INVESTIGATION OF THE BETA 2 ADRENERGIC RECEPTOR (Β2-AR) PATHWAY IN CANINE HEMANGIOSARCOMA

BY

ROBERTA PORTELA

THESIS

Submitted in partial fulfillment of the requirements

for the degree of Master of Science in VMS – Veterinary Clinical Medicine in the Graduate College of the

University of Illinois at Urbana – Champaign, 2014

Urbana, Illinois

Master’s Committee: Assistant Professor Jackie M. Wypij, Chair

Associate Professor Timothy M. Fan Assistant Professor Stephane Lezmi

  ii  

ABSTRACT

Canine hemangiosarcoma is a highly metastatic cancer arising from vascular endothelial cells. It

is one of the most aggressive canine cancers and most dogs die from this disease within a few

months of the diagnosis. Despite advancements in veterinary oncology, there has been minimal

improvement in the overall survival time even with standard treatment, which includes surgery

and chemotherapy. Propranolol, an oral drug originally developed for the treatment of

cardiovascular diseases, has been successfully used for the treatment of infantile hemangioma

which is a benign neoplasia of vascular endothelial cells. Propranolol blocks adrenergic

receptors, which would otherwise bind to catecholamines responsible for the “stress signal”

leading to many physiologic changes. Stress has been implicated in many models of

carcinogenesis and tumor progression. Given the relationship between stress and cancer, as well

as similarities between canine hemangiosarcoma and infantile hemangioma, we sought to

investigate the presence of the beta 2 adrenergic receptor and the effects of propranolol in canine

hemangiosarcoma. We demonstrated the presence of the beta 2 adrenergic receptor via

immunohistochemistry in all 18 tissue samples of spontaneous canine splenic hemangiosarcoma

and in canine hemangiosarcoma cell lines Fitz and DEN, cell line receptor expression was also

confirmed with Western blot. Src, a possible intermediary downstream protein involved in

adrenergic signaling was also investigated and both Fitz and DEN exhibited the presence of the

Src protein in Western blot. Activation of this pathway would involve phosphorylation of Src

upon catecholamine binding to the receptor, which was investigated through Western blot. Fitz

and DEN exhibited basal phosphorylation of Src and after treatment with norepinephrine, and

Fitz exhibited greater phosphorylation (23% increase compared to basal control) after 60-minute

exposure to the agonist. Fitz cells were pretreated with a biologically achievable dose of

  iii  

propranolol followed by the agonist, and a modest decrease in phosphorylation was observed

(11% decrease compared to basal level). Further investigation into the biological effects of

propranolol in Fitz and DEN revealed a decrease in VEGF secretion, increase in proliferation

and decrease in cell migration. Reduction in VEGF secretion was evaluated via ELISA and it

was present at propranolol doses greater than 10 µM for DEN and Fitz, achieving a maximum

reduction of 21% in DEN and 44% in Fitz compared to untreated cells. Cell proliferation was

measured through MTS assay, which revealed an increase in cell proliferation only in Fitz cells

treated with 0.1 µM of NE (25% increase) and cells treated with 0.1 µM of propranolol (30%

increase). Cell migration was evaluated with a scratch assay and was decreased only when cells

were treated with propranolol at a high dose (100 µM). Taken together, the findings of this study

show that beta 2 adrenergic receptors are expressed by canine hemangiosarcoma, Src may be

involved in the downstream signaling from the receptor and blockade of the receptor leads to

mild to moderate effects in cell angiogenesis, proliferation and migration.

  iv  

For my family,

Luiz, Thais, Rodrigo and Victoria Portela

for never letting me feel alone

and for always believing in me

  v  

ACKNOWLEDGEMENTS

Thank you to my advisor, Dr. Jackie Wypij, for all her help in this project.

My sincere thank you to Dr. Timothy Fan, for the enormous patience, without whom I

could not have accomplished this.

Thank you to Dr. Stephane Lezmi for the support not only with immunohistochemistry

but also with ideas and encouragement that inspired me to keep going.

A huge thank-you to my friend Evelyn Caporali for her incredible help and support when

I needed it most. Thank you so much.

Thanks to Kevin LeBoedec for taking the time to help me with statistics.

Special thank you to Holly Pondenis, for all the technical knowledge shared and

assistance with experiments. Her expertise is invaluable.

Thank you so much to my resident-mates, Sharon Shor, Zach Neumann and Alycen

Lundberg; technicians Jenny Byrd, Rebecca Kamerer and Tara Bailey, and Dr. Laura Garrett

from the Oncology Service for their words of encouragement throughout the years.

My gratitude to James R. Harkness, Wayne D. and Josephine H Spangler Endowment

fund for the financial assistance in this project.

Lastly, thank you to all my friends everywhere in the world, who at some point have

given me instructions, directions, shelter, a shoulder to cry on, and specially have pushed me to

move forward into success. Their kindness will never be forgotten.

  vi  

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION………………………………………………….. ……………. 1

CHAPTER 2: LITERATURE REVIEW………………………………………………………… 3

CHAPTER 3: MATERIALS AND METHODS.………………………………………………. 28

CHAPTER 4: RESULTS……………………………………………………………………….. 38

CHAPTER 5: DISCUSSION AND CONCLUSIONS…………………………………………. 43

CHAPTER 6: FIGURES.………………………………………………………………………. 49

REFERENCES……………………………………………………………………………….….58

  1  

CHAPTER 1

INTRODUCTION

Canine hemangiosarcoma (cHSA) is a malignant neoplasm of endothelial cells that

affects many species, being more frequent in dogs [1-3]. This disease is highly metastatic and the

prognosis is poor, with a median survival time of about 6 months with standard therapy, which

involves surgery and chemotherapy treatment [4]. Limited advancements have been made to

improve survival time therefore novel therapies are needed. Drug repositioning, which uses

known drugs for alternative purposes (treatment of new diseases), is a strategy that can be

pursued with the advantage of being of lower cost (pharmacokinetics and pharmacodynamics are

already known; drugs already in the market can be selected for their cost) and known tolerability.

Recently, propranolol has received attention for its effects on the treatment of infantile

hemangioma (IH), a benign neoplasm of endothelial cells. It was accidentally discovered while

treating infants with hypertension that propranolol (a β-adrenergic receptor antagonist) has

activity against IH, even in treatment-resistant cases [5]. This prompted many studies, mostly

clinical [6-11], looking at the effects of propranolol as an anti-neoplastic drug. Since both cHSA

and IH are tumors of endothelial cell origin, we developed an interest in investigating the effects

of propranolol in cHSA. Parallel to IH and the recent propranolol discovery, there is also a link

between stress and cancer. Stress is physiologically described as the binding of catecholamines

to the adrenergic receptors, which mediate the “fight-or-flight” response necessary for survival

of an organism. An interesting hypothesis is that this adrenergic signaling may also be

functioning to maintain survival and/or progression of neoplastic cells exposed to the same

catecholamines released during stressful situations, which are common during diagnosis and

  2  

treatment of cancer. To further support this hypothesis, studies linking breast cancer survival

and β-blockers have been published suggesting that inhibition of the β-adrenergic signaling

pathway can reduce breast cancer progression, recurrence, metastasis and mortality [12, 13].

Many other studies supporting the role of the β-adrenergic pathway in cancer development have

been described and will be discussed in the next section.

We hypothesized that cHSA expresses the receptors necessary for adrenergic signaling

(β2-AR), an intermediary kinase (Src) is involved in the signaling pathway downstream from the

receptor and propranolol treatment would lead to decrease in cell proliferation, angiogenesis and

migration.

  3  

CHAPTER 2

LITERATURE REVIEW

2.1. Stress and tumorigenesis

The process of tumorigenesis and metastasis has been extensively studied in the scientific

community. Many environmental components have been implicated in cancer development and

new compounds with carcinogenic activity are frequently being discovered. Stress is one

environmental component that may influence the carcinogenic process and potentiate

tumorigenesis. Stress is defined as “A state of mental or emotional strain or tension resulting

from adverse or very demanding circumstances” and “Pressure or tension exerted on a material

object” [14, 15]. In modern life, people are subjected to stress from many different aspects.

Stressful experiences include physical stressors such as pathogens and toxins, and psychological

stressors such as major life events, trauma, abuse, or factors related to the environment in the

home, workplace or family [16].

In 1936, Hans Selye defined stress physiologically as the state in which the

sympathoadrenomedullary system and the limbic-hypothalamic-pituitary-adrenal axis are co-

activated [16, 17]. From a biological standpoint, it is known that stress, whether physical or

mental, culminates in the release of hormones named catecholamines, which mediate the “fight-

or-flight” responses causing activation of the sympathetic nervous system [19]. The most

abundant catecholamines are epinephrine (E), norepinephrine (NE) and dopamine, which are

released from the chromaffin cells of the adrenal medulla and from neuro-muscular junctions

[19, 20]. In humans, acute stress can elevate the levels of norepinephrine and epinephrine by

  4  

greater than 10-fold in only a few seconds, although chronic stress can lead to fluctuations on the

catecholamine levels as well [19, 20].

Cancer patients are subjected to stress from different aspects, from dealing with the

acceptance of a terminal disease (sometimes associated with anxiety and depression) to the

homeostatic effects of stress on inflammation, immunocompetence, and cancer initiation or

progression [19, 21-28].

Evaluating stress in canine patients is difficult, especially the psychosomatic aspects.

Physiologically there have been attempts at quantifying mediators of stress signaling. Increase in

plasma epinephrine and norepinephrine levels in dogs was documented in a study that compared

levels of these catecholamines in groups subjected to environmental stress of blood collection

immediately after arrival to the laboratory versus 15 minutes after arrival, indicating the rapid

fluctuation of these hormones under different stress conditions [29]. Levels of catecholamines in

dogs subjected to various stressful stimuli were measured and in the literature they ranged

around 65 to 239 pg/mL for epinephrine, 145 to 524 pg/mL for norepinephrine and 20-437

pg/mL for dopamine [14, 29]. The release and degradation of catecholamines is variable and thus

the concentration of these hormones can differ significantly across different tissues and within

the circulating blood [2]. In humans, the difference can be in the order of nanomolar for

circulating NE versus micromolar concentrations in the tissue microenviroment [30].

2.2. Adrenergic pathway

Catecholamines mediate adrenergic signaling via receptors in target cells. Receptors

include α1A,B,D; α2A,B,C; β1,β2 and β3 and are present in a wide variety of cells and tissues,

  5  

including central nervous system (CNS), lungs, liver, pancreas, kidney, adrenal glands, breast,

ovary, prostate, vasculature, bone marrow and cells of the immune system (macrophages, mast

cells, lymphocytes) [19, 20]. The expression of β1 and β2 adrenergic receptors (AR) has been

extensively studied in the cardiovascular system, with high expression occurring in cardiac

myocytes and vascular smooth muscle cells [31]. These receptors belong to a family called G-

protein coupled receptors (GPCRs) [32], which contain seven transmembrane “serpentine”

segments with the N-terminal facing the extracellular space and the C-terminal immersed in the

cytoplasm. When the ligand binds to the extracellular domain of the receptor, the receptor

undergoes a conformational change, which will then activate the G-protein located in the

intracellular space. The G-protein is composed of a α, β and γ units, with the α and γ subunits

having lipid tails that covalently attach the protein to the cell membrane. In the absence of ligand

binding, the resting G-protein’s alpha subunit is bound to GDP (guanosine diphosphate). When

the signal is initiated by catecholamine binding, the receptor undergoes a conformational change

and GDP is replaced by GTP (guanosine triphosphate) in the alpha subunit of the G-protein. This

protein detaches from the transmembrane receptor and moves on to activate other downstream

proteins (Fig. 2.1). Inactivation of the protein occurs when phosphatase activity removes a

phosphate group from GTP transforming it back to GDP form [33].

Beta 2 adrenergic receptors (β2-AR) are the focus of this project due to its demonstrated

expression in different human vascular tumors [34]. These receptors activate adenylyl cyclase

which in turn will catalyze the conversion of ATP to cyclic AMP (cAMP). Cyclic AMP can act

as a secondary messenger itself or it can bind to other downstream proteins, such as protein

kinase A [33, 35, 36]. Protein kinase A can phosphorylate other target proteins including cAMP-

responsive element binding protein/activating transcription factor (CREB/ATF) and β-adrenergic

  6  

receptor kinase (BARK) [19]. Upon continuous stimulation, the receptor eventually becomes

inactivated by the inability to bind to G-proteins. That is achieved by phosphorylation of the

intracellular portion of the receptor by BARK, and the phosphorylated protein binds to β-

arrestin, which blocks the connection between the receptor and the G-proteins (Fig. 2.2). β-

arrestin also can function as an adaptor protein, and is this way that it recruits Src to become

bound to the β2-AR [37]. Furthermore, it has also been shown that cAMP/PKA mediates NE-

induced activation of Src [38].

Src is a non-receptor tyrosine kinase ubiquitously expressed in many tissues, and plays

an important role in transmitting signals from the cell surface to the nucleus via phosphorylation

of tyrosine residues of intracellular proteins [39]. Src is involved in the maintenance of normal

cell homeostasis including cell proliferation, survival, maintenance of cytoskeleton, cell adhesion

and motility [38, 39]. Src is overexpressed in many human malignancies, including colorectal,

breast, prostate, pancreas, head and neck carcinoma, lung carcinoma, glioma, melanoma and

sarcomas [38, 40, 41]. A recent study has demonstrated relationship between increased

phosphorylation of Src caused by NE stimulation and tumor growth and progression in ovarian

carcinoma murine models, assessed as number of nodules and tumor weight. Propranolol, a non-

selective β-blocker (described in detail later in this chapter) successfully counteracted the effects

of stress and of adrenergic agonists (isoproterenol and terbutaline) evidenced by decreased tumor

weight and number of nodules in mice, which was similar to control (non-stressed or non-treated

mice) [38]. Src has been also implicated in NE-stimulated vascular endothelial growth factor

(VEGF) production by adipocytes [42, 43]. VEGF is a cytokine known to stimulate angiogenesis

in both normal and neoplastic tissue. In addition, it was demonstrated that Src is present in

canine HSA cell lines and its reduced expression was linked to a decrease in cell viability [44].

  7  

2.3. Adrenergic pathway and molecular pathogenesis of cancer

2.3.1 Breast cancer

The β2-AR and its signaling pathways have been extensively studied in human oncology.

In breast cancer, preclinical models suggest that this pathway may influence breast cancer

progression through: 1) increased tumor cell survival after exposure to chemotherapeutic agents;

2) increased breast cancer cell proliferation; 3) altering the tumor microenvironment in

angiogenesis and inflammatory response [27].

At therapeutic levels, β-blockers do not seem to be directly cytotoxic; rather the β-AR

blockade may inhibit catecholamine signaling that normally would lead to pro-tumorigenic

effects (cell survival, proliferation, migration, angiogenesis) [45]. Therefore, research efforts are

now focused on the ability of β-blockers to inhibit cancer progression in already established

cancers. In one study using an orthotopic mouse model looking at the effect of chronic stress in

breast cancer progression, mice were injected with luciferase-transfected breast cancer cells into

the 4th mammary pad and physically restrained for 2 hours daily for 20 days, while the control

group was not subjected to stress. The results showed a 37-fold increase in metastasis to the

lungs and 67% increased metastasis to the lymph nodes in the stressed group compared to

control [30]. In a separate experiment, mice were pre-treated with propranolol prior to tumor

inoculation and stress stimuli. There was an increased expression of β2-AR in tumors in

propranolol-treated animals and propranolol significantly decreased the metastatic burden in

stressed animals. There was no change in metastasis for non-stressed animals, and treatment did

not affect primary tumor growth in either group [30]. This study provided some objective

evidence of the role of adrenergic signaling and breast cancer progression.

  8  

Catecholamines alter the cytokine profile of the bone microenvironment and promote the

incidence of metastatic colonization by breast cancer cells. Receptor activator of nuclear factor

kappa-B ligand (RANKL) is a protein involved in bone remodeling as well as dendritic cell

maturation and is secreted by osteoblasts in response to sympathetic activation. RANKL

stimulates breast cancer migration and bone colonization and is recognized as a crucial factor for

cancer cell motility in addition to its well-established role in tumor-induced osteolysis [46]. This

pathway was inhibited by propranolol therapy [47]. Activation of the sympathetic system

promoted bone metastasis in a mouse model of breast cancer. Stimulation of the β2-AR induced

RANKL expression in bone marrow osteoblasts and increased migration of metastatic MDA-

MB-231 mammary carcinoma cells in vitro. Mice models of bone metastasis subjected to chronic

stress had sympathetic activation blocked by propranolol. Sympathetic nerves inhibit osteoblast

proliferation and regulate hematopoietic stem cell proliferation, survival and trafficking [48, 49].

Norepinephrine released from the sympathetic nerves stimulate the formation of osteoclasts [50].

Perhaps one of the most commented serendipitous discovery was that women who

took β-blockers (specifically propranolol) for cardiovascular disease were found to be

significantly less likely to die of their cancer, and less likely to present at higher stages of the

disease, compared to women who did not take propranolol [12]. Another study showed a 57%

reduced risk of metastasis and a 71% reduction in breast cancer mortality after 10 years [13].

Further evidence also showed that β-blocker intake was associated with improved relapse-free

survival but not overall survival in patients with breast cancer, including triple-negative tumors

which are tumors negative for epidermal growth factor receptor (HER2), progesterone receptor

(PR) and estrogen receptor (ER) [51]. Relapse is more frequent among triple-negative breast

cancer patients, especially within the first 3 years of diagnosis, compared to patients with ER

  9  

positive tumors [52, 53]. High β2-AR expression was found in tumors containing estrogen

receptors that were hormonally positive, and patients carrying these tumors had a worse

prognosis five years later, coincidentally with the discontinuation of tamoxifen therapy. The

prognosis was good within the first five years, hinting at a possible advantage of combination

therapy with propranolol, which could have an impact in breast cancer survival [54].

Despite some encouraging published evidence on the benefits of propranolol in cancer

patients, there is still some controversy. There are studies suggesting an increased risk or no

survival advantage of various cancers including colon, lung, breast and prostate cancer in

patients taking β-blockers [55-57]. However, patients taking β-blockers for various medical

reasons may have already chronically elevated levels of catecholamines, which would promote

cancer progression and worse outcome in cancer-bearing patients. In that sense, the chronically

elevated NE would be a higher risk than the propranolol itself [58]. A large meta-analysis failed

to demonstrate strong evidence that β2-AR blockers promote tumor progression [59]. In

summary, these studies demonstrate that elevated levels of catecholamines as well as blockade of

the β-adrenergic receptor pathway by propranolol may have some impact in breast cancer

progression as well as patient survival.

