- Email: [email protected]
Esophageal Cancers and Helicobacter pylori: Do Host Genes Matter?
Editorials Esophageal Cancers and Helicobacter pylori: Do Host Genes Matter?
See “Association of Helicobacter pylori infection with reduced risk for esophageal cancer is independent of environmental and genetic modifiers,” by Whiteman DC, Parmar P, Fahey P, et al, on page 73.
E
sophageal adenocarcinoma has risen at an alarming rate in Western countries over the past few decades. This rise has not been met with any substantial improvement in treatment, making this cancer among the deadliest. It is likely that prevention offers the most credible strategy to defeat this cancer. As such, it is imperative to understand the basic pathogenesis of this condition and its precursors. The classic risk factors for this cancer include gastroesophageal reflux disease (GERD), smoking, obesity and a diet low in fresh fruit and vegetables.1 More recent work suggests that Helicobacter pylori infection, especially with more virulent cagA-positive strains, decreases the risk of esophageal adenocarcinoma.2 The mechanism of this protective effect is not clear but it has been postulated that it may be due to H pylori-induced gastric atrophy and hypochlorhydria, both of which reduce acid exposure of the lower esophagus.3 There is, therefore, an emerging consensus that pathologic changes at the gastroesophageal junction are determined to a large extent by the gastric pathophysiologic phenotype. There are 3 main gastric phenotypes that result from chronic H pylori infection: (1) the commonest by far is a mild pangastritis that does not affect gastric physiology and is not associated with significant human disease, (2) a corpus-predominant gastritis associated with multifocal gastric atrophy, hypochlorhydria, and increased risk of gastric cancer (the gastric cancer phenotype), and (3) an antral-predominant gastritis associated with high gastric acid secretion and increased risk of duodenal ulcer disease (the DU phenotype).4 The cause of these divergent clinical outcomes is most likely a combination of host genetic, bacterial, and environmental factors.5 Proinflammatory polymorphisms in the interleukin-1 beta gene (IL1-B) and tumor necrosis factor alpha gene (TNF-A) increase the risk of gastric atrophy, hypochlorhydria, and gastric cancer.6,7 IL-1 in particular is very relevant in this context because it is a very potent inhibitor of gastric acid secretion.8 Recent work by Waghray et al showed that IL-1 promotes gastric atrophy through suppression of sonic hedgehog9 and this contributes to the loss of the parietal cells capacity to secrete hydrochlo-
ric acid.10 TNF-␣ is also an acid inhibitor, albeit weaker than IL-1. It is tempting to speculate that the gastric cancer phenotype with its low acidity and association with acidreducing and atrophy-inducing proinflammatory IL-1 and TNF-␣ genotypes is protective against GERD and adenocarcinoma, but is there any evidence for this? At least 2 studies have addressed the role of these polymorphisms in reflux disease. Queiroz et al compared 98 GERD patients to 285 controls in Brazil and showed that the proinflammatory IL-1B-31 C/C and IL-1RN*2*2 genotypes were associated inversely with GERD in H pylori– positive individuals.11 In these subjects, cagA-positive status, IL-1B and IL-1RN proinflammatory genotypes, and the degree of corpus gastritis were negatively associated with GERD. They also demonstrated that the severity of the disease was inversely associated with the number of IL-1B and IL1-RN proinflammatory alleles. The second study by Ando et al examined 208 H pylori–positive and 112 H pylori–negative Japanese subjects and reported that infected subjects were less likely to report GERD symptoms and to have endoscopic evidence of erosive esophagitis.12 Furthermore, H pylori–positive subjects were more likely to have corpus gastric atrophy than H pylori– negative subjects. Among H pylori–positive patients, those without erosive oesophagitis or GERD symptoms were significantly more likely to have corpus atrophy than subjects with erosive esophagitis or GERD symptoms. Among H pylori–positive patients, subjects homozygous for the proinflammatory allele IL-1B-511T had a significantly lower risk of erosive oesophagitis and GERD symptoms compared with those homozygous for the -511C allele. The authors concluded that a proinflammatory IL-1B genotype was associated with increased risk of atrophy and decreased risk of GERD in H pylori– infected subjects in Japan.12 These data indicate that, in some genetically predisposed subjects, H pylori infection may protect against GERD through induction of gastric atrophy. Thus, these 2 studies clearly show that host genetic factors interact with H pylori infection leading to onset of gastric atrophy and protection against GERD. The crucial question that remains is whether this protection extends to esophageal and junctional adenocarcinomas. In this issue of GASTROENTEROLOGY, Whiteman et al attempted to answer this crucial question.13 They hypothesized that the association between H pylori infection and the risk of adenocarcinomas of the esophagus and GASTROENTEROLOGY 2010;139:17–30
