Helicobacter pylori infection induces a number of pro-inflammatory signaling pathways contributing to gastric inflammation and carcinogenesis. Among those, NF-κB signaling plays a pivotal role during infection and malignant transformation of the gastric epithelium. However, deficiency of the adaptor molecule myeloid differentiation primary response 88 (MyD88), which signals through NF-κB, led to an accelerated development of gastric pathology upon H. felis infection, but the mechanisms leading to this phenotype remained elusive. Non-canonical NF-κB signaling was shown to aggravate H. pylori-induced gastric inflammation via activation of the lymphotoxin β receptor (LTβR). In the present study, we explored whether the exacerbated pathology observed in MyD88-deficient (Myd88−/−) mice was associated with aberrant activation of non-canonical NF-κB. Our results indicate that, in the absence of MyD88, H. felis infection enhances the activation of non-canonical NF-κB that is associated with increase in Cxcl9 and Icam1 gene expression and CD3+ lymphocyte recruitment. In addition, activation of signal transducer and activator of transcription 3 (STAT3) signaling was higher in Myd88−/− compared to wild type (WT) mice, indicating a link between MyD88 deficiency and STAT3 activation in response to H. felis infection. Thereby, MyD88 deficiency results in accelerated and aggravated gastric pathology induced by Helicobacter through activation of non-canonical NF-κB.
Helicobacter pylori is one of the most prevalent infections affecting around half of the world’s population. Inflammation elicited by H. pylori is characterized by strong infiltration of neutrophils, monocytes and lymphocytes into the gastric mucosa1. If left untreated, H. pylori-induced chronic inflammation can eventually progress to gastric cancer. Therefore, H. pylori infection is considered a main risk factor for gastric cancer and mucosa-associated lymphoid tissue (MALT) lymphoma.
Activation of a number of signaling pathways, such as nuclear factor-kappa B (NF-κB) or signal transducer and activator of transcription 3 (STAT3) signaling, in response to the infection are reported to be involved in the carcinogenetic process leading to the development of gastric tumors2,3,4.
NF-κB signaling activation is one of the first events observed in gastric epithelial cells upon H. pylori infection5. Notably, NF-κB signaling can be executed by two different arms, the canonical and the non-canonical pathways6, both of which have been related to H. pylori-induced pathology by activating the expression of several genes contributing to the inflammatory response elicited by the bacterium. Canonical NF-κB signaling is mainly activated by engagement of bacterial products to immune receptors such as Toll-like receptors (TLRs) and NOD-like receptors (NLRs)7,8. Signaling by most TLRs is mediated by the adaptor molecule MyD88, which was reported to confer protection from inflammation-associated cancer induced by H. felis infection, through yet unidentified mechanisms. Thus, Myd88−/− mice showed a rapid progression of H. felis-induced inflammation to precancerous and cancerous lesions in the stomach when compared to wild type (WT) animals9.
Non-canonical NF-κB is activated by lymphotoxin α1β2 (LT) and LIGHT (TNFSF14; homologous to lymphotoxin exhibits inducible expression and competes with HSV glycoprotein D for binding to herpes virus entry mediator, a receptor expressed on T lymphocytes) produced in response to H. pylori infection and is dependent on a functional Type IV secretion system10. Binding of LT or LIGHT to LTβ receptor (LTβR) leads to NF-κB inducing kinase (NIK) activation and proteosomal processing of p100 to p52. p52 forms heterodimers with RelB that translocate into the nucleus of the cell regulating the expression of target genes involved in inflammation such as chemokines, CXCL13 and CXCL10 and cellular adhesion molecules such as vascular cell adhesion molecule (V-CAM) and intracellular adhesion molecule (I-CAM)6,11. Blocking of LTβR during H. pylori infection resulted in reduced gastric pathology, lower infiltration of inflammatory cells and decreased expression of inflammatory chemokines. These observations confirmed a direct link between LTβR signaling and H. pylori induced pathology in vivo10. In addition, analysis of human gastric tissue presenting with different degree of H. pylori-associated gastritis and gastric tumors further suggested an important role for non-canonical NF-κB signaling in gastric carcinogenesis in humans10.
