Development of new neural stem cell based tumor targeted gene therapy approaches

SUMMARY Neural stem cells NSCs have recently emerged as one of the most attractive cellular vehicles for targeted gene delivery to cancers due to their migratory capacity and tumor tropi

DEVELOPMENT OF NEW NEURAL STEM

CELL-BASED TUMOR-TARGETED GENE

THERAPY APPROACHES

ZHU DETU

NATIONAL UNIVERSITY OF SINGAPORE

2012

DEVELOPMENT OF NEW NEURAL STEM

CELL-BASED TUMOR-TARGETED GENE

THERAPY APPROACHES

ZHU DETU (B Sc.)

A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF BIOLOGICAL SCIENCES NATIONAL UNIVERSITY OF SINGAPORE

&

INSTITUTE OF BIOENGINEERING AND

NANOTECHNOLOGY

2012

DECLARATION

I hereby declare that this thesis is my original work and it has been written by me

in its entirety I have duly acknowledged all the sources of information which have

been used in the thesis

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

previously

ZHU DETU

21 August 2012

ACKNOWLEDGMENTS

I would like to acknowledge all who have helped and inspired me during my study

at the National University of Singapore and Institute of Bioengineering and Nanotechnology

I am very grateful to my supervisor, Dr Wang Shu, Associate Professor, Department of Biological Sciences, National University of Singapore, for his invaluable inspiration and guidance during my PhD study

I would like to dedicate my most sincere gratitude to my parents for their constant encouragement and support

I acknowledge the National University of Singapore, for honoring me with studentship and financial assistance in the form of scholarship

TABLE OF CONTENTS

ACKNOWLEDGMENTS I TABLE OF CONTENTS III SUMMARY VI LIST OF PUBLICATIONS VIII LIST OF TABLES IX LIST OF FIGURES X ABBREVIATIONS XII

CHAPTER 1 INTRODUCTION 1

1.1 Neural stem cells 2

1.1.1 Tumor tropism 2

1.1.2 Cell source 3

1.1.3 Genetic engineering 5

1.1.4 Side effects of intravenous injection 6

1.2 Fusogenic membrane glycoproteins 7

1.2.1 Bystander effect 7

1.2.1.1 Cell fusion 8

1.2.1.2 Antitumor immune response activation 9

1.2.2 Family members 10

1.2.2.1 GALV.fus 10

1.2.2.2 Syncytin-1 11

1.2.2.3 VSV-G 12

1.2.3 Applications in tumor gene therapy 12

1.2.3.1 Enhanced antitumor effect 12

1.2.3.2 Difficulties in large-scale clinical application 14

1.3 CD40-CD40 ligand interaction 14

1.3.2 CD40 expression and function in human cells 15

1.3.2 Direct growth inhibition of cancer 18

1.3.3 Antitumor immune response activation 22

1.4 Purpose 26

CHAPTER 2 SELECTIVE KILLING OF CANCER CELLS BY A NOVEL VSV-G MUTANT THAT PROMOTES LOW PH-DEPENDENT CELL FUSION 28

2.1 Introduction 29

2.2 Materials and Methods 32

2.2.1 Cell culture 32

2.2.2 Mutagenesis, baculovirus preparation and cell transduction 33

2.2.3 Indirect immunofluorescence microscopy 34

2.2.4 Syncytia Formation Assay 35

2.2.5 Cytotoxicity assays 35

2.2.6 Reverse transcriptase polymerase chain reaction 36

2.2.7 Western blot 36

2.2.8 In vitro Boyden chamber cell migration assay 37

2.2.9 Animal studies to evaluate therapeutic efficacy 37

2.2.10 Statistical analysis 38

2.3 Results 40

2.3.1 Whole body biodistribution of intravenously administered NSCs in a mouse 4T1 breast cancer model 40

2.3.2 Mutagenesis of VSV-G 44

2.3.3 pH-responsive Properties of VSV-G(H162R) 47

2.3.4 In vitro bystander effect of VSV-G(H162R) 53

2.3.5 Establishment of VSVG(H162R)-expressing NSCs using baculovirus .57

2.3.6 Metastatic breast cancer therapy using VSVG(H162R)-expressing NSCs 62

2.3.7 Side effects of cancer therapy using VSVG(H162R)-expressing NSC .68

2.4 Discussion 70

CHAPTER 3 SELECTIVE KILLING OF CD40-POSITIVE BREAST CANCER CELLS BY NSC-MEDIATED DELIVERY OF CD40 LIGAND IN A MOUSE MODEL 75

3.1 Introduction 76

3.2 Materials and Methods 79

3.2.1 Cell culture 79

3.2.2 Baculovirus preparation and cell transduction 80

3.2.3 Cytokine antibody array 80

3.2.4 Cytotoxicity assays 81

3.2.5 Reverse transcriptase polymerase chain reaction 82

3.2.6 Fluorescent-activated cell sorting analysis 82

3.2.7 Animal studies to evaluate therapeutic efficacy 83

3.2.8 Statistical analysis 83

3.3 Results 85

3.3.1 Establishment of CD40L-expressing iPS-NSCs using a baculovirus .85

2.3.2 CD40L induces cytokine production in CD40-positive 4T1 breast cancer cells 87

2.3.3 In vitro bystander effect of CD40L-expressing iPS-NSCs 92

2.3.4 Metastatic breast cancer therapy using CD40L-expressing NSCs 95

2.3.5 Side effects of cancer therapy using CD40L-expressing NSCs 102

3.4 Discussion 104

CHAPTER 4 CONCLUSION 108

References 113

SUMMARY

Neural stem cells (NSCs) have recently emerged as one of the most attractive cellular vehicles for targeted gene delivery to cancers due to their migratory capacity and tumor tropism However, because NSCs can be chemoattracted to non-target regions, especially after intravenous administration, off-target transgene expression is a concern for the clinical application of NSC-mediated cancer gene therapy To minimize this side effect, therapeutic transgenes that enable tumor cell-selective killing are needed In this project, I developed two novel approaches to target tumor cells