Editorials continued
esophagogastric junction is modified by polymorphisms in the proinflammatory genes IL-1B and TNF-A that regulate gastric acid secretion. As mentioned, this hypothesis assumes that a proinflammatory genetic makeup that reduces acid secretion would protect H pylori–infected subjects from developing esophageal and junctional adenocarcinoma, cancers that most likely develop after decades of acid-induced damage to the lower esophagus. Thus, one would expect that such infected and “protected” subjects would have the so-called gastric cancer phenotype, which is characterized by hypochlorhydria and gastric atrophy, whereas subjects who develop these proximal cancers are either free of H pylori infection or, if infected, have a normal or high gastric acid capacity. The authors set out to answer this question with a sound and adequately powered epidemiologic design. They studied IL-1B and TNF-A polymorphisms but, unfortunately, not IL-1RN. The latter has consistently been shown to be an excellent genetic marker for IL-1 levels and function, and most of the genetic studies executed out in Caucasians found positive associations with proinflammatory conditions when other markers were negative. As it turned out, the authors found no evidence that polymorphisms in IL-1B or TNF-A modified the association between H pylori and esophageal or junctional adenocarcinomas. They concluded that H pylori infection is associated inversely with risks of these 2 cancers, but the reduction in risk was similar across subgroups of potential modifiers, including the host genes studied. We need to consider potential explanations for these negative findings. The simplest to exclude are genotyping errors that might have affected the results. All SNPs were in Hardy-Weinberg equilibrium, suggesting no major genotyping problems. The 1 slight anomaly is in the degree of linkage disequilibrium between the IL-1B-31 and -511 loci, particularly in the controls. In other published reports, the 2 loci are in near total linkage disequilibrium, whereas in this study there seems to be a slight discrepancy. This may reflect population admixture (highly unlikely in this nearly homogeneous Caucasian population), genotyping error, or a genuine difference. It is unlikely that this small discrepancy could explain the negative findings of this study. The second crucial factor to consider is the prevalence of H pylori infection, particularly in the control population. This clearly has to be high enough to allow a demonstration of the interaction with the host genes. The reason for this is that, in Western populations, H pylori infection is actually more likely to induce a mild pangastritis or an antral predominant gastritis that is either associated with normal or increased acid output and certainly with little or no gastric atrophy. Relatively few subjects develop the corpus-predominant gastric cancer phenotype.14 It is therefore difficult to demonstrate a 18
modifying effect of host genes that interact with an infection if the prevalence of this infection is very low. So, what was the prevalence of H pylori in this Australian population? As it turned out, it was quite low at 23%. This is much lower than most populations studied so far and certainly lower than the Brazilian and Japanese populations cited (62.9% and 65%, respectively11,12). Furthermore, the difference in H pylori prevalence between those with the cancer outcome versus those without was also small (13% vs 23%). It is clear, therefore, that despite an adequate design with good numbers for cancer subjects and controls, the low prevalence of the infection was a major obstacle. Serologic diagnosis of the infection is bound to misclassify some subjects but this would impact equally on all participants, especially in the absence of major gastric atrophy. The third explanation is that the gastric phenotype was not assessed in this study, and this clearly reduced the ability to offer a mechanistic insight into the associations between host genotype and clinical outcome. There were no gastric biopsies to map out the type of gastritis, its severity and the presence of atrophy. The study did not report on serologic surrogate markers of atrophy, such as pepsinogen I levels or pepsinogen I/II ratios. This would have given some insight into gastric pathophysiology and this might have informed the statistical analyses a little better. One has to acknowledge, however, that such a study where biopsies are obtained from different parts of the stomach in addition to the primary tumor would only be meaningful if the controls also had a similar mapping of gastric inflammation and atrophy. For such