T cell responses in helicobacter pylori associated gastroduodenal diseases 1

Bibliography 126 SUMMARY This project begins with investigating the adaptive T cell immunology of Helicobacter pylori HP infection of the gastric mucosa in human patients, and concludes

T CELL RESPONSES IN HELICOBACTER PYLORI-ASSOCIATED

Thank you to the undergraduate students who have taken up parts of my research project - Christine Lee, Cassandra Ho, and Tan Hui Qing Your help in the lab has been immense and I have learnt much from you during the short course of your research career

My friends in the programme – Hazel, thank you for your help in every way,

everyday, I’m lost without you Kar Wai, thank you for always being one step ahead

of me Fei Chuin, thank you for your advice in both lab and non-lab stuff Veronique, thank you for your advice and help in picking up IF techniques All other lab

members, Daniel, Chenyu, Karen and Amos, thank you for the intellectual and not-so intellectual exchanges

Finally, I thank both my parents and aunties Rosie and Betty For letting me hog their car to travel between SGH and NUS I would not have achieved as much in this 4 years without this help Thank you all for your un-ending and unconditional support during the course of my (very long) education

TABLE OF CONTENTS

Acknowledgements II

Table of contents IV

Summary VII List of Tables IX List of Figures X List of Abbreviations XII 1 Introduction 1 1.1 Immune evasion of pathogens – failure of the human immune system 1

1.2 Helicobacter pylori-associated diseases 9

1.3 Helicobacter pylori and gastric cancer risk 13

1.3.1 Epidemiological studies 13

1.3.2 The COX-2 pathway 15

1.3.3 Genetic polymorphisms in immune genes 16

1.4 Adaptive immune responses to Helicobacter pylori 18

1.4.1 Sensing HP bacterium 18

1.4.2 Protective T cell responses 19

1.4.3 Mouse models of HP infection 20

1.4.4 T regulatory responses 22

1.5 T helper-17 responses and cytokines associated with Helicobacter pylori infection 24

1.6 Th17-Associated immunopathology 29

1.7 Rationale and hypotheses 34

2 Materials and Methods 37

2.1 Patient samples and classification 37

2.2 Isolation and culture of peripheral blood mononuclear cells 37

2.3 Isolation and expansion of primary human gastric epithelial cells and lamina propria mononuclear cells from gastric biopsies 38

2.4 GEC stimulation experiments 39

2.5 Culture of cell lines 39

2.6 Intracellular antibody labelling and flow cytometry 40

2.7 Immunofluorescent antibody labelling of frozen sections 41

2.8 HP lysate preparation co-culture experimental setup 41

2.9 mRNA isolation and real time PCR 42

2.10 Cytokine quantification 43

2.11 SDS-PAGE and western blot 44

2.12 Thymidine incorporation assay 45

2.13 Chemotaxis assay 45

2.14 Statistical analyses 46

3 Results 49

3.1 FUEL 49

3.1.1 Patient demographics and histopathology 49

3.1.2 Th17 frequencies remained elevated in the blood of individuals with previous HP infection up to 10 years post-HP eradication 51

3.1.3 Th17 and IL-17A levels remained high in the gastric mucosa of individuals with previous HP infection and this associates with PC lesions 54

3.1.4 Gene expression and cytokine environment in the gastric mucosa is highly pro-inflammatory in group A subjects and low levels of inflammation exists in group P subjects 57

3.1.5 IL-22 in the gastric mucosa 65

3.1.6 Strong intra-individual correlation of pro-inflammatory cytokines in group P subjects 66

3.1.7 Increased expression of IL-1R1 on IL-17A-expressing cells in group P individuals 68

3.1.8 Pro-inflammatory cytokines induced hBD-2 mRNA expression from human GECs 72

3.1.9 IL-1β-stimulated GEC supernatant is chemoattractive to autologous CCR6+ Th17 cellls 76

3.1.10 HP-specific Th17 responses persist in subjects with evidence of past HP infection 77

3.2 FLAMES 85

3.2.1 Expression of cytokine receptors IL-22R1, IL-1R1 and IL-17R1 in the human gastric epithelium and their regulation by inflammation 85

3.2.2 IL-22, IL-1β and IL-17A enhances gastric epithelial cell proliferation 87

3.2.3 IL-22-induced genes in gastric epithelium 91

3.3 SUMMARY 98

4 Discussion 103

4.1 Active IL-17A versus chronic IL-17A 103

4.2 Other sources of gastric IL-17A 109

4.3 The mixed CD4+ T cell response and T helper cell plasticity 112

4.4 Tight regulation of IL-1β production 119

4.5 Th17 cells in the development of gastric cancer – Role of STAT3 121

4.6 Concluding remarks and future directions 126

5 Bibliography 126

SUMMARY

This project begins with investigating the adaptive T cell immunology of

Helicobacter pylori (HP) infection of the gastric mucosa in human patients, and

concludes in the study of the effects of chronic inflammation on gastric epithelial cells HP causes a chronic on active gastritis and is associated with an increased risk

of developing the intestinal type of gastric cancer A large body of evidence suggests that the cancer develops via a chronic progression of precancerous changes in the gastric mucosa In many instances, risk of cancer progression is still evident after the bacteria is eradicated Concurrent with the cancer risk that precancerous changes associate with, the chronic immune cell infiltrate is apparent for years subsequent to antibiotic treatment Therefore, it is the main hypothesis of this project that the

immune response triggered by HP infection persists even after eradication of the bacteria, and that this chronic inflammation causes immunopathology to the gastric epithelium, which may subsequently predispose to the persistence of precancerous lesions and contribute to cancer risk

