T cell responses in helicobacter pylori associated gastroduodenal diseases 3

Given the elevated levels of IL-1β and IL-17A observed in group A and P individuals, and since IL-1β is a cytokine responsible for maintenance and proliferation of Th17 cells, the expres

3.1.5 IL-22 in the gastric mucosa

Th17 cells also secrete IL-22, a cytokine belonging to the IL-10 family, that enhances epithelial integrity and regeneration in the gut, but promotes inflammation under certain experimental settings (reviewed in Introduction, Chapter 1.6) Gastric mucosal IL-22 was quantified using real time PCR, immunofluorescence microscopy, and ELISA IL-22 mRNA levels were significantly higher in group A, as compared to

group P (figure 3.6A) However, for all other protein data (ex vivo cytokine

concentration and IF staining cell counts), such a suppression of IL-22 protein levels was not seen, in fact, there was no significant difference across the three groups (figure 3.6B and D) Additionally, there was no difference in the total amounts of IL-

22 in group P individuals, whether subjects had PC lesions in the gastric mucosa or not (figure 3.6C) Nevertheless, when CD4+IL-22+ cells counts were analysed from IF-stained frozen sections, an association with HP infection was seen (figure 3.6E), which is in line with the observation that there is a heightened Th17 response during

an active HP infection There was no significant difference between the CD4+IL-22+cell counts in groups P and N, suggesting that there might be a fairly low level of Th17-derived IL-22 in group P individuals However, within group P, a higher

median number of CD4+IL-22+ cells associated with PC lesions (figure 3.6G) The median and IQR values of the above-mentioned IL-22 protein data is summarized in table 3.5 Taken together, these data demonstrate that CD4+ T cells are not the only source of IL-22 in the gastric mucosa Although higher amounts of CD4+ T cell-dervied IL-22 associated with an ongoing HP infection, total IL-22 levels and cell counts showed that there were other sources contributing to what is observed in the

IL-22 levels were fairly similar across all three groups, and since it is reported to

play homeostatic and protective roles in the gut (Zenewicz, Yancopoulos et al 2008)

IL-22 in the absence of IL-17A has tissue-protective effects; however IL-22 and IL-17

synergistically provoke inflammation when co-expressed (Sonnenberg, Nair et al.,

2010) Intra-individual ratios of the pro-inflammatory IL-17A or IFNγ to IL-22 were

calculated as a measure of cytokine co-expression in all individuals IL-17A/IL-22

ratios were significantly elevated in samples from groups A and P compared to group

N (figure 3.6H), and IFNγ/IL-22 ratios were elevated in group P compared to N

(figure 3.6I) This suggests that the gastric cytokine milieu in group P subjects

continues to be biased against the baseline state of mucosal maintenance found in

HP-nạve patients, despite the absence of ongoing HP infection

169, 69-261

PC-

212, 105-301 Total IL-22 +

IF cell counts 31, 17-43

16, 11-22

17, 9.0-28 PC+

16, 11-25

PC-

18, 14-21 CD4 + IL-22 +

IF cell counts 11, 5.0-15

4.0, 3.0-5.1

3.0, 3.0-6.5 PC+

4.2, 3.2-6.6

PC- 3.3, 1.8-4.0 Table 3.6 Summary of median and interquartile range (IQR) values for IL-22 protein data reported in figure 3.6

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

subjects

Having established that the gastric microenvironment in group P patients had

increased Th17 infiltrate and a pro-inflammatory profile, it was of interest whether

Figure 3.6 IL-22 levels in human gastric biopsies (A) Semi-quantitative SYBR-Green real time PCR for IL-22 mRNA Gene expression was normalized to the housekeeping gene β-actin and arbitrary

values were further normalized to group N (B) ELISA for ex vivo IL-22 concentrations (C)

Subclassification of group P individuals into PC+ or PC- groups for analysis of ex vivo concentrations

there was any correlation across the pro-inflammatory cytokine levels within

individuals, that may increase the significance of the data The associations between several HP-associated cytokines and IL-17A were analysed The high levels of IL-1β

in group P patients were striking (figure 3.4G), as IL-1β is a well recognized risk factor for gastric cancer and is also important in the maintenance of human Th17 cells In line with this, there was a significant correlation between IL-1β and IL-17A

in group P individuals (figure 3.7A) There was no statistically significant correlation

in group A, although this might be due to the small sample size compared to group P the concurrently high levels of both IL-1β and IL-17A indicates that there may be an association between these two cytokines, and that during chronic inflammation, IL-1β and IL-17A are present at elevated levels at the same time There was also a

correlation between IL-17A and IFNγ, IL-8 and IL-10 levels in group P subjects (figures 3.7B, C and E respectively) However, there was no correlation between IL-17A and IL-22 nor IL-1β and IL-22 levels within individuals, in line with the finding that total gastric IL-22 levels were not affected by history of HP infection (figure 3.7D)