Gastric carcinogenesis has been linked to hyperactivation of STAT3 signaling, which has been shown to be induced by H. pylori infection12. STAT3 plays key roles in embryonic development, inflammation, cell differentiation, proliferation and apoptosis as well as in tumorigenesis. Interestingly, LT produced by tumor-infiltrating B cells was reported to signal through LTβR to induce IKKα nuclear translocation and STAT3 activation13, while LIGHT induced STAT3 phosphorylation through NIK in prostate cancer cells14.
Although the role of non-canonical NF-κB signaling in H. pylori-induced gastritis is clear, its possible involvement in induction of gastric precancerous lesions remained elusive, since mice infected with H. pylori do not develop severe gastric pathology. Therefore, we assessed non-canonical NF-κB activation in gastric lesions induced by H. felis infection. Moreover, and given the intimate crosstalk between both NF-κB pathways during H. pylori infection10, we compared activation levels between wild type and Myd88−/− mice, since the latter showed enhanced gastric pathology upon H. felis infection9. Our results indicate that activation of non-canonical NF-κB signaling might contribute to the early onset and rapid progression of gastric precancerous lesions observed in H. felis-infected Myd88−/− mice by enhancing the expression of pro-inflammatory chemokines, favoring the recruitment of inflammatory cells and activating STAT3 in the stomach.
Non-canonical NF-κB signaling is enhanced in Myd88 −/− mice upon H. felis infection
We previously observed that Myd88−/− mice develop more severe pathology than WT mice upon infection with H. felis9. Thus, severe gland atrophy, hyperplasia, and dysplasia was observed in the stomach of infected Myd88−/− mice (Supplementary Fig. 1a and ref.9). To identify possible molecular mechanisms involved, we first analyzed whether differences in NF-κB activation could be detected between the same previously reported WT and Myd88−/− infected mice. To this end, we stained gastric tissue samples for p65 (Supplementary Fig. 1b). We could not detect differences between control and infected WT and Myd88−/− mice, excluding the involvement of canonical NF-κB in more severe phenotype observed in Myd88−/− mice.
Recently, H. pylori was reported to induce non-canonical NF-κB signaling in gastric epithelial cells, which highly contributed to the inflammatory response to the bacterium10. To assess whether non-canonical NF-κB signaling could be related to the more severe pathology observed in Myd88−/− mice upon infection with H. felis, we first evaluated nuclear translocation of RelB into the nucleus of gastric epithelial cells. Mice infected with H. felis showed activation of non-canonical NF-κB (Fig. 1). Notably, mice deficient for MyD88 showed increased non-canonical NF-κB signaling in the stomach upon H. felis infection when compared to WT animals after 25 and 47 weeks of infection (Fig. 1). In contrast, there were no differences in activation of non-canonical NF-κB between WT and Myd88−/− mice when left uninfected (Fig. 1). We confirmed activation of non-canonical NF-κB by staining the same tissue samples for NIK (Supplementary Fig. 1c). These observations suggest that, at least partially, enhanced activation of non-canonical NF-κB in response to Helicobacter infection might contribute to the more severe pathology observed in mice lacking MyD88.
LIGHT might be the ligand responsible for non-canonical NF-κB activation upon H. felis infection
We recently showed that activation of non-canonical NF-κB by H. pylori was induced by LT through LTβR expressed on gastric epithelial cells10. Thus, we analyzed whether activation of non-canonical NF-κB signaling upon H. felis infection was driven by LT. Wild type- and Myd88−/− infected mice showed increased levels of Ltb in the stomach (Fig. 2a). After 25 weeks we detected a significant increase in Ltb expression upon H. felis infection, which was not observed in Myd88−/− infected mice. In addition, no differences in Ltb expression were detected between wild type and Myd88−/− infected mice after 25 or 47 weeks (Fig. 2a), indicating that LTβ is not the ligand inducing non-canonical NF-κB activation during H. felis infection. Thus, we next analyzed Tnfsf14, which is also expressed in response to H. pylori10 and is an alternative ligand of LTβR activating non-canonical NF-κB. Notably, H. felis infection significantly increased Tnfsf14 expression in the stomach of Myd88−/− mice (Fig. 2b), while only a slight increase was detected in WT animals after 25 weeks of infection. We also performed immunohistochemistry for LIGHT, however the antibody used did not work reliably on formalin fixed, paraffin-embedded tissue sections. These results indicate that LIGHT is the main ligand activating non-canonical NF-κB signaling in response to H. felis.
