ROLE OF THE IMMUNE SYSTEM IN TUMOR PROGRESSION 1

35 Different tumors from the same mouse have different immune infiltrates ...35 Granulocytes are increased in the periphery during tumor progression ...39 CXCL1, 2 and 5 and CCL19 are up

ROLE OF THE IMMUNE SYSTEM IN TUMOR

NUS Graduate School for Integrative Sciences and Engineering

NATIONAL UNIVERSITY OF SINGAPORE

2011

Acknowledgements

I would like to acknowledge my supervisor, Jean-Pierre Abastado, for his continued help and support throughout the course of my PhD He has been a great mentor and guide from whom I have learnt a great deal I would also like to thank my thesis advisory committee members, Ren Ee Chee and Laurent Renia, for their guidance and help

I would like to show my appreciation to all members of the Tumor Immunology Laboratory in the Singapore Immunology Network (SIgN) Special thanks go to Karen Khoo, Jeremy Wang and Jo Keeble for their help and discussions about the work presented in this thesis

Next, I would like to thank my collaborators for their help in this project Laurent Renia (SIgN) for the NIMP-R14 antibody, Ng Lai Guan (SIgN) for providing me with

the IL8R-KO and tdTomato mice, Esther Koh for help in the cell tracking software,

Poon Lai Fong (SIgN) for help with cell sorting, Josephine Lum (SIgN) for the microarray work and Wong Wing Cheong (BII) for the analysis of the microarray data Special thanks to Masashi Kato (Chubu University, Aichi, Japan) and Armelle Prevost-Blondel (Institut Cochin, Paris, France) for providing the RET mice I would also like to thank Jean-Paul Thiery (IMCB) and Sim Wen Jing (IMCB) for their help and discussion regarding the EMT assays

Last but not least, I would like to thank my family for their support, especially my wife, who painstakingly helped with the editing of this thesis

Table of Contents

Acknowledgements i

Summary v

List of Tables vi

List of Figures vii

List of Appendices ix

List of Videos ix

List of Publications x

List of Abbreviations xi

Introduction 1

Inflammation, immunity and cancer 1

Immunosurveillance theory 2

Immunoediting theory 3

Roles of immune cells 6

Melanoma 9

Diagnosis and treatment 11

Melanoma and the immune system 15

Objectives 16

Experimental Procedures 18

Mice 18

In vivo PMN-MDSCs depletion 18

Flow cytometry analysis 19

Isolation of PMN-MDSCs and macrophages 20

Microarray analysis 21

Cytospin and May-Grunwald/Giemsa stain 22

Tumor cell detection by qRT-PCR 23

Low density microarray 23

Immunohistochemistry and calculations 24

Migration assay 25

Tumor proliferation assay 26

OVA-specific T cell proliferation assay 27

E-Cadherin assays 27

MT assay 28

Statistics 29

Chapter 1: Not All Tumors Are Made Equal 30

Introduction 30

RETAAD – Spontaneous mouse model of melanoma 31

Results 35

Different tumors from the same mouse have different immune infiltrates 35

Granulocytes are increased in the periphery during tumor progression 39

CXCL1, 2 and 5 and CCL19 are up-regulated in primary tumors 41

CXCR2 ligands are able to attract PMN-MDSCs 42

CXCR2 is necessary for the attraction of PMN-MDSCs to the tumor in vivo 43

CXCL1 and 2 are expressed by the PMN-MDSCs while CXCL5 is expressed in tumor cells 44

Discussion 46

Chapter 2: Role of PMN-MDSCs in Tumor Progression 53

Introduction 53

Myeloid-derived suppressor cells 53

Metastasis and epithelial-mesenchymal transition 62

Results 65

Depletion of PMN-MDSCs reduces tumor growth in vivo 65

PMN-MDSCs promote cancer cell proliferation in the primary tumor 67

PMN-MDSCs secrete soluble factors that promote cancer cell proliferation in vitro 69

PMN-MDSCs favor multinodular development of primary tumors 70

PMN-MDSCs induce cancer cell dissemination to regional and distant sites 71

PMN-MDSCs favor metastatic outgrowth 73

PMN-MDSCs induce cancer cell MT in vitro 74

PMN-MDSCs induce down-regulation of epithelial marker,

E-Cadherin, in melanoma cells 77

PMN-MDSCs induce melanoma cell MT in vivo 78

Tumor cells express EGF, while PMN-MDSCs express HGF and TGF-β1 81

PMN-MDSCs induce MT through multiple pathways 82

Discussion 84

Induced proliferation of cancers cells by PMN-MDSCs 84

PMN-MDSCs induce MT in cancer cells 86

Chapter 3: Significance and Implications 92

MT, metastasis and phenotype-switching 92

Multiple roles of PMN-MDSCs 101

Tumor dormancy 105

Therapeutic implications 109

Lymph node excisions 109

Inflammation 112

PMN-MDSCs 114

Chemotherapy and immunotherapy 118

Conclusion 121

Bibliography 124

Appendices……… A

Summary

In order to metastasize, cancer cells need to acquire a motile phenotype Previously, development of this phenotype was thought to rely on the acquisition of selected, random mutations and thus occur late in cancer progression However, recent studies show that cancer cells disseminate early, implying the existence of a different, faster route to the metastatic motile phenotype Using a spontaneous murine model of melanoma, I show that a subset of bone marrow-derived immune cells (myeloid-derived suppressor cells or MDSCs) preferentially infiltrates the primary tumor and actively promotes cancer cell dissemination by inducing mesenchymal transition

(MT) In vitro and in vivo assays using purified MDSCs showed attraction of MDSCs

to the primary tumor is CXCR2-dependent and that TGF-β, EGF and HGF signaling pathways are all used by MDSCs to induce MT in cancer cells These findings explain how cancer cells acquire a motile phenotype early and provide a mechanistic explanation for the long recognized link between inflammation and cancer progression

List of Tables

Table 1: Roles of different immune cell subsets in cancer 7

Table 2: The ‘ABCD’ method of identifying early melanoma lesions 12

Table 3: List of antibodies used 20

Table 4: Summary of the characteristics and differences of MDSCs 61

Table 5: Summary of the characteristics and differences of MDSCs including the present research 91

