HEPATOCYTE GROWTH FACTOR IS a MAJOR CYTOKINE FOR NSC HOMING TO GLIOMA

VII Summary Neural stem cell NSC homing to brain tumor has been exploited for targeted gene therapy, but the underlying mechanism remains unclear.. Transplanted neural stem cells NSCs h

HEPATOCYTE GROWTH FACTOR

IS A MAJOR CHEMOTACTIC FACTOR FOR NEURAL STEM CELLS MIGRATION TO GLIOMA

NATIONAL UNIVERSITY OF SINGAPORE

2015

I

Declaration

I hereby declare that the thesis is my original work and it has been written by me in its entirety I have duly acknowledged all the sources of information that have been used in thesis

This thesis also not been submitted for any degree in any university

previously

Huang Sihua

16th Jan 2015

II

Acknowledgements

This dissertation would not have been possible without the opportunity given by Singapore-MIT Alliance of Research and Technology (SMART), Yong Loo Lin School of Medicine, and the constant support and encouragement from the following people:

First and foremost, I would like to thank my supervisors Professor Roger Kamm and Professor Hanry Yu In spite of his busy schedule, Professor Kamm always made time to talk to me about my problems in research He promptly and carefully replies all our email requests He will not blame me when the experiments are not working out Instead he is always there to help and guide me through my difficulties I am truly thankful to Professor Kamm for bringing me the courage and inspiration to pursue my studies

Professor Hanry Yu is my local supervisor He cares for students in a professional and personal level He would discuss with us about our projects for hours until we have a satisfying answer He patiently teaches us the skills

we needed in research and academia He has even made time to teach me the specific details on how to write a paper Professor Yu has given me valuable suggestions and creative ways to solve the problems He has supported me in

so many different ways

I would also like to thank my colleagues in BioSystems and Micromechanics unit (BioSyM) under the Singapore-MIT Alliance of Research and Technology (SMART), Yu Lab (Physiology, NUS), and Kamm Lab (Bioengineering, MIT) My special thanks go to: Dr Poon Zhiyong, Dr Lim Sei Hien, Dr Zhou Yan, Mr Tu

III

Ting Yuan, Dr Andrea Pavasi, Mr Ng Inn Chuan, Dr Jacky Lee, Dr Tan Chin Wen, Mr Evan Tan, Ms Michelle Chen, Mr Kwok Chee Keong, and Ms Averil Chen, who have been there to assist me strategically, technically, and socially

IV

Table of Contents

Declaration I Acknowledgements II

Table of Contents IV

Summary VII List of Tables IX List of Figures X

CHAPTER 1 1

INTRODUCTION 1

1.1 Brief background on neural stem cells 1

1.1.1 Origin 1

1.1.2 Sources of Neural stem cells 2

1.1.3 Therapeutic use of neural stem cells 6

1.2 Glioma 7

1.2.1 Types of Glioma 7

1.2.2 Therapies to treat glioma: traditional and new 8

1.3 Glioma tropism of neural stem cells 12

CHAPTER 2 16

LITERATURE REVIEW 16

2.1 Cytokines involved in NSC tumor homing 16

2.2 SDF1α/CXCR4 18

2.3 VEGF 20

2.4 uPA 21

2.5 IL-6 21

2.6 SCF/c-kit 23

2.7 HMGB1 24

2.8 EGF 25

2.9 HGF/c-Met 25

2.10 MCP1/CCR2 27

2.11 Glioma-produced ECM 27

2.12 Summary and Outlook 28

CHAPTER 3 30

OBJECTIVES 30

3.1 Specific Aim One: Develop in vitro Assays to Study NSC Homing to Glioma 31 3.2 Specific Aim Two: Identify the major homing signals using in vitro experiments in terms of gene and protein expression levels in glioma cells versus astrocytes 32

