in vitro Physico-chemical Characterization……………………………………………………..117

VI. Bloque II: Aplicación in vitro de nanopartículas poliméricas como sistemas de liberación controlada de fármacos.

Paper VI: Protein-Loaded PLGA Nanoparticles for Parenteral Administration ……………………………………………………………………………….………….139

Paper VII: Insulin-Loaded PLGA Nanoparticles for Oral Administration: an in vitro Physico-chemical Characterization…………………………………………………..…155

Paper VIII: Development of Novel Drug-Assembled Nanoparticles made from Amphiphilic Polymers .............………………………………….………………………173

Paper IX: Nanoparticles made from Novel Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers. Part 2.……………………………………………183

I Introducción

Un coloide o suspensión coloidal es un sistema compuesto por dos fases: una continua (normalmente fluida) y otra dispersa en forma de partículas de tamaño mesoscópico. En función del estado de agregación de ambas fases se pueden obtener diferentes tipos de suspensiones coloidales, tales como espumas, aerosoles, emulsiones, geles o soles. Esta memoria se centrará en el uso de dispersiones coloidales con una fase continua acuosa y una fase dispersa compuesta por nanopartículasa en estado sólido, o al menos cuya superficie o corteza puede considerarse sólida. Actualmente, debido a sus importantes aplicaciones industriales y biomédicas el estudio de los coloides ha cobrado una gran importancia dentro de los campos de la Química-Física, Física Aplicada así como de la Tecnología Farmacéutica.

El uso de estos coloides como sistemas de liberación controlada de fármacos presenta varias ventajas con respecto a la administración del fármaco libre. Cuando un fármaco es administrado in vivo su destino final está condicionado por varios procesos, tales como su absorción, distribución, metabolismo y finalmente su eliminación. Por otro lado, normalmente su efecto farmacológico está ligado con su toxicidad, estabilidad, solubilidad, capacidad para atravesar membranas o biodistribución. Por lo que si se consigue encapsular el fármaco en un transportador (fase dispersa de una suspensión coloidal) podríamos por un lado separar el efecto farmacológico (deseado) de las otras propiedades (no deseadas) así como tener cierto control sobre la absorción y distribución del mismo, lo cual afectará a su metabolismo y eliminación final [1,2].

Se puede decir que el concepto de nanopartículas aplicadas como sistemas para la liberación controlada de fármacos surgió cuando Paul Ehrlichb asistió a la opera “Der Freischüztz” de Carl Maria von Weber [3]. En esta opera el espíritu del demonio -“Freikugeln”- juega un papel central. Este “Freikugeln” siempre alcanzaba sus objetivos, independientemente de los obstáculos que tenía que

a Aunque actualmente no se puede encontrar una definición clara de nanopartícula, durante el desarrollo de esta tesis usaremos este término para denotar partículas que presentan una o más de sus dimensiones del orden de 100 nm. b Paul Ehrlich (14 de Marzo de 1854 – 20 de Agosto de 1915) fue un inmunólogo Alemán que ganó el premio Nobel en fisiología y medicina en 1908. Su trabajo se basó en el estudio de la barrera hematoencefálica.

2 I Introducción

superar para ello. Es por esto que Paul Ehrlich pensó en un sistema, bautizado como “Zauberkugeln” el cual podría transportar un fármaco hasta el órgano diana deseado a través del organismo sin dañar al resto de los tejidos y liberarlo de forma controlada [4].

Desde los tiempos de Paul Ehrlich hasta nuestros días el desarrollo y uso de nanopartículas como sistemas de liberación controlada de fármacos ha evolucionado considerablemente [5]. Como es evidente, la evolución de las nanopartículas ha estado ligada al desarrollo de nuevas técnicas para su preparación, así como, se han tenido que buscar nuevas materias primas en función del uso final que se les quisiera dar a las mismas [6]. Entre los nanosistemas de liberación controlada de fármacos más comunes podemos encontrar nanopartículas poliméricas, nanocápsulas, micelas, liposomas, dendrímeros, nanopartículas magnéticas y un largo etcétera. La elección de uno u otro dependerá del fármaco a encapsular y de la aplicación que se le quiera dar al sistema coloidal [1]. Un indicativo de la alta repercusión de estos sistemas es el gran número de medicamentos que podemos encontrar hoy en día en el mercado basados en sistemas coloidales [7,8]. Sin embargo, tanto los materiales empleados en el desarrollo de estos sistemas así como sus productos de degradación han de ser biocompatiblesc. Como consecuencia, uno de los grandes retos de de la investigación en el campo de la Tecnología Farmacéutica en la actualidad se centra en la búsqueda de nuevos materiales que cumplan estos requisitos. Por otro lado, pese al interés creciente por desarrollar este tipo de sistemas, la mayoría de los estudios realizados con sistemas biodegradables no incluyen una completa caracterización físico-química de los mismos. Sin embargo, las propiedades físico-químicas no sólo afectan al proceso de encapsulación y liberación del fármaco, sino que también gobiernan los procesos de interacción de las partículas con diferentes compuestos biológicos (proteínas o membranas) del tejido donde se introducen [10]. Realmente, el éxito en el desarrollo de estos polímeros actuando como transportadores depende en gran medida de conocer toda la información posible acerca de la naturaleza química y la estructura física de estos nuevos materiales y su interacción con el medio biológico en el que van a encontrarse.

En base a lo anterior, este proyecto de Tesis planteó una investigación centrada en el estudio de nuevos nanosistemas con potencial uso como sistemas de liberación controlada de fármacos que incluyera una completa caracterización físico-química de los mismos. Para llevarla a cabo hemos trabajado con tres tipos de nanosistemas, dos de ellos nanopartículas poliméricas preparadas, unas de ellas a partir de ácido poli-D-láctico-co-glicólico (PLGA), mientras que las otras se formularon utilizando polímeros anfifílicos. El tercer sistema lo formaron nanocapsulas lipídicas con una cubierta mixta de quitosano y Pluronic® F68.

c Se entiende por biocompatibilidad la capacidad de un material para no presentar efectos tóxicos o dañinos sobre los sistemas biológicos [9].

I Introducción 3

Presentamos a continuación un breve resumen con las principales características de los mismos.

- Nanopartículas mixtas de PLGA y poloxámeros o poloxaminas

El PLGA es un polímero biodegradable y biocompatible ampliamente utilizado en la preparación de micro/nanopartículas, mediante técnicas de emulsión, salting out, etc [11-13]. Éstas se han empleado como transportadores de diferentes principios activos tales como proteínas, péptidos o material genético entre otros [8]. Hoy en día existen varias formulaciones de PLGA que han sido aprobadas por la US Food and Drug Administration (US FDA) [14], como Suprecur® MP para el tratamiento del cáncer de próstata o Nutropin Depot® en el caso de trastornos de la hormona de crecimiento [8].

Pese a su amplio uso, las nanopartículas de PLGA presentan ciertos inconvenientes a la hora de encapsular y almacenar proteínas, péptidos y otros principios activos. La base de estos inconvenientes está relacionada con el mecanismo de degradación del polímero. Al tratarse de un poliéster su degradación se debe a la hidrólisis de los enlaces ester, lo que genera un microclima ácido dentro de la nanopartícula [15,16]. Esto puede comprometer la estabilidad del principio activo encapsulado, perdiendo por tanto su efecto farmacológico. Actualmente se han probado diferentes alternativas para minimizar este problema. Entre otras podemos destacar el uso de complejos de zinc o de tampones a la hora de encapsular moléculas [17]. Por otro lado, se han usado diferentes polímeros de bloque, tales como poloxámeros y poloxaminas [13] los cuales evitan la interacción de la molécula encapsulada con el PLGA y neutralizan la acidez generada en la degradación del polímero [18-21].

El uso de poloxámeros y poloxaminas como excipientes en la formulación de nanopartículas de PLGA ha permitido mejorar otros aspectos de estos sistemas coloidales. Por ejemplo, cuando estas partículas se preparan mediante técnicas de emulsión se ha observado que el uso de estos surfactantes disminuye el estrés que sufren las proteínas durante su encapsulación [22]. Asimismo, cuando se usan estos excipientes una gran parte se situará en la superficie de las nanopartículas [23,24]. Esta disposición del surfactante, recubriendo a la nanopartícula, hace que el sistema presente una mayor estabilidad coloidal, y ayudan claramente a evitar la opsonizaciónd [25].

d La opsonización es conocida como el proceso mediante el cual un agente exógeno, en nuestro caso las nanopartículas, es eliminado del flujo sanguíneo por parte del sistema fagocitario mononuclear. Este fenómeno se da mediante la adsorción de anticuerpos, específicamente las opsoninas, a través de su fragmento Fab a la superficie de las nanopartículas (fenómeno muy favorecido en sistemas hidrófobos, como ya veremos en las siguientes secciones de esta memoria), quedando expuesto hacia el medio el fragmento Fc el cual es reconocido por los fagocitos presentes en la sangre. Por este motivo

4 I Introducción

Finalmente se ha visto que estos excipientes también afectan a la liberación de los principios activos una vez encapsulados [26,27]. Mediante la incorporación de estos excipientes a la matriz de PLGA ha sido posible disminuir el efecto burste inicial, a la vez que se ha podido modular la velocidad de liberación de las moléculas encapsuladas en función del surfactante empleado como excipiente [27].

- Nanopartículas preparadas con polímeros anfifílicos

Un paso más en la evolución de las nanopartículas poliméricas ha sido el uso de polímeros anfifílicos -compuestos por fragmentos hidrófilos e hidrófobos- para su preparación. Como se verá en esta sección este tipo de polímeros puede presentar ciertas ventajas con respecto al uso de polímeros hidrófobos o hidrófilos.

Los polímeros anfifílicos han sido usados para preparar diferentes sistemas de liberación controlada de fármacos tales como micelas, nanocápsulas o nanopartículas [28-30]. Por un lado podemos destacar el uso de polímeros hidrófobos, tales como ácido poli-láctico (PLA), PLGA o poli-ε-caprolactona (PCL) modificados covalentemente con cadenas de poli-etilenglicol (PEG) [31,32]. De este modo ha sido posible obtener nanopartículas con una cubierta de PEG que no necesitan la adición de un surfactante para evitar su captación por parte del sistema fagocitario mononuclear [31], y además mejoran su interacción con las superficies biológicas, tales como la mucosa nasal e intestinal [32]. Con este tipo de sistemas se ha llegado incluso a obtener mejores resultados en comparación a nanopartículas formuladas con el polímero sin modificar a las que posteriormente se les adsorbió un surfactante [33].

Por otro lado podemos remarcar el uso de polímeros con un esqueleto hidrófilo que han sido modificados covalentemente con polímeros hidrófobos. En este grupo podemos encontrar polímeros de quitosano o dextrano modificados con diferentes restos hidrófobos [29,34]. Una importante ventaja de estas nanopartículas con respecto a las mencionadas anteriormente (PEG-poliester) es que en este caso quedará un número suficiente de grupos reactivos en el polímero los cuales pueden servir para vectorizar el sistema mediante la unión covalente de ligandos reconocibles por las superficies biológicas [35]. Además se ha comprobado que el uso de dextrano permite conseguir un efecto similar al PEG, obteniendo sistemas con una baja energía superficial en los que se previene la adsorción indeseada de proteínas [36].

es necesario evitar la adsorción de proteínas sobre las nanopartículas una vez que han sido administradas. e El efecto burst es conocido como la liberación producida en los primeros momentos de la incubación de las partículas de una forma incontrolada. Normalmente se asocia con la pérdida de las moléculas situadas en la parte más externa de la partícula, las cuales se disocian más fácilmente de la misma. Este fenómeno es más acentuado en nanosistemas debido a la gran cantidad de superficie que presentan.

I Introducción 5

Considerando los resultados comentados en los párrafos anteriores, el uso de polímeros anfifílicos para preparar nanopartículas es actualmente un campo de investigación muy atractivo [37]. Por este motivo una parte de la presente memoria se ha dedicado a la puesta a punto y aplicación in vitro de varios sistemas coloidales basados en polímeros anfifílicos. Sin embargo, como se explicará más adelante, debido a que los resultados obtenidos se están usando en la preparación de una patente, tanto el nombre de los polímeros empleados así como los detalles de la preparación de las nanopartículas han sido suprimidos de la presente memoria.

- Nanocápsulas lipídicas recubiertas de quitosano y poloxámero

Las nanocápsulas son sistemas vesiculares en los cuales el fármaco o principio activo se encuentra confinado en una cavidad generalmente oleosa la cual está rodeada por una membrana de polímero. Este polímero puede incorporarse bien por una reacción de polimerización en dicha interfase o por adsorción del mismo en esta zona de la nanoemulsión [38]. La incorporación de este polímero mejora la liberación del fármaco, aumenta la estabilidad del sistema en fluidos fisiológicos y lo hace más versátil debido a las posibles modificaciones del polímero que forma la cubierta de la nanocápsula [39].

Si las comparamos con las nanopartículas, generalmente estos sistemas presentan una mayor afinidad por el fármaco, lo que se traduce en una mayor capacidad de encapsulación [38,40]. Por otro lado, al encontrarse el fármaco exclusivamente confinado dentro de la cavidad oleosa y protegido por una membrana de polímero pueden ofrecer protección a la mucosa frente a la toxicidad del fármaco [41], así como se protegerá al fármaco de su degradación en el ambiente fisiológico y se puede disminuir el efecto burst asociado con el fármaco situado en la superficie de los nanotransportadores [40].

Actualmente pueden encontrarse en bibliografía nanocápsulas formuladas con diferentes tipos de polímero en su cubierta, tales como PLA, poli-alquilcianocrilato (PACA) [12], poli-isobutilcianocrilato (PIBCA) [42], o PCL [43]. Últimamente también está despertando cierto interés el uso de una cubierta de quitosano [44] debido a sus excelentes propiedades, tales como biocompatibilidad, mucoadhesividad [45] y capacidad de abrir la uniones íntimas de las células [46]. Por otro lado, al tratarse de un polisacárido presenta grupos hidroxilo que permiten la modificación covalente del mismo con ligandos específicos, obteniéndose de esta forma sistemas coloidales polifuncionales [47,48].

No obstante, al igual que ocurría con las nanopartículas, es de vital importancia diseñar sistemas coloidales “invisibles” a los macrófagos presentes en el flujo sanguíneo. Esto se ha conseguido mediante el uso de polímeros modificados covalentemente con cadenas de PEG [48,49] o bien por la co-adsorción del polímero y de Pluronic® F68 [38,50]. Esta última variante en la cual hemos focalizado nuestro

6 I Introducción

estudio, presenta en la actualidad resultados prometedores encaminados al tratamiento del cancer [50].

Como se ha visto en los párrafos anteriores los coloidales aplicados como sistemas de liberación controlada de fármacos presentan una gran relevancia desde un punto de vista tanto académico como industrial. También se desprende de lo anterior que uno de los principales parámetros que determina si los sistemas pueden ser utilizados o no como transportadores es la estabilidad que presenten en el medio en el que estén supendidos. Por este motivo, el análisis de los mecanismos que rigen su estabilidad es de suma importancia. Ésta dependerá de las interacciones entre las partículas coloidales. Dichas interacciones dependen a su vez tanto de las características superficiales de las partículas como de las condiciones del medio en el que se encuentran. Dada su importancia, gran parte de la caracterización físico-química llevada a cabo en este trabajo se ha centrado en el análisis de la estabilidad y comportamiento electrocinético de nuestros sistemas. Por consiguiente, dedicaremos el resto de la introducción a los distintos tipos de interacciones que pueden aparecer entre las partículas coloidales y las diferentes teorías desarrolladas para explicar el origen de su estabilidad coloidal.

- Estabilidad Coloidal

Una característica fundamental de las dispersiones coloidales es que presentan una gran área de contacto entre la fase dispersa y el medio de dispersión (agua en nuestro caso). La energía asociada a la creación de este área es elevada, por lo que termodinámicamente el sistema tenderá a la agregación para disminuir su área. Sin embargo, si entre las partículas que se aproximan existe una interacción repulsiva tal que supere su energía cinética, el sistema permanecerá cinéticamente estable. Hamaker [51] y Boer [52] fueron los primeros autores que establecieron una teoría de estabilización para coloides liófobos. Estas teorías fueron posteriormente ampliadas por Derjaguin y Landau [53], e independientemente por Verwey y Overbeek [54,55]. La teoría general de estabilidad coloidal, para sistemas liófobos, creada por estos autores es conocida como teoría DLVO, nombre que corresponde a las iniciales de sus creadores. A partir de ahora nos referiremos a los sistemas liófobos como hidrófobos, ya que en nuestro caso el medio de dispersión siempre será agua.

Los tipos de interacciones que se dan en coloides hidrófobos, según la teoría DLVO, se pueden clasificar en dos grandes grupos:

1. Fuerzas de atracción de van der Waals.

2. Fuerzas repulsivas de origen electrostático.

I Introducción 7

La teoría DLVO considera que el potencial total de interacción entre dos partículas, en función de la distancia existente entre ambas (H), se puede expresar como una suma de los potenciales independientes antes mencionados:

( ) ( ) ( )T A EV H V H V H= + (1)

Siendo VA el potencial atractivo debido a las fuerzas de dispersión de van der Waals y VE el potencial repulsivo debido al solapamiento de las dobles capas eléctricas.

1. Interacciones de van der Waals

Las interacciones de van der Waals se darán siempre y serán de naturaleza atractiva. Estas fuerzas pueden ser de diferente naturaleza:

a) Fuerzas de dispersión de London.

b) Fuerzas de orientación de dipolos o de Keesom.

c) Fuerzas de inducción de dipolos o de Debye.

Para el caso que nos ocupa (cuerpos macroscópicos en sistemas condensados) las fuerzas de dispersión de London son las más importantes [56,57].

Hamaker [51,58] extendió el tratamiento de London [59] con el objetivo de obtener la energía de atracción entre dos partículas. Para dos esferas con radio equivalente y simetría esférica el potencial de atracción de London-van der Waals puede expresarse como:

( ) ( )( )

( )

2 2

2

42 2( ) ln6 4 2 2

A 2

H a HA a aV HH a H a H a H

⎡ ⎤+= − + +⎢ ⎥

+ + +⎢ ⎥⎣ ⎦ (2)

donde A es la constante de Hamaker (A = π2N2B siendo B la constante de London y N la densidad de moléculas por unidad de material) para partículas de coloide en agua y a es el radio de las mismas

2. Interacciones de origen electrostático

Cuando dos partículas se aproximan, las partes difusas de su doble capa eléctrica (dce) comienzan a interpenetrarse originándose una fuerza repulsiva entre las mismas (VE). Si la magnitud de esta interacción es superior a la atracción de London-van der Waals el coloide se encontrará estabilizado electrostáticamente.

Esta interacción repulsiva puede ser considerada desde dos puntos de vista diferentes. Mediante un cambio de energía libre de Gibbs al solaparse la dce de ambas partículas, o bien, mediante la presión osmótica generada por la acumulación de iones entre las partículas.

8 I Introducción

De acuerdo con el modelo de potencial constante, VE puede ser expresado -para un potencial en la capa de Stern moderado, <50 mV- como [60-62]:

2( ( 2 ))

04( ) 2 ( ) HB

E ri

k TV H a ez e

κπε ε γ − − Δ⎛ ⎞= + Δ ⎜ ⎟

⎝ ⎠ (3)

donde Δ es el espesor de la capa de Stern, ε es la constante dieléctrica del medio, ε0 la permitividad del vacío, kB la constante de Boltzman, T la temperatura absoluta, zi la valencia del ion, e la unidad de carga elemental, κ la longitud de Debye, expresada como:

2 20

0

i ii

B

n e z

k Tκ

ε ε=∑

(4)

siendo ni0 la concentración tanto de contraiones como coiones en el seno de la disolución. Por otro lado γ queda definida como:

tanh4

i d

B

z ek Tψ

γ = (5)

siendo ψd el potencial de Stern. Hoy día la expresión (5) es válida para κa>>1.

0 1 2 3 4 5

-20

0

20

40

0 1 2 3 4 5-20

-10

0

10

20

Como se ha comentado anteriormente, la teoría DLVO establece que el potencial de interacción entre dos partículas se puede expresar según la ecuación (1). La Figura 1 muestra una curva de potencial de interacción típica obtenida por

VT(H)VA(H)

(a)

MínimoPrimario

MínimoSecundario

H (nm)

VT/k

T

VT/k

T

H (nm)

VT=dV/dH=0conc. crit. coag.

κH=1

(b)

VE(H)

Barrera

Figura 1. Energía libre de interacción entre dos partículas (a). Situación en la ccc (b).

I Introducción 9

combinación de las ecuaciones (2) (VA) y (3) (VE). En este gráfico podemos diferenciar tres zonas:

1. Una primera zona en la cual se pasa de un potencial de interacción nulo (H = ∞) a un mínimo secundario de no mucha profundidad. Si este mínimo es suficientemente profundo se producirá la floculación del sistema. Como se verá más adelante la floculación es un proceso reversible.

2. En la zona intermedia aparece un máximo en el potencial de interacción, la altura de esta barrera determinará la estabilidad del coloide y dependerá del potencial difuso de las dobles capas, de la constante de Hamaker y de la concentración de electrolito. Si por agitación térmica estas partículas tienen la energía suficiente como para superar esta barrera de energía, podrán acercarse hasta distancias muy próximas, donde se sitúa el mínimo primario.

3. En la tercera zona, para valores de H pequeños encontramos un mínimo primario. Si el sistema se encuentra en este mínimo habrá coagulado. Al contrario que la floculación la coagulación es un proceso difícilmente reversible.

Un aumento en la concentración de sal del medio provoca una compactación de la dce, por lo que dos partículas podrán aproximarse a una distancia tal que la intensidad de las fuerzas atractivas de London-van der Waals comienza a ser importante. Esto implica una disminución de la barrera de potencial de interacción entre dos partículas, por lo que el número de colisiones que conduce a la coagulación del sistema aumenta, disminuyendo la estabilidad del mismo. Una vez alcanzada una concentración de electrolito, conocida como concentración crítica de coagulación (ccc), la barrera energética desaparece y todas las colisiones entre partículas son eficaces, ver Figura 2. Esta situación, donde la velocidad del proceso de agregación está limitada por la difusión de las partículas, se conoce como DLCA (difusión limited colloidal aggregation), y difiere del proceso de agregación que se da a concentraciones menores de electrolito, donde no todas las colisiones entre las partículas generan agregados, conociéndose este último mecanismo como RLCA (reaction limited colloidal aggregation).

Fuchs [63] introdujo un factor que considera la efectividad de las colisiones entre las partículas. Se conoce como factor de estabilidad (W), ecuación (6), y es igual a la inversa del factor de eficiencia de la colisión (α). Para la ccc, W valdrá 1 y para disoluciones con una menor concentración de electrolito será superior a 1:

1 r

l

kWkα

= = (6)

siendo kr la constante de velocidad de agregación en condiciones DLCA y kl la constante de velocidad en condiciones RLCA.

Verwey y Overbeek [54] relacionaron W con la energía de interacción de las partículas a través de la siguiente expresión:

10 I Introducción

( )

( / )

202

2

TV kTeW a dHa H

∞=

+∫ (7)

Aunque la teoría DLVO explica bastante bien el comportamiento de numerosos sistemas coloidales, como comentó el propio Overbeek [64] presenta ciertos puntos débiles, siendo imposible explicar interacciones tales como las fuerzas de depleción, la interacción estérica o las fuerzas de hidratación. En esta memoria nos centraremos en el estudio de la estabilización estérica y las fuerzas de hidratación, dos tipos de interacciones que no contempla la teoría DLVO.

0 1 2 3 4 5-12

-10

-8

-6

-4

-2

0

2

4

6

8

10

12

- Fenómenos no-DLVO

i) Estabilización estérica

Son numerosos los autores que hablan de un mecanismo adicional al de la repulsión electrostática que estabiliza ciertas disoluciones coloidales; dicho mecanismo es conocido como estabilización estérica. Este tipo de estabilización se debe a la presencia de cadenas de un polímero, para el cual la fase continua actúe como un buen solvente, situadas sobre la superficie de las partículas. Este tipo de estabilización puede estar causada por dos mecanismos diferentes:

VT/K

T

H (nm)

Figura 2. Potencial total de interacción entre dos partículas de PLGA en función de laconcentración de NaCl del medio. 25 mM (línea negra), 75 mM (línea roja), 125 mM (líneaazul), 200 mM (línea verde), 300 mM (línea rosa), 500 mM (línea marron) 800 mM (línea azulmarino). ψd= 15,3 mV, A=0,5·10-20 J.

I Introducción 11

-Cuando dos partículas se acercan, se produce un solapamiento entre las capas más externas de sus superficies. Este solapamiento provoca un aumento de la concentración local de polímero, lo que implica un aumento de la energía libre, siempre que la fase continua sea un buen solvente. Como resultado de este aumento de energía libre habrá una tendencia de las moléculas de disolvente a entrar en esa zona y separar a las partículas. Este efecto se conoce como efecto osmótico.

-Un segundo mecanismo que puede darse está relacionado con el contacto entre las cadenas poliméricas situadas en la superficie de ambas partículas. A pequeñas distancias su libertad conformacional queda reducida enormemente, lo que implica una disminución de la entropía conformacional de las cadenas y esto genera un aumento en la energía libre del sistema. Este aumento de la energía libre del sistema hará que el acercamiento esté desfavorecido energéticamente. Este efecto se conoce como efecto elástico.

Si la contribución estérica es menor en valor absoluto que la interacción de London-van der Waals se habla de estabilización electroestérica. La estabilidad de un coloide estabilizado por esta vía es sensible a la concentración de electrolito. Cuando la contribución estérica sea superior a la de London-van der Waals se habla de estabilización estérica. Un coloide estéricamente estabilizado no puede ser floculado mediante la adición de un electrolito al sistema [65].

Vincent y col. [66] fueron los primeros que hicieron un tratamiento cuantitativo del problema. Según ellos, si sobre la superficie de la partícula existe una capa externa de cadenas poliméricas con un espesor δ, aparecerá un fenómeno osmótico cuando ambas partículas se encuentren a una distancia inferior a 2δ. En este caso el potencial de repulsión osmótica (Vosm) viene dado por:

22

1

4 1( ) ( )2 2osm

a HV H π φ χ δν

⎛ ⎞⎛= −⎜ ⎟⎜⎝ ⎠⎝

⎞− ⎟⎠

(8)

siendo ν1 el volumen molar del disolvente, φ2 la fracción de volumen efectiva del polímero y χ es el parámetro de solubilidad de Flory-Huggins para dichas cadenas de polímero. Cuando las partículas se encuentran a una distancia inferior a δ, aparece el efecto elástico, que genera un nuevo potencial de repulsión (Vel). Además, a esta distancia, la expresión del potencial osmótico varía, obteniéndose la siguiente expresión:

2 22

1

4 1 1( ) ( ) ln2 2 4osm

a HV H π φ χ δ Hν δ δ

⎡ ⎤⎛ ⎞ ⎛= − − −⎜ ⎟ ⎜⎞⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

(9)

12 I Introducción

2

2 2 22

( )

3 / 3 /ln 6 ln 3 1

2 2

el

m

aV H

P

H H H H H

πφ δ ρ

δ δ

δ δ δ

=

− −− +

⎛ ⎞⎜ ⎟⎝ ⎠

⎡ ⎛ ⎞⎛ ⎞ ⎛ ⎞ ⎛⎜ ⎟ ⎜ ⎟ ⎜⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠ ⎝⎣ ⎝ ⎠

i

i +⎤⎞⎟⎠⎦

(10)

donde ρ2 es la densidad del polímero y Pm es el peso molecular de la cadena extendida hacia el disolvente. En esta situación el potencial total de interacción entre 2 partículas vendrá dado como:

( ) ( ) ( ) ( ) ( )T A E osm elV H V H V H V H V H= + + + (11)

ii) Fuerzas de Hidratación

En los sistemas coloidales muy hidrófilos o hidrófobos el medio de dispersión no puede ser considerado como un continuo, sino que presenta una estructura característica a distancias próximas a la superficie de la partícula [67,68]. Cuando dos partículas que presentan capas estructuradas de agua a su alrededor interaccionan aparece una fuerza. Esta fuerza, llamada estructural o no-DLVO, va a ser repulsiva (fuerzas de hidratación) cuando las dos superficies son hidrófilas o atractiva cuando son hidrófobas (fuerzas hidrófobas). La teoría DLVO no puede explicar este tipo de interacción ya que considera al medio como un continuo. Las fuerzas de hidratación aparecerán en coloides cargados en disoluciones salinas por la interacción de las esferas de hidratación de los contrapones localizados en las proximidades de las dos superficies que interaccionan [69,70].

Las fuerzas estructurales son de corto alcance y su existencia se suponía desde hace tiempo aunque las dudas no desaparecieron hasta que no se midieron directamente con un aparato de fuerzas superficiales (SFA) [71]. Los datos experimentales publicados sobre las fuerzas de hidratación presentan, por lo general, una dependencia exponencial con la distancia de separación, H, entre dos superficies planas según la siguiente expresión [72,73]:

( / )0( ) HP H P e λ−= (12)

En el caso de fuerzas hidrófobas el valor de P0 es negativo, siendo positivo para las fuerzas de hidratación. El valor de la constante de decaimiento, λ, depende de la hidratación del ion [74]. La constante de hidratación P0 depende de la hidratación de la superficie [75] y ha sido determinada experimentalmente para diferentes materiales teniendo un valor entre 106-5·108 N/m2 [70]. Dependiendo de cómo los iones presentes en la disolución afecten a la estructura del agua que rodea a las partículas (destruyéndola o aumentándola), un incremento en la concentración iónica provocará una disminución o un aumento de las fuerzas de hidratación [67]. En cualquier caso estas fuerzas dependen en gran medida del tipo de electrolito

I Introducción 13

presente en la disolución, siendo más fuertes cuanto mayor es la energía de hidratación del ion.

En los años 90 se encontró que las fuerzas de hidratación pueden ser explicadas en parte si se considera en la ecuación de Poisson-Bolzmann el descenso en la constante dieléctrica de la doble capa al aumentar la concentración de electrolito [76]. Este descenso en la constante dieléctrica provoca un aumento de la energía libre de hidratación de los iones de la doble capa los cuales se repelen surgiendo una fuerza repulsiva. Esta fuerza varía exponencialmente con la distancia de separación entre las superficies. Su valor aumenta rápidamente a partir de una concentración dada de electrolito, conocida como csc (critic stabilization concentration) y depende de la energía de hidratación del ion.

Para establecer una justificación teórica sobre la estabilidad anómala que presentan los sistemas con un gran carácter hidrófilo se han de incluir las fuerzas de hidratación en la teoría clásica DLVO. A partir de la ecuación 12, que corresponde a las fuerzas de hidratación para dos superficies planas, y usando la aproximación de Derjaguin [77] se puede obtener la siguiente expresión para el potencial de interacción debido a las fuerzas de hidratación entre dos esferas de radio a:

( / ) 2 2 ( / )0 0( ) H

h H HV H a P e d H aP e Hλ λπ π

∞ ∞ −= =∫ ∫ λ − (13)

Con el objetivo de considerar la influencia de la concentración de electrolito sobre la estabilidad de los complejos en la zona no-DLVO, Molina-Bolívar [78] supuso, como aproximación inicial, que las fuerzas de hidratación variaban linealmente con la concentración de electrolito:

2 ( / )( ) ( ) Hh A h eV H a N C c e λπ λ −= (14)

donde ce es la concentración de electrolito (mM), NA el número de Avogadro y Ch una constante de proporcionalidad llamada constante de hidratación.

Para concentraciones de electrolito inferiores a la csc la contribución de las fuerzas de hidratación al potencial total de interacción se puede suponer que es despreciable, ya que estas fuerzas, como se puede ver en el factor pre-exponencial de la ecuación 14, presentan una dependencia con la concentración de electrolito del medio. Para concentraciones de electrolito superiores a la csc se usará la teoría DLVO extendida incluyendo el potencial de hidratación.

( ) ( ) ( ) ( )T A E hV H V H V H V H= + + (15)

A bajas concentraciones de electrolito la altura de la barrera de potencial disminuye conforme aumenta la fuerza iónica. A concentraciones superiores a la csc aparece de nuevo una barrera energética que dificulta la agregación, aumentando su altura con la fuerza iónica, Figura 3. Es de destacar la existencia de mínimos secundarios en la zona no-DLVO al introducir el potencial de hidratación. Su existencia puede provocar reversibilidad en la agregación coloidal.

14 I Introducción

0 1 2 3 4 5-8

-6

-4

-2

0

2

4

6

8

VT/K

T

H (nm)

Figura 3. Potencial total de interacción entre dos partículas de PLGA recubiertas dePluronic® F68 en función de la concentración de NaCl del medio. 50 mM (línea negra), 100 mM (línea roja), 200 mM (línea verde), 400 mM (línea azul), 540 mM (línea turquesa), 800mM (línea rosa), 1000 mM (línea marrón).

II Objetivos y Esquema

El principal objetivo de esta tesis ha sido la caracterización físico-química y el desarrollo de nuevos sistemas coloidales aplicados como sistemas de liberación controlada de fármacos. La investigación llevada a cabo en la misma se ha centrado en tres áreas interrelacionadas entre sí. i) Preparación de los sistemas coloidales; ii) caracterización físico-química de los mismos; iii) aplicación final de las nanopartículas mediante la encapsulación de moléculas, estudio de su comportamiento en condiciones fisiológicas y su administración in vitro.

Como se ha comentado en la introducción el conocimiento de las interacciones entre las partículas y el medio en el que se encuentran resulta de gran importancia para la comprensión de los procesos que gobiernan la degradación y la liberación de estos sistemas. No obstante, es difícil encontrar en bibliografía trabajos en los que se estudie de una manera profunda y sistemática el comportamiento de estos coloides en condiciones fisiológicas, así como el efecto de cada uno de sus componentes en las propiedades finales de los mismos. Es por esto que esta Tesis ha sido planificada con la idea de obtener un mejor entendimiento de los parámetros que rigen las interacciones en dichos sistemas, siendo éste, a nuestro parecer el mejor camino para poder optimizar el desarrollo de dichos sistemas coloidales.

Esta Tesis presenta un carácter tanto aplicado como básico. Por un lado el carácter claramente aplicado se pone de manifiesto a través del análisis de la encapsulación y posterior liberación in vitro de diversas macromoléculas en los sistemas coloidales previamente comentados. Por otro lado el estudio físico-químico de los coloides conlleva el carácter más básico de la investigación. Todo ello no hubiera sido posible sin la estrecha colaboración de diferentes grupos de investigación provenientes algunos de ellos del campo de la Tecnología Farmacéutica. De hecho, parte de esta investigación se ha llevado a cabo durante 6 meses en el grupo de Nanotecnologías Aplicadas al Diseño de Sistemas de Liberación de Fármacos perteneciente a la Universidad de Santiago de Compostela bajo la supervisión de la catedrática Mª José Alonso. Otra parte de la Tesis se ha desarrollado en el grupo de Biopharmazie und Pharmazeutische Technologie de la Universidad de Saarland en Alemania bajo al supervisión del catedrático Claus

16 II Objetivos y Esquema

Michael Lehr. Esta última estancia tuvo una duración de 12 meses y fue posible gracias a la concesión de una beca Marie Curie predoctoral dentro de la red europea Galenos para realizar estancias en centros de investigación extranjeros.

También ha sido necesario el uso de diversas técnicas experimentales, basándose la mayoría de ellas en dispersión de luz dinámica (DLS) y estática (SLS), tales como espectroscopía de fotocorrelación (PCS) o determinación de la movilidad electroforética, así como microscopía de fuerzas atómicas (AFM) o análisis de perfil de la forma de gotas axisimétricas (ADSA P). Por otro lado, cuando se encapsularon moléculas en estos sistemas se emplearon técnicas tales como la cromatografía líquida de alta resolución (HPLC) o células de difusión de Franz para realizar los estudios.

Centrándonos en la estructuración de esta memoria, la misma ha sido dividida en dos bloques principales. Los trabajos que componen cada bloque no se han ordenado de forma cronológica, si no de una forma tal que se facilite la comprensión del trabajo realizado. En este momento es preciso decir que el trabajo realizado en la Universidad de Saarland bajo al supervisión del catedrático C.M. Lehr, plasmado en los papers IV, VIII y IX, se está empleando como base para el desarrollo de una patente en colaboración con una compañía química internacional, razón por la que en la presentación de estos trabajos se ha obviado tanto el nombre propio de los materiales así como los detalles de la preparación de dichos sistemas coloidales.

El primer bloque de esta memoria presenta la caracterización físico-química de los sistemas aplicados como transportadores de fármacos comentados en la introducción, es decir, nanopartículas poliméricas preparadas con PLGA o bien con polímeros anfifílicos y por otro lado nanocápsulas lipídicas con una cubierta de quitosano y Pluronic® F68. Este bloque se compone de cinco trabajos:

• Paper I: Stability and Physico-Chemical Characteristics of PLGA, PLGA:poloxamer and PLGA:poloxamine Blend Nanoparticles: a Comparative Study.

• Paper II: Colloidal Stability of Pluronic® F68 Coated PLGA Nanoparticles: a Variety of Stabilization Mechanisms.

• Paper III: Electrophoretic Mobility and Colloidal Stability of PLGA Particles Coated with IgG.

• Paper IV: Nanoparticles made from Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers. Part 1.

• Paper V: Characterization of Core-Shell Lipid-Chitosan and Lipid-Poloxamer Nanocapsules.

Una vez que los sistemas fueron caracterizados se pasó a realizar un estudio más aplicado sobre las nanopartículas poliméricas comentadas anteriormente. En los trabajos englobados en este bloque se analizó, mediante la

II Objetivos y Esquema 17

encapsulación de diferentes moléculas modelo, el efecto de la matriz de dichos sistemas coloidales así como las propiedades de la molécula encapsulada en el comportamiento del sistema como transportador de fármacos, estando fuera de nuestro objetivo el uso de un principio activo dado para tratar una dolencia determinada. De esta forma consideramos que es posible conocer mejor el potencial de dichas nanopartículas. Este bloque está compuesto por cuatro trabajos:

• Paper VI: Protein-Loaded PLGA Nanoparticles for Parenteral Administration.

• Paper VII: Insulin-Loaded PLGA Nanoparticles for Oral Administration: an in vitro Physico-Chemical Characterization.

• Paper VIII: Development of Novel Drug-Assembled Nanoparticles made from Amphiphilic Polymers.

• Paper IX: Nanoparticles made from Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers. Part 2.

III Breve Resumen y Discusión de Resultados Paper I: Stability and Physico-Chemical Characteristics of PLGA, PLGA:poloxamer and PLGA: poloxamine Blend Nanoparticles: a Comparative Study.

J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

En este primer trabajo se llevó a cabo la caracterización electrocinética así como un estudio de la estabilidad de tres sistemas coloidales basados en PLGA. El primero estaba formado únicamente por PLGA, mientras que los otros dos, denotados genéricamente como formulaciones mezcla, presentaban una matriz compuesta por PLGA y un poloxámero, Pluronic® F68 (PLGA-PF68) o una poloxamina, Tetronic® 904 (PLGA-T904).

La caracterización electroforética de estos sistemas en función del pH mostró que tanto las nanopartículas PLGA como las formulaciones mezcla presentaban un comportamiento típico de sistemas coloidales con grupos carboxilo superficiales. Aunque las curvas de movilidad en función del pH fueron cualitativamente idénticas para los tres sistemas, la incorporación del surfactante en las formulaciones mezcla produjo un apantallamiento de la carga superficial de las mismas. Este apantallamiento se manifestó experimentalmente en una disminución de la magnitud de su movilidad electroforética en comparación con las nanopartículas compuestas únicamente por PLGA. Como se comenta en el artículo, la reducción de la movilidad electroforética de las formulaciones mezcla con respecto a las nanopartículas de PLGA se da en parte por el desplazamiento del plano de deslizamiento de la nanopartícula cuando un surfactante se coloca en la superficie de la misma. Midiendo la movilidad electroforética de estos sistemas en función de la concentración de sal del medio (NaCl en este caso) y haciendo uso de la ecuación de Eversole-Boardman fue posible calcular el desplazamiento del plano de deslizamiento para los tres sistemas. Este desplazamiento siguió el orden PLGA-PF68 > PLGA-T904 > PLGA. Las diferencias observadas entre ambas formulaciones mezcla puede explicarse en base a la estructura del poloxámero y la poloxamina. El primero presenta restos hidrófilos de mayor tamaño, los cuales pueden penetrar más profundamente en la fase acuosa produciendo por tanto un mayor desplazamiento del plano de deslizamiento de la nanopartícula.

20 III Breve Resumen y Discusión de Resultados

Con respecto a la estabilidad coloidal de los tres sistemas, las nanopartículas PLGA mostraron el comportamiento típico descrito por la DLVO para sistemas liófobos. Los valores de concentración crítica de coagulación (ccc) obtenidos mostraron la necesidad de la incorporación de surfactantes a dicho sistema para obtener nanopartículas suficientemente estables bajo condiciones físico-químicas similares a las encontradas en medios fisiológicos. Por otro lado, ambas formulaciones mezcla fueron totalmente estables independientemente de la concentración de sal del medio. Este resultado sugiere que ambas formulaciones mezcla se encontraban estabilizadas debido al efecto estérico producido por el surfactante presente en la superficie de las mismas. Esto implica una clara mejoría del sistema con respecto a las nanopartículas compuestas únicamente por PLGA a la hora de ser usadas en medios fisiológicos. No obstante, a muy altas concentraciones de iones divalentes, tales como Ca2+ o Ba2+ y en presencia de aniones polivalentes de gran tamaño (por ejemplo, fosfato) se observó la formación de grandes agregados que presentaban un tamaño estable en el tiempo. La formación de estos agregados fue debida a una interacción (similar a la formación de complejos en química) de los cationes divalentes con los átomos oxigeno presentes tanto en el poloxámero como en la poloxamina conjuntamente con los iones fosfato del medio.

Paper II: Colloidal Stability of Pluronic® F68 Coated PLGA Nanoparticles: a Variety of Stabilization Mechanisms. J. Colloid Interface Sci. 302, 2006, 522-529.

Basándonos en los resultados obtenidos en el trabajo anterior y con la idea de entender aún mejor los mecanismos que regían la estabilidad de las formulaciones mezcla se decidió llevar a cabo un segundo trabajo basado en la adsorción de Pluronic® F68 sobre nanopartículas de PLGA. Tanto el comportamiento electrocinético como la estabilidad de los complejos PLGA-poloxámero obtenidos por adsorción fueron estudiados en función de la cantidad de surfactante adsorbido.

El análisis de la movilidad electroforética en función del pH de los complejos PLGA-poloxámero mostró cómo según aumentaba el grado de recubrimiento de las nanopartículas PLGA su comportamiento electrocinético se asemejaba al de la formulación PLGA-PF68. Sin embargo, fue necesario superar el plateau de adsorción, esto es, alcanzar concentraciones de poloxámero cercanas a su concentración micelar crítica (CMC) (ver Figura 2 de este artículo) para obtener el mismo comportamiento electrocinético de la formulación mezcla. A concentraciones cercanas a su CMC se ha comprobado que los surfactantes pueden adsorberse no solamente como moléculas individuales, si no que también lo harán en forma de hemi-micelas, por lo que estos resultados sugieren que en la formulación PLGA-PF68 el poloxámero se encontraba en forma de agregados sobre la superficie de las nanopartículas, y no como una mono-capa molecular.

III Breve Resumen y Discusión de Resultados 21

No obstante, y con la intención de constatar esta hipótesis se estudió la estabilidad de estos complejos en función de la fuerza iónica del medio para los diferentes grados de recubrimiento de las nanopartículas. Cuando se adsorbieron pequeñas cantidades de poloxámero la ccc del complejo disminuyó en comparación con las nanopartículas no recubiertas. Este comportamiento puede explicarse considerando la teoría DLVO, a bajos recubrimientos el poloxámero situado en la superficie de las nanopartículas apantallará la carga de las mismas, lo cual disminuirá su estabilidad. Por otro lado, como se comenta en el artículo, en estas condiciones la conformación adquirida por el poloxámero no permite que el sistema contrarreste la disminución del potencial-ζ mediante un mecanismo de estabilización estérica. Al aumentar el grado de recubrimiento se observó un aumento de la estabilidad de los complejos, lo cual era indicativo de la aparición de un nuevo mecanismo de estabilización. Al aumentar el grado de recubrimiento el poloxámero muestra sus restos hidrófilos más extendidos hacia el agua, lo cual favorece la estabilización estérica de los complejos y por tanto un aumento de su estabilidad. No obstante, lo más sorprendente fue que estos complejos empezaron a mostrar fenómenos de re-estabilización a altas concentraciones de sal característicos de las fuerzas de hidratación descritas en la introducción de esta memoria. Es decir, al aumentar la fuerza iónica del medio los sistemas se desestabilizaban, pero, si la fuerza iónica seguía aumentando el complejo volvía a ser estable. El proceso de re-estabilización en sistemas coloidales depende del electrolito usado como agente coagulante y del carácter hidrófilo de la superficie de las nanopartículas. Es por esto que para un mismo electrolito la re-estabilización será más intensa cuanto más hidrófila sea la superficie de la nanopartícula. Según esto, la aparición de estos mecanismos de re-estabilización implica que la presencia del poloxámero en la superficie de las nanopartículas de PLGA cambió de hidrófobo a hidrófilo el carácter de su superficie. Finalmente, para recubrimientos cercanos a la CMC del poloxámero, condiciones en las cuales éste se encontraba sobre la superficie en forma de hemi-micelas, se obtuvieron complejos totalmente estables independientemente de la fuerza iónica del medio, al igual que ocurrió con la formulación mezcla PLGA-PF68. Este resultado sirvió para ratificar que la adsorción de hemi-micelas en lugar de mono-capas de poloxámero originó la estabilización estérica de los complejos y por tanto de las formulaciones mezcla.

Paper III: Electrophoretic Mobility and Colloidal Stability of PLGA Particles Coated with IgG.

J. Colloids and Surface B: Biointerfaces 60, 1, 2007, 80-88.

Este trabajo se realizó con el objetivo de evaluar la posible vectorización de las nanopartículas de PLGA. Esta tarea se llevó a cabo mediante la adsorción de un anticuerpo policlonal (IgG) sobre estas nanopartículas. Para caracterizar los complejos PLGA-IgG se estudió tanto el comportamiento electrocinético como la estabilidad de los mismos. Finalmente, para ver si la IgG mantenía su capacidad

22 III Breve Resumen y Discusión de Resultados

inmunológica tras el proceso de adsorción se midió la inmunoreactividad de las partículas. Para ello, como se comenta en profundidad en el artículo, se usó proteína C-reactiva (PCR) como agente agregante y nanopartículas de PLGA sensibilizadas con IgG-anti-PCR como nanopartículas reactivas. Para entender mejor los resultados obtenidos se realizó un estudio paralelo con nanopartículas modelo de poliestireno (PS) de características similares a las PLGA.

La adsorción de IgG en función del pH sobre las nanopartículas de PLGA y PS mostró la típica forma de campana con un máximo de adsorción a pHs cercanos al punto isoeléctrico del complejo “partícula-proteína”. Aunque cualitativamente ambas campanas de adsorción fueron idénticas, es necesario destacar que la cantidad de proteína adsorbida sobre PS fue aproximadamente tres veces superior a la adsorbida sobre PLGA. La adsorción de proteínas normalmente se encuentra controlada por las interacciones hidrófobas entre la molécula adsorbida y el sustrato. Por lo que se procedió a medir el carácter hidrófobo de ambos sustratos mediante medidas de ángulo de contacto usando la técnica ADSA-P. Estas medidas revelaron que el PS presenta un mayor carácter hidrófobo que el PLGA, lo cual justifica la mayor adsorción de la IgG sobre PS en comparación con el PLGA.

El estudio electrocinético de los complejos PLGA-IgG y PS-IgG en función del pH mostró cómo según aumentaba el recubrimiento de las nanopartículas por la proteína se pasó de un perfil típico de sistemas coloidales con grupos carboxilo superficiales al perfil teórico calculado para la IgG. La caracterización electrocinética de los complejos desveló que para altos recubrimientos de IgG ambas nanopartículas presentaban valores bajos de movilidad electroforética entre los pHs 4-9. Esto sugiere que si su estabilidad coloidal depende solamente del solapamiento sus dobles capas eléctricas los complejos presentarían una baja estabilidad en este intervalo de pHs.

Para constatar la estabilidad de estos complejos en el intervalo de pHs comentado en el párrafo anterior se midió la estabilidad de los mismos a dos recubrimientos de IgG en función de la concentración de sal del medio. Como era de esperar los complejos presentaron una baja estabilidad coloidal en el intervalo de pHs comentado anteriormente. No obstante a altas concentraciones de sal se observaron fenómenos de re-estabilización, los cuales fueron más acentuados cuanto mayor era el recubrimiento de la proteína, debido a que cuanto mayor era el recubrimiento más hidrófila se hacía la superficie de las nanopartículas.

Finalmente, para saber si el proceso de adsorción había desnaturalizado las moléculas de IgG se estudió la inmunoreactividad de los anticuerpos adsorbidos tanto en nanopartículas de PLGA como de PS. Ambos sistemas mostraron una buena respuesta inmune. Los resultados del inmonoensayo se usaron para calcular el porcentaje de IgG activa sobre las nanopartículas, el cual fue mayor para el caso de las nanopartículas de PLGA. Este resultado, que puede ser contradictorio en un principio, puede deberse a dos posibles razones. i) El mayor carácter hidrófobo del PS en comparación con el PLGA puede aumentar la adsorción de la proteína, pero a

III Breve Resumen y Discusión de Resultados 23

su vez también puede producir una mayor desnaturalización de la misma. ii) Al haber una mayor cantidad de proteína adsorbida sobre el PS pueden darse impedimentos estéricos entre las moléculas de anticuerpo vecinas disminuyendo la actividad de las mismas. No obstante ambos complejos mostraron una buena respuesta inmune, por lo que si se usase un anticuerpo monoclonal podría disminuirse la cantidad de anticuerpo adsorbido manteniendo una buena respuesta inmune y dejando a su vez espacio en la superficie de las nanopartículas para la adsorción de un surfactante, como por ejemplo Pluronic® F68 para aumentar la estabilidad de las nanopartículas. De esta forma podrían obtenerse sistemas coloidales estables y con una buena respuesta inmune capaz de vectorizar cualquier fármaco encapsulado en estas partículas.

Paper IV: Nanoparticles made from Novel Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers. Part 1.

Borrador para ser enviado a Biomaterials.

En este estudio se analizó el efecto del carácter hidrófilo-hidrófobo del polímero en la formación y posterior comportamiento de las nanopartículas. Para esto se usaron dos polímeros anfifílicos, P1 y P2, presentando P2 un carácter más hidrófobo. Las nanopartículas fueron preparadas por una emulsión simple de aceite en agua, disolviendo el correspondiente polímero en la fase orgánica y añadiendo un surfactante en la fase acuosa. Las nanopartículas formuladas con ambos polímeros, nanopartículas P1 y P2, presentaron un tamaño similar, siendo un poco mayores las preparadas con el polímero más hidrófobo. La medida de la estabilidad de la emulsión a partir de la cual se forman las nanopartículas mostró que en función del carácter más o menos hidrófobo de ambos polímeros se obtenía una emulsión más o menos estable. Además se pudo observar de forma clara el efecto del surfactante en la formación de las partículas, obteniéndose emulsiones prácticamente estables cuando éste se añadía a la emulsión.

El análisis de la estabilidad coloidal de las nanopartículas mostró que en un principio estos sistemas no podrían administrarse por rutas con una alta fuerza iónica. Por otro lado, a altas fuerzas iónicas se observaron fenómenos de re-estabilización. Estos resultados eran lógicos considerando que las nanopartículas presentaban en su superficie un surfactante con restos hidrófilos. Lo que fue más sorprendente es que las nanopartículas formuladas con el polímero más hidrófobo mostraron fenómenos de re-estabilización más pronunciados, esto es, presentaban una superficie más hidrófila. Como se ha comentado anteriormente (Paper III) los fenómenos de adsorción están controlados en gran medida por las interacciones hidrófobas entre la molécula adsorbida y el sustrato. Las gotas de la emulsión preparada con el polímero P2 tendrán una superficie más hidrófoba que aquella preparada con el polímero P1, por lo que es lógico pensar que se adsorberá una mayor cantidad de surfactante. Una vez adsorbido el surfactante a través de sus restos hidrófobos desplegará hacia la fase acuosa sus segmentos hidrófilos, lo cual

24 III Breve Resumen y Discusión de Resultados

explica por qué las nanopartículas con una matriz más hidrófoba tendrán una superficie más hidrófila.