List of Figures

Figure 1: Illustration of immunoediting theory .5Figure 2: Overall analysis of tumors .35Figure 3: Differential accumulation and morphology of macrophages and

granulocytes .37Figure 4: Comparison of tumors by immune subset .38Figure 5: Accumulation of granulocytes in tumor bearing mice .40Figure 6: Differential expression of chemokines and cytokines between

primary tumor and cutaneous metastases .41Figure 7: Migration of PMN-MDSCs to CXCR2 ligands .43Figure 8: CXCR2 ligands are important and necessary for PMN-MDSCs

migration to the primary tumor .44Figure 9: Expression of CXCR2 and its ligands in PMN-MDSCs and

tumor cells .45Figure 10: Depletion of intra-tumoral PMN-MDSCs .66Figure 11: Depletion of PMN-MDSCs reduces tumor growth but not

tumor vasculature .67Figure 12: Depletion of PMN-MDSCs reduces proliferation in young eye

tumors .68

Figure 13: PMN-MDSCs induce proliferation of tumor cells in vitro

through soluble factors .70Figure 14: Depletion of PMN-MDSCs reduces nodular structure of the

primary tumor .71

Figure 15: Depletion of PMN-MDSCs reduce metastasis to lung and

lymph nodes .72

Figure 16: Depletion of PMN-MDSCs reduces number of cutaneous metastases but not their size .73

Figure 17: PMN-MDSCs induce MT in vitro .75

Figure 18: PMN-MDSCs down-regulate E-Cadherin expression .77

Figure 19: S100A4 expression in primary tumors .80

Figure 20: Differential expression of TGF-β1, HGF and EGF in PMN-MDSCs and tumor cells .81

Figure 21: Inhibition of PMN-MDSC induced mesenchymal transition .82

Figure 22: Diagram illustrating the differences between the two models of tumor progression .96

Figure 23: Diagram illustrating the progression of tumors in RETAAD mice .99

Figure 24: In silico modeling of tumor growth in non-vascularised tumors .103

Figure 25: Graph illustrating that micrometastases occur before detection of the primary tumor .106

Figure 26: Illustration of interactions between PMN-MDSCs and tumor cells .122

List of Videos

Video 1: Time lapse of NBT-II cells without stimulation

Video 2: Time lapse of NBT-II cells with EGF

Video 3: Time lapse of NBT-II cells with PMN-MDSCs

List of Publications

1 Toh B, Wang X, Keeble J, Sim WJ, Khoo K, Wong WC, Kato M,

Prevost-Blondel A, Thiery JP and Abastado JP Mesenchymal Transition and Dissemination of Cancer Cells is driven by Myeloid-Derived Suppressor Cells Infiltrating the Primary Tumor PLoS Biol 9(9): e1001162 doi:10.1371/journal.pbio.1001162

2 Toh B, Nardin A, Dai X, Keeble J, Chew V and Abastado JP Detection, enumeration and characterization of immune cells infiltrating melanoma

tumors Chapter for “Molecular Methods in Dermatology” Humana Press,

USA (under Review)

3 Liang Zhi, Benjamin Toh, and Jean-Pierre Abastado Myeloid derived suppressor cells: subsets, expansion, and role in cancer progression Chapter

for “Tumor Microenvironment and Myelomonocytic Cells” ISBN:

979-953-307-100-7 Intech (under Review)

4 Bourgault-VilladaI, Hong M, KhooK, Tham M, Toh B, Wai LE and Abastado

JP (In Press) Current insight into the metastatic process and melanoma cell

dissemination Chapter for “Melanoma” ISBN: 978-953-307-293-7 Intech

(Accepted on March 23, 2011)

5 Eyles J, Puaux AL, Wang X, Toh B, Prakash C, Hong M, Yan TG, Zheng L, Ong LC, Jin Y, Kato M, Prevost-Blondel A, Chow P, Yang H and Abastado

JP (2010) Tumor cells disseminate early, but immunosurveillance limits

metastatic outgrowth, in a mouse model of melanoma J Clin Invest 120:

2030-2039

ECM = Extracellular matrix

EGF = Epidermal growth factor

EMT = Epithelial mesenchymal transition

MDSC = Myeloid-derived suppressor cells

MET = Mesenchymal epithelial transition

qRT-PCR = Quantitative real-time polymerase chain reaction

RNA = Ribonucleic acid

ROS = Reactive oxygen species

TAM = Tumor associated macrophages

TGF = Transforming growth factor

TNF = Tumor necrosis factor

Tregs = Regulatory T cells

TRP = Tyrosinase related protein

UV = Ultraviolet

Introduction

Inflammation, immunity and cancer

Tumors do not consist of a homogeneous population of cells; rather they are a composite of the cancer cells, mesenchymal and endothelial cells, and immune cell populations The link between cancer progression and inflammation was recognized

by R Virchow in the late 19th century [1,2,3,4,5] Now, the role of inflammation in tumorigenesis is generally accepted and an inflammatory microenvironment is an essential component in all tumors [3,4,6,7]

Inflammation is the response to an infection or tissue injury, whereby complex networks of chemical signals initiate and maintain a host response to ‘heal’ the afflicted or infected tissue [8] Acute inflammation is the initial early response of the body against harmful stimuli Chronic inflammation occurs when the body is unable

to resolve an acute inflammation or when inflammation itself has a slow onset However it starts, chronic inflammation can last for months and years Inflammation

is not beneficial all the time It has been shown that chronic inflammation is a significant risk factor in cancer development and tumor progression Up to 20% of cancers are directly linked to chronic infections, 30% can be attributed to tobacco smoking and inhaled pollutants (such as silica and asbestos), and 35% to dietary factors (20% of which includes obesity) [3] All these initiating events are able to

induce inflammation Persistent infections by Helicobacter pylori and human

papilloma virus have been identified as major risk factors in gastric cancer and cervical cancer respectively [9] Inflammation in these cases is actually part of the

normal host response to pathogens However, these microbes can subvert this host response to establish a chronic inflammatory environment for their own benefit [3] Interestingly, inducing acute inflammation can be effective in treating some cancers

like the treatment of superficial bladder cancer with Mycobacterium bovis bacillus

Calmette–Guérin (BCG) instillation [10] Chronic inflammatory diseases have also been linked to cancer i.e inflammatory bowel syndrome and colorectal cancer [11] Despite these associations between the host immune responses and cancer development, the exact molecular mechanisms and complex interactions between cells of the tumor microenvironment have yet to be fully understood