2.3.2. Pancreatic carcinoma

In pancreatic carcinoma, β2-AR was expressed in cell lines. NE promoted increased

proliferation of these cells in a dose-dependent fashion, also promoted S-phase cell cycle shift

and decreased cells in the G1/G2 phase. Migration was also increased with NE treatment as well

as increase in phosphorylation of p38 (MAPK pathway, important in cell proliferation), which

was blocked by treatment with propranolol [60]. In another study, invasiveness and angiogenic

  10  

potential of pancreatic carcinoma cell lines were assessed by measurement of matrix

metalloproteinases (MMPs) MMP-2, MMP-9 and VEGF. Norepinephrine promoted the

invasiveness of one of the cell lines in a concentration-dependent manner, and NE increased the

expression of MMP-2, MMP-9, and VEGF. These effects were inhibited by propranolol [61].

Furthermore, propranolol significantly suppressed cell invasion and proliferation in comparison

to β1-adrenergic specific antagonist metoprolol. Treatment with β2-adrenoceptor antagonists

inhibited activation of transcription factors nuclear factor κB (NF-κB), activator protein 1 (AP-1)

and cAMP response element binding protein (CREB). β2-adrenoceptor antagonists also

significantly altered VEGF, cyclooxygenase-2 (COX-2), MMP-2 and MMP-9 expression

confirming previous findings. The β2-adrenergic antagonists suppressed invasion and

proliferation by inhibiting both cAMP/PKA and Ras, which regulate activation of the MAPK

pathway and transcription factors, such as NF-κB, AP-1 and CREB, as well as expression of

their target genes, MMP-9, MMP-2 and VEGF. β1-adrenergic antagonists suppressed invasion

by inhibiting only the cAMP/PKA pathway [62]. Proliferation and migration of human

pancreatic ductal carcinoma is stimulated in vitro by β-AR and subsequent cAMP-dependent

signaling, which in turn leads to cAMP-dependent release of epidermal growth factor, and PKA-

dependent release of VEGF. These effects were also blocked with propranolol treatment [63].

Another study using isoproterenol (a non-selective β-AR agonist) significantly increased cell

proliferation of a pancreatic ductal carcinoma cell line in a dose-dependent manner, with

concomitant activation of ERK/MAPK signal pathway as well as increased levels of

phosphorylated ERK [64, 65]. It has been previously demonstrated that β2-AR causes activation

of the extracellular-signal regulated kinase 1/2 (ERK 1/2).

  11  

In vivo, mice carrying xenograft tumors had enhanced tumor growth caused by

isoproterenol, which was suppressed by propranolol [64]. Taken together, these studies suggest

that blocking the adrenergic pathway using propranolol may have some impact in pancreatic

carcinoma progression.

2.3.3. Ovarian carcinoma

Immunohistochemistry demonstrated stress-induced increases in levels of basic fibroblast

growth factor (bFGF), MMP-2 and MMP-9 proteins in ovarian carcinoma. Catecholamines have

previously shown to promote VEGF production by ovarian cancer cells in vitro [66, 67].

Chronic behavioral stress resulted in higher levels of tissue catecholamines, greater tumor

burden and more invasive pattern of ovarian cancer growth in a mouse model; effects which

were mediated by β2-AR activation of PKA signaling in the cancer cells [68]. Another

interesting study by Thacker et al used periodic physical restraint to stimulate chronic stress in

nude mice inoculated with human ovarian carcinoma cells in the peritoneal cavity. In mice

receiving 0, 2 or 6 hours of immobilization daily for 21 days, the number of tumor nodules

increased by 259% in the 2 hr stress group and 356% in the 6 hr stress group compared to

unstressed mice. Mean tumor weight increased 242% in the 2 hr stress group and 275% in the 6

hr group compared to unstressed mice. Social isolation as an alternative stressor also showed a

187% increase in tumor weight and a 255% increase in nodule count compared with group-

housed control animals. Stressed animals and non-stressed mice treated with isoproterenol and

terbutaline (β-agonists) had both significantly enhanced mean vessel density (MVD) counts (a

measure of angiogenesis [69]) and that effect was blocked by propranolol. This was

accompanied by significant elevation of VEGF protein and mRNA within tumor tissue.

  12  

In another model investigating the effects of surgical stress in tumor development, mice

were injected with ovarian carcinoma cells and 4 days later subjected to either a laparotomy

procedure (larger incision, more tissue handling and stress) or a laparoscopy procedure (less

invasive, less stress). Mice in the laparoscopy group had significantly lower tumor weight

compared to mice in the laparotomy group. In the immediate postoperative period, serum levels

of VEGF and MMP-2 were significantly lower in the laparoscopy group as well, further

supporting the mediators of stress-induced tumor progression [70]. Lee et al. had also

demonstrated previously that surgery significantly increased MVD and VEGF expression, which

were blocked by propranolol treatment in mice injected with ovarian carcinoma cell lines [71].

The stress of surgery is suggested to facilitate the post-surgery growth of pre-existing

micrometastases and small residual tumors [72-74]. Taken together, these studies indicate that

psychosomatic and surgical stress, both present in cancer patients, are linked to tumor growth

and angiogenesis and propranolol appears to ameliorate these effects.

2.3.4. Prostatic carcinoma

Many studies have also been performed investigating the role of β2-AR in prostatic

carcinoma. Among the studies investigating possible molecular pathways involved in β blockade

and prostatic carcinoma, there appears to be a relationship between β2-AR and histamine, as

histamine augmented β2-AR-induced cAMP accumulation independently of known histamine

receptors [75]. β2-AR activation also promoted prostate cancer cell proliferation and cell

migration through increasing cellular cAMP and ERK 1/2 activation in another study, where it

also demonstrated the involvement of β-arrestin in this process (β-arrestin participates in agonist-

mediated desensitization of G protein-coupled receptors). The formation of β-arrestin 2/c-Src

  13  

complex was a key factor in this process [76]. An experiment was performed where PC-3

xenografted mice were implanted with a norepinephrine-releasing micropump which

subsequently led to a 1.6 fold increase in tumor metastasis compared to control [77]. In addition,

behavioral stress activated β2-AR signaling and led to inhibition of apoptosis and accelerated

cancer development in mice. The effects of stress were prevented with treatment using a β2-AR

selective antagonist [78].

β2-AR is a known activator of the androgen receptor, and this receptor is upregulated in

androgen-independent prostatic carcinoma cell lines. Expression of this receptor is greater in

malignant cells compared to benign hyperplasia or normal tissue [79]. One study showed that NE

drives metastasis of PC-3 cells in BALB/C3 nude mice and this process is inhibited by β-

blockers [77]. Loss of BARK-1 in a group of patients with high-grade prostatic carcinoma could

indicate a possible pathway of prostate cancer development involving β2-AR signaling [80].

BARK-1 is a serine/threonine kinase that desensitizes the receptors from catecholamine over-

stimulation (negative feedback via recruitment of beta-arrestin) [81]. In another study, β-blocker

use was not associated with increased risk of prostatic carcinoma development or overall

mortality, however a subgroup of men treated with androgen deprivation therapy and a β-blocker

had a fivefold reduced prostatic carcinoma specific mortality [82].

In summary, these studies indicate that prostatic carcinoma express the β2-AR, β2-AR

activation promote prostate cancer cell proliferation and cell migration, mice exposed to

increased levels of catecholamines have increased metastasis and men treated with β-blockers

may have reduced cancer-related mortality.

2.3.5. Other malignancies

  14  

Stress has been suggested as a possible factor in human colorectal carcinoma

development. In one study looking at the effects of chronic restraint stress in nude mice bearing

human colorectal carcinoma xenografts, adrenergic signaling-dependent activation of ERK 1/2

promoted cell proliferation, and β-adrenergic antagonism inhibited proliferation and decreased

phosphorylation of ERK 1/2 in vitro and in vivo. Norepinephrine and epinephrine enhanced

colorectal carcinoma cell proliferation and viability in cell culture as well as tumor growth in

mice. These effects were antagonized by propranolol and phentolamine (α-AR antagonist) [83].

The first epidemiological investigation of the effect of post-diagnostic β-blocker usage on

colorectal cancer-specific mortality showed no association. There was some evidence of a weak

reduction in all-cause mortality in β-blocker users which was in part due to the marked effect of

atenolol on cardiovascular mortality [84].

Melanoma cell lines also expressed β1 and β2-ARs. Norepinephrine and E increased

metalloprotease-dependent motility, released interleukin-6 (IL-6), interleukin-8 (IL-8) and

VEGF. The effects of these catecholamines were inhibited by propranolol [85]. Interleukin-6 is

involved in the host immune defense mechanism as well as the modulation of growth and

differentiation in various malignancies [86]. Increased expression of IL-8 and/or its receptors has

been characterized in cancer cells, endothelial cells, infiltrating neutrophils, and tumor-

associated macrophages, suggesting that IL-8 may function as a significant regulatory factor

within the tumor microenvironment [87]. Expression of β-ARs had been previously identified in

melanoma cell lines as well as human melanoma biopsies and NE upregulated production of

VEGF, IL-8, and IL-6 in the cell lines [88]. Overall, the results from preclinical studies support

the suggestion that β-blockers could provide a clinical benefit in melanoma progression however

epidemiological studies are limited by sample size, which is also a limitation in breast, colorectal

  15  

and prostatic carcinoma studies [89]. In patients with melanoma, the β-blocker-treated group had

an overall improved survival after a median follow-up of four years and for each year of β-

blocker use, the risk of death was reduced by 38% [90]. Another study found that both β1 and

β2-ARs are expressed in tissues from benign melanocytic nevi, atypical nevi and malignant

melanomas and that expression was significantly higher in malignant tumors.

In an in vitro multiple myeloma study, propranolol IC50 values (concentration of

propranolol required for 50% inhibition of proliferation) were decreased over time. There were

also significant increases in caspase 3 activity, in apoptotic cell population, and a decrease in

expression levels of Bcl-2 (anti-apoptotic protein) in response to propranolol treatment [91].

Caspase-3 is an executioner caspase protein, involved in the process of apoptosis (programmed

cell death). Another study showed NE-induced secretion of VEGF in 3 multiple myeloma cell

lines [92].

In conclusion, although the relationship of stress and tumor development/progression has

been more extensively studied in a number of specific carcinomas (breast, ovarian, prostatic,

colorectal, pancreatic), it appears to influence non-carcinoma malignancies as well (melanoma,

myeloma). This illustrates how broad this research field is and indicates that propranolol may

have the potential to be of benefit in the treatment of many other tumors.

2.4. Propranolol

Propranolol is an attractive potential anti-neoplastic therapeutic option since it is

affordable, easily obtainable in the market and the pharmacokinetics have been established for

humans [93] and dogs [94]. It is a non-selective β-blocker, available as propranolol

  16  

hydrochloride, under the brand name Inderal® in North America by AstraZeneca®. It has been

mainly used for the treatment of hypertension, although other conditions reported include

tachyarrhythmias [95], tachycardia, essential tremor [96, 97], migraine [98, 99], post-traumatic

stress disorder [100-104], glaucoma [105], akathisia (restlessness) [106], schizophrenia (elevated

levels of norepinephrine may play a causative role in the development of this disease) [107],

adrenergic urticaria [108], treatment of burn patients [109-112] and anxiety [113, 114].

In human patients, receiving the diagnosis of cancer creates significant levels of

emotional distress, with intrusive thoughts (emotional memories, flashbacks, nightmares, and

intrusive images) being the most common manifestation among breast cancer survivors. Recently

diagnosed female breast and colorectal cancer patients using β-blockers reported less cancer-

related psychological distress [115]. Although direct psychological stress may not be an

important feature in canine cancer, propranolol could potentially help to control the anxiety

associated with repeated hospital visits required for treatment.

Side effects of propranolol reported in humans most commonly include hypotension,

hypoglycemia, bradycardia, bronchospasm, sleep disturbances (nightmares), acrocyanosis

(cyanotic extremities). Other fewer reported side effects include gastroesophageal reflux, nausea,

vomiting, diarrhea, somnolence, hyperkalemia, tumor lysis syndrome, psoriatic drug rash,

respiratory syncytial virus exacerbation, and dental caries [116-119]. Propranolol is

contraindicated in patients with asthma, heart block and sinus bradycardia. The most serious side

effect of propranolol is hypoglycemia [120, 121]. Propranolol is thought to cause hypoglycemia

by inhibiting glycogenolysis, glyconeogenesis, and lipolysis. Children have lower glycogen

stores and higher glucose consumption rates when fasting and, therefore, are more susceptible to

  17  

hypoglycemia than adults [122]. Side effect incidence, however, does not appear high. A recent

retrospective study revealed that only 7.6% (10/132) of the children who received propranolol

treatment had to discontinue it due to side effects. No adverse effects requiring hospitalization

were recorded. [123].

In dogs, propranolol is indicated for the treatment of hypertension, atrial fibrillation,

tachyarrhythmias [124, 125], myocarditis [126], ventricular premature contractions and

arrhythmias caused by digitalis toxicity [127]. It is also useful in the treatment of hypertrophic

cardiomyopathy especially when associated with hyperthyroid disease in cats (as it inhibits

conversion of thyroxine to triiodothyronine) [128] and for urinary retention (stimulates bladder

contraction) [129]. It is contraindicated in patients with congestive heart failure unless it is

secondary to a tachyarrhythmia responsive to beta-blockade. The drug is also contraindicated in

patients with 2nd or 3rd degree heart block, sinus bradycardia, asthma and thromboembolic

disease. It should be used with caution in diabetic patients due to risk of hypoglycemia as

described in humans. Side effects are similar to humans and include hypotension, bradycardia,

hypoglycemia, decreased cardiac contractility, bronchoconstriction, peripheral vasoconstriction

and diarrhea. It is recommended that propranolol therapy be gradually withdrawn due to possible

“rebound” effect of catecholamines to the chronically suppressed β2-AR, leading to tachycardia,

arrhythmias and hypertension.

Possible drug interactions with propranolol include [130]:

• Delayed gastrointestinal absorption of propranolol by antacids

• Additive toxic effects when used concomitantly with quinidine, procainamide and

lidocaine (although anti-arrhythmic effects are enhanced)

  18  

• Hypotensive effects are enhanced by chlorpromazine, cimetidine, furosemide,

phenothiazines and hydralazine

• Increased serum levels of lidocaine

• Increased effects of tubocurarine and succinylcholine

• Increased action of terbutaline

• Antagonizing effects on epinephrine and phenylpropanolamine

• Antagonizing bronchodilatory effects of theophylline

• Bradycardia may be potentiated by concurrent use of digitalis

• Anti-hypertensive effects of propranolol can be inhibited by concurrent use of

salicylates

• Propranolol’s effects can be decreased by concurrent use of thyroid hormone

supplementation and the dose of propranolol may require a decrease in animals

receiving methimazole

A pharmacokinetic study of propranolol in dogs used a racemic mixture as prepared for

pharmaceutical use. A half-life of 1.09 +/- 0.33 hours was observed, which was not different than

what was previously reported for a levo isomer given as an IV bolus, however the distribution

volumes were significantly greater. A mean value of 6.5 L/Kg was observed with a range from

3.44 to 10.47 L/Kg. Total body clearance was calculated to average 68 mL/min/Kg. Both hepatic

and extra-hepatic clearance of propranolol was suggested, and approximately 90% of propranolol

is extracted from the blood as it passes through the liver. This extensive extraction results in poor

bioavailability, with only 2-17% of the dose reaching the systemic circulation unchanged.

Systemic availability increased substantially with multiple dosing (2.1-5.5 mg/Kg every 6 hours),

reaching up to 10.7 times greater than predicted by single dose. Propranolol is rapidly absorbed

  19  

with an absorption half-life of 17 minutes. Mean CpMax measured in 5 dogs receiving 40 mg

(2.1-5.5 mg/Kg) in multiple oral doses was 79.54 ng/mL (range 21-162.5), which is equivalent to

a mean of 0.3 µM (range 0.08-0.62 µM) [127]. Another recent study with single administration

of three 40 mg tablets revealed a Cmax of 191.4 +/- 56.52 ng/mL (0.73 +/- 0.21 µM) [131].

Propranolol is available in tablets containing 10, 20, 40, 60, 80, 90 mg and oral solution

containing 4 and 8 mg/m. Injectable propranolol is available in 1 mg/mL concentration. Standard

dosage in dogs: 0.1-0.2 mg/Kg orally every 8 hours (maximum 1.5 mg/Kg), intravenously 0.02

mg/Kg over 5-10 minutes (maximum 1 mg/Kg) [132, 133].

2.5. Propranolol and infantile hemangioma

Fairly recently (2008), propranolol was serendipitously discovered to be effective for the

treatment of infantile hemangioma after two children who were treated with propranolol for

cardiomyopathy were observed to have their hemangioma regress [5]. Based on this observation,

the authors decided to treat nine additional children with propranolol and observed similar

results. Prior to this discovery, hemangiomas were treated with corticosteroids as a first line of

therapy, with other options including interferon alpha, vincristine or surgical removal if the

hemangiomas were progressing despite high dose of corticosteroids [134, 135]. Propranolol thus

came as an effective, inexpensive option for treatment of complicated hemangiomas.

Infantile hemangiomas (IH) are the most common tumors of early childhood, affecting

about 5-10% of infants. It is three times more common in females than males, and most prevalent

in Caucasian children [136]. In general is not considered a severe disease, since most IH will

spontaneously regress necessitating no additional treatment. However, approximately 12% of IH

  20  

are complicated cases in which the location or the speed of growth can promote significant

morbidity including disfigurement, ulceration, bleeding, visual compromise, airway obstruction,

congestive heart failure and rarely death [137]. Most IH undergo rapid proliferation during the

first months to year of life, reaching an average size of 2-20 cm. These lesions then undergo a

slow involution period over several years and are generally fully regressed by 5-10 years of age

however the duration and rate of growth are variable [136-138].

Histologically these tumors are composed of a mixture of clonal endothelial cells

associated with pericytes, dendritic cells and mast cells [139]. They are densely packed over-

proliferating capillaries with the absence of an open lumen. The origin is still unclear with some

studies suggesting aberrant transplantation of placental endothelial cells [140], predisposing

genetic factors [141, 142] and/or tumor stem cell components [136, 143].