These correlations provide the evidence that the persistent inflammation observed in individuals with a previous HP infection occurs as a collective presence of several pro-inflammatory cytokines, which may work together to worsen or prolong

inflammation

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

individuals

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

(A-E) Intra-individual cytokine concentrations correlated with IL-17A Spearmans’ test for correlation was used (F) Intra-individual cytokine correlation between IL-1β and IL-22

Given the elevated levels of IL-1β and IL-17A observed in group A and P

individuals, and since IL-1β is a cytokine responsible for maintenance and

proliferation of Th17 cells, the expression of the receptor for this cytokine, IL-1 receptor-1 (IL-1R1) was investigated alongside IL-17A expression by

immunostaining in frozen biopsy sections in three subject groups Frozen biopsy sections were co-stained with antibodies for the two antigens IL-1R1 and IL-17A and

some of this was co-expressed with IL-17A Not only was the intensity of the staining much brighter than in other samples, the frequency of positive stains were higher

out of 4 of these patients that had a bright staining and high frequency of IL-1R1 staining had severe IM in their gastric mucosa according to histopathology reports ? out of 2 of group A individuals also had severe IM in their gastric mucosa Figure 3.8A shows a section showing bright and frequent IL-1R1 staining from a group P individual having severe IM in the gastric mucosa, and this is contrasted with figure 3.8B, a group P individual with IM (not severe), having dimmer and much less IL-1R1 staining These staining patterns were not observed for any of the group N

subjects Figure 3.8C and D shows all the data points from subjects from the three groups Although there was no statistical significance between the three groups, it was clear that the bright and frequent IL-1R1 staining were only seen for group P subjects, less in group A and not in group N at all It is however, of note that histological diagnoses of IM in patient biopsies were made on separate biopsies from those used

in frozen section staining, and sampling variation within each patient may partially confound these results These results suggest that the expression of IL-1R1 in the gastric mucosa of some individuals may increase as PC lesions advance On top of the already elevated presence of IL-1β in the gastric mucosa of group P subjects, the increase in sensitivity to IL-1R1 may fuel precancerous changes IL-1R1

polymorphisms have been reported to be associated with increased HP infection incidences, but not with gastric cancer (Hartland, Newton et al 2004)

A

B

Figure 3.8 High frequencies of IL-1R1 bright cells in group P individuals with severe IM (A) 40X magnification image showing staining of IL-17A and IL-1R1 on frozen sections from a group P individual with severe IM (B) Same staining performed on a group P individual with IM (not severe) (C) Frequencies of IL-1R1 + cells determined by the average number of positively-stained cells counted per HPF of IF-labelled frozen sections (D) Frequencies of IL-1R1 + IL-17A + cells in the same sections

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

GECs

In Chapter 3.1.4, group P subjects were described to have an elevated level of

numerous pro-inflammatory cytokines, especially IL-17A and IL-1β These two cytokines have been reported to induce hBD-2 expression from epithelial cells (Jang,

Lim et al 2004; Kao, Chen et al 2004) Furthermore, ex vivo hBD-2 levels in group P

subjects were also slightly elevated Since IL-22 is also known to induce the

expression of anti-microbial peptides from epithelial cells in host innate immune responses (Wolk, Kunz et al 2004), the effects of IL-22 on gastric epithelial cells (GECs) were investigated alongside the pro-inflammatory cytokines GECs were isolated from fresh biopsy samples from several group P and N subjects and these were stimulated with IL-1β IL-1β significantly up-regulated hBD-2 expression in the gastric mucosa by a median of 50-fold (IQR 24-79) across all samples (figure 3.9A) There were no significant differences between the responses of GECs isolated from group P and group N samples towards IL-1β stimulation Due to the lack of clinical samples and scarce numbers of GECs isolated from each individual sample each time, stimulation of GECs with IL-22 and IL-17A in combination with IL-1β were not repeated as many times as for IL-1β alone For two individuals, IL-22 alone or in combination with IL-1β was used to stimulate GECs IL-22 alone did not result in any significant increase in hBD-2 gene expression However, IL-22 in combination with IL-1β resulted in a synergistic increase in hBD-2 gene expression level as compared

to either one of the cytokines alone (figure 3.9B) IL-17A resulted in a median of fold increase in hBD-2 gene expression, slightly higher than the IL-1β-induced