Immunosurveillance theory

In cancer, it has been proposed that the immune system plays a key role in preventing the natural occurrence of malignant cells, seen as foreign bodies, in humans This concept was first proposed in the 1900s by Paul Ehrlich However, it was only in

1957, with a deeper understanding of the immune system, that Burnet and Thomas formally proposed the hypothesis of cancer immunosurveillance This was defined by

Burnet as follows: “In large, long-lived animals, like most of the warm-blooded vertebrates, inheritable genetic changes must be common in somatic cells and a proportion of these changes will represent a step toward malignancy It is an evolutionary necessity that there should be some mechanism for eliminating or inactivating such potentially dangerous mutant cells and it is postulated that this mechanism is of immunological character [12]” The role that the immune system

plays in controlling cancer development can be seen in transplant patients These patients are under immunosuppressive drugs for prolonged periods to prevent

rejection of their transplant organ A 4-fold increase in incidence of in situ melanoma

can be found in patients that had undergone transplants [13] In a study of 162 patients that received transplants, melanoma constituted 5.2% of post-transplant skin cancers

in patients compared with the incidence of 2.7% in the general population [14] Transplant patients were also three times more likely to develop non-Kaposi’s sarcomas [15] In a study of 608 cardiac transplant patients, the prevalence of lung tumors was 25-fold higher than in the general population [16] In another study of 5,692 patients that had undergone renal transplants, 2- to 5-fold excess risks were seen for cancers of the colon, lung, bladder, kidney, ureter, endocrine tumors and malignant melanomas as opposed to the general population [17] Immunosuppression increases the risk of cancer, even for those of no known viral etiology, as seen in the examples above Of course, in cancer cases with known viral etiology, immunosuppression also results in an increased risk of cancer [13]

Immunoediting theory

However, the immunosurveillance theory only explains one half of the story The immunoediting theory proposed by Schreiber [18,19] is the updated version of previous theories that now takes into account evidence of tumor escape mechanisms

In this theory, shown in Figure 1, it is proposed that there are three distinct sequential stages of tumor progression – 1) elimination, 2) equilibrium and 3) escape 1) The elimination stage is an updated version of the immunosurveillance theory This is the stage in which innate and adaptive immune systems work to detect and destroy an incipient tumor before it becomes clinically apparent Examples that support this have been stated in the preceding paragraphs However, since this elimination event takes

place without clinical symptoms, this stage has never been directly observed in vivo

The next stage of tumor progression is 2) equilibrium At this stage, the immune

system, comprising tumor-specific CD8+ and CD4+ T cells and natural killer (NK) cells, is able to control the growth of the tumor This is typically the longest of all the three stages and is characterized by dormancy of the tumor [19] Dormancy refers to a state of cellular quiescence with cells in G0–G1 arrest [20] Latency may also occur during this period Latency is the perceived dormancy of the tumor due to the equal rates of proliferation and elimination of the cancer cells Latency is important to Schreiber’s immunoediting theory because proliferation of the tumor cells is required for the cells to gain sufficient mutations for escape Thus, during equilibrium, the tumors do not grow as the adaptive immune system prevents further tumor cell outgrowth Schreiber postulates that rare tumor cell variants may survive the elimination and enter the equilibrium stage He also postulates that due to the selection pressure of antigen-specific T cells, tumor cells that are not immunogenic are preferentially selected If the immune system is successful in controlling tumor development, this could be the stable end stage for some patients, whereby the tumor

is present but under tumor specific control [19]

The final stage is 3) escape In this stage, the escape from immune control can occur

in two ways – a tumor cell-intrinsic acquired mechanism that evades recognition by the immune system or the establishment of an immunosuppressive state within the tumor microenvironment The acquisition of an ‘escape’ phenotype is believed to be the end result of tumor cell population changes in response to the immune system’s selection or editing These include phenotypic changes such as down-regulation MHC Class I expression, disruption in the antigen processing machinery, down-regulation

of tumor antigens or up-regulation of anti-apoptotic genes such as Bcl-2 [21] These escaping tumor cells may also gain the ability to modulate the host immune system,

increasing cancer-induced immunosuppression [19] For example, the secretion of indoleamine 2, 3 dioxygenase (IDO) can deprive the microenvironment of tryptophan, inducing apoptosis of effector T cells [21] Up-regulation of FasL on tumor cells can also cause T cell apoptosis through Fas/FasL pathway [22]

Figure 1: Illustration of immunoediting theory

In Scheiber’s model of immunoediting [19], tumor progression occurs in three progressive and distinct stages – elimination, equilibrium and escape Newly formed tumors are eradicated by the immune system effectively most of the time However, in certain instances, rare tumor cells might evade this immune system mediated destruction and persist In these cases, tumor-specific mechanisms, such as CD8+ and CD4+ tumor-specific T cells are activated in the immune systems to control these tumors Immunogenic tumor cells are then eliminated from the system while non-immunogenic tumor cells are selected for and remain in equilibrium with the immune system in a state of dormancy When these dormant tumor cells have gained sufficient mutations to circumvent the immune system, they emerge

from dormancy and progress rapidly in the tissue [Taken from Scheiber (2011)] [19]

Roles of immune cells

Almost all immune cells have been implicated in both protective and suppressive roles during cancer progression (Table 1) Generally, macrophages and T cells are the most numerous and they are the most studied immune cell types in the tumor microenvironment In humans, intratumoral T cell infiltrate has commonly been associated with a good prognosis in several cancers such as colorectal cancer, hepatocellular carcinoma and melanoma [23,24,25] However, it has been shown that

T cells subsets, CD4+ or CD8+, Th1 (rarely) or Th2, and regulatory T cells (Tregs) have tumor promoting roles In mice with chemically-induced skin cancer, CD8+ T cells have been found to secrete IFN-γ, TNF-α, and cyclooxygenase-2 that contribute

to inflammation and cancer progression [26] IL-13 secreted by CD4+ Th2 T cells have also been shown to promote tumor progression in mice [27] Regulatory T cells have constantly been implicated in tumor progression due to their ability to suppress

an effective anti-tumoral immune response In humans, elevated levels of infiltrating regulatory T cells are correlated with a reduced overall survival in numerous cancers [28] T cells play an important role in tumor development They shape the immune response by secreting cytokines and one of the important immune subsets they can affect is the macrophage subset

tumor-Immune Cell Types Anti-Tumor Tumor Promoting

12 and type I IFN)