3.3 Specific Aim 3: Validation of the identified major NSC homing signal(s) in vitro and in vivo 33

CHAPTER 4 34

NEURAL STEM CELLS DISPLAY TROPISM TOWARDS GLIOMA IN TRADITIONAL TRANSWELL ASSAYS AND MICROFLUIDIC DEVICES 34

4.1 Introduction 34

4.2 Materials and Methods 35

V

4.2.1 Cell culture 35

4.2.2 Conditioned medium 37

4.2.3 Transwell Assay 37

4.2.4 Microfluidic devices 38

4.2.5 Seeding NSCs into microfluidic devices 39

4.2.6 Statistical Analysis 40

4.3 Results and Discussion 40

4.3.1 U87 glioma conditioned media induces NSC migration 40

4.4 Conclusions 48

CHAPTER 5 50

HGF, VEGF, AND IL6 ARE CANDIDATE CYTOKINES THAT REGULATE NSC HOMING TO GLIOMA; 50

HGF IS THE MOST PROMISING SIGNAL 50

5.1 Introduction 50

5.2 Materials & Methods 52

5.2.1 Cell culture 52

5.2.2 Conditioned media 52

5.2.3 Transwell Assay 52

5.2.4 Reverse transcription-PCR 53

5.2.5 ELISA (Sandwich Enzyme-Linked Immunosorbent Assay) 55

5.2.6 Flow Cytometry 55

5.2.7 Statistical Analysis 56

5.3 Results & Discussion 56

5.3.1 HGF, VEGF, and IL6 were upregulated in U87 glioma cells comparing to normal astrocytes 56

5.3.2 HGF has higher chemotactic potency than VEGF and IL6 59

5.3.3 NSCs express receptors for HGF, VEGF, and IL-6 64

5.3.4 Among three candidate signals, HGF induced maximum migration of NSCs in transwell assay 66

5.4 Conclusions 73

CHAPTER 6 75

BLOCKING HGF RECEPTOR INHIBITS NSC TROPISM TO GLIOMA; NSCs ARE CHEMOTACTIC TO HGF IN LIVE CONDITION 75

6.1 Introduction 75

6.2 Materials & Methods 76

6.2.1 Transwell Assay – Blocking 76

6.2.2 Dunn’s chamber 77

6.3 Results & Discussion 78

6.3.1 Blockades of CMET (HGF receptor) and VEGFR2 (VEGF receptor) were able to reduce NSC migration Blocking of CMET imposed greater inhibition on NSC migration than blocking of VEGFR2 78

6.3.2 NSCs are chemotactic to HGF gradient in Dunn Chamber 84

6.4 Conclusions 89

Chapter 7 91

Additional Results & Future Experiments 91

7.1 Summary 91

7.2 Methodology 94

7.2.1 Endothelial monolayer 94

7.2.2 Vascular network & NSC infusion 94

7.2.3 Transfecting U87 cells with RFP 96

VI

7.2.4 In-vivo protocol 97

7.3 Additional Results, Discussion & Future Plan 98

7.3.1 Forming endothelial monolayer in microfluidic devices 98 7.3.2 Forming vascular network within the 3-D gel region of microfluidic devices

VII

Summary

Neural stem cell (NSC) homing to brain tumor has been exploited for targeted gene therapy, but the underlying mechanism remains unclear Various signals were proposed by the literature We hypothesize that among the 9 known signals we chose from the literature, there would be a major homing signal regulating NSC homing to glioma

Our first specific aim is to observe the homing of NSCs towards glioma in vivo and in vitro Transwell and microfluidic assays are used and homing was

confirmed in both assays

Our second specific aim is to determine the major cytokine(s) that regulates NSC glioma tropism among the 9 known cytokines We screened these 9 factors for up-regulated gene expression in U87 glioma cells relative to non-cancerous astrocytes Only VEGF, IL6, and HGF were up-regulated in U87 glioma cells, among which HGF showed the highest up-regulation of >400 fold over

astrocytes, versus 2 fold for VEGF and 10 fold for IL6 Similar trends were observed in protein expression measured by ELISA In transwell assays, HGF induced significant NSC migration relative to VEGF and IL6 Through FACS experiment we also found that NSCs have the highest expression of HGF

receptor over the receptors for VEGF and IL6

VIII

Our third specific aim is to further explore the role HGF plays in NSC homing to glioma We found that blocking HGF receptor inhibited NSC migration towards glioma conditioned medium In live imaging, NSCs migrated along HGF gradient

We conclude that HGF is a major chemoattractant in NSC homing to glioma

IX

List of Tables

Table 1 List of pre-clinical trials that make use of NSCs as cellular carriers

to deliver therapeutic transgene into tumor bearing mice 13

Table 2 List of important signaling pathways involved in NSC homing to glioma 18

Table 3 Primer sequences of various cytokines used for RT-PCR experiments 54

Table 4 Amount of HGF, VEGF, SDF-1, and IL-6 secreted by astrocytes and U87 glioma cells 62

Table 5 List of dissociation constants of HGF, VEGF, and IL6 63

Table 6 Molecular Weight of HGF, VEGF, and IL6 68

Table 7 Parameter of migration in control condition and HGF gradient 87

X

List of Figures

Figure 1 Stem cell mediated CNS regeneration and therapeutic delivery [15] Stem cells, neural stem cells especially, have been used in cell replacement, remyelination, promotion of host tissue regeneration, enzyme replacement therapy for neuroprotection, and tumor-localized chemotherapy production 6

Figure 2 Cell based anti-cancer therapeutics Transplanted neural stem cells (NSCs) homing to tumor cells in rodent models of brain neoplasia [34] 13 Figure 3 Conditioned media derived from U87 glioma cells induced

significant migration of NSCs in a transwell assay Columns, cell

number per field; bars, SE; Statistical comparison were calculated between cells induced by conditioned medium and control using

Student’s t test Fluorescent images of migrated NSCs were shown in control, conditioned media, and 2% FBS 41

Figure 4 Growth of NSCs in different mixtures of gels Phase contrast

images were taken for cells cultured in 48-well plate 1 M/ml of NSCs were used Images were taken 4 h after gelation The concentration for each gel used is 2mg/ml 42

Figure 5 A) Top view of microfluidic device NSCs are seeded in the channel depicted in green and glioma cells are seeded in the other channel shown in red B) Dimensions of microfluidic device [77] 44

Figure 6 Figure A&B are magnification of central part of the microfluidic device Neural stem cells (shown in green) are seeded in the lower channel In Figure A, no cells are placed in the upper channel while in Figure B, U87 glioma cells are seeded together with neural stem cells Images were taken one week after cell seeding A) NSCs migrate to a minimal distance into the central gel region without any inducement B) NSCs migrate much further into the central gel region and more NSCs migrate in response to U87 glioma cells relative to control C) Measurement of migration distance under control condition and

condition with U87 cells With inducement by U87 cells, the migration distance is significantly greater Migration distance is the distance from the center of mass (COM) of the cell to the gel-channel interface 47