76-increase (55-fold) Together, IL-17A and IL-1β co-stimulation resulted in an additive effect, with a 126-fold increase in hBD-2 gene expression (figure 3.9C) Earlier data reported a presence of IL-22 across the three patient groups The effects of IL-22 together with pro-inflammatory IL-17A and IL-1β on GECs include the increase in hBD-2 expression In a single group P patient where there were enough GECs

resulted in the highest increase in hBD-2 gene expression (151-fold) (figure 3.9D) This observation is partly in line with previous reports that IL-22 can synergize IL-17A in the induction of human β-defensin production (Liang, Tan et al 2006), and shows for the first time, the cooperation between IL-22 and IL-1β in hBD- 2

induction This increase in hBD-2 levels (also seen in ex vivo biopsy analyses) may

enhance chemotaxis and retention of CCR6+

Th17 cells in the gastric mucosa of group P individuals

To establish a biological relevance of this finding, the ex vivo concentrations of

hBD-2 measured in biopsies were correlated with the cytokines used in stimulation, to determine if there was any biological association between the cytokines and the induced gene For IL-17A and IL-1β, there was no statistically significant correlation

in all three groups of individuals, but there was a trend towards an increased

correlation in group P subjects as compared to group N (figure 3.9E and F) There were probably insufficient data points for group A to see a statistically significant correlation hBD-2 and IL-22 levels were significantly correlated in group P

individuals (figure 3.9G), suggesting that the presence of IL-22 may be important for hBD-2 induction in this group of subjects, possibly by acting in synergy with IL-17A

and/or IL-1β, since it did not alone result in the increase in hBD-2 expression in the in

vitro experiments

1β was used at 20ng/ml (B-D) Same experimental setup, with the use of 22 (200ng/ml) and

17A (20ng/ml) where indicated (E-G) Correlation of ex vivo hBD-2 concentrations with 1β,

IL-17A and IL-22 respectively

3.1.9 IL-1β-stimulated GEC supernatant is chemoattractive to autologous

CCR6 +

Th17 cellls

In order to determine if the induced hBD-2 gene expression in GECs translated into increased chemotaxis of infiltrating Th17 cells, a chemotaxis assay was conducted Supernatant from the GEC stimulation assay was stored at -80°C while autologous LPMCs were being expanded in culture The expanded LPMCs were used as

responder cells and GEC supernatant was used as chemoattractant media Figure 3.10A shows a flow plot for a sample of expanded CD4+ LPMCs which was used in the chemotaxis assay A large proportion of the cells were CCR6+

and 2.37% were 17A+

IL- After a 24 hours chemotaxis assay, the responder cells that had migrated

through the barrier and into the lower chamber of the wells were harvested and subject to the same flow cytometry analyses (figure 3.10B) There were no significant differences in the proportions of CCR6 or IL-17A positive events across the

conditions for the assay as can be seen from figure 3.10B

Fluorescence was then used to measure the number of cells that had migrated through the barrier in the different conditions Due to the variation in the numbers of GECs and expanded LPMCs that could be obtained from each clinical sample, not all the same conditions could be repeated for the different samples Figure 3.10C-F shows the fluorescence intensity of migrated cells for the different conditions for different samples The LPMC sample that was analysed by flow cytometry has fluorescent cell measurement data in figure 3.10C In summary, for 2 (one group P and one group N)

out of 4 samples that the chemotaxis assay was performed, there was an increase in the number of cells that had migrated towards supernatant from GECs stimulated with IL-1β Controls were included, although limited by the amount of cells available for use The lack of a detectable difference in two of the samples may be due to the very low yield of GECs that is usually obtained from biopsy samples The small number of GECs may yield a smaller magnitude of response towards the cytokine stimulation and this may hence be insufficient to induce significant chemotaxis of the responder Th17 LPMCs

This demonstrates that CCR6-expressing IL-17A+

LPMCs isolated and expanded from gastric biopsies could migrate towards supernatant of GECs stimulated with IL-1β, which could contain hBD-2 amongst other possible chemotactic factors

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

infection

Whole lysate of HP strain 26695 was used to stimulate IL-17A and IFNγ responses in PBMCs from subjects from three groups As expected, HP-specific IFNγ responses were significantly elevated in PBMCs from group A individuals A median increase

of 54-fold (IQR 15-478) was observed in group A individuals for IFNγ production, when lysate-stimulated PBMCs were compared with the unstimulated PBMCs from the same individual A 6.4-fold (IQR 1.8-21) increase was seen in group P subjects, and 3-fold (IQR 1.6-20) in group N (figure 3.11A) In contrast, stimulation with HP

Figure 3.10 Expanded LPMCs and chemotaxis assay towards IL-1β-stimulated GEC supernatant (A)

Zebra-plot (showing outliers) of PMA and ionomycin-stimulated expanded CD4 + LPMCs from a group

N individual Plot shows the CCR6 and IL-17A expression of the CD3 + CD8 - sub-gate of ‘LIVE’ gated

cells (B) Zebra-plot (showing outliers) of PMA and ionomycin-stimulated CD4 + LPMCs that had

migrated across the membrane after chemotaxis assay (from the same individual) (C-F) Fluorescent

intensity of lysed migrated cells stained with fluorescent DNA dye Responder cells were expanded