Immunosuppression Production of cytokines, chemokines, proteases, growth factors, and angiogenic factors

antibodies

Production of cytokines Activation of mast cells Immunosuppression Inflammation

Production of cytotoxic cytokines

Production of cytokines

macrophages Production of cytokines

B cell activation

lymphocytes (CTLs) in tumor rejection

Production of cytokines (IFNγ)

Production of cytokines

Inflammation

(cytokines and other suppressive mechanisms)

Immunosuppression Production of cytokines

cancer cells Production of cytotoxic cytokines

cancer cells Production of cytotoxic cytokines

Regulation of CTL responses

Production of cytokines, proteases, and ROS

Table 1: Roles of different immune cell subsets in cancer

Immune cell can play paradoxical roles in cancer development depending on the environment and the polarization of the different subsets Different cancers induce different subsets of immune cells and some immune cells can play an important role in one type of cancer and not another The roles and functions of each subset are not fixed These can be altered during the course of the disease or during

treatment [Taken from Grivennikov et al (2010)][3]

Tumor associated macrophages (TAMs) play an important role in tumor development Macrophages are generally split into two subsets – M1 and M2; with TAMs

supposedly being of the M2-like subset The M2 subset is polarized in vitro by IL-4,

IL-10 and IL-13 [29] They down-regulate major histocompatibility complex (MHC) class II and IL-12 expression and show increased expression of the anti-inflammatory cytokine IL-10, scavenger receptor A, and arginase (Arg) These macrophages are mainly involved in wound healing processes and angiogenesis [29] This is in contrast with M1 macrophages which are more inflammatory and are activated by IFNγ and toll-like receptor (TLR) agonists These macrophages express high levels of pro-inflammatory cytokines like TNF-α, IL-1, IL-6, IL-12 and inducible nitric oxide synthase They are the main macrophages involved in killing pathogens and priming anti-tumor immune responses [29] However, it is now recognized that, although useful, segregation of the macrophage into distinct subsets is an over-simplification [30] This is due to the plasticity of the macrophages and their ability to differentiate along a continuous spectrum of phenotypes that lie between M1 and M2 However, what is consistent among TAMs, to a certain degree, is that they seem to primarily favor tumor growth They are capable of creating a suppressive immune environment through Arg expression They secrete angiogenic factors such as vascular endothelial growth factor (VEGF) and epidermal growth factor (EGF) that help in tumor angiogenesis TAMs also express matrix metalloproteases (MMP) like MMP2 that help break down extracellular matrix and (1) release ECM bound factors such as TGF-β and/or (2) allow metastasis of tumor cells These functions vary depending on the type of tumor and the location of the TAMs in the tumor e.g TAMs in hypoxic

areas of the tumor express more angiogenic factors compared to TAMs in other parts

of the tumor [8,30]

In summary, inflammation and the immune system play critical roles in both tumor suppression and tumor progression Understanding the complex interactions that occur within the tumor microenvironment is key to re-activating the immune system against the tumor However, we must also realize that not all tumor microenvironments are the same Tumors can occur throughout the body Each organ has its own vasculature system, lymphatic drainage and even organ-specific immune cells Tumor cells can further modulate this environment with its own secretions, making the microenvironment unique Therefore, the study of the cross-talk between cancer cells and the immune system must be analyzed in the context of the organ in which the tumor exists and the type of cells that the tumor recruits

Melanoma

In Singapore, cancer mortality rates are on the rise Cancer now accounts for more than a quarter of all deaths in Singapore, as compared to 14.8% in 1968-1973 [31] Skin cancer is the most common site of cancer development in humans, accounting for almost half of the new cancers cases in the United States of America [32] Skin cancers, including melanoma, are the 8th most common cancers in males and the 9thmost common in females in Singapore [31]

Melanoma is the malignancy of melanocytes – neural crest-derived pigmented cells that produce melanin Melanoma is generally found on the skin but can be found in any organ that contains melanocytes i.e eye (uveal tract), mouth, nasal cavities,

oesophagus, stomach, rectum and anus [31] In Singapore, melanomas form 6% of all skin cancers and 27% of all ocular cancers [31] but account for 75% of all skin cancer deaths [33] Worldwide, the incidence of melanoma has been climbing since the 1930s with a rate of increase higher than any type of cancer [32] Fair-skinned individuals are more at risk of melanoma Australia and Israel have the highest incidence of melanoma in the world (approximately 40 in 100,000 individuals per year) [32] In Singapore, the incidence is much lower, with only 106 cases diagnosed locally between 2003-2007 [31] Even though melanoma incidence is lower in darker-skinned populations, it is important to recognize that it does occur in these populations too Melanoma can be cured if excised early However, once it has disseminated to distant organs, the median survival of melanoma patients drops below nine months [34]

Two genes have been associated with an increased risk of melanoma –

cyclin-dependent kinase inhibitor 2A (CDKN2A) and cyclin-cyclin-dependent kinase 4 (CDK4) 10%

of melanoma patients have at least two first-degree relatives with melanoma This, in turn, corresponds with a 2-fold increase in the risk of melanoma [32] Ultra-violet (UV) radiation is also a major risk factor in melanoma development UV radiation causes genetic changes in the skin, impairs immune functions, increases local growth factor production and induces reactive-oxygen species production that affects keratinocytes and melanocytes [35,36] The incidence of melanoma among Caucasians is inversely related to latitude of residence [37], implying higher exposure

to the sun in tropical/sub-tropical areas as a risk factor UV B radiation (290 to 320 nm) is the cause of most of the DNA damage in the skin while UV A radiation (320 to

400 nm) is responsible for the oxidative damage and immunosuppression Indeed,

splenic immune cells transferred from mice exposed to UV A radiation are not able to reject UV-induced skin tumors [38] Melanomas are associated with intense intermittent exposure in contrast to other skin cancers like basal cell or squamous cell carcinoma which are associated with chronic UV exposure [35] Upon intense UV exposure, damaged melanocytes, unlike keratinocytes, do not undergo apoptosis but rather survive, mutate and divide [35] This division can be seen extrinsically in the form of freckles that appear in children after a bout of sun exposure These freckles are thought to be the clones of mutated dividing melanocytes and are associated with the risk of melanoma [39] Moles are benign tumors of melanocytes that, after an initial period of growth, have undergone oncogene-induced senescence [40]