Figure 7 HGF, VEGF, and IL6 were upregualted in U87 glioma cells

comparedµµ to normal astrocytes mRNA were collected from glioma cells and astrocytes and the amount of mRNA were determined using real time-PCR Columns, fold increase Bars, standard deviation ****, P<0.0001 57

XI

Figure 8 Glioma cells secreted more HGF, VEGF, and IL6, and less SDF1 than astrocytes Conditioned media were collected from the cells and

concentrated by ultra-centrifugal units ELISA was performed to

quantify the amount of cytokines secreted by the cells All data were obtained in triplicate and are represented as mean+ SD ****, p<0.0001 61

Figure 9 HGF has a higher chemotactic potency comparing to VEGF and IL6 Conditioned media were collected from the cells and concentrated by ultra-centrifugal units ELISA was performed to quantify the amount of cytokines secreted by the cells All data were obtained in triplicate and are represented as mean+ SD ****, p<0.0001 63

Figure 10 Expression of cMET, VEGFR2 and IL6R by NSCs The fluorescence signals detected from cells positively stained by the receptor

antibodies are shown c-MET is more highly expressed than VEGFR2 and IL6R *P<0.05; **P<0.01; bars, SD 65

Figure 11 HGF was able to induce the most significant migration of NSCs Transwell assay was used to study the effects of chemoattractants VEGF, IL6, and HGF on NSC migration Chemoattractants were added at the bottom well and NSC were seeded at the top well Columns, cell count per field Bars, standard deviation ****, P<0.0001 67

Figure 12 Dextran gradient established across the gel region of microfluidic devices Dextran with red fluorescence was introduced into one

channel, and media was placed into the other channel Two channels are separated by collagen-Matrigel mixture Fluorescence was

recorded over 6 h and analyzed Horizontal axis: Distance from the analyzed point to the gel-media interface Vertical axis: normalized fluorescence for each point 6 different colors of curves represent different point of time 69

Figure 13 NSCs are attracted to HGF gradient in microfluidic devices NSCs were seeded at the lower channel, media or HGF was added to the upper channel Cells were allowed to migrate for 11 days and images were taken at day 6, day 9, and day 11 71

Figure 14 Blocking VEGFR2 using 1uM Vatalanib reduces the migration, but not as much as 100nM JNJ Combination of 1uM Vatalanib and 100nM JNJ is not significantly different from 100nM JNJ alone in blocking NSC migration Increasing the concentration of both chemicals further reduces NSC migration Vatalanib is a VEGFR2 blockade with IC50 of

~40nM JNJ is a c-MET blockade with IC50 of ~4nM ****, p<0.0001;

***, p<0.001; ns, not significant 79

Figure 15 Blocking VEGF and CMET receptor alone is able to reduce NSC migration But blocking CMET receptor has a significantly greater inhibition on NSC migration ****, p<0.0001; ***, p<0.001; **, p<0.01 81

XII

Figure 16 Blocking HGF receptor c-MET using chemical JNJ is able to induce

a dose-dependent inhibition on NSC migration Transwell migration assay was used NSCs were incubated with chemical blockades before added to transwell Vatalanib, an inhibitor of VEGFR2with IC50 of 37

nM JNJ (JNJ-38877605), an inhibitor of c-Met with IC50 of 4 nM ****, p<0.0001; ***, p<0.001; ns, not significant 83

Figure 17 NSCs are chemotactic to HGF gradient Migration tracks of NSCs under no gradient (A) and HGF gradient (C) Black tracks – overall positive migration Red – overall negative migration Directionality histogram of NSC migration under no gradient (B) and HGF gradient (D) Dunn chamber was used and time lapse images were captured every 5 minutes for 6 h Cells were tracked using Metamorph software and analyzed by Ibidi Chemotaxis and Migration Tool 85

Figure 18 Forward Migration Index (FMI) of NSCs under control condition and HGF gradient 86

Figure 19 pRetroQ-mCherry-N1 Vector Map and Multiple Cloning Sites (MCS) 96

Figure 20 Forming endothelial monolayer in the channel of microfluidic devices 2.5M/ml of endothelial cells were seeded in Matrigel coated microfludic channel 3 days after seeding, cells were fixed and stained with DAPI A) Top view of the monolayer B) Side view of the

monolayer C) Picture of device D) Schematic view of high-throughput microfluidic device 99

Figure 21 A) Schematic diagram of 3-gel device (top view) B) Schematic diagram of 3-gel device (side view) C) Picture of 3-gel microfluidic device D) Endothelial cells form vascular network in the gel region of microfluidic devices E) Endothelial cells form vascular network under the presence of U87 glioma cells F) 3-D image of glioma vasculature G) Projected image of glioma vasculature Green: VE-cadherin; Red: U87-RFP cells; Blue: DAPI; White: Phalloidin 101

Figure 22 Process of NSC extravasation from endothelial network and homing towards U87 cells Image A, B, C, D represent images taken at 0.5 hr, 1.5hr, 2.5hr, and 3.5hr after introducing NSCs into the glioma vasculature Red: endothelial cells with RFP; Green: NSC GFP 103

Figure 23 A) 3-D image showing NSCs escaping from the vascular network and homing to glioma cells B) Projection image showing NSCs escaping from vascular network and homing to glioma cells Red: U87 RFP cells; Green: NSC GFP cells; Blue: DAPI; White: Phalloidin 104