CD4 + LPMCs and chemoattractant media was conditioned media from autologous GECs The

appropriate controls are labelled in the axes

in group N individuals (figure 3.11B) PBMCs from group A individuals produced a

6.2-fold (IQR 3.4-26) increase in IL-17A production when stimulated with HP lysate

but the difference compared to group N did not reach statistical significance Taken

together, these results show that HP lysate-specific IL-17A responses exist in the blood of individuals even after clearance of the HP infection

Antigen-specific central memory T cells are expected to be present in the peripheral blood of individuals who have been infected by HP, and a response to antigen is in line with theoretical expectations However, the responses observed in this

experimental setup (72-hour time point capturing an early response) show an even more rigorous cytokine response from group P individuals as compared to group A, indicating that there may be a large number of HP-specific central memory T cells found in the peripheral blood of group P individuals

To demonstrate that this cytokine response was mediated CD4+

T cells and was dependent on antigen presentation on MHC class II molecules, purified CD4+

T cells were used as responders, co-cultured with autologous APCs The IL-17A responses from a representative group P and group N subject is depicted in figure 3.11C, and the detailed experimental conditions are described in the figure legend HP lysate-pulsed APCs induced a 7-fold larger IL-17A response in CD4+

T cells isolated from a group

P individual compared to a group N individual In the presence of an MHC class II blocking antibody, the IL-17A response was decreased by 4-fold, showing that

interactions with surface MHC class II molecules was required for triggering the cytokine response This experiment was repeated for a number of individuals from the three subject groups and the results are summarized in table 3.7 In summary, IL-17A production was higher in PBMCs from groups A and P than from group N, and

blocking MHC class II resulted in a decrease in cytokine production, with the highest

CD4 T cells persist in the PBMCs of patients with a past HP infection, and the

elevated numbers of Th17 cells observed in group P subjects’ PBMCs could in part be due to the prolonged presence and even proliferation of these HP-specific T cells

To next demonstrate the presence of HP-specific CD4+

T cells in the gastric mucosal samples from patients with a past HP infection, the experiment was performed using expanded gastric lamina propria mononuclear cells (LPMCs) as responders HP-specific IL-17A responses were observed in 3 out of 3 individuals from group A Interestingly, we detected HP-specific IL-17A responses in 4 out of 5 individuals from group P All four responders had PC lesions Only weak responses were

detected from 4 individuals from group N The concentrations of IL-17A detected in the supernatants of the LPMC co-culture assays are reported in Table 3.8 Figure 3.11D depicts representative data from a group P individual, producing 11-fold more IL-17A when stimulated with autologous HP-pulsed APCs compared to LPMCs from

a group N individual The presence of the MHC class II blocking antibody resulted in

a 4-fold decrease in the IL-17A response from LPMCs from group P Figure 3.11E shows the other 3 group P responders not shown in figure 3.11D

These results demonstrate that there is a persistence of HP-specific, IL-17A producing CD4+

T cells, that were restricted by MHC class II amongst PBMCs as well as

LPMCs derived from group P patients The observed chronic cellular infiltrate is thus,

at least in part, made up of HP-triggered T cell responses during initial infection

Mean IL-17A production per 10 6

CD4 + T cells (pg/ml)

2594±686 n=3

2576±629 n=8

431±137 n=3 Mean fold-decrease in the presence

of MHC class II blocking antibody 2±0.3 5±3 1±0.3 Table 3.7 Summary of IL-17A responses of CD4 + PBMCs to HP-pulsed APCs

individual is depicted (D) HP-specific IL-17A response of gastric mucosal LPMCs 48 hours post culture Results from one representative group P individual who showed a response and one from group N is depicted (E-G) HP-specific LPMC responses from group A, P and N individuals

respectively

Mean IL-17A production per 10 6

LPMCs (pg/ml) 860 ± 546, n=3 169 ± 38, n=4 18 ± 8, n=4 Mean fold-decrease in the presence of

MHC class II blocking antibody 805 ± 581 73 ± 41 14 ± 7 Table 3.8 Summary of IL-17A responses of LPMCs to HP-pulsed APCs

This first part of the results first proved the hypothesis stated in Chapter 1.7 that

‘Th17 cells contribute to immunopathology associating with precancerous lesions, and perhaps indirectly, cancer risk’