Patient subjects were classified into three main groups for most of the analyses – group A: subjects with an active HP infection, group P: subjects with no active HP infection, but evidence of a past infection and group N: nạve individuals who have never had a HP infection Focus was given to the comparison of antigen-specific and non-antigen-specific Th17 responses and its associated cytokines in these subjects

Presence and persistence of HP-specific Th17 cells were demonstrated in the

peripheral blood and gastric mucosa of subjects in group P, and in some individuals,

the persistence lasted up to 10 years post HP eradication therapy Ex vivo cytokine

concentrations of a panel of human pro-inflammatory and T cell cytokines in gastric biopsies were quantified IL-1β, IFNγ and IL-17A were the cytokines found to be significantly elevated in group P subjects compared to group N Several other pro-inflammatory cytokines were also elevated in group P subjects compared to group N, and an association with precancerous lesions was found Expression of IL-22, another Th17 cytokine was demonstrated in the gastric mucosa of all subjects and no

difference was detected across the three groups In vitro experiments on gastric

epithelial cells demonstrated that IL-22 acted in synergy with IL-1β and IL-17A to elicit pro-inflammatory downstream effects like the induction of S100 proteins and human β-defensin 2

This thesis takes an old model of infection- and inflammation-associated cancer and studies the related immune responses in light of new advances in T cell immunology The clinical data presented here further contributes to the field of knowledge on T cell responses in HP-associated gastritis and precancerous lesions

LIST OF TABLES

1.1 IL-17 target genes 30

2.1 List of antibodies 47

2.2 List of PCR primers and their cycling conditions 47

3.1 Clinical sample demographic data 49

3.2 Demographic data of subjects who were followed-up after HP eradication therapy 50

3.3 Summary of median and IQR values for IL-17A and IFNγ values in the three subject groups 55

3.4 Summary of median and IQR values for ex vivo cytokine concentrations reported in figure 3.4 61

3.5 Summary of median and IQR values for ex vivo cytokine concentrations reported in figure 3.5 62

3.6 Summary of median and IQR values for IL-22 protein data reported in figure 3.6 66

3.7 Summary of IL-17A responses of CD4+ PBMCs to HP-pulsed APCs 80

3.8 Summary of IL-17A responses of LPMCs to HP-pulsed APCs 82

LIST OF FIGURES

1 General hypothesis of gastric carcinogenesis 11

3.1 Progression/regression of precancerous (PC) lesions 3 years later in subjects who underwent follow-up endoscopy 50

3.2 Th17 frequencies in PBMCs 53

3.3 IL-17A levels and Th17 frequencies in patient gastric biopsies 56

3.4 Pro-inflammatory cytokine and chemokine profile in gastric environment 63

3.5 Other pro-inflammatory cytokines and chemokines analysed 64

3.6 IL-22 levels in human gastric biopsies 67

3.7 Ex vivo IL-17A levels in biopsies correlates with several other pro-inflammatory cytokines 69

3.8 High frequencies of IL-1R1 bright cells in group P individuals with severe IM 71/72 3.9 IL-1β, IL-22 and IL-17A induces hBD-2 expression from freshly isolated human GECs 75

3.10 Expanded LPMCs and chemotaxis assay towards IL-1β-stimulated GEC supernatant 78

3.11 HP-specific IL-17A responses in PBMC and LPMCs 81

3.12 Cytokine receptor expression and regulation by inflammation 86

3.13 Regulation of IL-1R1 and IL-22R1 in AGS cells by inflammatory stimuli at various time points 88

3.14 Increase in epithelial cell proliferation in the presence of gastric cytokines 90

3.15 Human β-defensin expression regulation by IL-22 in AGS and KatoIII cells 92

3.16 S100A7-9 expression regulation by IL-22 in AGS and KatoIII cells 93

3.17 Other classes of proteins being regulated by IL-22 in AGS and KatoIII cells 94

3.18 Synergy between IL-22 and IL-17A in AGS cells 95 3.19 Synergy between IL-22 and IL-1β in AGS and KatoIII cells 96 3.20 Synergistic effects between IL-22 and IL-1β in primary GECs 97

LIST OF ABBREVIATIONS

PC Precancerous

SDS-PAGE Sodium dodecyl-sulphate polyacrylamide gel

1 INTRODUCTION

1.1 Immune Evasion By Pathogens – Failure of the Human Immune System

Since the beginning of life, pathogen-host interactions have shaped the co-evolution

of the host immune system and pathogen survival mechanisms, resulting in the

constant evolutionary development of the pathogens’ invasive or evasive mechanisms

to either protect itself from, or attack the host in order to survive Until today, the human immune system has not perfected its defense mechanisms against many pathogens, henceforth resulting in chronic disease and death Even though the

immune system is designed to fight infections and protect itself from pathogens, it seems that it is falling short in its role when pathogens acquire abilities to escape or counter immune responses, and chronic infections with some of these pathogens develop into diseases such as cancer This first chapter briefly reviews the

mechanisms of immune evasion of pathogens, focusing on Helicobacter pylori, with

the aim of highlighting the ability of the pathogen to survive within, and cause

damage to the human host, resulting in chronic infections and unresolved

inflammation It concludes in reviewing the link between chronic inflammation, immunopathology and cancer, thus establishing the hypothesis of this thesis – chronic

inflammation in the gastric mucosa succeeding H pylori infection contributes to the

progression of gastroduodenal diseases, in particular, pre-cancerous gastric lesions and cancer

The human immune system is made up of two compartments – the innate and the adaptive The innate immune system holds the ability to mount an acute response,

almost immediately when encountering a foreign pathogen, thereby conferring the host quick protection The innate immune response is made up of a number of

components – 1) mucosal epithelial barriers and the antimicrobial substances they produce, 2) the innate immune cells, neutrophils, macrophages, natural killer (NK) cells and gamma delta (γδ) T cells, and 3) the complement system and inflammatory mediators in the blood The innate immune system recognizes pathogen-associated molecular patterns and reacts to these foreign bodies