Debido a que el polímero a partir del cual se sintetizó nuestro polímero anfifílico presentaba fenómenos de hinchado también decidimos investigar si estos sistemas conservaban esta propiedad. Para ello se midió el tamaño de las nanopartículas en función de la concentración de NaCl del medio. El tamaño de las nanopartículas mostró una clara dependencia con la concentración de sal del medio, reduciéndose su tamaño según aumentaba la fuerza iónica del medio. No obstante, pese a presentar una dependencia tamaño-fuerza iónica ésta fue de menor intensidad que en el caso de los microgeles puros. Este resultado se explica en base a la modificación del polímero, la cual redujo sus propiedades de microgel.

Por último se estudió la estabilidad de estos sistemas almacenándolos a 4 y 25ºC. Ambas nanopartículas fueron estables por un período de un mes almacenadas a 4ºC. Sin embargo, cuando se almacenaron a 25ºC solamente las nanopartículas P2 fueron estables durante este período de tiempo. Este resultado puede atribuirse a la mayor cantidad de surfactante presente en esta formulación en comparación con las nanopartículas P1.

Paper V: Characterization of Core-Shell Lipid-Chitosan and Lipid-Poloxamer Nanocapsules.

Enviado al J. of Biomat. Sci. Polym. Ed.

Este trabajo constituye la primera incursión en el estudio de nanotransportadores de fármacos diferentes a las nanopartículas poliméricas. En este caso se caracterizaron nanocápsulas lipídicas las cuales son sistemas vesiculares compuestos por un núcleo oleoso formado por una mezcla de triglicéridos y lecitina recubierto por una capa de quitosano y poloxámero (Pluronic® F68). Con el objetivo de analizar el papel de cada componente del sistema en el comportamiento del mismo se estudiaron cuatro sistemas diferentes. El primero estaba compuesto únicamente por el núcleo oleoso y la lecitina (LC), el segundo fue recubierto con poloxámero (PX), el tercero con quitosano (CS) y el último se recubrió con quitosano y poloxámero (CS+PX).

Tanto la caracterización electrocinética como el estudio de la estabilidad coloidal de los sistemas LC, PX y CS mostraron que la incorporación del poloxámero a la superficie del núcleo oleoso era mínima en comparación con la incorporación del quitosano. Este resultado puede justificarse en base a las propiedades de ambos polímeros. Para el poloxámero, debido a su naturaleza de surfactante no iónico, su adsorción se da principalmente debido a las interacciones hidrófobas, por lo que si consideramos que la superficie de los núcleos oleosos contiene a los grupos polares negativos de la lecitina, los cuales poseen carácter hidrófilo, esta adsorción se verá desfavorecida. Por otro lado, la adsorción del quitosano se ve favorecida por la atracción electrostática entre los grupos amino

III Breve Resumen y Discusión de Resultados 25

positivos presentes en su estructura y los grupos polares negativos presentes en la lecitina. Teniendo esto en cuenta es lógico pensar que en el caso del sistema CS+PX se produjo una adsorción competitiva entre el quitosano y el poloxámero sobre los núcleos oleosos estando favorecida en todo momento la adsorción del primero sobre el segundo. Tanto la caracterización electrocinética como el estudio de su estabilidad coloidal así lo confirmaron. Debido a la poca cantidad de poloxámero presente en la superficie de las nanocapsulas no fue posible observar fenómenos de estabilización estérica en las mismas. No obstante, aunque los sistemas CS y CS+PX fueron inestables a pH 7, cuando se incrementó la fuerza iónica del medio para alcanzar condiciones isotónicas similares a la de los fluidos fisiológicos estos sistemas empezaron a ser estables. Este comportamiento se puede justificar en base a las fuerzas de hidratación debido al carácter hidrófilo de la superficie de estas nanocápsulas.

Paper VI: Protein-Loaded PLGA Nanoparticles for Parenteral Administration.

Borrador para ser enviado al J. of Biophysics.

En esta parte de la tesis se estudió la encapsulación de BSA e IgG en las formulaciones mezcla PLGA-PF68 y PLGA-T904. Se analizó el efecto de variables tales como la cantidad de proteína encapsulada, su carácter hidrófilo y el tipo de surfactante presente en la formulación mezcla en las características finales de las nanopartículas, así como en su comportamiento durante la posterior liberación de las proteínas encapsuladas.

Tanto BSA como IgG fueron encapsuladas en las nanopartículas PLGA-PF68 y PLGA-T904 a tres cargas teóricas diferentes, 1, 2, 4 % (p/p) en relación a la cantidad de polímero PLGA presente en las formulaciones mezcla, mostrando todas una eficacia de encapsulación superior al 80%. La encapsulación de BSA (proteína con carga neta negativa durante la encapsulación) en la formulación PLGA-PF68 no produjo cambios en su tamaño ni polidispersidad. Sin embargo, su potencial-ζ mostró una clara dependencia con la cantidad de proteína encapsulada, aumentando la magnitud del mismo según aumentaba la cantidad de proteína encapsulada. Por otro lado, la formulación PLGA-T904 no mostró un cambio aparente en ninguna de estas tres magnitudes. Cuando se encapsuló IgG (proteína con carga neta cercana a cero durante la encapsulación) ambas formulaciones presentaron un comportamiento similar que cuando fue encapsulada la BSA. No obstante, en este caso, la formulación PLGA-PF68 con un 4% de IgG presentó un claro aumento de su tamaño y polidispersidad. La encapsulación de IgG en PLGA-PF68 provocó un descenso en valor absoluto de su potencial-ζ tal que la formulación con un 4% de IgG encapsulada mostró un valor de esta magnitud cercano a cero, lo cual pudo producir la floculación de este sistema. El cambio del potencial-ζ observado para el PLGA-PF68 y no para PLGA-T904 cuando fueron encapsuladas las ambas proteínas sugiere una mayor acumulación de la proteína en la superficie de la formulación con poloxámero. Este resultado puede justificarse en

26 III Breve Resumen y Discusión de Resultados

base a las mejores propiedades surfactantes de este compuesto en comparación con la poloxamina.

Después de su preparación ambas formulaciones fueron incubadas en tampón fosfato salino (PBS) a 37ºC durante 14 días. Tamaño, potencial-ζ y cantidad de proteína liberada fueron monitorizadas durante la incubación. En general todos los sistemas mantuvieron un tamaño y potencial-ζ constante durante toda la incubación, lo cual concuerda con los resultados obtenidos por otros autores. Sin embargo, la formulación PLGA-PF68 con un 4% de IgG presentó un claro descenso en su tamaño y polidispersidad hasta valores normales durante el primer día de incubación. Dado que no se observaron fenómenos de sedimentación, este cambio en el tamaño cuando las nanopartículas pasaron de un medio con baja fuerza iónica-medio en el cual fueron preparadas- a otro con alta (~150 mM) -medio en el cual se realizó la incubación- sugiere que las nanopartículas sufrieron un proceso de re-estabilización debido a fuerzas de hidratación.

La última parte de este trabajo se centró en el estudio de la liberación de las proteínas desde ambas formulaciones mezcla. A los resultados de liberación experimentales se les aplicó un modelo matemático desarrollado por Ritger y Peppas para ver si la liberación de las proteínas estaba controlada por la difusión de las mismas a través de la matriz de la nanopartícula o se debía a la degradación de la misma. La BSA presentó un claro efecto burst, mientras que la IgG mostró una liberación más controlada en ambas formulaciones mezcla, eliminándose por completo el efecto burst para la formulación PLGA-T904. Los resultados experimentales revelaron el importante efecto de las interacciones hidrofóbicas en la liberación de las proteínas, mientras que el modelo matemático puso de manifiesto que la liberación de ambas proteínas estaba controlada por la difusión a través de los poros de las nanopartículas y no a la degradación de la matriz de las mismas.

Paper VII: Insulin-Loaded PLGA Nanoparticles for Oral Administration: an in vitro Physico-chemical Characterization.

Aceptado para publicación en el J. of Biomed. Nanotech.

El presente trabajo se focalizó en el uso de las nanopartículas de PLGA como sistemas de administración de insulina por vía oral y puede dividirse en dos partes bien diferenciadas. Por un lado se estudió la estabilidad de las nanopartículas PLGA y las formulaciones mezcla en fluidos gástrico e intestinal simulados. Por otro lado, teniendo en cuenta la estabilidad de los sistemas en la primera parte de este trabajo, se seleccionaron las formulaciones mezcla para encapsular insulina y caracterizar su liberación tanto en fluido gástrico como intestinal simulado libre de enzimas.

En primer lugar se estudió la estabilidad de los tres sistemas coloidales en fluido gástrico e intestinal simulado. Asimismo, se midió la producción de lactato,

III Breve Resumen y Discusión de Resultados 27

producto final generado en la degradación del PLGA, para ver el posible ataque de las enzimas presentes en estos medios sobre la matriz de PLGA de las nanopartículas. En fluido gástrico no se observó la producción de lactato, lo cual fue indicativo de una nula actividad por parte enzimas presentes en este medio. No obstante la nanopartículas de PLGA sin surfactante agregaron rápidamente, debido a la cancelación de la carga superficial de las mismas. Un resultado en principio chocante fue la baja estabilidad mostrada también por la formulación PLGA-T904, la cual se debió a la interacción electrostática atractiva con la pepsinaf presente en el medio de incubación. La formulación mezcla PLGA-PF68 fue totalmente estable. Cuando los tres sistemas fueron incubados en fluido intestinal simulado se vio cómo en el medio de incubación de las nanopartículas PLGA se generó lactato y se produjo la desestabilización del sistema coloidal. Esto se debió a la degradación de este polímero por parte de la pancreatínag presente en dicho medio. En el caso de las formulaciones mezcla no se observó la producción de lactato, lo cual indica que ambos surfactantes protegieron la matriz de la nanopartícula de la actividad enzimática. Tampoco se observó ningún tipo de agregación en las formulaciones mezcla.

La segunda parte de este trabajo se centró en la encapsulación de insulina en ambas formulaciones mezcla, la proteína fue encapsulada en una relación del 1% (p/p) con respecto a la cantidad de polímero PLGA presente en las formulaciones. Esta proteína fue encapsulada con dos cargas netas eléctricas diferentes, estudiando de este modo el efecto de la carga de la proteína encapsulada en las propiedades y comportamiento de estos sistemas coloidales. Cuando la insulina fue encapsulada con carga neta positiva se obtuvieron nanopartículas con un tamaño medio superior, lo cual se puede atribuir a la interacción electrostática atractiva entre la proteína y el polímero PLGA. Por otro lado, la eficacia de encapsulación dependió principalmente del surfactante presente en la formulación. La mayor encapsulación obtenida en presencia del poloxámero puede explicarse en base a que con este surfactante se obtienen emulsiones más estables, lo cual favorece una mayor incorporación de la proteína. Una vez formulados, estos sistemas fueron incubados en fluido gástrico e intestinal simulados libres de enzimas. Durante esta incubación se monitorizó la evolución del tamaño medio, potencial-ζ y la liberación de la insulina. Al igual que se vio en el Paper VI tanto el tamaño medio como el potencial-ζ fueron constantes durante toda la incubación. Con respecto a la liberación en fluido gástrico la formulación PLGA-PF68 mostró una clara dependencia con la carga neta de la insulina encapsulada, mostrando un claro efecto burst cuando la proteína fue encapsulada con carga neta negativa y una liberación más controlada cuando fue encapsulada con carga neta positiva. Este resultado indica que cuando

f La pepsina es una proteasa, una enzima digestiva que degrada las proteínas en el estómago. Esta enzima tiene un punto isoeléctrico cercano a 1, lo que justifica su interacción con las nanopartículas PLGA-T904, ver Paper I. g La pancreatina es una mezcla de enzimas producidas por células exocrinas en el páncreas. Está compuesta por amilasas, lipasas y proteasas.

28 III Breve Resumen y Discusión de Resultados

la proteína fue encapsulada con una carga diferente a la del PLGA se produce un mejor estructuramiento de la proteína en su interior. Por otro lado, en el caso de la formulación PLGA-T904 la insulina encapsulada tanto en forma positiva como negativa mostró un perfil de liberación similar y con un efecto burst más bajo que en el caso de la formulación PLGA-PF68. Esto puede justificarse en base al mayor carácter hidrófobo de la poloxamina en comparación con el poloxámero, lo cual disminuye la formación de poros en la matriz de PLGA y reduce por tanto la liberación de la proteína encapsulada. Finalmente, los sistemas incubados en fluido intestinal mostraron todos una liberación similar, lo cual puede justificarse en base a que a pHs cercanos a 7 la degradación del PLGA es menor, enmascarándose de este modo el efecto del surfactante.

Paper VIII: Development of Novel Drug-Assembled Nanoparticles made from Amphiphilic Polymers.

Borrador para ser enviado al Journal of Controlled Release.

Para preparar las nanopartículas se usó un método de emulsión simple de aceite en agua, disolviendo el polímero, P5, en la fase orgánica. No obstante, fue imposible obtener una distribución de tamaños estrecha mediante esta técnica. En este momento se pensó que la inclusión de una molécula hidrófoba en la fase orgánica de la emulsión podría atraer a los restos hidrófobos del polímero, actuando como un linker hidrófobo. De esta forma se podría obtener un comportamiento paralelo a la gelificación iónica pero haciendo uso de interacciones hidrófobas en lugar de electrostáticas. Como linker hidrófobo se usó ácido flufenámico, una molécula ampliamente conocida como agente anti-inflamatorio, la cual posee un valor de Log P ~ 5.0h y un pKa ~ 4.0. La inclusión de esta molécula en la fase orgánica sirvió para obtener nanopartículas con una buena distribución de tamaños, las cuales se denotaron como P5.

Una vez que las nanopartículas P5 fueron formuladas con el linker hidrófobo y para constatar el efecto del mismo en la estabilidad coloidal del sistema se estudió el efecto de su liberación y la estabilidad de las mismas. Para esto se llevaron a cabo dos estudios paralelos, por un lado se incubó una alícuota de nanopartículas en condiciones sinki mientras que otra alícuota se incubó en una disolución saturada en ácido flufenámico, lo cual dificultaba la liberación del mismo desde las nanopartículas. Las nanopartículas inmersas en una disolución saturada de ácido flufenámico mantuvieron constantes su tamaño y polidispersidad durante toda la incubación, mientras que las incubadas en condiciones sink fueron degradándose a

h El Log P se define como el logaritmo del coeficiente de partición octanol/agua. Este parámetro es usado para indicar el carácter hidrófilo-hidrófobo de una molécula, presentando un mayor valor de Log P las moléculas hidrófobas y un menor valor las hidrófilas. i Condiciones sink son aquellas en las que la concentración de la disolución es de 5-10 veces menor que la concentración requerida para hacer una solución saturada de un compuesto.

III Breve Resumen y Discusión de Resultados 29

medida que el ácido flufenámico era liberado. Estos resultados constatan nuestra hipótesis sobre el hecho de que el ácido flufenámico actuó como un linker hidrófobo en la formación de estas nanopartículas. No obstante, es necesario llevar a cabo más pruebas para constatar el papel del ácido flufenámico en la formación del sistema coloidal. Finalmente, estas nanopartículas mostraron un perfil de liberación bastante bueno, con una liberación de la molécula encapsulada prácticamente lineal y sin efecto burst inicial.

Paper IX: Nanoparticles made from Novel Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers. Part 2.

Borrador para ser enviado a Biomaterials.

Una vez caracterizadas las nanopartículas P1 y P2 (Paper IV). El siguiente paso fue el análisis de su posible aplicación como sistemas de liberación controlada de fármacos.

Debido a que una de las primeras restricciones que presentan estos sistemas es la estabilidad cuando son almacenados por un largo período de tiempo, en primer lugar se estudió la liofilización de las ambas nanopartículas usando trehalosa o sucrosa como agente crioprotector. Para el caso de las nanopartículas formuladas con el polímero más hidrófilo, P1, se vio que era necesario añadir alrededor de un 1% (p/p) de crioprotector en relación a la cantidad de nanopartículas para obtener un sistema con unas propiedades similares al sistema sin liofilizar. Por otro lado, las nanopartículas formuladas con el polímero más hidrófobo (P2) se pudieron resuspender sin necesidad de usar crioprotector. Este resultado puede justificarse considerando los resultados de estabilidad coloidal (Paper IV), los cuales mostraban que las nanopartículas P2 presentaban una mayor cantidad de surfactante en su superficie. El surfactante empleado en la preparación de estas nanopartículas posee propiedades crioprotectoras, por lo que es lógico pensar que el sistema que presentaba una mayor cantidad de éste en superficie no necesitaba crioprotectores para su liofilización.

El siguiente paso consitió en la encapsulación de moléculas modelo en las nanopartículas P1 y P2 para analizar su potencial como sistemas liberadores de fármacos. Para ello se eligieron tres moléculas diferentes; ácido flufenámico (Log P ~ 5.0; pKa ~ 4.0), testosterona (Log P ~ 3.5) y cafeína (Log P ~ -0.1 pkb ~ 10.0), mostrando ambos sistemas una eficacia de encapsulación superior al 80% en todos los casos. Con respecto a la liberación de las moléculas los dos tipos de nanopartículas presentaron un perfil liberación claramente dependiente del carácter hidrófobo de la molécula encapsulada e independiente del estado iónico de la misma. La cafeína fue rápidamente liberada mientras que las otras dos moléculas -que presentan un mayor carácter hidrófobo- fueron liberadas de una forma más sostenida y sin apreciarse un efecto burst inicial.

30 III Breve Resumen y Discusión de Resultados

En vista de los resultados obtenidos tanto en el Paper IV como en este trabajo, se decidió analizar el potencial de estas nanopartículas como sistemas para la liberación de fármacos por vía transdérmica, para lo cual se realizaron experimentos de paso de las moléculas tanto encapsuladas como libres a través de epidermis humana. Los resultados mostraron que cuando el ácido flufenámico era encapsulado en cualquiera de las nanopartículas se daba un claro aumento del paso a través de epidermis en comparación con la aplicación de una disolución del fármaco sin encapsular. Sin embargo, no fue posible obtener este efecto en el caso de la testosterona o la cafeína.

IV Conclusiones

Las conclusiones más relevantes, en función de los bloques que componen esta Tesis Doctoral, se enumeran a continuación:

Bloque I: Caracterización de nanopartículas aplicadas como sistemas de liberación controlada de fármacos.

1) Los resultados obtenidos tanto en la caracterización electrocinética como en el estudio de la estabilidad coloidal de las nanopartículas PLGA, PLGA-PF68 y PLGA-T904 revelaron el importante papel que juegan los surfactantes a la hora de obtener sistemas estables en condiciones fisiológicas. La similitud en la estabilidad coloidal de ambas formulaciones mezcla, pese a las grandes diferencias que existen entre ambos surfactantes, sugiere que hay una gran cantidad de surfactante en su superficie.

2) Aunque las formulaciones mezcla se encontraban estabilizadas estéricamente, encontramos un mecanismo de desestabilización poco corriente cuando se las incubó en presencia de altas concentraciones de Ca2+ y aniones HPO42-. Concluyéndose que este fenómeno se debió a la formación de agregados debido a una reacción de coordinación entre los átomos oxígeno presentes tanto en el poloxámero como la poloxamina, el catión divalente y el polianión fosfato.

3) Mediante la adsorción de Pluronic® F68 sobre las nanopartículas de PLGA fue posible entender mejor los procesos que gobiernan la estabilidad coloidal de las formulaciones mezcla. El análisis de la isoterma de adsorción del poloxámero sobre las nanopartículas PLGA así como de su caracterización coloidal nos permitió constatar que a altos recubrimientos se produce un cambio estructural en la capa de surfactante situada en la superficie de la nanopartícula, el cual contribuye en gran medida a su estabilización estérica. A medida que se aumentaba el recubrimiento de las nanopartículas de PLGA se observaron diferentes mecanismos de estabilización coloidal. Para bajos o nulos recubrimiento el comportamiento de los sistemas podía explicarse en base a la teoría DLVO, para recubrimientos intermedios

32 IV Conclusiones

fue necesario considerar las fuerzas de hidratación, mientras que para altos recubrimientos, a concentraciones cercanas a la CMC del poloxámero, los complejos estaban estabilizados de forma estérica, al igual que las formulaciones mezcla.

4) La adsorción de IgG sobre nanopartículas PLGA para testar su posible vectorización mostró que los complejos PLGA-IgG obtenidos presentaban una baja estabilidad coloidal a pHs fisiológicos. Sin embargo, las moléculas de IgG adsorbidas sobre las nanopartículas de PLGA mostraron una buena respuesta inmune, por lo que, si se usase una IgG monoclonal en lugar de policlonal se podría disminuir la cantidad de proteína adsorbida, manteniendo una alta inmunoreactividad. De este forma quedaría superficie libre sobre la nanopartícula para realizar la coadsorción de un surfactante (por ejemplo Pluronic® F68) obteniendo de este modo complejos sensibilizados además de estables.

5) Con respecto a la caracterización de las nanocápsulas lipídicas tanto el comportamiento electrocinético como las medidas de estabilidad coloidal demostraron que la incorporación de poloxámero a la cubierta del núcleo oleoso de las nanocápsulas era mínima en comparación con la incorporación del quitosano. Por lo que una posible alternativa para obtener sistemas estabilizados estéricamente puede ser el uso de quitosano modificado covalentemente con cadenas de polietilenglicol.

6) También se ha visto que las fuerzas de hidratación juegan un papel importantísimo en la estabilización de las nanocápsulas recubiertas de quitosano, ya que éstas, que eran inestables en condiciones de baja fuerza iónica, se volvían totalmente estables cuando se alcanzaban fuerzas iónicas típicas de condiciones fisiológicas.

7) Por otro lado, la caracterización físico-química llevada a cabo con las nanopartículas P1 y P2, la cual incluyó la determinación de la estabilidad de la emulsión usada para formar las nanopartículas así como el análisis de su estabilidad coloidal, sirvió para demostrar que en función del carácter hidrófilo-hidrófobo del polímero se podían obtener sistemas coloidales con diferentes propiedades superficiales.

8) La baja carga superficial de las nanopartículas P1 y P2 junto con la ausencia de una estabilización estérica por parte del surfactante hicieron que la estabilidad coloidal de estas partículas fuese bastante baja, desaconsejando su uso por vías de administración que presenten una alta fuerza iónica.

IV Conclusiones 33

Bloque II: Aplicación in vitro de nanopartículas poliméricas como sistemas de liberación controlada de fármacos.

1) La encapsulación de proteínas con diferentes características físico-químicas (BSA, IgG e insulina) en las nanopartículas PLGA-PF68 y PLGA-T904 mostró que el surfactante presente en la formulación mezcla presenta un claro efecto tanto en la eficacia de encapsulación como en la distribución de las proteínas dentro de la matriz de la nanopartícula. Esto puede atribuirse al diferente carácter hidrófilo del surfactante así como a sus propiedades tensioactivas.

2) Por otro lado, las características físico-químicas de las proteínas encapsuladas mostraron un claro efecto en su liberación desde las formulaciones mezcla, siendo clave el carácter hidrófobo de la proteína encapsulada.

3) Por último, cuando las nanopartículas PLGA, PLGA-PF68 y PLGA-T904 fueron incubadas en fluido gástrico simulado, únicamente el sistema PLGA-PF68 fue estable, mientras que los otros dos agregaron debido a interacciones electrostáticas entre las nanopartículas entre sí o con las enzimas presentes en el medio. Cuando estos sistemas se incubaron en fluido intestinal simulado se vio cómo las nanopartículas PLGA eran degradadas por las enzimas presentes en el medio. En el caso de las formulaciones mezcla esta degradación fue mínima, debido al efecto protector de los surfactantes presentes en la superficie de las mismas, los cuales impidieron la acción de las enzimas sobre las nanopartículas.

4) En relación a las nanopartículas P5 se puede destacar que gracias al uso de estos polímeros anfifílicos ha sido posible desarrollar un novedoso sistema de preparación de nanopartículas las cuales se forman debido a la presencia de un fármaco hidrófobo, en este caso ácido flufenámico, en su interior, que actúa de linker hidrófobo y se degradan cuando éste es liberado. Por otro lado, este sistema coloidal presenta una alta eficacia de encapsulación y un perfil de liberación prácticamente lineal para este fármaco.

5) Los resultados de la liofilización de las nanopartículas P1 y P2 confirmaron los resultados obtenidos durante la caracterización de las mismas, presentando las nanopartículas P2 una mayor cantidad de surfactante en su superficie que las P1. Por otro lado ambas partículas mostraron una alta eficacia de encapsulación para las tres moléculas modelo testadas (ácido flufenámico, testosterona y cafeína). Los resultados experimentales mostraron que la liberación de estas moléculas estuvo claramente influenciada por el carácter hidrófobo de las mismas y de la matriz de las nanopartículas. De esta forma, tanto la testosterona como el ácido flufenámico presentaron una liberación

34 IV Conclusiones

sostenida sin efecto burst inicial, mientras que la cafeína fue rápidamente liberada desde las nanopartículas.

6) Finalmente, los estudios llevados a cabo con piel humana mostraron que cuando el ácido flufenámico fue encapsulado tanto en P1 como en P2 se producía un claro aumento del paso de este principio activo a través de piel en comparación con la aplicación de la molécula libre.

Referencias

(1) Barratt G. Colloidal drug carriers: achievements and perspectives. Cell Mol Life Sci 2003; 60(1):21-37.

(2) Hervella P, Lozano MV, García-Fuentes M, Alonso MJ. Nanomedicine: New Challenges and Opportunities in Cancer Therapy. Journal of Biomedical Technology. In press.

(3) Greiling W. Paul Ehrlich. Econ Verlag. Duseldorf, Germany: 1954: 48.

(4) Kreuter J. Nanoparticles - a historical perspective. International Journal of Pharmaceutics 2007; 331(1):1-10.

(5) Farokhzad OC, Langer R. Nanomedicine: developing smarter therapeutic and diagnostic modalities. Adv Drug Deliv Rev 2006; 58(14):1456-1459.

(6) Soppimath KS, Aminabhavi TM, Kulkarni AR, Rudzinski WE. Biodegradable polymeric nanoparticles as drug delivery devices. J Control Release 2001; 70(1-2):1-20.

(7) Wagner V, Dullaart A, Bock AK, Zweck A. The emerging nanomedicine landscape. Nat Biotechnol 2006; 24(10):1211-1217.

(8) Mundargi RC, Babu VR, Rangaswamy V, Patel P, Aminabhavi TM. Nano/micro technologies for delivering macromolecular therapeutics using poly(D,L-lactide-co-glycolide) and its derivatives. J Control Release 2008; 125(3):193-209.

(9) Dorland's Medical Dictionary for healtcare consumers Elsevier, 2007.

(10) Alonso MJ. Nanoparticulate drug carrier technology. In: Cohen S, Berstein H, editors. Microparticulate Systems for the Delivery of Proteins and Vaccines. New York: Marcel Dekker Inc., 1996: 203-242.

(11) Allemann E, Gurny R, Doelker E. Preparation of aqueous polymeric nanodispersions by a reversible salting-out process: influence of process parameters on particle size. International Journal of Pharmaceutics 1992; 87:247-253.

(12) Fessi H, Puisieux F, Devissaguet JP, Ammouri N, Benita S. Nanocapsules formation by interfacial polymer deposition following solvent displacement. International Journal of Pharmaceutics 1989; 55:R1-R4.

(13) Csaba N, Gonzalez L, Sanchez A, Alonso MJ. Design and characterisation of new nanoparticulate polymer blends for drug delivery. J Biomater Sci Polym Ed 2004; 15(9):1137-1151.

(14) Rajeev AJ. The manufacturing techniques of various drug loaded biodegradable poly(lactide-co-glycolide) (PLGA) devices. Biomaterials 2000; 21:2475-2490.

(15) Fu K, Pack DW, Klibanov AM, Langer R. Visual evidence of acidic environment within degrading poly(lactic-co-glycolic acid) (PLGA) microspheres. Pharm Res 2000; 17(1):100-106.

(16) Luengo J, Weiss B, Schneider M, Ehlers A, Stracke F, Konig K, Kostka KH, Lehr CM, Schaefer UF. Influence of nanoencapsulation on human skin transport of flufenamic acid. Skin Pharmacol Physiol 2006; 19(4):190-197.

(17) Jiang HL, Jin JF, Hu YQ, Zhu KJ. Improvement of protein loading and modulation of protein release from poly(lactide-co-glycolide) microspheres by complexation of proteins with polyanions. J Microencapsul 2004; 21(6):615-624.

36 Referencias

(18) Zhu G, Mallery SR, Schwendeman SP. Stabilization of proteins encapsulated in injectable poly (lactide- co-glycolide). Nat Biotechnol 2000; 18(1):52-57.

(19) Sanchez A, Tobio M, Gonzalez L, Fabra A, Alonso MJ. Biodegradable micro- and nanoparticles as long-term delivery vehicles for interferon-alpha. Eur J Pharm Sci 2003; 18(3-4):221-229.

(20) Sanchez A, Villamayor B, Guo Y, McIver J, Alonso MJ. Formulation strategies for the stabilization of tetanus toxoid in poly(lactide-co-glycolide) microspheres. Int J Pharm 1999; 185(2):255-266.

(21) Tobio M, Schwendeman SP, Guo Y, McIver J, Langer R, Alonso MJ. Improved immunogenicity of a core-coated tetanus toxoid delivery system. Vaccine 1999; 18(7-8):618-622.

(22) Perez C, Castellanos IJ, Costantino HR, Al-azaam W, Griebenow K. Recent trends in stabilizing protein structure upon encapsulation and release from biodegradable polymers. J of Pharmacy and Pharmacology 2002; 54:301-313.

(23) Kiss E, Bertoti I, Vargha-Butler EI. XPS and wettability characterization of modified poly(lactic acid) and poly(lactic/glycolic acid) films. J Colloid Interface Sci 2002; 245(1):91-98.

(24) Kiss E, TaKács MG, Bertoti I, Vargha-Butler EI. Surface properties of poly(lactic/glycolic acid)-pluronic® blend films. Polymers for Advanced Technologies 2003;839-846.

(25) Moghimi SM, Hunter AC. Poloxamers and poloxamines in nanoparticle engineering and experimental medicine. Trends Biotechnol 2000; 18(10):412-420.

(26) Yeh MK, Jenkins PG, Davis SS, Coombes AGA. Improving the Delivery Capacity of Microparticle Systems Using Blends of Poly(Dl-Lactide Co-Glycolide) and Poly(Ethylene Glycol). J Control Release 1995; 37(1-2):1-9.

(27) Yeh MK, Davis SS, Coombes AGA. Improving protein delivery from microparticles using blends of poly(DL lactide co-glycolide) and poly(ethylene oxide)-poly(propylene oxide) copolymers. Pharmaceutical Research 1996; 13(11):1693-1698.

(28) Akagi T, Baba M, Akashi M. Preparation of nanoparticles by the self-organization of polymers consisting of hydrophobic and hydrophilic segments: Potential applications. Polymer 2007; 48:6729-6747.

(29) Rodrigues JS, Santos-Magalhaes NS, Coelho LCBB, Couvreur P, Ponchel G, Gref R. Novel core (polyester)-shell(polysaccharide) nanoparticles: protein loading and surface modification with lectins. J Control Release 2003; 92(1-2):103-112.

(30) Torchilin VP. Micellar nanocarriers: pharmaceutical perspectives. Pharm Res 2007; 24(1):1-16.

(31) Gref R, Luck M, Quellec P, Marchand M, Dellacherie E, Harnisch S, Blunk T, Muller RH. 'Stealth' corona-core nanoparticles surface modified by polyethylene glycol (PEG): influences of the corona (PEG chain length and surface density) and of the core composition on phagocytic uptake and plasma protein adsorption. Colloids Surf B Biointerfaces 2000; 18(3-4):301-313.

(32) Vila A, Gill H, McCallion O, Alonso MJ. Transport of PLA-PEG particles across the nasal mucosa: effect of particle size and PEG coating density. J Control Release 2004; 98(2):231-244.

Referencias 37

(33) Bazile D, Prud'homme C, Bassoullet MT, Marlard M, Spenlehauer G, Veillard M. Stealth Me.PEG-PLA nanoparticles avoid uptake by the mononuclear phagocytes system. J Pharm Sci 1995; 84(4):493-498.

(34) Kwon IC, Kim Y-H, Jeong SY. Physicochemical Characteristics of Self-Aggregates of Hydrophobically Modified Chitosans. Langmuir 1998; 14:2329-2332.

(35) Lemarchand C, Couvreur P, Besnard M, Costantini D, Gref R. Novel polyester-polysaccharide nanoparticles. Pharmaceutical Research 2003; 20(8):1284-1292.

(36) Osterberg E, Bergstrom K, Holmberg K, Schuman TP, Riggs JA, Burns NL, Vanalstine JM, Harris JM. Protein-Rejecting Ability of Surface-Bound Dextran in End-on and Side-on Configurations - Comparison to Peg. Journal of Biomedical Materials Research 1995; 29(6):741-747.

(37) Service RF. How far can we push chemical self-assembly? Science 2005; 309(5731):95.

(38) Couvreur P, Barratt G, Fattal E, Legrand P, Vauthier C. Nanocapsule technology: a review. Crit Rev Ther Drug Carrier Syst 2002; 19(2):99-134.

(39) Rübe A. Development and Physico-Chemical Characterization of Nanocapsules. Halle, Germany, 2006.

(40) Anton N, Benoit JP, Saulnier P. Design and production of nanoparticles formulated from nano-emulsion templates-a review. J Control Release 2008; 128(3):185-199.

(41) Legrand P, Barratt G, Mosqueira VC, Fessi H, Devissaguet JP. Polymeric nanocapsules as drug delivery systems. A review. S T P Pharm Sci 1999; 9:411-418.

(42) Aboubakar M, Puisieux F, Couvreur P, Deyme M, Vauthier C. Study of the mechanism of insulin encapsulation in poly(isobutylcyanoacrylate) nanocapsules obtained by interfacial polymerization. J Biomed Mater Res 1997; 47(4):568-576.

(43) Quintanar-Guerrero D, Allemann E, Doelker E, Fessi H. Preparation and characterization of nanocapsules from preformed polymers by a new process based on emulsification-diffusion technique. Pharm Res 1998; 15(7):1056-1062.

(44) Prego C, Garcia M, Torres D, Alonso MJ. Transmucosal macromolecular drug delivery. J Control Release 2005; 101(1-3):151-162.

(45) Lehr CM, Bouwstra JA, Schacht EH, Junginger HE. In vitro evaluation of mucoadhesive properties of chitosan and some others natural polymers. International Journal of Pharmaceutics 1992; 78:43-48.

(46) Artursson P, Lindmark T, Davis SS, Illum L. Effect of chitosan on the permeability of monolayers of intestinal epithelial cells (Caco-2). Pharm Res 1994; 11(9):1358-1361.

(47) Lemarchand C, Gref R, Couvreur P. Polysaccharide-decorated nanoparticles. European Journal of Pharmaceutics and Biopharmaceutics 2004; 58(2):327-341.

(48) Prego C, Torres D, Fernandez-Megia E, Novoa-Carballal R, Quinoa E, Alonso MJ. Chitosan-PEG nanocapsules as new carriers for oral peptide delivery. Effect of chitosan pegylation degree. J Control Release 2006; 111(3):299-308.

(49) Mosqueira VC, Legrand P, Morgat JL, Vert M, Mysiakine E, Gref R, Devissaguet JP, Barratt G. Biodistribution of long-circulating PEG-grafted nanocapsules in mice: effects of PEG chain length and density. Pharm Res 2001; 18(10):1411-1419.

38 Referencias

(50) Lozano MV, Torrecilla D, Torres D, Vidal A, Dominguez F, Alonso MJ. Highly Efficient System To Deliver Taxanes into Tumor Cells: Docetaxel-Loaded Chitosan Oligomer Colloidal Carriers. Biomacromolecules 2008.

(51) Hamaker HC. Recl Trav Chim Pays-Bas 1936; 55:1015.

(52) de Boer JH. Trans Faraday Soc 1936; 32:21.

(53) Derjaguin BV, Landau L. Theory of the stability of strongly charged lyophobic sols and of the adhesion of strongly charged particles in solution of electrolytes. Acta PhysicoChem 1941; 14:633.

(54) Verwey EJW, Overbeek JThG. Long-distance forces acting between colloidal particles. Transactions of the Faraday Society 1946; 42B:117-123.

(55) Verwey EJW, Overbeek JThG. Theory of the Stability of Lyophilic Colloids. elsevier, 1948: 108.

(56) Overbeek JThG. In: H.R.Kruyt, editor. Colloid Science I. Elsevier, Amsterdam, 1952: 263.

(57) Fowkes FM. In: Mittal, K.L., editors. Physicochemical. aspects of polymer surfaces 2. Plenum Press, New York, 2008: 583.

(58) Hamaker HC. Rec Trav Chim 1937; 56:727.

(59) London F. Z Phys 1934; 89:736.

(60) Matijevic E, Mathai KG, Ottewill RH, Kerker M. Detection of metal ion hydrolysis by coagulation. III. Aluminum. J Phys Chem 1961; 65(5):826-830.

(61) Hidalgo-Alvarez R. On the conversion of experimental electrokinetic data into double layer characteristics in solid-liquid interfaces. Adv Colloid Interface Sci 1991; 34:217-341.

(62) Hidalgo-Alvarez R, Martín A, Fernández A, Bastos D, Martinez F, de las Nieves FJ. Electrokinetic properties, colloidal stability and aggregation kinetics of polymer colloids. Adv Colloid Interface Sci 1996; 67:1-204.

(63) Fuchs N. Z Phys 1934; 89:736.

(64) Overbeek JThG. Strong and weak points in the interpretation of colloid stability. Advances in Colloid and Interface Science 1982; 16(1):17-30.

(65) Romero-Cano MS. Estudio de la estabilización electrostérica de partículas de Látex funcionalizadas después de la adsorción de Tritón X-100. University of Granada, 1998.

(66) Vincent B, Edwards J, Emmett S, Jones A. Depletion flocculation in dispersions of sterically-stabilised particles ("soft spheres"). Colloids and Surfaces 1986; 18(2-4):261-281.

(67) Derjaguin BV, Churaev NV. The current state of the theory of long-range surface forces. Colloids and Surfaces 1989; 41:223-237.

(68) Israelachvili JN, Pashley RM. Measurement of the hydrophobic interaction between two hydrophobic surfaces in aqueous electrolyte solutions. J Colloid Interface Sci 1984; 98(2):500-514.

(69) Pashley RM. DLVO and hydration forces between mica surfaces in Li+, Na+, K+, and Cs+ electrolyte solutions: A correlation of double-layer and hydration forces with surface cation exchange properties. J Colloid Interface Sci 1981; 83(2):531-546.

(70) Thompson DW, Collins IR. Electrolyte-Induced Aggregation of Gold Particles on Solid Surfaces. J Colloid Interface Sci 1994; 163(2):347-354.

Referencias 39

(71) Craig VSJ. An historical review of surface force measurement techniques. Colloids Surf A: Physicochemical and Engineering Aspects 1997; 129:75-93.

(72) Pashley RM. Hydration Forces Between Mica Surfaces in Electrolyte-Solutions. Advances in Colloid and Interface Science 1982; 16(JUL):57-62.

(73) Israelachvili JN, Adams GE. Measurement of forces between two mica surfaces in aqueous electrolyte solutions in the range 0–100 nm. J Chem Soc , Faraday Trans 1 1978; 74:975-1001.

(74) Salou M, Siffert B, Jada A. Study of the stability of bitumen emulsions by application of DLVO theory. Colloids Surf A: Physicochemical and Engineering Aspects 1998; 148(1):9-16.

(75) Israelachvili JN. Intermolecular and Surface Force. London: Academic Press, 1992.

(76) Basu S, Sharma MM. Effect of Dielectric Saturation on Disjoining Pressure in Thin Films of Aqueous Electrolytes. J Colloid Interface Sci 1994; 165(2):355-366.

(77) Derjaguin BV. Kolloid-Z 1934; 69:155.

(78) Molina-Bolívar JA. Mecanismos de estabilidad coloidal: Teoría y aplicación a inmunoensayos. University of Granada, 1999.

BLOQUE I:

Caracterización de Nanopartículas Aplicadas como Sistemas de Liberación Controlada de Fármacos

Stability and Physicochemical Characteristics of PLGA, PLGA:poloxamer and PLGA:poloxamine Blend

Nanoparticles. A Comparative Study

M. J. Santander-Ortegaa, N. Csabab, M.J. Alonsob, J.L. Ortega-Vinuesaa, and D. Bastos-Gonzáleza

a Biocolloid and Fluid Physics Group, Department of Applied Physics, University of Granada, Av. Fuentenueva S/N, 18071, Granada (Spain)

b Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy, University of Santiago de Compostela, 15706, Santiago de Compostela (Spain)

44 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

ABSTRACT

The physicochemical properties of new nanoparticulate carrier systems, which have been previously used for the delivery of DNA plasmids, have been study in this work. The new nanostructures consist of a blend matrix formed by poly(lactic-coglicolic) acid (PLGA) copolymer and polyoxyethylene derivatives. Two types of blend formulations, PLGA:poloxamer and PLGA:poloxamine, and also pure PLGA nanoparticles have been analyzed and their surface properties compared. Electrophoretic mobility data reflected the differences on surface characteristics among the three formulations. PLGA nanoparticles behaved as typical system with weak acid groups on their surface. For the blend formulations, mobility data corroborated the difference between the two surfactants employed, this is, particles containing poloxamers present a more hydrophilic character than those containing poloxamines. Stability data showed that pure PLGA particles exhibit the expected behavior for lyophobic colloids. In contrast, the stability of the blend formulations is governed by a steric mechanism. At high concentrations of calcium ion in phosphate buffer, however, an anomalous stability behavior was observed. This was explained on the basis of the interaction of the polyethylene oxide chains of the surfactant and the divalent cations, Ca2+, in presence of diphosphate anion. In all the stability experiments both blend nanoparticles behaved identically, this has been ascribed to a considerable amount of surfactants on the surface of the particles.

Paper I 45

I. INTRODUCTION

Poly(D,L-lactic-co-glycolic acid) (PLGA) micro- and nano-spheres have been extensively used as biodegradable colloidal drug carriers [1,2]. However, the rapid removal of intravenously administered colloidal drug carriers by the mononuclear phagocytic system (MPS), comprised mainly of the hepatic Kupffer cells of the liver and the macrophages of the spleen, has been identified as the major obstacle to the efficient targeting of colloidal carriers to target sites [3]. Nevertheless, the recognition of the carriers by the MPS can be significantly altered if the surface of the colloidal particles is modified using polyethylene oxide (PEO)/polypropylene oxide (PPO) block copolymers of the poloxamer and poloxamine series [4]. These copolymers are well known for their safety and biocompatibility. In addition, PEO/PPO/PEO block copolymers show a wide range of hydrophilicity/hydrophobicity as a function of the PEO:PPO ratio, so that it is possible to achieve different degrees of particle hydration [5]. The poloxamer and poloxamine copolymers are bound to the nanosphere surface by the hydrophobic interaction of the PPO chains while the hydrophilic PEO chains protrude into the surrounding medium creating a steric barrier [6,7]. It has been suggested that this barrier prevents or restricts the adsorption of plasma proteins onto the particle surface decreasing recognition by liver and spleen macrophages [8,9]. Further research has indicated that it is not only a reduction in the adsorption of plasma components but also the selective uptake of certain plasma components acting as dysopsonins that can prevent recognition by macrophages [10]. Therefore, the coating of nanoparticle carriers by this kind of ABA block copolymer has become a successful strategy in recent years to develop drug-delivery systems. In addition, recent work suggests improvements in developing colloidal drug carriers when the poloxamers and poloxamines are incorporated into the nanoparticles during preparation instead of adsorbing them onto the bare PLGA particles [11]. These new nanocarries containing hydrophilic PEO derivatives have allowed the encapsulating of plasmid DNA [12]. It has been also demonstrated that the presence of this type of polymer helps neutralize the acidity generated in the course of PLGA degradation, preserving DNA structural integrity and thus its biological activity [13]. Moreover, the presence of the polyoxyethylene derivative has been found to exert a positive effect on the release characteristics of nanoparticles [12].

Csaba et al. [11], using 1H-NMR and differential scanning calorimetry (DCS), reported that poloxamers and poloxamines were effectively incorporated into the particle composition when preparing the systems by an optimized emulsification-solvent diffusion technique. However, 1H-NMR and DCS techniques cannot distinguish the distribution of the non-ionic copolymers—that is, when located mainly inside the particles, preferably at the interface or at both sites. The analysis of some physical characteristics such as stability or electrokinetic behavior of colloidal systems as proved useful in ascertaining the final surface properties of nanoparticles [14]. Hence, these colloidal characteristics for PLGA:poloxamer (Pluronic® F68) and PLGA:poloxamine (Tetronic® 904) blend nanoparticles were

46 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

studied in the present paper. In addition, we have found no comparative studies concerning the final physicochemical properties of PLGA nanoparticles or those resulting from incorporating these amphiphilic copolymers into the nanospheres during production. Thus, the aim of the present work is a comparative characterization of physical properties of the particles related to their electrophoretic mobility and colloidal stability.

II. MATERIALS AND METHODS

2.1 Materials

The polymer poly(D,L-lactic acid/glycolic acid) 50:50 (PLGA) was purchased from Boehringer-Ingelheim under the commercial name of Resomer® RG 503. The poloxamer Pluronic® F68 was from Sigma Aldrich. The poloxamine

Tetronic® 904 was kindly donated by BASCOM Belgium. Fig. 1 shows the chemical structure and main characteristics of both surfactants. According to the HLB values, poloxamer F68 presents a more hydrophilic character than does poloxamine 904. All other solvents and chemicals used were of the highest grade commercially available. Buffered solutions presented a constant ionic strength of around 0.002 M.

CH3

HO(CH2CH2O)n(CHCH2O)m(CH2CH2O)nH

PEO PPO PEO (PEO)n(PPO)m

NCH2CH2N

(PPO)m(PEO)n

(PEO)n(PPO)m (PPO)m(PEO)n

Pluronic® F68 2*75 EO+30*PO HLB=29

Tetronic® 904 4*15 EO+4*17PO HLB=14,5

(a) (b)

Figure 1. Chemical structures of (a) poloxamer and (b) poloxamine. PEO = polyethylene oxide unit; PPO = propilene oxide unit. HLB = hydrophilia-lipophilia balance.

2.2 Methods

2.2.1 Preparation of PLGA nanoparticles

The PLGA nanoparticles were prepared by a modified emulsion-solvent diffusion technique. First, 50 mg of PLGA were dissolved in 2 ml of dichloromethane and this organic solution was mixed for 30 s with 0.2 ml of pure water by vortex (2400 min-1, Heidolph). The resulting emulsion was poured under moderate magnetic stirring into a larger polar phase (25 ml ethanol) leading to immediate polymer precipitation in the form of nanoparticles. This sample was diluted with 25 ml MilliQ water and stirred for 10 min more. After solvent

Paper I 47

evaporation under vacuum at 30oC (Rotavapor Büchi R-114, Flawil) nanoparticles were collected and dispersed in an aqueous medium.

Nanoparticles with PLGA:poloxamer and PLGA:poloxamine ratios of 50:50 were also prepared by a modified emulsion-solvent diffusion technique in a way similar to that explained above. Details of the synthesis are described elsewhere [11,12].

2.2.2 Electrophoretic mobility

The electrophoretic mobility was measured with a Zeta-Sizer IV (Malvern Instruments). The particles were diluted in buffered solutions of low ionic strength (I = 0.002 M) for 10 min just before measurement. The final particle concentration was equal to 3 x 109 particles/mL. The mobility data were taken from the average of six measurements at the stationary level in a cylindrical cell. Standard deviation was consistently lower than 5%.

2.2.3. Colloidal stability

Stability was studied as a function of salinity using NaCl and Ca2Cl as aggregating electrolytes. Particle aggregation was analyzed by photon correlation spectroscopy (PCS) using a 4700c System (Malvern Instruments). The PCS instrument had an Argon laser (λ0 = 488 nm) with perpendicular polarization and a power rating of 75 mW. After 2mL of very diluted samples were poured into a cylindrical cell, 1 mL of the saline solution at the desired ionic strength was added and rapidly mixed. The computer software analyzed the scattered-intensity auto-correlation function measured at 60º. The aggregation measurements lasted around 10 min. For information on the aggregation kinetics, the average diameter of the particles was plotted against time. The slopes of these curves (∂d/∂t) enabled determinations of the aggregation rate (k) for the different systems and hence calculations of the stability or Fuchs factor (W) defined by:

r

s

kWk

= (1)

where the rate constant kr corresponds to a rapid coagulation kinetic, and ks is the rate constant for a slow coagulation regime. The ratio of the two constants is equal to that of the ratio of the initial slopes in our coagulation experiments. Plotting the logarithm of W versus the logarithm of the salt concentration and locating that point where logW reduces to zero can easily give the critical coagulation concentration “CCC”; i.e. the minimum salt concentration needed for rapid aggregation of the colloidal system.

III. RESULTS AND DISCUSSIONS

3.1 Main Particle characteristics

Three type of systems—the pure PLGA and two blend formulations which differed in the surfactant used in their preparation (Table 1)—were used to compare

48 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

the influence of the different surfaces on the final physicochemical properties of the nanoparticles. Table 1 shows the incorporated amount of poloxamer F68 and poloxamine T904 with respect to the total quantity of polymers in the particle (data from reference 11). It can be seen that the most hydrophobic surfactant, poloxamine

Table 1. Nanoparticles theoretical polymer ratios, percentage of surfactant in the nanoparticlesa, mean particle size, shear plane position.

Surfactant Mol % Weight%

Size (nm)

Δ(nm)

PLGA PLGA:Poloxamer F68 PLGA:PoloxamineT904

50:0 50:50 50:50

--- 10.8 ±0.5 18.8 ±2.0

---- 2.8 ±0.1 4.2 ± 0.5

195±1 151±1 152±2

0.77±0.01 1.13±0.18 0.98±0.07

T904, was incorporated into PLGA matrix in a higher quantity than poloxamer F68, which presented a more hydrophilic character (see Fig.1). However, this result was not reflected by a difference in the average size of both blend nanoparticles. On the contrary, these differences appeared when the comparison was made with pure PLGA nanoparticles (Table 1). The larger size of PLGA nanoparticles can be explained by considering that the main source to achieve stability comes from the electric charge of the carboxyl groups of the PLGA, which implies that particles have to grow high enough to become colloidally stable. By contrast, in the blend formulations, the PEO molecules help to increase stability by being placed at the interface of the particles, so that, besides the charge there is a steric contribution that makes particles with smaller sizes more stable.

3.2 Electrophoretic mobility studies

The electric state of the different nanoparticles was determined from electrophoretic mobility (μe) measurements. First, μe was measured as a function of the pH of the medium. The results are shown in Fig. 2 reflecting a clear dependence of mobility as pH varied for each sample studied. In all cases the μe values increased when the pH changed to acid pH values, while more constant values were reached at basic pHs. This behavior is typical of systems with weak acid groups on their surface [15]. This weak character of the nanoparticles comes from carboxyl groups of the PLGA. The main effect when poloxamer or poloxamines are present in the particles is the reduction of μe in comparison with the pure PLGA nanospheres. This suggests that the surfactant molecules are, at least in part, located on the surface of the particles, which increases the roughness of the surface in PLGA:poloxamer and PLGA:poloxamine blend nanoparticles. As a consequence, a displacement of the shear plane of the diffuse layer is produced resulting in lower mobility values. In addition, the charge of PLGA could be partially screened by the surfactants, as they are capable of forming hydrogen bonds with carboxyl groups of the PLGA [5]. The difference between the two surfactants is also reflected in the

a Data from referente 11.

Paper I 49

curve for pH values higher than 6. Particles with poloxamer showed lower mobility values than those with poloxamine. This result can be attributed to the more hydrophilic character of the Pluronic® F68. In this case, the surfactant chains should present a more extended conformation towards the solvent causing greater displacement of the shear plane of the particles.