XIII

Figure 24 A) The set-up of in vivo experiment using stereotaxic instrument B) Isolated mouse brain with established glioma shown in red C) Migration of NSCs (Red) from the original cell mass towards the

established glioma (unseen) at the right frontal lobe 106

1

CHAPTER 1 INTRODUCTION

1.1 Brief background on neural stem cells

Neural stem cells (NSCs) are referring to cells that can self-renew and

differentiate into all three lineages of neuronal cells: neurons, astrocytes, and oligodendrocytes

1.1.1 Origin

Neural stem cells can be isolated from central nervous system of embryonic, fetal, and adult individuals These cells originate from the neuroectoderm of the early embryo The embryonic NSCs are the neuroepithelial cells (NEP) They are rapidly multiplying cells that form the neural plate and then neural tube, which later develop into the mature CNS [1] The molecular markers of this cell type include SOX1, an early NSC maker; as well as intermediate

filament nestin [2]; and CD133, a cell surface associated transmembrane protein [3] Oct4, a transcription factor responsible for pluripotency, is also expressed by early NEPs [4]

Towards the end of embryonic development, asymmetric division occurs and the neuroblast appears Meanwhile, the neuroepithelium converts into a

2

tissue consisting of different cell layers The most apical layer includes radial glial cells (RGCs), a secondary NSC that is different from NEPs as primary NSCs [5] Another type of secondary neural stem cell is basal stem cell This cell type is generated by asymmetric division of neuroepithelial cells and after depletion of this cell pool, also by asymmetric division of RGCs Basal stem cells are located at the subventricular zone (SVZ) and they are neural stem cells of restricted lineages, developing into neurons after one or two

symmetrical divisions [6]

Additional stem/progenitor cells exist to generate neurons, astrocytes, and oligiodendrocytes, not only during development, but also in adult CNS In the adult brain, neurogenesis mainly occurs in two regions: the SVZ of the basal forebrain along the lateral ventricle, and the subgranular zone (SGZ) of the hippocampal dentate gyrus Other areas of the brain as well as the spinal cord have also been reported to contain proliferating neurogenic cells

1.1.2 Sources of Neural stem cells

Embryonic stem-derived NSCs

Embryonic stem cells (ESCs) have the pluripotency to generate all types of cells Several protocols have been proposed to derive neural progeny from ESCs In summary, three methods are reported to generate NSCs from ESCs The first method is through the generation of embryoid bodies [7]; the second method leads to generating neural rosettes under adhesion conditions and

3

transiently supplementing Noggin to the culture medium [8]; the third

method involves differentiation of EBs into neural cells with sequential

addition of retinoic acid, Sonic hedgehog (Shh), and cAMP [9] The protocols have been used to generate a variety of neural cell types including astrocytes, oligodendrocytes, glutamatergic, GABAergic, and dopaminergic neurons However, the characterization and single-cell clonogenic potential still needs further clarification [10]

Fetal and adult NSCs

Adult NSCs divide slowly and therefore it is not feasible to establish a human NSC cell line from isolated adult NSCs However, NSCs can be successfully isolated from the telencephalic di-encephalic region or from the subgranular region of the fetal human brain [10]

Several clonal, genetically homogeneous human NSC cell lines have been established by genetic modification These cell lines are usually immortalized with V-MYC or C-MYC genes They have the advantages of their non-

transformed nature, human origin, mulitpotency, fast growth, high

availability, and feasibility for molecular manipulation These cells hold great potential for the development of cell replacement or gene transfer-based therapies It is also shown by a recent study that V-MYC modified neural progenitors derived from fetal human brain have no tumorigenic potential

4

either in vitro or in vivo[11] In our study, we will also be using a fetal NSC

cell line immortalized with V-MYC

Induced pluripotent stem cells

A major obstacle of utilizing NSCs in clinical applications is the need for fetal brain tissue There is limited availability of fetal tissue and it raises multiple moral and ethical complications Therefore having a renewable source of NSCs would greatly benefit potential therapeutic approaches Recently, induced Pluripotent Stem Cells (iPSCs) have been proposed as an alternative source of neural cells because they share similar properties as embryonic stem cells and the potential to differentiate into any somatic cell type And they may provide the benefits of minimal immune rejection

iPSCs have multiple advantages: first the ability to generate stem cells from skin fibroblasts without the need for human embryos to generate ESCs;

secondly, they are able to generate neurons, astrocytes and oligodendrocytes from adult patients; thirdly, they provide the opportunity to investigate how different cell types may be involved in a specific pathology; last but not least, they can be used to characterize the bio-molecular mechanisms that underlie the development of a chronic disorder[12]

5

However, the genetic manipulation of iPSCs may lead to tumor formation, which needs to be addressed before they are used as cell transplant into patients

CNS cancer stem cells

By applying the same culture system developed for neural stem cells, term stable cell lines can also be derived from human glioblastomas These cells can also be differentiated into the major three lineages neuronal cells: neurons, astrocytes, and oligodendrocytes [13] These cells, when they are injected subcutaneously and intracranially, form primary glioblastoma and have extensive migratory capacity, therefore these cells offer enormous

long-potential to establish pathological models [14]

6

1.1.3 Therapeutic use of neural stem cells

Figure 1 Stem cell mediated CNS regeneration and therapeutic delivery [15]

Stem cells, neural stem cells especially, have been used in cell replacement, remyelination, promotion of host tissue regeneration, enzyme replacement therapy for neuroprotection, and tumor-localized chemotherapy production