large-scale epidemiologic studies, this is quite a difficult prospect and is not feasible. Ironically, perhaps the best population to prove the author’s hypothesis would not be a Caucasian but rather an Asian one. In Asian countries such as Japan, Korea, and China, the rates of noncardiac gastric cancer are very high, whereas those for esophageal adenocarcinoma are very low. The default gastric phenotype that results from H pylori infection in the East is the gastric cancer phenotype with its low acid and gastric atrophy.14 More interestingly, the prevalence of proinflammatory IL-1B genotypes is significantly higher than in Caucasians. Could this be the reason why such proximal cancers are rare in the Far East? The work by Ando et al offers an insight into the potential answer and it is an open invitation to workers in the field to look into this question. This is indeed a very important issue; it lies at the heart of the pathogenesis of esophageal adenocarcinoma. The fast disappearance of H pylori globally (either spontaneously or owing to mass eradication) might be a logical strategy in the fight against distal gastric cancer, but we need to determine whether the return of normal to high gastric
Editorials continued
acid secretion in Eastern populations is likely to increase the risk of proximal cancers. Where do we go from here? The authors made an admirable attempt at answering an important question, but were perhaps thwarted by a fast disappearing H pylori infection. The major finding that a low infection rate increases the risk of proximal cancers confirms previous studies. The modifying influence of host genes on this risk was neither confirmed nor refuted. Two new studies are urgently needed. The first should look at the modifying effects of host genetics on risk of esophageal and junctional adenocarcinomas in Asia. This requires a large, multicenter, collaborative study pooling all resources across several Asian countries to collect adequate numbers of such a hitherto rare cancer. The results of such a study will allow workers in the field to predict and preempt a sudden rise in the incidence of such cancers in the future and will avoid the alarming increase seen in Western countries over the last 4 decades. The second study should include detailed phenotyping of the gastric environment in studies looking at esophageal cancers. The clue to these proximal cancers lies in the stomach. It might be difficult to stomach such studies, but they are essential if we are to make any significant progress!
MAIRI H. MCLEAN EMAD M. EL–OMAR Division of Applied Medicine Institute of Medical Sciences Aberdeen University Foresterhill, Aberdeen Scotland, United Kingdom References 1. Reid BJ, Li X, Galipeau PC, et al. Barrett’s oesophagus and oesophageal adenocarcinoma: time for a new synthesis. Nat Rev Cancer 2010;10:87–101. 2. Atherton JC, Blaser MJ. Coadaptation of Helicobacter pylori and humans: ancient history, modern implications. J Clin Invest 2009;119:2475–2487. 3. Blaser MJ. Disappearing microbiota: Helicobacter pylori protection against esophageal adenocarcinoma. Cancer Prev Res (Phila Pa) 2008;1:308 –311.
4. Snaith A, El-Omar EM. Helicobacter pylori: host genetics and disease outcomes. Expert Rev Gastroenterol Hepatol 2008;2: 577–585. 5. Amieva MR, El-Omar EM. Host-bacterial interactions in Helicobacter pylori infection. Gastroenterology 2008;134:306 –323. 6. El-Omar EM, Carrington M, Chow WH, et al. Interleukin-1 polymorphisms associated with increased risk of gastric cancer. Nature 2000;404:398 – 402. 7. El-Omar EM, Rabkin CS, Gammon MD, et al. Increased risk of noncardiac gastric cancer associated with proinflammatory cytokine gene polymorphisms. Gastroenterology 2003;124:1193– 1201. 8. El-Omar EM. The importance of interleukin 1beta in Helicobacter pylori associated disease. Gut 2001;48:743–747. 9. Waghray M, Zavros Y, Saqui-Salces M, et al. Interleukin-1beta promotes gastric atrophy through suppression of Sonic Hedgehog. Gastroenterology 2010;138:562–572. 10. van Dop WA, van den Brink GR. Sonic hedgehog: a link between inflammation, gastric atrophy, and acid suppression? Gastroenterology 2010;138:426 – 429. 11. Queiroz DM, Guerra JB, Rocha GA, et al. IL1B and IL1RN polymorphic genes and Helicobacter pylori cagA strains decrease the risk of reflux esophagitis. Gastroenterology 2004;127:73–79. 12. Ando T, El-Omar EM, Goto Y, et al. Interleukin 1B proinflammatory genotypes protect against gastro-oesophageal reflux disease through induction of corpus atrophy. Gut 2006;55:158 –164. 13. Whiteman DC, Parmar P, Fahey P, et al. Association of Helicobacter pylori infection with reduced risk for esophageal cancers is independent of environmental and genetic modifiers. Gastroenterology. 2010;139:73– 83. 14. Naylor GM, Gotoda T, Dixon M, et al. Why does Japan have a high incidence of gastric cancer? Comparison of gastritis between UK and Japanese patients. Gut 2006;55:1545–1552.