The effect of the ionic strength was analyzed by measuring μe as a function of the NaCl and CaCl2 concentrations. The results shown in Fig. 3 indicate that mobility of the particles declined for increasing electrolyte concentrations for both salts. These patterns are caused mainly by the double-layer compression as a result of the screening of the surface charge of the particles when the salt concentration increases [16]. This effect also depends on the charge of the ions. Thus, the reduction on mobility values was more pronounced for the double charge of the Ca2+ ions than for the monovalent Na+ ones [17]. Comparisons of the different formulations reveal lower μe values for PLGA:poloxamer and PLGA:poloxamine than for PLGA nanoparticles. Again, these differences can be attributed to the greater roughness of the blend nanostructures.

It is possible to estimate the thickness of the layer that coats the particle surface and shifts away the shear plane. Several authors [18-20] have employed the Eversole and Boardman equation to arrive at a rough estimation of the shear plane position, Δ

Δ−⎟⎠⎞

⎜⎝⎛ Ψ

=⎟⎠⎞

⎜⎝⎛ κζ

kTe

kTe

4tanhln

4tanhln 0 (2)

3 4 5 6 7 8 9

-5

-4

-3

-2

-1

0

1

μ e 108 (m

2 /Vs)

pH

Figure 2. Electrophoretic mobility (μe) as a function of pH. ( ) PLGA,( ) PLGA:poloxamer F68 and ( ) PLGA: Poloxamine T904. (The error bars represent the standard deviation).

50 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

-4,0 -3,5 -3,0 -2,5 -2,0 -1,5 -1,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5

4,0

-μe 1

08 (m2 /V

s)

Log [I/M]

Figure 3. Electrophoretic mobility (μe) as a function of ionic strength. ( ) PLGA, ( ) PLGA:poloxamer F68 and ( ) PLGA: PoloxamineT904 with solid lines are for NaCl. ( ) PLGA, ( ) PLGA:poloxamer F68 and ( ) PLGA: PoloxamineT904 with dotted lines forCaCl2. (The error bars represent the standard deviation).

In this equation, e is the elementary electric charge, ζ the experimental zeta potential, k the Boltzmann constant, T the temperature, ψo the surface potential, and κ the Debye parameter or the reciprocal electric double-layer thickness, which depends on the electrolyte concentration of the medium. The plot of ln(tanh(eζ/4kT)) versus κ should give a straight line with slope Δ. From the μe data with NaCl at 25ºC (Fig. 3), transformed into ζ-potential (given also by the zetasizer device), we calculated the Δ values for each system. To obtain a good linear fit (r > 0.9) only data in a concentration range from 10-3 M to 10-1 M were taken.

Table 1 presents the results. The sequence varied from lower to higher values of Δ as follows: PLGA < PLGA:Poloxamine < PLGA:Poloxamer. This order coincides with that expected taking into account the hydrophobicity of the different polymers. Pure PLGA nanoparticles showed the smoothest surface while particles with poloxamer surfactant, which possess the largest hydrophilic chains, had a rougher surface. We have found no similar experiments in the literature with which we could compare our Δ data. Only hydrodynamic adsorption-layer thicknesses, determined with photon correlation spectroscopy, are reported for different pluronics adsorbed onto latex particles [6,21].

3.3 Stability results

The stability of the samples was studied by analyzing the evolution of the diameter of the particles as a function of time at different salt concentrations. Fig. 4a-d displays the results for PLGA nanoparticles at pH 4 and 7 in presence of NaCl and CaCl2. In all curves the slopes became steeper with the electrolyte concentration to a maximum value, when the system aggregated. From these

Paper I 51

curves it is possible to calculate the stability factor and hence the CCC of the nanoparticles. Fig. 5 shows the logarithm of the stability factor, logW, versus salt concentration for NaCl and CaCl2. Among the results, it is noteworthy that CCC values at pH 7 were higher than that at pH 4 for both salts (CCCNaCl-pH7 = 270 mM, CCCNaCl-pH4 = 90 mM, CCCCaCl2-pH7 = 40 mM and CCCCaCl2-pH4 = 20 mM). This agrees with the weak acid character of the carboxyl groups and with the electrokinetic behavior found in Fig. 3. Second, CCC values with CaCl2 were lower than those with NaCl. This was due to the greater screening capacity of Ca2+ ions, which significantly reduce the repulsive potential among the particles, favoring the aggregation of the system. This effect was also evident from mobility data shown in Fig. 3.

Similar experiments were performed with the blend formulations. However, no aggregation of the particles was detected even at concentrations up to 4 M for NaCl. Therefore, a strong contribution of a steric stabilization mechanism is noted when poloxamer or poloxamine are present in the nanoparticles. On the

0 1 2 3 4 5 6 7 8

100

200

300

400

500

600

700

800

900

1000

1100

1200d/

nm

t/min

30 mM 66 mM 75 mM 90mM 100mM 2000 mM

0 1 2 3 4 5 6 7 8

100

200

300

400

500

600

700

800

900

1000

1100

1200 (b)(a) 150 mM 160 mM 170 mM 180 mM 190 mM 200 mM 2000 mM

d/nm

t/min

0 1 2 3 4 5 6 7 8

200

400

600

800

1000

1200

d/nm

t/min

2 mM10mM15mM20mM30mM40mM60mM

0 1 2 3 4 5 6 7 8

200

400

600

800

1000

1200 (c) (d) 2 mM 8 mM 10 mM 15 mM20 mM 30 mM 40 mM 100 mM

d/nm

t/min

500 mM 1000 mM

Figure 4. Aggregation kinetics of PLGA nanoparticles in: (a) NaCl at pH 4.1; (b) NaCl at pH7.0; (c) CaCl2 at pH 4.1 and (d) CaCl2 at pH 7.0. (In all cases the errors were smaller than thesymbol size).

52 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

100 1000

1

10

100

other hand, the experiments performed with CaCl2 indicated a different behavior in the stabilization of these systems. The blend particles remained stable up to 500mM; however, at higher salt concentrations small aggregates formed. Fig. 6 shows the

results at different CaCl2 concentrations. As can be observed by comparison with the Fig. 4a-d results, the aggregation rate of the system was very fast causing higher aggregates as the salt concentration increased. Surprisingly, the aggregate size

0 1 2 3 4 5 6 7 80

100

200

300

400

500

600

700

800

900

1000

d (n

m)

Time (min)

Figure 6. Average hydrodynamic diameter of aggregates, d, of PLGA:Poloxamer F68 nanoparticles as a function of CaCl2 concentration at 1 M ( ), 2 M ( ) and ( ) 3 M. (In all cases the errors were smaller than the symbol size).

W

NaCl concentration (mM)

(a)

10 100 1000

1

10

100

(b)

W

CaCl2 concentration (mM)

Figure 5. Log stability factor, W, as a function of log salt concentration of PLGAnanoparticles at pH 4 ( ) and pH 7 ( ): (a) in NaCl; (b) in CaCl2. (n = 1)

Paper I 53

remained constant over time for all the concentrations measured. These experiments were repeatedly confirmed for both blend formulations, and thus an unusual destabilization mechanism was deemed to have taken place in this case.

To understand the reasons of this behavior, we performed a different set of experiments. As the measured samples were buffered, we first studied the influence of the nature of the buffer. Fig. 7a shows that, when average size of the particles was measured as a function of the pH of the medium in absence of Ca2+, the diameters remained constant as the pH rose from 3 to 9 for the three samples. However, the average size of the aggregates depended on the pH of the medium when Ca2+ was present, Fig. 7b, the strongest effect occurring at pH 6 and 7. Therefore, it was concluded that this result must have been related to the CaCl2 added to the medium. It was also shown that both blend nanoparticles formulations behaved the same and had similar sizes within the pH range studied. From the results a relationship was established between the nature of the buffer and the size of the aggregates. The buffers used in the different solutions were: For pH 4 and 5, acetic acid (CH3COOH); for pH 6 and 7, sodium phosphate (NaH2PO4); and for pH 8 and 9, boric acid (H3BO3). Given that the concentration of the buffers in the solution were lower than 5 10-3 M while the concentration of CaCl2 was equal to 3 M, a possible explanation for this behavior could be the extreme conditions of the medium. In addition, the relatively small aggregates that remained constant over time, together with the strong steric contribution to stability observed in the blend nanoparticles, suggest that the system could have been aggregated by a mechanism of flocculation rather than coagulation [22]. The difference between the two terms refers to the irreversibility (coagulation) or reversibility (flocculation) of the aggregation process. If flocculation takes place between the particles, the aggregates can be broken down, for example, substantially diminishing the salt

Figure 7a. Average hydrodynamic diameterof aggregates, d, as a function of buffered solutions at various pH. ( )PLGA, ( ) PLGA:Poloxamer F68 and ( ) PLGA:Poloxamine T904. (The error bars represent the standard deviation).

3 4 5 6 7 8 9

200

300

400

500

600

700

800

d(nm

)

pH

d0

(b)

3 4 5 6 7 8 9 100

50

100

150

200

250

300

350

400d

(nm

)

pH

(a)

Figure 7b. Average hydrodynamic diameter of aggregates, d, as a function of buffered solutions at various pH in presence of CaCl2at 3 M. ( ) PLGA:Poloxamer F68 and ( ) PLGA:Poloxamine T904. (The error bars represent the standard deviation).

54 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

concentration of the medium. This phenomenon is known as repeptization. This supposition was tested by repeptization experiments, in which the system was aggregated and then the particles were diluted to a salt concentration at which the system stabilized. Fig. 8a presents data for time dependence of particle diameter before (3 M) and after (0.5 M) dilution for PLGA:poloxamer in presence of CaCl2 at different pH values. The results confirm that the aggregation process of the blend formulations is reversible, either completely for pH 5, 8 and 9 (as original diameters of the particles are reached after dilution), or partially for pH 7 and 8. Similar results were found for PLGA:poloxamine formulation. By contrast, PLGA nanoparticles showed no repeptization process, as reflected in Fig. 8b. In this case, the experiments were performed with NaCl and CaCl2 at pH 7. For both salts the diameter of the PLGA nanoparticles remained constant after dilution. Hence, coagulation instead of flocculation took place for PLGA nanoparticles, demonstrating that they exhibit the expected behavior for lyophobic colloids [23]. It can also be concluded that the blend formulations presented a steric stabilization mechanism, and thus a major part of both surfactants must be located on the surface of the particles.

However, the results for pH 6 and 7 indicate that another aggregation mechanism must be operating in this case. At this pH, sodium phosphate was used as buffer. Thus, some relationship could exist between the nature of the buffer and the results found. From literature [24-26], we determined the affinity of divalent cations, such as Ca2+ or Ba2+, for the oxygens present in the surfactant chains and their ability to form complexes as well as precipitates in presence of multivalent voluminous anions, such as tetraphenylborate [25] or molybdophosphoric acid [26]. Although divalent phosphate is not a large anion, we assumed that a similar behavior could apply to the aggregation of our blend formulations.

Figure 8a. Average hydrodynamic diameter of aggregates, d, as a function of time of PLGA:Poloxamer F68 nanoparticles at several pH in presence of CaCl2 (A) at 3 M and (B) after dilution at 0.5 M. ( ) pH 5 ( ) pH 6, ( ) pH 7, ( ) pH 8 and ( ) pH 9. (The error bars represent the standard deviation).

0 2 4 6 8 10 12 14 16

200

300

400

500

600

700

800

900 (B)

d (n

m)

Time (min)

d0

(A)

0 2 4 6 8 10 12 14 16

200

400

600

800

1000

1200

1400

1600

1800

2000 (B)

d (n

m)

Time (min)

(A)

Figure 8b. Average hydrodynamic diameter of aggregates, d, as a function of time of PLGA nanoparticles at pH 7 in presence of ( ) NaCl and ( ) CaCl2 (A) at 0.9 M of NaCl and 0.05 M of CaCl2 (B) after dilution at 0.150 M of NaCl and 0.008 M of CaCl2. (The error bars represent the standard deviation).

Paper I 55

0,0 0,5 1,0 1,5 2,0 2,5 3,0

0

500

1000

1500

2000

2500

d (n

m)

CaCl2 concentration (M)

d0

Figure 9. Average hydrodynamic diameter, d, as a function of CaCl2 concentration at pH 7 of ( ) PLGA:Poloxamer F68, ( ) PLGA:Poloxamine T904 nanoparticles and ( ) only Poloxamer F68 surfactant (no particles). (The error bars represent the standard deviation).

To test this hypothesis we performed other experiments. The size evolution of the aggregates was analyzed, on one hand, as a function of the calcium concentration in the medium at pH 7, and, on the other hand, as a function of the concentration of one monovalent anion, NaH2PO4, and another divalent, Na2HPO4 at 3M of CaCl2. Fig. 9 shows the results obtained when CaCl2 was added to the solution. A control solution of Poloxamer F68 of a concentration similar to that employed in the blend formulation was also included in this study. The nanoparticles and the poloxamer solution presented a similar behavior up to a concentration of 0.5 M, as the increase in the CaCl2 concentration did not affect their initial state. However, from 1M all the samples underwent aggregation, being the aggregates formed by the free surfactant much larger than those of the blend formulations. This can be explained taking into account that more oxygen must be available to bind to Ca2+ ions when the surfactant is free in the solution than when it is adsorbed onto the nanoparticles. It can also be observed that the behavior of both blend formulations was very similar over the entire range studied.

On the other hand, the influence of the divalent anion on aggregate formation was also confirmed. The measurements are shown in Fig. 10. A different behavior was found between mono- and divalent phosphate anion. The increase in monovalent phosphate did not affect the particle diameter whereas divalent phosphate did. In addition, the particle aggregates formed by di-phosphate were far larger than those from the buffer solutions, because, for the buffers pH 6 and 7, it was necessary to add a certain amount of HCl, which involves a displacement of the di-phosphate anion into the mono-phosphate one. Thus, the buffered solution contained two coexisting species—mono and divalent phosphate—but, as demonstrated, only divalent phosphate anions were responsible for the aggregation

56 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

0,0 0,1 0,2 0,3 0,4 0,50

1000

2000

3000

4000

5000

6000

7000

d (n

m)

Anion (mg/ml)

d0

Figure 10. Average hydrodynamic diameter, d, as a function mono- and divalent phosphate concentration at 3 M of CaCl2 and pH 7. ( ) PLGA:Poloxamer F68 and ( ) PLGA:Poloxamine T904 in Na2HPO4; ( ) PLGA:Poloxamer F68 and ( ) PLGA:PoloxamineT904 in NaH2PO4. (The error bars represent the standard deviation).

observed. Again, no significance differences were detected between the two blend formulations.

Lastly, another repeptization experiment was designed in order to corroborate that the affinity to form these precipitates also depends on the type of divalent cation. The cation selected was Ba2+ which reportedly has a greater capacity than Ca2+ to form these complexes [24,25]. The results in Fig. 11 verify that when the system is diluted the aggregates formed remain unaltered, signifying that the bonds created between Ba2+ and the oxygen of the surfactant are stronger than those of Ca2+ and oxygen. Larger aggregates are also found when Ba2+ is used instead of Ca2+, also indicating that Ba2+ has stronger affinity for surfactant chains. This suggests that the larger size of Ba2+ may provide more effective binding sites between the surfactants placed on the surface of the particles [26].

All these experiments have clearly demonstrated that the mechanism of aggregation of the blend formulations at pH 6 and 7 at high CaCl2 concentrations is due to the presence both of divalent cations (Ca2+) and of anions (HPO42-). These help bind several nanoparticles through the hydrophilic fragment of the surfactants. Nevertheless, it is important to highlight that this unusual mechanism appears only under very specific conditions. In fact, both blend formulations are highly stable and the mechanism that explains this stability is the steric contribution due to the surfactant chains on the surface of the nanoparticles, as mentioned above.

Paper I 57

0 2 4 6 8 10 12 14 160

500

1000

1500

2000

2500

3000

3500

4000

(B)

d (n

m)

Time (min)

d0

(A)

Figure 11. Average hydrodynamic diameter of aggregates, d, as a function of time of ( ) PLGA:Poloxamer F68 and ( ) PLGA:Poloxamine T904 nanoparticles in presence of BaCl2 at pH 7, (A) at 2 M (B) after dilution at 0.150 M. (The error bars represent the standarddeviation).

With respect to the blend nanoparticles, we found no differences in their stability behavior. However, both surfactants presented different degrees of hydrophobicity with Poloxamer F68 being more hydrophilic than Poloxamine T904. Consequently, we might expect a larger amount of poloxamer on the surface of the nanoparticles than in the poloxamine case. In fact, mobility data support this possibility. Nevertheless, the stability results appear to indicate that the amount of poloxamine on the surface of the particles is high enough to prevent colloidal destabilization (by a steric mechanism) even at very high salt concentrations. This would explain the similar stability of the two blend formulations, irrespective of the PEO derivative.

IV. CONCLUSIONS

Parameters describing the electrokinetic behavior and the stability of pure PLGA nanoparticles and the two blend formulations PLGA:poloxamer and PLGA:poloxamine have been determined in this work. Electrophoretic mobility measurements showed that this technique distinguishes the differences among the three samples studied and also helps to corroborate stability results. Mobility data confirmed the difference in the degree of hydrophobicity of the two blend formulations.

With respect to stability, PLGA nanoparticles presented a more distinctive behavior than did blend formulations. While the stability of pure PLGA samples can be described as expected for lyophobic colloids, the blend formulations presented a strong steric stabilization mechanism. Also, an interesting new mechanism of aggregation was detected in the blend formulations at high Ca2+

58 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

concentrations and pH 6 and 7. This has been explained by taking into account the interaction of the polyethylene oxide chains of the surfactant and the divalent cations, Ca2+, in the presence of a diphosphate anion. In all the stability experiments, both blend nanoparticles behaved identically. This last result, together with the high steric contribution observed in both blend formulations suggests that a major part of Poloxamer F68 and Poloxamine T904 are placed on the surface of the particles.

ACKNOWLEDGEMENTS

Authors thank the financial support given by the projects MAT2003-01257 and SAF2003-08765 from the Comisión Interministerial de Ciencia y Tecnología (CICYT), Spain.

Paper I 59

References

1. R. Langer, New Methods of Drug Delivery, Science 249 (1990)1527-1533.

2. M.J. Alonso, Nanoparticulate Drug Carrier Technology, in: S. Cohen, H. Bernstein (Eds.), Microparticulate Systems for the Delivery of Proteins and Vaccines, Marcel Dekker Inc., New York, 1996, pp 203-242.

3. G. Poste, Liposome Targeting in Vivo: Problems and Oportunities, Biol Cell 47 (1983) 19-38.

4. S. M. Moghimi, A.C. Hunter, Poloxamers and Poloxamines in Nanoparticle engineering and Experimental Medicine, Trends in Biotech, 18 (2000) 412-420.

5. T. G. Park, S. Cohen, R. Langer, Poly(L-lactic acid)/Pluronic blends: characterization of phase separation behavior, degradation, and morphology and use as protein-releasing matrixes, Macromolecules 25 (1992) 116-122.

6. J. B. Kayes, D. A. Rawlins, Adsorption Characteristics of Certain Poloexyethylene Polyoxypropylene Block Co-polymers on Polysterene Latex, Colloid Polym. Sci. 257 (1979) 622-629.

7. J. T. Li, K.D. Caldwell, N. Rapoport, Surface Properties of Pluronic –Coated Polymeric Colloids, Langmuir 10 (1994) 4475-4482.

8. L. Illum, L.O. Jacobsen, R.H. Muller, E. Mark, S.S. Davis, Surface Characteristics and the Interaction of Colloidal Particles with Mouse Peritoneal Macropahges, Biomaterials 8 (1987)113-117.

9. J. S. Tan, D.E. Butterfield, C.L. Voycheck, K.D. Caldwell, J.T. Li, Surface modification of nanoparticles by PEO/PPO Block Copolymers to Minimaze Interactions with Blood Components and Prolong Blood Circulation in Rats, Biomaterials 14 (1993) 823-833.

10. S.M. Moghimi, I.S. Muir, L. Illum, S.S. Davis, V. Kolb-bachofen, Coating particle with a Block- Copolymer ( Poloxamine 908) Supresses Opsonization but Permits the Activity of Dysopsonins in the Serum, Biochim. Biophys. Acta 1179 (1993) 157-165.

11. N. Csaba, L. González, A. Sánchez, M.J. Alonso, Design and Characterisation of New Nanoparticulate Polymer Blends for Drug Delivery, J Biomater. Sci. Polym. Edn. 15 (2004) 1137-1151.

12. N. Csaba, P. Camaño, A. Sánchez, F. Domínguez, M.J. Alonso, PLGA:Poloxamer and PLGA: Poloxamine Blend Nanoparticles: New Carriers for Gene Delivery, Biomacromolecules 6 (2005) 271-278.

13. M. Tobio, S. P. Schwendeman, Y. Guo, J. McIver, R. Langer, M. J. Alonso, Improved Immunogenicity of a Core-Coated Tetanus Toxoid Delivery System, Vaccine 18, (2000) 618-622.

14. J.L. Ortega Vinuesa, D. Bastos-González, A Review of Factors Affecting the Performances of Latex Agglutination Tests, J. Biomater. Sci. Polymer Edn. 12 (2001) 379-408.

15. D. Bastos-González, J.L. Ortega-Vinuesa, F.J. de las Nieves, R. Hidalgo-Álvarez, Carboxylated Latexes for Covalent Coupling Antibodies, J. Colloid Interface Sci. 176 (1995) 232-239.

16. R. Hidalgo-Álvarez, A. Martín, A. Fernández, D. Bastos, F. Martínez, F.J. de las Nieves, Electrokinetic Properties, Colloidal Stability and Aggregation Kinetics of Polymer Colloids, Adv. Colloid Interface Sci. 67 (1996) 1-118.

60 J. Colloids and Surfaces A: Physicochem. Eng. Aspects 296, 2007, 132-140.

17. D. Bastos, F. J. de las Nieves, Effect of Electrolyte Type on the Electrokinetic Behavior of Carboxylated Polystyrene Model Colloids, Colloid Polym. Sci. 274 (1996) 1081-1088.

18. L. Nabzar, D. Duracher, A. Elaïssari, G. Chauveteau, C. Pichot, Electrokinetic Properties and Colloidal Stability of Cationic Amino-Containing N-Isopropylacrylamide-Styrene Copolymer Particles Bearing Different Shell Structures, Langmuir 14 (1998) 5062-5069.

19. D. Bastos-González , F J. de las Nieves, Colloidal Stability of Model Polymer Colloids with Different Functional Groups, Prog. Colloid Polym. Sci. 98 (1995) 1-5.

20. L. A. Rosen, D. A. Saville, The Electronic Response of Surface-Modified polymer Latexes: Effects of Grafted Water-Soluble Polymer and Heat Treatment, J. Colloid Interface Sci. 149 (1992) 542-552.

21. J. A. Baker, J.C. Berg , Investigation of the adsorption configuration of polyethylene oxide and its copolymers with polypropylene oxide on model polystyrene latex dispersions, Langmuir 4 (1998) 1055-1061.

22. B. Vincent, J. Edwards, S. Emmett, A. Jones, Depletion Flocculation in Dispersions Sterically-Stabilised Particles (“Soft Particles”), Colloids and Surfaces 18 (1986) 261-281.

23. J. N. Israelachvili, Intermolecular & Surface Forces, Academic Press, London 1992.

24. R. J. Levins, Barium Ion-Selective Electrode Based on s Neutral Carrier Complex, Anal. Chem. 43 (1971) 1045-1047.

25. J. Nuyssink, K. Koopal, The Effect of Polyethylene Oxide Molecular Weight on Determination of its Concentration in Aqueous Solutions, Talanta 29 (1982) 495-501.

26. R. T. Thima, S. Tammishetti, Barium Chloride Crosslinked Carboxymethyl Guar Gum Beads for Gastrointestinal Drug Delivery, J. Applied Polym. Sci. 82 (2001) 3084-3090.

Colloidal Stability of Pluronic® F68 Coated PLGA Nanoparticles: A Variety of Stabilization Mechanisms

M.J. Santander-Ortega1, A.B. Jódar-Reyes2, N. Csaba3, D. Bastos-González1, J.L. Ortega-Vinuesa1

1 Biocolloid and Fluid Physics Group, Department of Applied Physics, University of Granada, Av. Fuentenueva S/N, 18071, Granada (SPAIN)

2 Department of Physics, University of Extremadura, Av. Universidad S/N, 10071, Cáceres (SPAIN)

3 Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy, University of Santiago de Compostela, 15706, Santiago de Compostela (SPAIN)

62 J. Colloid Interface Sci. 302 (2006) 522-529.

ABSTRACT

Poloxamers are a family of polypropylene oxide (PPO) and polyethylene oxide (PEO) tri-block copolymers that are usually employed in the micro and nano-particulate engineering for drug delivery systems. The aim of this work is to study the electrophoretic mobility (μe) and colloidal stability of complexes formed by adsorbing a poloxamer (Pluronic® F68) onto poly(D,L-lactic-co-glycolic acid) (PLGA) nano-particles. A variety of stabilization mechanisms have been observed for the Pluronic coated PLGA nanoparticles, where DLVO interactions, solvent-polymer segment interactions and hydration forces play different roles as a function of the adsorbed amount of Pluronic. In addition, the μe and stability data of these complexes have been compared to those obtained previously using a PLGA-Pluronic® F68 blend formulation. As both the μe and the stability data are identical between the two systems, a phase separation of both components in the PLGA-Pluronic blend formulation is suggested, being the PLGA located in the core of the particles and the Pluronic in an adsorbed shell.

Paper II 63

I. INTRODUCTION

The use of adsorbing macromolecules to modify the aggregation state, sedimentation behaviour, and rheological properties of colloidal dispersions represents an industrially significant, although largely empirical, technology [1]. The pharmaceutical industry also develops and works with colloidal systems for drug delivery purposes. In this case, any successful pharmaceutical application requires adjustment of the surface properties of the polymeric drug delivery system to be compatible with the biological environment. Thus, the use of biocompatible macromolecules adsorbed onto biodegradable nanoparticles is necessary, not only to avoid spontaneous particle aggregation under certain physico-chemical conditions of pH, ionic strength and temperature, but also to prevent the rapid uptake of intravenously injected particulate drug carriers by the cells of the reticuloendothelial system [2]. It has been proven that surface modification of the carriers by adsorption of non-ionic amphiphilic macromolecules (i.e. poloxamers, poloxamines or PEG derivatives) helps to overcome such a drawback [3]. In addition, the presence of these macromolecules in the drug carrier composition may also help to improve the release of the encapsulated materials (drugs, proteins, DNA, and so on) and even protects the proteins encapsulated into the carriers against partial or total denaturation [4,5]. This protective action can be explained as follows. The PLGA degradation is governed by hydrolytic processes and leads to the formation of acidic oligo- and monomers that cause an acidic microclimate. The use of protective excipients (i.e. polaxamers and poloxamines) could possibly prevent unwanted interactions between the drug and the PLGA as well as neutralize the acidity generated in the course of polymer degradation.

Although adsorption of this kind of surfactants is the most wide known procedure to modify the surface characteristics of the primitive carriers, the incorporation of these copolymers into the particles during the manufacturing process has become an alternative strategy. Successful incorporation of polypropylene oxide – polyethylene oxide (PPO-PEO) copolymers into Poly(D,L-lactic-co-glycolic acid) (PLGA) particles has been recently reported [5,6]. The extent of incorporation depends strongly on both the hydrophobic/hydrophilic degree of the carrier matrix and the hydrophilia-lipophilia balance (HLB) values of the PPO-PEO-derivatives. Therefore, whereas the incorporation of surfactants with high hydrophilicity (high HLB values) into hydrophobic nanoparticle matrixes is probably limited, an effective mixture and homogeneous distribution would be expected for surfactants with large hydrophobic and short hydrophilic moieties (low HLB values). This statement have been recently confirmed by Kiss et al. [7,8] who have studied the distribution of different poloxamers (Pluronics) into poly(lactic acid) (PLA) and PLGA blend films.

In the present work, the adsorption of Pluronic® F68 on PLGA nanoparticles has been studied and, subsequently, the electrophoretic mobility (μe) and colloidal stability of these complexes have been analysed. As will be shown, different stability mechanisms have been observed as a function of the surfactant

64 J. Colloid Interface Sci. 302 (2006) 522-529.

coverage, which suggests different spatial conformations of the poloxamer molecules in the adsorbed layer. In addition, a comparison between the μe and stability of our complexes with those of other particles obtained in another work [9], (where they were manufactured by simultaneously mixing PLGA and Pluronic®

F68 during the formulation of the particles,) is presented. As electrophoretic mobility and colloidal stability are exclusively dependent on the surface characteristics, the comparison may help us to gain an insight into the exact location of the polyoxide copolymers in the blend formulations. That is, if pure PLGA particles covered by Pluronic and blended PLGA-Pluronic particles present similar mobility and stability properties, it could be inferred that, in the blend formulation, a large part of the poloxamer must be located on the particle surface.

H2C

H2C

II. MATERIALS AND METHODS

2.1. Materials

The polymer poly (D,L-lactic acid / glycolic acid) 50:50 (PLGA) was purchased from Boehringer-Ingelheim, under the commercial name of Resomer® RG 503. Its average molecular weight was 35000 Da. The poloxamer Pluronic® F68 was obtained from Sigma Aldrich. It is a polyoxyethylene-polyoxypropylene-polyoxyethylene type polymer (see Figure 1) with a molecular weight equal to 8500 Da. All other solvents and chemicals used were of the highest grade commercially available. Buffered solutions presented a constant ionic strength of 0.002 M.

2.2. Preparation of PLGA nanoparticles

The PLGA nanoparticles were prepared by a modified emulsion-solvent diffusion technique. First, 50 mg of PLGA was dissolved in 2 ml of dichloromethane and this organic solution was mixed for 30 s with 0.2 ml of pure water by vortex (2400 min-1, Heidolph). This first volume of water is in pharmaceutical applications used to dissolve drugs to be incorporated in the particles. Then, the obtained emulsion was poured under moderate magnetic stirring onto a larger polar phase (25 ml ethanol), leading to immediate polymer precipitation in the form of nanoparticles. This sample was diluted with 25 ml

HO

H2C O O

CH2

CH O CH2

H

CH3n m n

PEO PPO PEO

Figure 1. Chemical structure of a poloxamer. It presents a central hydrophobic fragment of polyoxypropylene (PPO) and identical hydrophilic chains of polyoxyethylene (PEO) at both sides. For the Pluronic F68, n = 75 PEO units and m = 30 PPO units.

Paper II 65

MilliQ water and the stirring was maintained for 10 more minutes. Finally, the organic solvents (both ethanol and dichloromethane) were eliminated under vacuum at 30oC (Rotavapor Büchi R-114, Flawil).

2.3. Adsorption isotherm

The adsorption of Pluronic® F68 onto PLGA particles was performed as follows: 7.6 mg of PLGA (with a total surface area of 0.2 m2) was added to 8 mL of buffered solutions at pH 7 containing different concentrations of poloxamer. These solutions were poured in stoppered cylinders, which were then gently rotated end-over-end and set in cabinets agitated on a rotating plate at constant temperature (25oC) for 20 h. The adsorbed amounts reach a steady value well within this time [10]. The dispersions were centrifuged at 15000 g for 15 min and the supernatants were analyzed with the molybdophosphoric acid reagent [11]. It should be noted that the quantification of the poloxamer concentration by means of this reagent is only reliable for polymer concentrations below its critical micelle concentration (CMC). Consequently, the poloxamer concentration was spectrophotometrically determined using calibration curves constructed in the 0 – 400 mg/L range, where the errors are usually in the 5-10 % range.

2.4. Electrophoretic mobility

The electrophoretic mobility measurements were carried out with a Zeta-Sizer IV (Malvern Instruments). The PLGA-Pluronic particles were diluted in the desired buffered solutions for 10 min just prior to measuring. Final particle concentration was equal to 3 x 109 particles/mL. Then, the mobility data were taken from the average of six measurements at the stationary level in a cylindrical cell.

2.5. Hydrodynamic diameter and colloidal stability

The average size of stable PLGA and PLGA-Pluronic particles was analysed by photon correlation spectroscopy (PCS) using a 4700c System (Malvern Instruments). The same apparatus was used to study the aggregation of these particles in saline media, using NaCl and Ca2Cl as aggregating electrolytes. The PCS instrument had an Argon laser (λ0 = 488 nm) with a perpendicular polarization and a power rating of 75 mW. 2 mL of very diluted samples were poured into a cylindrical cell and 1 mL of the saline solution at the desired ionic strength was then added and rapidly stirred. The scattered intensity auto-correlation function measured at 60o was then analysed by the computer software. The aggregation measurements took about 10 min. Information about the aggregation kinetics was obtained by plotting the average diameter of the particles versus time.

III. RESULTS AND DISCUSSION

Four sets of experiments were performed to determine different properties of our Pluronic coated PLGA particles: adsorption isotherm, Pluronic adlayer thickness, electrokinetic behaviour of the complexes versus the pH, and colloidal stability as a function of the electrolyte concentration.

66 J. Colloid Interface Sci. 302 (2006) 522-529.

Adsorption isotherm results are shown in Figure 2. As expected from the work of Kayes and Rawlins [12], the specific adsorption values for Pluronic® F68 reach a clear plateau value at bulk polymer concentrations below its CMC. It is instructive to compare the plateau surface coverage determined in the present work to those reported in literature. Tadros and Vincent [10] and Baker and Berg [13] obtained a plateau adsorption value of approximately 8.5 10-4 g/m2 and 9.5 10-4 g/m2 respectively working with polystyrene particles. These values coincide with our plateau adsorption values (which are around 10 10-4 g/m2), although PLGA

0 100 200 300 4000,0

0,5

1,0

1,5

Γ ads (

mg/

m2 )

Ce (mg/L)

Figure 2. Adsorption isotherm of Pluronic® F68 onto PLGA particles. The adsorbed amount(Γads) is plotted versus the poloxamer equilibrium concentration (Ce). Vertical dotted line indicates the Pluronic CMC (400 mg/L).

(a) (b) (c)

Figure 3. Scheme of PLGA-Pluronic complexes. Particles and surfactant molecules arerepresent at different scale. The PPO block anchors the poloxamer to the PLGA surface,while the PEO chains extend to the solution. (a) Low coverage (below the adsorptionplateau). (b) Coverage at the adsorption plateau. (c) Hemimicelles adsorbed on the PLGAsurface when using Pluronic concentrations around the CMC.

Paper II 67

0,0 0,2 0,4 0,6 0,8 1,0 1,2150

200

250

300

350

50

100

150

200

250

Φ (n

m)

Γads(mg/m2)

I (Kc

ps)

Figure 4. Hydrodynamic diameter ( ) and scattered light intensity (at 60o) ( ) of different PLGA-Pluronic complexes.

instead of polystyrene surfaces are used. Nevertheless, the isotherm suffers a little change just below the CMC. AFM studies [14-18] have shown that surfactants do not always form homogeneous layers at any interface. Actually, non-ionic surfactants can form micellar-like surface aggregates under certain conditions [19,20], presenting different regimes for the formation of surface aggregates. For such surfactants, a step-isotherm is often found, which has also been explained theoretically [21,22]. Therefore, the observed change in the isotherm curve might be caused by an adsorption of hemimicelles onto the PLGA surface, as illustrated in Fig. 3. It should be noted that if experimental errors are taking into account, it is difficult to discern if this last point belongs to a single plateau or if it is out of this plateau. This conflictive point can be solved analysing differences on some colloidal properties associated to the surface characteristics (i.e. electrophoretic mobility or colloidal stability). As will be shown afterward, differences in the colloidal stability exist, and thus, it is more than likely that this last complex has a higher poloxamer coverage than that of the plateau.

The hydrodynamic diameter (∅) was measured by photon correlation spectroscopy for the complexes sensitised below the CMC. Results are shown in Fig. 4. It is possible to obtain the plateau adlayer thickness of the Pluronic® F68 comparing the ∅ values of the complexes with those of the bare particles. The calculated thickness was around 20 nm. This value is higher than that reported by Baker and Berg [13] using a 56-nm-diameter polystyrene latex, who obtained a 6 nm thickness for the same tri-block copolymer using the same optical technique. It should be noted that adsorption is driven by hydrophobic forces between the PPO moiety and the adsorbent surface. PLGA is not as hydrophobic as polystyrene, and thus, the adhesion of the surfactant onto the surface may be less compact in the

68 J. Colloid Interface Sci. 302 (2006) 522-529.

former than in the latter. This reasoning would explain, at least qualitatively, the above differences.

The electrokinetic behaviour was evaluated by measuring the electrophoretic mobility (μe) versus pH at different surfactant coverage. It should be noted that PLGA chains that form the particles have carboxylic groups at the extreme of them. It is more than likely that this charged groups tend to be located in an aqueous environment, that is, at the particle/water interface, creating a surface charge density capable of generating colloidal stability even without adsorbed Pluronic molecules. These weak acid groups explain the electrokinetic behaviour and the colloidal stability of the particles in absence of poloxamer. The (μe) results are shown in Fig. 5. Bare particles present a negative mobility value that increases in absolute value up to a certain pH (pH=5.5); this is due to the deprotonation of the superficial carboxyl groups. For pH values above 5.5, a plateau is reached. By adding Pluronic, the mobility of the complex remains negative. However, as a general rule, the mobility value decreases in absolute value by increasing the Pluronic load. The samples with coverage around the adsorption plateau present a similar electrokinetic behaviour, even the last point of the isotherm. However, differences are found when working with complexes obtained at concentrations above the CMC (see 1.5 x CMC data), which can be due to structural changes at the surface. In the case of the complex obtained just below the CMC (Γads = 1.33 mg/m2), hemimicelles might start appearing on the surface, and the polymer adsorbed amount increases slightly. Nevertheless, changes on the electrokinetic

2 3 4 5 6 7 8 9 10-4,5

-4,0

-3,5

-3,0

-2,5

-2,0

-1,5

-1,0

-0,5

μ e x 1

08 (m2 /V

s)

pH

Figure 5. Electrophoretic mobility versus pH of different PLGA-Pluronic complexes. Bare PLGA particles ( ). Γads = 0.38 mg/m2 ( ). Γads = 0.87mg/m2 ( ). Γads = 1.06 mg/m2 ( ). Γads = 1.13 mg/m2 ( ). Γads = 1.33 mg/m2 ( ). Above the CMC, Γadded = 1.5 x CMC ( ). PLGA-Pluronic blend formulation ( ), dash line.

Paper II 69

behaviour are not observed when comparing with the data corresponding to the plateau (Γads = 1.06 and 1.13 mg/m2). Only when the complex is formed at concentrations well above the CMC, the polymer adsorbed amount on the surface has reached its maximum value, which gives rise to differences in the electrokinetic behaviour that could come from structural changes of the adsorbed poloxamer chains. Mobility comes from a combination of electrical and frictional forces. The adsorption of this non ionic surfactant onto the PLGA particles affects both forces, as detailed below. The absorbed non-ionic Pluronic layer partially screens the surface charge of the PLGA particles, and thus, reduces the ζ-potential. The adsorption also changes a smooth PLGA surface into a rough one (with extended PEO chains into the solution) [23], which shifts the shear plane outward, diminishing the ζ-potential and causing an increase in the hydrodynamic friction when particles are in motion. Both effects simultaneously cause a decrease in mobility by increasing the surfactant coverage, as experimentally observed. Finally, the mobility curve of a PLGA-Pluronic® F68 sample obtained by blending both components during the particle synthesis process [9] is also shown in Fig. 5 (dashed line). The electrokinetic behaviour of this last sample resembles that of pure PLGA particles totally coated by adsorbed Pluronic® F68. This feature would support that in a binary mixture of PLGA and Pluronic® F68 (50:50 w/w) there must be a phase separation where PLGA forms the hydrophobic core of the particles and the poloxamer tends to accumulate at the solid/water interface. It is noteworthy that some theoretical considerations based on the Kiss et al.’s studies [7,8] support the following hypothesis: nanoparticles obtained by blending Pluronic® F68 and PLGA (50/50), (adding the same mass of both components during the manufacturing process), are formed by two phases: a core rich in PLGA and an external part consisting of an adhered poloxamer shell. If this heterogeneous distribution occurs, the electrokinetic behaviour of the blend formulation particles and those obtained by adsorbing the poloxamer onto a PLGA core should coincide, as experimentally found.

The colloidal stability of the complexes was analysed as ionic strength was increased, using independently NaCl and CaCl2 as aggregating electrolytes. Stability was evaluated by calculating the Fuchs factor (W) [24] versus the salt concentration from aggregation processes monitored by an optical technique. If W = 1, the system is completely unstable, while W = ∞ indicates total stability. Details about how to get experimental W values can be found elsewhere [25]. Fig. 6a and 6b show the results obtained with NaCl and CaCl2, respectively. Stability results for those complexes obtained by adding poloxamer at concentrations above the CMC are not shown, as these complexes were completely stable at any salt concentration. This extreme stability was also found with the last point of the isotherm (see Fig. 2), but not with the other complexes located at the plateau. Therefore, differences in the poloxamer coverage actually must exist between the former and the latter. In addition, the total stability shown by the particles coated with poloxamer at concentrations above the CMC also coincides with that found for the PLGA-

70 J. Colloid Interface Sci. 302 (2006) 522-529.

-1,0 -0,8 -0,6 -0,4 -0,2 0,0-0,5

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5

Pluronic blend formulations [9], and it is caused by a steric mechanism produced by the external PEO chains that protrude into the solution. It is therefore observed that not only the electrophoretic mobility but also the stability results suggest a heterogeneous distribution of PLGA and Pluronic® F68 in blend formulations. In addition, it is worth to highlighting that the PLGA-Pluronic particles show a

Figure 6a. Stability factor versus NaClconcentration. Bare PLGA particles ( ). Γads = 0.38 mg/m2 ( ). Γads = 0.64 mg/m2

( ). Γads = 0.87 mg/m2 ( ). Γads = 1.06 mg/m2 ( ).

log

W

log [NaCl] (M)-2,0 -1,5 -1,0 -0,5 0,0

-0,5

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5

log

W

log [CaCl2] (M)

Figure 6b. Stability factor versus CaCl2concentration. Bare PLGA particles ( ). Γads = 0.38 mg/m2 ( ). Γads = 0.64 mg/m2

( ). Γads = 0.87 mg/m2 ( ). Γads = 1.06 mg/m2 ( ). Γads = 1.13 mg/m2 ( ).

0 1 2 3 4 5

-30

-20

-10

0

10

20

30

V/k

T

H (nm)

Figure 7a. Interaction potentials versus distance according to the DLVO theory. Dotted line:attractive potential (VA) using a Hamaker constant A = 0.5 10-20 J. Different repulsive potentials (VE) using a Stern potential Ψδ = 15.3 mV at the following NaCl concentrations: 25mM (solid line); 270 mM (CCC) (dashed - dotted line); and 500 mM (dashed line). Totalinteraction potentials (VT) at 25 mM ( ), 270 mM (CCC) ( ), and 500 mM ( ).

Paper II 71

0 1 2 3 4 5-8

-6

-4

-2

0

2

4

6

8

relevant variety of stability mechanisms, depending on the surfactant coverage. Initially, the bare PLGA particles exhibit a stability behaviour that can be explained by the DLVO theory [24], which is applicable to lyophobic colloids. It is widely known that this theory considers the total interaction potential (VT) between two approaching particles dependent on two terms, one attractive (VA) and another repulsive (VE). VA corresponds to the interaction energy caused by the van der Waals dispersion forces, while VE represents the repulsive interaction created by the overlapping of the electrical double layers of the charged particles. The value of this repulsive term depends on the surface electrostatic potential (which in turn depends on the surface charge density of the particles), and it is modulated by the ionic strength of the medium in such a way that the higher the salt concentration, the lower the VE value. Those interested in reading further about the topic can find the analytical expressions of all these energetic terms in the Supplementary Material Section that can be found at the end of this paper. VT can be obtained simply by summing VA and VE. Thus, the colloidal stability of the bare PLGA particles can be reduced by diminishing the VE term (and thus VT) by adding electrolytes, as shown in Fig. 7a. This destabilization by salt is observed in Fig. 6a and 6b, where W diminishes to unity when salt concentration increases. The critical coagulation concentration (CCC), that is the minimum salt concentration needed to rapidly aggregate the system, can be obtained from these figures locating that point where log W reduces to zero. The CCC values so obtained are shown in Table 1. The screening effect exerted by calcium is more significant than that of the sodium. That explains why the CCC values obtained in presence of CaCl2 are lower than those of NaCl.

800 mM

1000 mM

550 mM

400 mM

200 mM

100 mM

50 mMV

T/kT

H(nm)

Figure 7b. Total interaction potential versus distance including the hydration interaction term(Vh). The following parameters were set at the values shown below: A = 0.5 10-20 J, Ψδ = 13.7 mV, Ch = 0.2 10-20 J , and λ = 0.6 nm [32].

72 J. Colloid Interface Sci. 302 (2006) 522-529.

0,0 0,5 1,0 1,5 2,0 2,5-200

-150

-100

-50

0

50

100

150

200

V/k

T

H(nm)

If PLGA particles are covered by small amounts of poloxamer, the surface electrostatic potential is partially reduced (in absolute value) when comparing it with that of the bare particles, as some charged carboxyl groups are screened by the surfactant adsorption. This agrees with the electrokinetic results shown in Fig. 5, where the μe of the complex (and thus the ζ-potential) decreased by adsorbing Pluronic. In addition, a low surface poloxamer concentration is still not capable of stabilizing the system by a steric hindrance mechanism. Actually, Li et al. [23] suggested that a low surface concentration allows a non-extended PEO chain conformation, (as depicted in Figure 3a). All this gives rise to more unstable particles (lower CCC values), as experimentally observed. Nevertheless, if the Pluronic surface concentration is increased, other stabilizing mechanisms appear. On the one hand, a crowded surfactant surface exhibits a more extended PEO layer [23] (see Fig. 3b), which favours the steric stabilization. This can explain that, once the minimum observed in the CCC values is surpassed, the stability increases when the complex coverage increases (see Table 1). However, the most surprising result is obtained at very high salt concentrations, where the system begins to re-stabilize (that is, W increases) for high Pluronic coatings. Neither the DLVO theory nor the steric hindrance of PEO chains predicts this sort of re-stabilization. This is caused by another stabilization mechanism based on repulsive hydration forces, which become significant when great amounts of hydrated ions are accumulated at the proximities of any hydrophilic surface. The physical origin of this type of force has been studied for the last 25 years and has been associated with the structuring of

Figure 7c. Interaction potentials for PLGA-Pluronic complexes represented in Figure 3(c).Attractive potential (VA) dash-dot-dot line. Repulsive potential (VE) solid line. Osmotic term (Vosm) dashed line. Elastic-steric repulsion (Vel) dotted line. Hydration potential (Vh) dashed-dotted line. Total interaction potentials (VT) ( ).The following parameters were set at the values shown below: A = 0.5 10-20 J, Ψδ = 13.7 mV, Ch = 0.2 10-20 J , and λ = 0.6 nm [32], χ= 0.48 [34], δ = 6 nm [13], φ2 = 0.01.

Paper II 73

water in the vicinity of hydrophilic surfaces [26]. It is a structural force that arises from the local order of water layer adjacent to the surface. It is not only correlated to the hydrophilicity of the surface but also depends strongly on the nature and concentration of the hydrated counterions that surround the surface [27-29]. There are several theoretical approaches that try to include the hydration interaction in the colloidal stability framework. Most of them are based on adding a “hydration” term (Vh), given by an exponential equation [30,31], to the usual DLVO VE and VA terms, that is: VT = VA + VE +Vh. The dependence of this new VT on the electrolyte (NaCl) concentration has been depicted in Fig. 7b. Therefore, the adsorption of a Pluronic® F68 has converted a hydrophobic surface (PLGA) into a hydrophilic one, promoting new stabilizing mechanisms at very high salt concentrations. The critical stabilization concentration (CSC) is defined as the minimum salt concentration at which re-stabilization by hydration forces starts. These data are also shown in Table 1. As expected, the presence of calcium (a highly hydrated cation) favours the re-stabilization more than sodium, a less hydrated ion.

Table 1. Critical coagulation concentration (CCC) and critical stabilization concentration (CSC) of different PLGA-Pluronic complexes in presence of NaCl and CaCl2.

Aggregating salt: NaCl Aggregating salt: CaCl2 ccc (mM) csc (mM) ccc (mM) csc (mM)

Bare PLGA 270 ---- 27 ---- 0.38 mg/m2 160 ---- 17 ---- 0.64 mg/m2 200 750 22 520 0.87 mg/m2 260 550 31 370 1.06 mg/m2 300 440 30 280 1.13 mg/m2 ---- ---- 45 175

Finally, if adsorption of the poloxamer is carried out at concentrations around or higher than the CMC, the complexes obtained in such a manner are completely stable at any salt concentration. The absence of aggregation phenomena by saline effects suggests that stability must be now controlled by steric interactions. For that reason, other non-DLVO stabilizing mechanism appears in the PLGA-Pluronic system. If a layer of structured polymer chains is located at the solid/liquid interface, stability by means of polymer-induced forces may occur. These forces can be enthalpically or entropically driven and are mainly dependent on the nature of the solvent-polymer segment interactions [24]. The steric stabilization effect usually includes two contributions: osmotic and polymer coil compression [33]. Thus, two new terms can be added to the total interaction potential (VT). The first one (Vosm) considers the osmotic effects that take place when the adsorbed polymer layers of dispersed particles overlap. The local concentration of polymers in the overlap zone exceeds that of the external regions, leading to a driving force for the spontaneous flow of solvent into the interparticle region, which pushes the particles apart. Of course, this potential is highly dependent on the solvent solubility of the polymer chains, represented by χ (the Flory-Huggins solvency parameter). The second term is a restriction volume potential (Vel). If two

74 J. Colloid Interface Sci. 302 (2006) 522-529.

approaching particles are closer than a distance shorter than the adhered polymer thickness, some polymer molecules will be forced to undergo elastic compression. Thermodinamically, this compression produces a net loss in configurational entropy, which renders repulsion between the particles, and in turn improves the colloidal stability. Fig. 7c reflects a combination of all the aforementioned interaction potentials, where we have used a value of χ = 0.48 for the solvent solubility of the PEO fragments of the Pluronic, which is in line with the calculations reported by Einarson and Berg for a similar polymer [34]. As can be seen, the osmotic contribution exerted by the adsorbed polymer chains immersed in good solvents becomes the most stabilizing term. It makes the potential barrier high enough regardless of the salt concentration, yielding extremely stable particles.

It is now possible to use these stability results and the adsorption isotherm data (Fig. 2) to support our hypothesis on the Pluronic layer structure (shown in Fig.3). That is, the steric stabilization mechanism only appears in the case of complexes formed above certain surfactant concentration around and above the CMC, where a step in the adsorption isotherm was observed. This result indicates a structural change in the surfactant adsorbed layer (i.e. the formation of surface aggregates that gives a highly enriched polymer layer). On the contrary, below the CMC single Pluronic molecules adsorb on the surface forming a monolayer, and a clear adsorption plateau is obtained. In these cases, no strong steric stabilizing phenomena appear.

IV. CONCLUSIONS

In this work has been demonstrated that both the electrophoretic mobility and the colloidal stability results of PLGA nanoparticles fully coated by Pluronic® F68 practically coincide those obtained with PLGA-Pluronic blend formulations. This would suggest a two phase separation of both components in PLGA-Pluronic (50:50 w/w) mixtures, which has also been predicted theoretically analysing the miscibility of these polymers at a molecular scale, where PLGA would form the core and the poloxamer would be located at the particle surface. On the other hand, the Pluronic coated PLGA nanoparticles have exhibited different stability patterns depending on the poloxamer coverage. At null or low coverage, pure DLVO interactions take place. Intermediate coatings stabilize the system at very high ionic strength by means of hydration forces. Finally, in the case of complexes formed above certain poloxamer concentration just below and above the CMC, strongly stabilizing steric mechanisms do manifest. The electrokinetic and stability results together with the adsorption behaviour of the system suggest a structural change in the surfactant layer at high poloxamer concentrations, (i.e. the formation of surface poloxamer aggregates).

ACKNOWLEDGEMENTS

Financial support from ‘Comisión Interministerial de Ciencia y Tecnología’ Projects MAT2003-01257 and AGL2004-01531/ALI (European FEDER support included) is gratefully acknowledged.

Paper II 75

References

[1] B. Dobias, X. Qiu, W. von Rybinski, Solid-Liquid Dispersions. Science Series, Vol. 81. Marcel Dekker, New York, 1999.