Lack of MyD88 relates to enhanced recruitment of T and B cells to the stomach upon H. felis infection
To analyze whether differential activation of non-canonical NF-κB pathway triggered by LIGHT translated into changes in the recruitment of inflammatory cells between WT and Myd88−/− mice during H. felis infection, we stained stomach tissue samples for T cell, B cells, and macrophages. We observed that mice lacking MyD88 showed higher infiltration of CD3+ cells into the stomach when compared to WT mice (Fig. 3a). No differences in CD4+ cells recruitment was detected after 25 weeks of infection (Fig. 3b), while after 47 weeks of infection, higher numbers of CD4+ cells were observed in the gastric tissue of Myd88−/− infected mice (Fig. 3b).
Wild type and Myd88−/− mice showed large lymphoid structures in the gastric mucosa after infection (Fig. 3c). Those were mostly composed of B cells, which were also found infiltrating the epithelium. Notably, more intraepithelial B cells were detected in Myd88−/− infected animals compared to WT mice (Fig. 3c), however the difference was not statistically significant.
No differences in macrophage infiltration between WT and Myd88−/− mice after 25 or 47 weeks post-infection, as detected by F4/80 staining (Supplementary Fig. 2).
These results suggest that lack of MyD88 favors the recruitment of T cells into the gastric epithelium in response to H. felis infection.
Cxcl9 and Icam1 are up-regulated in Myd88 −/− mice upon H. felis infection
We examined the expression of chemokines related to non-canonical NF-κB signaling that could explain the differences in T cell and B cell recruitment observed between WT and Myd88−/− mice. We first analyzed the levels of Cxcl13, which is a B cell chemoattractant regulated by non-canonical NF-κB6. H. felis infection up-regulated Cxcl13 expression in WT mice after 25 weeks of infection (Supplementary Fig. 3), while its expression was barely induced in MyD88−/− infected mice (Supplementary Fig. 3). After 47 weeks, Myd88−/− infected mice also showed up-regulation of Cxcl13 expression in the stomach, albeit at a lower level when compared to WT mice (Supplementary Fig. 3). This observation suggests that Cxcl13 might not be involved in the enhanced B cell recruitment observed in Myd88−/− during H. felis infection.
Next, we focused on chemokines involved in T cell recruitment. During H. pylori infection, Cxcl9 was shown to contribute to the recruitment of inflammatory T cells into the gastric mucosa15,16. Interestingly this chemokine has been previously related to LIGHT and NF-κB signaling6,17,18. Infection with H. felis led to Cxcl9 up-regulation only in Myd88−/− mice (Fig. 4a) after 25 and 47 weeks of infection, while no induction was detected in WT animals.
Lastly, we analyzed the expression of Icam1, an adhesion molecule involved in transmigration of leucocytes into tissues, which can be regulated by non-canonical NF-κB6 and induced by LIGHT19. Similarly to Cxcl9, WT mice infected with H. felis did not show increased expression of Icam1 in the stomach after 25 or 47 weeks of infection (Fig. 4b). In contrast, high levels of Icam1 were detected in Myd88−/− mice infected with H. felis for 25 weeks that were reduced but still high at 47 weeks post-infection (Fig. 4b). Together, these results suggest that CXCL-9 and ICAM-1, which have been previously reported to be induced by LIGHT, might be involved in the higher recruitment of T cells to the stomach of Myd88−/− mice during H. felis infection.