Firstly, it was demonstrated in a proportion of this cohort of subjects (made up mostly

of Chinese individuals above the age of 50) that were followed up for 2-3 time points, that after treatment for HP, there was little or no regression of PC lesions In most subjects, lesions had either not changed or had progressed to a more advanced stage Group P individuals who had been treated for HP up to 10 years ago still maintained a high frequency and infiltrate of Th17 cells in their peripheral blood and gastric

mucosa The presence of this inflammatory infiltrate correlated with PC lesions in their gastric mucosa suggesting a link between this chronic inflammation and gastric cancer risk Aside from cellular infiltrate, the elevated presence of many pro-

inflammatory cytokines like IL-1α, IL-1β, IL-6, IL-8, TNF-α, IL-17A and IFNγ will contribute to the chronic inflammatory environment in these individuals γ-chain cytokines like IL-2 and IL-15, which are known to induce lymphocyte proliferation were concurrently higher in group P individuals as compared to group N, suggesting

that proliferation may contribute to the increase in cellular infiltrate seen over time Finally, chemotaxis and retention of lymphocytes was proved to be another possible explanation for the large cellular infiltrate

Lastly, by demonstrating the presence of HP-specific responses in this group of individuals, the hypothesis that ‘HP-specific Th17 responses, which were primed during initial infection, persist in the gastric mucosa of patients even when they no longer have an active infection’ was proven CD4+

LPMCs from individuals with a past HP infection produced IL-17A in response to HP lysate, and this was mediated at least in part by MHC class II molecules

3.2.1 Expression of cytokine receptors IL-22R1, IL-1R1 and IL-17R1 in the human

gastric epithelium and their regulation by inflammation

The effects of three Th17-associated cytokines on the gastric epithelium will be studied in this next part of the results These cytokines are IL-17A, IL-1β and IL-22, and these will be studied due to their significant presence in the gastric mucosa in chronic inflammation and their reported affects on epithelial cells Furthermore, the effects of these cytokines, especially IL-22, on the gastric epithelium are not well studied thus far Whether IL-22 plays a pro-inflammatory or protective role on the gastric mucosa, is not known By studying their effects on gastric epithelial cells, more insight can be shed onto how the presence of these cytokines may contribute to immunopathological damage during chronic inflammation

Two cell lines and fresh human GECs were used as models to study the cytokine effects Firstly, the expression of the cytokine receptors on the gastric epithelial cells were verified by PCR (figure 3.12A) Expression of all cytokine receptors were detected in both cell lines as well as in GECs Protein levels of IL-22R1 on AGS cells lines were also determined by western blot analyses

It has been reported that expression of IL-22R1 in HT-29 colonic epithelial cell line was regulated in the presence of inflammation (Brand, Beigel et al 2006) Hence the regulation of IL-22R1 and IL-1R1 by inflammation in cell lines and fresh GECs were determined AGS cells were stimulated with various inflammatory stimuli over different time points and the expression of IL-1R1 and IL-22R1 were determined by

Figure 3.12 Cytokine receptor expression and regulation by inflammation (A) mRNA expression of IL-22R1, IL-10R2, IL-1R1 and IL-17R1 in AGS and KatoIII cell lines and freshly isolated GECs determined by RT-PCR No template control was used as a negative control (B) Cell lysates of AGS and KatoIII were resolved by on a 10% polyacrylamide gel and immunoblotted for IL-22R1 A 62kDa

band was detected (C&D) Semi-quantitative SYBR-Green real time PCR for IL-1R1 and IL-22R1 in human GECs respectively Gene expression was normalized to the housekeeping gene GAPDH and arbitrary values were further normalized to the unstimulated GECs (E) Expression of IL-22R1 on AGS cells following stimulation with cytokines Flow cytometry plots of AGS cells following stimulation with inflammatory stimuli

semi-quantitative real time PCR (figure 3.13) IL-1R1 was not significantly affected

by the inflammatory stimuli being investigated The largest effects were seen at the 24 hour time point – a down-regulation in IL-22R1 expression Hence the effects of a few of the inflammatory stimuli were tested in freshly isolated GECs As in cell lines, IL-1R1 was not significantly affected by inflammation (3.12C) IL-1β and TNF-α stimulation for 24 hours resulted in a significant down-regulation of IL-22R1 in GECs (figure 3.12D) Due to the insufficient numbers of GECs, protein level correlation of IL-22R1 was not able to be determined Hence AGS cell line was used and the cell surface regulation of IL-22R1 was investigated by flow cytometric analyses after stimulation with inflammatory stimuli Similar to the gene expression data, all

inflammatory stimuli resulted in a down-regulation of surface IL-22R1, although the magnitude was small (figure 3.12E)

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

Abnormal increase in cellular proliferation above the homeostatic level may result in the increase in cellular DNA damage and hence the chances of developing

malignancies Since IL-22 is important for homeostasis in the gastrointestinal tract, its effects on proliferation, together with IL-1β and IL-17A were investigated Next, the