The adaptive immune system, in contrast to the innate immune system, is not poised

to act immediately upon infection Infections trigger adaptive immune responses that are specific for a particular protein antigen from the pathogen, and this immune response takes 1-4 days to develop to its full strength However, once acquired, the adaptive immune system ‘remembers’ this particular pathogen and can respond more vigorously and quickly to subsequent re-infections The adaptive immune response is elicited by T and B lymphocytes The T cells are able to secrete cytokines and mount cytotoxic responses on infected cells, whilst B cells differentiate and mature into antibody secreting plasma cells, to secrete antibodies against the specific pathogen These cells are retained in the hosts’ circulation as ‘memory cells’ and are ready to respond to re-infections

Even with the elaborate structure and multitude of functional abilities our immune system has, humans are still prone to diseases The reason for this is that pathogens, in the face of pressures exerted by the human immune system, evolve mechanisms in attempts to evade or counteract the host’s immune responses The rest of this chapter

reviews some classic examples of mechanisms developed by pathogens to survive human immune responses

The human cytomegalovirus (hCMV) and NK cells of the human innate immune system are a classic example of the co-evolutionary struggles between pathogen and host, each to ensure its own survival (Loenen, Bruggeman et al 2001; Orange, Fassett

et al 2002; Lanier 2008) hCMV was found to express a MHC class I homologue UL18 (Farrell, Vally et al 1997), later found to function in immune evasion by the ligation of inhibitory receptors on NK cells, thereby preventing NK cells from lysing the infected cells (Leong, Chapman et al 1998) Unlike viruses, bacteria are much more complex organisms, having a cell wall and complex cellular processes Their

strategies for immune evasion are also very diverse For example, Pseudomonas

aeruginosa and Yersinnia enterocolitca are bacteria possessing a type III secretion

system that is utilized to introduce virulence factors into host cells to disrupt the actin cytoskeleton and hence inhibit the process of phagocytosis (Amiel, Lovewell et al ;

Grosdent, Maridonneau-Parini et al 2002; Davis, Rasmussen et al 2005) Listeria

monocytogenes escapes phagosomes by secreting a cytolysin listeriolysin O (LLO)

that creates pores in the phagosome, thereby allowing its escape (Beauregard, Lee et

al 1997) Other bacteria have developed resistance and counter mechanisms to the host’s bactericidal factors like reactive oxygen species, low pH, lysozymes,

lactoferrin, defensins, cathelicidins, endopeptidases and exopeptidases produced by phagocytes (Flannagan, Cosio et al 2009)

H pylori (HP) is a Gram-negative, microaerophilic spirochete which infects the

gastric mucosa of roughly 50% of the world’s population (Everhart 2000) It has a

huge array of mechanisms by which it copes with surviving in the human gastric mucosa and evading immune responses, hence making it very successful in

establishing long-term chronic infections in the human stomach Firstly, it adapts to the low pH in the human gastric environment by the secretion of urease, a highly immunogenic enzyme which breaks down urea in the gastric juice to ammonia and carbon dioxide to neutralize the acid in their surroundings (Hawtin, Stacey et al 1990) This allows the pathogen to survive the low pH (pH 1.0) in the gastric

environment The bacterium is usually found in the mucous layer of the stomach, in close contact with the gastric epithelia It seldom invades the epithelial barrier,

although some groups report the invasion of bacteria past the epithelium, and an isolated report found HP in the regional lymph nodes (Ito, Kobayashi et al 2008) The first step of HP’s immune evasion lies in its lipopolysaccharide (LPS) LPS is a

pathogen-associated molecular pattern0 (PAMP) expressed on bacterial cell walls and

is sensed by pattern recognition receptor (PRR) Toll-like receptor-4 (TLR4) on

antigen presenting cells (APCs), triggering the start of an immune response against most bacteria HP LPS was reported to be much less effective (500-1000 fold) in eliciting a cytokine response (TNFα, IL-1 and IL-6) as compared to LPS from other

bacterial pathogens like E coli and S typhimurium (Muotiala, Helander et al 1992;

Birkholz, Knipp et al 1993), hence resulting in a sub-optimal TLR4 stimulation Contributing to the lack of immunogenicity of HP LPS is the fact that its O antigen contains several human Lewis antigens (Wirth, Yang et al 1999) This may enhance molecular mimicry and cause the sub-optimal immunogenicity of the LPS Other groups have reported the activation of other PRRs by HP, both intracellular and extracellular sensors like TLR2 and TLR5 (Smith, Mitchell et al 2003), TLR9 (Rad, Ballhorn et al 2009) and Nod1 (Viala, Chaput et al 2004)

Several pathogenic components of HP act on both the gastric epithelium, to cause damage that drives carcinogenesis, as well as on cells of the immune system, to inhibit or deter immune responses against it

The most widely studied bacterial component of HP is the CagA protein, which is encoded within the Cag pathogenicity island This island of genes encodes the protein machinery required to form the bacteria’s type IV secretion system (T4SS) Via the T4SS, CagA is secreted into the host cell and becomes phosphorylated (Backert, Ziska et al 2000; Backert, Muller et al 2001; Moese, Selbach et al 2001) by Src and Abelson murine leukemia viral oncogene homolog (ABL) kinases on EPIYA motifs (Selbach, Moese et al 2002; Stein, Bagnoli et al 2002; Tammer, Brandt et al 2007) Phospho-cagA activates SHP2 phosphatase and ultimately results in a pro-

proliferative signal for the epithelial cell (Higashi, Tsutsumi et al 2002; Higashi, Nakaya et al 2004) CagA also induces the expression of pro-inflammatory cytokines and chemokines via NF-κB, p38 and ERK activation in gastric epithelial cells (Ogura, Takahashi et al 1998; Keates, Keates et al 1999; Meyer-ter-Vehn, Covacci et al 2000; Brandt, Kwok et al 2005) Pro-inflammatory cytokines like IL-6, IL-8 and chemokine CXCL2 are induced via activation of kinases (Sharma, Tummuru et al 1998)