STAT3 activation is increased in Myd88 −/− mice upon H. felis infection
STAT3 signaling has been linked to severe gastric pathology during H. pylori infection as well as to gastric cancer development3. Interestingly, LIGHT was reported to activate STAT3 through non-canonical NF- κB signaling14. Therefore, we analyzed whether changes in STAT3 activation could correlate with Tnfsf14 expression and the more severe pathology observed in Myd88−/− mice. H. felis infection led to activation of STAT3 in the stomach of WT and Myd88−/− mice after 25 and 47 weeks of infection (Fig. 5 and Supplementary Fig. 4). Notably, STAT3 activation was higher in Myd88−/− mice, suggesting that activation of STAT3 might be linked to the gastric pathology observed in Myd88−/− mice.
In a previous study we found MyD88 to play a protective role during H. felis infection9. Thus, mice deficient in MyD88 showed rapid progression of gastric pathology upon H. felis infection related to enhanced epithelial apoptosis and cell proliferation as well as altered expression of pro-inflammatory cytokines9. However, the molecular mechanisms driving early development of gastric malignancy in Myd88−/− mice remained mostly elusive. Two recently published clinical data sets of human stomach cancers from The Cancer Genome Atlas (TCGA)20,21 revealed the presence of MYD88 gene deletions and mutations in gastric and esophageal adenocarcinomas. These studies indicate for the first time a link between MyD88 deficiency and gastric cancer, thereby validating observations in H. felis-infected Myd88−/− mice.
We recently showed that up-regulation of the ligands LT and LIGHT by H. pylori is responsible for activating non-canonical NF-κB in gastric epithelial cells through LTβR, driving the inflammatory response to H. pylori10. Therefore, we sought to investigate whether activation of non-canonical NF-κB could be related to the more severe pathology observed in mice deficient in MyD88. We observed up-regulation of Ltb expression upon H. felis infection but no differences between WT and Myd88−/− mice that could explain the stronger activation of non-canonical NF-κB observed in the gastric epithelium of mice deficient for MyD88. In contrast, Myd88−/− mice showed higher levels of Tnfsf14 expression than WT mice, suggesting that LIGHT might be the ligand responsible for non-canonical NF-κB activation in these mice. What still remains unclear is how LIGHT is up-regulated in the context of H. pylori or H. felis infection and which H. felis bacterial factors can be involved, since H. felis lacks a type IV secretion system, which is important for Tnfsf14 induction in vivo during H. pylori infection. Our previous observations indicated that during H. pylori infection TNFSF14 expression was independent of canonical NF-κB activation10. Therefore, other signaling networks must be involved. In this context, regulation studies demonstrated an important role for Ca2+ signaling pathways and nuclear factor of activated T-cells (NFAT) transcription factors in LIGHT transcription induced by T cell activation. Moreover, Sp1 and Ets1 transcription factors were important for basal constitutive expression of LIGHT in T cells22. These pathways and transcription factors have been shown to be altered by H. pylori23,24,25 and thus might also be involved in LIGHT up-regulation. However, further experiments are needed to understand the mechanisms regulating LIGHT expression during Helicobacter infection.
The analysis of the gastric inflammatory infiltrates in H. felis-infected WT and Myd88−/− mice showed increased recruitment of T cells as well as B cells in the stomach of mice lacking MyD88. Notably, Tnfsf14 up-regulation in Myd88−/− mice did not only correlate with hyperactivation of non-canonical NF-κB signaling, but also with the expression of chemokines (Cxcl9) involved in the recruitment of immune cells to sites of inflammation as well as with adhesion molecules (Icam1). Forced expression of LIGHT in tumors was found to increase the expression of interferon gamma (IFNγ) as well as chemoattractant cytokines including CXCL917, while expression of ICAM1 was induced in bronchial epithelial cells upon treatment with LIGHT. Therefore, up-regulation of Tnfsf14 elicited by H. felis infection in Myd88−/− mice might contribute to immune cell recruitment by inducing the expression of chemoattractant molecules, specifically Cxcl9, since no differences in other chemokines such as Cxcl10 or Cxcl13 were detected between WT and Myd88−/−. Interestingly, CXCL9 was shown to contribute to the recruitment of inflammatory T cells into the gastric mucosa of H. pylori infected individuals15. In addition, H. pylori eradication in p27-deficient mice, which are prone to gastric cancer development upon infection, attenuated gastric pathology and this was associated with reduced expression of Cxcl9 in the stomach16. Thus, LIGHT might support early and rapid progression of gastric pathology observed in Myd88−/− mice by up-regulating the expression of chemokines important for the recruitment of inflammatory cells into the stomach.