[2] S.M. Moghimi, H.M. Patel, Biochim. Biophys. Acta 984 (1989) 379.

[3] S.M. Moghimi, A.C. Hunter, Tibtech 18 (2000) 412.

[4] R.L. Cleek, K.C. Ting, S.G. Eskin, A.G. Mikos, J. Control Release, 48 (1997) 259.

[5] N. Csaba, L. González, A. Sánchez, M.J. Alonso, J. Biomater. Sci. Polymer Edn., 15 (2004) 1137.

[6] S.E. Dunn, A.G.A. Coombes, M.C. Garnett, S.S. Davis, M.C. Davies, L. Illum, J. Control Release, 44 (1997) 65.

[7] E. Kiss, I. Bertóti, E.I. Vargha-Butler, J. Colloid Interface Sci., 245 (2002) 91.

[8] E. Kiss, M.G. Takács, I. Bertóti, E.I. Vargha-Butler, Polym. Adv. Technol., 14 (2003) 839.

[9] M.J. Santander-Ortega, N. Csaba, M.J. Alonso, D. Bastos-González, J.L. Ortega-Vinuesa, Colloid Surf. B, submitted (2006).

[10] T.F. Tadros, B. Vincent, J. Phys. Chem., 84 (1980) 1575.

[11] J. Nuysink, L.K. Koopal, Talanta, 29 (1982) 495.

[12] J.B. Kayes, D.A. Rawlins, Colloid Polym. Sci., 257 (1979) 622.

[13] J.A. Baker, J.C. Berg, Langmuir, 4 (1988) 1055.

[14] S. Manne, J.P. Cleveland, H.E. Gaub, G.D. Stucky, P.K. Hansma, Langmuir, 10 (1994) 4409.

[15] S. Manne, H.E. Gaub, Science, 270 (1995) 1480.

[16] E.J. Wanless, W.A. Ducker, J. Phys. Chem., 100 (1996) 3207.

[17] W.A. Ducker, L.M. Grant, J. Phys. Chem., 100 (1996) 11507.

[18] T. Svitova, R.M. Hill, C.J. Radke, Colloids Surf. A, 183-185 (2001) 607.

[19] P. Levitz, H. van Damme, D. Keravis, J. Phys. Chem., 88 (1984) 2228.

[20] P. Levitz, H. van Damme, J. Phys. Chem., 90 (1986) 1302.

[21] A.B. Jódar-Reyes, J.L. Ortega-Vinuesa, A. Martín-Rodríguez, F.A.M. Leermakers, Langmuir, 18 (2002) 8706.

[22] A.B. Jódar-Reyes, J.L. Ortega-Vinuesa, A. Martín-Rodríguez, F.A.M. Leermakers, Langmuir, 19 (2003) 878.

[23] J.T. Li, K.D. Caldwell, N. Rapoport, Langmuir, 10 (1994) 4475.

[24] P.C. Hiemenz, R. Rajagopalan, Principles of colloidal and surface chemistry. Marcel Dekker, New York, 1997.

[25] T. López-León, A.B. Jódar-Reyes, D. Bastos-González, J.L. Ortega-Vinuesa, J. Phys. Chem. B, 107 (2003) 5696.

[26] J.N. Israelachvili, G.E. Adams, J. Chem. Soc. Faraday Trans., 74 (1978) 975.

[27] J.N. Israelachvili, Intermolecular & Surface Forces. Academic Press, London, 1992.

76 J. Colloid Interface Sci. 302 (2006) 522-529.

[28] R.M. Pashley, Adv. Colloid Interface Sci., 16 (1982) 57.

[29] J.A. Molina-Bolívar, F. Galisteo-González, R. Hidalgo-Álvarez, Colloids Surf B, 21 (2001) 125.

[30] N.V. Churaev, B.V. Derjaguin, J. Colloid Interface Sci., 103 (1985) 542.

[31] J.A. Molina-Bolívar, F. Galisteo-González, R. Hidalgo-Álvarez, Phys. Rev. E, 55 (1997) 4522.

[32] J.A. Molina-Bolívar, J.L. Ortega-Vinuesa, Langmuir, 15 (1999) 2644.

[33] B. Vincent, J. Edwards, S. Emmet, A. Jones, Colloid Surf., 18 (1986) 261.

[34] M.B. Einarson, J.C. Berg, J. Colloid Interface Sci., 155 (1993) 165.

SUPPLEMENTARY MATERIAL

A brief discussion about the interaction potential energies versus the distance between colloidal particles along with the corresponding mathematical expressions used for plotting Figures 7a-c are shown below.

According to the DLVO theory, the total potential energy (VT) of interaction, which determines the colloid stability, between two lyophobic particles is defined as:

V V (1) VT A E= +

The term VA, the energy of interaction between the particles due to van der Waals dispersion forces, is expressed as [24]:

( ) ( )( )

( )V

A aH a H

aa H

H a H

a HA = − +

++

++

+

⎡

⎣⎢⎢

⎤

⎦⎥⎥6

24

22

4

2

2 2

2 2ln (2)

where A is the Hamaker constant for particles interacting in the medium (water), a the particle radius and H is the distance between the surfaces of the particles.

The quantity VE represents the repulsive interaction between the electrical double layers of the particles. According to the constant potential model, a reasonable expression to VE (for moderate potential in the Stern layer, < 50 mV), is [35]:

)2(2

04

)(2 Δ−−⎟⎟⎠

⎞⎜⎜⎝

⎛Δ+= H

i

BE e

ezTk

aV κγεεπ (3)

where Δ is the thickness of the Stern layer, ε the dielectric constant of the medium, ε0 the vacuum permittivity, kB the Boltzman constant, T the absolute temperature, zi the valence of the ion, e the elementary charge, κ is the Debye parameter,

1/ 22 2

0 0/i i Bi

n e z k Tκ ε ε⎛ ⎞= ⎜⎝ ⎠∑ ⎟ (4)

and γ is given by the following equation:

⎟⎟⎠

⎞⎜⎜⎝

⎛=

Tkez

tanhB

i

4δψ

γ (5)

where ψδ is the Stern potential. Actually, equation (5) is valid for κa>>1.

When hydrophilic instead of hydrophobic surfaces interact, hydration forces play an important role on the system stabilization. This non-DLVO

78 J. Colloid Interface Sci. 302 (2006) 522-529.

interaction can be quantified by an energetic term (Vh) that, as a rule, depends exponentially on the interlayer thickness. Molina-Bolívar et al. [31] proposed the following equation where the pre-exponential term is directly proportional to salt concentration (ce):

V a N C c eh A h eH= −π λ λ( ) /2 (6)

NA is the Avogadro number, Ch is a proportionality constant called “hydration constant”, its value depending on the surface hydrophilicity, the ce concentration is expressed in mM, and λ is the decay length. The value of λ depends on the hydration number of the hydrated counterion, and it is usually within the 0.2-1.1 nm range. Including Eq. 6 on Eq. 1, estimations of the total potential energy of interaction as a function of the separation distance were computed and represented in Figure 7b.

Finally, if a particle is coated by a polymer that is completely soluble in good solvents, a significant steric force arises and it prevents the coagulation of the system. As mentioned in the main text, there are two stabilizing mechanisms exerted by the adhered polymer layer, namely, an osmotic and a purely steric contribution. There are different analytical expressions to quantify both interaction energy terms. In this work, the expressions shown by Vincent et al. [33] have been used. According to these authors, if there are polymeric chains covering the external surface of a particle, the average thickness of such coils being δ, then an osmotic effect will appear when both particles are closer than a distance equal to 2δ. In that case the osmotic potential of repulsion (Vosm) can be considered as

2osm 2

1

4 a = ( (1/2 - ) ( - H/2 )V 2)π χ δφν

(7)

where ν1 represents the molecular volume of the solvent, φ2 is the effective volume fraction of polymer segments in the adlayer and χ is the Flory-Huggins solvency parameter.

However, if both particles are closer than a distance equal to δ, at least some of the polymer molecules will be forced to undergo elastic compressions. This effect originates a new repulsion potential (Vvr) related to the restriction volumen, which limites the movement of the hydrophilic coils extended towards the solvent. This elastic-steric repulsion is given as

ln ln2

2el 2 2

2 a H h 3 - H/ 3 - H/ = -6 +3(1+ H/ ) VMW 2 2π δ δ δφ ρδ

δ δ

⎛ ⎞⎡ ⎤⎛ ⎞ ⎛ ⎞ ⎡ ⎤⎜ ⎟⎢ ⎥⎜ ⎟ ⎜ ⎟ ⎢ ⎥⎜ ⎟⎝ ⎠ ⎝ ⎠ ⎣ ⎦⎢ ⎥⎣ ⎦⎝ ⎠(8)

where ρ2 and MW are the density and molecular weight of the adsorbed polymer.

At such a distance the expression for the osmotic potential must be modified, and now it is given by the following equation:

Suplementary Material Paper II 79

ln2 2osm 2

1

4 a H 1 H = ( (1/2 - ) - - )V 2 4π χφ δ

δ δν⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

(9)

The total interaction potential could be obtained as the sum of all the energetic terms: VT =VA + VE + Vh + Vosm + Vel , as shown in Figure 7c. Nevertheless, such a simplistic approach of adding all the terms may be questioned, as some repulsive terms are not entirely independent [34].

80 J. Colloid Interface Sci. 302 (2006) 522-529.

References

[24] P.C. Hiemenz, R. Rajagopalan, Principles of colloidal and surface chemistry. Marcel Dekker, New York, 1997.

[31] J.A. Molina-Bolívar, F. Galisteo-González, R. Hidalgo-Álvarez, Phys. Rev. E, 55 (1997) 4522.

[33] B. Vincent, J. Edwards, S. Emmet, A. Jones, Colloid Surf., 18 (1986) 261.

[34] M.B. Einarson, J.C. Berg, J. Colloid Interface Sci., 155 (1993) 165.

[35] E. Matijevic, K.G. Mathai, R.H. Ottewill, M. Kerker, J. Phys. Chem., 65 (1961) 826.

Electrophoretic Mobility and Colloidal Stability of PLGA Particles Coated with IgG

M.J. Santander-Ortega, D. Bastos-González and J.L. Ortega-Vinuesa

Biocolloid and Fluid Physics Group, Department of Applied Physics, University of Granada, Av. Fuentenueva S/N, 18071 Granada , Spain

82 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

ABSTRACT

Drug delivery systems based on polymeric nanocarriers have been widely exploited during the last years. However, one of the basic problems that is still not totally solved in this kind of systems is the ability of delivering drugs to specific target cells. Coating the nanocarrier with reactive antibodies against specific molecules presented in the external membrane of the target cells is a usual recommendation. In this paper, an ideal delivery system has been studied. Nanoparticles made of poly(D,L-lactic acid/glycolic acid) 50/50 (PLGA) polymers have been coated with polyclonal IgG. In the first part of the paper, some basic characteristics of these IgG-PLGA complexes have been analysed (i.e. size, electrophoretic mobility, and colloidal stability). Then, the immuno-reactivity of the immobilized IgG molecules was tested by using an optical device, monitoring the binding of a standard molecule (C-reactive protein, CRP) to the antibody (antiCRP-IgG) adsorbed on the PLGA particles. This allowed us to estimate the percentage of active IgG molecules on the PLGA particles by applying a simple kinetic model to the immuno-reactivity results. According to this model, the PLGA-IgG particles supply a good immunoresponse even if only less than 5% of the total IgG molecules on the surface were active. Despite the simplicity of the system, the results may be of potential interest for developing more realistic nanocarriers with targeting ability. That is, it can be inferred that it is possible to obtain a high targeting specificity in IgG-sensitized nanocarriers even working with a low coverage of active antibody molecules. The results have been compared with those similarly obtained with polystyrene (PS) particles used as a reference system.

Paper III 83

I. INTRODUCTION

A major effect in the pharmaceutical field has been caused by the use of colloidal dispersions of biodegradable polymers as nanoparticle drug carriers. This kind of systems, which emerged around 30 years ago, appears to be very promising to carry drugs, proteins or DNA molecules to specific organs within the body [1]. However, since the conception of this idea, a great deal of information about the potential and limitations of nanoparticles as drug carriers has been gathered. There is a general agreement about what the ideal carrier system must possess, which includes some specific common features, i.e. i) the size of the particles must be in the nanometric scale (≤ 200 nm) in order to be capable of crossing diverse biological barriers; ii) the external shell must be hydrophilic enough to evade the reticuloendothelial system (RES) and the mononuclear phagocytic system (MPS) and remain in the blood for a considerable amount of time; and iii) the particles must have reactive groups on the surface to be converted to stimuli responsive carriers that target to specific receptor ligands. In the case of carrying hydrophobic drugs, the use of hydrophobic particles becomes necessary to attain an appropriate drug load due the drug-polymer compatibility [2]. In these cases, the carrier particles must be made of a hydrophobic core that protects the drug from degradation and a hydrophilic shell (made of PEG, poloxamer or poloxamine derivatives) that prevents the recognition of the nanocarrier by the MPS cells. With regard to the targeting properties, the use of specific antibodies immobilized on the particle surface has become a feasible strategy usually recommended by numerous scientists.

The present work has been designed to model a simple drug delivery system with targeting ability. The particles have been made of poly(D,L-lactic acid/glycolic acid) 50/50 (PLGA) polymers, which is a material that has been extensively and successfully used in colloidal drug carriers [3-15]. In order to turn the hydrophobic PLGA surface into a hydrophilic one, the adsorption of a polypropylene oxide (PPO) and polyethylene oxide (PEO) tri-block copolymer (poloxamer) was performed in a previous work [16], where the colloidal stability of the PLGA-poloxamer complexes was also evaluated. In the present work, however, the potential capacity of PLGA nanoparticles to serve as specific reactive systems is studied. It should be noted that Kim et al. [17] have already developed PLGA particles with targeting specificity, although in this case specificity was obtained using a folate conjugate against folate receptors, which usually are overexpressed in cancer cells. Nevertheless, the use of immobilized antibodies in PLGA carriers is a targeting strategy that has not been sufficiently exploited at the present. In this paper, polyclonal IgG molecules have been adsorbed onto the PLGA particles in order to initially study some basic characteristics of these complexes, i.e. the electrophoretic mobility and colloidal stability. Subsequently, the targeting properties of the particles against any potential receptor have been analyzed by means of immunoprecipitation studies. C-reactive protein (CRP) has been used as a standard target molecule. Our goal in this point has been centred on quantifying

84 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

the reaction between the adhered immunoreactive antibody with the target analyte in solution. It has been possible to estimate how many anti-CRP-IgG molecules remain active after their adsorption onto the PLGA surface by modelling the kinetics of CRP-antiCRP-IgG reactions using particle enhanced optical immunoassays. Although this is a simple system, the results obtained in this work may be of potential interest for the development of real drug carriers for in vivo studies. Finally, a preliminary study about the colloidal stability of poloxamer-IgG-PLGA ternary systems is also shown. It should be noted that most of the results have been compared with those similarly obtained with polystyrene (PS) particles; PS-IgG complexes can be considered as a reference system, since they have extensively been used in numerous latex enhanced immunoassays during the last decades [18-21].

II. MATERIALS AND METHODS

2.1. PLGA nanoparticles

The poly (D,L-lactic acid / glycolic acid) 50:50 (PLGA) polymer was purchased from Boehringer-Ingelheim, under the commercial name of Resomer® RG 503. Its average molecular weight was 35000 Da. The PLGA nanoparticles were prepared in our labs by a modified emulsion-solvent diffusion technique. First, 50 mg of PLGA was dissolved in 2 ml of dichloromethane and this organic solution was mixed for 30 s with 0.2 ml of pure water by vortex (2400 min-1, Heidolph). This first volume of water is in pharmaceutical applications used to dissolve drugs to be incorporated in the particles. Then, the obtained emulsion was poured under moderate magnetic stirring onto a larger polar phase (25 ml ethanol), leading to immediate polymer precipitation in the form of nanoparticles. This sample was diluted with 25 ml MilliQ water and the stirring was maintained for 10 more minutes. Finally, the organic solvents (both ethanol and dichloromethane) were eliminated under vacuum at 30º C (Rotavapor Büchi R-114, Flawil). The mean hydrodynamic diameter of the PLGA particles was obtained by photon correlation spectroscopy using a 4700c System (Malvern Instruments, UK), and it was equal to 210 nm.

2.2. PS nanoparticles

The polystyrene (PS) latex particles were synthesized, cleaned and characterized by the Ikerlat S.A. laboratories (Spain). In order to maintain the same superficial charged groups as those existing in the PLGA particles, a carboxylic PS latex was chosen. The size of this PS sample was 260 nm and its polydispersity index was equal to 1.008.

2.3. Other materials

Different proteins have been used in this work: human C-Reactive Protein (CRP), bovine serum albumin (BSA) and polyclonal pre-immune immuno-γ-globulin (IgG) were purchased from Sigma. Polyclonal antiCRP-IgG was obtained, purified and kindly donated by Biokit S.A. (Spain). The isoelectric point (iep) of

Paper III 85

both polyclonal IgG samples was determined by isoelectric focusing, and it was in the 6.0 - 7.9 range. The Pluronic® F68 poloxamer was obtained from Sigma. It was a PEO-PPO-PEO type polymer with a molecular weight equal to 8500 Da. All other solvents and chemicals used were of the highest grade commercially available.

2.4. IgG adsorption

The adsorption of IgG molecules onto PLGA and PS particles followed the same procedure, which is described below. A small volume of the stock latex suspensions (containing a PLGA or PS total area equal to 0.2 m2) was added to the protein solutions buffered at the desired pH. These buffered solutions presented a constant ionic strength of 0.002 M. Initially, the IgG concentration was in the 0 – 8 mg/m2 range and the final volume of the samples was 8 mL. Incubation was carried out in a thermostatic bath where samples were gently agitated at 25ºC for 12 h. After incubation, samples were centrifuged and the supernatant filtered using a polytetrafluoroethylene filter (Millipore, pore diameter equal to 100 nm). The protein concentration in solution was determined, before and after adsorption, by direct UV spectrophotometry at λ = 280 nm (ΣIgG = 1.40 mL mg-1 cm-1), and then, the adsorbed amounts were calculated by subtracting the final from the initial values.

2.5. Electrophoretic mobility

The electrophoretic mobility measurements were carried out with a Zeta-Sizer IV (Malvern Instruments). The particles were diluted in the desired buffered solutions – all of them with ionic strength values equal to 0.002 M – for 10 min just before measuring. Final particle concentration was equal to 3 x 109 mL-1. The mobility data were taken from the average of six measurements at the stationary level in a cylindrical cell.

2.6. Colloidal stability

The aggregation of the latex particles immersed in different saline media was measured using a low-angle light scattering technique. In these experiments NaCl and CaCl2 were used as aggregating salts. These experiments were carried out in an apparatus working with a He/Ne laser, using a rectangular scattering cell with a 2 mm path length, and measuring the light scattered at an angle equal to 10º. Equal volumes (1mL) of salt and latex solutions were mixed and introduced into the cell by an automatic mixing device. The initial particle concentration was set at 5 1010 mL-1, and the intensity (I) of the scattered light during aggregation was analyzed for 120 s. The linearity in the aggregation kinetics was relatively good at the beginning, and the initial slopes (dI/dt) were easily obtained for every experiment. That allowed us to estimate the stability ratio W, (also called Fuchs factor), which can be calculated from the following expression:

r

s

(dI / dt)W(dI / dt)

= (1)

86 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

where the (dI/dt)r term corresponds to the initial slope of rapid coagulation kinetics, while (dI/dt)s is the same parameter for a slow coagulation regime. The critical coagulation concentration (ccc) can be easily determined by plotting the logarithm of W versus the logarithm of the salt concentration and locating that point where log W reduces to zero.

2.7. Immunoassays

The optimum experimental conditions for the immunoassays were established in reference [22]. All the assays were carried out in a standard medium, namely, BSA saline buffer (pH 8.0 borate (13 mM), NaCl (150 mM), NaN3 (1 mg/ml) as preservative, and BSA in a 1 mg/ml concentration). This buffer approximately simulates the pH and ionic strength values usually found in physiological fluids. The role played by the BSA molecules is to cover possible PLGA or PS bare patches to avoid particle bridging produced by an unspecific adsorption of the antigen (CRP) molecules. In addition, the adsorption of these albumin molecules significantly helps to increase the colloidal stability of our IgG-latex complexes at pH 8, avoiding any potential aggregation caused by saline effects in the immuno-reaction medium. Immunoagglutination was detected by turbidimetry working with a spectrophotometer (DU 7400, Beckman) at a λ = 570 nm. 950 μL of an antiCRP-IgG-latex suspension in BSA saline buffer were quickly mixed with 50 μL of a CRP solution, ranging a final CRP concentration from 0.025 μg/mL to 10 μg/mL. The initial particle concentration for PLGA and PS was 4.9 1010 mL-1, and 2.3 1010 mL-1, respectively. It should be noted that samples were colloidally stable in the BSA saline buffer, and thus, only when the antibody-antigen recognition occurred, the aggregation of the particles began. The increase in turbidimetry was then monitored for different times. Subsequently, the immunoagglutination kinetics were analyzed by using a kinetic model that will be deeply described in the next section. The experimental results were fitted to the theoretical model using a commercial computer program (Origin 7.0, Microcal, MA, USA), in which a nonlinear least squares fitting method is applied.

III. RESULTS AND DISCUSSION

Despite PLGA and PS particles share the same superficial charged group, its hydrophobic/hydrophilic character may not coincide. It is well known that one of the most important driving forces that control protein adsorption is the hydrophobic interaction [23]. Therefore, before analysing the IgG adsorption, the relative hydrophobicity of PLGA and PS surfaces was estimated by means of contact angle measurements. A water drop was deposited on a flat surface made of dry and compressed PLGA (or PS) particles. The contact angle was measured by using a goniometric technique based on the ADSA algorithm described in reference [24]. As expected, polystyrene gave a higher contact angle (82º) than PLGA (74º), which means that the former possess a higher hydrophobic character than the latter. As will be shown afterward, this feature will be the main responsible of the differences observed in the protein adsorption experiments.

Paper III 87

The maximum amount of adsorbed IgG as a function of the medium pH was then studied. Figure 1 shows the obtained results when using independently PLGA and PS as adsorbent materials. Both curves are qualitatively similar, although significant differences appear quantitatively, since adsorption in PS approximately triples that found in the PLGA sample. As mentioned above, this could be caused by the different hydrophobic character shown by the two polymeric surfaces. It should be noted that IgG is a “soft” globular protein, for which hydrophobic interactions exert a great influence on adsorption [25,26]. It is suggested that hydrophobic interaction comes from the entropy increase when structured water molecules located in the vicinity of hydrophobic components are released to the bulk after the protein-surface contact [27]. Therefore, for a given soft protein, the higher the hydrophobicity of the substrate, the higher the protein adsorption. This explanation based on hydrophobic attraction also justifies the IgG

adsorption observed at pH ≥ 8, where electrostatic interactions between IgG molecules and particle surface are repulsive. Nevertheless, although hydrophobic interactions play an important role in the protein adsorption, it must be noted that electrostatic forces also participate, above all in low ionic strength media. For example, the maximum in the IgG adsorbed amount observed around pH 6, which almost coincides for both systems, is a result of electrostatic effects. The maximum position is usually obtained near the iep of the protein, although it is slightly shifted toward more acid pHs for negative adsorbent surfaces and toward more basic pHs for positive ones [28-30]. In fact, this maximum matches the isoelectric point of the protein-latex complex, which can be experimentally obtained by electrophoretic mobility measurements. At pH values far from the iep of the protein, the molecules

2 3 4 5 6 7 8 9 100

1

2

3

4

5

6

7

8

9

Γ ads (

mg/

m2 )

pH

Figure 1. Maximum amounts of polyclonal IgG adsorbed on PS ( ) and PLGA ( ) particles as a function of pH.

88 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

2 3 4 5 6 7 8 9 10-4

-3

-2

-1

0

1

2

3

4

are identically charged, and thus, they repel each other; in addition, they are more distended due to internal electrostatic repulsions, which subsequently give higher area per molecule values. Both contributions make the adsorption decrease. Besides, if the sign of charge of the adsorbent surface and the protein also coincide, the adsorption will be hindered even more. On the contrary, at pH 6 the sign of charge of particles and IgG differs, and thus, there is a favourable electrostatic attraction between the PLGA (or PS) surface and the IgG molecules. In addition, as this pH is near the iep of the protein, the inter- and intra-molecular repulsive interactions are still low; this is why the maximum adsorption is obtained at this pH value in both samples.

Table 1a. Critical coagulation concentrations (ccc) and critical stabilization concentrations (csc) of different IgG-PLGA complexes given in mM units. Data were collected using different pH and protein load values, which are indicated in the first column.

pH IgG load (in mg/m2)

NaCl ccc / csc

CaCl2 ccc / csc

4 0.7 70 / -- 45 / -- 4 1.7 90 / -- 65 / -- 6 1.9 <5 / -- <5 / 235 6 2.9 Aggr. Aggr. 8 0.9 150 / -- <5 / -- 8 2.5 90 / 275 <5 / 170

The electrophoretic mobility (μe) of the IgG-PLGA and IgG-PS complexes as a function of pH is shown in Figures 2a and 2b, respectively. These measurements were carried out with particles having different protein load. A hypothetical mobility for the IgG molecules has also been depicted (dotted line). Such IgG μe data have been obtained from interpolating the mobility results of a cationic and an anionic PS latex totally covered with polyclonal IgG [28]. The mobility of the bare

μ e 108 (m

2 /Vs)

pH

(a)

2 3 4 5 6 7 8 9 10-6

-5

-4

-3

-2

-1

0

1

2

3

4

μ e 108 (m

2 /Vs)

pH

(b)

Figure 2. Electrophoretic mobility versus pH for (a) IgG-PLGA and (b) IgG-PS particles. Bare particles ( ); intermediate protein coverage ( ): ΓPLGA = 1.9 mg/m2 and ΓPS = 2.1 mg/m2; high protein coverage ( ): ΓPLGA = 2.9 mg/m2 and ΓPS = 6.5 mg/m2. Dashed line: hyphotetical electrophoretic mobility of IgG molecules.

Paper III 89

Table 1b. Critical coagulation concentrations (ccc) and critical stabilization concentrations (csc) of different IgG-PS complexes given in mM units. Data were collected using different pH and protein load values, which are indicated in the first column.

pH IgG load IgG load

(in mg/m2) (in mg/m2) NaCl NaCl

ccc / csc ccc / csc CaCl2 CaCl2

ccc / csc ccc / csc 4 4 3.1 3.1 Aggr. Aggr. Aggr. Aggr. 4 5.2 350 / -- 225 / -- 6 2.1 110 / -- <5 / -- 6 6.5 55 / -- <5 / -- 8 2.1 460 / -- 18 / 255 8 5.1 50 / 215 <5 / 195

particles reflects the weak acid character of the carboxylic groups located on both PLGA and PS surfaces. Protonation at acid pHs produces charge cancellation which gives low mobility values. This feature is more patent in the PLGA particles than in the PS sample. When particles are covered with IgG, the mobility tends to approach to that expected for the pure protein, although such a value is not attained. The higher the protein load, the more the approximation. It is worthy to remark the low

μe values obtained for the complexes with high IgG coverage. This suggests that, if only the overlap of electric double layers (but not steric hindrance effects) participates in the colloidal stability of the particles, the stability must be low in the pH 4-9 range.

0,01 0,1 1

1

10

100

W

Salt concentration (M)

Figure 3. Stability factor versus salt concentration using NaCl (squares) and CaCl2 (circles) as aggregating electrolytes. Open symbols and solid symbols represent IgG-PLGA (2.5 mg/m2) and IgG-PS (5.1 mg/m2) particles, respectively. Solid lines serve to guide the eyetoward the corresponding ccc values. Dashes lines serve to locate the critical stabilizationconcentration (csc) in restabilization phenomena. Measurements were carried out at pH 8.

90 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

In another set of experiments, the colloidal stability was evaluated at pH 4 (where the complexes presented a positive net charge), 6 (at which they were practically uncharged) and 8 (negatively charged particles) using NaCl and CaCl2 as aggregating agents. Figure 3 shows a representative experiment where the stability factor of IgG-PLGA and IgG-PS particles as a function of salt concentration is depicted at pH8. The ccc values of different complexes at different pH values are shown in Table 1a and 1b for the PLGA and PS samples, respectively. Let us first discuss the results found in acidic media. At pH4, IgG-PLGA complexes had a low mobility (see Fig 2a), which would mean a relatively low stability, provided that the stability depended exclusively on the electric characteristics (ζ-potential) of the particles. As expected, the ccc values at pH4 correlates with the mobility data at this pH, that is, the stability was low (< 100 mM), having the IgG-PLGA complex with more adsorbed protein (1.7 mg/m2) higher stability than that complex with lower protein load (0.7 mg/m2). In addition, the aggregating power of CaCl2 is stronger than NaCl due to the double valence of calcium. The same reasoning can be applied to the IgG-PS complexes, where those particles with the highest protein load (5.2 mg/m2) had a significant stability, while those with an intermediate IgG coverage (3.1 mg/m2) were completely unstable even in absence of salt. This last pattern can be explained if pH4 would correspond to the iep of this complex, which is more than likely taking into account the data shown in Fig. 2b. It should be noted that the complexes used for the mobility and the stability experiments differ in the protein load (2.1 mg/m2 and 3.1 mg/m2, respectively), and this is why no iep equal to pH4 is directly observed in Fig. 2b. Nevertheless, considering the electrophoretic behaviour of the IgG-PS samples, it is plausible to expect an iep around pH4 for the 3.1 mg/m2 complex. The mobility data also serves to explain the stability patterns observed at pH 6. This pH corresponds with the iep of the IgG-PLGA particles, which is translated into an extremely low stability. However, the IgG-PS complexes are more stable, since its corresponding μe values differ from zero. It is worthy to highlight the important destabilizing effect caused by the divalent cation (calcium) when it acts as a counterion, giving ccc values lower than 5 mM. Finally, the ccc results obtained at pH 8 also can be justified according to the electrophoretic behaviour of the systems. Nevertheless, the most striking result observed at this basic pH8 is the restabilization phenomenon found at high electrolyte concentrations. This stability cannot be explained by the ζ-potential values of the particles, since at high salt concentrations the electrical double layer is totally compressed. This restabilization mechanism has a structural origin based on the hydration forces that appear in hydrophilic surfaces [31]. A well-known interpretation of these hydration forces is that a polar (hydrophilic) surface induces an ordering of the solvent which exponentially decays away from the surface. The hydration of the surface reduces the free energy of the system. An overlap of the ordered-solvent layers near the two mutually approaching surfaces creates a structural force. Partial dehydration of the surface polar groups and the ions located near the interface due to the mutual approach will lead to an increase in the system energy, which results in a repulsive force [32]. The adsorption of the protein

Paper III 91

turns the original PLGA or PS hydrophobic surface into a hydrophilic one, where the hydrophilicity degree is governed by the protein coverage. This is why restabilization does not exist in bare latex particles (see Figure 4a and 4b for PLGA and PS, respectively) or in those complexes with a low IgG load. It should be noted that hydration forces does not only depend on the surface hydrophilicity, but also on the nature and concentration of the hydrated counterions that surround the particles. The critical stabilization concentration (csc) is defined as the minimum salt concentration at which restabilization starts. These csc data are also shown in Table 1. As cations are more hydrated that anions, hydrations forces are usually observed in negatively charged systems, where the cations acts as counter ions. This is why restabilization is observed at pH 8 but not at pH4. Finally, as calcium is a more hydrated cation than sodium, the restabilization power of the former is higher than that of the latter. Note that Ca2+ even restabilizes an uncharged IgG-PLGA sample (see csc data at pH6). Different authors have exhaustively studied this kind of hydration forces in colloidal systems; the interested reader can be found such information elsewhere [33-38].

In order to know if the adsorption process denatures the protein molecules – which would make the IgG-PLGA complexes completely useless for targeting purposes – the immunoreactivity of the adsorbed antibodies was performed. This kind of experiments allows us to test possible denaturation or unfavourable orientation of the molecules on the PLGA surface. Immunoreaction was then evaluated using CRP as a model target molecule and antiCRP-IgG-PLGA (and -PS) as reactive nanoparticles. In these experiments, the adsorption of antiCRP-IgG onto the particles was performed at pH8. The IgG adsorbed amounts were 1.0 mg/m2 and 2.1 mg/m2 for the PLGA and PS particles, respectively. It should be noted that immunoassays based on latex technology are usually performed at pH and ionic strength values reproducing the physiological conditions (pH ≈ 7.5 and I ≈150-170

Figure 4a. Stability factor of bare PLGA particles. NaCl at pH4 ( ) ccc = 85 mM; NaCl at pH8 ( ) ccc = 200 mM; CaCl2 at pH4 ( ) ccc = 20 mM; and CaCl2 at pH8 ( ) ccc = 30 mM.

0,01 0,1 1

1

10

100W

Salt concentration (M)0,01 0,1 1

1

10

100

W

Salt concentration (M)

Figure 4b. Stability factor of bare PS particles. NaCl at pH4 ( ) ccc = 550 mM; NaCl at pH8 ( ) ccc = 800 mM; CaCl2 at pH4 ( ) ccc = 15 mM; and CaCl2 at pH8 ( ) ccc = 35 mM.

92 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

mM). Under these conditions our IgG-PLGA (or -PS) particles were colloidally unstable when they are totally coated with antiboby (see in Table 1a the ccc value at pH8 when PLGA particles have an IgG load equal to 2.5 mg/m2, or see Table 1b for the 5.1 mg/m2 complex in the PS case). A false positive would be obtained if the system aggregates in absence of the specific antigen. Therefore, it is necessary to

work with stable particles that do not spontaneously coagulate at slightly basic pH and at 150 mM ionic strength media. There are different strategies to improve the stability of antibody-latex complexes, as reviewed in reference [21]. One of the possible solutions is to co-adsorb a stabilizing molecule (a surfactant, a lipid or a protein) together with the antibody. The immunological reactivity depends on the latter, whereas the colloidal stability is a function of the former. This is why we have used in our experiment PLGA and PS particles partially coated by IgG (approximately 50 % of the surface), instead of using nanospheres fully coated by antibody. In this way, there are PLGA and PS patches (free of IgG) available to adsorb the stabilizing molecule (BSA in our case). Therefore, the 1.0 mg/m2 and 2.1 mg/m2 for the IgG loading on PLGA and PS particles respectively allowed us to obtain stable particles (due to the spontaneous BSA adsorption on the surface patches when particles are immersed in the reactive medium: BSA saline buffer), but maintaining a good immunoreactivity. Figure 5 shows a typical immunoaggregation experiment. It should be noted that our particles were stable in the pure BSA saline buffer, and thus, they did not agglutinate in absence of CRP; therefore, they only collapsed due to the bridging effect caused by the CRP after the corresponding antiCRP-IgG recognition. According to Fig. 5, the immunoagglutination was high at intermediate CRP concentrations, but low at

0 10 20 30 40 50 600,35

0,40

0,45

0,50

0,55

0,60

0,65

0,70

Abs

orba

nce

time (min)

Figure 5. Immunoagglutination kinetics of antiCRP-IgG-PLGA particles for different CRP concentrations (in mg/l): 0.025 ( ); 0.05 ( ); 0.1 ( ); 0.15 ( ); 0.2 ( ); 0.25 ( ); 0.5 ( ); 1.0 ( ); 2.0 ( ); 3.0 ( ); 4.0 ( ); 5.0 ( ); 6.0 ( ); and 10.0 ( ).

Paper III 93

high and low CRP concentrations. This feature becomes evident in Figures 6a and 6b, where the absorbance increment versus CRP concentration is plotted for different time intervals. A typical bell-shape curve (also called precipitine curve) is obtained, which depends on the molar ratio of antigen to antibody [39,40]. For low CRP concentrations, few antigen molecules are available to bridge the IgG-latex

particles, and thus, small agglutination kinetics are obtained. On the other hand, at high CRP concentrations, the active sites of the antibodies are rapidly saturated with different CRP molecules, and thus the bridging process, which only can take place when the same antigen molecule is recognized by IgG molecules adsorbed in two different particles, is hindered. A maximum in agglutination efficiency exists between these two extreme regimes. It is observed that IgG-PS complexes are more reactive at low CRP concentrations than those of PLGA. This feature has a simple explanation: the amount of IgG molecules adsorbed on the PS doubles that of PLGA, and thus, the immunosensitivity must be also higher, provided that the orientation of the IgG molecules was similar in both samples.

The immunoassay results can also allow us to estimate the percentage of active IgG on the latex particles. This can be performed by using a simple kinetic model for antigen-antibody reactions in particle-enhanced immunoassays developed by Quesada et al. [41-43]. The model proposed by these authors can be applied to immunoagglutination processes detected by nephelometry [41, 42] or by spectrophotometry [43], although the basic equations slightly change depending on the optical technique. The model is based on the La Mer’s idea [44, 45], that says that the rate of agglutination must be equal to the product of the particle collision frequency and a collision efficiency factor which actually leads to aggregation of the particles (that is, to antibody-antigen binding). The model proposed by Quesada et al. required some hypothesis to find a simple relationship between the kinetics

0,1 1 100,00

0,05

0,10

0,15

0,20

0,25

0,30

ΔAb

s

CRP (mg/l)

(a)

0,1 1 100,00

0,05

0,10

0,15

0,20

0,25

0,30

ΔAb

s

CRP (mg/l)

(b)

Figure 6. Absorbance increment for the (a) IgG-PLGA and (b) IgG-PS systems versus CRP concentration for different time intervals (in min): 5 ( ); 10 ( ); 15 ( ); 25 ( ); 30 ( ); 35 ( ); 40 ( ); 45 ( ); 50 ( ); 55 ( ); and 60 ( ).

94 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

constant of aggregation and the initial concentration of antigen into the reaction buffer. The model can be summarized as follows. Note that this is only a resume, and thus, details must be found elsewhere [41-43]. The model conceptually divides the immunoaggregation process into two steps. In the first one, the antibody-latex particles are in the solution and the antigen molecules are added; then, these last molecules become distributed between the solution and the antibody sites reaching an equilibrium state before latex flocculation. In the second step, the formation of bridges between spheres through antigen molecules begins. Although this simplistic mechanism could be questionable, it can be justified by the fact that protein (CRP) molecules are much smaller than latex nanospheres, so the former diffuse much faster than the latter. The master equation that relates the immunoagglutination kinetics (that is, the initial slope of the curves shown in Fig. 5 – dAbs/dt –) with the total number of immunologically active IgG molecules per particle (n) is shown below:

0 0dAbs C 1 1N n A N n A

dt 4 K K⎛ ⎞⎛= + + − Δ − − +⎜ ⎟⎜⎝ ⎠⎝

⎞Δ ⎟⎠

(2)

where

2

01N n A 4N nAK

⎛ ⎞Δ = + + −⎜ ⎟⎝ ⎠

0

0,0 2,0x1013 4,0x1013 6,0x10130,0

3,0x10-6

6,0x10-6

9,0x10-6

1,2x10-5

1,5x10-5

1,8x10-5

(3)

dAbs

/dt (

min

-1)

[CRP] molec/ml

Figure 7. Initial slope of the immunoagglutination kinetics versus CRP concentration for thePLGA-IgG sample. Squares are experimental data, while the solid line represents thetheoretical fitting that comes from applying the kinetics model. The “C”, “K” and “n” values used for this fitting are given in Table 2.

Paper III 95

In equation 2, N0 is the initial concentration of latex particles, A is the initial concentration of antigen in the solution, K is the equilibrium constant describing the interaction of the free antigen with the antibody adsorbed onto the latex surface, and C is a parameter related to some optical characteristics and other factors (i.e. the particle diffusion coefficient, the electrostatic repulsion due to the overlap of the diffuse double layers of antigens and sensitized latex beads, the attractive interaction due to the London - van der Waals forces, etc) [43]. Equation 2 depends on three fitting parameters: n, K and C. Figure 7 shows the experimental immunoagglutination results (symbols) of the PLGA sample fitted by equation 2 (solid line). As mentioned in Materials and Methods section, the experimental data were adjusted to equation 2 by using the Origin 7.0 programme, which used a nonlinear least squares fitting algorithm with three independent variables (n, K and C). The best fitting for the PLGA system was obtained with the n, K and C values shown in Table 2. The same analysis was performed for the PS sample, and the corresponding values are also shown in Table 2. The antigen – antibody equilibrium constant K must be equal in both cases, as obtained. In addition, the estimated K value is of the same order of magnitude as those found previously for other polyclonal IgG immunoreactions [41-43,46]. Nevertheless, the most interesting parameter to discuss is the number of active IgG molecules per particle (n). This parameter is higher for the PS sample than for the PLGA one, which agrees with the better inmunoresponse shown by the former (see Fig. 6b, specially at low CRP concentrations) when compared to the latter (Fig. 6a). In

Table 2. Fitting parameters of the kinetics model and percentage of active IgG molecules on PLGA and PS particles.

C/4 (10-28abs units-1 ml2

molecules-2)

n (molecules/particle)

K-1

(1012 molecules ml-1)

% active IgG

PLGA 1.1 ± 0.6 21 ± 6 1.05 ± 0.20 3.8 ± 1.4 PS 1.9 ± 1.7 35 ± 5 1.05 ± 0.20 2.0 ± 0.3

addition, taking into account the particle surface area and the protein coverage, it is possible to calculate the percentage of active IgG molecules per particle. For example, a 210 nm diameter spherical PLGA particle possesses an area equal to 1.38 10-13 m2. Our IgG coverage (1.0 mg/m2) can be translated into molecules per square meter – by knowing the antibody molecular weight (150000 Da) and the Avogadro number – which was 4.015 1015 molecules/m2. Combining this last value with the calculated area per particle, one can determine the total number of IgG molecules per particle, which was 555 molecules/particle. According to the theoretical fitting, the number of active IgG molecules per particle – that is, the n parameter – was 21. This means that only around 4% of the total polyclonal IgG molecules are active and properly orientated outward for the PLGA sample. In the PS case, this results was even lower: 2%. Despite the n parameter indicates nPS > nPLGA, a lower value for the active IgG molecules in the PS particles is plausible. There are two factors that would contribute to reduce the percentage of reactive molecules in the PS particles.

96 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

i) On the one hand, the higher hydrophobicity of PS makes the surface-IgG hydrophobic interactions more intense, and thus, deformation of the adsorbed antibodies is favoured [23]; this partial denaturation reduces the native immunoreactivity. ii) On the other hand, a high adsorption coverage – which is favoured in the hydrophobic PS surface – also contributes to diminish the percentage of active molecules. A crowded polyclonal IgG layer hinders the union of a voluminous antigen (as CRP, which has a MW = 114.000 Da) to any active IgG, simple by steric impediments caused by the presence of non-reactive IgG molecules placed at the vicinity of the active one. Nevertheless and despite these low values, the reactivity of both types of complexes was very good, as shown in Figs. 6a and 6b. This means that it is possible to develop drug delivery systems with appropriate targeting properties using few active IgG molecules per particle. This information could become very useful when monoclonal active IgG molecules instead of polyclonal ones were used; at least, the cost for sensitizing the nanocarriers with the antibody can be reduced, since few IgG molecules would be needed to produce stimuli responsive particles. In addition, the colloidal stability of the system can be easily improved in this situation, since there would be a lot of available surface to be coated by any stabilizer molecule, which would increase the poor colloidal stability observed with these particles when they are totally coated by polyclonal antibodies. It should be noted that coadsorbing other molecules (surfactants, lipids, proteins) that act as stabilizer agents together with the IgG is a general strategy used to improve the stability of antibody-latex particles [21]. At the present, we are using a biocompatible non-ionic surfactant – a poloxamer (Pluronic® F68) – to enhance the stability of our IgG-latex complexes. These preliminary results obtained with a ternary system (PLGA-IgG-poloxamer) will be presented in a future paper.

IV. CONCLUSIONS

PLGA nanoparticles have been sensitized with IgG molecules in order to experimentally simulate a simple drug delivery system with targeting ability. The corresponding electrophoretic mobility and colloidal stability have been subsequently studied. It has been proved that adsorption of polyclonal IgG produces a reduction of the stability of the particles around physiological pHs. Nevertheless, this handicap could be overcome if coadsorption of IgG and any stabilizer molecule was performed. In addition, the potential targeting ability has been quantified by measuring the immunoreactivity of the IgG-PLGA complexes. The analysis of the results by using a simple theoretical model has allowed us to estimate the percentage of active IgG molecules per particle. A low value around 4% has been found, which is acceptable taking into account that we have worked with a polyclonal sample. The obtained results can be useful to design future nanocarriers vectorized with IgG molecules in order to deliver drugs to specific target cells. At least, according to the results shown in this paper, a potential delivery system based on PLGA can be prompted: The use of monoclonal IgG molecules is advised, as all of them will be immunologically active against the same

Paper III 97

target molecule. In this case, low antibody coverage may be enough to achieve a PLGA carrier with high targeting ability. Subsequently, the bare PLGA patches must be coated by a non-ionic surfactant (i.e. Pluronic® F68). In this way, it would be possible to obtain immunoreactive stable particles that would also avoid recognition by the MPS cells.

ACKNOWLEDGEMENTS

Financial support from the “Comisión Interministerial de Ciencia y Tecnología” Project MAT2003-01257 (European FEDER support included) and from the “Consejería de Innovación, Ciencia y Tecnología de la Junta de Andalucía” Project FQM 392 is gratefully acknowledged.

98 J. Colloids and Surface B: Biointerfaces 60, 1 (2007) 80-88.

References

[1] S.J. Douglas, S.S. Davis and L. Illum, CRC Crit. Rev., 3 (1987) 233.

[2] M.J. Alonso, in S. Cohen and H. Bernstein (Eds.), Microparticulate Systems for the Delivery of Proteins and Vaccines, Marcel Dekker Inc., New York, 1996, Chapter 7.

[3] D.H. Lewis, in M. Chasin and R. Langer (eds.), Biodegradable Polymers as Drug Delivery Systems, Marcel Dekker Inc., New York, 1990, Chapter 1.

[4] A. Schade, T. Niwa, H. Takeuchi, T. Hino and Y. Kawashima, Intern. J. Pharm., 117 (1995) 209.

[5] M.D. Blanco and M.J. Alonso, Eur. J. Pharm. Biopharm., 43 (1997) 287.

[6] T. Govender, S. Stolnik, M.C. Garnett, L. Illum and S.S. Davis, J. Control. Release, 57 (1999) 171.

[7] R. Graf, M. Luck, P. Quellec, M. Marchand, E. Dellacherie, S. Harnisch, T. Blunk and R.H. Muller, Colloids Surf. B, 18 (2000) 301.

[8] Y.P. Li, Y.Y. Pei, X.Y. Zhang, Z.H. Gu, Z.H. Zhou, W.F. Yuan, J.J. Zhou, J.H. Zhu and X.J. Gao, J. Control. Release, 71 (2001) 203.

[9] K. Avgoustakis, A. Beletsi, Z. Panagi, P. Klepetsanis, A.G. Karydas and D.S. Ithakissios, J. Control. Release, 79 (2002) 123.

[10] J. Panyam and V. Labhasetwar, Adv. Drug Delivery Rev., 55 (2003) 329.

[11] N. Csaba, L. González, A. Sánchez and M.J. Alonso, J. Biomater. Sci. Polym. Edn., 15 (2004) 1137.

[12] S.S. FENA, L. Mu, K.Y. Win and G.F. Huang, Current Med. Chem., 11 (2204) 413.

[13] M.L.T. Zweers, G.H.M. Engbers, D.W. Grijpma and J. Feijen, J. Control. Release, 100 (2004) 347.

[14] C.E. Astete and C.M. Sabliov, J. Biomater. Sci. Polym. Edn., 17 (2006) 247.

[15] E. Ricci-Junior and J.M. Marchetti, J. Microencapsulation, 23 (2006) 523.

[16] M.J. Santander-Ortega, A.B. Jódar-Reyes, N. Csaba, D. Bastos-González and J.L. Ortega-Vinuesa, J. Colloid Interface Sci., 302 (2006) 522.

[17] S.H. Kim, J.H. Jeong, K.W. Chun and T.G. Park, Langmuir, 21 (2005) 8852.

[18] A. Kondo, T. Kawano, F. Itoh and S. Higashitani, J. Immunol. Methods, 135 (1990) 111.

[19] C.P. Price and D.J. Newman, Principles and Practice of Immunoassay, Stockton Press, New Cork, 1991.

[20] J.L. Ortega-Vinuesa, J.A. Molina-Bolívar, J.M. Peula and R. Hidalgo-Álvarez, J. Immunol. Methods, 205 (1997) 151.

[21] J.L. Ortega-Vinuesa and D. Bastos-González, J. Biomater. Sci. Polym. Edn., 12 (2001) 379.

[22] J.L. Ortega-Vinuesa, J.A. Molina-Bolívar, and R. Hidalgo-Álvarez, J. Immunol. Methods, 190 (1996) 29.

[23] W. Norde, Adv. Colloid Interface Sci., 25 (1986) 267.

[24] O.I. del Río and A.W. Newmann, J. Colloid Interface Sci., 196 (1997) 136.

Paper III 99

[25] R.D. Milton and C.R. Robertson, Langmuir, 7 (1991) 2710.

[26] P. Warkentin, B. Wälivaara, I. Lundström and P. Tengvall, Biomaterials, 15 (1994) 786.

[27] J.N. Israelachvili, Intermolecular and Surface Forces, Academic Press, London, 1992.

[28] F. Galisteo-González, J. Puig, A. Martín-Rodríguez, J. Serra-Domènech and R. Hidalgo-Álvarez, Colloids Surf. B, 2 (1994) 435.

[29] J. Buijs, J.W.T. Lichtenbelt, W. Norde and J. Lyklema, Colloids Surf. B, 47 (1995) 633.

[30] J.L. Ortega-Vinuesa, M.J. Gálvez-Ruiz and R. Hidalgo-Álvarez, Langmuir, 12 (1996) 3211.

[31] J.N. Israelachvili, Intermolecular and Surface Forces, Academic Press, London, 1992.

[32] S. Leikin and A.A. Kornyshev, Phys Rev. A, 44 (1991) 2.

[33] J.N. Israelachvili and G.E. Adams, J. Chem. Soc. Faraday Trans., 74 (1978) 975.

[34] R.M. Pashley, Adv. Colloid Interface Sci., 16 (1982) 57.

[35] J.A. Molina-Bolívar, F. Galisteo-González and R. Hidalgo-Álvarez, Colloids Surf B, 21 (2001) 125.

[36] N.V. Churaev and B.V. Derjaguin, J. Colloid Interface Sci., 103 (1985) 542.

[37] J.A. Molina-Bolívar, F. Galisteo-González and R. Hidalgo-Álvarez, Phys. Rev. E, 55 (1997) 4522.

[38] J.A. Molina-Bolívar and J.L. Ortega-Vinuesa, Langmuir, 15 (1999) 2644.

[39] M. Heidelberger and F.W. Kendall, J. Exp. Med., 62 (1935) 467.

[40] R.P.O. Tengerdy, J. Immunol. Methods, 99 (1967) 126.

[41] M. Quesada, J. Puig, J.M. Delgado, J.M. Peula, J.A. Molina and R. Hidalgo-Álvarez, Colloids Surf. B, 8 (1997) 303.

[42] J.A. Molina-Bolívar, F. Galisteo-González, M. Quesada-Pérez and R. Hidalgo-Álvarez, Colloid Polym. Sci. 276 (1998) 1117.

[43] M. Quesada, J. Puig, J.M. Delgado, and R. Hidalgo-Álvarez, J. Biomater. Sci. Polym. Edn., 9 (1998) 961.

[44] R.H. Smellie and V.K. La Mer, J. Colloid Sci., 23 (1958) 589.

[45] T.W. Healy and V.K. La Mer, J. Phys. Chem., 66 (1962) 1835.

[46] J.L. Ortega-Vinuesa, R. Hidalgo-Álvarez, F.J. de las Nieves, C.L. Davey, D.J. Newman and C.P. Price, J. Colloid Interface Sci., 204 (1998) 300.

Nanoparticles made from Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers.

