BASIC SCIENCE

Nanomedicine: Nanotechnology, Biology, and Medicine14 (2018) 1169–1179

Original Article

PEGylated dendritic polyglycerol conjugate targeting NCAM-expressingneuroblastoma: Limitations and challenges

Laura Isabel Vossen, MSca,1, Ela Markovsky, PhDb,1, Anat Eldar-Boock, PhDb,Harald Rune Tschiche, PhDa, Stefanie Wedepohl, PhDa, Evgeny Pisarevsky, MScb,

Ronit Satchi-Fainaro, PhDb,⁎, Marcelo Calderón, PhDa,⁎⁎aFreie Universität Berlin, Institute of Chemistry and Biochemistry, Takustrasse 3, Berlin, Germany

bDepartment of Physiology and Pharmacology, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Received 11 August 2017; accepted 10 February 2018

nanomedjournal.com

Abstract

Neural cell adhesion molecule (NCAM) is found to be a stem-cell marker in several tumor types and its overexpression is known to correlatewith increased metastatic capacity. To combine extravasation- and ligand-dependent targeting to NCAM overexpressing-cells in the tumormicroenvironment, we developed a PEGylated NCAM-targeted dendritic polyglycerol (PG) conjugate. Here, we describe the synthesis,physico-chemical characterization and biological evaluation of a PG conjugate bearing the mitotic inhibitor paclitaxel (PTX) and an NCAM-targeting peptide (NTP). PG-NTP-PTX-PEG was evaluated for its ability to inhibit neuroblastoma progression in vitro and in vivo as comparedto non-targeted derivatives and free drug. NCAM-targeted conjugate inhibited the migration of proliferating endothelial cells, suggesting itwould be able to inhibit tumor angiogenesis. The targeting conjugate provided an improved binding and uptake on IMR-32 cells compared tonon-targeted control. However, these results did not translate to our in vivo model on orthotopic neuroblastoma bearing mice.© 2018 Elsevier Inc. All rights reserved.

Keywords: Neuroblastoma; Neural cell adhesion molecule; Polyglycerol; Polymeric nanomedicines; Paclitaxel; Polymer-drug conjugates

Despite great progress in many approaches to tumoreradication, including chemotherapy, radiotherapy, immunother-apy, and surgery, the current treatments still suffer frominsufficient therapeutic efficacy and severe toxicity. Cancer isone of the major fatal diseases, mainly due to complicationscaused by acquired drug-resistance and metastases in essentialorgans and not only due to the primary tumor mass. Since there isa lack of selectivity of the drugs to cancerous cells, the mainchallenge in current chemotherapy is to improve the pharmaco-kinetics, pharmacodynamics, and the bioavailability of the drugs,and therefore avoid drug toxicity in healthy tissue.

We gratefully acknowledge the financial support from the Focus Area NanosNanoMatFutur award (13N12561), The European Research Council (ERC) ConsoInitiative (INNI), The Leona M. and Harry B. Helmsley Nanotechnology ResearPersonalized Theranostics.

⁎Correspondence to: R. Satchi-Fainaro, Department of Physiology and PharmIsrael.

⁎⁎Correspondence to: M. Calderón, Freie Universität Berlin, Institut für CheE-mail addresses: ronitsf@post.tau.ac.il (R. Satchi-Fainaro), marcelo.calde

1 Equal contribution

https://doi.org/10.1016/j.nano.2018.02.0091549-9634/© 2018 Elsevier Inc. All rights reserved.

Please cite this article as: Vossen LI, et al, PEGylated dendritic polyglycerochallenges. Nanomedicine: NBM 2018;14:1169-1179, https://doi.org/10.1016/j

One approach to improve the properties of the pharmacoki-netic profile of chemotherapeutics is to convert them to amacromolecule by conjugating them to polymers, i.e. polymertherapeutics. Here, a low molecular weight (MW) drug iscoupled to a polymer, which can alter the biodistribution of thedrug following intravenous administration. The underlyingconcept of this approach is the enhanced permeability andretention (EPR) effect originally described by Maeda andMatsumara.1,2 Many of such polymer-drug systems have beendesigned, with the goal of improving the drugs’ stability in thebloodstream to prolong their circulation time and enhance their

cale, the Bundesministerium für Bildung und Forschung (BMBF) through thelidator Award (617445 - PolyDorm), and the Israeli National Nanotechnologych Fund and by Focal Technology Area (FTA) program: Nanomedicine for

acology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978,

mie und Biochemie, Takustrasse 3, 14195 Berlin, Germany.ron@fu-berlin.de (M. Calderón).

l conjugate targeting NCAM-expressing neuroblastoma: Limitations and.nano.2018.02.009

1170 L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

tumor accumulation due to leaky tumor vasculature.3 They alsohave the advantage of allowing concomitant conjugation of atargeting moiety to the delivery system.4 This type of targetingof the drug to malignant tissue through ligand-directed targetingthat can improve the selectivity of the drug delivery system.5

We selected dendritic polyglycerol (PG) as the polymericcarrier due to its outstanding properties to be utilized as a drugdelivery system, regarding structure, biocompatibility, and watersolubility. PG consists of a biocompatible polyether scaffold andits size and functional groups can be easily adjusted.6 It can beprepared on a kilogram scale in a controlled and easy manner viaanionic ring-opening multi-branching polymerization.7 Thegreat potential of PG conjugates for biomedical applicationshas been demonstrated in vitro and in vivo.8–13 Recently, wereported the therapeutic potential of a PEGylated PG conjugatecombining doxorubicin (DOX) and paclitaxel (PTX) through thepH-cleavable 6-maleimidocaprohydrazide (EMCH) linker.14

The hydrazone bond showed sufficient stability at physiologicalpH, avoiding drug release in the circulation. Thus, in this study,we also chose to employ the pH-cleavable hydrazone bond toconjugate PTX to PG.