In 2004, Viala et al discovered the sensing of HP peptidoglycan (PGN) by an

intracellular pattern recognition molecule, NOD1 (Viala, Chaput et al 2004) PGN also enters the host epithelial cell via the T4SS and elicits a similar response as CagA, the activation of NF-κB, p38 and ERK and subsequent pro-inflammatory cytokine and chemokine production (Allison, Kufer et al 2009) Intracellular CagA also causes

the disruption of cell-cell junctions, by associating with ZO1 and the junctional adhesion molecule A (JAMA) (Amieva, Vogelmann et al 2003) leading to a loss of epithelial barrier function and hence allowing further invasion of the pathogen

Another HP protein, that is widely studied due to its effects on epithelial cells and roles in immunosupression is the vacuolating cytotoxin (VacA) It causes vacuolation

in gastric epithelial cells and induces apoptosis in these cells upon binding (Cover and Blaser 1992; Kuck, Kolmerer et al 2001; Cover, Krishna et al 2003) VacA is

reported to induce immunosuppressive effects like the suppression of IL-2 production via preventing NFAT translocation into the nucleus, the inhibition of the IL-2

signalling pathway by forming anion channels, preventing calcium influx

(Boncristiano, Paccani et al 2003) and inhibition of cell proliferation (Sundrud, Torres et al 2004) VacA is recently reported to enter T cells via binding to an

integrin subunit CD18 (β2) and exploiting the cellular shuttling of integrins (LFA-1) and their receptors from the cellular surface to intracellular lipid rafts (Sewald,

Gebert-Vogl et al 2008)

Outer membrane proteins (OMPs) expressed by HP have specific receptors on the gastric epithelium in which they bind to For example, BabA, SabA and OipA are adhesins expressed on HP which binds to sialyl-Lewisx antigen on the gastric

epithelium (Mahdavi, Sonden et al 2002) OipA (also known as HopH) also

facilitates the binding of HP to the gastric epithelium Furthermore, it induces

proinflammatory cytokines IL-6 and IL-8 and chemokine RANTES secretion by gastric epithelial cells (Yamaoka, Kudo et al 2004; Kudo, Lu et al 2005; Lu, Wu et

al 2005) as well as activates β-catenin signaling, which is linked to carcinogenesis (Franco, Johnston et al 2008)

An important event in the pathogenesis of HP infections is the induction of nitric oxide (NO)-dependent macrophage apoptosis (Albina, Cui et al 1993; Gobert, Cheng

et al 2002; Gobert, Mersey et al 2002) This process highlights a fine example of the failure of the immune system’s defence mechanisms in fighting HP infection The production of free radicals like superoxide, hydrogen peroxide and NO is a basic innate immune response mounted against bacterial pathogens upon their entry into phagocytic cells These free radicals are produced within the phagosomes to kill the internalized bacteria However, it is ironical that NO itself triggers macrophage apoptosis, thus killing the cell (Albina, Cui et al., 1993) HP modulates the NO levels

in the cell by inducing the activity of inducible nitric oxide synthase (iNOS) in

macrophages (Wilson, Ramanujam et al 1996) and hence the production of NO Increase in NO concentrations in cells induces cellular stress and an increase in p53 levels within the cell Following this, p53-dependent apoptosis follows In addition to this, NO also induces apoptosis via the mitochondrial pathway – causing release of pro-apoptotic factors from the mitochondria It was recently shown that HP induces arginase II activity in macrophages during chronic HP infection (Lewis, Asim et al 2011) Arginase II competes for the L-arginine substrate with iNOS, thereby limiting the production of NO, and hence bacterial killing There is also a body of literature reporting the ability of HP to control and skew the T cell response to favor a

regulatory rather than pro-inflammatory response to itself The T cell responses to HP will be reviewed separately in another section

Considering the multitude of mechanisms that HP possesses to ensure in its survival

in the host, it is not surprising that infections can last for prolonged periods of time - even up to decades The long-term establishment of infections can have adverse

effects on the host – for example, chronic, unresolved inflammation, which is now established to cause pathology to the host In 1863, Rudolf Virchow first hypothesized the role of inflammation in cancer development He described the presence of large numbers of leukocytes in tumor tissue and hence established the link between cancer and inflammation, and inferred that sites of chronic inflammation were the ‘origins of cancer’ Succeeding this hypothesis, reports and findings up till today have supported this notion, and we now know that several types of cancers are associated with

chronic infections and inflammation The International Agency for Research on

Cancer (IARC) monographs 2011 lists several pathogens to be strong risk factors for one of more types of cancers: Epstein Barr virus, hepatitis B and C virus, human papilloma virus, human T-cell lymphotropic virus type 1, Kaposi Sarcoma

Herpesvirus (HHV-8), human immunodeficiency virus, liver flukes Opisthorchis

viverrini and Clonorchis sinensis, Schistosoma heamatobium and Helicobacter pylori