Part 1

M.J. Santander-Ortega1,2, T. Stauner3, B. Loretz2, Ortega-Vinuesa, J.L.1, Bastos-González, D.1, G. Wenz3, U.F. Schäfer2 and C.M. Lehr2

1Department of Applied Physics, University of Granada, Granada (SPAIN)

2Department of Biopharmaceutics and Pharmacological Technology, Saarland University, Saarbrücken (GERMANY)

3Department of Organic Macromolecular Chemistry, Saarland University, Saarbrücken (GERMANY)

102 Paper IV

ABSTRACT

The goal of this paper was aimed to the formulation of nanoparticles by using two different amphiphilic polymers –referred to as P1 and P2- presenting P2 a higher hydrophobic character. A simple o/w emulsion diffusion technique, avoiding the use of hazard solvents such as dichloromethane or dymethyl sulfoxide, was chosen to formulate nanoparticles with both polymers, producing the P1 and P2 nanoparticles. Once the nanoparticles were prepared, a deep physicochemical characterization was carried out, including the stability of the o/w emulsion during nanoparticle formation, nanoparticles stability once they were formed or swelling properties. Thanks to this characterization it was possible to understand that the hydrophobic character of each polymer was crucial in the formation and posterior properties displayed by each nanoparticle system.

Paper IV 103

I. INTRODUCTION

The aim of this work was the development of a promising nanoparticle system formulated with amphiphilic polymers, exploiting the advantages of these polymers, but using organic solvents with reduced hazard properties. Three polymers one hydrophilic P0 and the other two, amphiphilic polymers P1 and P2, synthesized by the insertion of hydrophobic moieties in the first one were used as basic constituents for the preparation of nanoparticles. Once nanoparticles were formulated, a thorough physicochemical characterization was carried out.

II. EXPERIMENTAL PART

2.1. Preparation of the Nanoparticles.

Nanoparticles with the different polymers were formulated by a simple o/w emulsion diffusion method. Briefly, specific polymer (P0, P1 or P2) was dissolved in an organic solvent and this organic solution was poured on an aqueous phase with different percentages (w/v) of a surfactant (0, 0.1, 0.5 and 1). This biphasic system was emulsified with a high speed homogenizer (Ultra Turrax® Ika®, Brasil Ltda, Taquara, Brasil). Then, MilliQ water was added to force the complete diffusion of the organic solvent to the aqueous phase. Finally, the organic solvent was evaporated under vacuum at 35ºC (Rotavapor Büchi®, Labortechnik AG, Flawil, Switzerland), generating stable nanoparticles. After nanoparticles preparation, MiliQ water was added to obtain a colloidal solution with a final volume of 10 ml.

2.2. Characterization of the Nanoparticles

Size and ζ–potential of the nanoparticles were analysed by photon correlation spectroscopy (PCS) using a Nano-ZS (Malvern Instruments, Malvern, UK). For ζ–potential measurement nanoparticles were diluted in NaCl 3mM. AFM images were obtained using an Atomic Force Microscopy Nanoscope IV BioscopeTM (Veeco Instruments, Santa Barbara, CA, USA). Imaging was done using Taping mode and a silicon cantilever with a spring constant of approximately 40 N/m and a resonance frequency of about 170 kHz. The scan speed applied was 0.2 Hz.

2.3. Stability of the O/W emulsion

In order to know the role of each component in the formation of the nanoparticles, the stability of the o/w emulsion was analyzed by PCS using a Turbiscan Classic MA 2000 (Formulaction, L’Union, France) as a function of its composition. The stability of this emulsion was determined for three different organic phases, pure organic solvent and P1 or P2 dissolved in organic solvent. Moreover, the effect of the surfactant in the aqueous phase was also tested. In order to obtain a better understanding of these results, the main characteristic of the turbiscan technique will be explain briefly. To determine the emulsion stability, samples of 5ml prepared as was commented above were placed in a cylindrical glass cell, which was inserted in the turbiscan device. In this instrument the light

104 Paper IV

source is an electro luminescent diode in the near infrared (λ=850nm). Two synchronous optical sensors receive respectively transmitted light (T) through the sample (180° from the incident light), and backscattered light (BS) by the sample (45° from the incident radiation). The detection head scanned the entire height of the sample, acquiring T and BS data at 40 µm height intervals at different time up to, at least, 10 h. The principle of the measurement was based on the variation of the droplet volume fraction (migration) or diameter (coalescence), resulting in a variation of T and BS signals [1-3].

2.4. Colloidal stability

Stability of the nanoparticles was studied as a function of salinity of the medium using NaCl and CaCl2 as aggregating salts. According to the classical DLVO theory [4], a salinity increment triggers the aggregation of lyophobic colloidal systems. Particle aggregation was analyzed by photon correlation spectroscopy (PCS) using a 4700c System (Malvern Instruments). The PCS instrument had a Helium laser (λ = 632 nm) with perpendicular polarization and a power rating of 35mW. After 0.3 ml of sample was poured into a cylindrical cell, 0.3 ml of the saline solution at the desired ionic strength was added and rapidly mixed. The computer software analyzed the scattered-intensity auto-correlation function measured at 60º. The aggregation measurements lasted around 10 min. For information on the aggregation kinetics, the average diameter of the particles was plotted against time. The slopes of these curves (∂d/∂t) enabled determinations of the aggregation rate (k) for the different systems and hence calculations of the stability or Fuchs factor (W) defined by:

( )( )

/

/f f

ss

d t kW

d t k

∂ ∂= =

∂ ∂ (1)

where the rate constant kf corresponds to a fastest aggregation kinetic, and ks is the rate constant for a slower aggregation regime. Therefore, when W=1 the colloidal system is totally unstable, while W ∞ indicates a stable colloidal system. Plotting W as a function of the medium salinity in a double-logarithmic scale becomes very useful to estimate the critical coagulation concentration (ccc) –the point where W reduces to 1- and the critical stabilization concentration (csc) –the point where W increase from 1 to ∞ when salt concentration increases even more- which are fundamental parameters in colloidal stability studies. The ccc value- defined as the minimum salt concentration needed for a rapid aggregation- is related to destabilization processes; a low ccc means low stability. However, the csc value –defined as the minimum salt concentration at which the system begins to re-stabilize when salinity is increased even more– is associated to the surface hydrophilicity. This kind of re-stabilization phenomenon at high salt concentration is well known, and it is governed by hydration forces [5-10]. When using a given electrolyte, the lower the csc, the higher the hydrophilicity of the particle surface.

Paper IV 105

2.5. Swelling studies

The extent of swelling of these nanoparticles was determined by osmotic effect as a function of ionic strength using NaCl, where the salt concentration in the bulk solution of the suspension was varied from 0.01 to 5 mM NaCl. The extent of swelling was determined from the change in the hydrodynamic diameter using PCS. It is experimentally convenient to measure swelling changes relative to the fully swollen hydrodynamic diameter d0. Hence, the extent of particle de-swelling is expressed as the de-swelling ratio (α) [11], which is simply:

3

0

dd

α⎛ ⎞

= ⎜ ⎟⎝ ⎠

(2)

where d is the measured hydrodynamic diameter at a given ionic strength.

2.6. Stability of stored nanoparticles suspension at 4ºC and 25ºC

Once both nanoparticles systems, P1 and P2, were formulated, one sample of these systems was stored at 4ºC and 25ºC and their size was measured by PCS at regular time period up to 25 days.

2.7. Cell culture

Caco-2 cell line clone C2BBe1 (ATCC No CRL-2102) between passage no 64 and 70 was used as test system. DMEM No 41965 supplemented with 1% Non-essential amino acids (both from Gibco, Karlsruhe, Germany) and 10 % FBS (F7524, Sigma-Aldrich, Taufkirchen, Germany) was used as growth medium. Subculture was done once a week at a subcultering ratio of 1:10 using trypsin-EDTA (No 25300, Gibco).

2.8. Lactat dehydrogenase (LDH) assay

Cytotoxicity detection Kit (LDH) from Roche (Mannheim, Germany) was used to assess the cell membrane damage induced by exposure to these nanoparticles. 5x104 Caco-2 cells were seeded into 24-well plates (Greiner, Frickenhausen, Germany). After 6 days propagation, at the state of a confluent monolayer the toxicity assay was performed. Cells were washed once with phosphate buffer saline (PBS) and incubated with the test samples. Samples were prepared by dilution of aqueous suspension of the nanoparticles with cell culture medium at ratios 1:3, 1:6, 1:12 and 1:24 resulting in concentrations of 0.0042 to 0.033mg/ml nanoparticles. As negative control cells were incubated with cell culture medium and as high control with medium plus 1 % of the detergent Trition-X100. Samples of 50 µL of supernatant were taken after 4 h or 24 h incubation time and transferred into a 96 well microtiter plate. Reaction mix was prepared following the manufactures protocol and mixed in a 1:1 ratio with the samples. After 10 min incubation on a shaker in a dark room the absorbance at λ = 492 nm was measured. Cytotoxicity was calculated according to the equation:

106 Paper IV

( )(exp )(%) 100 v c

c c

lCytotoxicityh l

−=

− (3)

where expv is the experimental value, lc is the low control and hc is the high control.

2.9. MTT assay

Thiazolyl Blue Tetrazolium Bromide (M5655, Sigma Aldrich) was dissolved in PBS pH 7.4 to yield a final concentration of 5 mg/ml for the stock solution.

Cells were seeded, maintained until reaching confluence and exposed to particle suspension as reported for LDH assay. After the incubation periode of 4 h the particle suspension was removed. Cells were washed once with PBS before fresh medium and 50µl MTT stock solution per well was added. After further 3 h incubation 500 ml lysis buffer (10% SDS in 0.01mM HCl) were added to lyse the cells and solubilize the tetrazolium crystals. The aborbance at λ = 550 nm was analysed in a plate reader. Viability was calculated in comparison to the positive control, untreated cells as 100% value, and negative control 1%-Trition solution as 0% value. For direct comparison with LDH assay results this viability values were transformed in toxicity values by subtraction of the viability value from the 100% control.

1 10 100 1000

0

5

10

15

20

25

Inte

nsity

(%)

Size (nm)

Figure 1. Hydrodynamic size distribution of P1 nanoparticles formulated with different percentages (w/v) of surfactant, (dotted line) 0%, (dash-dotted line) 0.1%, (dashed line) 0.5%, (solid line) 1%, (n≥3).

Paper IV 107

III. RESULTS

3.1. Characterization of the Nanoparticles

Figure 1 shows the hydrodynamic size distribution of nanoparticles formulated with P1 polymer and different amounts of surfactant. Similar results were obtained for P2 polymer (data not shown). However, in the case of non-amphiphilic polymer the formation of nanoparticle could not be observed.

As can be observed in Figure 1, the incorporation of the surfactant to the formulation improved both the nanoparticle hydrodynamic size distribution and the polydispersity index (PDI). A mono-modal size distribution was achieved when the aqueous phase contains a surfactant concentration equal or higher than 0.5% (w/v). The narrowest size distribution was reached when the aqueous phase presented a 1% (w/v) of surfactant. Consequently, nanoparticles formulated with an aqueous phase of 1% (w/v) of surfactant (P1 and P2 nanoparticles) were selected for the next studies. Table 1 summarizes the mean hydrodynamic size and PDI of these nanoparticles. P1 nanoparticles exhibited a lower hydrodynamic mean size than P2. As is illustrated by AFM images (Figures 2a and 2b), P1 and P2 nanoparticles showed a spherical shape with a narrow size distribution.

Table 1. Hydrodynamic mean size (nm), PDI and ζ-potential (mV) of P1 and P2 nanoparticles.

Size (nm) PDI ζ-Pot. (mV) P1 150.5 ± 3.5 0.12± 0.01 -8.3 ± 0.3 P2 182.7 ± 6.7 0.08 ± 0.02 -5.8 ± 0.5

Regarding to the negative ζ–potential values displayed by both colloidal systems, it is worthy to remark that neither amphiphilic polymers nor the surfactant present ionizable chemical groups. Hence, the low and negative ζ–potential summarized in Table 1 can be attributed to the specific interaction of the nanoparticles surface with the ions present in the medium (as ζ–potential measurements were developed in an aqueous solution of NaCl 3mM) [12].

3.2. Stability of the O/W emulsion

The role of each component in the final nanoparticle characteristics was analyzed by the study of the emulsion stability as a function of the components presented in the organic or aqueous phase. With this idea in mind the stability of five different emulsions was analyzed. In the first two of them, the aqueous phase presented only water while the organic phase was formed by the correspondent amphiphilic polymer, P1 or P2, and the organic solvent. In the third emulsion the organic phase was composed only by the organic solvent, while a solution of the surfactant 1% (w/v) was the aqueous phase. The last two emulsions were prepared with an organic phase of P1 or P2 polymer dissolved in the organic solvent and an aqueous phase of the surfactant 1% (w/v). Variation of transmission, ΔT(%), for these emulsions is shown in Figure 3a, while the variation of backscattered light ,

108 Paper IV

ΔBS (%), is shown in Figure 3b. From Figure 3a, the migration of the organic phase droplets through the aqueous phase could be obtained by the estimation of the slope of the linear part of the graphic (ΔT(%)/min), while from the Figure 3b it is possible to estimate the coalescence between the organic phase droplets [1-3]. The influence exerted by the macromolecule dissolve in the organic or aqueous phase will be commented in first place, while stability of the emulsion with a surfactant 1% (w/v) aqueous phase and an organic phase composed by P1 or P2 dissolved in the organic solvent will be commented afterward. P1 polymer presented a low speed of

Figure 2. AFM images of P1 (a) and P2 (b) nanoparticles.

(a)

(b)

Paper IV 109

phase separation (ΔT(%)/min ≈ 0.09), Figure 3a. However, Figure 3b illustrates a coalescence process of the organic phase drops (ΔBS ≈ 30%). P2 polymer showed worst surfactant properties than P1. In this case phase separation speed (ΔT(%)/min ≈ 0.18) and coalescence (ΔBS ≈ 40%) were higher than for the former. The highest phase separation speed (ΔT(%)/min ≈ 0.35) was obtained when a dissolution of the surfactant 1% (w/v) was used as aqueous phase and the organic solvent as organic phase. Nevertheless, this emulsion showed a low coalescence (ΔBS ≈ 10%). This means that the presence of this surfactant in the aqueous phase produce emulsions with a “fast” migration of the organic phase drops through the aqueous phase, but these droplets present a low coalescence [1-3].

With respect to the stability of the emulsions with an aqueous phase of the surfactant 1% (w/v) and an organic phase of P1 or P2 dissolved in the organic solvent, similar tendencies were obtained with respect to the previous experiments. Emulsions with P1 dissolved in the organic phase presented a lower phase separation speed. However, both emulsions, with P1 or P2 dissolved in the organic phase, presented a close to zero ΔBS values. This means that the incorporation of the surfactant to the aqueous phase prevented the coalescence between the organic phase droplets[1-3].

3.3. Colloidal stability

Stability of P1 and P2 nanoparticles as a function of the electrolyte concentration using NaCl and CaCl2 as aggregating salts was determined. Figure 4 illustrates the value of the Fuch factor (W) as a function of the salt concentration; the corresponding values of critical coagulation concentration (ccc) and critical stabilization concentration (csc) are summarized in Table 2. For both nanoparticle systems ccc values were lower when CaCl2 was used as aggregating salt. This was due to the greater screening capacity of Ca2+ ions, which significantly reduces the repulsive potential among the particles, favoring the aggregation of the system [13].

0 100 200 300 400 500

0

10

20

30

40

50

60

Δ T

(%)

Time (min)

(a)

0 100 200 300 400 500

0

5

10

15

20

25

Δ BS

(%)

Time (min)

(b)

Figure 3. ΔT (%) (a) and ΔBS (%) (b) for the different emulsion compositions. Water-P1 ( ), water-P2 ( ), surfactant-organic solvent ( ), surfactant-P1 ( ), surfactant-P2 ( ).

110 Paper IV

Nevertheless, although ccc values were slightly lower for P1 nanoparticles (with NaCl and CaCl2), it was not possible to find a clear difference between ccc values of both colloidal systems. Moreover, the low value of ccc for both systems indicates that only a low amount of the surfactant was incorporated on the nanoparticles surface, that it was not enough to produce a steric stabilization [14].

Table 2. ccc (mM) and csc (mM) values in presence of NaCl and CaCl2 of P1 and P2 nanoparticles. *Size and PDI after incubation for 25 days at room temperature.

ccc (mM) csc (mM) NaCl CaCl2 NaCl CaCl2 Size* (nm) PDI* P1 32 6 585 205 180.3 ± 1.1 0.24 ± 0.01 P2 36 9 185 55 185.5 ± 5.4 0.09 ± 0.01

Interesting information can be obtained when the csc values are analyzed. In this case, differences between both salts and both colloidal systems were clearly found. It should be noted that csc values depends on hydration forces and they do not only depend on the surface hydrophilicity, but also on the nature and concentration of the hydrated counterions that surround the particles. Then, as

calcium is a more hydrated cation than sodium, the re-stabilization power of the former is higher than that of the latter. This explains the lower csc values obtained for CaCl2 in comparison with NaCl for both colloidal systems. However, using a given electrolyte, the lower the csc, the higher the hydrophilicity of the particle surface [10]. Then, the lower csc values found for P2 nanoparticles, independently of the employed salt, with respect to P1 nanoparticles, indicate a higher hydrophilic

10 100 1000

1

10

100

1000

Fuch

s fa

ctor

(W)

[Salt]/mM

Figure 4. Stability factor as a function of salt concentration for P1 (square) and P2 (circle) nanoparticles. NaCl (closed symbols, solid line); CaCl2 (open symbols, dashed line).

Paper IV 111

surface of the former. But, P2 polymer has a higher hydrophobic character than P2 polymer.

3.4. Swelling studies

Due to the swelling properties of the P0 polymer, the swelling capacity of amphiphilic polymer nanoparticles was tested, in order to analyze the effect of the hydrophobic modification in these polymers. Change of size for P1 and P2 nanoparticles was evaluated as a function of the time, for different NaCl concentrations (data not shown). Hydrodynamic mean size of both nanoparticles suspensions displayed a negative exponential dependence with NaCl concentration. With these data, the change of swelling ratio (α) with the bulk salt concentration was calculated for both colloidal systems (Figure 5); α was calculated from 10-20 minutes after mixing, when the nanoparticles size was constant. The results illustrated in Figure 5 confirm the swelling properties of both types of nanoparticles, but it was not possible to appreciate a clear difference in the behaviour of both colloidal systems. P1 and P2 nanoparticles exhibited a very similar dependence of its size with the salt concentration d~n-0.008 (estimation not shown) and α parameter also presented a similar pattern in both cases.

3.5. Stability of stored nanoparticles suspension at 4ºC and 25ºC

Nanoparticles stored at 4ºC were totally stable during incubation time (data not shown), but both systems presented different patterns when stored at 25ºC. Table 2 summarized the hydrodynamic mean size and PDI of both colloidal systems incubated at room temperature for 25 days. In the P1 case a substantial

0 1 2 3 4 50,65

0,70

0,75

0,80

0,85

0,90

0,95

1,00

Swel

ling

ratio

(α)

NaCl (mM)

Figure 5. De-swelling ratio (α) for P1 ( ) and P2 ( ) nanoparticles as a function of NaCl bulk concentration.

112 Paper IV

change in both size and PDI was observed after 25 days. However, P2 nanoparticles showed a negligible change in size or PDI within the period of 25 days.

3.6. Cytotoxicity assay

Cytotoxicity was analysed using two in vitro tests. LDH assay is a test detecting the leakage of cell membrane and is useful in the investigation of nanoparticle toxicity since the plasma membrane is the main place of contact between particles and cells. MTT assay monitors the mitochondrial metabolism of cells as an indicator of their viability. The results of both tests are consistent and show that these particles up to a concentration of 0.033 mg/ml are not cytotoxic in the Caco-2 test model (data not shown). The same holds true for a longer time incubation (data not shown).

IV. DISCUSSION

Results illustrated in Figure 1 show the clear effect of the surfactant in the nanoparticles formation. In this case, the surfactant properties of this polymer improved the nanoparticles formation. In addition, the increase of surfactant concentration in the external aqueous phase gave a size reduction and a lower PDI. These results agree with those published in the 90’s by other authors, who found that the decrease in size and PDI when the concentration of this surfactant was increased can be explained by the increase of the external aqueous phase viscosity. Regarding to the low and negative ζ-potential shown by both colloidal systems, it is plausible that, insofar as neither amphiphilic polymers nor the surfactant present ionizable groups, this value proceed from the specific interaction of NaCl with the nanoparticles surface. The specific interaction of the ions with the nanoparticle surface depends on their polarizability, being higher for the anions. Then the results suggest a higher accumulation of the Cl- anions than Na+ cations although the nanoparticles surface must be neutral [12].

In order to analyze some aspects related to the formation of these nanoparticles the stability of the o/w emulsion as a function of the amphiphilic polymers and the surfactant was analyzed. The results described in section 3.2 (Figures 3a-b) showed that the emulsion prepared in presence of P1 derivative was more stable than that formed in presence of P2. Therefore, these results suggest that P1 polymer presents better tensoactive properties than P2. It means, P1 derivative exhibited a higher affinity for the emulsion interface. This observation is in agreement with the results published by other authors. Hence, the different affinity of both amphiphilic polymers for the interface of the emulsion can justify the lower nanoparticle size found with P1 polymer (amphiphilic polymer with the best tensoactive properties). Nevertheless, neither P1 nor P2 polymer avoided by themselves the coalescence of the organic phase droplets (see Figures 3a-b). Nevertheless, this problem was solved by the adsorption of the surfactant in the interface that helped to reduce the droplet coalescence.

Paper IV 113

The analysis of the colloidal stability presents additional information to know the role played by each polymer during the particle formation. As mentioned earlier (section 3.3), both colloidal systems presented a low stability (see ccc values in Table 2). This is indicative of a low incorporation of the surfactant in the interface during the emulsification step [14]. The amount of adsorbed surfactant can be enough to stabilize the emulsion but it was not enough to give a clear steric stabilization. On the other hand, the csc values are useful to obtain more information of the superficial composition of both nanoparticles systems. In this case, P2 nanoparticles displayed lower csc values than P1 with both salts (see Table 2). This result indicates that the nanoparticles formulated with the polymer with higher hydrophobic character (P2) displayed a more hydrophilic surface. The reason lies on the fact that the adsorption of molecules onto a substrate is mainly controlled by the hydrophobic forces [15, 16]. Hence, the adsorption through the hydrophobic moieties of the surfactant onto the drops of the organic phase during the emulsion step must be more favourable in the case of the more hydrophobic polymer, P2, than for P1. Concomitantly, due to the hydrophilic chains of the surfactant the higher accumulation of this molecule gives a higher hydrophilic character to the surface of the nanoparticles. Regarding to the null effect of the different amount of surfactant incorporated in both colloidal systems on the ccc magnitude, studies carried out previously focused on the adsorption of a surfactant on PLGA nanoparticles showed that the csc value presented a clear dependence with the amount of surfactant placed on the nanoparticles surface, while ccc values were less sensitive to these changes [14]. This observation suggests that although P1 and P2 nanoparticles presented similar ccc values, the obviously different csc values showed by both systems indicate a clearly different surface composition. Hence, stability results indicate that the hydrophobic character of the amphiphilic polymer had a clear effect in the final properties of these nanoparticles.

The next set of experiments was focused on the swelling properties of P1 and P2 nanoparticles. Considering the Flory-Huggins thermodynamic theory and Donnan theory, Fernández-Nieves et al [17] predicted that the hydrodynamic mean diameter of a microgel particle present a null dependence with the ionic strength for low salt concentrations and a d~n-0.2 asymptotic behaviour for high salt concentrations, irrespective of the particle network charge. None of these predicted patterns were observed for P1 and P2 nanoparticles. In fact, they presented a d~n-

0.008 dependence for low ionic strength and a null dependence for NaCl concentrations from 2 to 5 mM (estimation not shown). It is worthy to remark that it was not possible to used NaCl concentrations higher than 5 mM due to their low stability. It should be noted that the approximation of Fernández-Nieves et al was applied to pure microgel particles in which water behaves as a good solvent for the polymer chains. These results suggest a low hydration of the amphiphilic polymers that yield to a poor swelling capacity. Other authors found similar tendencies. Hence, the results published these authors support our assumption: the hydrophobic modification of P1 and P2 polymers reduces, but not totally, their swelling properties.

114 Paper IV

Finally, once both colloidal systems were characterized, more applied studies were carried out to know the potential application of these nanoparticles as drug delivery systems. The stability of P1 and P2 incubated at 4ºC and 25ºC for a period of time of 25 days was analyzed. As was commented above, nanoparticles incubated at 4ºC were totally stable, while those stored at room temperature presented different patterns. Others authors have previously demonstrated that the storage temperature is a crucial parameter for the long-term stability of colloidal systems, for more details see [18]. With respect to the systems incubated at 25ºC, the higher stability of P2 nanoparticles can be explain by the higher amount of surfactant adsorbed on the P2 nanoparticles surface [14].

With respect to the cytotoxicity results, these assays proof not only that the particles formulated with the amphiphilic polymers in the tested concentrations were not harmful for Caco-2 cells but also that the particle suspension was cleaned sufficiently from organic solvents.

V. CONCLUSIONS

In this study, the formulation of nanoparticles using amphipilic polymers with the same backbone but different hydrophobic character was analyzed. Thanks to the analysis of the o/w emulsion stability it was possible to know that the hydrophobic character of these polymers was a crucial point in the role of each one during the nanoparticles formation. From the colloidal stability, the different surface characteristics were determined, showing a higher amount of surfactant on the nanoparticles surface those prepared with the more hydrophobic polymer. Moreover, this study helped to choose the administration route of these nanoparticles as transdermal drug delivery systems, these studies will be shown in the Paper IX of this manuscript. Finally, the presence of hydrophobic moieties in the P1 and P2 polymers reduces their swelling properties.

ACKNOWLEDGEMENTS

Authors thank the financial support given by Galenos Fellowship in the Framework of the EU Project “Towards a European PhD in Advanced Drug Delivery”, Marie Curie Contract MEST-CT-2004-404992.

Paper IV 115

References

1. Buron H, Mengual O, Meunier G, Cayre I, Snabre P. Optical characterization of concentrated dispersions: applications to laboratory analyses and on-line process monitoring and control. Polymer International 2004;53(9):1205-1209.

2. Lemarchand C, Couvreur P, Vauthier C, Costantini D, Gref R. Study of emulsion stabilization by graft copolymers using the optical analyzer Turbiscan. International Journal of Pharmaceutics 2003;254(1):77-82.

3. Mengual O, Meunier G, Cayre I, Puech K, Snabre P. TURBISCAN MA 2000: multiple light scattering measurement for concentrated emulsion and suspension instability analysis. Talanta 1999;50(2):445-456.

4. Hiemenz PC, Rajagopalan R. Principles of Colloidal and Surface Chemistry. New york: Dekker, 1997.

5. Pashley RM. Hydration Forces Between Mica Surfaces in Electrolyte-Solutions. Advances in Colloid and Interface Science 1982;16(JUL):57-62.

6. Israelachvili JN. Intermolecular and Surface Forces. London: Academic Press, 1992.

7. Churaev NV, Derjaguin BV. Inclusion of Structural Forces in the Theory of Stability of Colloids and Films. Journal of Colloid and Interface Science 1985;103(2):542-553.

8. Molina-Bolivar JA, Galisteo-Gonzalez F, Hidalgo-Alvarez R. Colloidal stability of protein-polymer systems: A possible explanation by hydration forces. Physical Review E 1997;55(4):4522-4530.

9. Molina-Bolivar JA, Galisteo-Gonzalez F, Hidalgo-Alvarez R. Specific cation adsorption on protein-covered particles and its influence on colloidal stability. Colloids and Surfaces B-Biointerfaces 2001;21(1-3):125-135.

10. Molina-Bolivar JA, Ortega-Vinuesa JL. How proteins stabilize colloidal particles by means of hydration forces. Langmuir 1999;15(8):2644-2653.

11. Saunders BR, Vincent B. Microgel particles as model colloids: theory, properties and applications. Advances in Colloid and Interface Science 1999;80(1):1-25.

12. Lopez-Leon T, Elaissari A, Ortega-Vinuesa JL, Bastos-Gonzalez D. Hofmeister effects on poly(NIPAM) microgel particles: Macroscopic evidence of ion adsorption and changes in water structure. Chemphyschem 2007;8(1):148-156.

13. Santander-Ortega MJ, Csaba N, Alonso MJ, Ortega-Vinuesa JL, Bastos-Gonz lez D. Stability and physicochemical characteristics of PLGA, PLGA:poloxamer and PLGA:poloxamine blend nanoparticles: A comparative study. Colloids and Surfaces A: Physicochemical and Engineering Aspects 2007;296(1-3):132-140.

14. Santander-Ortega MJ, Jodar-Reyes AB, Csaba N, Bastos-Gonzalez D, Ortega-Vinuesa JL. Colloidal stability of Pluronic F68-coated PLGA nanoparticles: A variety of stabilisation mechanisms. Journal of Colloid and Interface Science 2006;302(2):522-529.

15. Tadros T. Colloid stability using polymeric surfactants. In: Tadros TF, editor. Colloid Stability: The role of the surface forces, Part 1. Weinheim: Wiley-vch Verlag, 2007.

16. Santander-Ortega MJ, Bastos-Gonzalez D, Ortega-Vinuesa JL. Electrophoretic mobility and colloidal stability of PLGA particles coated with IgG. Colloids and Surfaces B-Biointerfaces 2007;60(1):80-88.

116 Paper IV

17. Fernandez-Nieves A, Fernandez-Barbero A, las Nieves FJ. Salt effects over the swelling of ionized mesoscopic gels. Journal of Chemical Physics 2001;115(16):7644-7649.

18. Abdelwahed W, Degobert G, Stainmesse S, Fessi H. Freeze-drying of nanoparticles: Formulation, process and storage considerations. Advanced Drug Delivery Reviews 2006;58(15):1688-1713.

Characterization of Core-Shell Lipid-Chitosan and Lipid-Poloxamer Nanocapsules

M.J. Santander-Ortega1, M.V. Lozano-López2, D. Bastos-González1, J.M. Peula-García3, J.L. Ortega-Vinuesa1

1Biocolloid and Fluid Physics Group, Department of Applied Physics, University of Granada, 18071 Granada, Spain

2Department of Pharmacy and Pharmaceutical Technology, University of Santiago de Compostela, 15706 Santiago de Compostela, Spain.

3Department of Applied Physics, University of Málaga, 29071 Málaga, Spain.

118 Submitted to Journal of Biomat. Sci. Polym. Ed.

ABSTRACT

In this paper, different core-shell structured lipid nanocapsules have been synthesized and characterized looking for future applications as drug carriers. Despite this type of systems have already been used as delivery systems, it is difficult to find in the literature a deep physicochemical characterization of them. Hence, the aim of this work was to achieve a deeper knowledge in the properties of this kind of colloidal particles, paying special attention to the role played by the components that constituted the nanocapsules. Our lipid nanocapsules were formed by a triglyceride-lecithin core surrounded by a chitosan and/or poloxamer (Pluronic® F68) shell. Four different systems were formulated by varying the chitosan and poloxamer contents. The electrokinetic characterization together with colloidal stability studies revealed that Pluronic® F68 presented a secondary role during the nanocapsule formation, obtaining final systems with a low incorporation of poloxamer. However, the incorporation of chitosan was very significant in all cases. In addition, the stability studies, performed not only in ideal solutions but also in simulated physiological fluids, showed that hydration forces play a crucial role to maintain the integrity of these nanocapsules under some physicochemical conditions that match those found in real physiological fluids.

.

Paper V 119

I. INTRODUCTION

The research in Nanotechnology has intensively increased in the last three decades. Colloidal carriers, such as liposomes [1], nanoparticles [2] and polymeric micelles [3] have been widely investigated for drug delivery applications. Although vesicular systems, mainly nanoemulsions and nanocapsules, experimented a contemporary finding compared to other nanosystems, their development has not been so wide. Nevertheless, they are lately receiving increasing attention in different fields such as cosmetics or as drug carriers. Nanoemulsions are vesicular reservoir systems formed by an oil or aqueous core, which is surrounded by a thin polymeric membrane in case of nanocapsules. Therefore, highly hydrophobic drugs are likely to be encapsulated in lipidic cores [4] or hydrophilic molecules in case of aqueous core nanocapsules [5]. The versatility of these systems for the encapsulation of a wide variety of drugs such as low molecular weight, peptides or gene material enlarges their appealing characteristics [6-8]. On the other hand, the presence of a thin polymeric shell surrounding the inner compartment exhibited by nanocapsules in comparison to nanoemulsions, awards higher drug protection from degradation by preventing direct contact of the encapsulated drug with the environment. Moreover, the polymeric shell is crucial in the long term stability of the particles [9]. The shell can be formed by a wide variety of polymers capable to stabilize the oil/water emulsion and to confer stability, long-circulating properties after intravenous administration, and to modulate the interaction of the nanosystems with the biological environment at which they are immersed. A major drawback after intravenous injection of drug delivery systems is their recognition by the mononuclear phagocytes system (MPS) provoking undesirable accumulation of the colloidal carriers in the liver or the spleen. This problem can be overcome by using the so called Stealth® nanosystems, which reduce the opsonisation by grafting onto the carrier surfaces non-ionic amphiphilic macromolecules, for example polyethylene glycol (PEG) derivatives [10]. In addition to adsorb non-ionic surfactants, a cationization strategy is also traditionally designed for the surface of these carriers: whereas surface pegylated colloidal carriers exhibit a prolonged plasma residence time through an escaping tendency from reticuloendothelial uptake, the positive surface of colloidal carriers can offer additional advantages over the conventional negatively charged emulsions [11]. It has been proved that cationic nanosystems interact more intensely with the negatively charged cell membranes of skin, eye, and gastrointestinal mucosa [12, 13], leading to a better uptake of the encapsulated drug. Therefore, obtaining nanocapsules with a combined PEG-cationized shell is advisable.

With this scenario in mind, the goal of the present investigation was focussed to perform a deep characterization study of chitosan nanocapsules containing a polyoxyethylene/cationized shell, in order to analyze their in vitro properties as possible vehicles for either oral or intravenous administration. Therefore, the present study was mainly focused on analysing the colloidal stability of these systems, since any significant aggregation induced by the physicochemical

120 Submitted to Journal of Biomat. Sci. Polym. Ed.

characteristics of the physiological medium where they are immersed would make the carrier system useless. Our study was planned to work with four systems in which the complexity of the shell nature was increased step by step. The four systems shared the same oily core, composed by medium-chain triglycerides – that were liquid at room and physiological temperature – and emulsified by lecithin. The first system was constituted by this simple nanoemulsion where lecithin molecules located at the oil/water interface acted by themselves as stabilizers. The second system was formed adding a polyoxyethylene derivative emulsifier in order to supply some stealth® properties to our lipid core. Pluronic® F68 poloxamer was chosen for this purpose. A cationic shell was sought for the third system; in this case, chitosan was thought to be an ideal molecular candidate because of its advantageous biological properties, such as biodegradability, biocompatibility, and nontoxicity [14, 15]. Finally, the fourth system was formed by adding simultaneously both poloxamer and chitosan chains in order to achieve a heterogeneous shell where the previously described favourable effects given by both types of molecules would coexist. Note that, taking into account future in vivo studies and applications, our lipid nanosystems were always prepared with innocuous compounds (polysaccharides and lipids).

II. MATERIALS AND METHODS

2.1. Reagents.

In order to study the electrophoretic mobility and colloidal stability at different pH values, several buffered solutions with a low ionic strength (I = 0.002 M) were prepared: pH 4 and 5 were buffered with acetate, pH 6 and 7 with phosphate, and pH 8 and 9 with borate. In some cases, stability was evaluated in simulated protein free physiological fluids: simulated gastric (pH 1.2) and intestinal (pH 6.8) fluids were prepared according to the USP XXIX. Hanks buffer was used for simulating plasma physicochemical conditions. Nanocapsule hydrophobic core was formed by the oil Miglyol 812®, kindly donated by Sasol Germany GmbH (Germany), and by Epikuron 145V (which is a deoiled, wax-like, phosphatidilcholine enriched fraction of soybean lecithin (min 45% PC)) that was donated by Cargill (Spain). Pluronic® F68 was acquired from Sigma-Aldrich (Spain). Protasan® Cl 113, medium molecular weight chitosan chloride salt (medium Mw chitosan) with a deacetylation degree of 85%, was supplied from FMC Biopolymer Novamatrix (Norway). All the chemicals and electrolytes used were of the highest grade commercially available.

2.2. Nanocapsule synthesis.

The nanosystems herein studied were prepared by a solvent displacement technique following the procedure described previously by Prego et al. [16], this is a well-known technique widely reported for the preparation of nanocapsules. Traditionally, hydrophilic surfactants should be presented in the aqueous phase before the emulsion formation [17, 18]. These molecules were responsible of the nanocapsules properties and their future stability [19]. Briefly, an organic phase

Paper V 121

was prepared constituted by 40 mg of Epikuron 145V dissolved in 0.5 ml ethanol, 125 μl of Miglyol 812® and 9.5 ml of acetone. This organic phase was immediately poured over an aqueous phase that contained several compounds. Depending on the composition of this solution the final sample showed different interfacial properties. For example, when nothing was added to the aqueous phase, a nanoemulsion formed by (a Miglyol®) core stabilized by a lecithin external layer was obtained. This nanoemulsion will be referred to as “LC” hereafter. When the aqueous phase contained Pluronic® F68 at a 0.25% w/v concentration, a sample (called “PX”) coated by a poloxamer shell was generated. The presence of chitosan (0.05% w/v) as the only component in the aqueous phase gave nanocapsules that were named “CS”. Finally, the forth system, called “CS+PX”, was obtained by mixing chitosan and Pluronic® F68 (at the same concentrations mentioned above) in the aqueous phase. It should be noted that the formation of all our nanosystems was almost instantaneous, as indicated by the milky appearance of the mixture.

2.3. Size and storage stability.

The average size of the nanocapsules was determined by photon correlation spectroscopy (PCS) with a commercial light-scattering setup, 4700C, Malvern Instruments (Malvern, UK), with an argon laser of wavelength λ0 = 488 nm, working at a fixed angle (90º) at 25ºC. PCS gives information about the average diffusion coefficient of the particles, which can be easily related to the mean diameter (∅) by using the Stokes-Einstein equation for spheres. The mean diameter of our four systems in purified water (pH~5.8, 25ºC) was ∅LC = (118 ± 4) nm, ∅PX = (153 ± 3) nm, ∅CS = (212 ± 9) nm, and ∅CS+PX = (196 ± 5) nm. At present, there is no doubt on the fact that the size is a critical variable for the nanosystems to cross biological barriers and to elude their uptake by macrophages [14, 20, 21]. Generally, a mean diameter around 200 nm is advised in specialised literature. In this sense, all of our systems are potentially useful for future in vivo applications. The average size of our four samples – which was measured each week – was constant for months (data not shown), which is an indication of high stability, at least when kept in the storage medium (purified water, 4ºC).

2.4. Electrophoretic mobility.

A ZetaPALS instrument (Brookhaven, USA) was used to measure the electrophoretic mobility (μe). The study was focused on measuring the μe as a function of pH keeping constant a low ionic strength value (0.002 M). Each μe mobility data was obtained by averaging 45 individual measurements.

2.5. Colloidal stability.

NaCl and CaCl2 were used as destabilizing agents. According to the classical DLVO theory [15], a salinity increment triggers the coagulation of a lyophobic colloidal system. During aggregation, the turbidity of the system increases when the average size of the scattering particles increases. Therefore, a simple spectrophotometer (Beckman DU 7400) working with a visible wavelength

122 Submitted to Journal of Biomat. Sci. Polym. Ed.

0 20 40 60 80 100 120

0,5

1,0

1,5

2,0

2,5

3,0

Abs

orba

nce

(λ =

570

nm

)

time (s)

Figure 1a. Variation of the optical absorbance with time for the PX sample at different CaCl2concentrations (pH7): ( ) 2.5 mM; ( ) 5 mM; ( ) 7.5 mM; ( ) 12.5 mM; ( ) 18.8 mM; ( ) 37.5 mM; ( ) 50 mM; ( ) 87.5 mM; ( ) 125 mM; and ( ) 500 mM.

1 10 100 1000

1

10

100

1000

csc

W

CaCl2 concentration (mM)

ccc

Figure 1b. Dependence of the stability factor (W) on the CaCl2 concentration for the PX sample at pH7. Lines serve to guide the eye for locating the ccc and csc values.

(λ = 570 nm) is clearly able to detect and to analyze the aggregation kinetics. Figure 1a shows a typical coagulation experiment, in which the PX sample was forced to aggregate at different CaCl2 concentrations. Information about the kinetics aggregation constant “k” of dimmer formation can be derived from these curves.

Paper V 123

The initial slopes of the absorbance vs time curves (dAbs/dt) can be directly related to k by [22, 23]:

( ) 22 1 0/ 2

2.3C C N ldAbs k

dt−

= (1)

where C1 and C2 are the scattering cross-sections of a monomer and a dimmer, respectively, N0 is the initial particle concentration, and l is the optical path through the cuvette. Nevertheless, stability is usually evaluated by calculating the Fuchs factor (W), instead of calculating the k values by using equation 1. The Fuchs factor (also called “stability factor”) can be experimentally obtained by:

( / )( / )

f f

s s

k dAbs dtW

k dAbs dt= = (2)

where “kf” refers to the fastest aggregation kinetics constant, and the subscript “s” refers to slower coagulation rates. Therefore, when W = 1 the system is completely unstable, while W = ∞ indicates total stability. It is easy to calculate the Fuchs factor at every salt concentration from those data shown in Fig. 1a, and using equation 2. This is shown in Figure 1b, where the dependence of W on the CaCl2 concentration has been plotted. The double-logarithmic scale becomes very useful to estimate the critical coagulation concentration (ccc) and the critical stabilization concentration (csc), which are fundamental parameters in colloidal stability studies. The ccc value is related to destabilization processes and it indirectly gives information about the surface charge density of the particles; a low ccc means low stability. However, the csc value – defined as the minimum salt concentration at which the system begins to re-stabilize when salinity is increased even more – is associated to the surface hydrophilicity. This kind of restabilization phenomenon at high salt concentration is well known in colloidal systems, and it is governed by hydration forces [24-29]. For a given electrolyte, the lower the csc, the higher the hidrophilicity of the particle surface.

III. RESULTS AND DISCUSSIONS

The analysis of the particle sizes (shown in the previous section) can give initial information about the role played by the poloxamer and/or chitosan during the nanocapsule formation. Lecithin appears to be the best emulsifier, since the mean diameter was lower when this molecule acts by itself stabilizing the formulation (see the LC case). Poloxamer, however, seems to be a poorer emulsifier than lecithin because when both surfactants were mixed the particle size increased (see the PX diameter). This size increment became even higher when chitosan (instead Pluronic® F68) was added together with the lecithin. In this case, positive chitosan chains can interact electrostatically with negative lecithin molecules reducing the effective concentration of this last emulsifier and, consequently, producing bigger nanocapsules (see CS). Finally, when poloxamer was also added

124 Submitted to Journal of Biomat. Sci. Polym. Ed.

4 5 6 7 8 9

-6

-4

-2

0

2

4

μ e 108 (m

2 /Vs)

pH

ure 2. Electrophoretic mobility versus pH for the ( ) LC, ( ) PX, ( ) CS, and ( ).Fig

together with chitosan and lecithin (CS+PX) the mean diameter values practically coincided with the CS sample. It is as if our non-ionic surfactant does not play any role in this blend and the nanocapsule size only depends on the chitosan/lecithin mixture. As will be shown afterwards, electrophoretic mobility and stability data appear to corroborate that Pluronic® F68 is practically unable to anchor at the NC/water interface in the CS+PX system.

The next set of experiments was designed to know the electrical state of our nanocapsules at different pH values. The electrophoretic mobility data, obtained at low ionic strength media, are shown in Figure 2. We will analyse the results sample by sample. The LC nanocapsules showed a typical behaviour of colloids with weak acid groups at the surface, giving lower μe values at acid pH than those obtained at neutral and basic pH. These results are in agreement with the nature of the shell of these capsules, which is exclusively formed by lecithin; note that lecithin is a mixture of phospholipids that contains negatively charged components, although the major component is phosphatidylcholine. Surprisingly, when Pluronic® F68 is added to the formulation, the mobility of the resulting capsules (see the PX case in Fig. 2) is almost identical to the LC sample. A μe reduction was expected after the incorporation of this non-ionic surfactant onto the lecithin surface, since the presence of PEO chains would shift outwards the shear plane where the ζ-potential is defined, which subsequently would diminish the electrophoretic mobility. At least, for PLGA particles the μe reduction was significant and directly related to the poloxamer coating [30]. Only at pH4 a diminution of mobility (with regard to the LC case) is observed. The similar μe behaviour among these two samples (excepting the pH 4 data) suggests low poloxamer incorporation onto the particle surface, probably because the lecithin shell supplies an important hydrophilic character to the surface that hinders the adsorption of this non-ionic surfactant: Note that for

Paper V 125

uncharged polymers, as Pluronic® F68, adsorption only can be led by hydrophobic interactions. Similar reasoning can be applied when comparing the μe results of the CS and CS+PX systems. We will start analysing the CS sample. Incorporation of chitosan seems to be clearly effective when this polysaccharide is added to the formulation. Mobility of the LC nanocapsules radically changes in presence of chitosan chains (see the CS curve in Fig. 2), yielding to particle surfaces that show mobility behaviours totally similar to those obtained with pure chitosan nanogels [31]. It can be seen that mobility goes from positive values (at acid pH) to large negative values at more basic pH, presenting an isoelectric point (i.e.p.) at pH 7. The positive charge of nanocapsules is provided by the glucosamine groups of chitosan, which present at weak basic character. At basic pH chitosan chains are uncharged so the negative μe may come from phosphate groups of lecithin. These μe results suggest a shell structure practically formed (in its outer part) by a chitosan layer. The clear incorporation of this polycationic polymer at the nanocapsule shell can be understood by considering the attractive interaction existing between the (negative) lecithin and the (positive) chitosan molecules. Sonvico et al. [32] have experimentally evidenced a strong electrostatic interaction between these two components, which are capable to form by themselves self-organized lecithin/chitosan nanoparticles by means of purely ionic interactions. On the other hand, the CS+PX sample practically match the CS mobility, which would indicate a very low or almost negligible incorporation of poloxamer when this surfactant is added during the synthesis process. Note that both chitosan and Pluronic® F68 chains are simultaneously added to the formulation, and thus, a competitive adsorption onto an enriched lecithin layer must take place. In this competitive situation chitosan molecules are much more attracted by the negatively charged lecithin layer than poloxamer, which does not experiment any specific attraction (neither by means of electrostatic nor hydrophobic interactions) toward the hydrophilic surface of our nanocapsules. Therefore is plausible to believe that the surface will be mainly coated by chitosan, remaining the poloxamer molecules dissolved (or forming micelles) in the bulk solution. As will be shown, the stability experiment will confirm this assumption. Finally, as the isoelectric point for the CS and CS+PX nanocapsules was around pH 7, it is more than likely that these samples become unstable at neutral pH, unless the action of any steric contribution could prevent the system coagulation.

Table 1a. Critical coagulation concentration (ccc) and critical stabilization concentration (csc) data, in mM units of LC and PX at different pH values.

LC PX

NaCl CaCl2 NaCl CaCl2 ccc csc ccc csc ccc csc ccc csc

pH4

pH7

pH9

100 ---

310 1000

- stable -

10 220

10 150

12 >40

∼200 ∼

- stable -

- stable -

18 60

21 53

12 90

126 Submitted to Journal of Biomat. Sci. Polym. Ed.

Table 1b. Critical coagulation concentration (ccc) and critical stabilization concentration (csc) data, in mM units of CS and CS+PX at different pH values.

CS CS + PX

NaCl CaCl2 NaCl CaCl2 ccc csc ccc csc ccc csc ccc csc

pH4

pH7

pH9

32 280

- aggr. -

20 ---

10 20

- aggr. -

6 40

32 235

- aggr. -

30 ---

10 20

- aggr. -

7 40

In the third set of experiments the colloidal stability was studied at different pH values, in order to analyze several situations in which nanocapsules varied their corresponding electrical states. The selected pH values were 4, 7 and 9. Aggregations were induced by salinity using independently NaCl and CaCl2. The results obtained at each pH will be discussed separately. Figures 3a and 3b show the stability patterns at pH 4. The corresponding ccc and csc values are shown in Tables 1a-b. With regard to the ccc data, calcium exert a much higher destabilizating power than sodium, above all in those samples where these cations act as counterions (that is, in the LC and PX cases). The coincidence of ccc values in the CS and CS+PX samples appears to corroborate the conclusions extracted from the mobility experiments, that is to say: both systems practically share the same

superficial nature, indicating that our non-ionic surfactant was hardly able to adsorb onto the lecithin shell when competing for adsorption with chitosan chains. However, some differences exist between the ccc data of the LC and PX nanocapsules. The PX sample is more stable, suggesting that some poloxamer molecules achieve to adsorb onto the lecithin shell in formulation media free of chitosan.

10 100 1000

1

10

100

1000

PXLC

CS

W

NaCl concentration

CS+PX

(a)

1 10 100 1000

1

10

100

LC

PXCS + PX

W

CaCl2 concentration (mM)

CS

(b)

Figure 3. Stability factor versus (a) NaCl or (b) CaCl2 concentration at pH 4. ( ) LC, ( ) PX, ( ) CS, and ( ) CS+PX samples.

Paper V 127

With regard to the csc values, they are dependent on the hydration forces acting between two approaching nanocapsules. It should be noted that hydration forces are structural repulsive interactions that arises form the local order of water layers adjacent to a surface. The magnitude of this force is not only correlated to the hydrophilicity of the surface but also depends strongly on the concentration and hydration degree of the ions that surround the surface [24, 26, 28, 33]. This is why restabilization by means of hydration forces is often found when great amounts of hydrated ions are accumulated at the proximities of any hydrophilic surface.

According to data shown in Tables 1a-b, calcium clearly shows higher restabilization trends when compared to sodium. This is an expected result, since Ca2+ is much more hydrated than Na+. On the other hand, if only one electrolyte is considered (NaCl or CaCl2), differences in csc data give a qualitative information about the hydrophilicity of the nanocapsule surface. The clear restabilization phenomena observed in the CS and CS+PX samples, even with NaCl, indicate that these samples are much more hydrophilic than LC or PX ones. This result is expectable because water behaves as a good solvent at room temperature for the deacetylated chitosan chains that are located at the particle shell.

Additional information about the surface composition can be obtained if aggregation kinetics are also analysed. Figures 4a and 4b show the rapidest coagulation kinetics for every system at pH4, using NaCl and CaCl2, respectively; that is, coagulation regimes produced at salt concentrations above the ccc but below the csc. There are some expectable results, as those related to the aggregation kinetics given by CaCl2 (Fig. 4b). At this acid pH the CS and CS+PX particles present a positive surface, while the LC and PX nanocapsules are negative. Therefore, calcium acts as counterion for LC and PX, while chloride does for CS and

0 20 40 60 80 100 120

0,5

1,0

1,5

2,0

PX

LC

CS+PX

Abs

orba

nce

(λ =

570

nm

)

time (s)

CS(a)

0 20 40 60 80 100 120

0,4

0,6

0,8

1,0

Abso

rban

ce (λ

= 5

70 n

m)

time (s)

(b)

Figure 4. Aggregation kinetics at pH4 under maximum instability conditions, that is, at asalt concentration value above the ccc and below the csc: (a) 100 mM of NaCl (excepting the PX sample, in which 200 mM of NaCl was used); and (b) 12.5 mM of CaCl2, (excepting the PX sample, in which 20 mM of CaCl2 was used). ( ) LC, ( ) PX, ( ) CS, and ( ) CS+PX samples.