The molecular mechanisms behind the development of infantile hemangioma are still

under investigation. In the early phase of the disease development, angiogenic and growth

factors can contribute for tumor progression; in the late and involuting phase of the disease it is

thought that apoptosis is involved. The proposed mechanisms of propranolol effects include

inhibition of angiogenesis; early vasoconstriction of the tumor (attributed to the blocking of

nitric oxide production) which causes the softening of the tumor as well as change in color from

red to purple and later promotion of tumor apoptosis [5].

During the growth or proliferative phase, two main pro-angiogenic cytokines have been

implicated: VEGF and bFGF [139, 144-146]. Studies have shown that both endothelial and

interstitial cells are actively dividing in that phase. Propranolol has been implicated in

suppression of VEGF protein expression in hemangioma-derived endothelial cells in a dose-

  21  

dependent manner (25-100 µM) and induced apoptosis by activation of the caspase cascade

(caspase 3 and caspase 9) after 100 µM treatment of propranolol [147]. Serial serum VEGF

decrease was also demonstrated in clinical patients after initiating propranolol treatment. This

same study demonstrated a decrease in serum MMP-9 as well, suggesting another mechanism of

tumor control by propranolol [148]. Another study demonstrated a decrease in the VEGFR

(VEGF receptor) protein expression after 48 and 96 hours of cell treatment with higher

concentrations of propranolol (200-300 µM). Interestingly, there was upregulation of mRNA

expression but downregulation of VEGF by propranolol, indicating an inhibition in mRNA

translation into VEGF protein. This inhibition in VEGF was mediated by hypoxia-inducible

factor 1-alpha (HIF-1α), which is a transcription factor that becomes stabilized in hypoxic

situations. In a normoxic environment, HIF-1α is rapidly degraded; in hypoxia it persists as a

transcription factor for many genes including VEGF, ultimately contributing to angiogenesis.

Another finding was decrease in phosphorylation of PI3/Akt and p38/MAPK in a dose-

dependent fashion [149]. In another study, there was an increase in phosphorylation of p38 in

propranolol-treated IH. Since p38 regulates the production of inflammatory mediators such as

TNFα, IL1β and COX-2 [150], one proposed mechanism of IH remission could be due to

immune-mediated responses [136]. Propranolol treatment also caused decreased levels of

phosphorylated cofilin. Cofilin is a cytoskeletal-binding protein critical for actin microfilament

dynamics and reorganization that severs and depolymerizes actin filaments [151]. Cofilin

phosphorylation could lead to increased cofilin-mediated actin severing, which would disrupt

cell migration and affect cell proliferation because the actin cytoskeleton is intimately related to

regulation of the cell cycle progression [136, 152]. Propranolol also disrupted the cell cycle by

decreasing the expression of key cyclin proteins (cyclin A1, A2, B2, D2 and D3) and increased

  22  

the expression of important cell cycle inhibitors (p15, p21, p27). No change was noted on

expression of cyclin-dependent kinases Cdk2 and Cdk4 [136]. Propranolol inhibited cell

proliferation at an IC50 of 50 µM and induced an increase in the proportion of cells in the G1

phase while reducing the proportion of cells in S and G2/M phase [136]. Propranolol did not

induce apoptosis of endothelial cells at 50 µM, nor promoted cleavage of caspase 3 and caspase

9, in contrast with pro-apoptotic effects stated by other studies described in this text. Cell

migration (assessed by the scratch/wound healing assay) was significantly inhibited after

endothelial cells were treated with propranolol at 50 µM for 12 hours [136].

HIF-1α serum levels were found elevated in children with proliferative hemangiomas

[153, 154]. HIF-1α protein was decreased in a dose-dependent manner after treatment with

propranolol, which was evaluated both by Western blot and ELISA. Another possible pro-

angiogenic pathway, NF-κB (nuclear factor kappa-β), was analyzed and was upregulated by

treatment with propranolol, indicating a possible compensatory angiogenic mechanism by the

tumor [149]. NF-κB constitutes a non- HIF-1α dependent pathway resulting in VEGF-A

expression and angiogenesis [155]. Interestingly, the same experiments looking at cell viability,

migration, tubule formation, RT-PCR, flow cytometry, western blots and ELISA were performed

in hemangioma endothelial cells and repeated in human umbilical vein endothelial cells

(HUVEC) used as control cells, with the downregulation of VEGF, VEGFR and HIF-1α not

observed in control cells, indicating that the effects of propranolol were confined to the

hemangioma endothelial cells only. Propranolol was also observed to inhibit cell migration and

tubule formation. Apoptosis of hemangioma endothelial cells, but not hemangioma stem cells,

was demonstrated in another study [149]. This could explain the rebound growth of

hemangiomas after propranolol therapy has been discontinued [156-158]. Propranolol also

  23  

promoted adipogenic differentiation of hemangioma stem cells (involution) [159]. In another

study propranolol did not promote apoptosis of hemangioma stem cells at a concentration of 50

µM; the half maximal inhibition was 133 µM, which is well above the biologically achievable

dose in humans. Inhibition of proliferation was also not achieved in this study. High levels of

VEGF and bFGF were observed in proliferating IH tissue. At low concentrations (0-20 µM),

propranolol was capable of decreasing the VEGF mRNA and VEGF protein levels in

hemangioma-derived stem cells (HemSCs). There was downregulation of bFGF by HemSCs

however it was less pronounced than VEGF. The proportion of VEGF-positive HemSCs was

very low in IH tissue [160]. Zou et al showed that propranolol did not affect proliferation of

endothelial progenitor cells, but it did inhibit migration of these cells in a dose-dependent (0-100

µM) and time-dependent (24-72 hr) manner. This same study investigated the effects of

propranolol in the expression of CXCR4, which was suppressed via Akt and MAPK pathways

[161]. CXCR4 is a receptor for stromal cell derived factor 1α, which promotes mobilization of

the endothelial progenitor cells from the bone marrow to the site of vasculogenesis [154].

In summary, while there is still active research into the molecular pathways involved in

the response of infantile hemangioma to propranolol, most studies suggest that the effects of

propranolol are related to suppression of angiogenesis. Some studies also indicate an effect in

endothelial cell proliferation and survival.

2.6. Canine Hemangiosarcoma

Hemangiosarcoma (HSA) is a malignant neoplasm of vascular endothelial origin. This

cancer occurs more frequently in dogs than in any other species [2, 3]. It accounts for 2.3 to 3.6%

of skin tumors in dogs and 45 to 51% of splenic malignancies [2, 3, 162-164]. Like majority of

  24  

cancers, it affects mostly middle-aged to older animals [163, 165-168]. Breeds that appear

overrepresented include German shepherds, Golden retrievers and Labrador retrievers [163, 165,

168-170]. The most common primary site for HSA is the spleen, with other frequent sites include

the right atrium, skin, subcutis and liver [1, 2, 165, 166, 168, 169, 171-174].

Being a vascular malignancy, molecular pathways involving angiogenesis have been

explored to further clarify the development of this disease. VEGF expression has been

demonstrated in canine studies [175-177] and plasma levels of VEGF were higher in dogs with

HSA compared to healthy dogs [178]. There was no marked difference between VEGF levels in

effusions associated with malignant versus nonmalignant diseases [179]. Other angiogenic

cytokines important in HSA include bFGF and angiopoietin 1 (Ang-1) [177, 180]. These

cytokines can be secreted by the tumor cells leading to an autocrine growth signaling or can be

secreted by other cells within the microenviroment leading to paracrine stimulation. Another

study showed a significantly higher proportion of platelet derived growth factor (PDGFR-β)

expression in HSAs compared to cutaneous hemangiomas [181].

Mutations in tumor suppressor genes could provide a possible etiology for the disease,

however studies suggested that p53 (tumor suppressor gene) and Ras (oncogene) mutations are

infrequent in canine HSA [182-184]. PTEN (tumor suppressor gene) inactivation was

demonstrated in greater than 50% of evaluated canine HSA samples in one study [185]. Key

growth and apoptosis regulating proteins such as pRB, cyclin D1, Bcl-2 and survivin appear

overexpressed in HSA when compared with hemangiomas or normal tissues [183, 186].

Histopathology is often necessary to establish a diagnosis, since cytology is non-

diagnostic in the majority of time due to excessive bleeding and lack of exfoliation by the tumor

  25  

[187]. Histologically HSA consists of immature, pleomorphic endothelial cells forming vascular

spaces containing variable amounts of blood and/or thrombi [188]. Immunohistochemistry for

factor VIII (Von Willebrand) or CD31 can be used to confirm endothelial origin and support the

diagnosis [188]. Claudin-5 and CD117 have also been identified as potentially useful markers.

(Kit) [176, 189].

Surgery is the main method of treatment for HSA. Adjuvant chemotherapy is indicated in

most cases (except for dermal hemangiosarcoma) [190]. Protocols involving doxorubicin are

most commonly used, including VAC (vincristine, doxorubicin and cyclophosphamide) [191,

192]; vincristine, cyclophosphamide and methotrexate [168]; doxorubicin and

cyclophosphamide; doxorubicin and minocycline [193] and single-agent doxorubicin [194-197].

Ifosfamide has also been used [198, 199]. Epirubicin and intracavitary pegylated liposomal

encapsulated doxorubicin do not appear to have advantage over conventional doxorubicin [200,

201]. Immunotherapy has been investigated as possible treatment strategy, with one study using

a mixed killed bacterial vaccine following surgery showing some improvement in survival time

[168], and another study looking at doxorubicin/cyclophosphamide combination with liposome-

encapsulated muramyl-tripeptide-phosphatidylethanolamine (L-MTP-PE) which showed a

significant increase in median survival time (5.7 to 9 months) with 40% of the dogs experiencing

long term survival [202]. Radiation therapy is mainly used for non-visceral hemangiosarcoma

and although it can cause reduction in tumor size it does not significantly change the overall

survival [203]. A small study looking at a metronomic chemotherapy protocol using NSAID,

cyclophosphamide and etoposide showed a similar outcome to doxorubicin-based protocols

[204]. A combination of a dose-intensified doxorubicin protocol with deracoxib (COX2

inhibitor) was well tolerated but did not result in overall improvement of survival in dogs with

  26  

splenic HSA [205]. A study investigating the effects of toceranib phosphate (Palladia®), a

multiple kinase inhibitor (including VEGFR), administered after splenectomy and 5 doses of

doxorubicin given at 2-week intervals has been conducted and the results showed that the

administration of toceranib did not significantly improve disease-free interval or overall survival

time [206].

The prognosis for canine HSA is very poor. Hemangiosarcoma typically has a very

aggressive biologic behavior, with disseminated metastasis occurring very early in the

development of the disease, except for dermal hemangiosarcomas. Metastasis is typically

hematogenous or through transabdominal implantation following rupture. The most frequent

metastatic sites are the liver, omentum, mesentery and lungs. Other reported metastatic sites

include kidney, muscle, peritoneum, lymph nodes, adrenal gland, brain and diaphragm. Dogs

treated with splenectomy alone have a survival time that range from 19-86 days and less than

10% of dogs living one year [3, 164, 168, 207, 208]. Surgery plus adjuvant chemotherapy will

increase median survival times to 141-179 days, however the one year survival rate is still less

than 10% [191, 194, 195, 198, 200]. The average survival time of right atrial HSA undergoing

surgery is 1-4 months [209, 210]. Dermal HSA treated with surgery alone had a median overall

survival time of 987 days and lingual HSA had a median overall survival time of 553  days [211,

212]. One analysis of intramuscular and subcutaneous HSA identified 71 cases, the median time

to tumor progression and overall survival time were 116 and 172 days, respectively and 25%

survived to 1 year [213]. In summary, the literature on canine HSA describes an aggressive

cancer. Unfortunately, there has not been substantial improvement in the outcome of dogs with

HSA and therefore novel therapies are needed.

  27  

Investigations into the beta adrenergic presence in canine hemangiosarcoma as well as

the effects of its blockade on cell activity are starting to emerge. One study that evaluated canine

hemangiosarcoma cell lines as well as mouse angiosarcoma, hemangioendothelioma and human

dermal microvascular endothelial cell lines revealed that propranolol selectively inhibited

proliferation, survival, and migration of a panel of malignant vascular tumor cells, indicating that

the oncogenic properties of these tumor types are driven, in part, by beta adrenergic signaling.

Propranolol dramatically slowed the proliferation rate of all vascular tumor lines tested, and four

of the five tumor lines exhibited nearly 100% lethality at doses that had been previously reported

to be non-toxic for primary cultures of human endothelial cells. This finding suggested that

malignant endothelial tumors may be more sensitive to beta blockade [214].

  28  

CHAPTER 3

MATERIALS AND METHODS

3.1. Cells and reagents

Two cHSA cell lines (DEN and Fitz) were provided by Dr. Douglas Thamm (Colorado

State University). Additional cell lines used as controls included HeLa (human cervical

adenocarcinoma), MDA-MB-231 (human metastatic breast carcinoma), and MDCK (Madin-

Darby canine kidney) purchased from American Type Culture Collection (ATCC). Cells were

cultured at 37˚C in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with glutamine

(2 mmol/L), penicillin (100 IU/ml), streptomycin (100 IU/ml), and 10% fetal bovine serum

(FBS) in a humidified atmosphere supplemented with 5% CO2. Cell cultures were maintained in

subconfluent monolayers and passaged as deemed necessary by growth of cell line.

Propranolol hydrochloride (P) was purchased from Sigma-Aldrich (St. Louis, MO -

catalog # P0884). Norepinephrine bitartrate salt monohydrate was purchased from Sigma-

Aldrich (St. Louis, MO - catalog # A9512). Both reagents were diluted with deionized water

based on manufacturer’s instructions and solubility provided.

The anti-β2-AR antibodies (rabbit polyclonal; catalog #AB36956 and #AB135641) used

for Western blot (WB) and immunohistochemistry (IHC) experiments, respectively, were

purchased from Abcam Inc (Cambridge, MA). The anti-Src (mouse monoclonal; catalog #2110)

and anti-phospho-Src (Tyr 416) (rabbit monoclonal; catalog #2113 and rabbit monoclonal;

catalog #6943) were purchased from Cell Signaling Technology (Beverly, MA). Antibody anti-

tyrosine hydroxylase (rabbit polyclonal, Abcam #AB112) and anti-CD31 (mouse monoclonal,

  29  

Dako, North America Inc., #M0823) were used in splenic cHSA tissue samples. The anti-β actin

antibody used for loading control (rabbit polyclonal; catalog #AB8227) was purchased from

Abcam Inc. Horseradish peroxidase (HRP) conjugated anti-mouse and anti-rabbit secondary

antibodies used for WB were purchased from GE Healthcare, UK. Blocking peptide for the β2-

AR WB (catalog #38102) was purchased from Abcam Inc.

Other reagents used include WB developing solution (SuperSignal West Femto

Chemiluminescent Substrate, Thermo Scientific, Rockford, IL), bovine serum albumin (BSA),

milk, tris-buffered saline and Tween 20 (TBST) and hydrogen peroxide (H2O2).

3.2. Immunohistochemistry

3.2.1 Spontaneous canine HSA samples

Eighteen canine splenic HSA tissue blocks were obtained from the Veterinary Diagnostic

Laboratory of the University of Illinois at Urbana-Champaign. Paraffin-embedded cell pellets

were sectioned every 3 microns, placed on positively charged slides, and dried for one hour at

60°C. Slides were deparaffinized with two sequential xylene washes, and subsequently hydrated

using two ethanol baths for five minutes each; followed by a water rinse. Slides were placed in

3% H2O2 for 15 minutes and rinsed with water. Antigen retrieval was performed with citrate

buffer microwaved for five minutes. Nonspecific staining was minimized using a blocking

solution for 30 minutes (5% BSA in PBST). Slides were then incubated with rabbit polyclonal

anti-β2-AR antibody at a concentration of 1:35 for 30 minutes at 37°C followed by one hour at

room temperature. Slides were rinsed twice with PBST and then incubated with a biotinylated

secondary anti-rabbit antibody for 30 minutes at room temperature. Slides were washed in PBST

  30  

twice before incubation for 30 minutes with a streptavidin–biotinylated HRP complex, and

developed with 3,3-diaminobenzidine (DAB) (BioGenex, Freemont, CA). Slides were then

counterstained with Mayer’s hematoxylin. The negative control samples underwent the same

process but in the absence of the primary antibody. The isotypic negative control was performed

using tyrosine hydroxylase antibody (1:200), which is very specific for neuronal tissue. This was

performed to rule out non-specific Fc receptor binding or other cellular protein interactions, as

well as confirm the specificity of the primary antibody. The rabbit polyclonal anti-β2-AR

antibody was validated using canine heart tissue as a positive control as well as mouse spleen

and kidney [215-217]. All cHSA tissue samples were reviewed by a single pathologist (S.

Lezmi) and the diagnosis of hemangiosarcoma was confirmed with H&E staining and with IHC

using a CD31 antibody. CD31 antigen is a marker for endothelial cells and is expressed only by

endothelial cells, platelets and macrophages; it is conserved in neoplasia arising from endothelial

cells and validated in canine tissue [218]. The intensity of the β2-AR staining was evaluated

(scored as strong, mild or none) as well as the localization within the cell (nuclear vs.

cytoplasmic).

3.2.2. Cell pellets

Fitz and DEN cell lines and a human positive control cell line, MDA-MB-231 [219],

were used to evaluate β2-AR expression using IHC. Adherent cell cultures were collected and

washed in PBS, then pelleted by centrifugation (10,000 RPM for 10 minutes). Each pellet was

re-suspended in 1 ml 10% formalin for one hour. Formalin was removed and cell pellets were

re-suspended uniformly into 1 ml 4% melted agarose gel by vortexing, then immediately

centrifuged to create an agarose-embedded cell pellet. Cell pellets were then trimmed and

  31  

processed as for formalin-fixed paraffin-embedded tissue biopsy specimens. Briefly, paraffin-

embedded cell pellets were sectioned at three microns, placed on positively charged slides, and

dried for one hour at 60°C. Slides were deparaffinized with three sequential xylene washes, and

subsequently hydrated using 100%, 95%, and 70% ethanol for two minutes each; followed by a

water rinse. Slides were placed in 3% H2O2 in methanol for 15 minutes, and then treated with

citrate buffer (pH 6.0), placed in a decloaking chamber until the temperature reached 125 °C for

30 seconds with psi between 18-22, followed by 90°C for 10 seconds. Slides were removed

from decloaking chamber and cooled for two minutes followed by a wash in Super Sensitive

Wash Buffer (BioGenex, San Ramon, CA). The rabbit polyclonal anti- β2-AR antibody was used

at a concentration of 1:100 for one hour at room temperature. Slides were rinsed with

SuperSensitive Wash Buffer and treated with Super Enhancer (BioGenex, San Ramon, CA) for

20 minutes at room temperature, and subsequently treated with Polymer-HRP (BioGenex,

Fremont, CA) for 30 minutes at room temperature and incubated with DAB (BioGenex,

Fremont, CA) at room temperature for five minutes. Slides were washed with SuperSensitive

Wash Buffer and counterstained with Mayer’s hematoxylin for one minute. Negative control

samples underwent the same process but in the absence of the primary antibody.