These pathogens serve as examples demonstrating the failure of the human immune system in protecting itself from infections, and that there are short-comings in the mechanisms of our self-defense that leads to outcomes which may not be directly related to the initial infection (eg cancer) The next few chapters of this introduction will focus on the central theme of this thesis – the association between chronic T cell responses, inflammation and precancerous lesions The evidence in the literature suggesting a role for immunopathology and cancer progression and the nature of

cytokine and adaptive immune response against HP will be reviewed and discussed in relation to the central theme

1.2 Helicobacter pylori-Associated Diseases

Humans are the only known reservoir of HP infection and the incidence rate varies

from country to country A recent review by Goh et al in Helicobacter summarizes

the incidence rates reported in many different locations all over the world, showing the large variability The most recent reports in Singapore and Northern Malaysia reported 51% (Zhu, Loh et al 2009) and 30.4% (Sasidharan, Lachumy et al 2011) respectively In the rest of the world, the prevalence is at an average of about 50% However, the prevalence rates vary from country to country and are affected by factors such as ethnicity, socioeconomic status of the country and living conditions (Goh, Chan et al 2011)

HP infection results in a multitude of diseases These include duodenal or peptic ulcers (1-10% of infected patients), gastric cancer (0.1-3%) and gastric mucosa-associated lymphoid tissue (MALT) lymphoma (<0.01%) (McColl 2010)

Autoimmune gastritis is a less commonly diagnosed form of gastritis, which typically leads to gastric atrophy (similar to HP-associated gastritis), hypochloridria and

eventually, pernicious anemia The autoantigen described to be responsible for this autoimmunity is the H+

,K+

-adenosine triphosphatase (ATPase) proton pump The autoreactive peptides from this protein resemble several peptides derived from HP antigens and are thought to have become autoreactive as a result of molecular

mimicry (Amedei, Bergman et al 2003)

Gastric cancer is the second most common cause of cancer deaths in the world There are two histological types of gastric cancer: the well-differentiated or intestinal-type,

or the poorly differentiated or diffuse-type, both of which have different carcinogenic pathways and risk factors HP is strongly associated with the intestinal-type of gastric cancer and much less with the diffuse-type Since the focus of this thesis is on HP-associated diseases, focus here will be put on the intestinal type of gastric cancer, and the term ‘gastric cancer’ in this text refers to the intestinal-type, unless otherwise stated Large-scaled studies have now well established that gastric cancer develops in individuals infected with HP, and seldom in uninfected individuals (You, Zhang et al 2000; Uemura, Okamoto et al 2001; Hsu, Lai et al 2007)

The development of gastric cancer follows a multi-factorial and multi-step process, which involves step-wise histological changes in the gastric mucosa known as the

‘precancerous cascade’ (Figure 1) (Correa 1992) Moving down the precancerous cascade, as lesions get more advanced, the risk of developing gastric cancer increases (de Vries, van Grieken et al 2008) HP infection typically results in a chronic on active, superficial gastritis Similar to the occurrence of gastric cancer, the

development of precancerous lesions in the gastric mucosa does not occur in all individuals infected with HP, but rather, in just a small percentage Histological changes are well documented and described for the stages of precancerous lesions (Correa and Houghton 2007) The earliest precancerous change is atrophy of the tissue (chronic atrophic gastritis; CAG) Typically in atrophic gastritis, normal

glandular structures are replaced with fibrous tissue, and infiltrating lymphocytes are apparent Intestinal metaplasia (IM) follows atrophy, but may also develop in the absence of atrophy Mucus-secreting goblet cells are seen There are three histological

subtypes of IM, described by Jass and Filipe (Jass and Filipe 1980) Each type of IM

is characterized by the presence of different types of intestinal cells and by the

different types of mucins secreted by these cells Across the three types of IM – Type

I or type II (small intestinal type) or type III (large intestinal type) – risk of

developing gastric cancer is different, the significantly highest risk being in type III, followed by type II then type I (Wu and Shun et al., 1998) Several molecular and gene expression changes make the state of IM unique from a normal gastric mucosa and also from other precancerous lesions, hence making this precancerous lesion the most widely studied and reviewed The changes include the loss of heterozygosity, microsatellite instability, telomerase activation, the increase in gene expression of CDX1, CDX2, TFF1,2 and 3, villin, PDX1 and

OCT-1 (Busuttil and Boussioutas 2009) Dysplasia is the final step in the progression, closest to carcinoma – the epithelium is enlarged and hyperchromatic, and the lumen

Figure 1 General hypothesis of gastric carcinogenesis Correa, 1992

appear irregular Throughout the stages of precancerous changes, lymphocytic

infiltrate remains a constant feature, demonstrating that chronic inflammation is a hallmark of the precancerous cascade, and that immunopathology may be an

important factor driving carcinogenesis

Progression through the stages of the precancerous cascade increases the risks of developing gastric cancer – the more advanced the lesion, the higher the risk of

progression to gastric cancer In a study conducted in a high risk area in Linqu County

in the Shandong province in Northeast China, the odds ratio for gastric cancer

(comparing to individuals with superficial gastritis or chronic atrophic gastritis) were the highest in individuals with severe dysplasia, second highest in those with deep IM

or mild dysplasia, and third highest in those with superficial IM (You, Li et al 1999) Another older study in the same area reported 98.1% prevalence of chronic atrophic gastritis, 52.8% IM and 20.4% dysplasia in a population aged 35-64 years old,

showing that in this area where gastric cancer risk is almost the highest in the world, the prevalence of precancerous lesions are correspondingly high The same group also compared the prevalence of precancerous lesions in Linqu county with that of another nearby county (Cangshan) in the same province, not known to be a high risk area They reported incidences of 30% and 15% for IM and dysplasia respectively in Linqu county and incidences of 7.9% and 5.6% in Cangshan county (You, Zhang et al 1998) In 2 Japanese studies, individuals with fundal atrophic gastritis or severe atrophic gastritis have a higher risk of developing gastric cancer than those with no atrophic gastritis (Kato, Tominaga et al 1992; Tatsuta, Iishi et al 1993), again

associating gastric cancer risk with the presence of precancerous lesions All these studies indicate that having precancerous lesions in the stomach is a precursor for

gastric cancer, and it predisposes the individual to a higher chance of developing cancer The reason for this is not exactly clear, but there are many clinical studies which give insight into possible explanations for the higher risk associated with precancerous lesions