The first approach uses vesicular stomatitis virus G glycoprotein (VSV-G) to target tumor acidosis VSV-G is a viral fusogenic membrane glycoprotein that kills tumor cells via syncytia formation Tumor acidosis is one of the hallmarks of the tumor microenvironment Here, I have discovered a novel VSV-G mutant that functions specifically at an acidic tumor extracellular pH, thus enabling VSV-G to selectively kill tumor cells

The second approach uses CD40 ligand (CD40L) to target CD40+ tumor cells CD40 is a type I TNF receptor that is selectively expressed on a number of epithelial and mesenchymal tumors, such as breast tumors, bladder tumors and lymphomas, but not on most normal, non-proliferating epithelial tissues

Expression of CD40L in CD40+ tumors leads to tumor growth inhibition via apoptosis induction and immunity activation

In our studies, both approaches showed significant therapeutic effects in vitro and

in vivo In a mouse model of 4T1 metastatic breast cancer, both VSV-G and CD40L delivered by NSC-based vectors obtained greater therapeutic efficacy and reduced less toxicity to normal tissues than the conventional HSVtk/GCV suicide gene therapy

These findings are of crucial importance in terms of clinical trials of NSC-mediated cancer gene therapy This study is the first to deliver tumor-targeted VSV-G and cytokine into tumor sites using NSC-based vehicles, and provides a feasible solution to the current safety issues of intravenously administered NSC

LIST OF PUBLICATIONS

Manuscripts

Detu Zhu, Lam Dang Hoang and Shu Wang Systemic delivery of fusogenic

membrane glycoprotein-expressing neural stem cells to selectively kill tumor cells

through low pH-induced cell fusion 2012 (submitted to Molecular Therapy)

Detu Zhu, Lam Dang Hoang and Shu Wang Selective killing of CD40-positive

breast cancer cells by NSC-mediated delivery of CD40 ligand in a mouse model

2012 (in preparation)

The experiments in the above noted manuscripts were performed during my PhD study The major findings were presented in this thesis.

LIST OF TABLES

Table 2.1 Primer pairs used for mutagenesis of VSV-G 45 Table 2.2 Median survival time and log rank test in the survival study 67 Table 3.1 Differential expression of cytokines in 4T1 cell cultures after CD40L treatment 91 Table 3.2 Statistics of tumor metastasis sites 99 Table 3.3 Median survival time and log rank test in the survival study 101

LIST OF FIGURES

Fig 2.1 Dual-color, whole-animal imaging to demonstrate the tumor tropism of iPS-NSCs .42

Fig 2.2 Dual-color, ex vivo imaging to demonstrate tumor tropism of iPS-NSCs 43

Fig 2.3 Mutagenesis of VSV-G .46

Fig 2.4 Syncytia formation assay between VSVG-expressing iPS-NSCs and 4T1 cells 50

Fig 2.5 Dual-color syncytia formation assay between VSVG(H162R)-expressing NSCs and 4T1 cells .51

Fig 2.6 Dual-color syncytia formation assay between VSVG(H162R)-expressing NSCs and tumors of different lineages 52

Fig 2.7 Comparison of HSVtk and VSVG(H162R)-mediated cytotoxicity on 4T1 breast cancer cells .55

Fig 2.8 In vitro killing effects of VSVG(H162R) in tumors of different lineages 56

Fig 2.9 Establishment of VSVG(H162R)-expressing iPS-NSCs using a baculovirus 59

Fig 2.10 Cell viability assays for BV-transduced iPS-NSCs .60

Fig 2.11 Boyden chamber migration assays for BV-transduced NSCs 61

Fig 2.12 In vivo 4T1 breast cancer therapy using NSC-VSVG and NSC-TK/GCV 64

Fig 2.13 Bioluminescent images of tumor growth in representative animals 64

Fig 2.14 Survival analysis .66

Fig 2.15 Hepato- and nephro-toxicities of NSC-VSVG or NSC-TK/GCV treatment .69

Fig 3.1 Establishment of CD40L-expressing NSC using baculovirus .86

Fig 3.2 FACS analysis of CD40 expression on 4T1 breast cancer cell surface 89

Fig 3.3 Cytokine antibody array .90

Fig 3.4 In vitro bystander effect of CD40L-NSCs on 4T1 breast cancer cells 94 Fig 3.5 In vivo 4T1 breast cancer therapy using NSC-CD40L .97 Fig 3.6 Ex vivo organ imaging to demonstrate inhibition of 4T1 tumor metastasis by NSC-40L 98

Fig 3.7 Survival analysis .100 Fig 3.8 Hepato- and nephro-toxicities of NSC-VSVG or NSC-TK/GCV treatment .103