Enhanced proliferation was detected in the gastric epithelium of H. felis-infected Myd88−/− mice, contributing to the increased pathology observed in these mice9. Interestingly, interaction of LIGHT with LTβR was demonstrated to increase the survival and proliferation of bone marrow-derived mesenchymal stem cells through activation of survival and anti-apoptotic proteins26. In that study, activation of STAT3 by LIGHT played a central role by inducing the expression of platelet-derived growth factor (PDGF) and tumor growth factor β (TGFβ). We observed increased STAT3 activation in Myd88−/− mice upon H. felis infection. The role of STAT3 in gastric carcinogenesis has been extensively reported. Hyperactivation of STAT3 in mice bearing a knock-in mutation in the gp130 receptor was sufficient to induce the development of gastric tumors4. In humans, STAT3 phosphorylation has been associated with worse prognosis of gastric tumors27 and several studies reported activation of STAT3 by H. pylori in gastric epithelial cells12,28,29. Notably, LIGHT-induced activation of STAT3 was shown to be mediated by NIK14, an essential kinase for the activation of non-canonical NF-κB. Thus, LIGHT can enhance its effects on cellular proliferation, survival, and inflammation by concomitant activation of non-canonical NF-κB and STAT3 signaling. Our results suggest that the activation of these two signaling pathways in mice lacking MyD88 might be a driving force for the development of gastric pathology induced upon H. felis infection.
Although our results suggest a link between pathology induced by H. felis and activation of non-canonical NF-κB, we cannot conclude whether our findings are secondary to H. felis infection or consequent upon the development of atrophy/dysplasia. Studying an earlier time point may help to elucidate this.
Mouse gastric tissue samples
Formalin-fixed, paraffin-embedded (FFPE) samples prepared from gastric samples from mice previously described in ref.9 were used in the present study. Briefly, 7 to 8-week old male mice in the C57BL/6 background, wild type (WT) and Myd88 deficient (Myd88−/−) were infected with H. felis strain CS1 (ATCC 49179, Manassas VA) (8 mice per background) for 25 or 47 weeks9,30,31,32,33,34. Mice were inoculated with H. felis grown in Brain Heart Infusion (BHI) (109 organisms/mouse) by oral gavage every other day for a total of 3 inoculations. Control mice received BHI only. During infection three Myd88−/− mice died, one before and two after 25 weeks. At the selected time post-infection, mice were euthanized and stomachs were aseptically removed. Longitudinal cross-sections of gastric tissue from each mouse were cut, fixed in neutral buffered 10% formalin, and paraffin embedded. All H. felis-infected mice were confirmed colonized by assessing the flagella filament B (flaB) gene by real-time PCR, which was expressed similarly between WT and Myd88−/− mice9. All animal procedures were performed in accordance with institutional guidelines and regulations and approved by the Animal Care Committee at the University of California, San Diego.
Isolation of RNA from gastric tissue samples
RNA was isolated from FFPE gastric tissue samples using RecoverAll Total Nucleic Acid Isolation Kit (Life Technologies, Carlsbad CA) according to the manufacturer’s instructions. Briefly, a total of 20 pieces (20 μm thick/piece) of gastric tissue were cut using a microtome (Leica Biosystems, Buffalo Grove, IL). The gastric tissue sections were deparaffinilized with xylene followed by two washes in 100% ethanol to ensure complete paraffin removal and the pellet air dried. Protease digestion buffer was added to the pellet for 1 hr followed by RNA isolation buffer. The resulting RNA solution was passed through a filter cartridge and RNA eluted using nuclease-free water. RNA quality was determined using the Nanodrop system (Thermo Fisher, Waltham MA) by reading absorbance levels at 260 nm.