Figure 3.13 Regulation of IL-1R1 and IL-22R1 in AGS cells by inflammatory stimuli at various time points Semi-quantitative SYBR-Green real time PCR for IL-1R1 and IL-22R1 mRNA Gene

expression was normalized to the housekeeping gene GAPDH and arbitrary values were further normalized to the unstimulated cells

effects of cytokines on gastric epithelial cell proliferation were investigated in the cell lines Using an optimized range of concentrations of cytokines, AGS and KatoIII cell proliferation was assessed using tritiated thymidine (3H) incorporation Concentration

of cytokines used were determined by starting with the manufacturer’s recommended ED50 value, using concentrations that spanned above and below that value IL-17A induced an increase in AGS proliferation in all tested concentrations, although the highest concentration resulted in the lowest increase in proliferation (100ng/ml) This drop in proliferation rate in high cytokine dose was also observed when IL-22 was added to the cell culture, although for IL-22, at the highest cytokine dose there was still a higher proliferation rate as compared to the unstimulated cells For KatoIII, IL-

22 induced a dose dependent increase in proliferation rates for the concentrations of cytokines used (figure 3.14A,B and D) For IL-1β, only at 1ng/ml was there an increase in proliferation rate in AGS cells (figure 3.14C) For all other concentrations, there was a lower count per minute, indicating that there may have been apoptosis induced in those concentrations

Figure 3.14 Increase in epithelial cell proliferation in the presence of gastric cytokines (A-D) Tritiated thymidine ( 3 H) incorporation assay for AGS and KatoIII cells Concentrations of cytokines are shown

on a log10 scale (E) Cell lysates of AGS and Kato III, immuno-blot for p-Tyr705 STAT3 Cells were seeded and serum starved for 4 hours prior to a 10-minute stimulation with IL-22 (200ng/ml) in serum- free media, after which cells were lysed

Since IL-22 had an effect on the gastric epithelial cells, it was of interest to determine which cellular signaling pathway it utilized Gastric cancer is typically associated with an increase in STAT3 phosphorylation and IL-22 is known to activate

downstream STAT signaling STAT3 phosphorylation was detected in AGS and KatoIII cells following IL-22 stimulation, showing that IL-22 may increase cellular proliferation via STAT3 signaling

3.2.3 IL-22-induced genes in gastric epithelium

To measure if the effects of IL-22 in the gastric epithelium were pro-inflammatory or protective, the regulation of gene expression of several classes of proteins by IL-22 were studied These included the human β-defensins, the S100 proteins, matrix metalloproteinases (MMPs), serum amyloid-A proteins (SAAs), regeneration proteins (Reg1α), differentiation proteins and a few pro-inflammatory cytokines (Wolk, Kunz

et al 2004) Due to the limited time and resources (clinical samples) available,

extensive analyses of GEC responses to IL-22 could not be performed Hence, a few relevant proteins from each class mentioned were investigated

hBDs are typically induced in response to bacterial infections and IL-22 has been reported to also up-regulate these anti-microbial peptides in a few other tissue types (eg colon epithelial cells and keratinocytes) (Wolk, Kunz et al 2004; Zheng, Valdez

et al 2008) AGS and KatoIII cell lines were utilized as a first step in studying the regulation of gene expression by IL-22 at different doses These cells were seeded overnight in tissue culture plates, then serum starved for 4 hours prior to addition of IL-22 at various concentrations IL-22 at 200ng/ml resulted in the largest differences

in hBD-1, 2 or 3 in AGS and KatoIII cells (figure 3.15)

S100A7, A8 and A9 belong to the pleiotropic S100 family of calcium binding

proteins (Roth, Vogl et al 2003) and have been reported to act as pro-inflammatory mediators in a number of settings (Watson, Leygue et al 1998; Donato 1999) IL-22 did not significantly affect gene expression of either S100A7 or A8 (figure 3.16A, C

Figure 3.15 Human β-defensin expression regulation by IL-22 in AGS and KatoIII cells (A-C) quantitative SYBR-Green real time PCR for hBD-1, 2 and 3 mRNA in AGS cells Gene expression was normalized to the housekeeping gene GAPDH and arbitrary values were further normalized to the unstimulated cells Where significant, the Kruskal-Wallis p-value is shown, and Dunns’ multiple comparison post test results are depicted by the capped bars and asterisks above bar graphs (*p<0.05) (D-F) Same experiment repeated for KatoIII cells

Semi-Figure 3.16 S100A7-9 expression regulation by IL-22 in AGS and KatoIII cells (refer to figure 3.15 legend)