Reprint requests Address requests for reprints to: Professor Emad M El-Omar, Division of Applied Medicine, Institute of Medical Sciences, Aberdeen University, Foresterhill, Aberdeen AB25 2ZD, Scotland, UK. E-mail: [email protected]; Phone: ⴙ44-(0)1224-553021; Fax: ⴙ44-(0)1224-555766. Conflicts of interest The author declares no conflicts. © 2010 by the AGA Institute 0016-5085/$36.00 doi:10.1053/j.gastro.2010.05.022
19
See “Association of Helicobacter pylori infection with reduced risk for esophageal cancer is independent of environmental and genetic modifiers,” by Whiteman DC, Parmar P, Fahey P, et al, on page 73.
E
sophageal adenocarcinoma has risen at an alarming rate in Western countries over the past few decades. This rise has not been met with any substantial improvement in treatment, making this cancer among the deadliest. It is likely that prevention offers the most credible strategy to defeat this cancer. As such, it is imperative to understand the basic pathogenesis of this condition and its precursors. The classic risk factors for this cancer include gastroesophageal reflux disease (GERD), smoking, obesity and a diet low in fresh fruit and vegetables.1 More recent work suggests that Helicobacter pylori infection, especially with more virulent cagA-positive strains, decreases the risk of esophageal adenocarcinoma.2 The mechanism of this protective effect is not clear but it has been postulated that it may be due to H pylori-induced gastric atrophy and hypochlorhydria, both of which reduce acid exposure of the lower esophagus.3 There is, therefore, an emerging consensus that pathologic changes at the gastroesophageal junction are determined to a large extent by the gastric pathophysiologic phenotype. There are 3 main gastric phenotypes that result from chronic H pylori infection: (1) the commonest by far is a mild pangastritis that does not affect gastric physiology and is not associated with significant human disease, (2) a corpus-predominant gastritis associated with multifocal gastric atrophy, hypochlorhydria, and increased risk of gastric cancer (the gastric cancer phenotype), and (3) an antral-predominant gastritis associated with high gastric acid secretion and increased risk of duodenal ulcer disease (the DU phenotype).4 The cause of these divergent clinical outcomes is most likely a combination of host genetic, bacterial, and environmental factors.5 Proinflammatory polymorphisms in the interleukin-1 beta gene (IL1-B) and tumor necrosis factor alpha gene (TNF-A) increase the risk of gastric atrophy, hypochlorhydria, and gastric cancer.6,7 IL-1 in particular is very relevant in this context because it is a very potent inhibitor of gastric acid secretion.8 Recent work by Waghray et al showed that IL-1 promotes gastric atrophy through suppression of sonic hedgehog9 and this contributes to the loss of the parietal cells capacity to secrete hydrochlo-
ric acid.10 TNF-␣ is also an acid inhibitor, albeit weaker than IL-1. It is tempting to speculate that the gastric cancer phenotype with its low acidity and association with acidreducing and atrophy-inducing proinflammatory IL-1 and TNF-␣ genotypes is protective against GERD and adenocarcinoma, but is there any evidence for this? At least 2 studies have addressed the role of these polymorphisms in reflux disease. Queiroz et al compared 98 GERD patients to 285 controls in Brazil and showed that the proinflammatory IL-1B-31 C/C and IL-1RN*2*2 genotypes were associated inversely with GERD in H pylori– positive individuals.11 In these subjects, cagA-positive status, IL-1B and IL-1RN proinflammatory genotypes, and the degree of corpus gastritis were negatively associated with GERD. They also demonstrated that the severity of the disease was inversely associated with the number of IL-1B and IL1-RN proinflammatory alleles. The second study by Ando et al examined 208 H pylori–positive and 112 H pylori–negative Japanese subjects and reported that infected subjects were less likely to report GERD symptoms and to have endoscopic evidence of erosive esophagitis.12 Furthermore, H pylori–positive subjects were more likely to have corpus gastric atrophy than H pylori– negative subjects. Among H pylori–positive patients, those without erosive oesophagitis or GERD symptoms were significantly more likely to have corpus atrophy than subjects with erosive esophagitis or GERD symptoms. Among H pylori–positive patients, subjects homozygous for the proinflammatory allele IL-1B-511T had a significantly lower risk of erosive