Neural stem cells are used in multiple areas (Figure 1) They are used in cell replacement therapies to treat neurodegenerative diseases such as

Alzheimer’s disease They aid in re-myelination processes when there is

spinal cord injury or Pelizaeus-Merzbacher Disease Neural stem cells also promote host tissue regeneration by secreting neurotrophic factors In

enzyme replacement therapy, NSCs are employed as a cellular carrier for

lysosomal enzymes that can be released and protect host neurons In localized chemotherapy against brain tumor, NSCs are genetically engineered

tumor-to carry enzymes, which can transform systematically or orally administered

7

prodrug into an active form at the tumor foci and kill the tumor cells In this

study, we focus on tumor-localized chemotherapy and we aim to

improve the therapy through studying of the mechanisms of homing to brain cancers

According to the World Health Organization (WHO), gliomas are graded from

1 to 4 Grade 1 glioma is benign and circumscribed with a slow proliferation rate It includes the most common glioma of children: pilocytic astrocytoma Grade 2 glioma includes astrocytoma, oligodendroglioma, and

oligoastrocytoma They also have relatively slow growth rates but these

tumors start to diffuse to normal brain and have a higher potential of

malignant progression Grade 3 tumor is characterized by a higher cellular density and the existence of atypical and mitotic cells Grade 4 tumors are the most malignant and happen most frequently They include glioblastoma and

8

gliosarcoma They have properties of micro-vascular proliferation or

pseudo-palisading necrosis [17]

1.2.2 Therapies to treat glioma: traditional and new

The current standard of care for patients of glioma includes maximal safe resection, followed by radiation therapy (RT) to the resection cavity and chemotherapy (telozolomide or TMZ) Surgical resection alone will lead to survival of about 6 months, and when combined with RT survival extends to 12.1 months Adding telozolomide further extends survival to 14.6 months [18, 19]

Surgery is still an important aspect of treatment It helps to reduce

intracranial pressure, and sometimes leads to recovery of certain neurological function However the disadvantages are obvious too, due to poor

performance status, advanced age of the patients, and involvement of other parts of the brain [20]

The combination of RT and TMZ is most efficacious after the primary

resection Treatment after surgery usually includes 6 weeks of RT to the

surgical cavity and TMZ, and after that, 6 cycles of TMZ [19] RT of

glioblastoma multiforme (GBM) is focal, fractionated external beam radiation therapy to the surgical resection cavity and to a 2cm margin of surrounding

9

brain tissue [21] Ionizing radiation induces single strand and double strand DNA breaks in proliferating cells TMZ is an oral alkylating chemotherapeutic agent It derives its therapeutic effect from adding a methyl group to purine bases of DNA, causes DNA damage, and triggers a cascade of events, leading to tumor cell apoptosis [22] TMZ alone has very limited clinical benefits It is when applied concurrently with RT it prolongs the survival of glioma patients However, the progression-free survival time is still only 7 months despite all these efforts[23]

Molecular targeted therapies

Molecular targeted therapies can be divided into small molecule inhibitors and monoclonal antibodies Small molecule inhibitors are non-polymeric, organic compounds able to cross cell membrane and target specific

intracellular components Many of the small molecule inhibitors are tyrosine kinase inhibitors They act by selectively targeting the intracellular kinase domain of receptor tyrosine kinases (RTKs) In contrast, monoclonal

antibodies are usually too large to cross the cell membrane and they target cell surface proteins and extracellular peptides [22, 24]

These molecular targeted therapies can inhibit growth factor pathways such

as EGFR pathway Amplification of EGFR signaling is one the most common genetic alteration seen in GBM [25] Some of them inhibit angiogenic

10

pathways Glioma is a highly vascularized tumor characterized by extensive angiogenesis VEGF is critical for angiogenesis and is highly expressed in glioma Several small molecule inhibitors are used to block this pathway and inhibit angiogenesis in the tumor site Some of the molecules block

intracellular signaling pathways such as PI3K/AKT/mTOR and

RAS/RAF/MAPK as secondary messenger systems inside the cells

Immunotheray

Immunotherapy intends to make use of the immune system to selectively target glioma cells There are two types: passive immune therapy and active immune therapy Passive immune therapy uses antibodies, immune cells, or other components of the immune system to target the tumor It does not require activation of patients’ own immune system Immune cells are

activated outside the body and re-injected into the patients Active immune therapy requires activating the patients’ own immune system It includes peptide-based therapy and cell-based therapy These are usually referred to

as cancer vaccines

Gene therapy

Gene therapy for the treatment of cancer involves delivering genetic material, including transgenes, toxins, and viruses, into tumor cells to induce cell death The genetic material is usually packed within a vector to deliver into the cells

In addition to viruses, stem cells have been used as vectors to deliver genetic materials to the tumor Stem cell vectors are promising because of their innate tumor trophic properties Three types of stem cells have been studied

including neural, embryonic, and mesenchymal stem cells Moreover,

liposomes and nanoparticles have also been explored as vectors to deliver gene to glioma [27]

Gene therapy can also be divided into conditional cytotoxic gene therapy and

directly cytotoxic gene therapy Conditional cytotoxic gene therapy is also

known as an enzyme-prodrug activation system, which is commonly used in the treatment of glioma Usually the transgene of a non-cytotoxic enzyme is delivered and integrated into the tumor cells A non-cytotoxic prodrug is administered, and it will react with the enzyme expressed at the tumor site and become highly toxic, conferring the ability to kill the tumor cells Direct cytotoxic gene therapy utilizes the surface molecules

overexpressed in glioma to send toxins directly into the tumor cells This is usually done by viral-vector mediated delivery of transgenes for highly toxic