Quantitative real-time RT-PCR
Real-time RT-PCR (reverse transcription polymerase chain reaction) was performed as previously described in our studies9,35 to determine expression of selected genes including, Tnfsf14, Cxcl13, Cxcl9, Ltb, and Icam1. RNA isolated from FFPE samples (2 μg/sample) was reverse transcribed into cDNA using High Capacity cDNA Reverse Transcription Kit (Thermo Fisher, Waltham MA). 2 μl of cDNA was used per well in a total of 10 μl reaction mix for amplification using the StepOne Real Time PCR (Applied Biosystems, Carlsbad CA). The amplification conditions consisted of an initial cycle of 95 °C for 5 minutes followed by 40 cycles of amplification with denaturation as follows: 95 °C for 15 sec, 60 °C for 20 sec, 72 °C for 40 sec. Gene expression levels were normalized to Glyceraldehyde 3-phosphate dehydrogenase (Gapdh). Data were analyzed using a comparative cycle threshold calculations (ΔΔCT, Applied Biosystems) and plotted using GraphPad Prism software. Primers used are listed in Supplementary Table 1.
Gastric tissue samples from WT and Myd88−/− control and H. felis-infected mice were incubated with specific antibodies overnight (Supplementary Table 2) after antigen retrieval in a pressure cooker using 0.01 M sodium citrate. These antibodies have been previously validated by our group and others36,37,38. After incubation with HRP-conjugated secondary antibodies, sections were developed using SignalStain DAB substrate (Cell Signaling) following manufacturer’s instructions. A secondary antibody only control was included to ensure that the secondary antibody did not unspecifically bind to certain cellular compartments. Samples were counterstained with hematoxylin (Morphisto). For automated image acquisition, Olympus Virtual Slide System VS120 (Olympus) was used. Stainings were visually quantified in a blinded manner. Five random high power field (HPF) areas per sample were scored blinded.
Statistical analysis was performed using GraphPad Prism. Normal distribution of the data was tested using Kolmogorov-Smirnov and D’Agostino & Pearson omnibus normality tests. Normal distributed data was analyzed using T-test, while non-normally distribute data were analyzed using Mann Whitney U Test. For multiple comparisons, ANOVA with Bonferroni’s correction (normal distribution) or Kruskal-Wallis with Dunn’s comparison test (non-normal distribution) were used. Statistical significance was established when p < 0.05.
Moyat, M. & Velin, D. Immune responses to Helicobacter pylori infection. World journal of gastroenterology: WJG 20, 5583–5593, https://doi.org/10.3748/wjg.v20.i19.5583 (2014).
Ernst, M., Thiem, S., Nguyen, P. M., Eissmann, M. & Putoczki, T. L. Epithelial gp130/Stat3 functions: an intestinal signaling node in health and disease. Semin Immunol 26, 29–37, https://doi.org/10.1016/j.smim.2013.12.006 (2014).
Menheniott, T. R., Judd, L. M. & Giraud, A. S. STAT3: a critical component in the response to Helicobacter pylori infection. Cellular microbiology 17, 1570–1582, https://doi.org/10.1111/cmi.12518 (2015).
Jenkins, B. J. et al. Hyperactivation of Stat3 in gp130 mutant mice promotes gastric hyperproliferation and desensitizes TGF-beta signaling. Nature medicine 11, 845–852, https://doi.org/10.1038/nm1282 (2005).
Backert, S. & Naumann, M. What a disorder: proinflammatory signaling pathways induced by Helicobacter pylori. Trends in microbiology 18, 479–486, https://doi.org/10.1016/j.tim.2010.08.003 (2010).
Wolf, M. J., Seleznik, G. M., Zeller, N. & Heikenwalder, M. The unexpected role of lymphotoxin beta receptor signaling in carcinogenesis: from lymphoid tissue formation to liver and prostate cancer development. Oncogene 29, 5006–5018, https://doi.org/10.1038/onc.2010.260 (2010).
Viala, J. et al. Nod1 responds to peptidoglycan delivered by the Helicobacter pylori cag pathogenicity island. Nature immunology 5, 1166–1174, https://doi.org/10.1038/ni1131 (2004).