3.3. Cell protein extraction

Cells were grown in culture until confluence and then media was removed and cells were

washed twice with ice-cold PBS. Cells were then placed in trypsin for five minutes to detach

from culture plate, followed by addition of complete media to neutralize the effects of trypsin.

Non-adherent cells were centrifuged at 2,000 rpm at 4°C for five minutes. Supernatant was

removed and the pellet was washed twice in PBS until all trypsin and media was removed. The

  32  

pellet was then re-suspended in a solution containing Mammalian Protein Extraction Reagent

(M-PER, Pierce, Rockford, IL) and Pierce protease inhibitor cocktail solution (Pierce, Rockford,

IL) diluted 1:100 for final working solution, 10 µL of protease inhibitor to 1000 µL of M-PER.

This was placed on a shaker for 10 minutes, followed by centrifugation at 10,000 RPM, in 4°C

for 10 minutes. The supernatant was placed in separate Eppendorf tubes and placed in -80°C

until analysis.

3.4. Western blot analysis

Cellular protein concentrations were determined using a standard assay kit (Bicinchoninic

Acid Protein Assay, Rockford, IL). For each protein expression analysis, 50 µg samples were

electrophoresed on 12% polyacrylamide gel and then transferred to nitrocellulose membrane.

The membranes were blocked with either TBST with 5% milk (β2-AR and Src) or TBST with

5% BSA (p-Src), followed by addition of their primary antibody at their specific concentrations

and exposure time as described in the next sections. In all WBs, secondary HRP-conjugated

antibody was added at a concentration of 1:1000 for one hour at room temperature after the

membranes were washed three times with TBST for five minutes, and developed using

chemiluminescent substrate followed by detection with enhanced chemiluminescence (ECL)

detection system (Amersham). The blots were imaged using ChemiDoc XRS+ molecular imager

system (Bio-Rad Laboratories, Hercules, CA) with Image Lab software (version 5.1 build 8, Bio-

Rad Laboratories). The evaluation of protein loading within the blots was performed by

incubating the membranes with anti-β-actin antibody (1:2000) without the need for stripping the

membranes since the expected molecular weight of β-actin band was below (46 kDa) the other

bands evaluated. The membranes were subsequently incubated with the secondary antibody and

  33  

imaged as described previously. The β-actin loading control is important to verify that the

amount of protein loaded into the gel is approximately equal.

3.4.1 β2-AR

Protein expression of β2-AR was performed via Western blotting using cell lysates from

Fitz and DEN, as well as HeLa and MDA-MB-231 cells as positive controls [219, 220]. The

primary antibody was diluted in 5% milk in TBST at a concentration of 1:250. Band specificity

was confirmed with a matching blocking peptide. For evaluation of the blocking peptide, one

membrane containing duplicated samples was separated into two membranes becoming two

exact copies. One portion of the membrane was treated first with the blocking peptide (1:250)

diluted in 5% milk/TBST for one hour at room temperature, rinsed three times with TBST and

then incubated with the primary antibody for one hour at room temperature. The other portion of

the membrane was incubated with only the primary antibody (1:250) for one hour at room

temperature. Subsequent incubation with secondary antibody and analysis was performed as

described above.

3.4.2. Src

Protein expression of Src was performed via Western blotting using cell lysates from Fitz

and DEN, and MDCK as a positive control (per antibody manufacturer). Primary antibody

(1:1000) was diluted in TBST with 5% milk incubated overnight. Basal Src protein expression

was determined in all three cell lines. In addition, protein expression was assessed in Fitz and

DEN exposed to different concentrations of NE and P (0.1 and 10 µM) for 24 hours.

3.4.3. Phospho-Src

  34  

Protein expression of phosphorylated Src (p-Src) was performed via Western blotting

using cell lysates from Fitz and DEN. The membrane was incubated overnight with the anti-

phospho-Src (1:1000) antibody diluted in TBST with 5% BSA. For the analysis of

phosphorylation status of Src after cell treatment, several experiments were performed under

varying experimental conditions. For positive control of Src phosphorylation in all experiments,

cells were exposed to 3 mM of H2O2 for five minutes per manufacturer’s instructions. First

experiment: DEN and Fitz cells were serum-starved overnight then treated with NE (10 µM) or P

(0.1 µM) for different time exposures (5-30 minutes). Second experiment: To optimize

visualization of the agonism, Fitz cells were serum-starved for 24 hours then treated with NE (10

µM) at increasing time-exposure (5-60 minutes). The goal was to determine if phosphorylation

was positive upon contact with the agonist, if that difference was visible in the WB, and how

much exposure was necessary for optimal visualization. Third experiment: To determine if

propranolol could block NE agonism, Fitz cells were serum-starved for 24 hours, then pre-

treated with propranolol (0-0.1 µM) followed by NE treatment (10 µM) at increased time

exposures (5-60 minutes). Fourth experiment: To determine if propranolol could rescue NE

agonism, DEN and Fitz cells were serum-starved for 24 hours, then pre-treated with NE (0-10

µM) for 45 minutes followed by P (0-100 µM) for 24 hours. Band volume analysis was

performed using Image Lab software and results of the Phospho-Src proteins were adjusted for

β-actin expression levels used as loading control. For the third experiment, results are expressed

as p-Src/ β-actin ratio.

3.5 VEGF ELISA

  35  

Fitz and DEN cells were plated at a density of 5 x 103 cells/well in a 96-well plate. After

allowing cells to adhere to the plate overnight, the medium was removed and replaced with fresh

medium containing NE and/or P. In the first experiment, optimal agonism was assessed using 0-

10 µM NE for 24 hours. In the second experiment, cells were treated with 0-100 µM P with NE

agonism at 1 µM which was selected as the optimal agonist. Cell culture supernatants were

harvested and soluble VEGF was determined with a commercially available canine VEGF

ELISA immunoassay (Canine VEGF DuoSet, R&D Systems, Minneapolis, MN). Differences in

soluble VEGF secreted by the cHSA cells after each treatment were normalized to cell count

with the use of a colorimetric proliferation assay (CellTiter 96® AQueous One Solution Cell

Proliferation Assay – MTS, Promega, Madison, WI) in which optical density linearly correlates

with viable cell numbers. Specifically, normalized VEGF concentrations were based on the

average of quadruplicate samples for each experimental group expressed as the following ratio:

Normalized VEGF = [Calculated VEGF (pg/mL)]/optical density. Samples were performed with

six replicates per experiment and the results were replicated in three independent experiments.

3.6. Proliferation assay

Fitz and DEN cell proliferation was assessed using a commercial colorimetric assay per

manufacturer’s directions (CellTiter 96® AQueous One Solution Cell Proliferation Assay –

MTS, Promega, Madison, WI). This assay measures the number of viable cells by using a

tetrazolium compound that is bioreduced by metabolically active cells into a colored formazan

product, which is detected via spectrophotometry. Fitz and DEN were seeded into a 96-well plate

at a density of 5 x 103 cells/well. After allowing adherence overnight, the cells were treated with

0, 0.1, 1 and 10 µM of NE or P for 24 hours (first experiment) and with 0, 0.1 or 10 µM NE, 0.1

  36  

or 10 µM P, 0.1 or 10 µM NE + P (second experiment) for 48 hours. The supernatant was

removed, and 120 µL of fresh media containing 20 µL of the reagent was added to each well.

The plate was then incubated at 37°C in 5% CO2 for 1, 2 and 3 hours. At each hour,

spectrophotometer readings were obtained with a photometric plate reader (Bio-Tek EL 800) at a

490nm absorbance (wavelength correction set to 570 nm). The data collected for treated cells

was compared to untreated cells of the same experiment and the difference was calculated as a

percentage of control cells. The groups were subsequently compared to each other. Samples were

performed with six replicates per experiment and the results were replicated in three independent

experiments.

3.7. Scratch assay

Analysis of cell migration was done using the “scratch assay” method [221]. Fitz and

DEN were grown in culture in 6 well plates until reaching confluence. A scratch was then made

along the center of the cell culture area using a p200 pipette tip. The medium was then replaced

with fresh medium containing P (0.01-100 µM) or no treatment (control). An image of the

scratched cell monolayer was taken (“time 0”) using an inverted microscope (Nikon Eclipse

TS100) with a mounted digital camera (SPOT Insight QE model #4.2, SPOT Imaging Solutions,

Michigan, USA). Cells were then grown in culture for 24 hours and another image was taken at

the same location of the initial image. Captured images were analyzed with Image J software

(National Institutes of Health, Bethesda, MD). A total of 5 measurements of the gap between the

cells were taken for each image and average measurements were compared between time 0 and

24h. Experiment was performed in triplicate with replication.

3.8. Statistical analysis

  37  

Data from proliferation, VEGF assay and scratch assay were analyzed for normality with

the Kolmogorov-Smirnov test. Because of the small number of samples, histograms were also

performed for visual assessment. Data is reported as mean and standard deviation. One-way

ANOVA was used to evaluate for difference between groups, with Tukey’s multiple comparison

test to detect a difference between groups. GraphPad Instat 3.1 and GraphPad Prism 6.1 software

were used for analysis.

  38  

CHAPTER 4

RESULTS

4.1. β2-AR is expressed in spontaneous canine splenic hemangiosarcoma

To determine β2-AR protein expression in spontaneous canine hemangiosarcoma, IHC

was performed. First, the 18 selected tissue samples were confirmed cHSA based on H&E

staining and positivity for CD31. All tissue samples demonstrated positive β2-AR expression

(Fig 4.1). Expression of the receptor was identified in the nucleus and cytoplasm, with varying

stain intensity between tissue samples. Five of 18 (27%) samples had strong intensity and were

almost diffusely labeled. Eight of 18 (44%) of the samples had only cytoplasmic staining while

10/18 (56%) had staining of both the nucleus and the cytoplasm. The labeling was frequently

heterogeneous with some cases presenting large unlabeled areas. Isotype control did not reveal

any significant non-specific background staining.

4.2. β2-AR is expressed in canine hemangiosarcoma cell lines

To determine β2-AR protein expression in canine hemangiosarcoma cell lines Fitz and

DEN, WB and IHC were performed. Using WB, we identified a band at the expected molecular

weight of 55 kDa in both human positive control cell lines MDA-MB-231 and HeLa, as well as

the canine cell lines Fitz and DEN (Fig. 4.2b). Qualitatively, Fitz appears to have stronger

protein expression than DEN. Specificity was confirmed with abrogation of the band in the

presence of a blocking peptide. β2-AR protein expression was also confirmed via IHC on Fitz

and DEN cell pellets using the same antibody. Cell pellets demonstrated strong, uniform, diffuse

staining (Fig. 4.2a).

  39  

4.3. Investigating intermediate signaling via Src/p-Src

4.3.1. Src and p-Src are expressed in cHSA cell lines

Src and phosphorylated Src (Tyr416) are intermediaries involved in signal transduction

from β2-AR to the nucleus [19]. To determine their involvement in cHSA cells we performed

WB. We identified bands (Src and p-Src) at the expected molecular weight of 60 kDa in both

canine positive control cell line MDCK as well as the canine cell lines Fitz and DEN (Fig. 4.3a).

4.3.2. Propranolol attenuates norepinephrine-induced Src phosphorylation

Given that Src and phosphorylated Src (Tyr416) are intermediaries involved in signal

transduction from the β2-AR to the nucleus, we sought to determine if the β2-AR agonist (NE)

would lead to phosphorylation of the protein, and if this process could be blocked by the

antagonist (P). To determine the effects of NE and P on Src and p-Src, we performed WB on

cHSA cells. In DEN and Fitz cells treated with different concentrations of NE and P (up to 10

µM), there was no qualitative difference in the expression of total Src (Fig. 4.3b). Both DEN and

Fitz demonstrated significant basal phosphorylation of Src, suspected to be caused by the FBS

used in culture. Overnight serum-starvation in DEN and Fitz cell lines did not result in an

apparent difference in p-Src protein expression with P (low dose 0.1 µM) and increased time-

exposure to NE (Fig. 4.3c). To optimize phosphorylation, Fitz cells were serum-starved for 24

hours, followed by increasing time exposure to NE. Increased protein expression of p-Src was

identified, with phosphorylation more pronounced starting at 15 minutes and persisting through

60 minutes (Fig. 4.3d). Once the positive effects of the agonist NE on phosphorylation were

demonstrated, we proceeded to evaluate if pre-exposure to low-dose propranolol (0.1 µM) was

  40  

sufficient to block the receptor to the effects of the agonist. In Fitz cells, NE caused a modest

phosphorylation of Src and treatment with low-dose propranolol for 24 hours prior to exposure

led to a modest decrease in phosphorylation. These effects were visible in the WB and therefore

we proceeded to quantify these results by normalizing the adjusted volume of the p-Src band to

the matching β-actin band, to ensure that the increased intensity of the band observed was not

due to increased protein loading (Fig. 4.3e). To determine if propranolol could rescue NE

agonism, DEN and Fitz cells were serum-starved, pre-treated with NE (0-10 µM), then

propranolol (0-100 µM). In Fitz cells, low doses of propranolol reduce p-Src compared to

untreated cells, which is lost at the highest dose in unstimulated cells. In NE-stimulated Fitz

cells, propranolol reduces p-Src compared to basal levels at all doses. Conversely, in DEN cells,

propranolol increases p-Src in untreated and NE stimulated cells (Fig. 4.3f)

4.4. Propranolol affects VEGF secretion of cHSA in vitro

After demonstrating the presence of the receptor as well as a possible intermediate

protein involved in signal transduction, we proceeded to investigate the effects of NE and P on

cell function activities of cHSA cells. Since propranolol’s ability to inhibit tumor angiogenesis

has been largely implicated as a primary anti-neoplastic effect, we sought to evaluate the effects

of NE and P on the ability of cHSA cells to secrete VEGF, a key pro-angiogenic cytokine. In the

first experiment, we used NE at varying doses (0-10 µM) to assess optimal agonism. Normalized

to cell count, basal secretion of VEGF was 1098.83 pg/mL and 464.42 pg/mL for DEN and Fitz

cells, respectively. All doses of NE induced VEGF secretion compared to untreated cells for both

cell lines. Optimal NE agonism was obtained at 1 µM NE, with a significant increase in VEGF

secretion compared to control. At optimal NE agonism, there was an increase in VEGF of 29%

  41  

for DEN cells (p<0.05 for all treatments compared to control), and 2.6-fold increase for Fitz cells

(p<0.001 for all treatments compared to control) (Fig. 4.4a). Cells were then incubated with the

selected NE dose of 1 µM combined with varying doses of P starting at a biologically achievable

dose (0.1 µM) up to 100 µM. No significant differences were noted at low doses, however there

was a reduction in VEGF secretion in DEN at P ≥ 10 µM (p<0.01) with a maximum decrease in

VEGF of 21% at 100 µM. Similar results were seen with Fitz, with reduced VEGF secretion at P

≥ 10 µM (p<0.001) with maximum 44% reduction at 100 µM; a dose dependence was noted in

the Fitz cell line (p<0.05) (Fig 4.4b). Similar but less profound results were seen without NE

stimulation (results not shown).

4.5. Propranolol affects cell proliferation of cHSA in vitro

The metabolic activity of Fitz and DEN, an indirect measurement of cell viability and

proliferation, was evaluated upon exposure to different doses of NE and P. While there was no

significant difference noted after 24h incubation, when treated for 48h Fitz exhibited significant

difference in proliferation between untreated vs 0.1 µM of NE (25% increase) (p < 0.05) and

untreated vs 0.1 µM of P (30% increase) (p < 0.01). DEN did not exhibit significant difference

between groups (p > 0.05) (Fig. 4.5)

4.6. Propranolol affects cell migration of cHSA in vitro

The ability of cancer cells to move within the microenviroment towards the nearby

vasculature is important in the process of metastasis and therefore we elected to evaluate if P had

  42  

any effect on cell mobility by evaluating the cells ability to close a gap created between the cells.

Increasing concentrations of P were used (0-100 µM) and an artificially created gap was

measured before and 24 hours after treatment. Data is represented as a percentage of the original

gap (time 0), with being 100% a completely closed gap. The gap remained more noticeably open

at 10 and 100 µM, being statistically significant at 100 µM when compared to untreated control

cells (p<0.05) (Fig. 4.6).

  43  

CHAPTER 5

DISCUSSION AND CONCLUSIONS

In veterinary medicine, the emotional component of stress may not play an important role

as it does in humans, however the presence of cancer can lead to microenvironmental stressors

such as hypoxia and inflammation which may lead to catecholamine release [222, 223].

Therefore, stress can play a part in canine tumor progression. In this study we were able to

demonstrate the presence of β2-AR in all spontaneous splenic cHSA as well as cell lines

demonstrating the clinical relevance and consistency of expression, and suggesting that this

pathway may be activated by stress-related catecholamines. The Fitz cell line appeared to have

stronger protein expression than the DEN cell line. The expression of β2-AR in spontaneous

tumor samples was heterogeneous in intensity and localization within the cells. The variability of

IHC staining could be explained by a variable expression of the receptor within the neoplastic

cells, as well as integrity and detectability of the antigen depending on the time the tissue has

been stored and the fixative used to initially preserve it as described in a study looking at EGFR

[224, 225]. Since the samples tested were from different cases obtained over several years, the

intensity of the staining obtained could be a reflection of the particular sample rather than

variability within the tumors. However, since tumors undergo frequent mutations, some may

express more of a particular antigen or receptor compared to others and this variability may

account for the difference between cytoplasmic versus nuclear staining. Regarding cellular

localization, upon stimulation β2-AR undergoes internalization, intracellular trafficking and

recycling and thus variability in cytoplasmic localization may reflect variation in stimulation and

internalization [226]. Nuclear membrane localization of β1- and β3-AR as well as downstream

  44  

signaling partners, but not β2-AR, has been reported in normal cardiac myocytes [227]. Our IHC

results are in concordance with similar studies in human cancer. In a study of β2-AR IHC

expression in hepatocellular carcinoma, 60% of tumors had moderate/strong staining intensity,

31% weak staining intensity and 9% negative staining with nuclear and cytoplasmic distribution

[228]. Another study showed strong β2-AR staining in 41% of angiosarcoma samples that were

analyzed and a 77% total expression indicating heterogeneity in receptor expression, similar to

the results obtained in this project [34]. Further evaluation of localization of β2-AR and signaling

partners could include Western blotting of fractionated cell lysates, confocal microscopy with

immunofluorescence and other methods.