12

proteins or immunotoxins Immunotoxins are recombinant proteins that include a ligand conjugated to a toxin The ligand binds specifically to the tumor cell surface molecule leading to internalization of the toxin The toxin will then trigger a series of cellular response that leads to tumor cell apoptosis [27, 28]

1.3 Glioma tropism of neural stem cells

Neural stem cells were found to home to brain glioma by Aboody et al [29], which introduced a promising new method for treating the fatal glioma When injected directly into the brain or into the vein of the mouse model, NSCs migrate towards glioma [29, 30], medulloblastoma [31, 32], and melanoma brain metastases [31] Their unique capability to cross the blood brain barrier provides an additional advantage in treating CNS related cancer [33]

Researchers have been using NSCs as a cellular agent to carry transgenic products and target the tumor at the site Since brain tumor is highly

migratory, selective targeting of the residue by NSCs has the great advantage

of getting to the cancer cells without damaging the healthy brain tissue

13

Figure 2 Cell based anti-cancer therapeutics Transplanted neural stem cells

(NSCs) homing to tumor cells in rodent models of brain neoplasia [34]

A number of pre-clinical studies have been conducted to evaluate NSCs as

tumor selective therapy Various rodent and human NSC cell lines are

administered into a mouse model bearing glioma either intracranially or

through the circulation NSCs carry cytotoxic genes or an enzyme that

transforms a pro-drug into an active form at the tumor foci Test animals

receiving the NSC-mediated therapy experienced extended survival and

tumor reduction These trials are documented in Table 1

Table 1 List of pre-clinical trials that make use of NSCs as cellular carriers to

deliver therapeutic transgene into tumor bearing mice

Neural Stem

Cell type

Route of administration

Therapeutic transgene

Therapeutic

[37]

NSC-IL-12

Extended survival [38]

Significant tumor reduction/slowed disease

progression

[40]

A phase 1 clinical trial was initiated in September 2010 to use genetically modified NSC to treat glioma NSCs can be genetically modified to express cytosine deaminase (CD), an enzyme that can convert orally taken pro-drug 5-

FC to an active form 5-FU that preferentially kills tumor cells [15] Another year open-label phase I trial was published in October 2012 of using first trimester fetal NSC to treat Pelizaeus-Merzbacher disease (PMD), a rare

1-leukodystrophy caused by mutation of the proteolipid protein 1 gene [41] Axons are not properly myelinated in PMD, resulting in neurological

dysfunction Allogeneic hNSCs were surgically implanted into the frontal lobe white matter in four male subjects with PMD Immunosuppression was

administered for 9 months The results showed increased myelination in the

region of transplantation These phase I trials indicate a favorable safety

profile for using hNSCs to treat CNS diseases

15

The phenomenon of NSCs homing to brain tumor followed by these

pre-clinical trials opened up a new avenue for the treatment of deadly glioma It is

a promising and efficient method to use because of its targeting specificity that spares the healthy brain tissue Considerable effort has been devoted recently to improving this method Our study aims to increase the NSC

targeting efficiency by looking into the mechanism of homing

16

CHAPTER 2 LITERATURE REVIEW

2.1 Cytokines involved in NSC tumor homing

Studies have proposed a range of signals involved in NSC migration including stem cell derived factor-1 (SDF1), hepatocyte growth factor (HGF), monocyte chemoattractant protein-1 (MCP-1), urokinase plasminogen activator (uPA), interleukin 6 (IL-6), vascular endothelial growth factor (VEGF), stem cell factor (SCF), epidermal growth factor (EGF), and high mobility group box 1 protein (HMGB1)

SDF1 is an inflammatory chemoattractant that can be sensed by NSCs through

CX chemokine receptor 4 (CXCR4) NSCs have been shown to migrate towards brain infarcted areas where SDF1 is upregulated due to local inflammation For example, blocking CXCR4 inhibits migration of NSC towards brain injury [42, 43] HGF has also been shown to be critical in NSC glioma tropism via the Ras-phosphoinositide 3-kinase (PI3K) signaling Short hairpin RNA mediated deletion of the HGF receptor c-Met significantly reduced the response [44] Monocyte chemoattractant protein-1 (MCP-1) is found to induce NSC

migration in transwell assay and increase the migration of single NSCs from neurospheres There is evidence of NSCs migrating towards rat brain regions infused with MCP-1 [45, 46] Cytokine interleukin 6 (IL-6) is identified to be a

17

major factor mediating NSC tropism to invasive breast cancer cells There are elevated IL6 levels in breast cancer cells with high invasiveness in contrast to those with low invasiveness Blocking IL6 receptor or si-RNA assisted IL6 ablation has attenuated NSC migration in vitro [47] uPA is found to be

important in stem cell migration to malignant solid tumors A study has

shown that depletion of uPA from prostate cancer conditioned medium

blocked NSC migration and overexpression of uPA and its receptor in a

neuroblastoma cell line induced robust NSC migration [48] VEGF is reported

to promote long distance migration of NSCs in the brains of adult mice;

therefore tumor up-regulated VEGF is a relevant guidance signal for NSC tropism [49] Other possible signals that regulate NSC homing include SCF which induces NSC to migrate to areas of brain injury or glioma conditioned medium [50, 51], EGF that confers a motile phenotype of NSCs [52], and