Functionalization with poly(ethylene glycol) (PEGylation) isused to enhance size and solubility of a nanoparticle and obtain ahigher tumor accumulation after intravenous administration, butalso to reduce immune activation.15 PEG layers (corona) lead toa longer blood circulation time of the compound, as PEG inhibitsthe attachment of opsonins and subsequent uptake by themacrophages.16

Neuroblastoma is a solid tumor derived from primitive cellsof the sympathetic nervous system. It usually arises in aparaspinal location in the abdomen or chest.17 It is the third mostcommon and deadliest tumor amongst childhood malignancies.When metastatic at diagnosis, or with molecular and geneticabnormalities, neuroblastoma represents a clinical challengehaving a 5-year event free survival of less than 50%.18,19

An effective way of targeting this childhood cancer could beby the neural cell adhesion molecule (NCAM/CD56). NCAM isa cell adhesion molecule that exists in various isoforms. It is amember of the immunoglobulin superfamily and mediatescell-cell interactions. 20,21 It has been found to be acancer-associated antigen expressed on neuroblastoma, smallcell lung cancer, melanoma, glioblastoma, Wilms tumor andothers. Moreover, increased NCAM expression is associatedwith increased metastatic capacity, expression of stem cellmarkers and poor prognosis in several tumor types.22 Sinceneuroblastoma is characterized by high expression of NCAMthat is associated with cancer progression, we chose to exploit ahuman neuroblastoma cell line, IMR-32, to evaluate theanticancer activity of an NCAM-targeted nanomedicine. Thiscell line is derived from primary stage 4 neuroblastoma withMYCN gene amplification and is known to form highlyangiogenic tumors when injected orthotopically into the adrenalgland. The IMR-32 cell line represents an appropriate metastaticdisease setting, as well as relevant microenvironment wheninoculated into the adrenal gland to evaluate anticanceragents.4,23,24

There have been several reports on NCAM-targeted therapiesfor cancer, using either antibody-based immunotherapies or

NCAM-targeting peptides (NTPs).4,25–28 The most advancedimmunoconjugate is IMGN901, which has reached phase IIclinical trials for small-cell lung cancer. However, it wasdiscontinued in 2013 due to insufficient improvement overexisting therapy. C3 peptide, originally developed as an NCAMagonist,29,30 has been successfully utilized as a targeting peptidein liposomal 26 and polymer-based4,31 drug deliveryformulations.

In this study, we chose to conjugate PTX, a microtubulestabilizing agent, as a chemotherapeutic agent. This drug hasbeen deemed one of the most significant advances in anticancertherapy.32,33 It has also shown anti-angiogenic and pro-apoptoticproperties.34 However, its poor water solubility necessitates theuse of Cremophor® EL as a solubilizing agent (CrEL, Taxol®,CrEL-paclitaxel). The side effects of Cremophor® EL includeneurotoxicity, vasodilation lethargy, and hypersensitivity reac-tions mediated by histamine release.35 Therefore, to reduce theside effects and increase the drug’s safety and aqueous solubility,an appropriate drug delivery system is required. Moreover, amethod for selective delivery of PTX to tumor cells, preventingits distribution to healthy cells immediately after administrationis equally important. In addition, only a small amount of the druglocalizes in the tumor and the drug is a substrate for effluxpumps, p-glycoprotein, resulting in multiple drug resistance.36

A 130 nm albumin-bound (nab™) PTX (Abraxane®;nab-paclitaxel) was approved by the FDA in 2005 and hasshown improved efficacy without increasing overall toxicitycompared to Cremophor® EL formulated PTX.37 Abraxane® hasbeen marketed by Celgene and was evaluated in phase III clinicaltrials for metastatic melanoma as a single agent,38 fornon-small-cell lung cancer in combination with carboplatin39

and as a monotherapy for metastatic breast cancer. It became afirst-line treatment for pancreatic cancer and non-small-cell lungcancer.40 The first PTX polymer-drug conjugate in phase IIIclinical trials for ovarian cancer or non-small cell lung cancer isPGA-paclitaxel (PGA-PTX, OPAXIO™).41,42 However, asurvival benefit was surprisingly observed in non-small celllung cancer female patients but not males. This is related to thecathepsin B-mediated activation of OPAXIO™ and therelationship between estrogen levels and cathepsin B activity.43

In this study, we wanted to achieve a selective delivery ofPTX to NCAM expressing tumors with an enhanced therapeuticeffect and reduced toxicity. Therefore, we propose the design ofa novel NCAM-targeted dendritic conjugate of PTX and anNCAM targeting peptide bound to PEGylated PG. We describethe synthesis and characterization of the conjugate and evaluateits anticancer activity on an NCAM-expressing neuroblastomamodel in vitro and in vivo.

Methods

Ethics statement

All animals were housed in the Tel Aviv University animalfacility and the experiments were approved by our institutionalanimal care and use committee (IACUC) and conducted inaccordance with NIH guidelines.

Figure 1. Synthetic routes of prodrug PTX-EMCH, PEGylated peptides NTP-PEG-mal and cNTP-PEG-mal, and the dendritic targeted PG-NTP-PTX-PEG conjugate.(A) EMCH linker was coupled to PTX-bz to obtain PTX-EMCH. i) 4-acetylbenzoyl chloride, DCM, Et3N, rt, 16 h. ii) MeOH, rt, 3 h, 93%. (B) NTP-PEG-mal andcNTP-PEG-mal were synthesized by SPPS. (C) Conjugate PG-NTP-PTX-PEG and controls were synthesized in a one-pot synthesis. DCM= dichloromethane; DIPEA =diisopropylethylamine; DMF = dimethylformamide; Et3N = triethylamine; MeOH = methanol; rt = room temperature; TFA = trifluoroacetic acid.

1171L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

For materials, general methods and synthesis see thesupplementary information.

Results

Synthesis of NCAM targeted and non-targeted PG conjugates

NTP was conjugated to PG to target cancer stem cells andtumor endothelial cells of neuroblastoma. PG with a molecular

weight of approximately 10 kDa and 30% hydroxyl groupsconverted to amine groups was selected as nanocarrier forconjugation of PTX and NTP. The prodrug PTX-EMCH wassynthesized by a three-step synthesis according to proceduresdescribed in the literature.44 Briefly, an ester derivative ofPTX at the C-2’-OH-position, PTX benzoic acid (PTX-bz),was reacted with a hydrazone linker to obtain PTX-EMCH(Figure 1). NTP and non-targeting control peptide cNTP weresynthesized using a solid phase peptide synthesis method (SPPS) on

Figure 2. Chemical structure of (A) PG-NTP-PTX-PEG conjugate and (B) the control PG-cNTP-PTX-PEG, PG-PTX-PEG and PG-PEG.