The non-receptor tyrosine kinase Src has been reported as one of the many downstream

proteins involved in β2-AR activation and therefore we elected to investigate if this pathway is

indeed activated and blocked by β2-AR agonism/antagonism [19]. After initially demonstrating

the presence of Src within the cell, we proceeded to verify if this protein is phosphorylated by

NE. The initial challenge was to overcome the strong basal phosphorylation of this protein

within the cHSA, which did not allow for a difference in visualization between untreated and

treated cells. Part of this phosphorylation was probably due to the presence of FBS, which

contains growth factors that can stimulate the Src pathway. After 24h serum-starvation there was

still p-Src activity within the untreated cells, however it was diminished enough that an

appreciable difference was seen compared to NE-treated cells. Norepinephrine agonism of Src

phosphorylation was confirmed, supporting a role for catecholamine-induced tumor stimulation.

The subsequent experiment was performed with pre-treatment of cells with P, showing a

decrease in phosphorylation at low doses of P. With NE agonism prior to P treatment (i.e.

propranolol “rescue”), which may most closely resemble the clinical scenario, P successfully

  45  

decreased Src phosphorylation in NE pretreated cells, while phosphorylation was variable in

DEN cells. This may be a factor of β2-AR expression, which was qualitatively higher in Fitz

cells. The preliminary results of this experiment warrants further investigation of this pathway,

however due to multiple pathways involved in β2-AR signaling it is possible that alternative

pathways are stimulated or inhibited with propranolol leading to the effects seen. These effects

were related to proliferation, VEGF secretion and cell migration. With respect to VEGF

secretion, cHSA cells secrete basal levels of VEGF with NE causing increased VEGF secretion,

more evident in the Fitz cell line. This supports our hypothesis that NE would increase VEGF

secretion as demonstrated in other studies [229]. The increase in VEGF may be a result of Src

phosphorylation as hypothesized, and the more profound reduction in VEGF in the Fitz cell line

may be related to the increased β2-AR expression as well as the more consistent phosphorylation

effects of P. Taken together, this data supports a role for the β2-AR and the potential role of

propranolol in inhibiting VEGF-mediated HSA angiogenesis.

The effects of NE and P on cell proliferation were also investigated. The initial

hypothesis was that NE would lead to increased proliferation and P would decrease proliferation

[230]. Interestingly, we found that both NE and P caused an increase in proliferation in Fitz cells

at 0.1 µM. There is one report in the literature of P inhibiting proliferation of pulmonary artery

smooth muscle cells at a 10 µM but not at 0.1 µM. The mechanism for this effect on proliferation

remained unexplained [231]. One possible explanation may be that this β-blocker could act as an

agonist to a different receptor or pathway involved in cell proliferation. Additional studies such

as time-dependent effects may elucidate this further. In one rodent study, propranolol caused a

transient increase in markers of cell proliferation (PCNA, mitotic index) that were abrogated

with longer term treatment [232]. The variation observed between the different cell lines in this

  46  

project was also observed in another study, where beta blockade effectively reduced proliferation

rate among multiple cell lines however some lines were more resistant to these effects than

others [214].

Our last finding was that P at a high dose (100 µM) had an effect on cell migration. This

effect has been described previously in human infantile hemangioma endothelial cells and it also

decreased phosphorylation of cofilin, as a possible mechanism involved (since cofilin is an actin-

severing protein and phosphorylation is an inhibitory event) [136, 233]. The dose necessary for a

significant effect is not biologically achievable, however a time-dependent effect at lower doses

was not investigated in this study and is an avenue for further investigation. Decrease in cell

migration was detected at a 50 µM dose in one study, which was careful to evaluate the effects of

this dose on net cell proliferation to eliminate inhibition of proliferation as a contributing factor

[214]. It is possible that our dose of 100 µM is affecting cell proliferation, which is indirectly

leading to part of the migration effects seen on the scratch assay.

In relating this study to the clinical scenario of canine hemangiosarcoma, the finding of

β2-AR expression in both spontaneous tumors and two cell lines, strong basal phosphorylation of

Src, as well as consistent agonism of downstream signaling by NE supports a role of this

pathway in tumor progression. Variability in expression may suggest a variable response to

pathway blockade. Effects on VEGF secretion and cell migration were most significant at the

highest doses of propranolol, which are not biologically achievable with single-dose treatment at

standard doses. The high doses of propranolol needed to impair tumor cell function is consistent

with that seen in infantile hemangioma – most of these studies show in vitro effect at 100-300

µM, which is not clinically achievable. However, in infantile hemangioma, continuous treatment

is used clinically, which may account for the discordancy between in vitro and in vivo effect. The

  47  

finding of positive effects on Src phosphorylation at low doses in cHSA may also support a

potential for low-dose propranolol time-dependent effects on VEGF secretion or cell migration,

although this was beyond the scope of this study. Continuous low dose treatment with

propranolol would be reasonable in an adjunctive setting in cHSA due to its low cost and side

effect profile.

Although this study indicated the presence of the β2-ARs in cHSA and that propranolol

exerts some biological effects in cHSA cell lines, it is important to acknowledge some

limitations of this project. There is no direct evidence that Src is the intermediary pathway

connecting β2-AR agonism and cell proliferation, VEGF secretion or cell migration since there

are many other possible pathways involved [19]. We attempted to elicit phosphorylation of Src

by exposing the cells to NE over time in order to detect the optimal moment at which most of the

protein is phosphorylated and thus more easily detectable. We were only able to detect a small

percentage of phosphorylation and even smaller attenuation by propranolol in one experiment.

The necessary amount of phosphorylated Src protein required for downstream activation of

subsequent proteins (such as STAT3 [234]) is unknown and could be as low as 1% and as high

as 100%, therefore the amount of necessary blockade of phosphorylation for biological effect

could also be variable. The use of an agonist with a stronger affinity for β2-AR, such as

isoproterenol [235], could lead to greater phosphorylation of Src and allow for more objective

visualization of the effects of propranolol. The cells were treated with 0.1 µM of propranolol for

one hour, which is a biologically achievable dose; however it does not represent the exposure

time to which an animal is subjected to, being a drug of continuous use. The same critique can be

applied to the other assays in which the cells were treated for either 24 or 48 hours, revealing

only statistically significant results at higher doses, which are not biologically achievable. Most

  48  

studies looking at the in vitro effects of norepinephrine or propranolol used high doses of

agonist/antagonist, probably because it was necessary to achieve significant findings. Drug

absorption and delivery to the tumor can be variable in each patient depending on tumor blood

flow and microenviroment, which is one limitation of translating in vitro findings to in vivo

effects. Additionally, as noted in one study where propranolol did not completely abolish tumor

cell proliferation in vivo as noted by positive PCNA staining in propranolol-treated tumors,

single-agent propranolol may not be as effective in tumor control however synergistic effects

may be observed with chemotherapeutics warranting further investigation into combination

therapy [214].

In conclusion, our study showed that cHSA express β2-ARs, Src kinase is one pathway

activated by this receptor and agonism/antagonism of this receptor exerts anti-angiogenic effects

in VEGF secretion and minimal effect on cell proliferation. Propranolol at high doses affects cell

migration. Further studies looking at other possible intermediate pathways as well as other

mechanisms involved in cancer development (matrix metalloproteinases or apoptosis) are areas

for further investigation.

  49  

CHAPTER 6

FIGURES

Figure 2.1. G Protein-coupled receptor activation (Belmonte S.L and Blaxall B.C., Circulation Research 109 (3), 309-319, July 2011).

Figure 2.2. G Protein-coupled receptor downstream signaling. (Cole S.W. and Sood A.K., Clin Cancer Res 18 (5), 1201-1206, March 2012).

  50  

Figure 4.1. Expression of β2-AR in spontaneous cHSA. Immunohistochemistry was performed on formalin-fixed paraffin-embedded sections containing tissue from splenic cHSA. All samples stained positive for the β2-AR antibody and the degree of staining was variable. Examples of strong β2-AR stain intensity (top row) and mild β2-AR stain intensity (bottom row) are shown. The diagnosis of splenic cHSA was confirmed after review of H&E staining and CD31. Negative control was performed using an isotypic control antibody (1000x).

Figure 4.2a. Expression of β2-AR in cHSA cell lines. Immunohistochemistry was performed on formalin-fixed paraffin-embedded sections of agar-embedded cell pellets containing cHSA cells Fitz and DEN. Human mammary carcinoma cell line MDA-MB-231 was used as a positive control. Negative control was the absence of the primary antibody (10x).

  51  

Figure 4.2b. Expression of β2-AR protein in cHSA cell lines. Western blot was performed using whole cell lysates of Fitz and DEN. Cell lines HeLa and MDA-MB-231 were used as positive control.

Figure 4.3a. Expression of Src and p-Src in cHSA cell lines. Western blot using an antibody for Src and p-Src was performed using whole cell lysates of Fitz, DEN and MDCK (control). For better visualization of p-Src the cells were serum-starved and stimulated with 3 mM of H2O2 for 5 minutes. All cells demonstrated the presence of basal Src protein as well as phosphorylated Src protein.

Figure 4.3b. Expression of Src in cHSA cell lines exposed to different concentrations of NE and P. Fitz and DEN cells were treated with different concentrations of NE and P for 24 hours as indicated below. Western blot revealed bands of similar size across all treatments, indicating that the presence of total Src protein is independent of the stimulus.

  52  

Figure 4.3c. Expression of p-Src in cHSA cell lines during increasing exposure to NE and P. Fitz and DEN cells were serum-starved overnight and treated with 10 µM of NE and 0.1 µM of P for different lengths of time as indicated below. Despite 12hr serum-starvation there was still marked baseline phosphorylation of Src, which did not allow for qualitative distinction between control and treated groups.

Figure 4.3d. Expression of p-Src in Fitz cell line during increasing exposure to NE. Fitz was serum-starved for 24 hr and subsequently increasingly exposed to 10 µM of NE. There was a qualitative difference between untreated control and 15-60 minute exposure.

  53  

Figure 4.3e. Expression of p-Src in Fitz cell line during increasing exposure to NE after pretreatment with P. Fitz was serum-starved for 24 hr and cells received either no pre-treatment or 0.1µM of propranolol for 24 hours. Both groups were subsequently increasingly exposed to 10 µM of NE. There was a modest increase in phosphorylation of Src over time (23% at 60’) and there was modestly decreased phosphorylation when the cells had been pretreated with propranolol (11% at 60’). This difference was quantified as the ratio of chemiluminescence emitted by p-Src and β-actin for each band.

Figure 4.3f. Expression of p-Src in DEN and Fitz cell line pretreated with NE and rescued with propranolol. DEN and Fitz were serum-starved for 24 hr and cells received either no pre-treatment or 10µM of NE for 45 minutes. Cells were subsequently exposed to 0-100µM of P. In non-pretreated Fitz cells, low doses of propranolol reduce p-Src, which is lost at the highest dose. With NE agonism in Fitz cells, propranolol reduces p-Src compared to basal levels at all doses. In DEN cells, propranolol increases p-Src in untreated and NE stimulated cells.

  54  

Figure 4.4a. VEGF secretion by cHSA cell lines after treatment with NE. The graph represent an average of VEGF secretion by Fitz and DEN quantified by ELISA and normalized to cell count. Cells were treated with increasing doses of NE (0-10 µM) for 24 hours. In both cell lines, compared to untreated cells, there is a significant difference in VEGF secretion (DEN p<0.05, Fitz p<0.001). Error bars represent SD.

Figure 4.4b. VEGF secretion by cHSA cell lines after treatment with NE and P. The graphs represent an average of VEGF secretion by Fitz and DEN quantified by ELISA and normalized to cell count. All cells were treated with 1 µM NE (agonist) and P at 0-100µM for 24 hours. At propranolol doses of ≥10  µM, there was a significant difference of VEGF secretion by both cell lines compared to untreated cells (DEN p<0.01, Fitz p<0.001). For Fitz, there was a significant difference between 10 and 100 µM. p<0.05. Error bars represent SD.

  55  

Figure 4.5. Proliferation activity of cHSA cell lines after treatment with NE and P. Fitz and DEN were treated with different concentrations of NE and P for 24 hours (top two graphs) and subsequently analyzed by MTS assay. There was no statistical difference between the groups in both cell lines (p > 0.5). When cells were treated for 48 hours (bottom 2 graphs), Fitz exhibited significant difference between untreated vs 0.1 µM of NE (p < 0.05) and untreated vs 0.1 µM of P (p < 0.01). Error bars represent SD.

  56  

Figure 4.6. Migration of cHSA cells when exposed to increasing doses of propranolol. A scratch assay was performed with Fitz and DEN exposed to increasing concentrations of propranolol for 24 hours, and an average measurement of the residual gap in each treatment group was compared to an untreated control. There was a significant difference between the untreated group and 100 µM (p < 0.05). Error bars represent SD.

  57  

Figure 4.6 (cont). Migration of cHSA cells when exposed to increasing doses of propranolol. A scratch assay was performed with Fitz and DEN exposed to increasing concentrations of propranolol for 24 hours, and an average measurement of the residual gap in each treatment group was compared to an untreated control. There was a significant difference between the untreated group and 100 µM (p < 0.05). Error bars represent SD.

  58  

REFERENCES

1.   Priester,  W.A.,  Hepatic  angiosarcomas  in  dogs:  an  excessive  frequency  as  compared  with  man.  J  Natl  Cancer  Inst,  1976.  57(2):  p.  451-­‐4.  

2.   Priester,  W.A.  and  F.W.  McKay,  The  occurrence  of  tumors  in  domestic  animals.  Natl  Cancer  Inst  Monogr,  1980(54):  p.  1-­‐210.  

3.   Spangler,  W.L.  and  M.R.  Culbertson,  Prevalence,  type,  and  importance  of  splenic  diseases  in  dogs:  1,480  cases  (1985-­‐1989).  J  Am  Vet  Med  Assoc,  1992.  200(6):  p.  829-­‐34.  

4.   Thamm,  D.H.,  Hemangiosarcoma,  in  Small  Animal  Clinical  Oncology,  S.J.  Withrow,  Vail  D.M.,  Page,  R.L.,  Editor.  2013,  Elsevier:  St.  Louis,  MO.  p.  679-­‐688.  

5.   Leaute-­‐Labreze,  C.,  et  al.,  Propranolol  for  severe  hemangiomas  of  infancy.  N  Engl  J  Med,  2008.  358(24):  p.  2649-­‐51.  

6.   Leboulanger,  N.,  et  al.,  Propranolol  in  the  therapeutic  strategy  of  infantile  laryngotracheal  hemangioma:  A  preliminary  retrospective  study  of  French  experience.  Int  J  Pediatr  Otorhinolaryngol,  2010.  74(11):  p.  1254-­‐7.  

7.   Kurzyna,  A.,  et  al.,  [Propranolol  for  treatment  of  subglottic  hemangioma].  Otolaryngol  Pol,  2010.  64(6):  p.  388-­‐91.  

8.   Jadhav,  V.M.  and  S.N.  Tolat,  Dramatic  response  of  propranolol  in  hemangioma:  report  of  two  cases.  Indian  J  Dermatol  Venereol  Leprol,  2010.  76(6):  p.  691-­‐4.  

9.   Fay,  A.,  et  al.,  Propranolol  for  isolated  orbital  infantile  hemangioma.  Arch  Ophthalmol,  2010.  128(2):  p.  256-­‐8.  

10.   Baetz,  J.,  et  al.,  [Infantile  hemangioma.  Successful  treatment  with  propranolol].  Hautarzt,  2010.  61(4):  p.  290-­‐2.  

11.   Denoyelle,  F.,  et  al.,  Role  of  Propranolol  in  the  therapeutic  strategy  of  infantile  laryngotracheal  hemangioma.  Int  J  Pediatr  Otorhinolaryngol,  2009.  73(8):  p.  1168-­‐72.  

12.   Barron,  T.I.,  et  al.,  Beta  blockers  and  breast  cancer  mortality:  a  population-­‐  based  study.  J  Clin  Oncol,  2011.  29(19):  p.  2635-­‐44.  

13.   Powe,  D.G.,  et  al.,  Beta-­‐blocker  drug  therapy  reduces  secondary  cancer  formation  in  breast  cancer  and  improves  cancer  specific  survival.  Oncotarget,  2010.  1(7):  p.  628-­‐38.  

14.   Buhler,  H.U.,  et  al.,  Plasma  adrenaline,  noradrenaline  and  dopamine  in  man  and  different  animal  species.  J  Physiol,  1978.  276:  p.  311-­‐20.  

15.   http://www.oxforddictionaries.com/us/definition/american_english/stress).  .  16.   Reiche,  E.M.,  S.O.  Nunes,  and  H.K.  Morimoto,  Stress,  depression,  the  immune  system,  and  

cancer.  Lancet  Oncol,  2004.  5(10):  p.  617-­‐25.  17.   Chrousos,  G.P.,  The  hypothalamic-­‐pituitary-­‐adrenal  axis  and  immune-­‐mediated  inflammation.  N  

Engl  J  Med,  1995.  332(20):  p.  1351-­‐62.  18.   Dimsdale,  J.M.,  M,  Plasma  catecholamines  in  stress  and  exercise.  JAMA,  1980.  243(2):  p.  340-­‐

342.  19.   Cole,  S.W.  and  A.K.  Sood,  Molecular  pathways:  beta-­‐adrenergic  signaling  in  cancer.  Clin  Cancer  

Res,  2012.  18(5):  p.  1201-­‐6.  20.   Daly,  C.J.  and  J.C.  McGrath,  Previously  unsuspected  widespread  cellular  and  tissue  distribution  of  

beta-­‐adrenoceptors  and  its  relevance  to  drug  action.  Trends  Pharmacol  Sci,  2011.  32(4):  p.  219-­‐26.  

21.   Carroll,  B.T.,  et  al.,  Screening  for  depression  and  anxiety  in  cancer  patients  using  the  Hospital  Anxiety  and  Depression  Scale.  Gen  Hosp  Psychiatry,  1993.  15(2):  p.  69-­‐74.  

22.   Sellick,  S.M.  and  A.D.  Edwardson,  Screening  new  cancer  patients  for  psychological  distress  using  the  hospital  anxiety  and  depression  scale.  Psychooncology,  2007.  16(6):  p.  534-­‐42.  