IL-22 did not have significant effects on MMP1, SAA-1 and 4, gelsolin and filaggrin (figures 3.17A, E, F, H-J and N) MMP-3 was slightly up-regulated in KatoIII cells but not AGS cells (figures 3.17B and K) There was a slight down-regulation in pro-inflammatory IL-8, but not TNFα gene expression in AGS cells (figures 3.17C and D) This was not seen in KatoIII cells (figures 3.17L and M) The largest increase in gene expression was observed for Reg1α, where IL-22 at 200ng/ml resulted in a 2.9-fold increase in mRNA expression in AGS cells and a 29-fold increase in KatoIII cells (figures 3.17G and O) Reg1α belongs to a family of regeneration proteins that are known to be up-regulated in response to damage, inflammation and infections (Dieckgraefe, Crimmins et al 2002; Sekikawa, Fukui et al 2005) Furthermore,

Figure 3.17 Other classes of proteins being regulated by IL-22 in AGS and KatoIII cells quantitative SYBR-Green real time PCR for hBD-1, 2 and 3 mRNA in AGS cells Gene expression was normalized to the housekeeping gene GAPDH and arbitrary values were further normalized to the unstimulated cells

Semi-Figure 3.18 Synergy between IL-22 and IL-17A in AGS cells (A-D) Semi-quantitative SYBR-Green real time PCR for hBD-1, 2 and 3 mRNA in AGS cells Gene expression was normalized to the housekeeping gene GAPDH and arbitrary values were further normalized to the unstimulated cells both gastrin and presence of HP (Steele, Dimaline et al 2007) The results here demonstrate that IL-22 may be another factor inducing the expression of Reg1α in the gastric environment, and possibly under normal steady-state conditions IL-22

induced Reg1α may be important in gut homeostasis

Effects of IL-22 in conjunction with IL-17A and IL-1β on gene expression was investigated In Chapter 3.1.8, the effects of these cytokines on hBD-2 gene

expression was investigated, and it was concluded that they played synergistic effects

in the induction of hBD-2 The

effects of synergy was investigated in the hBDs and S100 genes only In AGS, IL-22 and IL-17A synergized in the induction of hBD-1 expression (figure 3.18A) For the other genes,

Figure 3.19 Synergy between IL-22 and IL-1β in AGS and KatoIII cells (A-F) Semi-quantitative SYBR-Green real time PCR for hBD-1, 2 and 3 mRNA in AGS cells Gene expression was normalized

to the housekeeping gene GAPDH and arbitrary values were further normalized to the unstimulated cells

there was no significant synergistic effect between IL-22 and IL-17A being observed (figures 3.18B-D, data not shown)

IL-22 together with IL-1β had a synergistic effect on the induction of hBD-3 and S100A7 in AGS cells (figures 3.19A and B), and of hBD-2, 3, S100A7 and 9 in KatoIII cells (figures 3.19C-F) The effects of IL-22 and IL-1β were then further confirmed in primary GECs isolated from clinical samples The largest synergistic effects were observed for hBD-2 gene expression (figure 3.20B), and to a lesser extent, S100A7 (figure 3.20E) Figure 3.20C shows the 3 individual data points

depicted in figure 3.20B The other genes were induced singly by either 1β or

IL-22 but these effects did not synergise The tissue protective versus pro- inflammatory properties of IL-22 seems to be governed by co-expression of other cytokines For example, IL-22 in the absence of IL-17A has tissue-protective effects; however IL-22

Figure 3.20 Synergistic effects between IL-22 and IL-1β in primary GECs (A-G) Semi-quantitative SYBR-Green real time PCR for hBD-1-3 and S100A7-9 mRNA in primary GECs Gene expression was normalized to the housekeeping gene GAPDH and arbitrary values were further normalized to the unstimulated cells

and IL-17 cooperatively provoke inflammation when co-expressed in the airway

epithelium (Sonnenberg, Nair et al 2010) In the gastric mucosa, the elevated

presence of both IL-17A and IL-1β in infected subjects and in subjects with a history

environment, while on the other hand, this pro-inflammatory effect may be less apparent or non-existent in normal healthy individuals despite the presence of IL-22

in their gastric environment

3.3 Summary

The results presented in this thesis have demonstrated that there are persistent Th17 responses accompanied by other inflammatory responses in individuals with a past history of HP infection Furthermore, HP-specific responses make up part of the persistent responses

Firstly, a number of volunteers from this study had been subject to yearly followup after detection of a HP infection at their first visit Analysis of histology reports from individuals who had been followed up showed that only 3 out of 18 subjects (with varying stomach lesions the beginning) had regressed PC lesions in their gastric mucosa All other subjects had either remained with their existing PC lesion, or progressed to a more advanced lesion despite being given antibiotic therapy for the

HP infection This mirrors the findings from previous reports that in older subjects above the age of 45, treating for HP does not reduce the risk for advancement of PC lesions