AcMNPV Autographa californica multiple nucleopolyhedrovirus

bFGF basic fibroblast growth factor

DMEM dulbecco’s modified eagle’s medium

eGFP enhanced green fluorescent protein

FBS fetal bovine serum

GM-CSF granulocyte-macrophage colony-stimulating factor

HERV-W human endogenous retrovirus W

HIV-1 human immunodeficiency virus type 1

HLA human leukocyte antigen

HSV-1 herpes simplex virus type 1

HSVtk herpes simplex virus-thymidine kinase

ICAM intercellular adhesion molecule

iPSC induced pluripotent stem cell

MOI multiplicity of infection

NF-κB nuclear factor kappa light-chain-enhancer of activated B cells

NK cell natural killer cell

PBS phosphate buffered saline

RT-PCR reverse transcription polymerase chain reaction

SLAM signaling lymphocytic activation molecule

TNF tumor necrosis factor

TRAF tumor necrosis factor receptor-associated factor

CHAPTER 1

INTRODUCTION

1.1 Neural stem cells

Neural stem cells (NSCs) are a self-renewing and multipotent population that gives rise to the three major neural lineages (neurons, astrocytes and oligodendrocytes) throughout the central nervous system (CNS) NSCs are highly migratory and display an innate tropic behavior towards neoplastic lesions

(Aboody et al., 2000; Benedetti et al., 2000) Hence, NSCs are promising gene

delivery vectors for tumor-targeted therapy

1.1.1 Tumor tropism

Metastatic tumors are the most aggressive type of neoplasm in humans, characterized by a high infiltrative ability and resistance to conventional therapeutic treatments such as surgical excision, radiotherapy and chemotherapy Currently, gene therapy has emerged as a promising new approach for treating aggressive malignant tumors; however, the low efficiency of transgene delivery towards tumor sites limits the clinical application of this therapy To overcome this barrier, traditional viral vectors, such retrovirus (Rainov and Kramm, 2003; Rainov

and Ren, 2003), adenovirus (Immonen et al., 2004) and herpes simplex virus type

I (HSV-1) (Varghese and Rabkin, 2002) have been extensively explored as gene delivery vectors Nevertheless, the use of viral vectors raises several safety concerns, such as the risk of tumorigenesis caused by viral integration into the host genome and the wide stimulation of the host immune system Therefore, alternative gene delivery vectors are urgently required

Recently, it was reported that transplanted NSCs have a tropism not only toward a tumor mass, but also toward infiltrative “satellite” tumor cells in animal models

(Aboody et al., 2000; Benedetti et al., 2000; Glass et al., 2005), which makes

NSCs particularly promising for targeted therapies for metastatic tumors Furthermore, genetically modified NSCs carrying suicide genes, such as

thymidine kinase (TK) (Benedetti et al., 2000; Li et al., 2005; Uhl et al., 2005) and cytosine deaminase (CD) (Kim et al., 2006), anti-tumorigenic cytokines (Sims et al., 2008) or oncolytic viruses (Herrlinger et al., 2000; Tyler et al., 2009) were shown to

exert a strong cytotoxicity toward metastatic tumors via bystander effects

1.1.2 Cell source

The great potential of NSCs in regenerative medicine highlights the need for consistent and renewable sources for the collection or production of uniform human NSCs suitable for clinical applications Fetal or adult human brain tissues provide possible sources of primary human NSCs However, derivation of primary NSCs from either the adult human brain or from a fetus is an extremely invasive procedure and raises ethical and regulatory issues As cell therapy products, primary human NSCs are variable in quality and hold limited passaging capacity, posing a significant challenge for the large-scale preparation of cells with stable characteristics Although oncogene-mediated immortalization provides a means to overcome the drawback of limited life span of primary human NSCs, the suitability

of therapy application remains questionable due to the well-documented oncogenic potential of these genes The ability of human pluripotent stem cells such as human embryonic stem cells (hESCs) to generate virtually any differentiated cell type provides the possibility of using these cells as new sources

of human NSCs Self-renewing hESCs are inherently immortal, and their proliferation capacity is preserved during long-term cell culture Hence, they have become a reliable and accessible source of unlimited amounts of uniform human

stem cells (Zhang et al., 2001; Ben-Hur et al., 2004) However, despite their

unique potential, the use of hESCs remains ethically controversial, because the process of generating hESC lines involves the destruction of human embryos

To circumvent this problem, another type of human pluripotent stem cells, human induced pluripotent stem cells (iPSCs), which are generated through the reprogramming of adult somatic cells by forced expression of several

transcriptional factors, can be used as a new source of NSCs (Takahashi et al., 2007; Yu et al., 2007) Standardization of the large-scale mass production of cell

therapeutics is a prerequisite for the widespread application of cell therapy Thus, standardization in generating human iPSC-derived cells offers the potential for manufacturing large batches of uniform, allogeneic cell therapy products that can

be used in a similar way as pharmaceutical products and are sufficient for repeated treatments in multiple patients This standardization will help eliminate variability in the quality of cell therapeutics and facilitate the reliable comparative

analysis of clinical outcomes

1.1.3 Genetic engineering

In recent years, the insect baculovirus has captured great interest as a gene

transfer vector for in vitro and in vivo applications (Hofmann et al., 1995; Kost et al.,