1.3 Helicobacter pylori and Gastric Cancer Risk

1.3.1 Epidemiological Studies

Although HP is the strongest known risk factor for malignancies in the stomach, not everyone infected with HP is at risk of developing gastric cancer In fact, only a minor percentage of infected individuals develop malignancies in the gastric mucosa

succeeding HP infection The risk is multi-factorial – ethnic group, genetic

polymorphisms in immune or non-immune related genes, existing precancerous changes in the stomach and environmental factors such as diet, smoking and alcohol consumption are some of the factors influencing risk, in addition to HP infection itself The risk factors involved in gastric cancer development will be reviewed in this chapter Individuals of East Asian origins – Chinese, Japanese and Korean have been shown to harbor a significantly higher risk of developing gastric cancer, as compared

to individuals of the rest of Asia, and in Western countries despite the similar or even higher infection rates, indicating that genetic makeup is an important factor

contributing to cancer risk (Leung, Wu et al 2008; Goh, Chan et al 2011) The focus

of this thesis lies in the chronic immune responses leading up to gastric cancer

development in a Singaporean Chinese population, hence, the review of gastric cancer risk will focus mostly on studies relevant to this ethnic group

Since HP is one of the strongest risk factors for gastric cancer, many large-scale studies have been conducted to study the effects of HP eradication on reducing

precancerous lesion progression and cancer development However, the literature presents conflicting results Some groups reported beneficial effects of HP eradication

on the prevention of precancerous lesion progression (Sung, Lin et al 2000; Zhou, Sung et al 2003; Leung, Lin et al 2004), while others reported little or no efficacy in

HP eradication on preventing progression of precancerous lesions or development of gastric cancer, suggesting that HP is not the only factor driving carcinogenesis

(Correa, Fontham et al 2000) A randomized, controlled trial conducted in high risk area in Fujian province in China reported no difference in the gastric cancer incidence between the placebo controlled group and the HP-eradicated group during 7.5 years of follow-up However, they found precancerous lesions to be a confounding factor – for

a subgroup of participants who did not have any precancerous lesions, HP eradication treatment significantly decreased the development of gastric cancer during the follow-

up period (Wong, Lam et al 2004) In another study conducted in Hong Kong, risk factors for the progression of IM were being assessed (Leung, Lin et al 2004) All subjects presented with HP infection at baseline, and were randomly given either HP antibiotic triple therapy, or placebo In this study, in contrast to the previous one, IM progression is significantly reduced upon HP eradication treatment The incidence and progression of IM was higher in male subjects, and reported to be associated with persisting HP infection, age (>45years), alcohol consumption and drinking water from

a well HP eradication has also been assessed when given prophylactically, post-early gastric cancer resection Such a protocol was successful in the prevention of

secondary cancers arising in the gastric mucosa (Fukase, Kato et al 2008)

A review of the literature in human clinical trials presented above shows that HP eradication may not completely reduce the risk of precancerous lesions advancing and gastric cancer developing Since precancerous lesions and chronic inflammation are almost always co-existent, the persisting inflammation may be one of the factors that significantly contribute to risk of cancer progression A study by Mera et al in 2005 provided evidence that the gastric mucosa was chronically inflamed despite clearance

of HP infection, and that this may be a factor driving progression of precancerous lesions (Mera, Fontham et al 2005) The group reported that chronic inflammation, measured by the presence of stromal mononuclear cells in the gastric mucosa was observed 12 years post HP eradication; even when active inflammation was

diminished, chronic inflammation persisted in the absence of HP in the gastric

mucosa Their findings correlate well with the fact that leukocyte infiltrate is typically seen in precancerous lesions Chronic presence of infiltrating lymphocytes may be one of the reasons for precancerous lesion progression despite HP eradication

1.3.2 The COX-2 Pathway

The hypothesis that chronic inflammation drives cancer progression also draws a degree of substantiation from epidemiological studies on the association between use

of NSAIDs and cancer risk Several studies have demonstrated, in a number of types

of cancers, that the use of non-steroidal anti-inflammatory drugs (NSAIDs) is

associated with a lower cancer risk, possibly by limiting inflammation (inhibition of cyclo-oxygenase enzymes), and limiting cancer growth via a number of mechanisms (Muscat, Stellman et al 1994; Husain, Szabo et al 2002; Huls, Koornstra et al 2003)

A meta-analyses of studies performed on gastric cancer patients and NSAIDs revealed that the use of NSAIDs was associated with a decrease in gastric cancer risk (Wang, Huang et al 2003) A recent study further supports a possible role of inflammation in gastric cancer progression (Wong, Zhang et al 2011) It assessed the effects of HP eradication together with a 24-month cyclo-oxygenase-2 (COX-2) inhibitor

(celecoxib) treatment on regression of precancerous lesions in a high-risk region in China COX-2 is frequently reported to be up-regulated in gastric cancers and

precancerous lesions It is an enzyme that catalyzes the conversion of arachidonic acid

to prostaglandins, which then contribute to more inflammation by inducing vascular permeability and allowing more inflammatory cellular infiltrate Their results show that celecoxib treatment alone had resulted in significantly more regression of

precancerous lesions in individuals This study further supports the hypothesis that inflammation is associated with precancerous lesion progression, and that celecoxib treatment may result in regression of precancerous lesions, possibly by limiting

inflammation

1.3.3 Genetic Polymorphisms in Immune Genes

Further illustrating the role of inflammation in fuelling carcinogenesis is the fact that genetic polymorphisms in several inflammation-related genes are associated with gastric cancer The commonly reported cytokine genes include interleukin (IL) genes