Smith, S. M. Role of Toll-like receptors in Helicobacter pylori infection and immunity. World journal of gastrointestinal pathophysiology 5, 133–146, https://doi.org/10.4291/wjgp.v5.i3.133 (2014).
Banerjee, A. et al. Deficiency of the myeloid differentiation primary response molecule MyD88 leads to an early and rapid development of Helicobacter-induced gastric malignancy. Infection and immunity 82, 356–363, https://doi.org/10.1128/IAI.01344-13 (2014).
Mejias-Luque, R. et al. Lymphotoxin beta receptor signalling executes Helicobacter pylori-driven gastric inflammation in a T4SS-dependent manner. Gut. https://doi.org/10.1136/gutjnl-2015-310783 (2016).
Ansel, K. M. et al. A chemokine-driven positive feedback loop organizes lymphoid follicles. Nature 406, 309–314, https://doi.org/10.1038/35018581 (2000).
Lee, I. O. et al. Helicobacter pylori CagA phosphorylation status determines the gp130-activated SHP2/ERK and JAK/STAT signal transduction pathways in gastric epithelial cells. The Journal of biological chemistry 285, 16042–16050, https://doi.org/10.1074/jbc.M110.111054 (2010).
Ammirante, M., Luo, J. L., Grivennikov, S., Nedospasov, S. & Karin, M. B-cell-derived lymphotoxin promotes castration-resistant prostate cancer. Nature 464, 302–305, https://doi.org/10.1038/nature08782 (2010).
Nadiminty, N. et al. LIGHT, a member of the TNF superfamily, activates Stat3 mediated by NIK pathway. Biochem Biophys Res Commun 359, 379–384, https://doi.org/10.1016/j.bbrc.2007.05.119 (2007).
Eck, M. et al. CXC chemokines Gro(alpha)/IL-8 and IP-10/MIG in Helicobacter pylori gastritis. Clin Exp Immunol 122, 192–199 (2000).
Zhang, S. et al. Early or late antibiotic intervention prevents Helicobacter pylori-induced gastric cancer in a mouse model. Cancer letters 359, 345–351 (2015).
Kanodia, S. et al. Expression of LIGHT/TNFSF14 combined with vaccination against human papillomavirus Type 16 E7 induces significant tumor regression. Cancer research 70, 3955–3964, https://doi.org/10.1158/0008-5472.CAN-09-3773 (2010).
Miyagaki, T. et al. Low herpesvirus entry mediator (HVEM) expression on dermal fibroblasts contributes to a Th2-dominant microenvironment in advanced cutaneous T-cell lymphoma. The Journal of investigative dermatology 132, 1280–1289, https://doi.org/10.1038/jid.2011.470 (2012).
Qiu, H. N., Wong, C. K., Dong, J., Lam, C. W. & Cai, Z. Effect of tumor necrosis factor family member LIGHT (TNFSF14) on the activation of basophils and eosinophils interacting with bronchial epithelial cells. Mediators Inflamm 2014, 136463, https://doi.org/10.1155/2014/136463 (2014).
Cancer Genome Atlas Research, N. Comprehensive molecular characterization of gastric adenocarcinoma. Nature 513, 202–209, https://doi.org/10.1038/nature13480 (2014).
Cancer Genome Atlas Research, N. et al. Integrated genomic characterization of oesophageal carcinoma. Nature 541, 169–175, https://doi.org/10.1038/nature20805 (2017).
Castellano, R. et al. Mechanisms regulating expression of the tumor necrosis factor-related light gene. Role of calcium-signaling pathway in the transcriptional control. The Journal of biological chemistry 277, 42841–42851, https://doi.org/10.1074/jbc.M207689200 (2002).
Marlink, K. L. et al. Effects of Helicobacter pylori on intracellular Ca2+ signaling in normal human gastric mucous epithelial cells. Am J Physiol Gastrointest Liver Physiol 285, G163–176, https://doi.org/10.1152/ajpgi.00257.2002 (2003).