oesophagitis and GERD symptoms compared with those homozygous for the -511C allele. The authors concluded that a proinflammatory IL-1B genotype was associated with increased risk of atrophy and decreased risk of GERD in H pylori– infected subjects in Japan.12 These data indicate that, in some genetically predisposed subjects, H pylori infection may protect against GERD through induction of gastric atrophy. Thus, these 2 studies clearly show that host genetic factors interact with H pylori infection leading to onset of gastric atrophy and protection against GERD. The crucial question that remains is whether this protection extends to esophageal and junctional adenocarcinomas. In this issue of GASTROENTEROLOGY, Whiteman et al attempted to answer this crucial question.13 They hypothesized that the association between H pylori infection and the risk of adenocarcinomas of the esophagus and GASTROENTEROLOGY 2010;139:17–30
Editorials continued
esophagogastric junction is modified by polymorphisms in the proinflammatory genes IL-1B and TNF-A that regulate gastric acid secretion. As mentioned, this hypothesis assumes that a proinflammatory genetic makeup that reduces acid secretion would protect H pylori–infected subjects from developing esophageal and junctional adenocarcinoma, cancers that most likely develop after decades of acid-induced damage to the lower esophagus. Thus, one would expect that such infected and “protected” subjects would have the so-called gastric cancer phenotype, which is characterized by hypochlorhydria and gastric atrophy, whereas subjects who develop these proximal cancers are either free of H pylori infection or, if infected, have a normal or high gastric acid capacity. The authors set out to answer this question with a sound and adequately powered epidemiologic design. They studied IL-1B and TNF-A polymorphisms but, unfortunately, not IL-1RN. The latter has consistently been shown to be an excellent genetic marker for IL-1 levels and function, and most of the genetic studies executed out in Caucasians found positive associations with proinflammatory conditions when other markers were negative. As it turned out, the authors found no evidence that polymorphisms in IL-1B or TNF-A modified the association between H pylori and esophageal or junctional adenocarcinomas. They concluded that H pylori infection is associated inversely with risks of these 2 cancers, but the reduction in risk was similar across subgroups of potential modifiers, including the host genes studied. We need to consider potential explanations for these negative findings. The simplest to exclude are genotyping errors that might have affected the results. All SNPs were in Hardy-Weinberg equilibrium, suggesting no major genotyping problems. The 1 slight anomaly is in the degree of linkage disequilibrium between the IL-1B-31 and -511 loci, particularly in the controls. In other published reports, the 2 loci are in near total linkage disequilibrium, whereas in this study there seems to be a slight discrepancy. This may reflect population admixture (highly unlikely in this nearly homogeneous Caucasian population), genotyping error, or a genuine difference. It is unlikely that this small discrepancy could explain the negative findings of this study. The second crucial factor to consider is the prevalence of H pylori infection, particularly in the control population. This clearly has to be high enough to allow a demonstration of the interaction with the host genes. The reason for this is that, in Western populations, H pylori infection is actually more likely to induce a mild pangastritis or an antral predominant gastritis that is either associated with normal or increased acid output and certainly with little or no gastric atrophy. Relatively few subjects develop the corpus-predominant gastric cancer phenotype.14 It is therefore difficult to demonstrate a 18