IL1B, IL1RN, IL8, IL10, IL17A, (Wu, Zeng et al 2010) IL23A (Langowski, Zhang et

al 2006), transforming growth factor-β (TGFB) (Li, Cao et al 2008) and its receptors (TGFBR1 and TGFBR2) (Guo, Dong et al 2011) and tumor necrosis factor-α

(TNFA) Polymorphisms associated with a phenotype that favors more inflammation

are generally associated with gastric cancer risk Apart from the cytokine genes, polymorphisms in the pattern recognition receptor TLR4 have also been associated with gastric cancer risk (Hold, Rabkin et al 2007) and severe gastric atrophy

(Hishida, Matsuo et al 2009) Studies on these polymorphisms often give conflicting results when conducted on different ethnic backgrounds, again highlighting the

importance in the link between ethnic origins and gastric cancer risk In 2000, Omar genotyped several gastric cancer cases and controls in a Scottish and Polish

El-population and reported that single nucleotide polymorphisms in the IL-1 gene cluster (IL1B -31, IL1B -511 and IL1RN) are associated with gastric cancer risk In particular, the T-T haplotypes of IL1B -31 and IL1B -511 and IL1RN*2 have the highest odds

ratio for gastric cancer (El-Omar, Carrington et al 2000) They also demonstrated by

an electrophoretic mobility shift assay that there was a five-fold increase in

transcription factor binding to the IL1B-31T oligonucleotide A later study conducted

on a Chinese population validated the association between IL1B-511T and gastric

cancer risk (Li, Xia et al 2007) Complementing the data on genetic associations with cancer risk, a Japanese group reported a correlation between the T-T haplotype of

IL1B-511 and IL1RN*2 allele with increased protein levels of IL-1β (Hwang, Kodama

et al 2002) Sugimoto et al have conducted several studies in Japan and reported

polymorphisms in TNFA-857T, TNFA-863A, TNFA-1031C and 592C and

IL10-819C but not IL1B-511/-31 being significantly associated with gastric cancer

development (Sugimoto, Furuta et al 2007; Sugimoto, Furuta et al 2007)

The studies above have provided evidence that chronic inflammation and cytokine responses in HP-associated gastric cancer are an important factor in cancer risk The failure of HP eradication to curb cancer risk, the presence of chronic inflammation

years after clearance of infection, the association between polymorphisms in related genes with gastric cancer all indicate that the chronic immune responses in the gastric mucosa may play a significant role in cancer development Since the

immune-inflammation is almost always initially triggered by HP infection, it is important to next understand the immune responses during a HP infection

1.4 Adaptive Immune Responses to Helicobacter pylori

1.4.1 Sensing HP bacterium

HP is generally a non-invasive pathogen and is usually found in the extracellular mucus layer on the gastric mucosa However, some studies have found it invading epithelial cells (Amieva, Salama et al 2002), and by using transmission electron microscopy, HP was observed within the lamina propria where it was in direct contact with immune cells (Amieva, Salama et al 2002; Jhala, Siegal et al 2003; Necchi, Candusso et al 2007) In this context, the bacterium can then initiate immune

responses by binding to pattern recognition receptors on immune cells such as

monocytes, macrophages and dendritic cells The recognition of HP by the immune system via TLRs 2,4 and 5, and by intracellular sensor NOD1 was reviewed in

chapter 1.1 A vast body of literature can be found relating to the activation of

immune responses by HP in antigen-presenting cells Different antigenic components

of the bacteria have also been reported to activate immune responses via different pathways For example, LPS is recognized by TLR4, peptidoglycan is recognized by TLR2 and intracellular NOD1, and flagellin is recognized by TLR5 Other HP

antigens that are immunogenic include CagA, VacA, urease B, and HP-neutrophil

activating protein (NAP) to name a few Since the topic of this thesis is the associated chronic immune responses (which is comprised mainly of lymphocytic responses), the review of HP immune responses will focus on the adaptive immune responses, in particular, the T cell responses and cytokines associated with chronic gastritis and gastric cancer In understanding chronic inflammation and cancer progression, there are always two antagonizing concepts – 1) that inflammation is protective, and required for bacterial clearance and in scenarios when a cancer is present, the presence of inflammation is important to counter ‘regulatory’ responses (eg Tregs and myeloid-derived suppressor cells) so that immune responses can be mounted against the cancer cells (theory of immunosurveillance); 2) that

HP-inflammation causes immunopathology – direct damage to surrounding tissue

microenvironment, fuelling cancer progression Hence the immune system has to perform a balancing act to protect itself from disease Studies on inflammation and T cell responses associated with HP infection supporting either one of the two concepts above is reviewed below