Yokoyama, K. et al. Functional antagonism between Helicobacter pylori CagA and vacuolating toxin VacA in control of the NFAT signaling pathway in gastric epithelial cells. Proceedings of the National Academy of Sciences of the United States of America 102, 9661–9666, https://doi.org/10.1073/pnas.0502529102 (2005).
Strowski, M. Z. et al. Helicobacter pylori stimulates host vascular endothelial growth factor-A (vegf-A) gene expression via MEK/ERK-dependent activation of Sp1 and Sp3. FASEB journal: official publication of the Federation of American Societies for Experimental Biology 18, 218–220, https://doi.org/10.1096/fj.03-0055fje (2004).
Heo, S. K. et al. LIGHT (TNFSF14) Increases the Survival and Proliferation of Human Bone Marrow-Derived Mesenchymal Stem Cells. PloS one 11, e0166589, https://doi.org/10.1371/journal.pone.0166589 (2016).
Li, M. X. et al. Prognostic Role of Phospho-STAT3 in Patients with Cancers of the Digestive System: A Systematic Review and Meta-Analysis. PloS one 10, e0127356, https://doi.org/10.1371/journal.pone.0127356 (2015).
Bronte-Tinkew, D. M. et al. Helicobacter pylori cytotoxin-associated gene A activates the signal transducer and activator of transcription 3 pathway in vitro and in vivo. Cancer research 69, 632–639, https://doi.org/10.1158/0008-5472.CAN-08-1191 (2009).
Lee, K. S. et al. Helicobacter pylori CagA triggers expression of the bactericidal lectin REG3gamma via gastric STAT3 activation. PloS one 7, e30786, https://doi.org/10.1371/journal.pone.0030786 (2012).
Cai, X. et al. Helicobacter felis eradication restores normal architecture and inhibits gastric cancer progression in C57BL/6 mice. Gastroenterology 128, 1937–1952 (2005).
Cai, X. et al. Overcoming Fas-mediated apoptosis accelerates Helicobacter-induced gastric cancer in mice. Cancer research 65, 10912–10920, https://doi.org/10.1158/0008-5472.CAN-05-1802 (2005).
Fox, J. G. et al. Hypertrophic gastropathy in Helicobacter felis-infected wild-type C57BL/6 mice and p53 hemizygous transgenic mice. Gastroenterology 110, 155–166 (1996).
Houghton, J. et al. Gastric cancer originating from bone marrow-derived cells. Science 306, 1568–1571, https://doi.org/10.1126/science.1099513 (2004).
Tu, S. et al. Overexpression of interleukin-1beta induces gastric inflammation and cancer and mobilizes myeloid-derived suppressor cells in mice. Cancer cell 14, 408–419, https://doi.org/10.1016/j.ccr.2008.10.011 (2008).
Obonyo, M., Rickman, B. & Guiney, D. G. Effects of myeloid differentiation primary response gene 88 (MyD88) activation on Helicobacter infection in vivo and induction of a Th17 response. Helicobacter 16, 398–404, https://doi.org/10.1111/j.1523-5378.2011.00861.x (2011).
Mejias-Luque, R. et al. Lymphotoxin beta receptor signalling executes Helicobacter pylori-driven gastric inflammation in a T4SS-dependent manner. Gut 66, 1369–1381, https://doi.org/10.1136/gutjnl-2015-310783 (2017).
Mejias-Luque, R. et al. Inflammation modulates the expression of the intestinal mucins MUC2 and MUC4 in gastric tumors. Oncogene 29, 1753–1762, https://doi.org/10.1038/onc.2009.467 (2010).
Jin, Z., Liang, F., Yang, J. & Mei, W. hnRNP I regulates neonatal immune adaptation and prevents colitis and colorectal cancer. PLoS genetics 13, e1006672, https://doi.org/10.1371/journal.pgen.1006672 (2017).
This work was supported by the National Cancer Institute of the National Institute of Health under award R21CA210227 and the Deutsche Forschungsgemeinschaft (DFG) GE 2042/6-1. M.H. is supported by ERC-CoG and the SFBTR projects 36, 179 and 209.
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.