Diagnosis and treatment

Melanoma can be detected early Early melanoma development can be assessed by observing asymmetry, border irregularity, color and diameter of the lesion (Table 2) Suspected melanoma lesions are excised with narrow margins of 1-3mm Histological evaluation is then carried out on these excised lesions to determine the extent of invasion and the stage of the disease Melanomas generally have two growth phases – the radial growth phase and the vertical growth phase In the radial growth phase, the tumor cells are in the epidermis and papillary dermis with raised irregular growth at the surface During the vertical growth phase, the melanoma tumor grows to the deeper layers of skin with increasing nodularity The tumors in both the radial and vertical phases are considered to be invasive However, only the vertical growth phase and the depth of invasion correlate with prognosis [32]

Feature Benign Mole Melanoma

Table 2: The ‘ABCD’ method of identifying early melanoma lesions

Recommended method of differentiating early melanoma from benign moles Other signs can include

itching, bleeding ulceration or changes in pre-existing benign mole [Adapted from Jensen et al

(2007)][32]

Staging of melanoma is determined by the TMN classification of melanoma T represents the primary tumor stage where thickness and ulceration of the primary tumor is assessed The thickness or invasiveness of the tumor is measured in two ways The first method is known as Clark’s levels whereby the invasion of the tumors

is determined by the histologically defined layers of the skin However, this can be inaccurate as the thickness of the different layers of the skin can vary between individuals The Breslow’s thickness is the alternate method of measurement This method measures the distance from the epidermal granular layer to the base of the tumor There is an inverse correlation between Breslow’s thickness and survival N refers to the extent of lymph node metastases found The presence of lymph node metastases is a bad sign regardless of the invasiveness of the primary tumor The number of lymph nodes detected with melanoma cells and the size of the metastases (microscopic or macroscopic) help determine the aggressiveness of the disease M represents the distant metastases of melanoma cells to other organs Stage I and II patients have no lymph node or distant metastases As such, severity is determined by the Breslow’s thickness and Clark’s levels In stage III patients, the number of nodes and metastases are prognosis factors Stage IV patients have distant metastasis Factors such as serum lactic dehydrogenase levels and visceral sites of metastases are the prognosis factors [32]

Cutaneous melanoma is treated primarily by excision of the primary tumor as it is readily accessable An excision margin around the tumor is always recommended but the thickness of the tumor determines the actual size of the margin taken [41] However, for melanomas with a thickness of 2mm or less, excision margins of 2 or 5cm had no difference in local reoccurrence [42] For melanomas in other parts of the body such as the eye, treatment options include enucleation, radiation therapy, laser hyperthermia and surgical resection Although treatment of the primary tumor is generally successful, hematogenous spreading is known to occur This hematogenous spreading of the tumor cells can be found in almost half of melanoma patients [43] Patients with intermediate or thick primary melanomas (Breslow’s depth above 1mm) are advised to undergo sentinel lymph node biopsy (SLNB) [44] The sentinel lymph node (SLN) is defined as the node that first drains the lymph from the primary tumor site [45] Injection of both a radioactive colloid and a blue dye is done around the primary tumor to identify the SLN Hematoxylin and eosin staining, as well as immunohistochemistry, are done on the biopsied lymph node to determine the presence of melanoma cells S100B, HMB45 and Melan-A/MART1 are used as markers for melanoma cells with the latter two used more for their specificity and S100B for its sensitivity Reverse transcriptase polymerase chain reaction is now emerging as a molecular staging tool for missed micrometastases Tyrosinase, tyrosinase related protein (Trp)-2, Melan-A/MART-1 as well as other markers are used to detect melanoma-specific transcripts [46,47,48] Tyrosinase and Trp-2 both play important roles in melanogenesis and are specifically expressed in melanocytes and melanoma [49] In cases of metastatic cells found in the lymph nodes, complete

regional lymph node dissection is recommended [32], although this practice is controversial

In advanced melanoma, median survival is only two to eight months with only 5% surviving more than five years [50] Contributing to this high mortality rate is the fact that melanoma is often resistant to most chemotherapeutic agents Dacarbazine (DTIC) remains one of the most effective therapies against metastatic melanoma However, the effect of this treatment is still limited with response rates of 15-20% and median response duration of five to six months Even then, less than 5% of these responses are complete responses Long term treatment with DTIC shows that less than 2% of patients survive for more than six years [51] Cisplatin is another drug with modest activity on metastatic melanoma with a 15% response rate and median duration of only three months [51] Generally, combinations of various chemotherapies have also shown no advantage to DTIC alone in terms of response or survival [52] However, recent studies have shown that Fotemustine has a better overall response rate than DTIC (15.2% vs 6.8%) with a trend for better overall survival [53] This treatment option is also slightly more conveniant than DTIC as it requires weekly injections of Fotemustine instead of daily injections of DTIC [51] Immunotherapies such as IFN-α and IL-2 treatment have also been used in metastatic melanoma In the case of IL-2, due to the high dose required to elicit a clinical response, only a select group of patients are able to withstand the toxicity that comes along with high-dose IL-2 therapy IFN-α has a modest effect on melanoma with approximately 15% responses with a duration between six to nine months Less than 5% of patients show complete response Recently, Ipilimumab, an anti-human CTLA-4, treatment have been shown

to improve overall survival of metastatic melanoma patients [54] Combinations of

DTIC with Ipilimumab have been shown to improve response rates and survival [55] Combined treatments of chemotherapy and immunotherapy have clearly shown to improve the treatment of metastatic melanoma This is because melanoma is one of the cancers that show a strong association with the host immune system