2005; Hu, 2008) The most commonly used baculoviral vectors are derived from

Autographa californica multiple nucleopolyhedrovirus (AcMNPV) This insect DNA

viral vector has proven to be very effective in transducing many mammalian cells and providing a high level expression of transgenes The vector has a high cloning capacity, allowing for the accommodation and delivery of large functional genes or multiple genes Unlike many other gene therapy viral vectors, baculoviral vectors can be produced in serum-free cell culture medium, which eliminates the potential hazard of serum contamination with viral and prion agents from the donating animal An important advantage of using non-human, insect baculoviral vectors as gene therapy vectors is that it circumvents several inherent problems associated with using human viral vectors that are currently commonly used, such as the

pre-existing immunity directed against infectious human viruses For example,

specific antibodies against adenoviruses are detectable in 97% of individuals

(Bessis et al., 2004), which may eliminate viral particles and impair viral

transduction, decreasing both the level and duration of therapeutic gene expression Thus, baculoviral vectors are considered novel and promising gene therapy vectors In previous studies, baculoviral vectors have been demonstrated

to be useful for hESC engineering It was found that baculoviral vectors effectively mediate gene transfer to hESCs Using baculoviral vectors containing eGFP at an MOI of 100 PFU, up to 80% of the cells in the infected hESC clumps were eGFP-positive at day 2, as determined using flow cytometry In a following study,

the baculoviral vectors were reconstructed by including the rep 78/68 genes and

ITR sequences derived from AAV, which enabled them to integrate into the AAVS1 site in the human chromosome 19 and realize stable transgene expression in hES

cells (Zeng et al., 2007)

1.1.4 Side effects of intravenous injection

The systemic intravenous administration of NSCs is an attractive option, given that intravenous injection is a minimally invasive procedure, and the injected cells may home in on multiple intracranial tumor foci and solid tumors of a non-neural origin through the circulation However, the majority of intravenously injected cells become trapped in the lung and liver because of the narrow diameters of lung capillaries and liver sinusoids In a previous study, it was found that only a small proportion of NSCs (10%) injected into the systemic circulation could reached the tumor sites, whereas most of the injected cells were trapped in non-target organs

including the lung, liver and spleen (Yang et al., 2012; Zhao et al., 2012) One of

the safety concerns here is damage to these healthy organs caused by NSC-delivered cancer therapeutics As an example, in the HSVtk/GCV system that employs HSVtk to phosphorylate GCV and interfere with DNA replication in

tumor cells, the suicide gene products will eliminate not just cancer cells but also

other proliferating normal cells (Johnson et al., 2005) To avoid the possible side

effects that deteriorate the patient situation, tumor-targeted therapeutic gene is urgently needed

1.2 Fusogenic membrane glycoproteins

Fusogenic membrane glycoproteins (FMGs) are a class of glycoproteins derived from viral envelope genes that can induce cell membrane fusion and cytotoxicity in mammalian cells In the year 2000, a research group found that overexpression of FMGs in tumor cell cultures would induce the formation of gigantic, multinucleated syncytia that recruited 50-200 cells, thus leading to massive tumor cell death

(Bateman et al., 2000; Diaz et al., 2000; Higuchi et al., 2000) Since then, FMGs

have been extensively explored as antitumor agents

5-fluorocytosine (CD/5-FC) (Bateman et al., 2000; Diaz et al., 2000) The powerful

bystander killing effects of FMGs arise from the induction of local syncytia formation, the activation of antitumor immune response and the spread of

pro-apoptotic agents via syncytiosomes

1.2.1.1 Cell fusion

Overexpression of FMGs in tumor cells leads to massive cell to cell fusion, formation of gigantic, multinucleated syncytia and subsequent cell death in 2-5

days (Higuchi et al., 2000) The exact mechanism underlying the

syncytia-mediated cell death remains poorly defined, but both necrosis and apoptosis are involved In experiments using melanoma models, necrosis appeared to play a major role in the FMG-mediated syncytia death In these studies, it was described that signs of mitochondrial failure, ATP depletion and

autophagic degeneration were observed during the death of syncytia (Bateman et

al., 2002) Indeed, nuclear fusion was also observed, but the nuclei present in the syncytia could not be stained using the terminal deoxynucleotidyl transferase

dUTP nick-end labeling (TUNEL) assay (Bateman et al., 2000) However, in

experiments using glioma, pancreatic and colorectal cancer models, apoptosis appeared to play a major role in the FMG-mediated syncytia death In these studies, addition of the caspase inhibitor Z-VAD-fmk blocked syncytia death In contrast, feeding of fructose to the cultured tumor cells failed to inhibit syncytia

death (Hoffmann et al., 2006; Hoffmann and Wildner, 2006; Hoffmann et al., 2007)

Moreover, the hallmarks of apoptosis were also observed during syncytia death in the cultured tumor cells, in which 80-90% FMG-mediated syncytia were

TUNEL-positive on day 6 post-transfection (Galanis et al., 2001) Taken together,

these combined results suggest that the underlying mechanism of FMG-mediated cytotoxicity could be variable and cell line-dependent