1172 L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

Sieber amide resin.4 NTP was modified with N-hydroxysuccinimide-PEG-maleimide (NHS-PEG-mal, 1.4 kDa) prior to the detachmentfrom the resin (Figure 1, B). Then, the amine groups of PG aminewere reacted with 2-iminothiolane, followed by a selective Michaeladdition between the maleimide group of PTX-EMCH,NTP-PEG-mal or cNTP-PEG-mal, and PEG-mal (2 kDa), and thesulfhydryl groups of thiolated PG (Figure 1, C). As a proof ofcoupling, an Ellman’s test was performed where the thiolconcentration was monitored via UV/Vis spectroscopy at 412nm.45 The highest thiol concentration was found after 20 min of2-iminothiolane activation.8 Conjugation formation was also con-firmed by chromatography on reverse phase thin-layer chromatogra-phy (70% acetonitrile/ 30% phosphate buffer (PB) pH 7.4) andappearance of a faster band on a Sephadex G-25 column. Theconjugates were characterized by dynamic light scattering (DLS), zetapotential measurements and the shapes were analyzed withtransmission electron microscopy (TEM).

The structures of the final PG-NTP-PTX-PEG conjugate andthe control conjugates PG-cNTP-PTX-PEG, PG-PTX-PEG andPG-PEG are depicted in Figure 2, A.

Characterization of the targeted and non-targeted conjugates

Hydrodynamic diameter, zeta potential and shapeThe hydrodynamic diameter of the conjugates was measured

by DLS. The conjugates without peptide showed Z-average

values of 44 and 23 for PG-PEG and PG-PTX-PEG, respective-ly. The NCAM-targeted (PG-NTP-PTX-PEG) and non-targetedcontrol conjugate (PG-cNTP-PTX-PEG) showed Z-averagevalues of around 70 and 102 nm (Table 1) indicating aggregationinduced by the combination of peptide and drug conjugation.Conjugates carrying only a peptide without PTX and labeledwith an indodicarbocyanide (IDCC) dye (PG-NTP-IDCC-PEGand PG-cNTP-IDCC-PEG) showed smaller Z-average valuesthan their analogues with PTX (SI Table 2). All conjugatesshowed some degree of aggregation and they have relativelyhigh polydispersity indices (PDIs) (Table 1). Present aggregationis demonstrated by comparing the different types of sizedistributions during a DLS measurement. Size distributions byvolume are shown in Figure S6. Here, PG-cNTP-PTX-PEGseems to aggregate more in solution than PG-NTP-PTX-PEG.Zeta potential at pH 7.4 was found to be close to zero forPG-PEG and PG-PTX-PEG and conjugates with NTP and cNTPshowed, as expected, a stronger positive charge (+ 25 mV) due tothe free amines of the peptides (Table 1). The shape of theconjugates was studied by TEM. In all cases, the conjugates’shape is spherical.

Drug loading and drug releaseBound PTX-bz concentrations were determined by measuring

the total released amount of PTX-bz by RP-HPLC at low pH.

Table 1Physico-chemical characterization of PG-NTP-PTX-PEG and control conjugates.

PTX loading (w%) Peptide loading (w%) Z-average [nm] PDI Zeta potential [mV] Molecular weight (theoretical) [kDa]

PG-NTP-PTX-PEG 1.8 2.7 70 0.38 + 27.9 58PG-cNTP-PTX-PEG 1.9 2.4 102 0.43 + 25.6 58PG-PTX-PEG 2.2 - 23 0.35 + 4.3 67PG-PEG - - 44 0.34 + 2.7 70

Figure 3. A) Binding (%) of PG-NTP-PTX-PEG conjugate and controls to NCAM-expressing IMR-32 neuroblastoma cells at 37 °C. B-D) Internalization of theconjugates (PG-NTP-IDCC-PEG or PG-cNTP-IDCC-PEG) visualized by confocal imaging together with LysoTracker® Green DND-26 at 37 °C. B) Livesuspended IMR-32 cells after 7 min, C-D) fixed IMR-32 cells after 7 min (C) or 30 min (D). Lysosomes are in green and conjugate material in red.

1173L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

Therefore, the conjugates were incubated in Britton-Robinsonbuffer (pH 2.0) and after 6 h, a plateau was reached and theconcentration of released PTX-bz was determined. The calcu-lated loadings are summarized in Table 1. Additionally, in vitrostability studies were performed at pH 4.0 and pH 7.4 byRP-HPLC. At pH 7.4 a release of 12% of the drug after 48 hcould be observed, while at acidic pH 4.0 the release was 49%(Figure S4). These properties should enable the drug to bereleased in the slightly acidic tumor microenvironment andintracellular environment of tumor cells (e.g. pH 5-6 in theendosomes and down to 4-5 in the lysosomes).46 PTX-bz releasefrom PG-cNTP-PTX-PEG and PG-PTX-PEG conjugatesshowed similar results (Figure S4).

Binding and internalization of NTP-targeted conjugate toNCAM-expressing cancer cells

Markovsky et al. had previously evaluated NCAM expressionin various cancer cell lines and IMR-32 neuroblastoma cell lineshowed a high percentage of NCAM-expressing cells (90%) and

was thus selected for our studies.4 PG-NTP-PTX-PEG conjugateand control PG-cNTP-PTX-PEG conjugate were labeled withIDCC dye and binding to IMR-32 cells was examined byfluorescence assisted cell sorting (FACS) (Figure 3, A).PG-NTP-PEG showed the highest binding of 83% and wasmuch higher than that of control PG-cNTP-PEG conjugate (15%,Figure 3, A). Moreover, binding of PG-NTP-PTX-PEG conju-gate to the cells was higher than control PG-cNTP-PTX-PEGconjugate (56% vs. 20%). However, when comparingPG-NTP-PEG and PG-NTP-PTX-PEG, conjugation of PTX tothe conjugate seems to interfere with the binding of the conjugateto the NCAM peptide expressed on the cell’s surface (83% vs.56%). The addition of PTX to the conjugate may affect itsmacromolecular structure and alter its spatial orientationpotentially leading to steric hindrance. NTP competes with theextracellular matrix for the adhesion molecules of the cells,hence the interaction of the conjugate with the cells results intheir rapid (1-2 min) detachment. In order to further follow theinternalization of the IDCC-labeled conjugates into IMR-32

Figure 4. Cell proliferation inhibition of IMR-32 neuroblastoma cells bydendritic PG-NTP-PTX-PEG and control conjugates.

Figure 5. HUVEC migration inhibition at 50 nM PTX-bz-equivalentconcentration following incubation of 13 hours. The ability to close thegap was measured relative to time zero. * p b 0.05, ** p b 0.01. Results arerepresentative of three experiments.