Melanoma and the immune system

Melanomas are immunogenic tumors that evade the immune system Tumor infiltrating lymphocytes (TILs) have been correlated to better prognosis and a better five-year survival rate [56] However, the prognostic value is only valid at an early stage of melanoma; in thick lesions this prognostic value is lost TILs isolated from melanoma patients are able to lyse MHC-matched allogeneic tumors [57,58] In fact, most melanoma-specific antigens are non-mutated peptides from proteins involved in melanin synthesis such as MelanA/MART-1, tyrosinase related protein (TRP) -1, TRP-2, gp100 and tyrosinase [59] Unfortunately, there has been limited success in vaccination of patients with these antigens [60] MelanA/MART-1 specific T cells have been found in the blood and tumors of melanoma patients However, only the T cells found in the blood were able to produce IFNγ and granzyme B when stimulated with antigens, but not the T cells found in the tumor [61] This illustrates local immunosuppression at the tumor site This immunosuppression is likely to be caused

by immune cells recruited into the tumor Lymphodepletion has been shown to be effective in enhancing adoptive T cell transfer therapy in melanoma patients [62] In a clinical trial testing the effectiveness of adoptive T cell transfer with lymphodepleting NonMyeloablative Chemotherapy (NMC), better objective responses and complete remission were obtained when NMC was combined with total body irradiation or high

dose irradiation alone [63] This implies that the immunosuppressive effect does not arise from lymphocytes alone but also myeloid cells

There are four reasons why melanoma is an ideal cancer to study interactions between the immune system and cancer – 1) it can be responsive to the immune system, 2) a large number of the antigens have been studied, 3) melanoma tumors share common antigens and 4) tumors are accessible It is clear that melanoma is a cancer that has close interactions with the immune system Its ability to evade the immune system is dependent on components of the immune system itself Much research has been done

on immunotherapy focusing on enhancing T cell effectiveness against the tumor More research is required to uncover further interactions between the melanoma cells and the immune systems Mechanisms would also have to be identified such that therapeutic options can be devised to provide better treatment for the patient

Objectives

In this thesis, I use a unique mouse model that spontaneously develops melanoma, RETAAD, to investigate the interactions between tumor cells and immune cells In Chapter 1, I investigate the tumor immune infiltrates in RETAAD mice I find that the immune infiltrates are all different in three different subtypes of tumors – primary tumors, cutaneous metastases and visceral metastases Interestingly, polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) preferentially accumulated in the primary tumor I find that this is due to the differential expression

of chemokines CXCL1, CXCL2 and CXCL5 in the primary tumor that specifically attracted the PMN-MDSCs In Chapter 2, I delve deeper into the role PMN-MDSCs play in tumor progression Notably, PMN-MDSCs are able to affect two important

aspects of tumor progression – tumor growth and metastasis PMN-MDSCs are able

to directly induce tumor cell proliferation by secreting a soluble factor PMN-MDSCs are also able to induce epithelial-mesenchymal transition, the first step in the metastatic processes, at an early stage in cancer These are two unique roles for PMN-MDSCs that had not been previously described In Chapter 3, I shall discuss the significance and implications of these findings and the impact they would have on therapy

Experimental Procedures

Mice

Animal care and experimental procedures were approved by the IACUC of the Biological Resource Center, 20 Biopolis Way, Singapore 138668 The generation of RETAAD mice was previously described [64] In this particular transgenic line (304/B6 mice), RET expression is driven by the metallothionein gene promoter and is expressed by melanocytes [65] RET mice were crossed once with AAD mice to yield tumor-bearing RETAAD mice

For the in vivo migration experiments, Rosa mT/mG reporter mice expressing tdTomato and Il8rb-KO mice were obtained from JAX Laboratories (Cat No 007576 and 006848 respectively) Il8rb-KO mice were crossed with mice expressing EGFP

under the Lysozyme promoter [66] to obtain Gr1hi cells lacking CXCR2 and expressing EGFP

In vivo PMN-MDSCs depletion

PMN-MDSCs were depleted using rat anti-mouse Ly6G depleting antibody, R14 [67], isolated from hybridoma cultures (provided by Renia L., Singapore Immunology Network, A-STAR) Mice were injected intraperitoneally with 0.25 mg

NIMP-of anti-Ly6G antibody or control rat IgG (Sigma-Aldrich) at one or five weeks NIMP-of age, continuing twice a week until seven and 20 weeks of age respectively The efficiency

of PMN-MDSCs depletion was monitored by flow cytometry analysis of circulating blood Mice were clinically assessed once per fortnight for palpable tumors On the

day of sacrifice, mice were examined and measured for superficial and internal macroscopic tumors

For the early depletion studies, primary tumors were collected in 10% neutral buffered formalin with all other palpable tumors collected in RNAlater (Qiagen) For the late depletion studies, all tumors were collected and individually analyzed by flow cytometry for the immune cell subsets Tumors that were too small for flow cytometric analyses were collected in RNALater (Qiagen) In both experiments, one half of the lung was analyzed by IHC, and the other half by qRT-PCR

Flow cytometry analysis

Tumors were dissected from RETAAD mice and cut up into small pieces in RPMI medium without serum Single cell suspensions were obtained by incubating the tumor pieces for 20mins at 37°C in digestion buffer consisting of RPMI medium, Collagenase A (1 mg/ml; Roche) and DNase I (0.1 mg/ml, Roche), constantly stirring with a magnetic stirrer bar Tumor cells were filtered through a 70µm filter (BD Biosciences) and red blood cell lysis was carried out using red blood cell lysis buffer (155 mM NH4Cl, 10 mM KHCO3, 0.1 mM EDTA) Fc receptors were blocked with CD16/CD32 antibodies (eBioscience) before antibody labeling was carried out on ice for 30mins Antibodies used are listed in Table 3 Countbright beads (Invitrogen, C36950) were added when an accurate cell count was needed Calculations were as such:

𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨 𝒏𝒏𝑨𝑨 𝑨𝑨𝒐𝒐 𝒄𝒄𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨 𝒊𝒊𝒏𝒏 𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨

=𝑁𝑁𝑁𝑁 𝑁𝑁𝑜𝑜 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑝𝑝𝑏𝑏𝑝𝑝 𝑡𝑡𝑡𝑡𝑏𝑏𝑏𝑏𝑁𝑁𝑁𝑁 𝑁𝑁𝑜𝑜 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑐𝑐𝑁𝑁𝑡𝑡𝑐𝑐𝑡𝑡𝑏𝑏𝑏𝑏 × 𝑁𝑁𝑁𝑁 𝑁𝑁𝑜𝑜 𝑐𝑐𝑏𝑏𝑐𝑐𝑐𝑐𝑏𝑏 𝑐𝑐𝑁𝑁𝑡𝑡𝑐𝑐𝑡𝑡𝑏𝑏𝑏𝑏