23.   Mantovani,  A.,  et  al.,  Cancer-­‐related  inflammation.  Nature,  2008.  454(7203):  p.  436-­‐44.  

  59  

24.   Candido,  J.  and  T.  Hagemann,  Cancer-­‐related  inflammation.  J  Clin  Immunol,  2013.  33  Suppl  1:  p.  S79-­‐84.  

25.   Pollock,  R.E.  and  J.A.  Roth,  Cancer-­‐induced  immunosuppression:  implications  for  therapy?  Semin  Surg  Oncol,  1989.  5(6):  p.  414-­‐9.  

26.   Jenkins,  F.J.,  B.  Van  Houten,  and  D.H.  Bovbjerg,  Effects  on  DNA  Damage  and/or  Repair  Processes  as  Biological  Mechanisms  Linking  Psychological  Stress  to  Cancer  Risk.  J  Appl  Biobehav  Res,  2014.  19(1):  p.  3-­‐23.  

27.   Obeid,  E.I.  and  S.D.  Conzen,  The  role  of  adrenergic  signaling  in  breast  cancer  biology.  Cancer  Biomark,  2013.  13(3):  p.  161-­‐9.  

28.   Yang,  E.V.  and  T.D.  Eubank,  The  impact  of  adrenergic  signaling  in  skin  cancer  progression:  possible  repurposing  of  beta-­‐blockers  for  treatment  of  skin  cancer.  Cancer  Biomark,  2013.  13(3):  p.  155-­‐60.  

29.   Bridle,  P.A.,  et  al.,  Basal  levels  of  plasma  epinephrine  and  norepinephrine  in  the  dog.  Hypertension,  1983.  5(6  Pt  3):  p.  V128-­‐33.  

30.   Sloan,  E.K.,  et  al.,  The  sympathetic  nervous  system  induces  a  metastatic  switch  in  primary  breast  cancer.  Cancer  Res,  2010.  70(18):  p.  7042-­‐52.  

31.   Taylor,  M.R.,  Pharmacogenetics  of  the  human  beta-­‐adrenergic  receptors.  Pharmacogenomics  J,  2007.  7(1):  p.  29-­‐37.  

32.   Audet,  M.  and  M.  Bouvier,  Insights  into  signaling  from  the  beta2-­‐adrenergic  receptor  structure.  Nat  Chem  Biol,  2008.  4(7):  p.  397-­‐403.  

33.   Rosenbaum,  D.M.,  S.G.  Rasmussen,  and  B.K.  Kobilka,  The  structure  and  function  of  G-­‐protein-­‐coupled  receptors.  Nature,  2009.  459(7245):  p.  356-­‐63.  

34.   Chisholm,  K.M.,  et  al.,  beta-­‐Adrenergic  receptor  expression  in  vascular  tumors.  Mod  Pathol,  2012.  25(11):  p.  1446-­‐51.  

35.   Lipka,  E.,  et  al.,  In  vivo  non-­‐linear  intestinal  permeability  of  celiprolol  and  propranolol  in  conscious  dogs:  evidence  for  intestinal  secretion.  Eur  J  Pharm  Sci,  1998.  6(1):  p.  75-­‐81.  

36.   Vauquelin,  G.a.v.M.,  B.  ,  G-­‐protein  coupled  receptors.  G  Protein-­‐Coupled  Receptors:  Molecular  Pharmacology  from  Academic  Concept  to  Pharmaceutical  Research,  2007:  p.  77-­‐230.  

37.   Luttrell,  L.M.,  et  al.,  Beta-­‐arrestin-­‐dependent  formation  of  beta2  adrenergic  receptor-­‐Src  protein  kinase  complexes.  Science,  1999.  283(5402):  p.  655-­‐61.  

38.   Armaiz-­‐Pena,  G.N.,  et  al.,  Src  activation  by  beta-­‐adrenoreceptors  is  a  key  switch  for  tumour  metastasis.  Nat  Commun,  2013.  4:  p.  1403.  

39.   Guarino,  M.,  Src  signaling  in  cancer  invasion.  J  Cell  Physiol,  2010.  223(1):  p.  14-­‐26.  40.   Summy,  J.M.  and  G.E.  Gallick,  Src  family  kinases  in  tumor  progression  and  metastasis.  Cancer  

Metastasis  Rev,  2003.  22(4):  p.  337-­‐58.  41.   Shor,  A.C.,  et  al.,  Dasatinib  inhibits  migration  and  invasion  in  diverse  human  sarcoma  cell  lines  

and  induces  apoptosis  in  bone  sarcoma  cells  dependent  on  SRC  kinase  for  survival.  Cancer  Res,  2007.  67(6):  p.  2800-­‐8.  

42.   Fredriksson,  J.M.,  et  al.,  Norepinephrine  induces  vascular  endothelial  growth  factor  gene  expression  in  brown  adipocytes  through  a  beta  -­‐adrenoreceptor/cAMP/protein  kinase  A  pathway  involving  Src  but  independently  of  Erk1/2.  J  Biol  Chem,  2000.  275(18):  p.  13802-­‐11.  

43.   Fredriksson,  J.M.  and  J.  Nedergaard,  Norepinephrine  specifically  stimulates  ribonucleotide  reductase  subunit  R2  gene  expression  in  proliferating  brown  adipocytes:  mediation  via  a  cAMP/PKA  pathway  involving  Src  and  Erk1/2  kinases.  Exp  Cell  Res,  2002.  274(2):  p.  207-­‐15.  

44.   Dickerson,  E.B.,  et  al.,  Imatinib  and  Dasatinib  Inhibit  Hemangiosarcoma  and  Implicate  PDGFR-­‐beta  and  Src  in  Tumor  Growth.  Transl  Oncol,  2013.  6(2):  p.  158-­‐68.  

45.   Powe,  D.G.  and  F.  Entschladen,  Targeted  therapies:  Using  beta-­‐blockers  to  inhibit  breast  cancer  progression.  Nat  Rev  Clin  Oncol,  2011.  8(9):  p.  511-­‐2.  

  60  

46.   Guise,  T.A.,  et  al.,  Evidence  for  a  causal  role  of  parathyroid  hormone-­‐related  protein  in  the  pathogenesis  of  human  breast  cancer-­‐mediated  osteolysis.  J  Clin  Invest,  1996.  98(7):  p.  1544-­‐9.  

47.   Campbell,  J.P.,  et  al.,  Stimulation  of  host  bone  marrow  stromal  cells  by  sympathetic  nerves  promotes  breast  cancer  bone  metastasis  in  mice.  PLoS  Biol,  2012.  10(7):  p.  e1001363.  

48.   Katayama,  Y.,  et  al.,  Signals  from  the  sympathetic  nervous  system  regulate  hematopoietic  stem  cell  egress  from  bone  marrow.  Cell,  2006.  124(2):  p.  407-­‐21.  

49.   Calvi,  L.M.,  Osteoblastic  activation  in  the  hematopoietic  stem  cell  niche.  Ann  N  Y  Acad  Sci,  2006.  1068:  p.  477-­‐88.  

50.   Elefteriou,  F.,  Neuronal  signaling  and  the  regulation  of  bone  remodeling.  Cell  Mol  Life  Sci,  2005.  62(19-­‐20):  p.  2339-­‐49.  

51.   Melhem-­‐Bertrandt,  A.,  et  al.,  Beta-­‐blocker  use  is  associated  with  improved  relapse-­‐free  survival  in  patients  with  triple-­‐negative  breast  cancer.  J  Clin  Oncol,  2011.  29(19):  p.  2645-­‐52.  

52.   Haffty,  B.G.,  et  al.,  Locoregional  relapse  and  distant  metastasis  in  conservatively  managed  triple  negative  early-­‐stage  breast  cancer.  J  Clin  Oncol,  2006.  24(36):  p.  5652-­‐7.  

53.   Lin,  N.U.,  et  al.,  Sites  of  distant  recurrence  and  clinical  outcomes  in  patients  with  metastatic  triple-­‐negative  breast  cancer:  high  incidence  of  central  nervous  system  metastases.  Cancer,  2008.  113(10):  p.  2638-­‐45.  

54.   Powe,  D.G.,  et  al.,  Alpha-­‐  and  beta-­‐adrenergic  receptor  (AR)  protein  expression  is  associated  with  poor  clinical  outcome  in  breast  cancer:  an  immunohistochemical  study.  Breast  Cancer  Res  Treat,  2011.  130(2):  p.  457-­‐63.  

55.   Jansen,  L.,  et  al.,  Beta  blocker  use  and  colorectal  cancer  risk:  population-­‐based  case-­‐control  study.  Cancer,  2012.  118(16):  p.  3911-­‐9.  

56.   Friedman,  G.D.,  N.  Udaltsova,  and  L.A.  Habel,  Norepinephrine  antagonists  and  cancer  risk.  Int  J  Cancer,  2011.  128(3):  p.  737-­‐8;  author  reply  739.  

57.   Shah,  S.M.,  et  al.,  Does  beta-­‐adrenoceptor  blocker  therapy  improve  cancer  survival?  Findings  from  a  population-­‐based  retrospective  cohort  study.  Br  J  Clin  Pharmacol,  2011.  72(1):  p.  157-­‐61.  

58.   Fitzgerald,  P.J.,  Beta  blockers,  norepinephrine,  and  cancer:  an  epidemiological  viewpoint.  Clin  Epidemiol,  2012.  4:  p.  151-­‐6.  

59.   Bangalore,  S.,  et  al.,  Antihypertensive  drugs  and  risk  of  cancer:  network  meta-­‐analyses  and  trial  sequential  analyses  of  324,168  participants  from  randomised  trials.  Lancet  Oncol,  2011.  12(1):  p.  65-­‐82.  

60.   Huang,  X.Y.,  et  al.,  Norepinephrine  stimulates  pancreatic  cancer  cell  proliferation,  migration  and  invasion  via  beta-­‐adrenergic  receptor-­‐dependent  activation  of  P38/MAPK  pathway.  Hepatogastroenterology,  2012.  59(115):  p.  889-­‐93.  

61.   Guo,  K.,  et  al.,  Norepinephrine-­‐induced  invasion  by  pancreatic  cancer  cells  is  inhibited  by  propranolol.  Oncol  Rep,  2009.  22(4):  p.  825-­‐30.  

62.   Zhang,  D.,  et  al.,  beta2-­‐adrenergic  antagonists  suppress  pancreatic  cancer  cell  invasion  by  inhibiting  CREB,  NFkappaB  and  AP-­‐1.  Cancer  Biol  Ther,  2010.  10(1):  p.  19-­‐29.  

63.   Al-­‐Wadei,  H.A.,  M.H.  Al-­‐Wadei,  and  H.M.  Schuller,  Prevention  of  pancreatic  cancer  by  the  beta-­‐blocker  propranolol.  Anticancer  Drugs,  2009.  20(6):  p.  477-­‐82.  

64.   Lin,  X.,  et  al.,  Beta-­‐adrenoceptor  action  on  pancreatic  cancer  cell  proliferation  and  tumor  growth  in  mice.  Hepatogastroenterology,  2012.  59(114):  p.  584-­‐8.  

65.   Wisler,  J.W.,  et  al.,  A  unique  mechanism  of  beta-­‐blocker  action:  carvedilol  stimulates  beta-­‐arrestin  signaling.  Proc  Natl  Acad  Sci  U  S  A,  2007.  104(42):  p.  16657-­‐62.  

66.   Thaker,  P.H.,  et  al.,  Chronic  stress  promotes  tumor  growth  and  angiogenesis  in  a  mouse  model  of  ovarian  carcinoma.  Nat  Med,  2006.  12(8):  p.  939-­‐44.  

67.   Lutgendorf,  S.K.,  et  al.,  Stress-­‐related  mediators  stimulate  vascular  endothelial  growth  factor  secretion  by  two  ovarian  cancer  cell  lines.  Clin  Cancer  Res,  2003.  9(12):  p.  4514-­‐21.  

  61  

68.   Jemal,  A.,  et  al.,  Cancer  statistics,  2004.  CA  Cancer  J  Clin,  2004.  54(1):  p.  8-­‐29.  69.   Hlatky,  L.,  P.  Hahnfeldt,  and  J.  Folkman,  Clinical  application  of  antiangiogenic  therapy:  

microvessel  density,  what  it  does  and  doesn't  tell  us.  J  Natl  Cancer  Inst,  2002.  94(12):  p.  883-­‐93.  70.   Lee,  J.W.,  et  al.,  The  effect  of  surgical  wound  on  ovarian  carcinoma  growth  in  an  animal  model.  

Anticancer  Res,  2013.  33(8):  p.  3177-­‐84.  71.   Lee,  J.W.,  et  al.,  Surgical  stress  promotes  tumor  growth  in  ovarian  carcinoma.  Clin  Cancer  Res,  

2009.  15(8):  p.  2695-­‐702.  72.   Allendorf,  J.D.,  et  al.,  Increased  tumor  establishment  and  growth  after  open  vs  laparoscopic  

bowel  resection  in  mice.  Surg  Endosc,  1998.  12(8):  p.  1035-­‐8.  73.   Abramovitch,  R.,  et  al.,  Stimulation  of  tumour  growth  by  wound-­‐derived  growth  factors.  Br  J  

Cancer,  1999.  79(9-­‐10):  p.  1392-­‐8.  74.   Belizon,  A.,  et  al.,  Major  abdominal  surgery  increases  plasma  levels  of  vascular  endothelial  

growth  factor:  open  more  so  than  minimally  invasive  methods.  Ann  Surg,  2006.  244(5):  p.  792-­‐8.  75.   Ramos-­‐Jimenez,  J.,  et  al.,  Histamine  augments  beta2-­‐adrenoceptor-­‐induced  cyclic  AMP  

accumulation  in  human  prostate  cancer  cells  DU-­‐145  independently  of  known  histamine  receptors.  Biochem  Pharmacol,  2007.  73(6):  p.  814-­‐23.  

76.   Zhang,  P.,  et  al.,  beta-­‐arrestin2  mediates  beta-­‐2  adrenergic  receptor  signaling  inducing  prostate  cancer  cell  progression.  Oncology  reports,  2011.  26(6):  p.  1471-­‐7.  

77.   Palm,  D.,  et  al.,  The  norepinephrine-­‐driven  metastasis  development  of  PC-­‐3  human  prostate  cancer  cells  in  BALB/c  nude  mice  is  inhibited  by  beta-­‐blockers.  Int  J  Cancer,  2006.  118(11):  p.  2744-­‐9.  

78.   Hassan,  S.,  et  al.,  Behavioral  stress  accelerates  prostate  cancer  development  in  mice.  J  Clin  Invest,  2013.  123(2):  p.  874-­‐86.  

79.   Ramberg,  H.,  et  al.,  Hormonal  regulation  of  beta2-­‐adrenergic  receptor  level  in  prostate  cancer.  Prostate,  2008.  68(10):  p.  1133-­‐42.  

80.   Prowatke,  I.,  et  al.,  Expression  analysis  of  imbalanced  genes  in  prostate  carcinoma  using  tissue  microarrays.  Br  J  Cancer,  2007.  96(1):  p.  82-­‐8.  

81.   Pippig,  S.,  et  al.,  Overexpression  of  beta-­‐arrestin  and  beta-­‐adrenergic  receptor  kinase  augment  desensitization  of  beta  2-­‐adrenergic  receptors.  J  Biol  Chem,  1993.  268(5):  p.  3201-­‐8.  

82.   Grytli,  H.H.,  et  al.,  Use  of  beta-­‐blockers  is  associated  with  prostate  cancer-­‐specific  survival  in  prostate  cancer  patients  on  androgen  deprivation  therapy.  Prostate,  2013.  73(3):  p.  250-­‐60.  

83.   Lin,  Q.,  et  al.,  Effect  of  chronic  restraint  stress  on  human  colorectal  carcinoma  growth  in  mice.  PLoS  One,  2013.  8(4):  p.  e61435.  

84.   Hicks,  B.M.,  et  al.,  beta-­‐Blocker  usage  and  colorectal  cancer  mortality:  a  nested  case-­‐control  study  in  the  UK  Clinical  Practice  Research  Datalink  cohort.  Ann  Oncol,  2013.  24(12):  p.  3100-­‐6.  

85.   Moretti,  S.,  et  al.,  beta-­‐adrenoceptors  are  upregulated  in  human  melanoma  and  their  activation  releases  pro-­‐tumorigenic  cytokines  and  metalloproteases  in  melanoma  cell  lines.  Lab  Invest,  2013.  93(3):  p.  279-­‐90.  

86.   Guo,  Y.,  et  al.,  Interleukin-­‐6  signaling  pathway  in  targeted  therapy  for  cancer.  Cancer  Treat  Rev,  2012.  38(7):  p.  904-­‐10.  

87.   Waugh,  D.J.  and  C.  Wilson,  The  interleukin-­‐8  pathway  in  cancer.  Clin  Cancer  Res,  2008.  14(21):  p.  6735-­‐41.  

88.   Yang,  E.V.,  et  al.,  Norepinephrine  upregulates  VEGF,  IL-­‐8,  and  IL-­‐6  expression  in  human  melanoma  tumor  cell  lines:  implications  for  stress-­‐related  enhancement  of  tumor  progression.  Brain  Behav  Immun,  2009.  23(2):  p.  267-­‐75.  

89.   De  Giorgi,  V.,  et  al.,  beta-­‐adrenergic-­‐blocking  drugs  and  melanoma:  current  state  of  the  art.  Expert  Rev  Anticancer  Ther,  2012.  12(11):  p.  1461-­‐7.  

  62  

90.   De  Giorgi,  V.,  et  al.,  Effect  of  beta-­‐blockers  and  other  antihypertensive  drugs  on  the  risk  of  melanoma  recurrence  and  death.  Mayo  Clin  Proc,  2013.  88(11):  p.  1196-­‐203.  

91.   Kozanoglu,  I.,  et  al.,  New  indication  for  therapeutic  potential  of  an  old  well-­‐known  drug  (propranolol)  for  multiple  myeloma.  J  Cancer  Res  Clin  Oncol,  2013.  139(2):  p.  327-­‐35.  

92.   Yang,  E.V.,  et  al.,  VEGF  is  differentially  regulated  in  multiple  myeloma-­‐derived  cell  lines  by  norepinephrine.  Brain  Behav  Immun,  2008.  22(3):  p.  318-­‐23.  

93.   Shand,  D.G.,  Pharmacokinetics  of  propranolol:  a  review.  Postgrad  Med  J,  1976.  52  Suppl  4:  p.  22-­‐25.  