1174 L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

cells, they were examined by confocal microscopy together withLysoTracker® Green DND-26 (used as intracellular marker) at37 °C. Since the conjugates interfere with the attachment of thecells, live images of floating cells were taken following 7 min oftreatment with the conjugates. Increased attachment of PG-NTP-IDCC-PEG to the cell’s surface, compared to non-targetingcontrol is visualized by the red color of the labeled conjugatematerial. Both PG-NTP-IDCC-PEG and PG-cNTP-IDCC-PEG arenot colocalized with the lysosome following short exposure to thecompounds (Pearson correlation of 0.067 and 0.13, respectively, n=6). In a second experiment, IMR-32 cells were treated with theconjugates for 7 or 30 min, and then, cells were fixed with 4%paraformaldehyde (Figure 3, C and D). Here, a higher internaliza-tion of the targeted conjugate into the cells can be observed after 30min (Figure 3, D).

Evaluation of the antiproliferative activity of PG-NTP-PTX-PEG conjugate on cancer cells

The antiproliferative activity of the conjugates was evaluatedin vitro on NCAM expressing IMR-32 neuroblastoma cells by aproliferation assay. The cells were incubated with the conjugatesfo r 72 hou r s and coun t ed . PG-NTP-PTX-PEG,PG-cNTP-PTX-PEG and PG-PTX-PEG retained their antipro-liferative activity compared to PTX-bz alone (Figure 4).Conjugates without a drug did not show antiproliferative activityat the concentrations examined. Similar results were obtainedwith NCAM expressing cell line MEL-526 (Figure S5). The IC50

values are summarized in SI Table 1.

Evaluation of the antiangiogenic activity of PG-NTP-PTX-PEGconjugate

Markovsky et al. had previously shown that HUVEC expresshigh level of NCAM during angiogenesis, represented by highestNCAM expression during capillary formation as demonstratedby a capillary-like tube formation assay.4 The ability ofPG-NTP-PTX-PEG conjugate to inhibit the migration ofHUVEC through a gap in the monolayer was evaluated. Ascratch was done on a 90% confluent cell monolayer. Then, cellswere incubated with the conjugates or the free drug atPTX-bz-equivalent concentrations of 50 nM for 13 hours. Our

PG-NTP-PTX-PEG conjugate significantly inhibited the migra-tion of HUVEC (Figure 5).

Evaluation of antitumor activity and toxicity

In vivo study at 15 mg/kg and 30 mg/kg PTX-bz-equivalent dosemCherry-labeled IMR-32 neuroblastoma cells were in-

oculated orthotopically in the adrenal gland of SCID mice. Thedecreased fluorescence mCherry signal correlates with tumorgrowth inhibition. Treatments were initiated 2 weeks after tumorcell inoculation and administered intravenously (i.v.) everyother day at PTX-bz-equivalent concentrations of 15 mg/kg or 30mg/kg per treatment, for a total of five treatments. Eighteen daysafter treatments initiation, 3 out of 8 mice in each group wereeuthanized and tumors were weighed. Treatment withNCAM-targeted PG-NTP-PTX-PEG conjugate successfullyinhibited the growth of neuroblastoma tumors. ControlPG-cNTP-PTX-PEG conjugate and non-targeted PG-PTX-PEGconjugate inhibited tumor growth to a similar extent at aconcentration of 15 mg/kg PTX-bz-equivalent per treatment.When the dose was increased to 30 mg/kg PTX-bz-equivalent, abetter growth inhibition for both the NTP and cNTP-targetedconjugates could be observed (Figure 6). At this dose,PG-NTP-PTX-PEG was better tolerated than controlPG-cNTP-PTX-PEG conjugate as monitored by significantweight loss of the mice (Figure 7). Free PTX-bz, at 15 mg/kgdose, efficiently inhibited tumor growth and unexpectedly, micedid not suffer from significant weight loss following treatments(Figure 7). However, increasing the dose of free PTX-bz beyondits maximum tolerated dose (MTD) is unbearable following i.v.injection.

Treatment efficacy, defined as tumor growth inhibition ratiobetween treated versus control (T/C) groups was determined atthe end of the study (Table 2). Mice treated with the targeted andnon-targeted conjugates had similar survival after 30 days at aPTX-bz-equivalent dose of 15 and 30 mg/kg (Figure 7, B).

Discussion

Throughout the last decades, polymeric nanocarriers havebeen exploited as a strategy to overcome several problems

Figure 6. (A) Antitumor activity of PG-NTP-PTX-PEG conjugate and controls on orthotopic neuroblastoma mouse model (n = 6). Conjugates were administeredat a PTX-bz-concentration of 15 mg/kg. PG-NTP-PTX-PEG and PG-cNTP-PTX-PEG were also administered at a PTX-bz-concentration of 30 mg/kg. (B)Tumor weight following treatment withdrawal measured at n = 3 mice/group. Data represents mean ± SEM.

Figure 7. (A) Body weight change of mice treated with PG-NTP-PTX-PEG conjugate at a concentration of 15 mg/kg PTX-bz-equivalent treatment and 30 mg/kgPTX-bz equivalent treatment. (B) Survival of mice treated with PG-NTP-PTX-PEG and control conjugates.

1175L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

related to conventional chemotherapy. Extravasation-dependenttargeting through the leaky tumor vessels through the EPR effecthas been the focal point in several studies. However, theleakiness of the vasculature in malignant tissues can varysignificantly between tumors and even in different parts of thesame tumors.5 Therefore, ligand-dependent targeting of a drugwould be necessary.5,47 Neuroblastoma is a childhood cancerwith metastatic behavior and has poor long-term prognosis oncethe tumor becomes metastatic and drug-resistant despiteaggressive and multimodal therapies. Advances in understandingthe biochemical and genetic mechanisms of neuroblastoma’sclinical behavior, have allowed the identification of proteins,genes and pathways as effective targets for biologically-based

therapy.48 Hence, scientific research has not only focused ondeveloping nanosystems that are solely based on passivelytargeting neuroblastoma but also on potential targets for itstherapy that increase the affinity of nanocarriers to the tumortissue.5,47 There are several molecules overexpressed on tumorcells which can act as receptors for targeting moieties.49 Eachexpression is tumor-type related and a model must be chosencarefully. An approach for actively targeting neuroblastoma, isthrough an NCAM targeting peptide (NTP, C3), since NCAM ishighly expressed on neuroblastoma. Recently, Markovsky et al.showed great in vivo efficacy of the first NCAM-targetedpolymer-drug conjugate4 using NTP and paclitaxel as therapeu-tic drug. We chose to employ the same targeting peptide in our

Table 2Ratio of tumor volume in treated vs control mice at study endpoint

Saline PG-PEG PTX-bz15 mg/kg

PG-PTX-PEG15 mg/kg

PG-NTP-PTX-PEG15 mg/kg

PG-cNTP-PTX-PEG15 mg/kg

PG-NTP-PTX-PEG30 mg/kg

PG-cNTP-PTX-PEG30 mg/kg

1.00 0.99 0.01 0.44 0.39 0.38 0.03 0.06

1176 L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

system to obtain selective targeted nanomedicine for neuroblas-toma therapy.