Biolegend

Table 3: List of antibodies used

Isolation of PMN-MDSCs and macrophages

PMN-MDSCs and macrophages were isolated using an immune-magnetic separation kit – EasySep Mouse PE Positive Selection Kit (STEMCELL Technologies) The manufacturer’s protocol was followed Briefly, PMN-MDSCs and macrophages from the blood and spleen were labeled with Ly6G-PE (1A8; BD Bioscience) or F4/80-PE

(BM8; eBioscience) antibodies respectively PE-labeled cells were magnetically labeled and positively selected using an EasySep magnet >80% cell purity was confirmed via flow cytometric analysis on a BD FACS Calibur (BD

immuno-Bioscience) before proceeding with in vitro assays

Microarray analysis

Primary tumors and metastatic tumors were pooled from two to four mice and dissociated as described Tumors were incubated with anti-CD45 microbeads (Miltenyi Biotec) for 20mins at 4°C and subsequently sorted using LS magnetic bead columns CD45+ fraction was labeled with antibodies and PMN-MDSCs were then sorted with the FACSAria (BD Bioscience) into PBS 1X with 5% FCS An average of 80% purity was obtained from the sorted fractions Cells were centrifuged at 4000rpm for 10mins and the cell pellet was lysed in Trizol (Invitrogen)

Total RNA was extracted following the double extraction protocol: RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction followed by a Qiagen RNeasy clean-up procedure RNA yield was assessed using Quant-iT Ribogreen RNA assay and measured by Tecan Infinite M200 Monochromator microplate reader RNA purity was assessed by spectrophotometer (Nanodrop): 260/280 and 260/230 ratios were evaluated for selected samples If the 260/280 and 260/230 ratios were less than 1.8, RNA samples were further ethanol-precipitated Total RNA integrity was assessed by Agilent Bioanalyzer and only high quality RNAs, with RNA Integrity Number (RIN) >7, were used for microarray analysis Biotinylated cRNA was prepared according to the protocol by Epicentre TargetAmp™ 2-Round Biotin-aRNA

Amplification Kit 3.0 cRNA was analysed by Agilent Bioanalyzer Labelled cRNAs were hybridized to the standard Illumina Mouse-6 arrays for 16h at 58°C; the arrays were then washed and stained base on Illumina Wash Protocol and then scanned using BeadArray Scanner 500GX at BSF Microarray Facility The images were analyzed using Beadstudio Gene Expression v 3.3.7 and comparison analyses were carried out according to the instructions provided by Illumina Standard array quality controls

The intensities of the microarray data were acquired at the following scan factors of 1.2 and 1.7 For each array, the two scan factors were linearly regressed to detect saturation in intensity Rightfully, the regressed slope should be around 1.4 (1.7/1.2) The array data was normalized using first quantile normalization, followed by scaling method The quantile normalization variant was applied to groups of technical replicates with the goal to achieve equal spread in the distribution of the bead intensities for each array Then, the scaling method was applied to all the arrays to ensure that the medians of all arrays were equal

Cytospin and May-Grunwald/Giemsa stain

Approximately 5x104 cells were placed in the cytospin and spun down onto a lysine coated slide Excess moisture was allowed to evaporate.Cells were stained with 0.3% May-Grunwald solution diluted in methanol for 5mins Subsequently slides were immersed in 0.3% Giemsa Solution for 15mins Slides were washed with tap water, dehydrated and mounted

poly-Tumor cell detection by qRT-PCR

Lung samples were collected in RNAlater (Qiagen) and homogenized in TRIZOL (Qiagen) RNA was extracted using the Qiagen RNeasy 96-well Universal Tissue Kit cDNA was reverse transcribed (Roche Applied Biosystems reagents) from 2µg RNA The presence of melanocytic tumor cells in tissues and organs was determined by qRT-PCR with SYBR green (Bio-rad iTaq SYBR Green Supermix

with ROX) and specific primers against the melanocyte-specific gene Dct (primers

5′-CTTCCTGAATGGGACCAATG-3′ and

5′-ACGGCGTAATTGTAGCCAAG-3′) Gene expression was normalized against GAPDH expression (primers TGCGACTTCAACAGCAACTC-3′ and 5′-ATGTAGGCCATGAGGTCCAC-3′) Real-time PCR cycling conditions were as follows: 95°C for 10mins; 40 cycles of 95°C for 30s, 55°C for 1 minute, and 72°C for 30s; 95°C for 1 minute; 55°C for 30s; and 95°C for 30s All qRT-PCR reactions were carried out in Stratagene Mx3005P and analyzed with Stratagene MxPro QPCR Software

5′-Low density microarray

One primary tumor in the eye and one cutaneous tumor were taken from each mouse and individually homogenized in TRIZOL (Qiagen) RNA was extracted using the Qiagen RNeasy 96-well Universal Tissue Kit CXCL2 (primers 5’-AGTGAACTGCGCTGTCAATG-3’ and 5’-GAGAGTGGCTATGACTTCTGTCTG-3’) was analyzed separately from the low density microarray due to the absence of this gene from the arrays PCR array kits – Mouse Common Cytokines PCR Array (SABiosciences, PAMM-021) and Mouse Inflammatory Cytokines & Receptors PCR Array (SABiosciences, PAMM-011) – were used to analyze gene expression in the

tumors Manufacturers’ protocols were followed RT2 First Strand Kit (SABioscience) was used for cDNA conversion and RT2 SYBR Green/ROX qPCR Master Mix (SABioscience) was used for the qRT-PCR reaction in Stratagene’s Mx3005P Gene expression was normalized to GAPDH expression and p-values were calculated in Microsoft Excel using a two-tailed paired t-test

Immunohistochemistry and calculations

Formalin-fixed paraffin-embedded sections (5μm) were immunolabeled for S100B (Dako Z0331; diluted 1:4,000), Ki67 (Dako, M7249; 1:25), S100A4 (Abcam, Ab27957; 1:100), vimentin (Proteintech,103661-AP; 1:50) and HMB45/MART1 (Abcam, ab732; 1:50) for use in immunofluorescent (IF) and immunohistochemistry (IHC) studies Briefly, sections were heat-treated with Target Retrieval Solution (Dako, S1699) or, in the case of HMB45/MART-1, a Tris-EDTA Buffer pH9.0 (10mM Tris base, 1mM EDTA, 0.05% Tween-20) After which they were blocked with 3% (v/v) hydrogen peroxide in methanol for 30mins and 10% (v/v) normal goat

or donkey serum (Dako) in PBS for 2h, and then incubated overnight at 4°C Slides were washed and secondary antibodies were incubated for 45mins after