1.2.1.2 Antitumor immune response activation

In addition to FMG’s fusogenic capacity, there is a rising interest in the ability of FMGs to activate an antitumor immune response Previous studies have found that the dying tumor syncytia induced by FMGs will release exosome-like vesicles,

termed syncytiosomes (Bateman et al., 2002) These vesicles not only spread

pro-apoptotic agents derived from the dying tumor syncytia, but also load dendritic cells (DCs) with tumor-associated antigens (TAAs) and then mediate T-cell

priming (Bateman et al., 2002; Errington et al., 2006) In a past animal study, a

cancer cell vaccine was produced by using a FMG to fuse allogeneic and autologous murine melanoma cells at a 1:1 ratio and was then applied to a mouse therapy model The results showed that these FMG-mediated fusing cells promoted a specific DC cross-priming against the tumor-associated antigens more efficiently than the conventional allogeneic melanoma vaccine, and significantly prolonged mice survivals in a second challenge with the same type of tumor

(Linardakis et al., 2002) Further studies have demonstrated that FMGs are

immunogenic and can activate DCs to potentiate IL-12 production and T cell

priming (Errington et al., 2006)

1.2.2 Family members

FMG is a broad class of proteins including gibbon ape leukemia virus (GALV)

FMG (Bateman et al., 2000; Diaz et al., 2000; Higuchi et al., 2000; Galanis et al., 2001; Bateman et al., 2002), vesicular stomatitis virus G glycoprotein (VSV-G) (Bateman et al., 2000), measles virus H and F proteins (MV H+F) (Bateman et al., 2000; Galanis et al., 2001), Syncytin-1 (Lin et al., 2010), human immunodeficiency virus-1 (HIV-1) gp120 (Ferri et al., 2000; Li et al., 2001; Scheller and Jassoy, 2001; Andreau et al., 2004; Nardacci et al., 2005; Perfettini et al., 2005) and others In

spite of the common membrane fusogenic capacity, the diversity in species sources of FMGs has provided a variety of additional properties, some of which may be beneficial for cancer gene therapy clinical trials The followings are several examples of FMGs that benefit therapy with their different aspects

its clinical application in cancer gene therapy remains limited due to the lack of

tumor specificity and the side effects induced (Guedan et al., 2011) Recently,

efforts have been made to enhance its tumor-targeting ability by regulation of

protease interaction (Johnson et al., 2003; Allen et al., 2004) or responsive promoters (Brade et al., 2003; Fu et al., 2003; Guedan et al., 2008) Despite using

these transgene regulatory methods, the leakage of GALV.fus fusogenic activity in

non-target cells is yet to be avoided (Kirkham et al., 2002)

1.2.2.2 Syncytin-1

Syncytin-1, also known as EnvERVWE1, is of great interest because it is an FMG derived from the type-D-related human endogenous retrovirus-W (HERV-W),

whose genome has been sequestered by the human host during evolution (Blond

et al., 1999) Syncytin-1 fuses human cells that express the type-D mammalian retrovirus receptor and is known to play a critical role in human placental

morphogenesis (Mi et al., 2000) A previous study has shown that overexpression

of syncytin-1 in tumors produces a strong bystander effect and leads to tumor

regression in vivo, although its therapeutic efficacy both in vitro and in vivo is not

as high as that of GALV.fus (Lin et al., 2010) Nonetheless, as syncytin-1 is a

human endogenous protein, using it for cancer gene therapy will not induce the pre-existing immunity problem observed in other external viral proteins

1.2.2.3 VSV-G

Vesicular stomatitis virus (VSV) is a member of the Rhabdoviridae, a family of enveloped, negative strand RNA viruses The virus enters human cells via a low pH-dependent membrane fusion within endosomes, which is mediated by a

homotrimerized envelope glycoprotein termed VSV-G (Matlin et al., 1982) Similar

to other viral fusogenic proteins, VSV-G undergoes a fusogenic structural transition during viral entry, and there is a pH-dependent equilibrium between the

different conformational states (Roche et al., 2006; Roche et al., 2007; Roche et

al., 2008) Therefore, the VSVG-mediated membrane fusion can be regulated in a pH-dependent manner With the well-known fact that acidosis is a hallmark of the

tumor microenvironment (Lee et al., 2008), VSV-G has been reported to preferentially function in tumor sites (Ebert et al., 2004)

1.2.3 Applications in tumor gene therapy

Since the year 2000, FMGs have been extensively explored as antitumor agents owing to both their potent cytotoxic effect on cancer cells and their capacity to

stimulate an antitumor immune response in vivo (Bateman et al., 2000; Diaz et al., 2000; Higuchi et al., 2000) In addition, FMGs are being intensively developed as

an adjuvant agent with other therapeutic strategies such as oncolytic virotherapy, chemotherapy and immunotherapy

1.2.3.1 Enhanced antitumor effect

Interestingly, during FMG-mediated cytotoxicity assays, caspase activities were

observed not only in the dying syncytia, but also in the non-fusing cells around the

syncytia (Bateman et al., 2000) Subsequent studies found that the dying syncytia

mediated by FMGs can release exosome-like vesicles, also known as syncytiosomes, and spread pro-apoptotic agents produced in the syncytia to the

neighboring cells (Bateman et al., 2002) Further studies have revealed that these

syncytiosomes can also spread chemokines, cytokines and progeny virons to the adjacent tumor cells In a previous study, an oncolytic herpes simplex virus (HSV) incorporated with fusogenic function showed an enhanced potency of viral oncolysis in various tumor cell lines including glioma, non-small cell lung

carcinoma and hepatoma (Simpson et al., 2006) A further study using a

combination of GALV.fus, CD/5-FC prodrug activation and oncolytic HSV showed

a significantly higher therapeutic efficacy than treatment with prodrug activation or