HMGB1 that recruits stem cells through endothelial barriers [53]

The candidate chemokine-receptor pairs involved in NSC’s tumor tropism are summarized in Table 2 Currently the SDF-1 alpha/CXCR4 was considered to

be the strongest chemoattractant SCF, MCP-1, HGF, and VEGF were also

suggested by other studies to be the major homing chemoattractants Few

studies have made comparison among these different cytokines in their ability to induce NSC homing There is no study to compare the cytokines regarding their genetic and proteomic expression between glioma and

neural progenitors expressing surface CXCR4 migrate along the

concentration gradient of SDF1 secreted by tumor cells [58], which when inhibited, leads to a truncation of this migratory behavior [43]

19

A study by Imitola et al reported that human NSCs move toward regions of

CNS injury where high levels of SDF1 are secreted by local astrocytes and

endothelium SDF-1 promotes NSC proliferation, migration, and can trigger a

series of intracellular molecular processes such as phosphorylation of

p38MAPK (involved in regulating cytokine-induced migration) [59-61],

phosphorylation of ribosomal S6 kinase (or p90 RSK, related to

phosphorylation of cytoskeletal molecules in neurite outgrowth) [62, 63], phosphorylation of c-Jun (a kinase involved in migratory responses) [64]; rapid activation of extracellular response kinase (involved in proliferative responses); and rapid phosphorylation of paxilin (a scaffold molecule that is critical for migration) [65] These processes are closely linked to cell

proliferation and migration

Another study showed that murine C17.2 NSC cells exhibited greater

migration towards tumor-derived endothelium (TEC) than normal

endothelium, and that blocking SDF-1α attenuated C17.2 tropism to TEC It has been suggested that NSCs are attracted to TEC by SDF-1α, which may also promote transendothelial migration under flow conditions [54]

SDF-1 seems to play an important role in NSC tumor tropism, which has been claimed by a range of studies However, SDF-1 not only home to brain tumors but also to brain infarcted areas or injuries Local endothelium at the tumor site or injury is usually involved In our study, we are going to look at how NSCs react to SDF-1 without the effect of endothelium These results will

20

make a significant contribution to the effect of SDF-1 on the process If NSC homing occurs without the involvement of endothelium, SDF-1 itself is an important cytokine in NSC homing to brain tumors

2.3 VEGF

VEGF is able to stimulate NSC migration in vitro in a dose-dependent manner The expression of VEGFR2 in human NSCs is confirmed by mRNA expression and immunocytochemistry It has been demonstrated by Schmidt et al that VEGF is a potent signal guiding NSCs migration from a distant site in the adult mouse brain When neutralizing antibody to VEGF is added to the protein extract derived from three astrocytomas and brain metastases, NSC migration

is reduced up to 80% However, the incomplete blocking of NSC migration indicates the presence of other chemotactic signals involving in NSC homing

to glioma [49]

VEGF plays an important role in tumor vasculogenesis and it is also expressed

by most cell types including brain endothelial cells and tumor cells VEGF helps the formation of vascular protrusion through out the brain tumor so that nutrients are supplied to the tumor cells Here we aim to study the other aspect of VEGF functionality in NSC tumor tropism in order to shed light to its involvement in NSC chemotaxis

21

2.4 uPA

Urokinase plasminogen activator (uPA) and urokinase plasminogen activator receptor (uPAR) are up-regulated in invasive tumors In the study by Gutova

et al [48], they reported that activation of uPA and uPAR in brain, lung,

prostate, and breast tumor induces neural and mesenchymal stem cell

tropism Cytokine expression profiles were examined for these

tumor-conditioned media 79 cytokines were investigated, among which interleukin 6(IL-6), interleukin 8 (IL-8), and monocyte chemoattractant protein-1 (MCP-1) were most abundant in uPAR-positive tumors The study also observed that human recombinant uPA induced stem cell migration and depletion of uPA from PC-3 prostate cancer cell-conditioned medium stopped the migration of stem cells Meanwhile, overexpressing uPA and uPAR in neuroblastoma

(NB1691) through retrovirus induced robust migration of stem cells toward NB1691 cell conditioned medium Therefore the study concludes that uPA activation of uPAR in cancer cells is a critical mechanism in regulating stem cell tropism to malignant solid tumors [48]

2.5 IL-6

Zhao et al [47] has identified IL-6 as a major cytokine responsible for NSC homing to breast cancer cells Firstly, cytokine levels in breast cancer cell-conditioned media were detected by cytokine antibody arrays It was found that Il-6 and IL-8 were the most highly expressed cytokines in invasive breast

22

cancer cells in contrast to non-invasive breast cancer cells They further

confirmed the results by quantitative real-time PCR and ELISA

In order to further investigate the functionality of IL-6 and IL-8, Zhao et al added both cytokines solely, or in addition, to the conditioned medium

derived from breast cancer cells with low invasiveness As a result, IL-6

increased NSC migration while 8 made little difference Addition of both

IL-6 and IL-8 didn’t affect the migration more than adding IL-IL-6 alone They also added neutralizing antibodies for both IL-6 and IL-8 to breast cancer cell-conditioned medium Neutralizing IL-6 reduced the migration of NSC by

approximately 50% while neutralizing IL-8 did not show a significant

difference Later, they further confirmed the importance of IL-6R by

neutralizing IL-6R or using IL-6R si-RNA As a result, both methods resulted in significant reduction in NSC migration Therefore, the study proposed IL-6 as

an important cytokine in regulating NSC tropism towards breast cancer [47]