Secondary antibodies used for IHC were biotinylated donkey polyclonal anti-rat (Jackson Lab; AffiniPure F(ab′)2 Fragment; diluted 1:300), alkaline phosphatase–conjugated streptavidin (Rockland Inc.; diluted 1:2,000), and Envision anti-rabbit HRP (Dako, K4003) For IF, anti-mouse Alexa-Fluor 488 (Invitrogen, 412974) and anti-rabbit Alexa Fluor 594(Invitrogen, 56948A) were used as antibodies and 4′,6-diamidino-2-phenylindole, dilactate (DAPI, dilactate) was used as a counterstain

(Invitrogen, D3571) AEC peroxidase substrate (Vector Laboratories) and Alkaline Phosphatase Substrate Kit III (Vector Laboratories) were used as substrates in IHC

Brightfield images for IHC were taken with the Olympus CX31 upright microscope and multiple images stitched together with ImagePro Analyzer 6.2 software (Media Cybernetics Inc.) Fluorescent images were taken with an Olympus Fluoview FV1000 confocal system Tumor areas were assessed using ImagePro Analyzer 6.2 software Total cell number was calculated by dividing the tumor area by the average area of one cancer cell (estimated to be 82µm2) The mitotic index was calculated as follows:

MDSCs were treated 1h before and during the migration assay with inhibitors SB225002 and SB265610 from Tocris Bioscience Flow cytometry was carried out to identify CFSE+Gr1hi PMN-MDSCs

For the in vivo migration assay, 5x106 bone marrow cells from Rosa mT/mG reporter

mice expressing tdTomato and an equal number of Il8r-KO bone marrow cells

expressing GFP were injected intra-orbitally into tumor-bearing RETAAD mice After 18h, the ratio of tdTomato+ cells to GFP+Gr1hi cells infiltrating the contra-lateral tumor was measured by flow cytometry

Tumor proliferation assay

Melan-ret cells, previously described in Lengagne et al (2004) [64], were seeded

(5x103 cells/well) in a flat bottom 96 well plate with 5% RPMI Dilutions of isolated PMN-MDSCs or macrophages were added accordingly and were incubated for 24h at 37°C, 5% CO2 At 24h, 1µCi of [3H]thymidine (PerkinElmer) was added per well and further incubated for 24h Cells were harvested onto a filtermat with the MicroBeta FilterMate-96 Harvester (PerkinElmer) and scintillant luminescence read on a 1450 MicroBeta TriLux (PerkinElmer) PMN-MDSCs and macrophages were irradiated at 2000rads and washed at least twice before co-culture For Transwell cultures, 0.4µm PCF membrane inserts (Millipore) were used and Melan-ret cells were cultured in the bottom well while PMN-MDSCs or macrophages were seeded in the upper well

OVA-specific T cell proliferation assay

CD8+ T cells were isolated from the spleen of C.Cg-Rag2tm1Fwa Tg(DO11.10)10Dlo (Taconic #4219) using a CD8a+ T Cell Isolation Kit (Miltenyi Biotec) 1x105 isolated

T cells were plated with 2x104 irradiated (2000 rads) cells from the CD8- fraction PMN-MDSCs isolated from tumor-bearing mice (12-week-old) were added at a ratio

of 1:1, 1:4, 1:16 and 1:64 to T cells 5µg/ml of SIINFEKL peptide (GL Biochem) was added After 96h incubation, T cell proliferation was measured by [3H]thymidine incorporation as described above

E-Cadherin assays

Fresh tumors were isolated from tumor-bearing RETAAD mice and dissociated as stated above MDSCs were also isolated from the spleen and blood of the same mouse Tumor cells and PMN-MDSCs were co-cultured at a 2:1 tumor to MDSCs ratio for 24h Flow cytometry was carried out on the cells with E-Cadherin antibody (R&D Systems, FAB7481A) and CD45-FITC to determine expression level of E-Cadherin

on tumor cells

A human melanoma cell line, 888mel, was used to determine the effect of MDSCs on E-Cadherin expression 888mel was seeded at 80% confluency and serum starved before co-culture with 0.5x106 PMN-MDSCs isolated from RETAAD tumors per well in a six-well plate After 24h, cells were washed and placed into Trizol RNA was extracted and a qPCR was done to determine the expression level of E-Cadherin Primers used for E-Cadherin were 5’-TTGACGCCGAGAGCTACAC-3’ and 5’-ACTTTGAATCGGGTGTCGAG-3’ The actin gene (primers: 5’-

PMN-CCAACCGCGAGAAGATGA-3’ and 5’-TAGCACAGCCTGGATAGCAA-3’) was used as a normalizing control

For IF staining of cells in culture, actin fibres were stained with Alexa Fluor 488 phalloidin (Invitrogen, A12379) and E-Cadherin was stained with anti-E-Cadherin Alexa Fluor 488 (BD Pharmingen, 610182) Cells were plated on Ibidi 8 well µ-slides and left in culture for four days to form colonies Purified PMN-MDSCs (1x104cells/ml) were added on day 4 and left in co-culture for 24h Cells were washed, fixed

in 4% paraformaldehyde and permeabilized with 0.5% Triton-X for 5mins Cells were blocked with 10% fetal calf serum and incubated with antibodies Fluorescent images were captured with an Olympus IX-81 Inverted microscope and Retiva-SRV CCD Camera (QImaging) and analyzed with ImagePro analysis software (MediaCybernetics)

MT assay

NBT-II cells, expressing histone H2B-mCherry, were seeded at a density of 100 cells per well on Ibidi 8 well µ-slides and left in culture for four days to form colonies Inhibitors were added on day 3 Purified PMN-MDSCs (1x104 cells/ml) were added

on day 4 and left in co-culture for 24h HGF (5ng/ml; Calbiochem) was used as a positive control Inhibitors used were TGF-β1 receptor inhibitor, SB 525334 (Tocris; final concentration 10µM), EGF receptor inhibitor, PD 153035 hydrochloride (Tocris; final concentration 8µM) and HGF receptor inhibitor, JNJ38877605 (Selleck Chemicals; final concentration 5µM) Time-lapse images were captured with an Olympus FV-1000 confocal system and analyzed with MacBiophotonics ImageJ [68]