1.4.2 Protective T cell responses

Literature on T cell responses to HP covers a few subsets of T cells, the T-helper 1 (Th1), Th17 as well as T regulatory cells (Tregs), implying that the adaptive T cell response towards HP is complex, and may change with chronicity of the infection Earliest studies on gastric T cell clones isolated from the lamina propria of HP-associated peptic ulcer disease patients showed that most of the T cells were skewed

to a Th1-type, producing large amounts of interferon-γ (IFNγ) and little IL-4 About 50% of the clones were specific for CagA antigen, assessed by the proliferative

response of the clone to the antigen (D'Elios, Manghetti et al 1997) In contrast to what is observed in ulcer patients, in patients with non-ulcer gastritis, there is a less striking skew towards a Th1 response Their immune response also encompasses a small but detectable IL-4 response, indicative of a Th2 response Furthermore, their responses are detected against a broad range of HP antigens which include urease, VacA and heat shock proteins (HSPs), along with CagA (Di Tommaso, Xiang et al 1995; D'Elios, Manghetti et al 1997; Itoh, Wakatsuki et al 1999) Priming of Th1 cells requires IL-12p40 production by antigen-presenting cells, and not surprisingly, many groups have found the production of IL-12p40 in response to HP infection (Haeberle, Kubin et al 1997; Guiney, Hasegawa et al 2003; Pellicano, Sebkova et al 2007)

1.4.3 Mouse models of HP infection

C57BL/6J mice are frequently used as an animal model for further in-depth studies on the adaptive immune responses in chronic HP infection Most of these studies utilize

either H pylori or H felis for infection and this results in inflammation and gastric

pathology, similar to that seen in humans (Marchetti, Arico et al 1995; Lee, O'Rourke

et al 1997) Studies in these mice have shown that IFNγ plays an inflammatory and protective role in defense against HP infection (Sawai, Kita et al 1999; Akhiani, Pappo et al 2002) In IFNγ-/-

mice, gastric inflammation was markedly reduced, indicating that IFNγ was the main source of inflammation in that model In contrast,

in IL-4

mice, gastritis was more severe, and production of IFNγ was elevated,

suggesting that a balance between Th1 and Th2 cytokines was essential for limiting inflammation (Smythies, Waites et al 2000) This phenomenon is also mirrored in

mice deficient in IL-10 In the absence of the regulatory IL-10, there is increased hyperplasia of the epithelium and Th1-dependent inflammation (Berg, Lynch et al 1998; Chen, Shu et al 2001; Matsumoto, Blanchard et al 2005) Adoptive transfer of CD4+

T cells in murine experiments demonstrated that the Th1 cells were responsible for inducing inflammation, as well as reducing bacterial load (Eaton, Mefford et al 2001; Eaton and Mefford 2001; Sayi, Kohler et al 2009)

Despite the protective roles IFNγ and Th1 responses may have on HP infection, the inflammation may also contribute to ‘bystander damage’, or immunopathology One study demonstrated that the adoptive transfer of CD4+

T cells resulted in an increase

in gastric epithelial cell apoptosis and proliferation in recipient mice, suggesting that the Th1-dependent gastritis may be responsible for epithelial damage, and subsequent metaplastic changes (Peterson, Hoepf et al 2003) Sayi et al described IFNγ to play

opposing dual roles in H felis infected gastric mucosa – a strong IFNγ response aids

in enhancing bacterial clearance in the gastric mucosa, but at the same time, increases the risk of developing precancerous changes like atrophy and epithelial hyperplasia (Sayi, Kohler et al 2009) Gene expression studies performed by this group revealed that IFNγ was most strongly correlated with hyperplasia but not metaplasia A recent study identified the T cells responsible for the promotion of gastric neoplasia to be specific for epitoptes from the CagA protein (Arnold, Hitzler et al 2011) In this paper, the adoptive transfer of CagA-specific T cells to T cell deficient recipients resulted in the induction of precancerous immunopathology in the recipients,

demonstrating a direct link between T cell-mediated inflammation and precancerous lesion development

In contrast to the inflammatory and carcinogenic properties described above, IFNγ is also known to play anti-tumor roles and has been reported to inhibit IL-1β-dependent gastric carcinogenesis in mice over-expressing IFNγ in the gastric mucosa (Tu, Quante et al 2011) The conflicting data strongly suggests that there may be

additional T cell subsets and factors that may also influence the balance between protection and immunopathology in the gastric lamina propria Therefore, it is

essential to consider other T cell responses that may occur during HP infection

recruitment, and the activation of NF-κB by HP in gastric epithelial cells (Robinson, Kenefeck et al 2008) CD4+

CD25hi

T cells isolated from the gastric mucosa were found to express high levels of FOXP3 and CTLA-4 (Lundgren, Stromberg et al 2005) Upon eradication of the infection, the amount of FOXP3 mRNA decreases in

the gastric mucosa, suggesting that the presence of the bacteria was required for the localisation of the Tregs in the lamina propria (Chen, Jin et al 2010)

The role of Tregs in limiting inflammation is reiterated in the finding that an

inadequate Treg response is usually seen in peptic ulcer disease In peptic ulcer

disease, the T cell response is skewed towards a strong Th1 response, which

associates with abundant IFNγ production The lack of a sufficient Treg response possibly contributes to the excessive inflammation in this disease condition

(Robinson, Kenefeck et al 2008) Hence, by inducing a regulatory response, HP is able to limit inflammation and allow itself to survive in the gastric mucosa for long periods of time

A recent study in the murine model of HP infection reports that HP controls the skewing of T cell responses via acting on dendritic cells (DCs) (Kao, Zhang et al 2010) In this study, the adoptive transfer of HP-pulsed DCs decreases the IL-

17A/FOXP3 mRNA ratio, suggesting a skew towards Treg, away from Th17, another pro-inflammatory T cell response induced by HP (that will be reviewed in depth in the next chapter) The role of epithelial cells in shaping immune responses is also

recognized in the model of HP infection HP-infected gastric epithelial cells (GECs) secrete TGF-β1 and TGF-β2, and express programmed death ligand-1 (PDL-1) and this results in a decrease in proliferation of activated CD4+

T cells, and enhanced development of nạve CD4+ T cells into CD4+CD25+FOXP3+ Tregs (Beswick,

Pinchuk et al 2007; Beswick, Pinchuk et al 2011)