Our conjugate consists of NTP, PTX covalently bound todendritic PG through a cleavable hydrazone bond, and PEG toenhance solubility and increase the size of the conjugate. Thepeptide is modified with a PEG-spacer and bound to PG via astable thiol-ene bond. In a previous study, with the aim oftargeting cancer cells expressing the epidermal growth factor,we showed that a PEG spacer between PG and the targetingmoiety is crucial to obtain active targeting to the cells. 10

Therefore, we used the same PEG spacer for conjugation ofNTP to PG. PTX is cleaved under acidic conditions andproved sufficient stability under physiological conditions inprevious studies.14,50 This conjugation should enable atriggered release of PTX in the endolysosomal compartmentsof tumor cells and enhance the half-life of the drug. Ourin vitro drug release experiments were designed to mimic thisscenario (Figure S4).

PG-NTP-PTX-PEG and its non-targeting controlsPG-cNTP-PTX-PEG, PG-PTX-PEG and PG-PEG were success-fully synthesized and characterized. The determined peptideloadings of NCAM-targeted (NTP) and non-targeted (cNTP)conjugates were 2.7 w% and 2.4 w%. The measured drugloadings were similar for all conjugates (1.8 w%PG-NTP-PTX-PEG, 1.9 w% PG-cNTP-PTX-PEG, and 2.2 w%PG-PTX-PEG). The hydrodynamic diameter of the conjugateswith NTP or cNTP was much larger (70 and 102 nm) comparedto conjugates without a peptide (44 nm for PG-PEG and 23 nmfor PG-PTX-PEG). These results suggest that the combination ofpeptide and PTX conjugation induces aggregation. Thisaggregation behavior is, however, only observed when PTXand a peptide are conjugated to the same nanocarrier. In the case,that PG is only carrying a peptide without the drug, lessaggregation occurs (SI Table 2). This could possibly explain whywe barely see a therapeutic improvement in vivo whencomparing targeted vs. non-targeted conjugates. We speculatethat due to aggregate formation, the peptides could be hidden andtherefore hindered from binding to NCAM tumor surfacereceptors. On the other side, in vitro binding studies toIMR-32 neuroblastoma cell line at short incubation times (10min) evidenced a three times higher binding (56%) of theNCAM-targeted conjugate (PG-NTP-PTX-PEG) compared tothe non-targeted conjugate (PG-cNTP-PTX-PEG, 20%) to thecells (Figure 3, A). However, these binding studies alreadyrevealed, that PTX conjugation reduces the binding ability whencomparing PG-NTP-PEG (83%) and PG-NTP-PTX-PEG (56%).We assume, that this effect could be attributed to sterichinderance induced by PTX. Since we designed our conjugatewith a pH-cleavable hydrazone linker, we evaluated itsinternalization into IMR-32 cells. Confocal microscopy showed

an enhanced attachment of the targeting conjugate(PG-NTP-IDCC-PEG) on the cell surface after 7 min incubationtime and an enhanced internalization into the cells after 30 min(Figure 3, B-D). Moreover, PG-NTP-PTX-PEG significantlyinhibited the migration of HUVEC compared to controlconjugates and free PTX-bz. HUVEC express high level ofNCAM during the angiogenic process, which was previouslyshown by Markovsky et al.4 and the results are also in line withthe work of Bussolati et al..21 These results suggest thatPG-NTP-PTX-PEG conjugate can inhibit proliferation of tumorvasculature as well. In summary, all in vitro results indicated thatPG-NTP-PTX-PEG would have adequate properties in order towork in vivo.

Surprisingly, in vivo, all conjugates bearing PTX had a similargrowth inhibition activity at a PTX-bz-equivalent dose of 15 mg/kg on mice bearing an orthotopic neuroblastoma tumor (Table 2).When the dose was increased to a PTX-bz-equivalent dose of 30mg / k g f o r c o n j u g a t e s PG -NTP - PTX -PEG a ndPG-cNTP-PTX-PEG, the antitumor activity was enhanced butno difference between targeted (T/C = 0.03) and non-targetedconjugate (T/C = 0.06) was detected. However ,PG-NTP-PTX-PEG conjugate caused less weight loss and wasbetter tolerated in treated mice compared to the non-targetedPG-cNTP-PTX-PEG conjugate (Figure 7). We assume that ourtargeted conjugate at a PTX-bz-equivalent dose of 30 mg/kg ispossibly up taken via receptor-mediated endocytosis throughactively targeting NCAM, then conjugate accumulation in thetumor would be increased and hence the toxicity of PTX inhealthy tissue would be reduced even when the administereddose is increased. It has already been shown in the past, that theuse of targeting moieties such as alendronate (for boneneoplasms), folic acid (for ovarian cancer), RGD peptidomi-metics (for αvβ3 integrin targeting in breast cancer andglioblastoma) and NCAM (for Wilms tumor, neuroblastoma)abrogates the neurotoxicity and body weight-loss associated withPTX or cardiotoxicity associated with doxorubicin.4,31,51,52

Moreover, PTX-bz showed a great tumor growth inhibition aswell as no additional toxicity to treated mice (Figures 6 and 7) atan administered dose of 15 mg/kg. We chose PTX modified witha benzoic acid moiety (PTX-bz) as our control free drug becausePTX-bz will be released first and only after some time, slowhydrolysis of the ester bond can occur releasing free PTX. In fact,the electron donating aromatic ring stabilizes both, hydrazoneand ester bond.53 This prodrug of PTX seems to inhibitneuroblastoma growth causing less side effects to treated micethan the original non-modified PTX. It has been demonstratedbefore, that PTX causes severe side effects and is very toxic totreated mice with orthotopic neuroblastoma tumors.4 In anotherstudy where we evaluate the anti-cancer activity of a PEGylatedPG conjugate bearing a combination of PTX and DOX in a

1177L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

murine model of mammary of adenocarcinoma we couldestablish that conjugation of PTX to our PG carrier prolongsmice survival and overall improves the therapeutic effectcompared to free PTX-bz.14 In the case of our NCAM-targetedconjugate we could overcome PTX-related toxicity byPTX-bz-conjugation to our PG nanocarrier, too. However, animproved overall therapeutic effect of the targeted conjugatecompared to the free drug could not be obtained on aneuroblastoma model.