94.   Mills,  P.C.,  G.A.  Siebert,  and  M.S.  Roberts,  A  model  to  study  intestinal  and  hepatic  metabolism  of  propranolol  in  the  dog.  J  Vet  Pharmacol  Ther,  2004.  27(1):  p.  45-­‐8.  

95.   Akin,  A.,  et  al.,  The  efficacy  of  amiodarone-­‐propranolol  combination  for  the  management  of  childhood  arrhythmias.  Pacing  Clin  Electrophysiol,  2013.  36(6):  p.  727-­‐31.  

96.   Hedera,  P.,  F.  Cibulcik,  and  T.L.  Davis,  Pharmacotherapy  of  Essential  Tremor.  J  Cent  Nerv  Syst  Dis,  2013.  5:  p.  43-­‐55.  

97.   Schneider,  S.A.  and  G.  Deuschl,  The  treatment  of  tremor.  Neurotherapeutics,  2014.  11(1):  p.  128-­‐38.  

98.   Hepp,  Z.,  L.M.  Bloudek,  and  S.F.  Varon,  Systematic  review  of  migraine  prophylaxis  adherence  and  persistence.  J  Manag  Care  Pharm,  2014.  20(1):  p.  22-­‐33.  

99.   Shamliyan,  T.A.,  et  al.,  Preventive  pharmacologic  treatments  for  episodic  migraine  in  adults.  J  Gen  Intern  Med,  2013.  28(9):  p.  1225-­‐37.  

100.   Hruska,  B.,  P.K.  Cullen,  and  D.L.  Delahanty,  Pharmacological  modulation  of  acute  trauma  memories  to  prevent  PTSD:  Considerations  from  a  developmental  perspective.  Neurobiol  Learn  Mem,  2014.  

101.   Fitzgerald,  P.J.,  J.R.  Seemann,  and  S.  Maren,  Can  fear  extinction  be  enhanced?  A  review  of  pharmacological  and  behavioral  findings.  Brain  Res  Bull,  2013.  

102.   Tawa,  J.  and  S.  Murphy,  Psychopharmacological  treatment  for  military  posttraumatic  stress  disorder:  an  integrative  review.  J  Am  Assoc  Nurse  Pract,  2013.  25(8):  p.  419-­‐23.  

103.   de  Kleine,  R.A.,  B.O.  Rothbaum,  and  A.  van  Minnen,  Pharmacological  enhancement  of  exposure-­‐based  treatment  in  PTSD:  a  qualitative  review.  Eur  J  Psychotraumatol,  2013.  4.  

104.   Searcy,  C.P.,  et  al.,  Pharmacological  prevention  of  combat-­‐related  PTSD:  a  literature  review.  Mil  Med,  2012.  177(6):  p.  649-­‐54.  

105.   Black,  J.W.,  et  al.,  A  NEW  ADRENERGIC  BETARECEPTOR  ANTAGONIST.  Lancet,  1964.  1(7342):  p.  1080-­‐1.  

106.   Basu,  B.,  et  al.,  A  Case  of  Akathisia  induced  by  Escitalopram:  Case  Report  &  Review  of  Literature.  Curr  Drug  Saf,  2014.  9(1):  p.  56-­‐9.  

107.   Fitzgerald,  P.J.,  Is  elevated  norepinephrine  an  etiological  factor  in  some  cases  of  schizophrenia?  Psychiatry  Res,  2014.  215(3):  p.  497-­‐504.  

108.   Hogan,  S.R.,  J.  Mandrell,  and  D.  Eilers,  Adrenergic  urticaria:  Review  of  the  literature  and  proposed  mechanism.  J  Am  Acad  Dermatol,  2013.  

109.   Finnerty,  C.C.  and  D.N.  Herndon,  Is  propranolol  of  benefit  in  pediatric  burn  patients?  Adv  Surg,  2013.  47:  p.  177-­‐97.  

110.   Herndon,  D.N.,  et  al.,  Long-­‐term  propranolol  use  in  severely  burned  pediatric  patients:  a  randomized  controlled  study.  Ann  Surg,  2012.  256(3):  p.  402-­‐11.  

111.   Olah,  G.,  et  al.,  Increased  poly(ADP-­‐ribosyl)ation  in  skeletal  muscle  tissue  of  pediatric  patients  with  severe  burn  injury:  prevention  by  propranolol  treatment.  Shock,  2011.  36(1):  p.  18-­‐23.  

112.   Rojas,  Y.,  et  al.,  Burns:  an  update  on  current  pharmacotherapy.  Expert  Opin  Pharmacother,  2012.  13(17):  p.  2485-­‐94.  

  63  

113.   Nandhra,  H.S.,  C.L.  Murphy,  and  A.  Sule,  Novel  pharmacological  agents  targeting  memory  and  cognition  in  the  treatment  of  anxiety  disorders.  Hum  Psychopharmacol,  2013.  

114.   Altamura,  A.C.,  et  al.,  Understanding  the  pharmacokinetics  of  anxiolytic  drugs.  Expert  Opin  Drug  Metab  Toxicol,  2013.  9(4):  p.  423-­‐40.  

115.   Lindgren,  M.E.,  et  al.,  Beta-­‐blockers  may  reduce  intrusive  thoughts  in  newly  diagnosed  cancer  patients.  Psychooncology,  2013.  22(8):  p.  1889-­‐94.  

116.   Lawley,  L.P.,  E.  Siegfried,  and  J.L.  Todd,  Propranolol  treatment  for  hemangioma  of  infancy:  risks  and  recommendations.  Pediatr  Dermatol,  2009.  26(5):  p.  610-­‐4.  

117.   Pavlakovic,  H.,  et  al.,  Hyperkalemia  complicating  propranolol  treatment  of  an  infantile  hemangioma.  Pediatrics,  2010.  126(6):  p.  e1589-­‐93.  

118.   Abbott,  J.,  et  al.,  Diarrhea  associated  with  propranolol  treatment  for  hemangioma  of  infancy  (HOI).  Pediatr  Dermatol,  2010.  27(5):  p.  558.  

119.   Giron-­‐Vallejo,  O.,  et  al.,  Dental  caries  as  a  side  effect  of  infantile  hemangioma  treatment  with  propranolol  solution.  Pediatr  Dermatol,  2010.  27(6):  p.  672-­‐3.  

120.   Bonifazi,  E.,  et  al.,  Severe  hypoglycemia  during  successful  treatment  of  diffuse  hemangiomatosis  with  propranolol.  Pediatr  Dermatol,  2010.  27(2):  p.  195-­‐6.  

121.   McBride,  J.T.,  M.C.  McBride,  and  P.H.  Viles,  Hypoglycemia  associated  with  propranolol.  Pediatrics,  1973.  51(6):  p.  1085-­‐7.  

122.   Holland,  K.E.,  et  al.,  Hypoglycemia  in  children  taking  propranolol  for  the  treatment  of  infantile  hemangioma.  Arch  Dermatol,  2010.  146(7):  p.  775-­‐8.  

123.   Kwon,  E.K.,  et  al.,  Retrospective  review  of  adverse  effects  from  propranolol  in  infants.  JAMA  Dermatol,  2013.  149(4):  p.  484-­‐5.  

124.   Volmer,  P.A.,  Human  Drugs  of  Abuse,  in  Kirk's  Current  Veterinary  Therapy  XIV,  J.D.B.D.C.  Twedt,  Editor.  2009,  Saunders  Co.:  Philadelphia.  p.  144-­‐145.  

125.   Wright,  K.N.,  Assessment  and  Treatment  of  Supraventricular  Tachyarrhythmias,  in  Kirk's  Current  Veterinary  Therapy  XIV,  J.D.B.D.C.  Twedt,  Editor.  2009,  Saunders  Co:  Philadelphia.  p.  722-­‐727.  

126.   Schober,  K.E.,  Myocarditis,  in  Kirk's  Current  Veterinary  Therapy  XIV,  J.D.B.D.C.  Twedt,  Editor.  2009,  Saunders  Co:  Philadelphia.  p.  804-­‐808.  

127.   Kates,  R.E.,  B.W.  Keene,  and  R.L.  Hamlin,  Pharmacokinetics  of  propranolol  in  the  dog.  Journal  of  Veterinary  Pharmacology  and  Therapeutics,  1979.  2(1):  p.  21-­‐26.  

128.   Trepanier,  L.A.,  Medical  Treatment  of  Feline  Hyperthyroidism,  in  Kirk's  Current  Veterinary  Therapy  XIV,  J.D.B.D.C.  Twedt,  Editor.  2009,  Saunders  Co:  Philadelphia.  p.  175-­‐179.  

129.   Westropp,  I.F.L.J.L.,  Urinary  Incontinence  and  Micturition  Disorders:  Pharmacologic  Management,  in  Kirk's  Current  Veterinary  Therapy  XIV,  J.D.B.D.C.  Twedt,  Editor.  2009,  Saunders  Co.:  Philadelphia.  p.  955.  

130.   Dana  G.  Allen,  J.K.P.,  Dale  Smith,  Handbook  of  Veterinary  Drugs.  1993.  131.   Feng,  X.M.,  et  al.,  Preparation  and  evaluation  of  a  novel  delayed-­‐onset  sustained-­‐release  system  

of  propranolol  hydrochloride.  J  Pharm  Pharmacol,  2008.  60(7):  p.  817-­‐22.  132.   Papich,  M.G.,  Table  of  Common  Drugs:  Approximate  Dosages,  in  Kirk's  Current  Veterinary  

Therapy  XIV,  J.D.B.D.C.  Twedt,  Editor.  2009,  Saunders  Co:  Philadelphia.  p.  1329.  133.   Plumb,  Plumb's  Veterinary  Drug  Handbook.  Vol.  6th.  2008.  134.   Fonseca  Junior,  N.L.,  et  al.,  [Therapeutical  effectiveness  of  interferon  alpha  in  a  child  with  

craniofacial  giant  hemangioma:  case  report].  Arq  Bras  Oftalmol,  2008.  71(3):  p.  423-­‐6.  135.   Enjolras,  O.,  et  al.,  [Vincristine  treatment  for  function-­‐  and  life-­‐threatening  infantile  

hemangioma].  Arch  Pediatr,  2004.  11(2):  p.  99-­‐107.  136.   Stiles,  J.,  et  al.,  Propranolol  treatment  of  infantile  hemangioma  endothelial  cells:  A  molecular  

analysis.  Exp  Ther  Med,  2012.  4(4):  p.  594-­‐604.  

  64  

137.   Drolet,  B.A.,  et  al.,  Initiation  and  use  of  propranolol  for  infantile  hemangioma:  report  of  a  consensus  conference.  Pediatrics,  2013.  131(1):  p.  128-­‐40.  

138.   Shah,  S.  and  I.J.  Frieden,  Treatment  of  infantile  hemangiomas  with  beta-­‐blockers:  a  review.  Skin  Therapy  Lett,  2013.  18(6):  p.  5-­‐7.  

139.   Frieden,  I.J.,  et  al.,  Infantile  hemangiomas:  current  knowledge,  future  directions.  Proceedings  of  a  research  workshop  on  infantile  hemangiomas,  April  7-­‐9,  2005,  Bethesda,  Maryland,  USA.  Pediatr  Dermatol,  2005.  22(5):  p.  383-­‐406.  

140.   Haggstrom,  A.N.,  et  al.,  Measuring  the  severity  of  infantile  hemangiomas:  instrument  development  and  reliability.  Arch  Dermatol,  2012.  148(2):  p.  197-­‐202.  

141.   Boye,  E.  and  B.R.  Olsen,  Signaling  mechanisms  in  infantile  hemangioma.  Curr  Opin  Hematol,  2009.  16(3):  p.  202-­‐8.  

142.   Grimmer,  J.F.,  et  al.,  Familial  clustering  of  hemangiomas.  Arch  Otolaryngol  Head  Neck  Surg,  2011.  137(8):  p.  757-­‐60.  

143.   Khan,  Z.A.,  et  al.,  Multipotential  stem  cells  recapitulate  human  infantile  hemangioma  in  immunodeficient  mice.  J  Clin  Invest,  2008.  118(7):  p.  2592-­‐9.  

144.   Guimaraes,  S.  and  D.  Moura,  Vascular  adrenoceptors:  an  update.  Pharmacol  Rev,  2001.  53(2):  p.  319-­‐56.  

145.   Wolter,  N.E.,  et  al.,  Propranolol  as  a  novel  adjunctive  treatment  for  head  and  neck  squamous  cell  carcinoma.  J  Otolaryngol  Head  Neck  Surg,  2012.  41(5):  p.  334-­‐44.  

146.   Neufeld,  G.,  et  al.,  Vascular  endothelial  growth  factor  (VEGF)  and  its  receptors.  Faseb  j,  1999.  13(1):  p.  9-­‐22.  

147.   Ji,  Y.,  et  al.,  Effects  of  propranolol  on  the  proliferation  and  apoptosis  of  hemangioma-­‐derived  endothelial  cells.  J  Pediatr  Surg,  2012.  47(12):  p.  2216-­‐23.  

148.   Zhao,  Z.F.,  et  al.,  [The  change  of  serum  vascular  endothelial  growth  factor  and  matrix  metalloproteinases-­‐9  in  proliferative  hemangioma  treated  with  propranolol].  Zhonghua  Zheng  Xing  Wai  Ke  Za  Zhi,  2011.  27(5):  p.  359-­‐61.  

149.   Chim,  H.,  et  al.,  Propranolol  induces  regression  of  hemangioma  cells  through  HIF-­‐1alpha-­‐mediated  inhibition  of  VEGF-­‐A.  Ann  Surg,  2012.  256(1):  p.  146-­‐56.  

150.   Schieven,  G.L.,  The  biology  of  p38  kinase:  a  central  role  in  inflammation.  Curr  Top  Med  Chem,  2005.  5(10):  p.  921-­‐8.  

151.   Bamburg,  J.R.,  Proteins  of  the  ADF/cofilin  family:  essential  regulators  of  actin  dynamics.  Annu  Rev  Cell  Dev  Biol,  1999.  15:  p.  185-­‐230.  

152.   Street,  C.A.  and  B.A.  Bryan,  Rho  kinase  proteins-­‐-­‐pleiotropic  modulators  of  cell  survival  and  apoptosis.  Anticancer  Res,  2011.  31(11):  p.  3645-­‐57.  

153.   Kelly,  B.D.,  et  al.,  Cell  type-­‐specific  regulation  of  angiogenic  growth  factor  gene  expression  and  induction  of  angiogenesis  in  nonischemic  tissue  by  a  constitutively  active  form  of  hypoxia-­‐inducible  factor  1.  Circ  Res,  2003.  93(11):  p.  1074-­‐81.  

154.   Kleinman,  M.E.,  et  al.,  Hypoxia-­‐induced  mediators  of  stem/progenitor  cell  trafficking  are  increased  in  children  with  hemangioma.  Arterioscler  Thromb  Vasc  Biol,  2007.  27(12):  p.  2664-­‐70.  

155.   Greenberger,  S.,  et  al.,  Targeting  NF-­‐kappaB  in  infantile  hemangioma-­‐derived  stem  cells  reduces  VEGF-­‐A  expression.  Angiogenesis,  2010.  13(4):  p.  327-­‐35.  

156.   Bagazgoitia,  L.,  A.  Hernandez-­‐Martin,  and  A.  Torrelo,  Recurrence  of  infantile  hemangiomas  treated  with  propranolol.  Pediatr  Dermatol,  2011.  28(6):  p.  658-­‐62.  

157.   Chang,  L.,  et  al.,  Recurrence  of  infantile  hemangioma  after  termination  of  propranolol  treatment.  Ann  Plast  Surg,  2014.  72(2):  p.  173-­‐5.  

158.   Shehata,  N.,  et  al.,  Late  rebound  of  infantile  hemangioma  after  cessation  of  oral  propranolol.  Pediatr  Dermatol,  2013.  30(5):  p.  587-­‐91.  

  65  

159.   Wong,  A.,  et  al.,  Propranolol  accelerates  adipogenesis  in  hemangioma  stem  cells  and  causes  apoptosis  of  hemangioma  endothelial  cells.  Plast  Reconstr  Surg,  2012.  130(5):  p.  1012-­‐21.  

160.   Zhang,  L.,  et  al.,  Propranolol  inhibits  angiogenesis  via  down-­‐regulating  the  expression  of  vascular  endothelial  growth  factor  in  hemangioma  derived  stem  cell.  Int  J  Clin  Exp  Pathol,  2013.  7(1):  p.  48-­‐55.  

161.   Zou,  H.X.,  et  al.,  Propranolol  inhibits  endothelial  progenitor  cell  homing:  a  possible  treatment  mechanism  of  infantile  hemangioma.  Cardiovasc  Pathol,  2013.  22(3):  p.  203-­‐10.  

162.   Rostami,  M.,  et  al.,  Tumors  in  domestic  animals  examined  during  a  ten-­‐year  period  (1980  to  1989)  at  Miyazaki  University.  J  Vet  Med  Sci,  1994.  56(2):  p.  403-­‐5.  

163.   Day,  M.J.,  V.M.  Lucke,  and  H.  Pearson,  A  review  of  pathological  diagnoses  made  from  87  canine  splenic  biopsies.  J  Small  Anim  Pract,  1995.  36(10):  p.  426-­‐33.  

164.   Spangler,  W.L.  and  P.H.  Kass,  Pathologic  factors  affecting  postsplenectomy  survival  in  dogs.  J  Vet  Intern  Med,  1997.  11(3):  p.  166-­‐71.  

165.   Schultheiss,  P.C.,  A  retrospective  study  of  visceral  and  nonvisceral  hemangiosarcoma  and  hemangiomas  in  domestic  animals.  J  Vet  Diagn  Invest,  2004.  16(6):  p.  522-­‐6.  