128 Submitted to Journal of Biomat. Sci. Polym. Ed.

CS+PX, and thus, the surface potential screening given by the divalent cation is much more effective, producing more rapid aggregation, than that produced by the monovalent anion [15]. In addition, the existence of repulsive structural interactions (mediated by hydration forces) in the most hydrophilic systems (CS and CS+PX) – even at low calcium concentrations (note that the csc is around 20 mM for both samples) – contributes to slow down the kinetics in the CS and CS+PX samples with regard to the LC and PX ones. The high aggregation kinetics shown by the LC sample can be understood not only taking into account the strong screening effect exerted by this divalent counterion, but also considering that Ca2+ form an poorly soluble ionic pair with phosphate groups. Consequently, the phosphate charges in the lecithin shell would be rapidly neutralized by calcium, and thus, the LC particles would rapidly collapse in absence of repulsive electrostatic forces. In addition, these experiments permit to infer that incorporation of poloxamer molecules onto the nanocapsule surface was not negligible, since the PX aggregation kinetics was significantly slower than that of LC. That is, kinetics results appear to indicate that some PEO chains are present in the particle/water interface, creating some kind of steric repulsive barrier, which would explain why the PX and CS+PX kinetics are lower than those of the LC and CS cases, respectively. Note that this feature is also observable when working with NaCl (see Fig. 4a). Nevertheless, a striking result is observed when working with the CS sample in the 100 mM NaCl solution (see Fig. 4a): an extremely rapid coagulation process is found. If there was acting only one aggregating mechanism – i.e. the well-known electric double layer compression – a lower kinetics would be obtained (as that observed for the LC case in NaCl). The very rapid aggregation occurring in the CS system suggests the participation of other specific destabilization mechanism. It is known that chitosan easily tends to form ion pairs with non-monovalent anions, a property which is broadly used in ionic gelation processes by using tripolyphosphate [34-36] or sulphate [37, 38] as cross-linking molecules. However, to our knowledge, chitosan is not able to form this type of specific ion pairs with chloride; consequently, giving a convincing explanation to this point becomes rather complicated. Table 2 summarizes, following a qualitative scale, the aggregation kinetics patterns obtained under maximum instability conditions, that is, at salt concentration values above the ccc and below the csc.

Table 2. Qualitative classification of aggregation kinetics above the ccc and below the csc: (-) no aggregation, (+) very low, (++) low, (+++) rapid, (++++) very rapid, and (+++++) extremely rapid kinetics.

LC PX CS CS + PX

NaCl CaCl2 NaCl CaCl2 NaCl CaCl2 NaCl CaCl2

pH4

pH7

pH9

++

+

-

+++

+++

++++

+

-

-

++

++++

+++++

+++++

- aggr. -

+

+

- aggr. -

+++

+++++

- aggr. -

+

+

- aggr. -

++++

Paper V 129

The next set of experiments was aimed to analyze the stability at neutral pH. It should be noted that aggregation studies by using NaCl and CaCl2 were not carried out with the CS and CS+PX samples, since these two systems coagulated as soon as they were immersed into the pH7 buffer. This feature is shown in Figure 5a. The isoelectric point of both samples coincided with the neutral pH (see Fig 2), and thus, this spontaneous aggregation is a result of charge cancellation. Once more, the CS+PX system shows lower aggregation kinetics than that of the CS one, suggesting that, at least, some poloxamer molecules were able to adsorb together with the chitosan chains to form the external shell of the nanocapsules. With regard to the other two systems (LC and PX), their electrokinetic charges were higher than those at pH4 (see Fig. 2), and thus, it is presumable to find a higher stability. The stability results are shown in Figure 5b, and the corresponding ccc and csc values are given in Table 1. The PX sample became completely stable when using NaCl. This can be explained as follows. Data obtained with CaCl2 clearly inform us about the presence of Pluronic® F68 molecules on the PX surface, since its ccc is higher than that of LC (which must be a result of the action of a stabilizing agent), and also its csc is lower than that of LC, which is a signal of a higher hydrophilic character that is given by the PEO fragments. When NaCl is used instead, the ccc values become higher than those at pH4, even making the ccc and csc overlap for the most hydrophilic system (PX). That is, restabilization mechanisms effectively act on this hydrophilic system even for moderate NaCl concentrations, avoiding aggregation in all the NaCl concentration range. This does not occur in the less hydrophilic sample (LC), where restabilization phenomena also take place at very high salt concentrations, although in this case the ccc and csc values did not attain to overlap completely; nevertheless, the rapidest LC aggregation kinetics became rather low when working with this salt (see Table 2).

Figure 5a. Aggregation kinetics of ( ) CSand ( ) CS+PX nanocapsules when they were immersed at pH 7 in a low strengthbuffered solution (I = 0.002 M).

0 50 100 150 2000,25

0,30

0,35

0,40

0,45

0,50

0,55

0,60

0,65A

bsor

banc

e (λ

= 5

70 n

m)

time (s)10 100 1000

1

10

100

LC (Na+)

PX (Ca2+

)W

Salt concentration (mM)

LC (Ca2+

)

Figure 5b. Stability factor versus salt concentration at pH 7. ( ) LC with NaCl, ( ) LC with CaCl2, and ( ) PX with CaCl2.

130 Submitted to Journal of Biomat. Sci. Polym. Ed.

10 100 1000

1

10

Finally, the aggregation at pH9 was studied. Figures 6a and 6b show the stability factor versus NaCl and CaCl2 concentration, respectively. At this basic pH the lecithin ionizable groups must be completely charged, making the surface even more hidrophilic than that at pH 7. This would explain why not only the PX but also the LC nanocapsules were completely stable, since (as just explained previously) the ccc and csc values overlap in both samples, and thus, no aggregation was found in all the NaCl concentration range. On the other hand, a low stability is found for the CS and CS+PX systems (see the ccc data in Table 1). It should be noted that at pH9 the chitosan shell is not charged, as the pKa values for chitosan usually goes from 6 to 7, depending on the deacetylation degree [31, 39]. In this situation, electrostatic repulsive forces become weak, which favours the colloidal instability. This lack of charge does not only explain the low ccc values, but also the absence of csc, since the uncharged chitosan shell is much less hydrophilic than a fully charged one. It is necessary the participation of a very hydrated ion (i.e. calcium) at moderate and high concentrations to observed restabilization by means of hydration forces (see Fig. 6b). Once more, the minor differences existing among the CS and CS+PX results suggest a low incorporation of poloxamer at the particle interface. Additionally, at this basic pH a partial desorption of uncharged chitosan might take place, since the electrostatic attraction between the (negative) lecithin layer and the (positive) chitosan chains would disappear when our polysaccharide becomes uncharged (which occurs at pH9). This partial desorption would explain why the mobility of CS and CS+PX is almost equal to that of the LC sample at pH9 (see Fig. 2); or why the stability patterns of both samples are similar to that of LC at pH9 (i.e. compare the ccc and csc data with CaCl2). Nevertheless, chitosan desorption must not be fully complete as differences appears with NaCl: LC sample was totally stable while some unstability was found with the CS and CS+PX systems, although it should be noted that they showed very low aggregation kinetics (see Table 2).

CS + PX

W

NaCl concentration (mM)

CS

(a)

1 10 100 1000

1

10

100

1000

LC

PX

CS + PX

W

CaCl2 concentration (mM)

CS

(b)

Figure 6. Stability factor versus (a) NaCl or (b) CaCl2 concentration at pH 9. ( ) LC, ( ) PX, ( ) CS, and ( ) CS+PX samples.

Paper V 131

If one combines all the stability results shown in Tables 1 and 2, it is possible to extract some general conclusions. With regard to the ccc values, all of them are concordant, at least qualitatively, with the electrophoretic mobility data shown in Fig. 2, indicating that electric repulsive forces are the main responsible of the stability of our systems. Incorporation of Pluronic® F68 at the nanocapsule surface is not high in the PX case or even very poor in the CS+PX case, and consequently no stabilization mechanisms based on steric hindrance of PEO chains have been observed. However, the hydrophilic character of our surfaces has become an advantage against aggregation due to the effect caused by stabilization processes governed by hydration forces at moderate and high ionic strengths. The csc results permit to conclude that hidrophilicity increased as follows: LC < PX < CS ≤ CS+PX. With regard to the maximum aggregation kinetics (Table 2), calcium exerts rapid aggregation mechanisms only when this divalent cation acts as counter-ion. When it acts as co-ion (see the CS and CS+PX cases at pH4) the kinetics become rather low, although in such cases NaCl rapidly destabilized both systems. Finally, if the kinetics between LC and PX (on the one hand), and CS and CS+PX (on the other hand) are compared (see Table 2), the presence of some poloxamer molecules at the nanocapsule interface slightly speeds up aggregation at neutral and basic pH with CaCl2. This feature may be caused by the natural tendency of calcium to form chemical complexes with the oxygens of PEO groups, a reaction catalyzed by traces of multivalent voluminous anions (as phosphate or borate which are present in our buffers) [40-42]. In these cases, additionally to the coagulation caused by charge screening, a bridging mechanism – mediated by complexation of calcium with PEO groups of two different particles – also participates, speeding up somehow the aggregation kinetics.

Despite stability has been analysed in ideal solutions, the obtained ccc values may help to predict aggregation or stability regimes of our nanocarriers in some physiological solutions. For example, taking into account that sodium and calcium concentrations in blood (pH 7.4) are 145 mM and 1.2 mM, respectively, or that they are 140 mM and 2.5 mM in the intestinal fluid (pH 6.8) [43], a stability of LC and PX systems and a destabilization of CS and CS+PX ones may be predictable according to the data shown in Table 1. Nevertheless, the best way to check these assumptions, at least in vitro, is to analyze the potential aggregation kinetics in simulated fluids. Figures 7a-d show the results in three different simulated solutions: gastric, intestinal and plasma. As will be justified soon afterwards, a forth solution was also used, namely, a simulated intestinal fluid in which the sodium concentration was reduced ten-fold. We will start discussing the results in the gastric fluid, where the pH was 1.2 and the sodium concentration was 34 mM. Note that this sodium concentration matched the ccc found at pH4 for the CS and CS+PX samples. However, we have significantly shifted the pH towards a more acidic value in which CS and CS+PX present a fully charged positive shell, making these nanocapsules stable even at 34 mM in Na+. The other two samples (LC and PX) clearly aggregate by means of a charge cancellation mechanism, since the

132 Submitted to Journal of Biomat. Sci. Polym. Ed.

0 20 40 60 80 100 120

0,5

1,0

1,5

2,0

Abso

rban

ce (λ

= 5

70 n

m)

Time (s)

(a)

0 20 40 60 80 100 120

0,5

1,0

1,5

2,0

Abso

rban

ce (λ

= 5

70 n

m)

Time (s)

(b)

0 20 40 60 80 100 120

0,5

1,0

1,5

2,0

phosphate groups in the lecithin molecules become totally protonated at pH 1.2. The stability behaviour of our systems changed when they were immersed in intestinal fluid. In this medium LC and PX were stable, as predicted by the data shown in Tables 1a-b. Surprisingly, CS and CS+PX systems became completely stable. This could be a striking result, as the intestinal pH matches the CS and CS+PX isoelectric points (see Fig. 2). It is worth reminding that both samples aggregated as soon as they were immersed into our low ionic strength pH7 buffer (see Fig 5a). There is only one logical explanation to justify the clear stability found in the intestinal fluid: the sodium concentration is around 150 mM, which is a moderate (but high enough) value capable to restabilize both hydrophilic systems by means of hydrations forces. To corroborate this hypothesis, aggregation kinetics were repeated in a simulated intestinal medium where the sodium concentration was reduced to 15 mM. As can be seen in Fig. 7c and 7d, both systems immediately began to coagulate in this low salinity medium, which confirms our assumption. Therefore, hydration forces are capable to convert by themselves an uncharged and unstable system into a stable one, provided that the particle surface is hydrophilic enough. Finally, the stability in Hanks buffer shows that LC nanocapsules are stable

Abso

rban

ce (λ

= 5

70 n

m)

Time (s)

(c)

0 20 40 60 80 100 120

0,5

1,0

1,5

2,0

2,5

3,0

3,5

4,0

Abso

rban

ce (λ

= 5

70 n

m)

Time (s)

(d)

Figure 7. Aggregation kinetics of (a) LC, (b) PX, (c) CS, and (d) CS+PX nanocapsules when they were immersed in ( ) simulated gastric fluid, ( ) simulated intestinal fluid, ( ) simulated intestinal fluid with a low Na+ concentration (15 mM), and ( ) Hanks buffer.

Paper V 133

(as predicted), while PX is partially unstable (not predicted initially), and CS and CS+PX are totally unstable. It may be surprising not to observe restabilization phenomena mediated by hydration forces in our last two samples (above all reminding the stability in intestinal fluid), since the buffer contains hydrated Na+, Ca2+ and Mg2+, that significantly contribute to create a repulsive barrier against coagulation. The reason may lie on the fact that Hanks buffer also contains sulphate and phosphate anions, with which (as mentioned previously) chitosan easily tends to form ion pairs [34-38] and, consequently, they can induce coagulation not only by charge cancellation, but also by a bridging mechanism. In addition, when comparing the results between LC and PX (on the one hand), and CS and CS+PX (on the other hand), the presence of Ca2+ and Mg2+ in Hanks speeds up the aggregation kinetics of those particles that have PEO chains adhered on their surfaces. This is due to the (previously mentioned) tendency of divalent cations to form complexes with the oxygens of PEO fragments, increasing the aggregation of the system by bridging. It should be noted that PX initially aggregates and then it reaches a steady state, which reproduces the patterns observed with the aggregation of PLGA-Pluronic® F68 complexes in presence of calcium (see Fig. 6 of reference [42]). This would confirm the complexation reaction between Ca2+ and Mg2+ and PEO groups as responsible of the rare aggregation behaviour found with the PX sample in Hanks.

As a general conclusion, administration of LC and PX particles via oral is not advisable, since both types of nanocapsules, despite being stable in the intestinal tract, would aggregate in the stomach losing their colloidal identity. However, nanocapsules covered by chitosan (either CS or CS+PX) are potentially useful for oral administration, since they are completely stable not only in the stomach (by means of repulsive electrostatic forces), but also in the small intestine thanks to the action of the hydration forces. In fact, various drug delivery carriers based on chitosan formulations have been successfully used in vivo experiments when using oral, ocular and nasal routes [16, 44, 45]. With regard to intravenous administration, nanocapsules coated by chitosan would aggregate rapidly, and, in principle, their use would be discarded. Stability is only observed for LC and PX (at least partially), although the presumable low concentration of poloxamer molecules in their interface would make these carriers to be rapidly uptaken by the MPS. This handicap may be solved if PEG, instead of PEO, derivatives were used. For example, following the usual strategy of covalent-modification with polyethylenglycol, chitosan pegylation could be an interesting approach to improve the stability of cationized nanocapsulate systems with stealth® properties. Moreover, the covalent pegylation would become an advantage for intravenous administrations because the PEG desorption or dilution after contact with blood components could be reduced [13].

IV. CONCLUSIONS

134 Submitted to Journal of Biomat. Sci. Polym. Ed.

Four core-shell lipid nanocapsules differing in the shell nature have been synthesized with innocuous compounds. In all cases, the mean diameter was optimum and appropriate to be potentially used as drug delivery carriers.

The colloidal characterization (performed by electrophoretic mobility and stability measurements) has allowed us to conclude that incorporation of Pluronic® F68 at the particle/water interface was not high. It should be noted that this non-ionic surfactant is only capable to adsorb by means of hydrophobic forces, and the external (hydrophilic) lecithin layer in the nanocapsules does not favour such type of adsorption mechanism. The poloxamer incorporation was even poorer when chitosan chains were added together with the surfactant molecules during the synthesis. In this situation, the competitive adsorption between both types of polymers is much more favourable for chitosan, which is strongly adhered to the lecithin coat by means of intense attractive electrical forces. This is why performing a sequential adsorption (adding first the PEO derivative) instead of a competitive adsorption would be recommended to increase the surfactant/chitosan ratio at the particle interface. Nevertheless, there are available better strategies in order to obtained cationized stealth® nanocapsules using shells based on chitosan: For example, 1) coadsorbing chitosan and a poloxamine instead of a poloxamer, since the poloxamines possess positive charges that would help to enhance the incorporation of this PEO derivative on the outer lecithin layer; or 2) forming a shell of chitosan covalently modified by polyethylenglycol derivatives. In this manner, a higher durable shell would be obtained and, moreover, changing the PEO by PEG groups the undesired complexes formed by PEO fragments with divalent cations (i.e. Ca2+, Mg2+) in the presence of phosphate traces would disappear.

It has been shown in this work that hydration forces play a crucial role in the colloidal stability of hydrophilic nanocapsules. In fact, total destabilized nanocapsules in low salinity media (i.e. CS and CS+PX in the pH7 buffer) become completely stable at physiological ionic strength values (see data in simulated intestinal fluid) due to the action exerted by these structural forces.

Finally, it should be noted that it is absolutely mandatory to carry out in vivo studies to test the viability of lipidic nanocapsules as potential drug delivery systems. However, in vitro studies, as those presented in this paper, become very useful, since they serve to delimit and diminish the number of variables to study in in vivo analysis, which in turn also helps to reduce the number of sacrificed animals.

ACKNOWLEDGEMENTS

The authors wish to thank the financial support granted by the “Ministerio de Educación y Ciencia” (MEC, Spain), project MAT2007-66662-C02-01 (European FEDER support included), and from the “Conserjería de Innovación, Ciencia y Tecnología de la Junta de Andalucía” (Spain), projects of excellence FQM 392 and FQM03099. Moreover, M.V. Lozano-Lopez acknowledges the fellowship received from de Spanish Government (AP2005-1701). Finally, we wish to thank Professor M.J. Alonso for her supervision.

Paper V 135

References

1. D. C. Drummond, O. Meyer, K. Hong, D. B. Kirpotin and D. Papahadjopoulos, Pharmacol. Rev. 51, 691 (1999).

2. I. Brigger, C. Dubernet and P. Couvreur, Adv. Drug Deliv. Rev. 54, 631 (2002).

3. M. C. Jones and J. C. Leroux, Eur. J. Pharm. Biopharm. 48, 101 (1999).

4. K. H. Bae, Y. Lee, T. G. Park, Biomacromolecules. 8, 650 (2007).

5. G. Lambert, E. Fattal, H. Pinto-Alphandary, A. Gulik, P. Couvreur, Pharm. Res., 17, 707 (2000).

6. M. N. Khalid, P. Simard, D. Hoarau, A. Dragomir, J-C Leroux, Pharm. Res. 23 752 (2006).

7. C. Prego, D. Torres, E. Fernandez-Megia, R. Novoa-Carballal, E. Quiñoá, M.J. Alonso, J. Control. Rel. 111, 299 (2006).

8. S. H. Lee, S. H. Choi, S. H. Kim, T. G. Park, J. Control. Rel. 125, 25 (2008).

9. Andrea Rübe. Development and physico-chemical characterization of nanocapsules. Thesis, Martin-Luther-University, Germany (2006).

10. A. Gabizon, R. Catane, B. Uziely, B. Kaufman, T. Safra, R. Cohen, F. Martín, A. Huang and Y. Barenholz, Cancer Res. 54, 987 (1994).

11. M. Tobio, A. Sánchez, A. Vila, I. Soriano, C. Evora, J.L. Vila-Jato and M.J. Alonso, Colloids Surf. B 18, 315 (2000).

12. S. M. Moghimi and A. C. Hunter, Crit. Rev. Ther. Rug Carrier Syst. 18, 527 (2001).

13. G. Barrat, Cell. Mol. Life Sci. 60, 21 (2003).

14. M. P. Desai, V. Labhasetwar, E. Walter, R. J. Levy and G. L. Amidon, Pharm. Res. 14, 1568 (1997).

15. P. C. Hiemenz and R. Rajagopalan, Principles of Colloidal and Surface Chemistry. Dekker, New York (1997).

16. C. Prego, M. Fabre, D. Torres, M.J. Alonso, Pharm. Res. 23, 549 (2006).

17. K. S. Soppimath, T. M. Aminabhavi, A. R. Kulkarni, W. E. Rudzinski, J. Control. Rel. 70, 1 (2001).

18. K. Letchford, H. Burt, Eur. J. Pharm. Biopharm. 65, 259 (2007).

19. H. Fessi, F. Puisieux, J.Ph. Devissaguet, N. Ammoury, S. Benita, Int. J. Pharm 55, R1 (1989).

20. T. M. Allen, G. A. Austin, A. Chonn, L. Lin and K. C. Lee, Biochim. Biophys. Acta 1061, 56 (1991).

21. A. T. Florence, Pharm. Res. 14, 259 (1997).

22. J. M. Singer, F. C. A. Vekermans, J. W. Th. Lichtenbelt, Th. Hesselkink and A. P. H. Wiersema, J. Colloid Interface Sci. 45, 608 (1973).

23. M. Quesada, J. Puig, J. M. Delgado and R. Hidalgo-Álvarez, J. Biomater. Sci. Polymer Edn 9, 961 (1998).

24. R. M. Pashley, Adv. Colloid Interface Sci. 16, 57 (1982).

136 Submitted to Journal of Biomat. Sci. Polym. Ed.

25. N. V. Churaev and B. V. Derjaguin, J. Colloid Interface Sci. 103, 542 (1985).

26. J. N. Israelachvili, Intermolecular and Surface Forces. Academic Press, London (1992).

27. J. A. Molina-Bolívar, F. Galisteo-González and R. Hidalgo-Álvarez, Phys. Rev. E 55, 4522 (1997)

28. J. A. Molina-Bolívar and J. L. Ortega-Vinuesa, Langmuir 15, 2644 (1999).

29. J. A. Molina-Bolívar, F. Galisteo-González and R. Hidalgo-Álvarez, Colloids Surf. B 21, 125 (2001).

30. M. J. Santander-Ortega, A. B. Jódar-Reyes, N. Csaba, D. Bastos-González and J. L. Ortega-Vinuesa, J. Colloid Interface Sci. 302, 522 (2006).

31. T. López-León, E. L. S. Carvalho, B. Seijo, J. L. Ortega-Vinuesa and D. Bastos-González, J. Colloid Interface Sci. 283, 344 (2005).

32. F. Sonvico, A. Cagnani, A. Rossi, S. Motta, M. T. Di Bari, F. Cavatorta, M. J. Alonso, A. Deriu and P. Colombo, Int. J. Pharm. 324, 67 (2006).

33. Christenson, H. K. J. Dispersion Sci. Technol. 1988, 9, 171.

34. P. Calvo, C. Remuñán-López, J. L. Vila-Jato and M. J. Alonso, J. Appl. Polym. Sci. 63, 125 (1997).

35. K. A. Janes, M. P. Fresneau, A. Marazuela, A. Fabra and M. J. Alonso, J. Controlled Rel. 73, 255 (2001).

36. A. Vila, A. Sánchez, M. Tobío, P. Calvo and M. J. Alonso, J. Controlled Rel. 78, 15 (2002).

37. A. Berthold, K. Kremer and J. Kreuter, J. Controlled Rel. 39, 17 (1996).

38. I. M. van der Lubben, J. C. Verhoef, A. C. Van Aelst, G. Borchard and H. E. Junginger, Biomaterials 22, 687 (2001).

39. C. Schatz, C. Pichot, T. Delair, C. Viton and A. Domard, Langmuir 19, 9896 (2003).

40. J. Nuyssink and K. Koopal, Talanta 29, 495 (1982).

41. R. T. Thima and S. Tammishetti, J. Appl. Polym. Sci. 82, 3084 (2001).

42. M. J. Santander-Ortega, N. Csaba, M. J. Alonso, J. L. Ortega-Vinuesa and D. Bastos-González, Colloids Surf. A 296, 132 (2007).

43. J. S. Fordtran and T. W. Locklear, Am. J. Digestive Diseases 11, 503 (1966).

44. P. Calvo, J. L.Vila-Jato, M. J. Alonso, Int. J. Pharm 153, 41 (1997).

45. C. Prego, D. Torres, M.J. Alonso, J.D.D.S.T. 16, 331 (2006).

BLOQUE II:

Aplicación in vitro de Nanopartículas Poliméricas como Sistemas de Liberación Controlada deFármacos

Protein-Loaded PLGA Nanoparticles for Parenteral Administration

M.J. Santander-Ortega1, D. Bastos-González1, J.L. Ortega-Vinuesa1 and M.J. Alonso2

1Biocolloid and Fluid Physics Group, Department of Applied Physics, University of Granada, Av. Fuentenueva S/N, 18071, Granada (Spain)

2Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy, University of Santiago de Compostela, 15706, Santiago de Compostela (Spain)

140 Submitted to Journal of Biophysics

ABSTRACT

Two blend formulations based on a poly(d,l-lactic-co-glycolic) matrix have been used to encapsulate proteins (Albumin and Immuno-γ-Globulin G). One formulation was prepared in presence of Pluronic® F68 while the second one was prepared with Tetronic® 904. Bovine serum albumin (BSA) and Immuno-γ-Globulin G (IgG) were encapsulated in these nanoparticles with three theoretical loadings (1, 2 and 4%) w/w. Once loaded nanoparticles were obtained, some characteristics were analyzed, such as size, ζ-potential.... The different protein loadings did not show a clear effect in the nanoparticles characteristics. Only, when IgG was encapsulated with the higher loading in the formulation prepared with Pluronic® F68 a flocculation of the system was observed. Afterwards one aliquot of each colloidal system was incubated in phosphate buffer saline at 37ºC under horizontal agitation for a period of time of 14 days to study the protein release. Size and ζ-potential were also monitored during the incubation, being constant in this period of time. During the incubation the formulation prepared with Pluronic® F68 whit the highest IgG loading was re-stabilized by hydration forces. Finally, the release profiles were fitted with a power law to test the release mechanism involved in the protein leakage. Both proteins encapsulated in both nanoparticles showed a release mechanism controlled by the diffusion of such proteins through the nanoparticles pores.

Paper VI 141

I. INTRODUCTION

The interest in the application of nanoparticles as controlled drug delivery systems has been increased in the last decades (1). The reasons are obvious, if the carrier can transport the drug to the target zone, it is possible to reach an optimal pharmacologic effect and to reduce undesirable side effects. Concomitantly with the nanoparticles development it has been necessary to develop different types of raw materials and preparation methods to improve their biocompatibility, encapsulation-release or tissue interaction properties (2).

The use of polyesters to produce nanoparticles is well known for their availability, safety and biocompatibility. Among different polyesters, poly(d,l-lactic-co-glycolic) acid (PLGA) micro- and nanospheres have been extensively used as biodegradable colloidal drug carriers (3-5). In this context, the biodegradable polyester, in our case PLGA, becomes a very attractive matrix constituent for parenteral drug administration, since it would allow to control drug release not only by diffusion through the polymer matrix, but also by erosion of the polymer matrix (6). On the other hand, in relation to the preparation methods the double emulsion method (w/o/w) has demonstrated to be a good alternative to encapsulate hydrophilic molecules in nanoparticles prepared with hydrophobic polymers, such as PLGA (4,7).

However, the rapid removal of intravenously administered colloidal drug carriers by the mononuclear phagocytic system (MPS) has been identified as the major obstacle to the efficient targeting of colloidal carriers to target sites (8). Nevertheless, the recognition of the carriers by the MPS can be significantly altered if the surface of the colloidal particles is modified using polyethylene glycol (PEG) derivatives or polyethylene oxide (PEO)/polypropylene oxide (PPO) block copolymers of the poloxamer and poloxamine series (9). These last copolymers are well known for their safety and biocompatibility. It has been suggested that the steric barrier given by these surfactant molecules prevents or restricts the adsorption of plasma proteins onto the particle surface, decreasing recognition by liver and spleen macrophages (10,11).

In addition, it should be noted the effect of the w/o/w formulation technique in the stability of the encapsulated molecule. It has been published for various proteins that when CH2Cl2 is used as organic phase, as in our case, CH2Cl2/water interface causes their aggregation and/or denaturation (12). There are two different manners to avoid this problem. One of them is the increase of the protein concentration (13), while the second one is based in the addition of surfactant molecules to the nanoparticle formulation. The presence of surfactant; such as poloxamers and poloxamines, reduces the protein-interface interactions, but also avoids the protein aggregation (14). Moreover, in the case of nanoparticles formulate with PLGA, the presence of surfactants (i.e. Pluronic® F68 or Tetronic® 904) in the nanoparticle matrix avoids the water induced aggregates and even can help to reduce the initial burst effect (15,16), as well as help to neutralize the acidity

142 Submitted to Journal of Biophysics

generated in the course of PLGA degradation, preserving structural integrity of the encapsulated molecule and thus its biological activity (17).

In a previous study, the physico-chemical characteristics of unloaded blend formulations was analyzed (18). Hence, the aim of this paper has been to study the encapsulation of two well known model proteins in order to investigate the influence of the hydrophobic/hydrophilic degree of these molecules in the final properties of the delivery systems. The proteins used were BSA and IgG at three different theoretical loadings in two blend formulations. The first of them was prepared with PLGA and Pluronic® F68 (PLGA-PF68) and the second one with PLGA and Tetronic® 904 (PLGA-T904). Afterwards, these nanoparticles were incubated in phosphate buffer saline (PBS) pH 7 to see the effect of the specific surfactant and protein in the release properties of these colloidal carriers. Moreover, their size and ζ-potential were monitored during the period of incubation.

II. MATERIALS AND METHODS

2.1. Materials

Polymer poly(D,L-lactic acid/glycolic acid) 50:50 (PLGA) was purchased from Boehringer-Ingelheim, under the commercial name of Resomer® RG 503. The poloxamer Pluronic® F68 (HLB = 29) was from Sigma Aldrich. The poloxamine Tetronic® 904 (HLB = 14.5) was kindly donated by BASCOM Belgium. BSA and IgG were purchased from Sigma Aldrich. All other solvents and chemicals used were of the highest grade commercially available.

2.2. Preparation of PLGA nanoparticles

PLGA blend formulations were prepared by a modified emulsion-solvent diffusion technique (14,19). First, 50 mg of PLGA and 50 mg of the specific polyoxide derivative (Pluronic® F68 or Tetronic® 904) were dissolved in 2 ml of dichloromethane and this organic solution was mixed for 30 s with 0.2 ml of pure water by vortex (2400 min-1, Heidolph). The resulting emulsion (w1/o) was poured under moderate magnetic stirring into a larger polar phase (25 ml ethanol), leading to immediate polymer precipitation in the form of nanoparticles (w1/o/w2). This sample was diluted with 25 ml MilliQ water and stirred for 10 min more. After solvent evaporation under vacuum at 30ºC (Rotavapor Büchi R-114, Flawil) PLGA-PF68 and PLGA-T904 nanoparticles were collected and dissolved in an aqueous medium.

2.3. Physicochemical characterization of nanoparticles

Size, polydispersity (PDI) and ζ-potential of the nanoparticles were determined, respectively, by photon correlation spectroscopy (PCS) and laser Doppler anemometry using a Zetasizer® Nano-ZS (Malvern Instrument). Briefly, ζ-potential was calculated from electrophoretic mobility data by using the equation of Henry:

Paper VI 143

32 ( )

epotentialf aμ ηζε κ

− = (1)

where μe is the electrophoretic mobility, η the viscosity of the medium, ε the dielectric constant and f(κa) the Henry function. A value of 1.5 was estimated by Smoluchowski approximation for f(κa).

Size was calculated from the Brownian movement of the nanoparticles by using the Stokes-Einstein equation:

6bk TD

aπη= (2)

where D is the diffusion coefficient, kb is the constant of Boltzman, T the temperature, η the viscosity of the medium and a the radius of the nanoparticles.

2.4. Encapsulation of BSA and IgG

BSA or IgG was introduced into the internal aqueous phase (w1) of the PLGA-PF68 and PLGA-T904 nanoparticles before the emulsification step and encapsulated with three increasing theoretical loadings 1, 2 and 4% (w/w). Loadings were calculated with respect to the total weight of the PLGA in the formulation.

The exact quantity of encapsulated protein was calculated by an indirect way analyzing the amount of protein that remained in the recollected supernatant after centrifugation (30 min at 15000×g and 15ºC) (Avanti 30, Beckman) the nanoparticles suspension. Protein concentration in the supernatant was determined spectrophotometrically by the micro BCA protein assay (Pierce, IL, USA).

2.5. In vitro release study of BSA and IgG

After the loaded nanoparticles preparation, they were recollected and concentrated in phosphate buffer saline (PBS) pH 7. The final concentration of this suspension was adjusted to 2mg nanoparticles/ml. Samples were incubated at 37ºC with horizontal shaking (120 min−1, Promax 1020, Heidolph, Schwabach, Germany), and they were analyzed at predetermined time intervals (3h, 1, 2, 3, 7 and 14 days) by centrifugation for 30 min at 15.000×g and 15ºC (Avanti 30, Beckman). The amount of protein released at each time point was determined from the isolated supernatants by the micro BCA protein assay mentioned above. The evolution of size and ζ-potential of the nanoparticles were also determined, respectively, by PCS and laser Doppler anemometry using a Zetasizer® Nano-ZS (Malvern Instrument).

2.6. Kinetics analysis of BSA and IgG release

To obtain a better understanding of the process involved in the release of the proteins from PLGA nanoparticles, the semi-empirical power law expression for

144 Submitted to Journal of Biophysics

the analysis of the drug release from non-swelling spherical nanoparticles was used (20):

ntM ktM∞

= (3)

Where Mt/M∞ is the drug fraction released at time t, k is a constant that incorporates characteristics of the polymeric network and the drug and n is the diffusional exponent, which indicates the transport mechanism. Usually, Fickian diffusion from a thin film is defined by n = 0.5 and non-Fickian diffusion is found for n > 0.5 (according to the classical Higuchi equation (21)). However, these values can not be used to characterize the release from spheres, due to the fact that k and the diffusional exponent present a geometric dependency. Ritger and Peppas (22), defined a new limit values for n as a function of the geometry of the release device and its size distribution. These authors published that Fickian diffusion from spheres it is characterized by n = 0.43, non-Fickian diffusion obeys 0.43 < n < 1.0, and a Zero-order release velocity is reached for n = 1.0. Moreover, it is worthy to remark that these values were calculated for mono-disperse spheres, and that calculation for polydisperse samples showed lower values of n (22). The quality of the fit was evaluated by using the goodness-of-fit parameter (χ2).

III. RESULTS AND DISCUSSION

3.1. Encapsulation of BSA and IgG

Tables 1 and 2 illustrate, respectively, the effect of the BSA and IgG encapsulation in the main physicochemical characteristics of PLGA-PF68 and PLGA-T904 nanoparticles, showing all the systems an encapsulation efficacy higher than 75%. Let’s start with the encapsulation of BSA (see Tab. 1). Regarding to the PLGA-PF68 systems, encapsulation of BSA did not produces a clear effect in their size or PDI. However, the ζ-potential magnitude showed a slight increase when BSA was incorporated to the nanoparticles. This change was almost independent of the protein loading. This increase can be attributed to the accumulation of, at least, part of the protein on the surface of the nanoparticles (23,24). For PLGA-T904 formulations neither size nor ζ-potential magnitude displayed a clear dependence with the encapsulation of BSA. The almost null effect of the encapsulation of BSA on the ζ–potential magnitude of this system could suggest a lower accumulation of the protein in the nanoparticle-water interface (23,24).

With respect to the encapsulation of IgG (see Tab. 2), the main physicochemical characteristics of PLGA-PF68 nanoparticles displayed a clear correlation with the protein loading. Their size and PDI increased substantially as a function of the IgG loading. On the other hand, its ζ-potential magnitude decreased as a result of the incorporation of the macromolecule to the nanoparticles. The dependence of the ζ–potential with the loading suggests that the protein was accumulated on the surface of these nanoparticles (25) as IgG is practically

Paper VI 145

uncharged at neutral pH (at which the ζ–potential was measured). Moreover, the low ζ–potential magnitude of the nanoparticles found with the highest loading, around 5 mV (see Tab. 2), was the responsible of the low stability showed by these PLGA-PF68 nanoparticles (25,26). Under these pH conditions nanoparticles presented a net charge close to zero, and thus, the repulsive electrostatic interaction between them practically did not exist; consequently, the system tended to flocculate (25).

Table 1. Hydrodynamic mean size, PDI, ζ-potential and theoretical loading (TL) of encapsulated BSA nanoparticles.

Size (nm) PDI ζ-Pot(mV) TL (%) PLGA-PF68 130,1 ±1,1 0.142 ± 0.020 - 12,4 ± 0,4 ---- 154,9 ± 8,6 0.150 ± 0.086 -19,4 ± 1,5 1 146,6 ± 5,3 0.151 ± 0.011 -20,5 ± 0,5 2 155,1 ± 7,5 0.155 ± 0.023 -20,0 ± 0,7 4 PLGA-T904 137,1 ± 6,6 0.168 ± 0.032 -17,2 ± 2,6 ---- 131,8 ± 3,1 0.161 ± 0.016 -14,9 ± 0,6 1 133,7 ± 5,2 0.164 ± 0.006 -18,5 ± 0,8 2 130,8 ± 9,2 0.168 ± 0.030 -19,0 ± 1,0 4

Table 2. Hydrodynamic mean size, PDI, ζ-potential and theoretical loading (TL) of encapsulated IgG nanoparticles.

Size (nm) PDI ζ-Pot(mV) TL (%) PLGA-PF68 130,1 ±1,1 0.142 ± 0.020 - 12,4 ± 0,4 ---- 192,8 ± 2,8 0.097 ± 0.013 -8,5 ± 0,1 1 259,7 ± 9,0 0.119 ± 0.031 -7,8 ± 0,4 2 354,1 ± 17,8 0.644 ± 0.078 - 5,1 ± 0,5 4 PLGA-T904 137,1 ± 6,6 0.168 ± 0.032 -17,2 ± 2,6 ---- 198,7 ± 11,4 0.166 ± 0.030 -18,6 ± 1,2 1 200,8 ± 11,8 0.137 ± 0.040 -13,3 ± 0,3 2 192,3 ± 6,4 0.221 ± 0.035 -12,9 ± 0,9 4

It is worthy to remark the behaviour showed by the PLGA-T904 formulations. Although this system presented a parallel behaviour to that of PLGA-PF68, the change in size, PDI and ζ-potential as a function of the protein loading was less sharp than that observed with the poloxamer. Hence, the obtained results with both proteins suggest a higher surface localization of the protein for PLGA-PF68 formulation than for PLGA-T904. Encapsulation of macromolecules in nanoparticles it is a complex process controlled by several factors, such as the tensioactive properties of the encapsulated macromolecule and the stability of the w1/o emulsion (27,28). Studies carried out previously in our laboratory demonstrated that, for these nanoparticles, the first emulsion prepared in presence

146 Submitted to Journal of Biophysics

of the poloxamer was much more stable than that prepared with the poloxamine (19), that is, this poloxamer presents higher surfactant properties than the poloxamine. For this reason, during the formation of these systems it is possible that the poloxamer displaces the proteins easier than the poloxamine from the w1/o interface toward the external phase. Accordingly, it could be possible to obtain different protein distributions as a function of the surfactant used in the nanoparticle formation.

3.2. Size and ζ-Potential evolution of the loaded nanoparticles

Once the colloidal systems were formulated in the presence of the specific protein, the next step was incubation in PBS at 37ºC under horizontal stirring for a period of 14 days. The hydrodynamic mean size and PDI of these systems were measured at different times during the incubation.

0 2 4 6 8 10 12 14150

200

250

300

350

400

Siz

e (n

m)

Time (days)

(b)

0 2 4 6 8 10 12 14150

200

250

300

350

400

Siz

e (n

m)

Time (days)

(a)

Figure 1. Size (a) and PDI (b) evolution for PLGA-PF68 nanoparticles with encapsulated IgG incubated in PBS at 37ºC. 1 % ( ), 2 %( ) and 4 %( ).

When BSA was chosen as encapsulated protein, both systems, PLGA-PF68 and PLGA-T904, did not show a clear change in their size and PDI during the incubation time (data not shown). Analogous results were obtained when IgG was encapsulated in the PLGA-T904 nanoparticles (data not shown). Nevertheless, as it is depicted in Figures 1a-b, PLGA-PF68 loaded with IgG displayed a size and PDI evolution dependent on the protein loading. In this case, the nanoparticles with the highest loading (4 %) showed a clear decrease in their size and PDI during the first day of incubation. Both magnitudes changed from values of 350 nm to 225 nm in size and 0.65 to 0.15 in PDI. After this first incubation stage, size and PDI remained constants. As no sedimentation was observed in the incubation samples, it is necessary to analyze de effect of the incubation medium in the re-dispersation of the flocculated system. Nanoparticles were prepared in MilliQ water -medium with a close to zero ionic strength- obtaining an un-stable colloidal system. However, when the system was incubated in PBS, with a high ionic strength of around 150 mM, its hydrodynamic mean size and PDI started to decrease. Once PLGA-PF68

Paper VI 147

nanoparticles with a 4% of IgG reach the size and PDI values of the formulations with a lower loading the values were constants during the rest of the incubation. This kind of re-stabilization phenomenon when nanoparticles are incubated in mediums with high salt concentration is well known, and it is governed by hydration forces (29-34). Re-stabilization due to hydration forces can not be explaining on the basis of the typical potentials of interaction –van der Wall and electrostatic- involved in the classical DLVO theory (26). In this case the hydrophilicity of the nanoparticles surface and the ionic strength of the medium must be considered. It has been demonstrated previously that the incorporation of Pluronic® F68 and IgG onto PLGA nanoparticles changes from hydrophobic to hydrophilic the character of their surface (35,36). This change in the hydrophilicity of the PLGA surface together with the high ionic strength of PBS suggests that the re-stabilization process observed for PLGA-PF68 nanoparticles with an IgG loading of 4% was totally mediated by hydration forces.

As was commented above, once this system was totally re-stabilized -one day after its incubation in PBS- its behaviour was the same that those showed by the other systems, it is mean, their size and PDI were constant. According to these results, an incubation of two weeks seems to be a short time period to appreciate degradation of the blend formulations. Similar results were also obtained by other authors working with PLGA nanoparticles (16,37). However, it is known that the PLGA nanoparticles degradation in the first stages is controlled by a surface-core mechanism, due to the acid conditions generated by the rupture of the ester bonds in the nanoparticles core (38). Hence these results suggest that these nanoparticles loaded with proteins did not suffer matrix degradation.

To obtain a better understanding of the patterns involved in the protein release, the ζ-potential value of the loaded nanoparticles was also monitored as a function of the incubation time. Nanoparticles loaded with BSA showed a constant

0 2 4 6 8 10 12 14 16-40

-35

-30

-25

-20

-15

-10

-5

0

ζ-Po

tent

ial (

mV)

Time (days)

(a)

0 2 4 6 8 10 12 14-40

-35

-30

-25

-20

-15

-10

-5

0

ζ-Po

tent

ial (

mV)

Time (days)

(b)

Figure 2. ζ–Potential evolution for PLGA-PF68 (a) and PLGA-T904 (b) nanoparticles with encapsulated IgG incubated in PBS at 37ºC. 1 % ( ), 2 %( ) and 4 %( ).

148 Submitted to Journal of Biophysics

ζ-potential value during the incubation time (data not shown). Similar results were obtained with the IgG loaded nanoparticles, Figures 2a-b illustrate the behaviour of these complexes. Both formulations presented similar behaviours that in the case of the BSA. However, in this situation, the ζ-potential magnitude of PLGA-PF68 and PLGA-T904 was ordered along the in vitro study as a function of the protein loading, being higher (in absolute value) for the nanoparticles with a lower loading. As it will be shown in the next section the release of the proteins from both types of nanoparticles seemed to be not enough to change the superficial electrokinetic state of the nanoparticles.

3.3. In vitro release of BSA and IgG

The release profiles of BSA and IgG from both blend formulations with a loading of 0.8% are depicted in Figure 3. The experimental results are represented by dots and the fitting to the power law by lines -see equation (3)-. Similar results were found with the others loadings. This kind of behaviour agree with that published previously by Wong et al (39). With this study it was possible to see the effect of the encapsulated molecule and the poly-oxide derivative -Pluronic® F68 or Tetronic® 904- in the liberation of the protein from PLGA nanoparticles. BSA presented a biphasic pattern with both blend formulations, characterized by and initial burst effect of around 50% followed by a slower release phase. Regarding to the IgG release, both blend formulations presented different patterns. PLGA-PF68 displayed an initial burst effect of around 20 %, much lower than in the BSA case, followed by a smooth release kinetic. However, PLGA-T904 presented a negligible

0 2 4 6 8 10 12 140

10

20

30

40

50

60

70

80

90

100

Rel

ease

(%)

Time (days)

Figure 3. Release profile of BSA (rhombus) and IgG (triangle) from PLGA-PF68 (close symbols) and PLGA-T904 (open symbols) nanoparticles incubated in PBS at 37ºC with a loading of 1 %(w/w).Lines -solid for PLGA-PF68 and dashed for PLGA-T904- are the fitting of the power law to the experimental data.

Paper VI 149

burst effect followed by a nearly linear release velocity. Sandor et al (40) found similar differences between the burst effect displayed by BSA and IgG. The burst effect is associated with the leakage of the surface-associated protein. At pH 7 BSA is negatively charge (i.e.p. 4.7) while IgG with an i.e.p. in the 6-8 range is practically uncharged. Hence, BSA presented a higher hydrophilic character and the same sign of the negative PLGA matrix under these conditions. As a consequence BSA molecules do not show a good affinity for the surface of the nanoparticles, which favors the leakage of the protein (19). Moreover, it has been also demonstrated that the diffusion coefficient (Dc) of each protein is a relevant parameter during the first stage of protein release (41). BSA presents a Dc in the 10-7-10-9 cm2/s range (42,43), while for IgG Dc is around 10-12cm2/s (39). It is important to note that although IgG display a lower Dc, its value is much higher than those of hydrophobic molecules (41). Therefore, parameters such as, hydrophilic character, i.e.p. and Dc of the encapsulated molecule must be considered for the development of drug delivery systems.

The effect of the surfactant in the release profile was only clear in the case of the IgG, while with BSA, due to the high burst effect, it was not possible to appreciate any difference between both surfactants. Regarding to the IgG the major release was found when Pluronic® F68 was used instead of Tetronic® 904 as surfactant. It can be attributed to the higher hydrophobic character of the latter, in comparison to the former. It has been shown that the presence of poloxamines with low HLB values (i.e. Tetronic® 904) in PLGA nanoparticles reduce the degradation of the PLGA matrix slowing down the pore formation (16). Consequently, this reason might be responsible for the observed delay in the IgG release in PLGA-T904 nanoparticles.

The application of the power law to the experimental release data for BSA (see Fig. 3) gave n values of 0.10 (χ2~2 10-4) and 0.15 (χ2~20 10-4) for PLGA-PF68 and PLGA-T904 respectively. When IgG was the encapsulated protein PLGA-PF68 had a n value of 0.18 (χ2~6 10-4), while PLGA-T904 had a value of 0.33 (χ2~6 10-4). These n values are lower than the 0.43-1 range calculated by Ritger and Peppas (22). Leo et al (44) also obtained n values lower than 0.43. The decrease of n with respect to the theoretical value can be attributed to the polydispersity of these nanoparticles (22). Hence the obtained n value suggest that the release of the proteins it was not controlled by the degradation of the matrix. Therefore, the release was governed by the diffusion of the proteins through the nanoparticles pores. Consequently, the application of the power law to the experimental results confirms that during the incubation time both nanoparticles –those formulated with Pluronic® F68 or Tetronic® T904- did not loss their integrity, and the release of both proteins was controlled by diffusion through the nanoparticles pores, being higher the diffusion of the smallest protein (BSA).

IV. CONCLUSIONS

150 Submitted to Journal of Biophysics

The effect of the encapsulation of two different proteins in two different PLGA blend formulations has been studied. The ζ-potential measurement reveals that the type of surfactant present in the blend formulation cause different protein distribution, presenting a higher protein surface concentration when Pluronic® F68 was chosen as surfactant. Regarding to the stability of these nanoparticles, both blend formulations were stable for an incubation period of 14 days.

On the other hand, the amount of the encapsulated protein in the studied range did not have a clear effect in the nanoparticles behaviour. However, the type of encapsulated protein had a clear effect in the release profiles, presenting a reduction in the burst effect when the more hydrophobic protein was encapsulated. This burst effect was even eliminated in the PLGA-IgG complexes when Tetronic® 904 was used in the blend formulation.

ACKNOWLEDGEMENTS

Authors thank the financial support from the “Comisión Interministerial de Ciencia y Tecnología” Projects MAT2006-13646-C03-03 and MAT2007-66662-C02-01 (European FEDER support included) and from the “Consejería de Innovación, Ciencia y Tecnología de la Junta de Andalucía” Projects P07-FQM-2496 and P07-FQM03099.

Paper VI 151

References

1. Kreuter, J. Nanoparticles - a historical perspective. International Journal of Pharmaceutics 331: 1-10; 2007.

2. Soppimath, K. S.; Aminabhavi, T. M.; Kulkarni, A. R.; Rudzinski, W. E. Biodegradable polymeric nanoparticles as drug delivery devices. J.Control.Release 70: 1-20; 2001.

3. Langer, R. New Methods of Drug Delivery. Science 249: 1527-1533; 1990.

4. Csaba, N.; Caamano, P.; Sanchez, A.; Dominguez, F.; Alonso, M. J. PLGA : poloxamer and PLGA : poloxamine blend nanoparticies: New carriers for gene delivery. Biomacromolecules 6: 271-278; 2005.

5. Alonso, M. J. Nanoparticulate drug carrier technology. In: Cohen, S.; Berstein, H., eds. Microparticulate Systems for the Delivery of Proteins and Vaccines. New York: Marcel Dekker Inc.; 1996: 203-242.

6. Schindler, A.; Jeffcoat, R.; Kimmel, G. L.; Pitt, G. C.; Wall, M. E.; Zweidinger, R. Biodegradable polymers for sustained drug delivery. Contemporary Topics in Polymer Sciences 2: 251-289; 1977.

7. Csaba, N.; Gonzalez, L.; Sanchez, A.; Alonso, M. J. Design and characterisation of new nanoparticulate polymer blends for drug delivery. Journal of Biomaterials Science-Polymer Edition 15: 1137-1151; 2004.

8. Poste, G. Liposome Targeting in Vivo: Problems and Oportunities. Biology of the Cell 47: 19-38; 1983.

9. Moghimi, S. M.; Hunter, A. C. Poloxamers and poloxamines in nanoparticle engineering and experimental medicine. Trends Biotechnol. 18: 412-420; 2000.

10. Illum, L.; Jacobsen, L. O.; Muller, R. H.; Mak, E.; Davis, S. S. Surface characteristics and the interaction of colloidal particles with mouse peritoneal macrophages. Biomaterials 8: 113-117; 1987.

11. Tan, J. S.; Butterfield, D. E.; Voycheck, C. L.; Caldwell, K. D.; Li, J. T. Surface modification of nanoparticles by PEO/PPO block copolymers to minimize interactions with blood components and prolong blood circulation in rats. Biomaterials 14: 823-833; 1993.

12. Brigger, I.; Dubernet, C.; Couvreur, P. Nanoparticles in cancer therapy and diagnosis. Adv.Drug Deliv.Rev. 54: 631-651; 2002.

13. Perez, C.; Castellanos, I. J.; Costantino, H. R.; Al-Azzam, W.; Griebenow, K. Recent trends in stabilizing protein structure upon encapsulation and release from bioerodible polymers. J.Pharm.Pharmacol. 54: 301-313; 2002.

14. Csaba, N.; Gonzalez, L.; Sanchez, A.; Alonso, M. J. Design and characterisation of new nanoparticulate polymer blends for drug delivery. Journal of biomaterials science 15: 2004.

15. Kumar, P. S.; Ramakrishna, S.; Saini, T. R.; Diwan, P. V. Influence of microencapsulation method and peptide loading on formulation of poly(lactide-co-glycolide) insulin nanoparticles. Pharmazie 61: 613-617; 2006.

16. Yeh, M. K.; Davis, S. S.; Coombes, A. G. A. Improving protein delivery from microparticles using blends of poly(DL lactide co-glycolide) and poly(ethylene oxide)-poly(propylene oxide) copolymers. Pharmaceutical Research 13: 1693-1698; 1996.

17. Tobio, M.; Schwendeman, S. P.; Guo, Y.; McIver, J.; Langer, R.; Alonso, M. J. Improved immunogenicity of a core-coated tetanus toxoid delivery system. Vaccine 18: 618-622; 1999.

152 Submitted to Journal of Biophysics

18. Santander-Ortega, M. J.; Bastos-González, D.; Ortega-Vinuesa, J. L.; Alonso, M. J. Insulin-Loaded PLGA nanoparticles for oral administration: An in vitro physico-chemical characterization. Journal of Biological Nanotechnology.Submited 2007.

19. Santander-Ortega, M. J.; Bastos-Gonzalez, D.; Ortega-Vinuesa, J. L.; Alonso, M. J. Insulin-Loaded PLGA nanoparticles for oral administration: An in vitro physico-chemical characterization. Journal of Biological Nanotechnology.Accepted 2008.