As mentioned before, all conjugates showed aggregationmeasured by DLS and conjugates PG-NTP-PTX-PEG andPG-cNTP-PTX-PEG displayed bigger hydrodynamic diameter.These different sizes could have also influenced the in vivobehavior of the conjugates. Extravasation-dependent accumula-tion in the tumor due to the leaky vasculature could havedominated over the internalization into the cells via activelybinding to NCAM receptors expressed on the tumor surface.54

Shapes of the conjugates were analyzed by TEM measurementsshowing that all conjugates retain a spherical shape (FiguresS7-S9). Additionally, zeta potentials of the conjugates, revealedan overall positive charge for the conjugates with targeting andnon-targeting peptides. We assume that the positive charge ofthese two conjugates also influenced the uptake into cells, as hasbeen shown in the literature.11,55 However, the degree ofaccumulation of a polymeric drug delivery system in the tumor ismuch more complex and will always depend on many factorssuch as size, molecular weight, overall charge andhydrophobic-hydrophilic characteristics.56

Finally, a preliminary biodistribution study of theIDCC-labeled targeted and non-targeted conjugate was evaluat-ed in mCherry-labeled tumor-bearing mice. The results suggestthat 30 min following i.v. injection, NCAM-targeted conjugatePG-NTP-IDCC-PEG begins to accumulate in the tumor area,whereas PG-cNTP-IDCC-PEG is distributed randomly andcirculating in the blood (Figure S10A). Additionally, after 2.5h, we assume that PG-NTP-IDCC-PEG accumulates mainly inthe tumor and is cleared through the kidneys as shown in FigureS10B).

In conclusion, our results show that an NCAM-targetedPEGylated dendritic PG has its limitations and needs furtherimprovement to be a good platform for drug delivery toneuroblastoma. In this study, a targeting moiety for an increaseduptake of the nanocarrier into NCAM-expressing neuroblastomadid not lead to an improved therapeutic effect in vivo comparedto non-targeted controls and free drug. Although we coulddemonstrate a targeting effect in vitro, this was not translatableto our in vivo cancer model. This is despite the fact that anincreased administered dose of our targeted PG-NTP-PTX-PEGabrogated body weight loss associated with PTX and apreliminary biodistribution study with IDCC-labeled targetedconjugate suggests a great accumulation in the tumor.

Markovsky et al . have previously shown that anNCAM-targeted polymer-drug conjugate based on linear poly-glutamic acid (PGA) can increase the efficacy of the nanome-dicines when treating appropriate tumor models.4,31 Since ourPEGylated dendritic polymeric conjugate has fundamentaldifferences compared to the linear polymeric PGA, such asshape, zeta potential, hydrodynamic diameter, release of the drug

(pH- vs. esterase-mediated) the results are not quite comparable.Furthermore, other groups could demonstrate that a targetingmoiety does have an influence and upgrades the nanosystem forneuroblastoma treatment. To this regard, G. Zuccari et al. havedeveloped an NGR-targeted liposomal formulation, bearingencapsulated bortezomib (BTZ), a proteasome inhibitor, forneuroblastoma targeted therapy.57 NGR targets aminopeptidaseN (APN) of positive tumor endothelial cells. With their system,they could show an improved therapeutic effect in vivo, byminimizing side effects of BTZ. In another work, Cossu et al.identified a novel peptide sequence as a specific ligand foraggressive neuroblastoma and showed that their targetednanoparticles could selectively increase tumor binding andpenetration, reducing tumor growth and preventing metastaticspreading by selective delivery of doxorubicin.58 They claimedthat their system could be part of a multi-target approach, sincethere is a lot to consider from the variety of signaling pathwaysinvolved in the crosstalk of tumor and microenvironment to thevariety of cell types.

It is well known that although the experimental settings invitro for receptor-mediated targeting give mostly promisingresults, the clinical settings in vivo are still challenging.Receptor-mediated targeting requires homogeneity of receptorexpression on tumor cells, and nanocarriers should be able topenetrate the intra-tumoral extracellular matrix bypassing thebinding site barrier.59 Radioimmunotherapy for solid tumorsreveals the limitations of active targeting in the clinic, since onlya very small amount of administered antibody (0.001-0.01%)localizes in the tumor and a higher dose even causesmyelotoxicity.60

Our PEGylated dendritic PG system requires optimization toeffectively target neuroblastoma and we suggest a multi-targetPG system with an additional tumor microenvironment targetingmoiety as a suitable alternative.

Appendix A. Supplementary data

Supplementary data to this article can be found online athttps://doi.org/10.1016/j.nano.2018.02.009.

References

1. Matsumura Y, Maeda H. A new concept for macromoleculartherapeutics in cancer chemotherapy: mechanism of tumoritropicaccumulation of proteins and the antitumor agent Smancs: a newconcept for macromolecular therapeutics in cancer chemotherapy:mechanism of tumoritropic accum. Cancer Res 1986;46:6387-92.

2. Maeda H, Wu J, Sawa T, Matsumura Y, Hori K. Tumor vascularpermeability and the EPR effect in macromolecular therapeutics: areview. J Control Release 2000;65(1-2):271-84.

3. Markovsky E, Baabur-Cohen H, Eldar-Boock A, Omer L, Tiram G,Ferber S, et al. Administration, distribution, metabolism and eliminationof polymer therapeutics. J Control Release 2012;161(2):446-60.

4. Markovsky E, Eldar-Boock A, Ben-Shushan D, Baabur-Cohen H, YeiniE, Pisarevsky E, et al. Targeting NCAM-expressing neuroblastoma withpolymeric precision nanomedicine. J Control Release 2017;249:162-72.

5. Maeda H. Toward a full understanding of the EPR effect in primary andmetastatic tumors as well as issues related to its heterogeneity. Adv DrugDeliv Rev 2015;91:3-6.

1178 L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

6. Roller S, Zhou H, Haag R. High-loading polyglycerol supportedreagents for Mitsunobu- and acylation-reactions and other usefulpolyglycerol derivatives. Mol Divers 2005;9(4):305-16.

7. Sunder A, Hanselmann R, Frey H, Mu R. Controlled synthesis ofhyperbranched polyglycerols by ring-opening multibranching polymer-ization. Macromolecules 1999;32:4240-6.