166.   Oksanen,  A.,  Haemangiosarcoma  in  dogs.  J  Comp  Pathol,  1978.  88(4):  p.  585-­‐95.  167.   Arp,  L.H.  and  R.L.  Grier,  Disseminated  cutaneous  hemangiosarcoma  in  a  young  dog.  J  Am  Vet  

Med  Assoc,  1984.  185(6):  p.  671-­‐3.  168.   Brown,  N.O.,  A.K.  Patnaik,  and  E.G.  MacEwen,  Canine  hemangiosarcoma:  retrospective  analysis  

of  104  cases.  J  Am  Vet  Med  Assoc,  1985.  186(1):  p.  56-­‐8.  169.   Srebernik,  N.  and  E.C.  Appleby,  Breed  prevalence  and  sites  of  haemangioma  and  

haemangiosarcoma  in  dogs.  Vet  Rec,  1991.  129(18):  p.  408-­‐9.  170.   Moe,  L.,  et  al.,  Canine  neoplasia-­‐-­‐population-­‐based  incidence  of  vascular  tumours.  APMIS  Suppl,  

2008(125):  p.  63-­‐8.  171.   Hargis,  A.M.,  et  al.,  A  retrospective  clinicopathologic  study  of  212  dogs  with  cutaneous  

hemangiomas  and  hemangiosarcomas.  Vet  Pathol,  1992.  29(4):  p.  316-­‐28.  172.   Ward,  H.,  et  al.,  Cutaneous  hemangiosarcoma  in  25  dogs:  a  retrospective  study.  J  Vet  Intern  

Med,  1994.  8(5):  p.  345-­‐8.  173.   Ware,  W.A.  and  D.L.  Hopper,  Cardiac  tumors  in  dogs:  1982-­‐1995.  J  Vet  Intern  Med,  1999.  13(2):  

p.  95-­‐103.  174.   Aupperle,  H.,  et  al.,  Primary  and  secondary  heart  tumours  in  dogs  and  cats.  J  Comp  Pathol,  2007.  

136(1):  p.  18-­‐26.  175.   Goritz,  M.,  et  al.,  Canine  splenic  haemangiosarcoma:  influence  of  metastases,  chemotherapy  and  

growth  pattern  on  post-­‐splenectomy  survival  and  expression  of  angiogenic  factors.  J  Comp  Pathol,  2013.  149(1):  p.  30-­‐9.  

176.   Sabattini,  S.  and  G.  Bettini,  An  immunohistochemical  analysis  of  canine  haemangioma  and  haemangiosarcoma.  J  Comp  Pathol,  2009.  140(2-­‐3):  p.  158-­‐68.  

177.   Yonemaru,  K.,  et  al.,  Expression  of  vascular  endothelial  growth  factor,  basic  fibroblast  growth  factor,  and  their  receptors  (flt-­‐1,  flk-­‐1,  and  flg-­‐1)  in  canine  vascular  tumors.  Vet  Pathol,  2006.  43(6):  p.  971-­‐80.  

178.   Clifford,  C.A.,  et  al.,  Plasma  vascular  endothelial  growth  factor  concentrations  in  healthy  dogs  and  dogs  with  hemangiosarcoma.  J  Vet  Intern  Med,  2001.  15(2):  p.  131-­‐5.  

179.   Clifford,  C.A.,  et  al.,  Vascular  endothelial  growth  factor  concentrations  in  body  cavity  effusions  in  dogs.  J  Vet  Intern  Med,  2002.  16(2):  p.  164-­‐8.  

180.   Kato,  Y.,  et  al.,  Gene  expressions  of  canine  angiopoietin-­‐1  and  -­‐2  in  normal  tissues  and  spontaneous  tumours.  Res  Vet  Sci,  2006.  81(2):  p.  280-­‐6.  

181.   Asa,  S.A.,  et  al.,  Expression  of  platelet-­‐derived  growth  factor  and  its  receptors  in  spontaneous  canine  hemangiosarcoma  and  cutaneous  hemangioma.  Histol  Histopathol,  2012.  27(5):  p.  601-­‐7.  

  66  

182.   Mayr,  B.,  et  al.,  Tumour  suppressor  gene  p53  mutation  in  a  case  of  haemangiosarcoma  of  a  dog.  Acta  Vet  Hung,  2002.  50(2):  p.  157-­‐60.  

183.   Yonemaru,  K.,  et  al.,  The  significance  of  p53  and  retinoblastoma  pathways  in  canine  hemangiosarcoma.  J  Vet  Med  Sci,  2007.  69(3):  p.  271-­‐8.  

184.   Tamburini,  B.A.,  et  al.,  Gene  expression  profiling  identifies  inflammation  and  angiogenesis  as  distinguishing  features  of  canine  hemangiosarcoma.  BMC  Cancer,  2010.  10:  p.  619.  

185.   Dickerson,  E.B.,  et  al.,  Mutations  of  phosphatase  and  tensin  homolog  deleted  from  chromosome  10  in  canine  hemangiosarcoma.  Vet  Pathol,  2005.  42(5):  p.  618-­‐32.  

186.   Murakami,  M.,  et  al.,  Expression  of  the  anti-­‐apoptotic  factors  Bcl-­‐2  and  survivin  in  canine  vascular  tumours.  J  Comp  Pathol,  2008.  139(1):  p.  1-­‐7.  

187.   Bertazzolo,  W.,  et  al.,  Canine  angiosarcoma:  cytologic,  histologic,  and  immunohistochemical  correlations.  Vet  Clin  Pathol,  2005.  34(1):  p.  28-­‐34.  

188.   Gamlem,  H.  and  K.  Nordstoga,  Canine  vascular  neoplasia-­‐-­‐histologic  classification  and  inmunohistochemical  analysis  of  221  tumours  and  tumour-­‐like  lesions.  APMIS  Suppl,  2008(125):  p.  19-­‐40.  

189.   Jakab,  C.,  et  al.,  Claudin-­‐5  protein  is  a  new  differential  marker  for  histopathological  differential  diagnosis  of  canine  hemangiosarcoma.  Histol  Histopathol,  2009.  24(7):  p.  801-­‐13.  

190.   Thamm,  D.H.,  Miscellaneous  tumors  (hemangiosarcoma),  in  Withrow  and  MacEwen's  Small  Animal  Clinical  Oncology,  S.J.W.D.M.V.R.L.  Page,  Editor.  2013,  Saunders  Co.  p.  679-­‐688.  

191.   Hammer,  A.S.,  et  al.,  Efficacy  and  toxicity  of  VAC  chemotherapy  (vincristine,  doxorubicin,  and  cyclophosphamide)  in  dogs  with  hemangiosarcoma.  J  Vet  Intern  Med,  1991.  5(3):  p.  160-­‐6.  

192.   Alvarez,  F.J.,  et  al.,  VAC  protocol  for  treatment  of  dogs  with  stage  III  hemangiosarcoma.  J  Am  Anim  Hosp  Assoc,  2013.  49(6):  p.  370-­‐7.  

193.   Sorenmo,  K.,  et  al.,  Canine  hemangiosarcoma  treated  with  standard  chemotherapy  and  minocycline.  J  Vet  Intern  Med,  2000.  14(4):  p.  395-­‐8.  

194.   Sorenmo,  K.U.,  et  al.,  Efficacy  and  toxicity  of  a  dose-­‐intensified  doxorubicin  protocol  in  canine  hemangiosarcoma.  J  Vet  Intern  Med,  2004.  18(2):  p.  209-­‐13.  

195.   Ogilvie,  G.K.,  et  al.,  Surgery  and  doxorubicin  in  dogs  with  hemangiosarcoma.  J  Vet  Intern  Med,  1996.  10(6):  p.  379-­‐84.  

196.   Ogilvie,  G.K.,  et  al.,  Phase  II  evaluation  of  doxorubicin  for  treatment  of  various  canine  neoplasms.  J  Am  Vet  Med  Assoc,  1989.  195(11):  p.  1580-­‐3.  

197.   Wiley,  J.L.,  et  al.,  Efficacy  of  doxorubicin-­‐based  chemotherapy  for  non-­‐resectable  canine  subcutaneous  haemangiosarcoma.  Vet  Comp  Oncol,  2010.  8(3):  p.  221-­‐33.  

198.   Payne,  S.E.,  et  al.,  Treatment  of  vascular  and  soft-­‐tissue  sarcomas  in  dogs  using  an  alternating  protocol  of  ifosfamide  and  doxorubicin.  Vet  Comp  Oncol,  2003.  1(4):  p.  171-­‐9.  

199.   Rassnick,  K.M.,  et  al.,  Evaluation  of  ifosfamide  for  treatment  of  various  canine  neoplasms.  J  Vet  Intern  Med,  2000.  14(3):  p.  271-­‐6.  

200.   Kim,  S.E.,  et  al.,  Epirubicin  in  the  adjuvant  treatment  of  splenic  hemangiosarcoma  in  dogs:  59  cases  (1997-­‐2004).  J  Am  Vet  Med  Assoc,  2007.  231(10):  p.  1550-­‐7.  

201.   Sorenmo,  K.,  et  al.,  Clinical  and  pharmacokinetic  characteristics  of  intracavitary  administration  of  pegylated  liposomal  encapsulated  doxorubicin  in  dogs  with  splenic  hemangiosarcoma.  J  Vet  Intern  Med,  2007.  21(6):  p.  1347-­‐54.  

202.   Vail,  D.M.,  et  al.,  Liposome-­‐encapsulated  muramyl  tripeptide  phosphatidylethanolamine  adjuvant  immunotherapy  for  splenic  hemangiosarcoma  in  the  dog:  a  randomized  multi-­‐institutional  clinical  trial.  Clin  Cancer  Res,  1995.  1(10):  p.  1165-­‐70.  

203.   Hillers,  K.R.,  et  al.,  Effects  of  palliative  radiation  therapy  on  nonsplenic  hemangiosarcoma  in  dogs.  J  Am  Anim  Hosp  Assoc,  2007.  43(4):  p.  187-­‐92.  

  67  

204.   Lana,  S.,  et  al.,  Continuous  low-­‐dose  oral  chemotherapy  for  adjuvant  therapy  of  splenic  hemangiosarcoma  in  dogs.  J  Vet  Intern  Med,  2007.  21(4):  p.  764-­‐9.  

205.   Kahn,  S.A.,  et  al.,  Doxorubicin  and  deracoxib  adjuvant  therapy  for  canine  splenic  hemangiosarcoma:  a  pilot  study.  Can  Vet  J,  2013.  54(3):  p.  237-­‐42.  

206.   Gardner,  H.L.L.,  C.A;  Portela,  R.F.;  Nguyen,  S.;  Rosenberg,  M.P.;  Klein,  M.K.;  Clifford,  C.;  Thamm,  D.H.;  Vail,  D.M;  Bergman,  P.J.;  Crawford-­‐Jakubiak,  M.;  Henry,  C.;  Locke,  J.;  Garrett,  L.D.;  Cronin,  K.L.,  Maintenance  therapy  with  toceranib  following  doxorubicin-­‐based  chemotherapy  for  canine  splenic  hemangiosarcoma.  Manuscript  submitted  for  publication.,  2014.  

207.   Prymak,  C.,  et  al.,  Epidemiologic,  clinical,  pathologic,  and  prognostic  characteristics  of  splenic  hemangiosarcoma  and  splenic  hematoma  in  dogs:  217  cases  (1985).  J  Am  Vet  Med  Assoc,  1988.  193(6):  p.  706-­‐12.  

208.   Wood,  C.A.,  et  al.,  Prognosis  for  dogs  with  stage  I  or  II  splenic  hemangiosarcoma  treated  by  splenectomy  alone:  32  cases  (1991-­‐1993).  J  Am  Anim  Hosp  Assoc,  1998.  34(5):  p.  417-­‐21.  

209.   Dunning,  D.,  et  al.,  Analysis  of  prognostic  indicators  for  dogs  with  pericardial  effusion:  46  cases  (1985-­‐1996).  J  Am  Vet  Med  Assoc,  1998.  212(8):  p.  1276-­‐80.  

210.   Aronsohn,  M.,  Cardiac  hemangiosarcoma  in  the  dog:  a  review  of  38  cases.  J  Am  Vet  Med  Assoc,  1985.  187(9):  p.  922-­‐6.  

211.   Szivek,  A.,  et  al.,  Clinical  outcome  in  94  cases  of  dermal  haemangiosarcoma  in  dogs  treated  with  surgical  excision:  1993-­‐2007*.  Vet  Comp  Oncol,  2012.  10(1):  p.  65-­‐73.  

212.   Burton,  J.H.,  B.E.  Powers,  and  B.J.  Biller,  Clinical  outcome  in  20  cases  of  lingual  hemangiosarcoma  in  dogs:  1996-­‐2011.  Vet  Comp  Oncol,  2012.  

213.   Shiu,  K.B.,  et  al.,  Predictors  of  outcome  in  dogs  with  subcutaneous  or  intramuscular  hemangiosarcoma.  J  Am  Vet  Med  Assoc,  2011.  238(4):  p.  472-­‐9.  

214.   Stiles,  J.M.,  et  al.,  Targeting  of  beta  adrenergic  receptors  results  in  therapeutic  efficacy  against  models  of  hemangioendothelioma  and  angiosarcoma.  PLoS  One,  2013.  8(3):  p.  e60021.  

215.   Rada,  T.,  L.  Okruhlicova,  and  J.  Slezak,  Immunohistochemical  localization  of  beta-­‐adrenergic  receptors  in  the  heart.  Bratisl  Lek  Listy,  1991.  92(3-­‐4):  p.  138-­‐41.  

216.   Vroon,  A.,  et  al.,  Taxol  normalizes  the  impaired  agonist-­‐induced  beta2-­‐adrenoceptor  internalization  in  splenocytes  from  GRK2+/-­‐  mice.  Eur  J  Pharmacol,  2007.  560(1):  p.  9-­‐16.  

217.   Neubauer,  B.,  et  al.,  Renin  expression  in  large  renal  vessels  during  fetal  development  depends  on  functional  beta1/beta2-­‐adrenergic  receptors.  Am  J  Physiol  Renal  Physiol,  2011.  301(1):  p.  F71-­‐7.  

218.   Ferrer,  L.,  et  al.,  Immunohistochemical  detection  of  CD31  antigen  in  normal  and  neoplastic  canine  endothelial  cells.  J  Comp  Pathol,  1995.  112(4):  p.  319-­‐26.  

219.   Slotkin,  T.A.,  et  al.,  Beta-­‐adrenoceptor  signaling  and  its  control  of  cell  replication  in  MDA-­‐MB-­‐231  human  breast  cancer  cells.  Breast  Cancer  Res  Treat,  2000.  60(2):  p.  153-­‐66.  

220.   Tallman,  J.F.,  C.C.  Smith,  and  R.C.  Henneberry,  Induction  of  functional  beta-­‐adrenergic  receptors  in  HeLa  cells.  Proc  Natl  Acad  Sci  U  S  A,  1977.  74(3):  p.  873-­‐7.  

221.   Liang,  C.C.,  A.Y.  Park,  and  J.L.  Guan,  In  vitro  scratch  assay:  a  convenient  and  inexpensive  method  for  analysis  of  cell  migration  in  vitro.  Nat  Protoc,  2007.  2(2):  p.  329-­‐33.  

222.   Avni,  R.,  B.  Cohen,  and  M.  Neeman,  Hypoxic  stress  and  cancer:  imaging  the  axis  of  evil  in  tumor  metastasis.  NMR  Biomed,  2011.  24(6):  p.  569-­‐81.  

223.   Coussens,  L.M.  and  Z.  Werb,  Inflammation  and  cancer.  Nature,  2002.  420(6917):  p.  860-­‐7.  224.   Ramos-­‐Vara,  J.A.,  Technical  aspects  of  immunohistochemistry.  Vet  Pathol,  2005.  42(4):  p.  405-­‐

26.  225.   Atkins,  D.,  et  al.,  Immunohistochemical  detection  of  EGFR  in  paraffin-­‐embedded  tumor  tissues:  

variation  in  staining  intensity  due  to  choice  of  fixative  and  storage  time  of  tissue  sections.  J  Histochem  Cytochem,  2004.  52(7):  p.  893-­‐901.  

  68  

226.   Xiao,  K.  and  S.K.  Shenoy,  Beta2-­‐adrenergic  receptor  lysosomal  trafficking  is  regulated  by  ubiquitination  of  lysyl  residues  in  two  distinct  receptor  domains.  J  Biol  Chem,  2011.  286(14):  p.  12785-­‐95.  

227.   Boivin,  B.,  et  al.,  Functional  beta-­‐adrenergic  receptor  signalling  on  nuclear  membranes  in  adult  rat  and  mouse  ventricular  cardiomyocytes.  Cardiovasc  Res,  2006.  71(1):  p.  69-­‐78.  

228.   Chen,  D.,  et  al.,  The  beta2-­‐adrenergic  receptor  is  a  potential  prognostic  biomarker  for  human  hepatocellular  carcinoma  after  curative  resection.  Ann  Surg  Oncol,  2012.  19(11):  p.  3556-­‐65.  

229.   Yang,  E.V.,  et  al.,  Norepinephrine  up-­‐regulates  the  expression  of  vascular  endothelial  growth  factor,  matrix  metalloproteinase  (MMP)-­‐2,  and  MMP-­‐9  in  nasopharyngeal  carcinoma  tumor  cells.  Cancer  Res,  2006.  66(21):  p.  10357-­‐64.  

230.   Hajighasemi  F.  ,  M.A.,  Propranolol  effect  on  proliferation  and  vascular  endothelial  growth  factor  secretion  in  human  immunocompetent  cells.  Journal  of  Clinical  Immunology  and  Immunopathology  Research,  2010.  2(2):  p.  22-­‐27.  

231.   Fujio,  H.,  et  al.,  Carvedilol  inhibits  proliferation  of  cultured  pulmonary  artery  smooth  muscle  cells  of  patients  with  idiopathic  pulmonary  arterial  hypertension.  J  Cardiovasc  Pharmacol,  2006.  47(2):  p.  250-­‐5.  

232.   Cardani,  R.,  et  al.,  Influence  of  beta-­‐adrenergic  antagonists  on  cell  proliferation  rates  in  the  kidney  of  untreated  and  diethylnitrosamine-­‐treated  male  F344  rats.  Chem  Biol  Interact,  1999.  118(3):  p.  217-­‐31.  

233.   Arber,  S.,  et  al.,  Regulation  of  actin  dynamics  through  phosphorylation  of  cofilin  by  LIM-­‐kinase.  Nature,  1998.  393(6687):  p.  805-­‐9.  

234.   Schreiner,  S.J.,  A.P.  Schiavone,  and  T.E.  Smithgall,  Activation  of  STAT3  by  the  Src  family  kinase  Hck  requires  a  functional  SH3  domain.  J  Biol  Chem,  2002.  277(47):  p.  45680-­‐7.  

235.   Lefkowitz,  R.J.,  E.  Haber,  and  D.  O'Hara,  Identification  of  the  cardiac  beta-­‐adrenergic  receptor  protein:  solubilization  and  purification  by  affinity  chromatography.  Proc  Natl  Acad  Sci  U  S  A,  1972.  69(10):  p.  2828-­‐32.