20. Peppas, N. A. Analysis of Fickian and non-Fickian drug release from polymers. Pharmaceutica acta Helvetiae 60: 1985.

21. HIGUCHI, T. Rate of release of medicaments from ointment bases containing drugs in suspension. J.Pharm.Sci. 50: 874-875; 1961.

22. Ritger, P. L.; Peppas, N. A. A simple Equation for Description of Solute Release I: Fickian and non-Fickian Release from non-Sweliable Devices in the form of Slabs, Spheres, Cyliders or Discs. J.Control.Release 5: 1987.

23. Peula, J. M.; Hidalgo-Alvarez, R.; de las Nieves, F. J. Coadsorption of IgG and BSA onto sulfonated polystyrene latex: II. Colloidal stability and immunoreactivity. J.Biomater.Sci.Polym.Ed 7: 241-251; 1995.

24. Peula, J. M.; de las Nieves, F. J. Adsorption of monomeric BSA on sulfonated polystyrene model colloids. 3. Colloid and Surfaces A:Physicochemical and Engineering aspects 90: 55-62; 1994.

25. Santander-Ortega, M. J.; Bastos-Gonzalez, D.; Ortega-Vinuesa, J. L. Electrophoretic mobility and colloidal stability of PLGA particles coated with IgG. Colloids and Surfaces B-Biointerfaces 60: 80-88; 2007.

26. Hiemenz, P. C.; Rajagopalan, R. Principles of Colloidal and Surface Chemistry. New york: Dekker; 1997.

27. De Rosa, G.; Iommelli, R.; La Rotonda, M. I.; Miro, A.; Quaglia, F. Influence of the co-encapsulation of different non-ionic surfactants on the properties of PLGA insulin-loaded microspheres. Journal of Controlled Release 69: 283-295; 2000.

28. Schugens, C.; Laruelle, N.; Nihant, N.; Grandfils, C.; Jerome, R.; Teyssie, P. Effect of the Emulsion Stability on the Morphology and Porosity of Semicrystalline Poly L-Lactide Microparticles Prepared by W/O/W Double Emulsion-Evaporation. Journal of Controlled Release 32: 161-176; 1994.

29. Pashley, R. M. Hydration Forces Between Mica Surfaces in Electrolyte-Solutions. Advances in Colloid and Interface Science 16: 57-62; 1982.

30. Churaev, N. V.; Derjaguin, B. V. Inclusion of Structural Forces in the Theory of Stability of Colloids and Films. Journal of Colloid and Interface Science 103: 542-553; 1985.

31. Israelachvili, J. N. Intermolecular and Surface Forces. London: Academic Press; 1992.

32. Molina-Bolivar, J. A.; Galisteo-Gonzalez, F.; Hidalgo-Alvarez, R. Specific cation adsorption on protein-covered particles and its influence on colloidal stability. Colloids and Surfaces B-Biointerfaces 21: 125-135; 2001.

33. Molina-Bolivar, J. A.; Ortega-Vinuesa, J. L. How proteins stabilize colloidal particles by means of hydration forces. Langmuir 15: 2644-2653; 1999.

34. Molina-Bolivar, J. A.; Galisteo-Gonzalez, F.; Hidalgo-Alvarez, R. Colloidal stability of protein-polymer systems: A possible explanation by hydration forces. Physical Review E 55: 4522-4530; 1997.

Paper VI 153

35. Santander-Ortega, M. J.; Jodar-Reyes, A. B.; Csaba, N.; Bastos-Gonzalez, D.; Ortega-Vinuesa, J. L. Colloidal stability of Pluronic F68-coated PLGA nanoparticles: A variety of stabilisation mechanisms. Journal of Colloid and Interface Science 302: 2006.

36. Santander-Ortega, M. J.; Bastos-Gonzalez, D.; Ortega-Vinuesa, J. L. Electrophoretic mobility and colloidal stability of PLGA particles coated with IgG. Colloids and Surfaces B-Biointerfaces 60: 2007.

37. Wu, X. S.; Wang, N. Synthesis, characterization, biodegradation, and drug delivery application of biodegradable lactic/glycolic acid polymers. Part II: Biodegradation. Journal of Biomaterials Science-Polymer Edition 12: 21-34; 2001.

38. Fu, K.; Pack, D. W.; Klibanov, A. M.; Langer, R. Visual evidence of acidic environment within degrading poly(lactic-co-glycolic acid) (PLGA) microspheres. Pharmaceutical Research 17: 100-106; 2000.

39. Wong, H. M.; Wang, J. J.; Wang, C. H. In vitro sustained release of human immunoglobulin G from biodegradable micro-spheres. Industrial and Engineering Chemistry Research 40: 933-948; 2001.

40. Sandor, M.; Enscore, D.; Weston, P.; Mathiowitz, E. Effect of protein molecular weight on release from micron-sized PLGA microspheres. J.Control Release 76: 297-311; 2001.

41. Arifin, D. Y.; Lee, L. Y.; Wang, C. H. Mathematical modeling and simulation of drug release from microspheres: Implications to drug delivery systems. Advanced Drug Delivery Reviews 58: 1274-1325; 2006.

42. Van, T., Sr.; De Geest, B. G.; Braeckmans, K.; De Smedt, S. C.; Siepmann, F.; Siepmann, J.; van Nostrum, C. F.; Hennink, W. E. Mobility of model proteins in hydrogels composed of oppositely charged dextran microspheres studied by protein release and fluorescence recovery after photobleaching. J.Control Release 110: 67-78; 2005.

43. Zhang, Y.; Chu, C. C. Biodegradable dextran–polylactide hydrogel network and its controlled release of albumin. J.Biomedical Materials Research 54: 1-11; 2000.

44. Leo, E.; Scatturin, A.; Vighi, E.; Dalpiaz, A. Polymeric nanoparticles as drug controlled release systems: a new formulation strategy for drugs with small or large molecular weight. J Nanosci.Nanotechnol. 6: 3070-3079; 2006.

Insulin-Loaded PLGA Nanoparticles for Oral Administration: An in vitro Physico-chemical

Characterization

M.J. Santander-Ortega1, D. Bastos-González1, J.L. Ortega-Vinuesa1 and M.J. Alonso2

1Biocolloid and Fluid Physics Group, Department of Applied Physics, University of Granada, Av. Fuentenueva S/N, 18071, Granada (Spain)

2Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy, University of Santiago de Compostela, 15706, Santiago de Compostela (Spain)

156 Accepted for publication in J. of Biomed. Nanotech.

ABSTRACT

The aim of this work was to study poly(d,l-lactic-co-glycolic) acid (PLGA) nanoparticles – formulated by a modified solvent diffusion technique – applied as model nanocarriers for insulin in potential oral administrations. These nanostructures consisted of a blend matrix formed by PLGA copolymer and polyoxyethylene derivatives. Two types of blend formulations, PLGA:poloxamer (Pluronic® F68) and PLGA:poloxamine (Tetronic® T904), were analyzed, and the results compared to those obtained with pure PLGA nanoparticles. The work has been divided into two parts. a) Firstly, the stability of the unloaded nanoparticles in simulated gastric and intestinal fluids was studied. Degradation studies reflected a strong interaction between the pure PLGA nanoparticles and the digestive enzymes. However, this interaction was considerably reduced in the blend formulations, although the PLGA:poloxamine system became colloidally unstable in the simulated gastric fluid. b) Secondly, the effect of the net charge of the encapsulated macromolecule in the final properties of the blend formulations was studied by encapsulating insulin below and above its corresponding isoelectric point. The net charge of the encapsulated protein showed a clear effect in the final size of the nanoparticles, while the encapsulation efficiency was controlled by the polyoxyethylene derivative presents in the blend formulation. The obtained results show that those carriers formed with encapsulated insulin in PLGA-Pluronic® F68 particles are capable, at least in vitro, to overcome the gastrointestinal barrier. Therefore, these nanocarriers seem to be appropriate for oral administration of insulin, although performing in vivo studies becomes necessary to corroborate such statement.

Paper VII 157

I. INTRODUCTION

At present, the typical mode to administrate insulin is via intramuscular, subcutaneous or intravenous injection. Injections in general are not ideal methods for the administration of drugs, especially for long-term treatments, being the oral route far more accepted by patients. However, several problems must still be solved before insulin can be efficiently administered by the oral route, such as acid-catalyzed degradation in the stomach, proteolytic breakdown in the gastrointestinal tract, and absorption through the intestinal epithelium [1; 2]. Encapsulation of drugs or proteins in nanoparticle carriers can help to overcome these problems as the therapeutic molecule can be protected against degradation in biological environments. The use of polyesters as polymeric drug delivery systems is well known for their availability, safety and biocompatibility. Among different polyesters, poly(d,l-lactic-co-glycolic) acid (PLGA) micro- and nanospheres have been extensively used as biodegradable colloidal drug carriers [3-5]. In this context, the biodegradable polyester, in our case PLGA, becomes a very attractive matrix constituent for oral administration of insulin, since it would allow to control drug release not only by diffusion through the polymer matrix, but also by erosion of the polymer matrix [6]. However, there is evidence that biodegradable polyester nanoparticles, such as those made with poly(lactic acid) PLA, suffer a significant degradation in intestinal fluids and this degradation is affected by the surface composition of the particles [7; 8]. The incorporation of polyoxyethylene derivatives to the PLA or PLGA matrix can reduce the degradation of these systems [7; 9; 10]. Furthermore, the use of this type of surfactants can improve the stability of the encapsulated protein or drug [11-13]. In this work, we have chosen two polyoxyethylene derivatives that differ in their hydrophobic/philic character, namely, a poloxamer (Pluronic® F68, HLB=29) and a poloxamine (Tetronic® 904, HLB=14.5), obtaining PLGA-PF68 and PLGA-T904 blend formulations, respectively. These systems have been studied previously with respect to their physico-chemical characteristics [14], and with respect to their parenteral and nasal administration [4; 14 -16], obtaining promising results. For this reason, it could be interesting to increase the characterization of these systems, in order to know their possible application to other administration routes. On the other hand, it could be interesting to carry out the encapsulation of a unique macromolecule (e.i. a protein), but under different conditions that may change its physico-chemical properties, for example its charge state. The goal would be focused on the evaluation of possible modifications produced in the final properties of the nanocarriers by inserting the protein with different sign of charge. These results could serve to get a better understanding of the process that control the encapsulation and release of macromolecules in PLGA nanoparticles.

With these ideas in mind, the present work was carried out presenting two main objectives. The first objective was to investigate the colloidal stability and degradation of PLGA, PLGA-PF68 and PLGA-T904 nanoparticles (without encapsulated insulin) in simulated gastro and intestinal media, so that, it is well

158 Accepted for publication in J. of Biomed. Nanotech.

established that factors as pH [17; 18], and the HLB of the surfactant present in the nanoparticles [11] can modify the stability and degradation patterns of polyester matrix. The second objective of this study was aimed to the encapsulation of insulin into these PLGA nanocarriers. Specifically, the effect of the net charge of the protein in the final properties of the blend formulations was studied. This last study was carried out loading positive or negative insulin in PLGA-F68 and PLGA-T904 nanoparticles by controlling the medium pH. Afterwards, these formulations with encapsulated insulin were incubated in simulated gastro-intestinal fluids, in order to know the effect produced by both the net charge of the protein and by the polyoxyethylene derivative in the final characteristics of the nanoparticles.

II. MATERIALS AND METHODS

2.1. Materials

Polymer poly(D,L-lactic acid/glycolic acid) 50:50 (PLGA) was purchased from Boehringer-Ingelheim, under the commercial name of Resomer® RG 503. The Pluronic® F68 poloxamer was from Sigma Aldrich. The Tetronic® 904 poloxamine was kindly donated by BASCOM, Belgium. Bovine insulin, pepsin A from porcine stomach mucosa, and pancreatin from porcine pancreas were purchased from Sigma Aldrich. All other solvents and chemicals used were of the highest grade commercially available.

2.2. Preparation of PLGA nanoparticles

PLGA nanoparticles were prepared by a modified emulsion-solvent diffusion technique. First, 50 mg of PLGA were dissolved in 2 ml of dichloromethane and this organic solution was mixed for 30 s with 0.2 ml of pure water by vortex (2400 min-1, Heidolph). The resulting emulsion (W1/O) was poured under moderate magnetic stirring into a larger polar phase (25 ml ethanol), leading to immediate polymer precipitation in the form of nanoparticles. In some cases, the stability of this primary W1/O emulsion was analyzed by PCS using a Turbiscan Classic MA 2000 (Formulaction). The solution with the nanoparticles was then diluted with 25 ml MilliQ water and stirred for 10 min more (W1/O/W2). After solvent evaporation under vacuum at 30ºC (Rotavapor Büchi R-114) nanoparticles were collected and dissolved in an aqueous medium.

Nanoparticles with PLGA:poloxamer and PLGA:poloxamine ratios of 50:50 were also prepared by a modified emulsion-solvent diffusion technique in an identical way to that explained above. The unique difference was that, in this case, the organic phase was composed by 50 mg of a specific surfactant, poloxamer or poloxamine, together with the PLGA polymer (50 mg) dissolved in 2 ml of dichloromethane. More details of the synthesis are described elsewhere [4; 19].

2.3. Physicochemical characterization of nanoparticles

Particle size and polydispersity were determined by photon correlation spectroscopy (PCS), while ζ-potential was measured by means of laser Doppler

Paper VII 159

anemometry, using in both cases the same apparatus: Zetasizer® Nano-ZS (Malvern Instrument). Mean ζ-potential values were taken from the average of six measurements.

2.4. Degradation studies of PLGA, PLGA-PF68 and PLGA-T904 nanoparticles in model gastrointestinal fluids.

PLGA, PLGA-PF68 and PLGA-T904 nanoparticles were incubated in simulated gastric (pH 1.2, pepsin 0.32%w/v) and intestinal (pH 7, pancreatin 1%w/v) fluids prepared according to USP XXIX at a concentration equal to 1 mg nanoparticles/ml. Samples were collected at predetermined time intervals (0.2, 0.5, 1.5, 3 and 6 hours). Nanoparticles that interacted with gastrointestinal fluids components were separated by centrifugation for 2 min at 1000xg and 15ºC (Avanti 30, Beckman). Size of the remaining nanoparticles was measured by photon correlation spectroscopy (PCS) using a Zetasizer® Nano-ZS (Malvern Instruments). Likewise, nanoparticles degradation was also followed by using a lactate detection procedure (Lactate diagnostic kit, Sigma).

2.5. Encapsulation of insulin

Bovine insulin was introduced into the internal aqueous phase (0.2 ml) of the PLGA-PF68 and PLGA-T904 formulations before the emulsification step and encapsulated with a 1% w/w theoretical loading. Theoretical loadings were calculated with respect to the total PLGA weight in the formulation. The isoelectric point of insulin was equal to 5.6. In order to study the effect of the charge of this protein during its encapsulation, it was dissolved in non-buffered solutions – HCl 0.01M (pH 2) or NaOH 0.01M (pH 12) – working in this way with positive and negative insulin molecules.

Encapsulation efficiency (EE) was calculated from the amount of insulin that remained in the supernatant samples after centrifugating the nanoparticles suspension for 30 min at 15000×g and 15ºC (Avanti 30, Beckman). The amount of this insulin was determined spectrophotometrically by a micro BCA protein assay (Pierce, IL, USA).

The evolution of size and ζ-potential of these nanoparticles were independently determined by using the Zetasizer® Nano-ZS (Malvern Instrument).

2.6. In vitro release of insulin

As just mentioned, PLGA-PF68 and PLGA-T904 nanoparticles were prepared with a 1% theoretical loading of insulin, according to the solvent diffusion technique described above. After the solvent evaporation step, nanoparticles were collected and concentrated in simulated gastric and intestinal media, both of them free of enzymes to avoid degradation of the released insulin. The final concentration of this suspension was adjusted to 2mg nanoparticles/ml. These samples were incubated at 37ºC and horizontally shaken (120 min−1, Promax 1020, Heidolph). Subsequently, samples were collected at predetermined time intervals

160 Accepted for publication in J. of Biomed. Nanotech.

(3h, 1, 2, 3, 7 and 14 days) and centrifugated for 30 min at 15.000×g and 15ºC (Avanti 30, Beckman). The amount of insulin released at each time was determined from the isolated supernatants by the micro BCA protein assay (Pierce, IL, USA).

III. RESULTS AND DISCUSSION

3.1. Physicochemical characterization of nanoparticles

Table 1. Mean nanoparticle size, ζ-potential and EE values for PLGA, PLGA-PF68, PLGA-T904, PLGA-PF68 ins-pH12, PLGA-PF68 ins-pH2, PLGA-T904 ins-pH12 and PLGA-T904 ins-pH2 nanoparticles in GF and IF (n=9, ± s.d.).

Gastric Fluids Intestinal Fluids d (nm) ζ-Pot (mV) d (nm) ζ-Pot (mV) EE (%)

PLGA Totally unstable 154,0 ± 4,0 - 50 ± 5 ----- PLGA-PF68 126,5 ± 0,7 1 ± 5 128,1 ± 1,1 -12 ± 5 ----- PLGA-T904 128,5 ± 0,7 14 ± 5 130,0 ± 7,0 -17 ± 5 ----- PLGA-PF68 ins-pH12 126,0 ± 0,1 4 ± 5 127,1 ± 6,0 -8 ± 5 34,7± 1,1 PLGA-PF68 ins-pH2 255,2 ± 2,4 1 ± 5 253,2 ± 2,3 -4 ± 5 40,6 ± 1,3 PLGA-T904 ins-pH12 128,0 ± 0,1 19 ± 5 128,4 ± 5,0 -18 ± 5 9,6 ± 2,0 PLGA-T904 ins-pH2 174,9 ± 1,3 17 ± 5 179,3 ± 7,0 -16 ± 5 11,1 ± 1,1

In order to separate aggregation and degradation phenomena in the simulated gastrointestinal fluids, the first set of experiments was performed in absence of enzymes. The three PLGA systems were separately incubated in gastric fluid (GF) and intestinal fluid (IF); if any aggregation process took place, it would be caused exclusively by the physicochemical conditions (pH and ionic strength) of these simulated fluids. Table 1 shows the mean diameter and ζ-potential of the three systems in GF and IF. The pure PLGA nanoparticles were totally unstable in GF. It should be noted that the pH value of this medium is 1.2; at this acidic pH, the PLGA carboxylic endgroups are completely protonated [14]. Therefore, our PLGA nanoparticles present a net charge very close to zero, and thus, the repulsive electrostatic interaction between them practically does not exist; consequently, the system tends to aggregate rapidly. The PLGA-PF68 particles also suffer charge cancellation in GF, (i.e. see the corresponding ζ-potential value, which is almost null), as the presence of a non-ionic surfactant (poloxamer) does not alter the PLGA electrical state. However, this system remains totally stable in GF. The origin of this colloidal stability is purely steric, and it is given by the aqueous solvency of the poloxamer polar chains located at the particle/water interface [15]. The high stability found in GF is a clear indication of the incorporation of the poloxamer molecules to the PLGA particles during the synthesis process. The other blend formulation (PLGA-T904) showed a clear positive net charge, according to its ζ-potential value. This positive value is due to the presence of poloxamine molecules on the surface of these nanoparticles. The Tetronic® T904 organic structure presents nitrogen atoms, which are fully protonated under extreme acidic conditions. Therefore, the high stability of this sample in GF does not come only from the steric

Paper VII 161

hindrance given by the surfactant polyoxyethylene chains extended toward the aqueous phase, but also by electrical repulsions.

Table 1 also reflects that our three systems were stable in IF. Under the neutral pH of this medium, PLGA surface becomes highly charged. As no aggregation occurs, the ionic strength of this simulated solution (I ≈ 65 mM) must not be high enough to make the typical DLVO interaction barrier between two approaching particles disappear [14]. The higher size of the PLGA system compared to the blend formulations does not come from any partial aggregation. This size difference is originated during the particle synthesis. The unique source to achieve stability in the growing PLGA nanoparticles during their synthesis comes from electric charges of weak carboxylic groups placed at the particle surface. This implies that particles have to grow up to a critical surface charge density is reached; if not, the small growing nanospheres would collapse among them giving bigger particles. Once this critical size is attained, the system stops growing as the formed particles become colloidally stable. By contrast, in the blend formulations, the PEO molecules help to increases the stability of the growing nanospheres at lower sizes, due to the high stabilizing power exerted by these surfactants due to steric contributions [14]. All this explains the differences of size found at IF. With respect to the net charge of the systems, all of them presented negative ζ-potential values, which reflects the presence of acid groups on the particle surface [20]. The lower negative ζ-potential values obtained for the blend formulations compared to the pure PLGA particles are attributed to the presence of a surfactant coating layer, which moves the shear plane of the diffuse layer (at which the ζ-potential is

Figure 1. Size evolution for PLGA (square), PLGA-PF68 (circle) and PLGA-T904 (triangle) incubated in GF (open symbols with dashed line) and IF (closed symbols with solid line),USP XXIX at 37ºC.

1 2 3 4 5 60

200

400

600

800

1000

1200

1400

1600

Size

(nm

)

Time (hours)

162 Accepted for publication in J. of Biomed. Nanotech.

defined) away from the particle surface. This surfactant mediated shear-plane displacement even justifies the little ζ-potential differences observed between our two blend formulations: PLGA-PF68 possesses a more hydrophilic surfactant (Pluronic® F68) with PEO chains formed by 75 ethyleneoxide units, while the length of the PEO moieties in the poloxamine molecule is only made of 15 EO units. In this situation, the poloxamer chains present a more extended conformation towards the solvent, causing a greater displacement of the shear plane than in the PLGA-T904 case [14], and thus, lowering (in absolute value) the PLGA-PF68 ζ-potential.

3.2. Incubation in GF and IF: Degradation studies with enzimes

The second set of experiments was carried out with enzymes dissolved in the simulated gastrointestinal fluids. In order to quantify the particle degradation, the mean size of the nanoparticles and the percentage of PLGA converted in lactate by the action of the enzymes were measured as a function of time. The size evolution of our three systems incubated in GF and IF is depicted in Figure 1, while the production of lactate in IF, which is an indication of the particle degradation, is

shown in Figure 2. It should be noted that no lactate production was observed after incubating the three systems in GF, as pepsin is not able to cut the ester bounds presented in the PLGA backbone. Moreover, enzymatic degradation of PLGA nanoparticles would diminish their mean diameter [7], since the specific enzymatic - substrate recognition and reaction should be performed at the outer part of the particle (that is, the surface). Therefore, degradation phenomena cannot explain the significant size increase observed with the PLGA sample in both fluids (GF and IF) and with the PLGA-T904 system in GF. Consequently, these size increments must

0 1 2 3 4 5 60

10

20

30

40

50

60

70

80

Lact

ate

prod

uctio

n (%

)

Time (hours)

Figure 2. Lactate production for PLGA ( ), PLGA-PF68 ( ) and PLGA-T904 ( ) incubated in IF (USP XXIX) at 37ºC.

Paper VII 163

exclusively come from aggregation processes. Let’s start discussing the PLGA case. Aggregation of pure PLGA particles in GF by means of a charge cancellation mechanism (that is, pH mediated) was expected, as discussed previously. It may be surprising, however, that after the first incubation hour, the size of the PLGA nanoparticles decreased. The explanation of this feature must be sought in a sedimentation process of big PLGA clusters formed during the first hour, which leaves in suspension the smallest aggregates that are detected by the PCS apparatus. Results in IF are relatively similar to those found in GF. PLGA aggregation in intestinal fluid was no observed in absence of enzymes, (see Table 1), but now the system coagulates in presence of pancreatin. According to these results, it is more than likely that aggregation of PLGA particles arises from an adsorption mechanism of the enzymes that constitute the pancreatin onto the PLGA surface. Protein adsorption onto a relatively hydrophobic surface (PLGA) is a spontaneous process that may be speeded up if attractive electrostatic interactions between the protein and the sorbent surface exists [21; 22]. It should be noted that pancreatin is a mixture of enzymes composed by amylase, lipase and protease. In a mixture of proteins, adsorption of those molecules with a sign of charge different to that shown by the colloidal particle is favoured [23]. This adsorption reduces the global surface potential of the particles, and consequently, the colloidal stability of the system decreases. According to our experimental results, it is plausible that part of the pancreatin proteins become adsorbed on the PLGA surface aggregating the system not only by reducing the surface potential, but also by means of bridging mechanisms that usually occur in protein-coated latex particles [24]. Additional to the protein adsorption, the enzymatic action of lipases generates a significant amount lactate (see Fig. 2). Previous authors [7; 10] have shown, using both lactate detection and chromatography methods, that hydrophobic nanoparticles with a polyester matrix, as that existing in our PLGA nanospheres, may undergo an enzymatic degradation in simulated digestive fluids leading to water-soluble oligomers. Both processes – aggregation mediated by protein adsorption and enzymatic degradation – suggest that the PLGA surface is clearly accessible to the pancreatin molecules. These results also agree with those previously reported by Landry et al. [7] and Tobio et al. [10] with PLA nanoparticles. On the other hand, it can be also seen in Figure 1 that PLGA-T904 nanoparticles showed a very low stability in GF while PLGA-PF68 nanoparticles were totally stable. As the PLGA-T904 sample was stable in absence of enzymes, this destabilization must be again related to a possible interaction between these particles and the protein (pepsin) molecules. As suggested previously, the destabilization would be attributed to electrostatic interactions. In this case, the pepsin molecules are negatively charged (as pepsin has an isoelectric point around 1 [25]) while the PLGA-T904 nanoparticles were positively charged at the GF pH (1.2), as reflected in Table 1. In addition, our poloxamine is not very hydrophilic since it has short length PEO chains, and thus, protein (enzyme) adsorption appears not to be totally avoided. Above explanation is also applicable to the PLGA-F68 behaviour: these particles were not charged at this pH; therefore, no attractive electrostatic interactions

164 Accepted for publication in J. of Biomed. Nanotech.

between pepsin and particle can take place. In addition, the long PEO backbones of this poloxamer protrude into the aqueous phase preventing any potential protein adsorption and yielding to a stable system.

According to the results shown in Fig. 2, the presence of surfactants on the particle surface significantly reduces the enzymatic action on the PLGA matrix in both PLGA-PF68 and PLGA-T904 samples. It should be noted that lactate release is only produced in the first instants, and then lactate liberation practically stops. This is an indication that pancreatin might initially access to some PLGA patches to generate lactate by means of an enzymatic degradation based in surface erosion [7]. However, after this initial degradation, the accessibility of the enzyme to the PLGA core is hindered by a rough surface rich in poloxamer and poloxamine molecules. The recognized protein repellent effect of these polyoxyethylene derivatives must then preserve the particles against further enzymatic erosion [10; 26; 27].

3.3. Encapsulation of insulin

The goal of this section differs from those discussed previously. Specifically, the effect of the net charge of the encapsulated protein in the final characteristics of the blend formulations is now studied. Prior to discuss our results it should be note the effect of the w/o/w formulation technique in the stability of the encapsulated macromolecule. It has been published for various proteins that water/CH2Cl2 interface causes their aggregation and/or denaturation [27]. There are two different manners to avoid this problem. One of them consists in the increase of the protein concentration [28], while the second one is based in the addition of surfactant molecules to the nanoparticle formulation. The presence of surfactant reduces the protein-interface interactions, but also avoids the protein aggregation [19]. In this paper we have chosen the incorporation of a surfactant (Pluronic® F68 or Tetronic® 904) to solve this problem. The above discussion explains why encapsulation of insulin in pure PLGA nanoparticles was not performed.

The encapsulation of positively charged insulin (referred to as ins-pH2) was carried out dissolving this protein in HCl 0.01M during the particle synthesis. Likewise, insulin with negative charge (ins-pH12) was incorporated to the particles using a NaOH 0.01M solution. The main characteristics of these insulin loaded blend formulations incubated in GF and IF are also shown in Table 1. The mean size of both formulations showed a clear dependence with the net charge of the encapsulated insulin. The size was much lower when both PLGA and insulin were negatively charged (ins-pH12), independently on the surfactant nature. On the contrary, when PLGA and insulin have opposite charge (ins-pH2), the charge cancellation during the synthesis favors the aggregation of small growing particles, yielding bigger nanospheres that are finally stabilized by the poloxamer or poloxamine action. With respect to the ζ-potential values, the presence of insulin does not almost alter the original values found for the blend formulations without incubated insulin. This feature would mean either that insulin is not incorporated

Paper VII 165

to the particles during the synthesis, or it is effectively incorporated but mainly in the PLGA matrix and not in the particle interface, which must be crowded by surfactant molecules. The evaluation of the encapsulation efficiency (see EE in Table 1) shows that the second option occurs, as these quantitative values indicates that insulin is effectively incorporated – at least partially – to the PLGA nanocarriers. As can be seen, the EE mainly depends on the surfactant nature used in the original mixture, but it is almost independent on the initial charge state of the insulin. In fact, encapsulation is slightly lower when PLGA and insulin share the same sign of charge (ins-pH12), as electrostatic repulsions may hinder in some manner the protein approach. One may wonder why insulin is encapsulated much more in the PLGA-PF68 samples (around 35%) than in the PLGA-T904 ones (10% approximately). It should be noted that insulin is a small globular protein with a hydrophobic core although with a surface rich in polar residues. Therefore, the outer part of the insulin is relatively hydrophilic [29]. According to the “like seeks like” rule based on entropic interactions [30], insulin appears to interact much better with the most hydrophilic PEO derivative (Pluronic® F68), whereas the interaction with more hydrophobic molecules (as PLGA or Tetronic® 904) in the PLGA-T904 sample seems not to be very significant. Besides, there is another potential reason that can justify the protein incorporation. It is based on the stability of the primary W1/O emulsion obtained during the particle synthesis, as such emulsion stability is considered the key point for the entrapment of hydrophilic compounds [30; 31]. With this idea in mind, the stability of this W1/O emulsion was evaluated for our systems analyzing the time necessary to produce a liquid phase

0 2 4 6 8 10 12 14

120

140

160

180220

240

260

280

300

Siz

e (n

m)

Time (days)

Figure 3. Size evolution for PLGA-PF68 ( , solid line), PLGA-PF68 ins-pH2 ( , dashed line); PLGA-PF68 ins-pH12 ( , dotted line); PLGA-T904 ( , solid line); PLGA-T904 ins-pH2 ( , dashed line) and PLGA-T904 ins-pH12 ( , dotted line) incubated in IF (USP XXIX, without enzymes) at 37ºC.

166 Accepted for publication in J. of Biomed. Nanotech.

separation. For the PLGA-T904 blend formulation, both organic and aqueous phases were well separated after 22 h, while the emulsion was stable even for a week for the PLGA-PF68 blend formulation. Large emulsion stability permits insulin to interact with the emulsion components much better, and thus, enhancing the encapsulation. This would also explain why higher encapsulation efficiencies are found when working with more stable emulsions, that is, with those formed by using Pluronic® F68 instead of Tetronic® 904.

3.4. Incubation in GF and IF: Degradation studies without enzymes

The next set of experiments was focused to measure the size and ζ-potential time evolution when insulin loaded nanoparticles were incubated in gastrointestinal fluids free of enzymes. Figure 3 shows the size evolution in IF – evolution in GF (Figure not shown) was totally similar – while the ζ-potential data in GF are shown in Figure 4; note that ζ-potential data in IF (Figure not shown) were also constant. Data were collected for 14 days, and it can be observed that no significant variation on size or zeta-potential was found in this period of time. According to the results, an incubation of two weeks seems to be a short time period to appreciate degradation of our blend formulations in absence of enzymes. Similar results were also obtained by other authors working with particles based on a PLGA matrix [11; 18]. It is known that the PLGA nanoparticles degradation in the first stages is controlled by a surface-core mechanism, due to the acid conditions generated by the rupture of the ester bonds in the nanoparticles core [32]. Therefore, the constancy in size and ζ-potential must not be thought as an

0 2 4 6 8 10 12 14

-10

-5

0

5

10

15

20

25

ζ-po

tent

ial (

mV

)

Time (days)

Figure 4. ζ-potential evolution for PLGA-PF68 ( , solid line), PLGA-PF68 ins-pH2 ( , dashed line); PLGA-PF68 ins-pH12 ( , dotted line); PLGA-T904 ( , solid line); PLGA-T904ins-pH2 ( , dashed line) and PLGA-T904 ins-pH12( , dotted line) incubated in GF (USP

X, without enzymes) at 37ºC. XXI

Paper VII 167

0 1 2 3 4 5 6 70

20

40

60

80

100R

elea

se (%

)

Time (days)

(a)

0 1 2 3 4 5 6 70

20

40

60

80

100

Rel

ease

(%)

Time (days)

(b)

Figure 5. Cumulative release profile of insulin from PLGA-PF68 ins-pH2 ( , dashed line); PLGA-PF68 ins-pH12 ( , dotted line); PLGA-T904 ins-pH2 ( , dashed line) and PLGA-T904 in ( , dotted line) incubated in GF (a) and IF (b) (USP XXIX, withouts-pH12

indication about a lack of insulin release, since this release process might take place without altering the mean diameter or the surface properties of our nanocarriers. Consequently, in order to evaluate any potential insulin release, the detection of this hormone in GF and IF was also performed.

3.5. In vitro release of insulin

The release profile of macromolecules from hydrophobic PLGA matrix presents usually an initial burst effect, due to the surface-associated molecules which are easily released from the nanoparticles. On a second phase, macromolecules diffuse for a network of newly generated water filled pores. It has been reported [33] that due to the formation of insulin aggregates after water penetration into the matrix of the nanoparticles; it is not possible to appreciate the second release phase in the case of this macromolecule. However, the presence of surfactants (i.e. Pluronic® F68) in our nanoparticle matrix avoids the water induced aggregates and even can help to reduce the initial burst effect [33, 34]. Figures 5a and 5b show the in vitro release profile of the insulin for PLGA-PF68 and PLGA-T904 blend formulations in simulated GF and IF, respectively. It should be noted that only the first 7 days of incubation has been depicted to observe better the initial burst effect. The release profiles maintained their tendencies from the 7th to the 14th day. As can be observed in GF (Fig. 5a) the insulin released at long time is higher in the PLGA-PF68 samples than in the PLGA-T904 ones. The PLGA-T904 blend formulations did not show any difference regardless the insulin was encapsulated with positive or negative charge. These systems showed an initial burst effect, around 20%, followed by a close to zero release velocity. It has been shown that the presence of poloxamines with low HLB values (i.e. Tetronic® 904) in PLGA nanoparticles decreases the degradation of the PLGA matrix slowing down the pore formation [11]; consequently, this reason might be responsible for the observed delay in the insulin release in our PLGA-T904 samples. The net charge of the protein during the encapsulation step, however, had a clear effect in the release

168 Accepted for publication in J. of Biomed. Nanotech.

profile for the PLGA-PF68 blend formulations. The most striking result is the rapid insulin liberation detected in the PLGA-PF68 ins-pH12 system. In this sample the basic pH of the encapsulation process is completely inverted to a highly acidic medium in the GF. This marked pH change alters the electrical states of both PLGA and insulin producing a significant release which is patent even after 24 h. It should be noted that if there is no drastic pH change, as that observed in the PLGA-PF68 ins-pH2 sample – since the pH of encapsulation almost coincides with the pH of the incubation medium – the insulin liberation becomes low during the first days. This result can be due to the best arrangement of the protein when it was incorporated to the nanoparticle with opposite charge to that of the PLGA. Therefore, although the encapsulation pH did not affect in the EE values, it appears to be crucial to speed up or to slow down the insulin release at short times in the PLGA-PF68 samples.

Release profiles in IF (shown in Fig. 5b) were similar for both blend formulations, as they were practically independent on the surfactant nature and on the insulin charge during its encapsulation. Nevertheless, both formulations appear to show a more progressive release when the insulin was positively encapsulated. The similitude between the PLGA-PF68 and PLGA-T904 formulations may be due to a lower degradation of the polyester matrix at pH close to 7 [17,18,35]. This lower PLGA degradation at neutral pH could mask the effect of the surfactant in the degradation of the systems, and consequently, in the release of the protein.

IV. CONCLUSIONS

Stability of pure PLGA nanoparticles and two blend formulations – PLGA:poloxamer and PLGA:poloxamine – in simulated gastro-intestinal fluids has been determined in this work. In GF only the PLGA-PF68 formulation was stable, while pure PLGA and PLGA-T904 particles aggregated due to electrostatic reasons associated to surface charge cancellations induced by the medium pH (in the PLGA case) or by the adsorption of enzymes with opposite sign of charge. In IF, all our systems were degraded by the enzimatic action of pancreatin, although the presence of surfactants in the blend formulations reduced significantly the lactate production.

With respect to the insulin encapsulation, the net charge of the protein during the encapsulation process showed a clear effect in the final size of the particles, while the encapsulation efficiency was determined by the polyoxyethylene derivative present in the nanoparticles. Finally, the release profile in GF presented certain dependence on the surfactant nature present in the blend formulation, while the charge of the encapsulated protein clearly affected to the insulin release at short times in the PLGA-PF68 samples. However, release differences were not almost observed in IF.

Taking into account all the results shown in this paper, a recommendation can be suggest with regard to use PLGA nanoparticles as oral delivery systems for insulin. The use of Pluronic® F68 for encapsulating insulin in PLGA based

Paper VII 169

nanocarriers generates stable particles that are able to maintain their integrity in simulated gastrointestinal fluids. The release of the insulin can be speeded up or slowed down just by tuning the encapsulation pH. Nevertheless, it will be necessary to perform in vivo studies to corroborate the suggestions that can arise after considering our in vitro results.

ACKNOWLEDGEMENTS

Authors thank the financial support given by the project MAT2007-66662-CO2-01 from the Comisión Interministerial de Ciencia y Tecnología (Spain), European FEDER support included, and from the “Conserjería de Innovación, Ciencia y Tecnología de la Junta de Andalucía” (Spain), projects of excellence FQM 392 and FQM 03099.

170 Accepted for publication in J. of Biomed. Nanotech.

References

1. J. F. Woodley, Enzymatic barriers for GI peptide and protein delivery., Crit Rev Ther Drug Carrier Syst, 11, 61-95 (1994).

2. M. Saffran, B. Pansky, G. C. Budd, and F. E. Williams, Insulin and the gastrointestinal tract, Journal of Controlled Release, 46, 89-98 (1997).

3. R. Langer, New Methods of Drug Delivery, Science, 249, 1527-33 (1990).

4. N. Csaba, P. Caamano, A. Sanchez, F. Dominguez, and M. J. Alonso, PLGA : poloxamer and PLGA : poloxamine blend nanoparticies: New carriers for gene delivery, Biomacromolecules, 6, 271-8 (2005).

5. M. J. Alonso, Microparticulate Systems for the Delivery of Proteins and Vaccines, eds. S.Cohen and H.Berstein, New York: Marcel Dekker Inc., 1996, 203-42.

6. A. Schindler, R. Jeffcoat, G. L. Kimmel, C. G. Pitt, M. E. Wall, and R. Zweidinger, Biodegradable polymers for sustained drug delivery, Contemporary Topics in Polymer Sciences, 2, 251-89 (1977).

7. F. B. Landry, D. V. Bazile, G. Spenlehauer, M. Veillard, and J. Kreuter, Influence of coating agents on the degradation of poly(D,L-lactic acid) nanoparticles in model digestive fluids (USP XXII), Stp Pharma Sciences, 6, 195-202 (1996).

8. C Ropert, D. V. Bazile, J. Brendenbach, M. Marlard, M. Veillard, and G. Spenlehauer, Fate of 14C radiolabeled poly(DL-lactic acid) nanoparticles following oral administration to rats, Colloids and Surfaces B: Biointerfaces, 1, 233-9 (1993).

9. S. M. Moghimi and A. C. Hunter, Poloxamers and poloxamines in nanoparticle engineering and experimental medicine, Trends in Biotechnology, 18, 412-20 (2000).

10. M.Tobio, A.Sanchez, A.Vila, I.Soriano, C.Evora, J.L.Vila-Jato, and M.J.Alonso, The role of PEG on the stability in digestive fluids and in vivo fate of PEG-PLA nanoparticles following oral administration, Colloids and Surfaces B: Biointerfaces, 18, 315-23 (2000).

11. M. K. Yeh, S. S. Davis, and A. G. A. Coombes, Improving protein delivery from microparticles using blends of poly(DL lactide co-glycolide) and poly(ethylene oxide)-poly(propylene oxide) copolymers, Pharmaceutical Research, 13, 1693-8 (1996).

12. M. Tobio, J. Nolley, Y. Y. Guo, J. McIver, and M. J. Alonso, A novel system based on a poloxamer PLGA blend as a tetanus toxoid delivery vehicle, Pharmaceutical Research, 16, 682-8 (1999).

13. D. Blanco and M. J. Alonso, Protein encapsulation and release from poly(lactide-co-glycolide) microspheres: effect of the protein and polymer properties and of the co-encapsulation of surfactants, European Journal of Pharmaceutics and Biopharmaceutics, 45, 285-94 (1998).

14. M. J. Santander-Ortega, N. Csaba, M. J. Alonso, J. L. Ortega-Vinuesa, and D Bastos-González, Stability and physicochemical characteristics of PLGA, PLGA:poloxamer and PLGA:poloxamine blend nanoparticles: A comparative study, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 296, 132-40 (2007).

15. M. J. Santander-Ortega, A. B. Jodar-Reyes, N. Csaba, D. Bastos-Gonzalez, and J. L. Ortega-Vinuesa, Colloidal stability of Pluronic F68-coated PLGA nanoparticles: A variety of stabilisation mechanisms, Journal of Colloid and Interface Science, 302, 522-9 (2006).

Paper VII 171

16. N. Csaba, A. Sanchez, and M. J. Alonso, PLGA: Poloxamer and PLGA: Poloxamine blend nanostructures as carriers for nasal gene delivery, Journal of Controlled Release, 113, 164-72 (2006).

17. T. Ivanova, I. Panaiotov, F. Boury, J. E. Proust, J. P. Benoit, and R. Verger, Hydrolysis kinetics of poly(D,L-lactide) monolayers spread on basic or acidic aqueous subphases, Colloids and Surfaces B-Biointerfaces, 8, 217-25 (1997).

18. X. S. Wu and N. Wang, Synthesis, characterization, biodegradation, and drug delivery application of biodegradable lactic/glycolic acid polymers. Part II: Biodegradation, Journal of Biomaterials Science-Polymer Edition, 12, 21-34 (2001).

19. N. Csaba, L. Gonzalez, A. Sanchez, and M. J. Alonso, Design and characterisation of new nanoparticulate polymer blends for drug delivery, Journal of Biomaterials Science-Polymer Edition, 15, 1137-51 (2004).

20. D. Bastos-Gonzalez, J. L. Ortega-Vinuesa, F. J. D. Nieves, and R. Hidalgoalvarez, Carboxylated Latexes for Covalent Coupling Antibodies .1, Journal of Colloid and Interface Science, 176, 232-9 (1995).

21. W. Norde, Adsorption of Proteins from Solution at the Solid-Liquid Interface, Advances in Colloid and Interface Science, 25, 267-340 (1986).

22. J. L. Ortega-Vinuesa, Galvez-RuizM.J., and R. Hidalgo-Alvarez, F(ab')(2)-coated polymer carriers: Electrokinetic behavior and colloidal stability, Langmuir, 12, 3211-20 (1996).

23. F. Galisteo-González, J. Puig, A. Martín-Rodriguez, J. Serra-Domènech, and R. Hidalgo-Alvarez, Influence of electrostatic forces on IgG adsorption onto polystyrene beads, Colloids and Surfaces B: Biointerfaces, 2, 435-41 (1994).

24. M. Tirado-Miranda, A. Schmitt, J. Callejas-Fernandez, and A. Fernandez-Barbero, The aggregation behaviour of protein-coated particles: a light scattering study, European Biophysics Journal with Biophysics Letters, 32, 128-36 (2003).

25. F. Bovey and S. Yanari, Pepsin, the Enzimes NY: Academic Press, 1960, 63.

26. R. Gref, Y. Minamitake, M. T. Peracchia, V. Trubetskoy, V. Torchilin, and R. Langer, Biodegradable Long-Circulating Polymeric Nanospheres, Science, 263, 1600-3 (1994).

27. I. Brigger, C. Dubernet, and P. Couvreur, Nanoparticles in cancer therapy and diagnosis, Advanced Drug Delivery Reviews, 54, 631-51 (2002).

28. C. Perez, I.J. Castellanos, H.R. Costantino, W. Al-Azaam, k. Griebenow, Recent trenes in stabilizing protein structure upon encapsulation and release from biodegradable polymers. Journal of Pharmacy and Pharmacology, 54, 301-313 (2002).

29. K Huang and B Xu, How Insulin Binds: the B-Chain a-Helix Contacts the L1 b-Helix of the Insulin Receptor, Journal of Molecular Biology, 341, 529-50 (2004).

30. J. N. Israelachvili, Intermolecular and surface force London: Academic Press, 1992.

31. C. Schugens, N. Laruelle, N. Nihant, C. Grandfils, R. Jerome, and P. Teyssie, Effect of the Emulsion Stability on the Morphology and Porosity of Semicrystalline Poly L-Lactide Microparticles Prepared by W/O/W Double Emulsion-Evaporation, Journal of Controlled Release, 32, 161-76 (1994).

32. K. Fu, D.W. Pack, A.M. klivanov and R. Langer, Visual Evidence of Acidic Environment Within Degrading Poly(Lactic-co-glycolic acid) (PLGA) Microspheres, Pharmaceutical Research, 17, 100-106 (2000).

172 Accepted for publication in J. of Biomed. Nanotech.

33. P.S. Kumar, S. Ramakrishna, T.R. Saini, P.V. Diwan, Influence of microencapsulation method and peptide loading on formulation of poly(lactide-co-glycolide) insulin nanoparticles, Pharmazie, 61, 613-617 (2006).

34. M.K. Yeh, S.S. Davies and A.G.A. Combes, Improving Protein Delivery from Microparticles using Blends of Poly(DL lactide co-glycolide) and Poly (ethylene oxide)-poly(propylene oxide) Copolymers, Phermaceutical Research, 13, 1693-1698 (1996).

35. A. Belbella, C. Vauthier, H. Fessi, J. P. Devissaguet, and F. Puisieux, In vitro degradation of nanospheres from poly(D,L-lactides) of different molecular weights and polydispersities, International Journal of Pharmaceutics, 129, 95-102 (1996).

Development of Novel Drug-Assembled Nanoparticles made from Amphiphilic Polymers

M.J. Santander-Ortega1,2, Ortega-Vinuesa, J.L.1, Bastos-González, D1, U.F. Schäfer2, G. Wenz3 and C.M. Lehr2

1Department of Applied Physics, University of Granada, Granada (SPAIN)

2Department of Biopharmaceutics and Pharmacological Technology, Saarland University, Saarbrücken (GERMANY)

3Department of Organic Macromolecular Chemistry, Saarland University, Saarbrücken (GERMANY)

174 Paper VIII

ABSTRACT

Nanoparticles were formed by an o/w emulsion method dissolving the polymer in the organic phase. The nanoparticle formation was clearly conditioned by the addition of a hydrophobic molecule (Flufenamic acid) to the organic phase. This result suggests that nanoparticles were formed by an assemble of the amphiphilic polymer and flufenamic acid. Nanoparticles showed a size of around 200 nm and a low polydispersity index. Regarding to the release profile of the Flufenamic acid it presented a close to lineal profile without showing any burst effect. Concomitantly with the leakage of this molecule, the size evolution of the colloidal system during the release was also monitored. The degradation of the nanoparticles when the molecule was released suggests that flufenamic acid was the main responsible of the nanoparticles formation, as it seems to act as crosslinking between the hydrophobic moieties of the amphiphilic backbone.

Paper VIII 175

I. INTRODUCTION

The aim of this work was the development of a new drug delivery system based in the use of an amphiphilic polymer, named as P5 and composed by hydrophilic and hydrophobic segments. These nanoparticles were formulated by a novel method in which the encapsulated molecule acts as a “hydrophobic linker” to form the nanoparticles. Flufenamic acid, a well known anti-inflammatory drug, was selected as hydrophobic model drug. Hence, the release of this molecule from the nanoparticles matrix produces their degradation, making easier the elimination of the polymer from the body. Formation and characterization of the nanoparticles as well as the effect of the release of the molecule from the nanoparticles has been studied.

II. EXPERIMENTAL PART

2.1. Nanoparticles preparation

Nanoparticles, denoted as P5 with the amphiphilic polymer (P5) were formulated by simple o/w emulsion method. Briefly, P5 polymer and flufenamic acid (FFA) in a 1:1 ratio (w/w) were dissolved in an organic solvent, and this organic solution was poured in an aqueous phase with different percentages (w/v) of a surfactant (0, 0.1, 0.5 and 1). This biphasic system was emulsified with a high speed homogenizer (Ultra Turrax® T25, Ika®,Stauffen, Germany). Then, MilliQ water was added to force the complete diffusion of the organic solvent to the aqueous phase. Finally, the organic solvent was evaporated under vacuum at 35ºC (Rotavapor Büchi®, Labortechnik AG, Flawil, Switzerland).

Encapsulation efficiency (EE) was calculated using a Franz diffusion cella and can be defined as:

100(%) 100 D

D

FEET

⎛ ⎞= − ⎜

⎝ ⎠⎟

(1)

where FD is the amount of free drug present in the receptor compartment at time close to zero and TD is the total amount of drug dissolved in the organic phase. The exact amount of the drug was calculated by HPLC [1].

2.2. Nanoparticles Characterization

Size and ζ–potential of the nanoparticles were analysed by photocorrelation spectroscopy (PCS) using a Nano-ZS (Malvern Instruments, Malvern, UK). AFM images were obtained using an Atomic Force Microscopy Nanoscope IV BioscopeTM (Veeco Instruments, Santa Barbara, CA, USA). Imaging was done using Taping mode and a silicon cantilever with a spring constant of

a For more details see section Release of FFA from the nanoparticles

176 Paper VIII

approximately 40 N/m and a resonance frequency of about 170 kHz. The scan speed applied was 0.2 Hz.

2.3. Release of FFA from the nanoparticles

Once the nanoparticles were formulated the release of the FFA as well as the effect of this release on the stability of the nanoparticles were studied. Release of FFA from assembled P5 nanoparticles was studied using Franz-diffusion Cells (FdC) type 6G-01-00-15 (Perme-Gear, Riegelsville, PA, USA). Briefly, 0.5 ml of the nanoparticle dispersion and 1.5 ml of phosphate buffer pH 7 (2mM) were poured in the donor compartment of the FdC, while the receptor compartment was filled with 12 ml of the same buffer solution. Donor and receptor compartment were separated by a cellulose membrane with a pore size of 12-14 kDa (Medicell Int. Ltd, London, UK.). FdCs were incubated in an oven at 32ºC for a period of time of 72 hours. Samples of 0.4 ml were removed from the receptor compartment at regular time intervals up to 72h and replaced with an equal volume of fresh buffer. The concentration of free drug from the receptor was analyzed by HPLC [1].

In order to know the effect of the FFA leakage on the stability of the assembled nanoparticles, a parallel study was developed under non-sink conditions. In this case, nanoparticles were immersed in a FFA saturated aqueous medium avoiding the release of the FFA from the assembled P5 nanoparticles. Size evolution of the nanoparticles incubated under both conditions, sink and non-sink, was monitored as a function of the time by PCS technique.

10 100 1000

0

5

10

15

20

25

Inte

nsity

(%)

Size (nm)

Figure 1. Hydrodynamic size distribution of P5 nanoparticles as a function of the components present in the blend. P5 alone (1mg/ml) (solid line); Surfactant (1% w/v) and P5(dashed line); Surfactant (1% w/v), FFA (1:1) and P5 (1mg/ml) (dotted line (n≥3).

Paper VIII 177

(a)

(b)

Figure 2. AFM image of P5 nanoparticles formed without FFA (a) and with FFA (b).

2.4. Kinetics analysis of FFA release

To obtain a better understanding of the process involved in the release of the FFA from these nanoparticles, the semi-empirical power law expression for the analysis of the drug release from non-swelling spherical nanoparticles was used [2]:

ntM kt= (2) M∞

178 Paper VIII

Where incorporatdiffusional exponent, which indicates the transport mechanism. Ritger and Peppas [3] publish

e nanoparticles formation, being impossible to er with the same backbone but without

amphip

present protonable groups, the low

Unmodified. Polymer. No formation of nanoparticles

Mt/M∞ is the drug fraction release at time t, k is a constant ing characteristics of the polymeric network and the drug, and n is the

ed that Fickian diffusion from non-swelling spheres is characterized by n = 0.43, while anomalous (non-Fickian) diffusion by 0.43 < n < 1.0, and a Zero-order release velocity by n = 1.0.