Next, it was the main hypothesis of this thesis that chronic inflammation commonly associated with PC lesions contributed to the continuing risk for PC lesions

advancement as well as gastric cancer This thesis focused on characterizing the role

of Th17 cells, a relatively new subset of CD4+ T helper cells that is commonly

reported to play pathogenic roles in autoimmune diseases and inflammatory

conditions Persistent Th17 cells are observed both in the peripheral blood (CD4+

17A+

associated with PC lesions Aside from cellular counts, ex vivo cytokine

concentrations of various inflammatory cytokines remained high even in the subjects with a past HP infection (group P) Although statistical significance was not reached

in most cytokine analyses due to the small sample size, there was a trend towards elevated amounts of inflammatory cytokines in individuals with PC lesions There was also elevated amounts of hBD-2 which was chemotactic for CCR6+ Th17 cells,

as well as an elevated level of IL-1β, important for maintaining Th17 cells, and several other cytokines which could contribute to the inflammatory environment that

is conducive for the persistence of Th17 cells The definite reason or mechanism for the observed increase in Th17 infiltrate is however, not clear from the results

presented

IL-22, another Th17 derived cytokine was also studied at depth due to many recent reports implicating a vital role for it in gut homeostasis and innate immunity Total IL-22 levels across three subject groups were not different, indicating that this

cytokine was perhaps not involved in chronic inflammation in the gastric mucosa However, in line with its role in innate immunity, CD4+IL-22+ cells were significantly elevated in group A subjects, but not group P and N

Intra-individual correlations between levels of pro-inflammatory cytokines were performed to determine the biological significance of high levels of the high levels of pro-inflammatory cytokines Correlations between IL-17A, IL-1β, IFNγ, IL-8 and interestingly IL-10 were found, demonstrating that these cytokines make up the pro-inflammatory environment accompanying the chronic infiltrate in the gastric mucosa

Furthermore, increased expression of IL-1R1 was found on IL-17A+ cells in the gastric infiltrate in group P subjects with severe intestinal metaplasia, further

emphasizing that the IL-1β/IL-1R1 axis may significantly contribute to chronic inflammation and hence persisting risk, especially since polymorphisms in IL-1β is well known to contribute to gastric cancer risk

The effects of IL-1β were further studied on gastric epithelial cells (GECs) freshly isolated from biopsies In particular, the effects of IL-1β on hBD-2 expression was studied given the evidence in literature that hBD-2 is induced in epithelial cells by IL-1β and given that hBD-2 is biologically significant in this context (chemoattractive for CCR6+ Th17 cells) hBD-2 was induced by IL-1β and IL-17A and where IL-22 was present, there was a synergistic effect on hBD-2 induction There was a slight intra-individual correlation between hBD-2 and IL-17A and IL-1β and a statistically significant correlation between hBD-2 and IL-22, suggesting that the presence of IL-

22 correlates with increased hBD-2 presence IL-1β-stimulated GEC supernatant was also found to be chemoattractive for autologous expanded LPMCs in 2 individuals

The antigen-specificity of persistent T cells in group P subjects was investigated and HP-specific responses were detected in both the PBMCs and expanded LPMCs of

group A and P subjects This demonstrated that at least part of the chronic

inflammatory infiltrate is made up of HP-specific cells and that HP infection plays a role in triggering initial immune responses that have the capacity to persist in the infected individual, and perhaps contribute to chronic inflammation even after

clearance of the infection

Finally, the effects of the cytokines of recent interest, IL-17A, IL-1β and IL-22, on gastric epithelial cell and immunopathology were studied IL-22R1 but not IL-1R1 and IL-17R1 was found to be down-regulated in the presence of inflammation All three cytokines resulted in an increase in proliferation of AGS cell lines and IL-22 activated the STAT3 pathway in AGS and KatoIII cells Presence of these cytokines

in the gastric environment also resulted in the up-regulation of hBDs and S100

proteins as well as Reg1α Synergistic effects was seen with IL-1β and IL-22, but less with IL-17A and IL-22 The induction of these downstream genes also contribute to inflammation in the gastric mucosa

Altogether, the findings of this research project has provided evidence that chronic inflammation does persist in individuals even after they have been treated for HP, especially in older individuals – this proves the first hypothesis of this thesis, ‘that HP-specific Th17 responses, which were primed during initial infection, persist in the gastric mucosa of patients even when they no longer have an active infection’ This chronic inflammation is made up of persistent Th17 infiltrate accompanied by an array of pro-inflammatory cytokines and chemokines, including IL-17A, IL-1β, IFNγ, hBD-2 and IL-22 These cytokines are able of further inducing downstream pro-

partially addresses the second hypothesis of this thesis ‘that these Th17 cells

contribute to immunopathology associating with precancerous lesions, and perhaps

indirectly, cancer risk’ A direct in vivo link between the Th17 cells and

immunopathology was not proved However, several lines of evidences in the

literature suggest a role for Th17 cells in driving carcinogenesis These are discussed below