oncolytic HSV individually in a colon cancer model both in vitro and in vivo (Simpson et al., 2012) Another independent study using a combination of MV H+F,

gemcitabine chemotherapy and oncolytic adenovirus showed an enhanced synergistic therapeutic efficacy on colon and pancreatic cancers (Hoffmann and Wildner, 2006) In addition, FMGs are also intensively explored in other

researches as an adjunctive agent for oncolytic virotherapy (Ahmed et al., 2003; Ebert et al., 2004; Grisson et al., 2004), chemotherapy (Hoffmann et al., 2006; Hoffmann et al., 2007) and immunotherapy by cytokine genes (Eslahi et al., 2001; Hoffmann et al., 2007; Hoffmann et al., 2007), allogeneic tumor cell vaccines (Linardakis et al., 2002; Errington et al., 2006), DNA vaccines (Mao et al., 2010) or

exosomal vaccines (Temchura et al., 2008), because the syncytiosome released

from the FMG-infected syncytia can help distribute the oncolytic viruses, drugs, chemokines or cytokines throughout the solid tumor, or promote the cross presentation of tumor-associated antigens

1.2.3.2 Difficulties in large-scale clinical application

Despite the powerful bystander effects mentioned above, the application of FMG

in tumor gene therapy is impeded due to the lack of tumor targeting It has been reported that the generation of high-titer viral vectors encoding the GALV.fus is

difficult due to the rapid fusion of the producer cells (Diaz et al., 2000) Moreover,

animals treated with a GALV.fus-expressing virus have shown a significant increase in serum levels of liver enzymes, indicating side effects on the liver

(Guedan et al., 2011) As the FMGs may destruct the producer cells before

harvest or may damage healthy tissues if injected intravenously into the body, currently they are only applied via intratumoral injection, thus not suitable for treating metastatic tumors Therefore, methods for both local gene control and better gene delivery system are still required for a broader clinical application of FMGs

1.3 CD40-CD40 ligand interaction

CD40 is a type I membrane glycoprotein receptor of the tumor necrosis factor receptor (TNF-R) superfamily The natural ligand of CD40 is CD40L, a type II

membrane glycoprotein of the TNF family, also known variously as CD154, TRAP and T-BAM The CD40-CD40L interaction is best known for its multifaceted growth-regulatory functions in normal B cells and DCs In recent years, the CD40-CD40L interaction was also found to have a direct growth-inhibitory effect

on human breast, ovarian, cervical, bladder, non-small cell lung and squamous epithelial carcinoma cells via cell cycle blockage and/or apoptotic induction without obvious side effects on their normal counterparts (Tong and Stone, 2003) Thus, CD40L treatment is considered a promising tumor-targeted gene therapy for CD40+ cancers

1.3.2 CD40 expression and function in human cells

In the year 1985, CD40 was identified as a surface marker on bladder carcinoma

cells and on B cells independently (Paulie et al., 1989) The CD40 receptor is

constitutively expressed on highly proliferating cells (hematopoietic progenitors, epithelial cells and endothelial cells), all antigen-presenting cells (DCs, activated monocytes and activated B cells), CD8+ T cells and eosinophil granulocytes (van Kooten and Banchereau, 2000) CD40L, the natural ligand of CD40 receptor, was identified and isolated in activated T cells in 1992 by CD40-Fc fusion protein and

expression cloning techniques (Armitage et al., 1992) CD40L is transiently

expressed on activated leukocytes (mature CD4+ T cells, CD8+ T-cell subsets and

γδ T cells), IL-2 activated NK cells, mast cells and eosinophil granulocytes

(Gauchat et al., 1993)

The CD40-CD40L interaction is best known for its multifaceted activation functions

in normal B cells CD40 emerges early in CD34+ B-cell precursors in the bone marrow before immunoglobulin gene rearrangement and is expressed on B cells

until their eventual differentiation into plasma cells (Uckun et al., 1990) CD40

ligation leads to mainly pro-proliferative effects for resting B cells but inhibits the

growth and immunoglobulin production of activated B cells (Funakoshi et al., 1994; Garrone et al., 1995; Miyashita et al., 1997) The extent of CD40 activation

ultimately determines the differentiation path of the antigen-activated B cells Shortened CD40L exposure skews the mature B cells to eventually differentiate into plasma cells, whereas prolonged CD40L exposure generates CD40+ memory

B cells (Arpin et al., 1995; Quiding-Jarbrink et al., 1995)

Likewise, the interaction of CD40L+, activated T cells with CD40+ DCs has a critical role in the induction and maintenance of cellular immune responses (Grewal and Flavell, 1998) CD40L produces a pro-survival signal in CD40+ DCs and up-regulates the expression of co-stimulatory molecules (MHC class II, CD58 and B7) that enhance their antigen presentation functions This interaction in turn

‘‘primes’’ CD40L+ helper and cytotoxic T cells via increasing IL-2 receptor expression, thus leading to the expansion of both class I- and class II-dependent

pools of tumor-reactive effector and memory T cells (Grewal et al., 1995; Roy et al., 1995; Sin et al., 2001) Following CD40 activation, the antigen-presenting DCs

further promote the cellular immune response by up-regulating the expression levels of TNFα, MIP-1a, IL-8, IL-12 and SLAM that coordinately activate the

neighboring DCs and T cells (Kiener et al., 1995; Cella et al., 1996; Bleharski et al.,