8. Calderon M, Graeser R, Kratz F, Haag R. Development of enzymaticallycleavable prodrugs derived from dendritic polyglycerol. Bioorg MedChem Lett 2009;19(14):3725-8.

9. Calderón M, Welker P, Licha K, Graeser R, Kratz F, Haag R.Development of efficient macromolecular prodrugs derived fromdendritic polyglycerol. J Control Release 2010;148(1):e24-5.

10. Hussain AF, Krüger HR, Kampmeier F, Weissbach T, Licha K, Kratz F,et al. Targeted delivery of dendritic polyglycerol–doxorubicin conju-gates by scFv-SNAP fusion protein suppresses EGFR+ cancer cellgrowth. Biomacromolecules 2013;14(8):2510-20.

11. Calderon M, Reichert S, Welker P, Licha K, Kratz F, Haag R. Receptormediated cellular uptake of low molecular weight dendritic polyglycer-ols. J Biomed Nanotechnol 2014;10(1):92-9.

12. Fatemeh Sheikhi M, Juliane A, Shilpi G, Stefanie W, Marcelo C, UlrichB, et al. Bispecific antibodies for targeted delivery of dendriticpolyglycerol (dPG) prodrug conjugates. Curr Cancer Drug Targets2016;16(7):639-49.

13. Gutierrez-Corbo C, Dominguez-Asenjo B, Vossen LI, Perez-Pertejo Y,Munoz-Fenandez MA, Balana-Fouce R, et al. PEGylated dendriticpolyglycerol conjugate delivers doxorubicin to the parasitophorous vacuolein Leishmania infantum infections.Macromol Biosci 2017;17(10):1700098.

14. Baabur-Cohen H, Vossen LI, Kruger HR, Eldar-Boock A, Yeini E,Landa-Rouben N, et al. In vivo comparative study of distinct polymericarchitectures bearing a combination of paclitaxel and doxorubicin at asynergistic ratio. J Control Release 2017;257:118-31.

15. Wang M, Thanou M. Targeting nanoparticles to cancer. Pharmacol Res2010;62(2):90-9.

16. Miceli E, Kar M, Calderon M. Interactions of organic nanoparticles withproteins in physiological condit ions. J Mater Chem B2017;5(23):4393-405.

17. Brodeur GM. Neuroblastoma: biological insights into a clinical enigma.Nat Rev Cancer 2003;3(3):203-16.

18. Ronn LC, Berezin V, Bock E. The neural cell adhesion molecule insynaptic plasticity and ageing. Dev Neurosci 2000;18(2-3):193-9.

19. Cohn SL, Pearson ADJ, London WB, Monclair T, Ambros PF, BrodeurGM, et al. The International Neuroblastoma Risk Group (INRG)Classification System: An INRG Task Force Report. J Clin Oncol2009;27(2):289-97.

20. Pode-Shakked N, Metsuyanim S, Rom-Gross E, Mor Y, Fridman E,Goldstein I, et al. Developmental tumourigenesis: NCAM as a putativemarker for the malignant renal stem/progenitor cell population. J CellMol Med 2009;13(8b):1792-808.

21. Bussolati B, Grange C, Bruno S, Buttiglieri S, Deregibus MC, Tei L, etal. Neural-cell adhesion molecule (NCAM) expression by immature andtumor-derived endothelial cells favors cell organization into capillary-like structures. Exp Cell Res 2006;312(6):913-24.

22. Wachowiak R, Rawnaq T, Metzger R, Quaas A, Fiegel H, Kähler N, etal. Universal expression of cell adhesion molecule NCAM inneuroblastoma in contrast to L1: implications for different roles int umo r b i o l ogy o f neu rob l a s t oma? Ped ia t r Su rg In t2008;24(12):1361-4.

23. Daudigeos-Dubus E, L.E.D. L, Rouffiac V, Bawa O, Leguerney I,Opolon P, et al. Establishment and characterization of neworthotopic and metastatic neuroblastoma models. In Vivo2014;28(4):425-34.

24. Jensen M, Berthold F. Targeting the neural cell adhesion molecule incancer. Cancer Lett 2007;258(1):9-21.

25. Bourne SP, Patel K, Walsh F, Popham CJ, Coakham HB, Kemshead JT.A monoclonal antibody (ERIC-1), raised against retinoblastoma, thatrecognizes the neural cell adhesion molecule (NCAM) expressed on

brain and tumours arising from the neuroectoderm. J Neurooncol1991;10(2):111-9.

26. Grange C, Geninatti-Crich S, Esposito G, Alberti D, Tei L, Bussolati B,et al. Combined delivery and magnetic resonance imaging of neural celladhesion molecule-targeted doxorubicin-containing liposomes in exper-imentally induced Kaposi's sarcoma. Cancer Res 2010;70(6):2180-90.

27. Ishitsuka K, Jimi S, Goldmacher VS, Ab O, Tamura K. Targeting CD56by the maytansinoid immunoconjugate IMGN901 (huN901-DM1): apotential therapeutic modality implication against natural killer/T cellmalignancy. Haematol 2008;141(1):129-31.

28. Whittington HA, Hancock J, Kemshead JT. Generation of a humanisedsingle chain Fv (Scfv) derived from the monoclonal Eric-1 recognisingthe human neural cell adhesion molecule. Med Pediatr Oncol2001;36(1):243-6.

29. Berezin V, Bock E. NCAM mimetic peptides: Pharmacological andtherapeutic potential. J Mol Neurosci 2004;22(1-2):33-9.

30. Ronn LC, Olsen M, Ostergaard S, Kiselyov V, Berezin V, MortensenMT, et al. Identification of a neuritogenic ligand of the neural celladhesion molecule using a combinatorial library of synthetic peptides.Nat Biotechnol 1999;17(10):1000-5.

31. Markovsky E, Vax E, Ben-Shushan D, Eldar-Boock A, Shukrun R,Yeini E, et al. Wilms tumor NCAM-expressing cancer stem cells aspotential therapeutic target for polymeric nanomedicine. Mol CancerTher 2017;16(11):2462-72.

32. Arbuck SG. Taxol (paclitaxel): future directions. Ann Oncol1994;5(Suppl 6):S59-62.

33. Markman M, Mekhail TM. Paclitaxel in cancer therapy. Expert OpinPharmacother 2002;3(6):755-66.

34. Oldham EA, Li C, Ke S, Wallace S, Huang P. Comparison of action ofpaclitaxel and poly(L-glutamic acid)-paclitaxel conjugate in humanbreast cancer cells. Oncol 2000;16(1):125-32.