Here IL6 was proven to play an important role in NSC homing towards breast cancer How IL6 functions in NSC brain tumor tropism is not clear Our study will address this question by looking at IL6 secretion from brain tumor and studying the inducement of IL6 together with 8 other cytokines on NSC

homing

23

2.6 SCF/c-kit

Pluripotent mouse embryonic stem cells (ES) can be induced to become

neural progenitors in vitro by retinoic acid When transplanted into mouse brain, neural progenitors migrate towards areas of inflammation and damage Neural progenitors share the same characteristics with neural stem cells except that they can be self-renewed for a limited number of times It is

observed that neural progenitors derived from ESCs selectively migrate to glioma cell conditioned media as well as to synthesize SCF RT-PCR results show that three glioma cell lines N1321, U87, and rat glioma cell line C6 produce SCF, and that the SCF receptor c-Kit is expressed by both

undifferentiated ES cells and derived neural progenitors[51]

Erlandsson et al used a microchemotaxis assay to study the effect of SCF on NSCs from embryonic rat cortex They demonstrated that NSC chemotaxis can

be triggered by SCF and that antibodies to SCF inhibited the migratory

response They have confirmed that c-kit was expressed in NSCs and their differentiated progeny and that SCF acts as a survival factor for NSCs [55] Another study also shows that SCF induces that migration of NSCs to the areas

of brain injury [50]

The above studies concluded that SCF is important in NSC migration to

damaged areas, various types of tumors, and brain inclammation, and that SCF

24

is involved However how SCF is involved and to what extent SCF triggers NSC chemotaxis is unknown It is shown that SCF is expressed by U87 glioma cells (cells used in our study) But is SCF expression by U87 cells higher than that

by the surrounding cells, for example non-cancerous astrocytes?

2.7 HMGB1

HMGB1 is a non-histone protein required to maintain chromatin structure Recently HMGB1 was observed to play a role in inflammation, cell migration and metastasis Palumbo et al reported that HMGB1 attract mesoangioblast to migrate across an endothelial monolayer and reach damaged muscle HMGB1 therefore appears to induce stem cell transmigration across an endothelial barrier [53] In addition, HMGB1 acts as a pro-inflammatory factor, and is released by damaged cells as a chemoattractant for vascular smooth muscle cells and fibroblasts and induces cytoskeleton reorganization and migration [66, 67] It is suggested that HMGB1 promotes invasion and metastasis of cancer cells, most likely by activating metalloproteases MMP2 and MMP9 [68, 69]

These studies suggest an important role of HMGB1 in stem cell transmigration across the endothelial barrier There is, however, no direct evidence of its involvement in NSC brain tumor tropism We hope to extend the current

25

knowledge of HMGB1 and its attraction of mesangioblast across the

endothelial monolayer to whether it plays a role in stem cells tumor homing

2.8 EGF

Epidermal Growth Factor Receptor (EGFR) signaling inhibits NSC

differentiation Increase in EGFR signaling induces reorganization of the

cytoskeleton and focal adhesion disassembly [52] It is also suggested that EGFR kinase activation offers transplanted NSCs the necessary cellular signals

to migrate and proliferate particularly into the cortical gray matter, this is significant because the intact adult brain matter usually does not support NSC migration outside rostral migratory stream[29] C17.2 NSCs expressing EGF receptors orient along blood vessels in contact with laminin in the basal

lamina and migrate long distances [70] This is consistent with the work demonstrating that EGF receptor transcriptionally up-regulates vascular endothelial growth factor expression [71] Moreover, EGF increases the

invasiveness of glioma The EGF receptor gene is amplified up to 40% in

malignant gliomas EGF receptor expression level is correlated with the

quantitative results of invasiveness [56]

2.9 HGF/c-Met

An in vitro experiment conducted by Heese et al [57] found that HGF is the strongest chemoattractant for NSCs out of 13 different growth factors tested

26

including EGF and FGF-2 An antibody to HGF prevented this action It has also been suggested that the ECM of the brain is relatively similar in all patients so the NSC behavior to pass through brain ECM is more or less consistent [57] However, infiltration of NSCs into tumor masses can be different because tumor ECM has a variable composition Therefore additional factors may be involved for NSCs to enter 3D glioma spheroids possibly involving interaction with tumor ECM [57]

Kendall et al [44] also tested a range of chemotactic factors including HGF, VEGF, EGF, and TGF-α and found that HGF has the highest efficacy and induces the highest migration capacity of NSCs at all concentrations tested VEGF has the second highest impact on the dose-dependent response It is also observed that a short hairpin RNA-mediated ablation of c-Met significantly inhibited the response Moreover, blocking of Ras-phosphoinositide 3-kinase (PI3K)

signaling impaired NSC responding to HGF and other factors [44]

HGF is involved in cancer cell proliferation and migration It has also been reported in NSC glioma tropism The experiments have been conducted in vitro and PI3K mechanism was suggested to be involved What we hope to test is how strong a candidate HGF is comparing to the range of other

cytokines and does HGF functions in the same way in vivo Extending the study in vivo is challenging but it is intriguing and has the potential to

improve the current innovative ways to cure brain tumor