2001) In addition, the IL-2 activated, CD40L+ NK cells exhibit an enhanced cytotoxic ability after engagement with the CD40+ target cells (Carbone et al.,

1997)

Epithelial CD40 receptor expression is mainly restricted to the self-renewing stem cells residing in the basal layer such as the basal/proliferative layer of the nasopharyngeal, tonsillar and ectocervical epithelium CD40 activation leads to a

uniform growth-inhibitory effect (Young et al., 1998) and phenotypic alterations that may impact the local inflammatory response (Denfeld et al., 1996) It has been

speculated that the activation of CD40, Fas or β1-integrin receptor, all of which are localized on basal epithelium, may direct the self-renewing epithelial cells towards

a pro-differentiation path (Bata-Csorgo et al., 1993) As CD40L is barely detected

in normal and malignant epithelial cells, the CD40 activation likely originates from CD40L+, infiltrating T cells, particularly the γδ T cells that congregate preferentially

at the epithelium during inflammation Alternatively, a CD40L analog that is unique

to epithelial sites may be expressed (Young et al., 1998)

The CD40 receptor expression on malignant cells generally mirrors their normal, proliferating counterparts, and has been detected on a variety of carcinoma cell

types such as melanomas, bladder carcinomas and breast carcinomas (Hess and

Engelmann, 1996; Thomas et al., 1996; Young et al., 1998) In accordance with

the growth- and immune-regulatory characteristics of the CD40-CD40L interaction

in normal cells, agonistic anti-CD40 monoclonal antibodies (mAbs), recombinant CD40L and CD40L analogs have been suggested to promote corresponding direct and indirect growth-inhibitory effects on these CD40+ tumors

1.3.2 Direct growth inhibition of cancer

Similar to the multifaceted activation effects of CD40-CD40L interaction on normal cells, CD40L has a diverse growth-regulatory effect on CD40+ tumor cells In low-grade B-cell malignancies such as follicular lymphomas, hairy cell leukemia and B chronic lymphocytic leukemia, CD40 ligation promotes tumor survival and

resistance to chemotherapy (Schattner et al., 1996) Likewise, CD40 ligation

promotes the growth of HIV-related lymphomas, probably via an increase in vascularization and the activation of NF-κB signaling, which in turn stimulates HIV

replication (Berberich et al., 1996) In addition, treatment with soluble, recombinant

CD40L leads to dose-dependent proliferation and colony formation of leukemia

blasts in vitro In contrast, CD40 ligation induces cell cycle arrest in B lymphoma

cells, which is important for the initiation and maintenance of tumor dormancy

(Marches et al., 1995) In post-transplant, high-grade aggressive lymphomas,

CD40 ligation leads to the inhibition of growth via apoptosis induction

(Vyth-Dreese et al., 1998) Likewise, in non-HIV, high-grade malignancies such as

Burkitt’s lymphoma, CD40 activation results in growth arrest, up-regulation of Fas

and increased apoptosis (Schattner et al., 1996) Similarly, treatment with soluble,

recombinant CD40L induces growth arrest and apoptosis in human multiple

myeloma cells which are highly expressing the CD40 receptor (Pellat-Deceunynck

et al., 1994)

In the year 1996, a direct growth-inhibitory effect through CD40 ligation of epithelial cancers was first demonstrated, in which CD40L transgene expression

in HeLa cervical carcinoma cells led to a dramatic growth arrest in vitro (Hess and

Engelmann, 1996) This CD40-mediated growth-inhibitory effect appears not to be mediated by apoptosis in studies using bladder, ovarian and skin carcinoma models, and is additive with the pro-apoptotic activities of cytotoxic cisplatin and

other immune stimuli such as TNFα, Fas ligand (FasL) and ceramide (Eliopoulos

et al., 1996) For non-small cell lung carcinomas, CD40L treatment results in a dose-dependent, reversible cell cycle S-phase arrest in the tumor subset that

expresses a high level of CD40 receptor (Yamada et al., 2001) For breast and

squamous carcinoma lines, apoptosis induction appears to play a major role in the

CD40-dependent growth-inhibitory effect (Eliopoulos et al., 1996) Likewise, CD40L inhibits melanoma cell proliferation in vitro via the induction of apoptosis,

the enhancement of tumor-specific lymphocyte-mediated tumor cytotoxicity and the stimulation of pro-inflammatory cytokine production, including IL-6, IL-8 and

TNFα, by melanoma cells (von Leoprechting et al., 1999)

Previous animal studies using severe combined immunodeficiency (SCID) mice

the in vivo antitumor effect of CD40L in the absence of T- and B-cell activation, for

which different constructs of soluble CD40L produced a uniform growth-inhibitory

effect (Funakoshi et al., 1997; Hirano et al., 1999; Tong et al., 2001) Co-treatment with the CD40L-blocking antibody LL48 abolished these antitumor effects in vitro and in vivo, indicating the dependence on CD40 in these treatments In contrast,

CD40L is ineffective in altering the growth rates of the CD40- tumors Altogether, these animal experimental results confirm the direct growth-inhibitory effect of CD40L on CD40+ cancer cells in vivo