35. Gelderblom H, Verweij J, Nooter K, Sparreboom A. Cremophor EL.Cancer 2001;37(13):1590-8.

36. Ajabnoor GM, Crook T, Coley HM. Paclitaxel resistance is associatedwith switch from apoptotic to autophagic cell death in MCF-7 breastcancer cells. Cell Death Dis 2012;3(1):e260.

37. Green MR, Manikhas GM, Orlov S, Afanasyev B, Makhson AM, BharP, et al. Abraxane®, a novel Cremophor®-free, albumin-bound particleform of paclitaxel for the treatment of advanced non-small-cell lungcancer. Ann Oncol 2006;17(8):1263-8.

38. Hersh EM, Del Vecchio M, Brown MP, Kefford R, Loquai C, Testori A,et al. A randomized, controlled phase III trial of nab-Paclitaxel versusdacarbazine in chemotherapy-naïve patients with metastatic melanoma.Ann Oncol 2015;26(11):2267-74.

39. Miyauchi E, Inoue A, Usui K, Sugawara S, MaemondoM, Saito H, et al.Phase II study of modified carboplatin plus weekly nab-paclitaxel inelderly patients with non-small cell lung cancer: North Japan LungCancer Study Group Trial 1301. Oncologist 2017;22(6):640-e59.

40. Hare JI, Lammers T, Ashford MB, Puri S, Storm G, Barry ST.Challenges and strategies in anti-cancer nanomedicine development: Anindustry perspective. Adv Drug Deliv Rev 2017;108:25-38.

41. Galic VL, Wright JD, Lewin SN, Herzog TJ. Paclitaxel poliglumex forovarian cancer. Expert Opin Investig Drugs 2011;20(6):813-21.

42. Langer CJ, O'Byrne KJ, Socinski MA, Mikhailov SM, Lesniewski-Kmak K, Smakal M, et al. Phase III trial comparing paclitaxelpoliglumex (CT-2103, PPX) in combination with carboplatin versusstandard paclitaxel and carboplatin in the treatment of PS 2 patients withchemotherapy-naive advanced non-small cell lung cancer. J ThoracOncol 2008;3(6):623-30.

43. Melancon MP, Wang W, Wang Y, Shao R, Ji X, Gelovani JG, et al. Anovel method for imaging in vivo degradation of poly(L-glutamic acid), abiodegradable drug carrier. Pharm Res 2007;24(6):1217-24.

44. Willner D, Trail PA, Hofstead SJ, King HD, Lasch SJ, Braslawsky GR,et al. (6-Maleimidocaproyl)hydrazone of doxorubicin – a new derivativefor the preparation of immunoconjugates of doxorubicin. BioconjugChem 1993;4(6):521-7.

1179L.I. Vossen et al / Nanomedicine: Nanotechnology, Biology, and Medicine 14 (2018) 1169–1179

45. Riener CK, Kada G, Gruber HJ. Quick measurement of proteinsulfhydryls with Ellman's reagent and with 4,4'-dithiodipyridine. AnalBioanal Chem 2002;373(4-5):266-76.

46. Lammers T, Aime S, Hennink WE, Storm G, Kiessling F. Theranosticnanomedicine. Acc Chem Res 2011;44(10):1029-38.

47. Bertrand N, Wu J, Xu X, Kamaly N, Farokhzad OC. CancerNanotechnology: The impact of passive and active targeting in the eraof modern cancer biology( ). Adv Drug Deliv Rev 2014;66:2-25.

48. Cheung N-KV, Dyer MA. Neuroblastoma: developmental biology, cancergenomics, and immunotherapy. Nat Rev Cancer 2013;13(6):397-411.

49. Bazak R, Houri M, El Achy S, Kamel S, Refaat T. Cancer activetargeting by nanoparticles: a comprehensive review of literature. JCancer Res Clin Oncol 2015;141(5):769-84.

50. Ferber S, TiramG, Sousa-Herves A, Eldar-BoockA,KrivitskyA, ScomparinA, et al. Co-targeting the tumor endothelium and P-selectin-expressingglioblastoma cells leads to a remarkable therapeutic outcome, eLife 6; 2017.

51. Miller K, Clementi C, Polyak D, Eldar-Boock A, Benayoun L, BarshackI, et al. Poly(ethylene glycol)–paclitaxel–alendronate self-assembledmicelles for the targeted treatment of breast cancer bone metastases.Biomaterials 2013;34(15):3795-806.

52. Scomparin A, Salmaso S, Eldar-Boock A, Ben-Shushan D, Ferber S,Tiram G, et al. A comparative study of folate receptor-targeteddoxorubicin delivery systems: dosing regimens and therapeutic index.J Control Release 2015;208:106-20.

53. Etrych T, Šírová M, Starovoytova L, Říhová B, Ulbrich K. HPMACopolymer Conjugates of Paclitaxel and Docetaxel with pH-ControlledDrug Release. Mol Pharm 2010;7(4):1015-26.

54. Fox ME, Szoka FC, Frechet JMJ. Importance of molecular architecture.Acc Chem Res 2009;42(8):1141-51.

55. Reichert S, Welker P, Calderón M, Khandare J, Mangoldt D, Licha K, etal. Size-dependant cellular uptake of dendritic polyglycerol. Small2011;7(6):820-9.

56. Khandare J, Calderón M, Dagia NM, Haag R. Multifunctional dendriticpolymers in nanomedicine: opportunities and challenges. Chem Soc Rev2012;41(7):2824-48.

57. Zuccari G, Milelli A, Pastorino F, Loi M, Petretto A, Parise A, et al.Tumor vascular targeted liposomal-bortezomib minimizes side effectsand increases therapeutic activity in human neuroblastoma. J ControlRelease 2015;211(Supplement C):44-52.

58. Cossu I, Bottoni G, Loi M, Emionite L, Bartolini A, Di Paolo D, et al.Neuroblastoma-targeted nanocarriers improve drug delivery and pene-tration, delay tumor growth and abrogate metastatic diffusion. Bioma-terials 2015;68(Supplement C):89-99.

59. Duncan R. Polymer therapeutics: Top 10 selling pharmaceuticals —What next? J Control Release 2014;190(Supplement C):371-80.

60. Jain M, Venkatraman G, Batra SK. Optimization of Radioimmunother-apy of Solid Tumors: Biological Impediments and Their Modulation.Clin Cancer Res 2007;13(5):1374-82.