III. RESULTS AND DISCUSSION

Regarding to the nanoparticles preparation, the amphiphilic character of the polymer was a critical point in thform nanoparticles with a polym

hilic character (data not shown). With respect to the formation of nanoparticles with the P5 polymer, Figure 1 illustrates the clear effect of the incorporation of FFA to the organic phase. When the P5 polymer was the unique component dissolved in the organic phase and miliQ water formed the aqueous phase, a very broad size distribution was obtained. The incorporation of the surfactant to the aqueous phase improved somehow the size distribution, being the best result that obtained with a surfactant concentration of 1% (w/v) (see Fig 1). Nevertheless, the size distribution of this formulation was not narrow at all (145 ± 8 nm; PDI 0.45 ± 0.05). The incorporation of the FFA together with the amphiphilic polymer to the organic phase generated the narrowest size distribution. These results agree with those previously published by Pan et al [4] and suggest that FFA acts as a “hydrophobic linker”, performing a formulation-encapsulation step during the nanoparticles formation. AFM pictures shown in Figure 2a-b, give more intuitive information than the PCS data about the effect of FFA in the nanoparticles formation. As mentioned earlier, when FFA was not present in the formulation a large variety of structures were found with broad size distribution (Fig. 2a). In contrast, the incorporation of FFA to the organic phase produced highly spherical nanoparticles with a regular size distribution (Fig. 2b).

The main characteristics of the assembled P5 nanoparticles are summarized in Table 1. These nanoparticles showed a hydrodynamic mean size of 207 nm with low PDI (0.117). As neither P5 nor surfactant polymers

negative ζ-potential value displayed by this colloidal system must be attributed to the accumulation of, at least, parts of the FFA on the nanoparticles surface [5]. It is worthy to remark the high affinity of the FFA for the nanoparticles matrix, presenting an EE higher than 90%.

Table 1. Size, PDI, ζ-potential and Encapsulation efficacy (EE) of the P5 nanoparticles.

Size (nm) PDI ζ-pot (mV) E.E. (%)

P5 0,11 ,04207,2 ± 3,3 7 ± 0 -5,65 ± 0,98 >90

Paper VIII 179

50

The next set of experiments was focused on the release of the FFA from the assembled P5 nanoparticles and its effect on the colloidal stability. Nanoparticles incubated in FdC showed a promising release pattern, with a sustained release profile and without initial burst effect, see Figure 3. Although sometime a burst

0 10 20 30 40 50 60 70 800

10

20

30

40

Rel

ease

(%)

Time (h)

Figure 3. Release of FFA from P5 nanoparticles incubated at 32ºC.

0 10 20 30 40 50 60 70

160

180

200

220

240

260

280

300

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

Siz

e (n

m)

Time (h)

PDI

Figure 4. Size (solid line and close symbols) and PDI (dashed line and open symbols)evolution of FFA loaded P5 nanoparticles incubated at 32ºC under non-sink conditions (circles) and under sink conditions (squares).

180 Paper VIII

effect is desirable as part of the drug administration strategy [6], usually this burst effect can be pharmacologically dangerous and economically inefficient [7]. Studies developed in our laboratory with similar matrixes and drugs with different Log Pb values suggest that in this case the release of FFA from P5 matrix may be controlled by the hydrophobic interactions of the encapsulated molecules and the hydrophobic moieties of the amphiphilic polymer [8]. The effect of the FFA release on the nanoparticles stability is clearly manifested and it is graphically illustrated in Figure 4. By comparison of Fig 3 and 4, it is easy to see that the leakage of the FFA from the nanoparticles implied the lost of the nanoparticles integrity. On the other hand, in basis of the data depicted in Fig. 4, when the release of FFA was hindered by using a saturated FFA solution as receptor phase (see circular symbols of Fig. 4), the size of the assembled nanoparticles were constant during the incubation time. Moreover, in order to know the type of release of FFA, the power law for non swelling spherical devices was applied to the experimental results [3], obtaining a n=0.83 (r2=0.988). In spite of the simplicity of this mathematical model, it is possible to obtain an estimation of the mechanism that controlled the release of FFA. An n=0.83 suggests that the release of FFA it was not controlled by a diffusion mechanism. This non-Fickian diffusion release mechanism can be attribute to the breaking of the nanoparticle matrix when FFA is released [3]. Hence, the release results

“hydrophobic linker” by hydrophobic interactions with the hydrophobic moieties of P5 polymer allows the nanoparticle formation. When this linker molecule is released the nanoparticle integrity becomes to disappear yieling inhomogeneous meshes of P5 matrix. The formation and release of this type of system is still subject of investigations such as IR-spectroscopy or DSC but it is more than likely that it is controlled by hydrophobic interactions [4].

IV. CONCLUSIONS

A novel drug delivery system has been developed by using a novel matrix and an original preparation method which implicates the degradation of the nanoparticle when the drug is released. Moreover, this colloidal carrier presents promising applications due to the high encapsulation capacity of FFA and the peculiar controlled release profile of this molecule.

ACKNOWLEDGEMENTS

Authors thank the financial support given by Galenos Fellowship in the Framework of the EU Project “Towards a European PhD in Advanced Drug Delivery”, Marie Curie Contract MEST-CT-2004-404992.

and the stability of the assembled nanoparticles under non-sink conditions are in the same direction that the nanoparticles preparation results, namely, the formation of this colloidal system is controlled by the incorporation of a

b Log P is known as the logarithm of the octanol-water partition coefficient.

Paper VIII 181

References

(1) Luengo J, Weiss B, Schneider M, Ehlers A, Stracke F, Konig K, Kostka KH, Lehr CM, Schaefer UF. Influence of nanoencapsulation on human skin transport of flufenamic acid. Skin Pharmacology and Physiology 2006; 19(4):190-197.

(2) Peppas NA. Analysis of Fickian and non-Fickian drug release from polymers. Pharmaceutica acta Helvetiae 1985; 60(4).

(3) Ritger PL, Peppas NA. A simple Equation for Description of Solute Release I: Fickian and non-Fickian Release from non-Sweliable Devices in the form of Slabs, Spheres, Cyliders or Discs. J Control Release 1987; 5.

(4) Pan X, Yao P, Jiang M. Simultaneous nanoparticle formation and encapsulation driven by hydrophobic interaction of casein-graft-dextran and beta-carotene. J Colloid Interface Sci 2007; 315(2):456-463.

(5) Santander-Ortega MJ, Bastos-Gonzalez D, Ortega-Vinuesa JL. Electrophoretic mobility and colloidal stability of PLGA particles coated with IgG. Colloids Surf B Biointerfaces 2007; 60(1):80-88.

(6) Development of encapsulated antibiotics for topical administration to wounds. Second World Congress on Biomaterials 10th Annual Meeting of the Society for Biomaterials; 1984.

(7) Huang X, Brazel CS. On the importance and mechanisms of burst release in matrix-controlled drug delivery systems. J Control Release 2001; 73(2-3):121-136.

(8) Santander-Ortega MJ, Stauner T, Loretz B, Ortega-Vinuesa JL, Bastos-Gonzalez D, Wenz G, Schaefer UF, Lehr CM. Nanoparticles made from Novel Amphiphilic Polymers for Advanced Drug Delivery across Biological Barriers. 2008. (Unpublished Work)

Nanoparticles made from Novel Amphiphilic polymers for Advanced Drug Delivery across Biological

Barriers. Part 2

M.J. Santander-Ortega1,2, T. Stauner3, B. Loretz2, Ortega-Vinuesa, J.L.1, Bastos-González, D.1, G. Wenz3, U.F. Schäfer2 and C.M. Lehr2

1Department of Applied Physics, University of Granada, Granada (SPAIN)

2Department of Biopharmaceutics and Pharmacological Technology, Saarland University, Saarbrücken (GERMANY)

3Department of Organic Macromolecular Chemistry, Saarland University, Saarbrücken (GERMANY)

184 Paper IX

ABSTRACT

The goal of this paper was aimed to the formulation of nanoparticles (P1 and P2) by using two different amphiphilic polymers –referred to as P1 and P2- presenting P2 a higher hydrophobic character. These nanoparticles, characterized in a previous study (see Paper IV), were prepared by a simple o/w emulsion diffusion technique, avoiding the use of hazard solvents such as dichloromethane or dymethyl sulfoxide. In this second part more applied studies were developed. Firstly, the lyophilization of these nanoparticles was analyzed, showing the surface composition of each system a clear effect on their resistance to the lyophilization stress. Secondly, the encapsulation and release properties of these nanoparticles were tested, showing high encapsulation efficiency for three tested drugs (flufenamic acid, testosterone and caffeine); in addition a close to linear release profile was observed for hydrophobic drugs with a null initial burst effect. Finally, the potential use of these nanoparticles as transdermal drug delivery systems was also tested, displaying a clear enhancer effect for the flufenamic acid.

Paper IX 185

I. INTRODUCTION

The aim of this work was the development and use of a promising nanoparticle system formulated with amphiphilic polymers for advanced drug delivery across biological barriers. On the one hand, as an important limitation of nanoparticles suspended in aqueous mediums is their low stability when conserved for a longer time, the lyophilization of P1 and P2 nanoparticles was studied. On the other hand, the capacity of these nanoparticles as drug delivery systems was tested by the encapsulation and release of three different model drugs, flufenamic acid (FFA), testosterone (Test) and caffeine (Caff). Finally, considering the nanoparticles characterization results (see Paper IV), it was possible to establish the use of these nanoparticles as transdermal drug delivery system (TDDS). Hence, the potential application of these nanoparticles as TDDS was analyzed by studying their permeation capacity through human heat-separated epidermis (HSE).

II. MATERIALS AND METHODS

2.1. Preparation of the Nanoparticles.

Nanoparticles with the different polymers were formulated by a simple o/w emulsion diffusion method. Briefly, specific polymer (P1 or P2) was dissolved in an organic solvent and this organic solution was poured on an aqueous phase with different percentages (w/v) of a surfactant (0, 0.1, 0.5 and 1). This biphasic system was emulsified with a high speed homogenizer (Ultra Turrax® Ika®, Brasil Ltda, Taquara, Brasil). Then, MilliQ water was added to force the complete diffusion of the organic solvent to the aqueous phase. Finally, the organic solvent was evaporated under vacuum at 35ºC (Rotavapor Büchi®, Labortechnik AG, Flawil, Switzerland), generating stable nanoparticles. After nanoparticles preparation, MiliQ water was added to obtain a colloidal solution with a final volume of 10 ml.

2.2. Characterization of the Nanoparticles

Size and ζ–potential of the nanoparticles were analysed by photon correlation spectroscopy (PCS) using a Nano-ZS (Malvern Instruments, Malvern, UK). For ζ–potential measurement nanoparticles were diluted in NaCl 3mM. AFM images were obtained using an Atomic Force Microscopy Nanoscope IV BioscopeTM (Veeco Instruments, Santa Barbara, CA, USA). Imaging was done using Taping mode and a silicon cantilever with a spring constant of approximately 40 N/m and a resonance frequency of about 170 kHz. The scan speed applied was 0.2 Hz.

2.3. Lyophilization assay

P1 and P2 nanoparticles, prepared as was commented previously, were lyophilized by using a Freeze-drier Alpha 2-4 LSC (Christ, Osterode, Germany) and sucrose or trehalose as cryoprotectant agent. Briefly, different volumes of a solution of Sucrose or Trehalose of 10% (w/v) were poured onto different aliquots of nanoparticles in order to obtain a cryoprotectant range between 0-1% (w/v).

186 Paper IX

Previously to the lyophilization step, the different samples were frozen at -80ºC for 2h. Hydrodynamic mean size of these aliquots was measured before and after the lyophilization step.

2.4. Encapsulation of flufenamic acid, testosterone and caffeine

In order to reach a good knowledge of the potential use of these nanoparticles as drug delivery system three different drugs were chosen to study the encapsulation behaviours of P1 and P2 nanoparticles. These drugs present different hydrophobicity (expressed as the logarithm of the octanol-water partition coefficient; Log P) and net charge. Exactly, the chosen drugs were: FFA (Log P = 4.80; pKa = 3.90; Mw = 281.23), testosterone (Log P = 3.47; non-ionizable molecule; Mw = 288.40) and caffeine (Log P = -0.08; pKb = 10.40; Mw = 194.20). The preparation method of drug loaded P1 and P2 nanoparticles was not modified with respect to the un-loaded nanoparticles. The specific molecule and the P1 or P2 polymer were dissolved in the organic phase (1:1 ratio), following the same procedure described above. The characterization of these nanoparticles was performed as was comment previously for the un-loaded nanoparticles (see Paper IV). Encapsulation efficiency (EE) was calculated using a Franz diffusion cella and can be defined as:

100(%) 100 D

D

FEET

⎛ ⎞= − ⎜

⎝ ⎠⎟

(1)

where FD is the amount of free drug present in the receptor compartment at time close to zero and TD is the total amount of drug dissolved in the organic phase. The exact amount of each drug was calculated by HPLC [1, 2].

2.5. In-vitro release of flufenamic acid, testosterone and caffeine

Release of FFA, testosterone and caffeine from loaded P1 and P2 nanoparticles was studied using Franz-diffusion Cells (FdC) type 6G-01-00-15 (Perme-Gear, Riegelsville, PA, USA). Briefly, a solution of 0.5 ml of nanoparticles dispersion and 1.5 ml of phosphate buffer pH 7 (2mM) were poured in the donor compartment of FdC, while the receptor compartment was filled with 12 ml of the same buffer solution. In the case of the testosterone, the buffer solution was changed due to the low solubility of this drug. In this case, the medium consisted in phosphate buffer pH 7 (2mM) with addition of 2% (v/v) Igepal® CA-630 and 0.4% (v/v) ethanol. In all cases, donor and receptor compartment were separated by a cellulose membrane with a pore size of 12-14 kDa. (Medicell Int. Ltd, London, UK.). FdCs were incubated at 32ºC for a period of time of 72 hours. Samples of 0.4 ml were removed from the receptor compartment at regular time intervals up to 72h and replaced with an equal volume of fresh buffer. The concentration of free drug from the receptor was analyzed by HPLC [1, 2].

2.6. Skin preparation

a For more details see section 2.5

Paper IX 187

Excised human skin from Caucasian female patients, who had undergone abdominal plastic surgery, was used. The Ethical Committee of the Caritas-Traegergesellschaft (6th July 1998), Trier, Germany, approved the procedure used. Adequate health and no medical history of dermatological disease were required. After excision, the skin was cut into 10 × 10 cm2 pieces and the subcutaneous fatty tissue was removed from the skin specimen using a scalpel. Afterwards, the surface of each specimen was cleaned with water, wrapped in aluminium foil and stored in polyethylene bags at –26°C until use. Previous investigations have shown that no change in the penetration characteristics occurs during a storage time of 6 months [3, 4].

Disks of 25 mm in diameter were punched out from frozen skin, thawed, cleaned with MilliQ water, and directly used to prepare heat separated epidermis sheets for the permeation experiments.

2.7. Heat separated epidermis preparation

The epidermis was separated placing the thawed and cleaned skin disk in water at 60°C for 90 seconds. After that, the skin was removed from the water and placed, dermal side down, on a filter paper. The epidermal layer was peeled off from the skin using forceps.

2.8. Permeation of flufenamic acid, testosterone and caffeine through human heat separated epidermis

Permeation studies of the three encapsulated drugs in P1 or P2 nanoparticles were carried out using FdC. In this case, a HSE disk was mounted on a cellulose membrane disk. Donor and receptor compartment were separated by these disks. All experimental conditions were the same that in the case of the release, except in the case of the testosterone. In this case the receptor compartment was filled only with phosphate buffer pH7 (2 mM), due to that the interaction of the epidermis with Igepal® CA-630 can generate epidermis degradation products that interfere in the quantification of the testosterone. FdCs were incubated at 32ºC for a period of time of 30 hours. Samples of 0.4 ml were removed from the receptor at regular time intervals up to 30h and replaced with an equal volume of fresh buffer. The concentration of free drug from the receptor was analyzed by HPLC [1, 2].

From the drug permeation studies trough HSE the apparent permeation constant (Papp) was calculated. Papp is defined as the drug transport speed trough the membrane, and can be used to test the potential applicability of a colloidal system as TDDS. Briefly, if the transport through stratum corneum is the rate limiting process, after a certain lag-time steady-state conditions will be achieved. Assuming a homogeneous membrane, the drug permeation can be described by diffusion and Fick’s first law (for more details see [5]). Which can be expressed as:

( )app v lagQ P AC t t= − (2)

188 Paper IX

0,01 0,1 1

140

160

180

200

220

240

260

Siz

e (n

m)

Cryoprotectant %(w/v)

(a)

0,01 0,1 10,0

0,1

0,2

0,3

0,4

0,5

0,6

PD

I

Cyoprotectant % (w/v)

(b)

Figure 1. Hydrodynamic mean size (a) and PDI (b) of reconstituted P1 nanoparticles after lyophilization as a function of cryoprotectant concentration. Trehalose (closed squares, solidline); sucrose (open squares, dashed line). Straight solid line represents the size (a) and PDI (b) of un-lyophilised nanoparticles.

where Q is the amount of solute permeated, A is the membrane area, Cv is the drug concentration of the donor, t is the exposure time, tlag is the lag-time. The rearrangement of the equation (2) leads to:

app v sslag

QP C Jt t

= =−

(3)

where Jss is the flux at steady state. Jss correspond to the slope of the linear part of the diagram of the permeated amount of drug per area as a function of the time. Then, Papp was calculated by equation (3) knowing Cv and Jss.

Similar experiments were also undertaken in parallel to obtain the Paap value of the free drugs (Papp*). To achieve this purpose the donor compartments of FdC, prepared as was commented above, were filled with a solution of free drug at the same concentration of the nanoparticles suspension. By comparison of Papp and Papp* it was possible to determine the potential enhancer effect of these nanoparticles.

2.9. Determination of flufenamic acid,, testosterone and caffeine

As was comment above drug content of each sample was determined by HPLC: Chromeleon ™ version 6.5 SP2, build 968; P580 pump; ASI-100 automated sample injector; STH 585 column oven; UVD 170S detector (Dionex Softro GmbH, Germering, Germany); column LiChrospher 100 RP-18, 5 μm, 125*4 mm (Merck, Darmstadt, D). The used methods for each drug have been previously validated [1, 2].

Paper IX 189

III. RESULTS

3.1. Lyophilization assay

A typical limitation of colloidal systems suspended in aqueous mediums is their low stability when conserved for a longer time. For this reason, in order to obtain promising drug carriers this barrier must be solved. Lyophilization is a good alternative to obtain long-term stable systems [6]. With this idea in mind, the lyophilization of P1 and P2 nanoparticles was studied as a function of the type and concentration of cryoprotectant agent. Figures 1a-b and 2a-b display the obtained results. As shown in Figure 1a (P1), sucrose seemed to be a better cryoprotectant agent than trehalose for concentrations up to 0.5% (w/v). However, regarding the PDI data of the resuspended nanoparticles, Figure 1b, both cryoprotectant agents showed a PDI higher than 0.2 for concentrations lower than 0.5% (w/v).

Table 1. Hydrodynamic mean size (nm), PDI, ζ-potential (mV), EE (%) and Papp (10-6 cm/s) values of P1 and P2 nanoparticles with FFA, testosterone or caffeine encapsulated. Papp*, Papp value calculated for a solution of un-encapsulated drug.

On other hand, reconstituted P2 nanoparticles (Figures 2a-b) displayed a similar size and PDI as the un-lyophilize nanoparticles. This pattern was totally independently of the concentration and type of cryoprotectant agent. Figure 3a

Encap. Drug

Size (nm) PDI ζ-Pot.

(mV) EE (%)

Papp (10-6 cm/s)

Papp* (10-6 cm/s)

P1 FFA 159.7 ± 1.8 0.14 ± 0.02 -15.6 ± 1.1 >95 2.43 ± 0.42 0.21 ± 0.05 Test 148.4 ± 0.7 0.10 ± 0.01 -14.1 ± 3.3 >95 0.40 ± 0.10 0.32 ± 0.07 Caff 158.8 ± 1.5 0.11 ± 0.01 -16.4 ± 2.3 >80 0.08 ± 0.01 0.09 ± 0.01 P2 FFA 185.5 ± 3.4 0.06 ± 0.02 -12.3 ± 1.5 >95 3.11 ± 0.41 0.21 ± 0.05 Test 176.6 ± 6.0 0.11 ± 0.04 -12.7 ± 0.6 >95 0.43 ± 0.11 0.32 ± 0.07 Caff 183.3 ± 5.4 0.11 ± 0.01 -10.3 ± 1.2 >80 0.11 ± 0.03 0.09 ± 0.01

0,01 0,1 1

130

140

150

160

170

180

190

200

210

220

Siz

e (n

m)

Cryoprotectant %(w/v)

a

0,01 0,1 10,00

0,05

0,10

0,15

0,20

0,25

0,30

0,35

0,40

b

PDI

Cryoprotectant % (w/v)

Figure 2. Hydrodynamic mean size (a) and PDI (b) of reconstituted P2 nanoparticles after lyophilization as a function of cryoprotectant concentration. Trehalose (closed circles, solidline); sucrose (open circles, dashed line). Straight solid line represents the size (a) and PDI(b) of un-lyophilised nanoparticles.

190 Paper IX

(a)

(b)

Figure 3. AFM images of lyophilized P2 (a) and FFA loaded P1 (b) nanoparticles.

depicts an AFM image of reconstituted P2 nanoparticles that were lyophilized without cryoprotectant agent. Comparison of Figures 3a and 2b of the Paper IV shows that the reconstituted nanoparticles presented a similar appearance to the un-lyophilised nanoparticles.

3.2. Encapsulation of flufenamic acid, testosterone and caffeine

Table 1 summarizes the main characteristics of P1 and P2 nanoparticles as a function of the encapsulated molecule. When compared to un-loaded nanoparticles (Table 1 Paper IV), it is easy to see that the encapsulation of FFA, testosterone or caffeine in P1 or P2 nanoparticles did not alter their original hydrodynamic mean size and PDI. In other respect, the increase of the magnitude of ζ-potential of both colloidal systems can be attributed to the accumulation of at least one part of these drugs on the nanoparticles surface [7]. As can be seen in the AFM image (Figure 3b), encapsulation of FFA did not produce any intelligible change in the spherical

Paper IX 191

10 20 30 40 50 60 70 800

10

20

30

40

50

60

70

80

90

100R

elea

se (%

)

Time (h)

a

10 20 30 40 50 60 70 800

10

20

30

40

50

60

70

80

90

100

Rel

ease

(%)

Time (h)

(b)

Figure 4. Release profile of FFA ( , solid line), testosterone ( , dotted line) and caffeine ( , dashed line) from P1 (a) and P2 (b) nanoparticles.

shape and soft surface of P1 nanoparticles; similar results were found for testosterone and caffeine and with P2 nanoparticles (data not shown).

Finally, it is also interesting to note that P1 and P2 nanoparticles exhibited a high EE for the three tested drugs, >95% for FFA and testosterone; and >80% for caffeine.

0 5 10 15 20 25 30-1

0

1

2

3

4

5

6

7

8

0 5 10 15 20 25 30 350

2

4

6

8

10

12

μg F

FA/c

m2 H

SE

Time (h)

μg F

FA/c

m2 C

ellu

lose

Time (h)

Figure 5. Permeation profile of FFA through HSE. Encapsulated in P1 ( ) or P2 ( )nanoparticles and free drug ( ). In the permeation profile through cellulose (little graph) thesymbols have their usual significant but are represented in an empty form.

192 Paper IX

3.3. In-vitro release of flufenamic acid, testosterone and caffeine

Once FFA, testosterone and caffeine were encapsulated in P1 and P2 nanoparticles, next step was to study their release patterns. For P1 nanoparticles (Figure 4a), FFA and testosterone presented a sustained release without any burst effect and a nearly linear profile. On other hand, a clear biphasic release profile was shown by caffeine. For this drug, an initial burst effect is followed by close to zero release kinetic. P2 nanoparticles, Figure 4b presented different release pattern as a function of the encapsulated drug. FFA presented the lowest release rate, followed by testosterone. Both drugs showed a negligible burst effect and a nearly linear profile. Finally, caffeine showed an initial burst effect, more pronounced than in the case of P1, followed by close to zero release kinetic.

3.4. Permeation of flufenamic acid, testosterone and caffeine

For a better comparison, release profiles of each drug from both nanoparticle systems -plotting the permeation through cellulose instead of Release (%)- have been inserted on the HSE permeation graphics.

Permeation profiles of FFA are illustrated in Figure 5, Papp values calculated from these profiles are summarized in Table 1. Clear differences were found between the patterns and Papp values of free and encapsulated molecules. Moreover, Papp values calculated with both colloidal systems were significantly different (ANOVA, α < 0.05), being higher for P2 nanoparticles. It is worthy to

0 5 10 15 20 25 30-0,2

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

1,8

0 5 10 15 20 25 30 350

2

4

6

8

10

12

14

μg T

esto

ster

one/

cm2 H

SE

Time (h)

μg T

esto

ster

one/

cm2 C

ellu

lose

Time (h)

Figure 6. Permeation profile of testosterone through HSE. Encapsulated in P1 ( ) or P2 ( ) nanoparticles and free drug ( ). In the permeation profile through cellulose (little graph)the symbols have their usual significant but are represented in an empty form.

Paper IX 193

0 2 4 6 8 10 12

0,00

0,05

0,10

0,15

0,20

0 2 4 6 8 10 120

10

20

30

40

50

60

μg C

affe

ine/

cm2 H

SE

Time (h)

Time (h)

μg C

affe

ine/

cm2 C

ellu

lose

Figure 7. Permeation profile of caffeine through HSE. Encapsulated in P1 ( ) or P2 ( )nanoparticles and free drug ( ). In the permeation profile through cellulose (little graph) thesymbols have their usual significant but are represented in an empty form.

remark the similitude between both permeation experiments, through cellulose and HSE, see Figure 5. Regarding to testosterone, Figure 6, both types of drug, encapsulated or not, displayed very similar permeation profiles, indicating a null nanoparticle effect in the permeation. As was expected from Figure 6, not significant differences (ANOVA, α < 0.05) were found among the different Papp values (Table 1). Comparison with the permeation through cellulose gives us the idea that the released testosterone hardly can pass through the HSE. Similar results were obtained with caffeine (Figure 7); its Papp values indicate that this molecule showed the lowest permeation velocities of the three drugs. It is worthy to remark for this drug the huge differences found between the release patterns (Fig. 4a-b) and the permeation experiments (Figure 7).

IV. DISCUSSION

As can be deduced from the previous lines, the long-term colloidal stability is a serious handicap against the clinical use of nanoparticles systems [6]. For this reason, lyophilization is advisable. This is why we studied the freeze-drying behaviour of P1 and P2 nanoparticles. It is well known that lyophilization may generate a high stress that could destabilize colloidal systems. Hence, the use of cryoprotectant agents to protect the nanoparticles integrity from the freezing stress is often necessary [6]. With this idea in mind, we selected two different disaccharides, sucrose and trehalose, as cryoprotectant agents, as they are well known for their good cryoprotectant properties [8-10].

After lyophilization and redispersation in water, the average size of the nanoparticles was analyzed (see Figures 1a-b and 2a-b). P1 nanoparticles with a low

194 Paper IX

cryoprotectant concentration, lower than 0.5% (w/v), sucrose presented better results than trehalose, but not optimal. At higher disaccharide concentrations similar results were found with both of them. Then, in order to recover the initial properties of the un-lyophilized P1 nanoparticles, it was necessary to add at least a concentration of cryoprotectant agent higher than 0.5% (w/v), either sucrose or trehalose. Similar behaviour between both disaccharides was previously published for others authors [11]. However, it is easy to find in the literature that trehalose seems to be the preferable cryoprotectant due to its better properties in comparison with other sugars. Among these properties, it is remarkable its higher glass transition temperature Tg [12, 13]. Crowe et al [14] found that, for long storage times trehalose is better cryoprotectant than sucrose, due to its higher Tg. At this moment it is necessary to remark that P1 nanoparticles were resuspended in MilliQ water just after lyophilization and may be, for this reason, it was not possible to find a clear different between sucrose and trehalose.

For P2 nanoparticles, from Figures 2a-b and 3a, it is possible to conclude that it was not necessary the use of any cryoprotectant agent to obtain reconstructed nanoparticles with similar properties than the un-lyophilized nanoparticles. Experimental results commented in the Paper IV indicated that both nanoparticles present different surface composition, being the P2 nanoparticles enriched in surfactant in comparison with the P1. Different publications have shown that a fraction of this surfactant is strongly attached to the nanoparticle surface. Moreover, several authors have demonstrated that this surfactant shell can enhance the nanoparticles stability during freeze-drying, being even innecessary the use of additional cryoprotectants. Then, not only bibliography but also our own results suggest that the excess of surfactant presented in the surface of the P2 nanoparticles, in comparison with P1, is the responsible of its higher stability.

Finally, encapsulation and release properties of P1 and P2 nanoparticles were analyzed. Encapsulation of FFA, testosterone or caffeine did not produce a clear effect in the size and PDI of both colloidal systems, see Table 1 and Figures 3b and 2a from Paper IV. Regarding to their ζ-potential, the change of this magnitude can be attributed to the accumulation of, at least, part of the drug on the surface of the nanoparticles [7]. On other hand, considering the chemical characteristics of the amphiphilic polymers and encapsulated drugs, EE summarized in Table 1 can be classified as a function of the hydrophobic character of the encapsulated macromolecule, being higher and very similar for FFA and testosterone (Log P ~ 5.0 and ~ 3.5, respectively) and lower for the less hydrophobic drug, caffeine (Log P ~ -0.1).

Figures 4a-b depict the release profile of the three drugs from the P1 and P2 nanoparticles respectively. At was commented in the Paper IV these nanoparticles presented a swelling behaviour. For this reason, prior to start the discussion it is worthy to remark that during the release experiences swelling pattern was not exhibited by any of the colloidal systems. Just after mixing of both nanoparticles systems with the release medium (phosphate buffer pH7, 2mM) the size of both

Paper IX 195

systems rapidly decreased but, afterwards, it was constant during the rest of the incubation time (data not shown). For this reason we have not considered the swelling effect in the release patterns.

To simplify the discussion, P1 nanoparticles will be analyzed firstly. The three release profiles illustrated in Figure 4a can be clearly separated in two different groups. The first group was composed by FFA and testosterone, while caffeine forms the second one. As was commented above FFA and testosterone present hydrophobic character, being higher for FFA (LogP ~ 5.0) than for testosterone (LogP ~ 3.5). On other hand, caffeine presents hydrophilic character (LogP ~ -0.1) and positive net charge under release conditions (pkb ~ 10.5). Hence, from Figure 4a there is evidence to suggest that the release pattern from P1 nanoparticles was mainly controlled by the hydrophobic interaction between the encapsulated macromolecule and the nanoparticle matrix. It is also interesting to note that the ionic drug state did not play an evident role in the release pattern. Therefore, considering the amphiphilic polymer structure, without any ionizable group, it was plausible that the diffusion of the encapsulated drug from the nanoparticle to the bulk water phase was governed by drug-matrix hydrophobic interactions [15].

Figure 4b illustrates the release profiles of P2 nanoparticles. In this case three drugs displayed three different release patterns. Release velocity followed the sequence caffeine >> testosterone > FFA. Considering the drug characteristics described above, this sequence fits perfectly with the drug hydrophilic character, caffeine >> testosterone > FFA, independently of the ionic drug state. Hence, for both nanoparticles suspensions the control of the release pattern can be explained in base of the drug-polymer hydrophobic interactions, which were more intense for the P2 case. This can be attributed to the higher hydrophobic character of P2 matrix with respect to P1 [15]. This allows discrimination between FFA and testosterone, in spite of their similar Log P, being more favourable the FFA-P2 than the testosterone-P2 interaction. Concomitantly, it is also logic the faster release of caffeine showed for P2 nanoparticles due to the worst interaction caffeine-P2 than the caffeine-P1 matrix.

As a function of the stability results discussed previously (see Paper IV), topical application seems to be the best administration route for these nanoparticles. Hence, the goal of the last section was to analyze the capacity of these colloidal systems as TDDS. For a better understanding, discussion will be separated as a function of the encapsulated drug. Previously to discuss these results, it is worthy to remark that nowadays the mechanism by which nanoparticulate formulations could facilitate skin transport remains ambiguous [16].

For FFA, the use of nanoparticles showed a very sensitive increase in its Paap value, more than 10-fold (see Table 1). Similar results were published by Luengo et al using PLGA nanoparticles [1], and by other authors with other drugs

196 Paper IX

[16-18]. Moreover, Papp values from FFA encapsulated in both nanoparticles showed significant differences, being higher for P2 nanoparticles. As was reported previously by Kaur et al, a higher presence of the surfactant in the nanoparticles shell confers a higher viscosity of this shell. This can increase the drug-skin contact time and then a better drug penetration will be obtained. On the other hand, although testosterone presents good properties to across the stratum corneum (Log P ~ 3.5 and Mw < 400), a null enhancer effect was obtained when testosterone was encapsulated. However, it is worthy to note that these permeation studies through HSE were carried out with a donor medium with Igepal® CA-630 and a receptor medium free of Igepal® CA-630. As was commented earlier, this surfactant was necessary to increase the solubility of the testosterone, which allowed us to work under sink conditions [19]. However it was not possible to use it in these experiences due to its interaction with HSE. For this reasons, it is possible to postulate that the similar Papp values for loaded and un-loaded testosterone could be attributed to the different capacity of donor and receptor medium to solubilize the testosterone [19]. Finally, similar results were obtained with caffeine, which are in agreement with those previously reported by Schäfer-Korting et al [2]. Nevertheless, in this case the difference between permeation through cellulose and HSE can not be attributed to the medium composition. The low Papp values of loaded and un-loaded caffeine can be due to its inherent physicochemical characteristics [20, 21]. Considering Log P and Mw of caffeine the expression calculated by Potts and Guy [20] estimates a Paap ~ 0.03·10-6 cm/s for this molecule, which is too similar to that obtain for us. Then, the permeation results from the three drugs suggest that both nanoparticles can transport the drug toward the skin surface but can not produce any clear effect on the permeation of the drug through HSE. Hence, although we do not have images from the nanoparticles skin penetration, the low drug permeation obtained with testosterone and caffeine and the physicochemical characteristics of P1 and P2 nanoparticles hint that these colloidal systems did not cross the HSE [21].

V. CONCLUSIONS

In this study, the application of nanoparticles formulated with amphiphilic polymers as TDDS was analyzed. Superficial differences between P1 and P2 nanoparticles, previously determined in the Paper IV, were obvious concerning to their behaviour when were lyophilized. Regarding to the drug delivery properties of both colloidal systems, the different hydrophobic character of P1 and P2 polymers was evident in the release profiles, but not in EE. Showing both systems a close to linear release pattern for hydrophobic molecules. Finally, P1 and P2 nanoparticles showed a clear enhancer effect as TDDS for FFA.

ACKNOWLEDGEMENTS

Authors thank the financial support given by Galenos Fellowship in the Framework of the EU Project “Towards a European PhD in Advanced Drug Delivery”, Marie Curie Contract MEST-CT-2004-404992.

Paper IX 197

References

1. Luengo J, Weiss B, Schneider M, Ehlers A, Stracke F, Konig K, et al. Influence of nanoencapsulation on human skin transport of flufenamic acid. Skin Pharmacology and Physiology 2006;19(4):190-197.

2. Schafer-Korting M, Bock U, Gamer A, Haberland A, Haltner-Ukomadu E, Kaca M, et al. Reconstructed human epidermis for skin absorption testing: Results of the German prevalidation study. Atla-Alternatives to Laboratory Animals 2006;34(3):283-294.

3. Wagner H, Kostka KH, Adelhardt W, Schaefer UF. Effects of various vehicles on the penetration of flufenamic acid into human skin. European Journal of Pharmaceutics and Biopharmaceutics 2004;58(1):121-129.

4. Bronaugh RL, Stewart RF, Simon M. Methods for Invitro Percutaneous-Absorption Studies .7. Use of Excised Human-Skin. Journal of Pharmaceutical Sciences 1986;75(11):1094-1097.

5. Luengo J. Human skin drug delivery using biodegradable PLGA-nanoparticles: Saarland University; 2007.

6. Abdelwahed W, Degobert G, Stainmesse S, Fessi H. Freeze-drying of nanoparticles: Formulation, process and storage considerations. Advanced Drug Delivery Reviews 2006;58(15):1688-1713.

7. Santander-Ortega MJ, Bastos-Gonzalez D, Ortega-Vinuesa JL. Electrophoretic mobility and colloidal stability of PLGA particles coated with IgG. Colloids and Surfaces B-Biointerfaces 2007;60(1):80-88.

8. Chacon M, Molpeceres J, Berges L, Guzman M, Aberturas MR. Stability and freeze-drying of cyclosporine loaded poly(D,L lactide-glycolide) carriers. European Journal of Pharmaceutical Sciences 1999;8(2):99-107.

9. Quintanar-Guerrero D, Ganem-Quintanar A, Allemann E, Fessi H, Doelker E. Influence of the stabilizer coating layer on the purification and freeze-drying of poly(D,L-lactic acid) nanoparticles prepared by an emulsion-diffusion technique. Journal of Microencapsulation 1998;15(1):107-119.

10. Sameti M, Bohr G, Kumar MNVR, Kneuer C, Bakowsky U, Nacken M, et al. Stabilisation by freeze-drying of cationically modified silica nanoparticles for gene delivery. International Journal of Pharmaceutics 2003;266(1-2):51-60.

11. Liao YH, Brown MB, Quader A, Martin GP. Protective mechanism of stabilizing excipients against dehydration in the freeze-drying of proteins. Pharmaceutical Research 2002;19(12):1854-1861.

12. Crowe JH, Hoekstra FA, Crowe LM. Anhydrobiosis. Annual Review of Physiology 1992;54:579-599.

13. Simperler A, Kornherr A, Chopra R, Bonnet PA, Jones W, Motherwell WDS, et al. Glass transition temperature of glucose, sucrose, and trehalose: An experimental and in silico study. Journal of Physical Chemistry B 2006;110(39):19678-19684.

14. Crowe LM, Reid DS, Crowe JH. Is trehalose special for preserving dry biomaterials? Biophysical Journal 1996;71(4):2087-2093.

15. Arifin DY, Lee LY, Wang CH. Mathematical modeling and simulation of drug release from microspheres: Implications to drug delivery systems. Advanced Drug Delivery Reviews 2006;58(12-13):1274-1325.

198 Paper IX

16. Alvarez-Roman R, Naik A, Kalia YN, Guy RH, Fessi H. Enhancement of topical delivery from biodegradable nanoparticles. Pharmaceutical Research 2004;21(10):1818-1825.

17. Alvarez-Roman R, Naik A, Kalia Y, Guy RH, Fessi H. Skin penetration and distribution of polymeric nanoparticles. Journal of Controlled Release 2004;99(1):53-62.

18. Toll R, Jacobi U, Richter H, Lademann J, Schaefer H, Blume-Peytavi U. Penetration profile of microspheres in follicular targeting of terminal hair follicles. Journal of Investigative Dermatology 2004;123(1):168-176.

19. Kaca M, Bock U, Jalal MT, Harms M, Hoffman C, M�ller-Goymann C, et al. Physicochemical parameters of marker compounds and vehicle properties for in vitro percutaneous absorption studies. Atla-Alternatives to Laboratory Animals 2008.

20. Potts RO, Guy RH. Predicting Skin Permeability. Pharmaceutical Research 1992;9(5):663-669.

21. Cevc G. Lipid vesicles and other colloids as drug carriers on the skin. Advanced Drug Delivery Reviews 2004;56(5):675-711.

Curriculum Vitae

CURRICULUM VITAE Manuel J. Santander-Ortega

Manuel J. Santander-Ortega CURRICULUM VITAE

PERSONAL INFORMATION Name and Surname MANUEL J. SANTANDER-ORTEGA Date and Place of Birth DECEMBER , 27TH, 1980. GRANADA Nationality SPANISH Marital Status SINGLE Address APPLIED PHYSICS DEPARTMENT, SCIENCES FACULTY, UNIVERSITY

OF GRANADA. CAMPUS DE FUENTENUEVA, S/N, 18071, GRANADA, SPAIN.

Telephone Number 0034 958 24 61 75 E-mail Address mjsant@ugr.es EDUCATION UNIVERSITY Master Degree Science and Technology of Colloids and Interfaces, Faculty of Sciences, University of Granada, Spain. September 2003-September, 2005 Graduate in Chemistry Faculty of Sciences, University of Granada, Spain. September, 1998 – September, 2003 LANGUAGE SKILLS Mother tongue Spanish Intermediate level in English

TECHNICAL SKILLS EXPERIMENTAL TECHNIQUES

CURRICULUM VITAE Manuel J. Santander-Ortega

- Turbiscan Classic MA2000 - Beckman DU7040 Spectrophotometer - ALV®-NIBS/HPPS. (High Performance Particle Sizer/ Non Invasive Back Scattering) - Brookhaven Zeta-PALS. - Malvern Nano-ZS - Malvern 4700c (Photon Correlations Spectroscopy-PCS) - AFM (Atomic Forces Microscopy) - HPLC - Franz-cell Diffusion

KNOWLEDGE OF COMPUTER SCIENCE

- Office tools: Microsoft Office, Word Perfect, StarOffice - Scientific tools: Microcal Origin. - Graphics tools: Paint Shop Pro, Adobe Photoshop

GRANTS RECEIVED EST Marie Curie Fellowship April, 15th, 2007–April, 14th, 2008 Department of Biopharmaceutics and Pharmaceutical Technology, Universität des Saarlandes (Germany). Financial support from: “Galenos Network MEST-CT-2004-504992” Short Term Fellowship. January, 11th - March, 23rd, 2006. Department of Pharmacy and Pharmaceutical Technology. University of Santiago de Compostela (Spain). Financial support from: “Universidad de Granada, Vicerrectorado de investigación y tercer ciclo” Research Introductory fellowships. October, 1st, 2002- June, 30th, 2003. University of Granada (Spain). Financial support from: “Ministerio de Educación, Cultura y Deporte”. TEACHING EXPERIENCE

24 hours of Teacher of Lab course II: Preparation of Polymeric Nanoparticles of 7th Conference and Workshop on Biological Barriers and Nanomedicine - Advanced Drug Delivery and Predictive non vivo Testing Technologies. February 20th – 29th, 2008, Department of Biopharmaceutics and Pharmaceutical Technology, Universität des Saarlandes (Germany).

Manuel J. Santander-Ortega CURRICULUM VITAE

STAGES IN OTHER LABORATORIES Department of Biopharmaceutics and Pharmaceutical Technology, Universität des Saarlandes. Person in charge: Professor C.M. Lehr. April, 15th, 2007 - April, 14th, 2008 Department of Pharmacy and Pharmaceutical Technology. University of Santiago de Compostela. Person in charge: Professor M.J. Alonso. January, 11th -July, 11th, 2006 COURSES 7th Conference and Workshop on Biological Barriers and Nanomedicine - Advanced Drug Delivery and Predictive non vivo Testing Technologies. February 20th – 29th, 2008, Department of Biopharmaceutics and Pharmaceutical Technology, Universität des Saarlandes, Saarbrücken (Germany). Skin barrier function: pharmaceutical and cosmetic applications European IP Galenos course. September 16th- October 2nd, 2007. University Claude Bernard de Lyon and University of Geneva, Lyon (France). 5th Thematic Workshop and Training Course on “GALENOS-TOWARDS A EUROPEAN PhD IN ADVANCED DRUG DELIVERY”. September 8th-11th, 2007. Trinity Collage, Dublin. (Ireland). PUBLICATIONS 1- Protein-Loaded PLGA nanoparticles for parenteral administration Journal of Biophysics, Submitted M.J. Santander-Ortega, D. Bastos-González, J.L. Ortega-Vinuesa and M.J. Alonso. 2- Characterization of core-shell lipid-chitosan and lipid-poloxamer nanocapsules Journal of Biomaterial Science: Polymer Edition, Submitted M.J. Santander-Ortega, M.V. Lozano-López, D. Bastos-González, J.M. Peula-García and J.L. Ortega-Vinuesa 3- Hofmeister effects in colloidal systems: Influence of the surface Nature. PhysChem-ChemPhys, Accepted T. López-León, M.J. Santander-Ortega, J.L. Ortega-Vinuesa and D. Bastos-González 4- Insulin-Loaded PLGA nanoparticles for oral administration: An in vitro physico-chemical characterization Journal of Biomedical Nanotechnology, Accepted M.J. Santander-Ortega, D. Bastos-González and J.L. Ortega-Vinuesa and M.J. Alonso

CURRICULUM VITAE Manuel J. Santander-Ortega

5- Electrophoretic Mobility and Colloidal Stability PLGA particles coated with IgG Colloids an d Surfaces B: Biointerfaces,Volume 60, Issue 1, October 2007, Pages 80-88 M.J. Santander-Ortega, D. Bastos-González and J.L. Ortega-Vinuesa 6- Stability and Physicochemical Characteristics of PLGA, PLGA:poloxamer and PLGA:poloxamine Blend Nanoparticles. A Comparative Study Colloids and Surfaces A: Physicochemical and Engineering Aspects, Volume 296, Issues 1-3, March 2007, Pages 132-140 M.J. Santander-Ortega, N. Csaba, M.J. Alonso, J.L. Ortega-Vinuesa and D. Bastos-González 7- Colloidal Stability of Pluronic F68-coated PLGA nanoparticles: A variety of stabilisation mechanisms Journal of Colloid and Interface Science, Volume 302, Issue 2, October 2006, Pages 522-529 M.J. Santander-Ortega, A.B. Jódar-Reyes, N. Csaba, D. Bastos-González and J.L. Ortega-Vinuesa BOOK CHAPTERS Caracterización de nanopartículas de PLGA-Poloxámero y PLGA-Poloxamina. M.J. Santander-Ortega; N. Csaba; M.J. Alonso; J.L. Ortega-Vinuesa; D. Bastos-González. Ediciones Universidad de Salamanca. Ed: M. Mercedes Velázquez Salicio y M. Dolores Merchán Moreno, ISBN: 84-7800-524-2 (Depósito Legal: S.915-2005), 2005. CONTRIBUTIONS TO CONFERENCES 1. 1st Conference on “Innovation in drug delivery: From biomaterials to devices”. April 7th-11th,

2008. Barcelona, Spain. CONTROLLED RELEASE OF NANO-ENCAPSULATED FLUFENAMIC ACID: ELIMINATION OF INITIAL BURST EFFECT. M.J. Santander-Ortega; C. Thiele; G. Wenz; U.F. Shaefer and C.M. Lehr. POSTER.

2. 7th Conference and Workshop on Biological Barriers and Nanomedicine - Advanced Drug

Delivery and Predictive non vivo Testing Technologies. February 20th – 29th, 2008, Saarbrücken, Germany. CONTROLLED RELEASE OF NANO-ENCAPSULATED FLUFENAMIC ACID: ELIMINATION OF INITIAL BURST EFFECT. M.J. Santander-Ortega; C. Thiele; G. Wenz; U.F. Shaefer and C.M. Lehr. POSTER.

3. 1st Conference on “Innovation in drug delivery: From biomaterials to devices”. September

30th-October 3rd, 2007. Naples, Italy. STABILITY AND POTENTIAL VECTORIZATION OF PLGA NANOPARTICLES. M.J. Santander-Ortega; D. Bastos-González and J.L. Ortega-Vinuesa. POSTER.

Manuel J. Santander-Ortega CURRICULUM VITAE

4. XXXI Reunión Bienal de la Real Rociedad Española de Física y 17º encuentro Ibérico para la Enseñanza de la Física. September 22nd-24th, 2007. Granada, Spain. ESPECIFICIDAD IÓNICA EN SISTEMAS COLOIDALES: INVERSIONES INDUCIDADAS POR LA SUPERFICIE EN DISPERSION. T. López-León, M.J. Santander-Ortega; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

5. XXXI Reunión Bienal de la Real Rociedad Española de Física y 17º encuentro Ibérico para la

Enseñanza de la Física. September 22nd-24th, 2007. Granada, Spain. APLICACIÓN DE PARTÍCULAS DE PLGA COMO SISTEMAS DE LIBERACIÓN CONTROLADA DE FÁRMACOS. M.J. Santander-Ortega; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

6. Skin barrier function: pharmaceutical and cosmetic applications European IP Galenos

course. September 16th- October 2nd 2007. University Claude Bernard de Lyon and University of Geneva. (France). CHARACTERIZATION OF PLGA NANOPARTICLES COATED WITH IgG.

M.J. Santander-Ortega, D. Bastos-González and J.L. Ortega-Vinuesa. ORAL PRESENTATION. 7. 21st Conference of the European Colloid and Interface Society.. September 10th-14th, 2007.

Geneva, Switzerland. HOFMEISTER EFFECTS ON HIGHLY HYDROPHILIC CHITOSAN NANOCAPSULES. M.J. Santander-Ortega; T. López-León; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

8. 5th Thematic Workshop and Training Course on “GALENOS-TOWARDS A EUROPEAN PhD

IN ADVANCED DRUG DELIVERY”. Trinity Collage, Dublin, September 8th-11th, 2007. CHARACTERIZATION OF PLGA NANOPARTICLES COATED WITH IgG.

M.J. Santander-Ortega, D. Bastos-González and J.L. Ortega-Vinuesa. ORAL PRESENTATION. 9. VII Spanish-Portuguese Conference on Controlled Drug Delivery. October 22nd-24th, 2006.

Pamplona, Spain. ELECTROPHORETIC MOBILITY AND COLLOIDAL STABILITY OF PLGA PARTICLES COATED WITH IGG. M.J. Santander-Ortega; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

10. 20th Conference of the European Colloid and Interface Society. September, 17th-22nd, 2006.

Budapest, Hungary. HOFMEISTER EFFECTS ON AGREGATION KINETICS AND FRACTAL DIMENSIONS OF POLYSTIREN NANOPARTICLES. T. Lopez-León; J.M. López-López; M.J. Santander-Ortega; A. Schmitt; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

11. 33rd Annual Meeting of Controlled Release Society. July 22nd-26th, 2006. Vienna, Austria.

CURRICULUM VITAE Manuel J. Santander-Ortega

PHYSICOCHEMICAL BEHAVIOR OF PLGA:POLOXAMER AND PLGA:POLOXAMINE BLEND NANOPARTICLES WITH PROTEINS ENCAPSULATED. M.J. Santander-Ortega; N. Csaba; M.J. Alonso; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

12. VI Reunión del Grupo Especializado de Coloides e Interfases. I Reunión Ibérica de Coloides

e Interfases. July, 13rd-15th, 2005. Salamanca, Spain. CARACTERIZACIÓN DE NANOPARTÍCULAS DE PLGA-POLOXÁMERO Y PLGA-POLOXAMINA. M.J. Santander-Ortega; N. Csaba; M.J. Alonso; J.L. Ortega-Vinuesa and D. Bastos-González. ORAL.

13. Euresco Conference. Biological Surfaces and Interfaces Euroconference on: Biomaterials.

June, 18th-23rd, 2005. Sant Feliu de Guixols, Spain. CHARACTERIZATION OF POLOXAMERS-PLGA AND POLOXAMINE-PLGA NANOPARTICLES. M.J. Santander-Ortega, N. Csaba, M.J. Alonso; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

14. VI International Symposium on Frontiers in Biomedical Polymer, FBPS05. June, 16th-19th,

2005. Granada, Spain. COLLOIDAL STABILITY OF PLGA NANOPARTICLES COVERED BY POLOXAMERS AND IGG. M.J. Santander-Ortega, N. Csaba, M.J. Alonso; J.L. Ortega-Vinuesa and D. Bastos-González. POSTER.

15. 18th Conference of the European Colloid and Interface Society. September, 19th-24th, 2004. El

Ejido, Spain. CHARACTERIZATION OF POLOXAMERS-PLGA AND POLOXAMINE-PLGA NANOPARTICLES. M.J. Santander-Ortega, N. Csaba, M.J. Alonso; J.L. Ortega-Vinuesa and D. Bastos-González . POSTER.