Currently, the underlying molecular mechanisms of CD40-mediated cancer growth regulation are still unclear Recent studies have shown that CD40L inhibits cancer

cell growth via activation of endogenous apoptotic pathways (Eliopoulos et al., 1996; Eliopoulos et al., 1997; Gallagher et al., 2002) Whereas CD40 lacks the

death domain and its link to caspase 1 activation, apoptosis may nevertheless be initiated through the interaction of TRAF, through the transactivation of other TNF-R members that express the death domain, or by collaterally activating the

caspase cascade (Baker et al., 1998; Wingett et al., 1998; Grell et al., 1999; Gallagher et al., 2002) These events depend on the integrity of the CD40 membrane proximal domain but not its TRAF-interacting motif (Werneburg et al.,

2001) An up-regulation of pro-inflammatory cytokines, including TNFα, IL-6, IL-8

and GM-CSF (Alexandroff et al., 2000), and the activation of corresponding caspase (Keane et al., 1996; Srinivasan et al., 1998; Grell et al., 1999) may also

contribute to CD40-mediated tumor growth inhibition

The CD40 receptor expression has been detected in breast, colorectal, bladder, ovarian, and liver carcinomas, whereas CD40L is not endogenously expressed or

is expressed at very low levels Strong expression levels of the CD40 receptor in the tumor vasculature of renal and breast carcinomas indicate a potential role of

this receptor in tumor angiogenesis (Kluth et al., 1997; Tong et al., 2001)

Particularly, CD40 is expressed on all primary breast tumor histologic subtypes, including infiltrating ductal and lobular carcinomas, and carcinomas in situ Although CD40L is detected within the cytoplasm of the tumor population in the focal areas, this molecule is barely membrane stained, suggesting that expression

of the membrane-bound CD40L may be uncommon, or occurs transiently Restricted endogenous CD40L expression may limit its capacity for inhibiting

tumor progression (Tong et al., 2001) Furthermore, CD40L was rarely expressed

among tumor-infiltrating lymphocytes (TILs) in the majority of breast cancer cases tested, suggesting that TILs in established breast cancers may lack the ability to

down-regulate tumor cell growth via the CD40-CD40L interaction (Tong et al.,

2001)

1.3.3 Antitumor immune response activation

The CD40-CD40L interaction has been shown to stimulate an antitumor immune response in a number of mouse tumor models In a previous animal study, CD40 knock-out mice had a significantly increased occurrence of spontaneous tumors and, unlike mice with a wild-type genetic background, could not mount a protective tumor-specific immune response against a secondary tumor challenge following sensitization with a GM-CSF-transfected, syngeneic B16 melanoma cancer cell

vaccine (Mackey et al., 1997) In contrast, intratumoral injection of an agonistic

anti-CD40 mAb induced a substantial, systemic, tumor-specific cytotoxic T lymphocyte (CTL) response that was sufficient to destroy pre-existing CD40+

tumors (van Mierlo et al., 2002) Moreover, co-immunization with an anti-CD40

cross-linking antibody and the cRL1 tumor peptide provoked a cytotoxic immune

response against tumor and prolonged survival of tumor-bearing mice (Ito et al.,

2000) In another animal study, 94% of syngeneic mice bearing MB-49 bladder cancer xenografts exhibited tumor regression following immunization with CD40L-transduced tumor cell vaccines, which protected against a subsequent parental, untransfected tumor challenge Furthermore, rapid infiltration of large numbers of CD4+ and CD8+ T cells in the CD40L-expressing tumor anlage was observed, indicating that immunosensitization contributed significantly to this

antitumor effect (Loskog et al., 2001)

The immunostimulatory effects of CD40L correlate with its capacity to activate

DCs in vivo (Grewal et al., 1995; Kikuchi and Crystal, 1999; Urashima et al., 2000;

Todryk et al., 2001) CD40 ligation produces a pro-survival effect on the CD40+

tumor-infiltrating DCs, attained partially via the activation of the anti-apoptotic

serpin serine protease inhibitor 6 in the DCs (Esche et al., 1999; Medema et al., 2001) In vitro treatment of soluble, recombinant CD40L promotes

monocyte-derived DC maturation, with an increased expression of adhesion and co-stimulatory molecules (ICAM-1, CD83 and CD80/86), and an up-regulated production of pro-inflammatory cytokines and chemokines (IL-12, IL-6, TNFα and

MIP-1a) (Wurtzen et al., 2001) Furthermore, CD40L exposure is required to enhance the antigen-presenting functions of DCs following in vitro maturation treatments of GM-CSF and IL-4 (Kikuchi et al., 2000) Activation of CD40 signaling

in the presence of pro-inflammatory cytokines such as IFNγ is necessary for the

‘‘cross-priming’’ function of DCs, by which the ingested tumor apoptotic body-derived antigens are presented to T cells in the context of HLA class I, thus

leading to the activation of tumor-specific effector CTLs (French et al., 1999)

expression and acquire an enhanced capacity of producing IL-12 in response to CD40 engagement by CD40L+, activated T cells (MacDonald et al., 2001), which

is important for initiating a TH1 response (De Becker et al., 1998) Moreover,

CD40-activated, monocyte-derived DCs play a determinant role of skewing the CD4+ T cell from a TH2 phenotype to an IFNγ+ TH1 phenotype (Terheyden et al.,

2000) An equally important role of CD40-CD40L interaction has also been described for initiating an immune response against TH2 infectious antigens