modifying effect of host genes that interact with an infection if the prevalence of this infection is very low. So, what was the prevalence of H pylori in this Australian population? As it turned out, it was quite low at 23%. This is much lower than most populations studied so far and certainly lower than the Brazilian and Japanese populations cited (62.9% and 65%, respectively11,12). Furthermore, the difference in H pylori prevalence between those with the cancer outcome versus those without was also small (13% vs 23%). It is clear, therefore, that despite an adequate design with good numbers for cancer subjects and controls, the low prevalence of the infection was a major obstacle. Serologic diagnosis of the infection is bound to misclassify some subjects but this would impact equally on all participants, especially in the absence of major gastric atrophy. The third explanation is that the gastric phenotype was not assessed in this study, and this clearly reduced the ability to offer a mechanistic insight into the associations between host genotype and clinical outcome. There were no gastric biopsies to map out the type of gastritis, its severity and the presence of atrophy. The study did not report on serologic surrogate markers of atrophy, such as pepsinogen I levels or pepsinogen I/II ratios. This would have given some insight into gastric pathophysiology and this might have informed the statistical analyses a little better. One has to acknowledge, however, that such a study where biopsies are obtained from different parts of the stomach in addition to the primary tumor would only be meaningful if the controls also had a similar mapping of gastric inflammation and atrophy. For such large-scale epidemiologic studies, this is quite a difficult prospect and is not feasible. Ironically, perhaps the best population to prove the author’s hypothesis would not be a Caucasian but rather an Asian one. In Asian countries such as Japan, Korea, and China, the rates of noncardiac gastric cancer are very high, whereas those for esophageal adenocarcinoma are very low. The default gastric phenotype that results from H pylori infection in the East is the gastric cancer phenotype with its low acid and gastric atrophy.14 More interestingly, the prevalence of proinflammatory IL-1B genotypes is significantly higher than in Caucasians. Could this be the reason why such proximal cancers are rare in the Far East? The work by Ando et al offers an insight into the potential answer and it is an open invitation to workers in the field to look into this question. This is indeed a very important issue; it lies at the heart of the pathogenesis of esophageal adenocarcinoma. The fast disappearance of H pylori globally (either spontaneously or owing to mass eradication) might be a logical strategy in the fight against distal gastric cancer, but we need to determine whether the return of normal to high gastric
Editorials continued
acid secretion in Eastern populations is likely to increase the risk of proximal cancers. Where do we go from here? The authors made an admirable attempt at answering an important question, but were perhaps thwarted by a fast disappearing H pylori infection. The major finding that a low infection rate increases the risk of proximal cancers confirms previous studies. The modifying influence of host genes on this risk was neither confirmed nor refuted. Two new studies are urgently needed. The first should look at the modifying effects of host genetics on risk of esophageal and junctional adenocarcinomas in Asia. This requires a large, multicenter, collaborative study pooling all resources across several Asian countries to collect adequate numbers of such a hitherto rare cancer. The results of such a study will allow workers in the field to predict and preempt a sudden rise in the incidence of such cancers in the future and will avoid the alarming increase seen in Western countries over the last 4 decades. The second study should include detailed phenotyping of the gastric environment in studies looking at esophageal cancers. The clue to these proximal cancers lies in the stomach. It might be difficult to stomach such studies, but they are essential if we are to make any significant progress!
MAIRI H. MCLEAN EMAD M. EL–OMAR Division of Applied Medicine Institute of Medical Sciences Aberdeen University Foresterhill, Aberdeen Scotland, United Kingdom References 1. Reid BJ, Li X, Galipeau PC, et al. Barrett’s oesophagus and oesophageal adenocarcinoma: time for a new synthesis. Nat Rev Cancer 2010;10:87–101. 2. Atherton JC, Blaser MJ. Coadaptation of Helicobacter pylori and humans: ancient history, modern implications. J Clin Invest 2009;119:2475–2487. 3. Blaser MJ. Disappearing microbiota: Helicobacter pylori protection against esophageal adenocarcinoma. Cancer Prev Res (Phila Pa) 2008;1:308 –311.
4. Snaith A, El-Omar EM. Helicobacter pylori: host genetics and disease outcomes. Expert Rev Gastroenterol Hepatol 2008;2: 577–585. 5. Amieva MR, El-Omar EM. Host-bacterial interactions in Helicobacter pylori infection. Gastroenterology 2008;134:306 –323. 6. El-Omar EM, Carrington M, Chow WH, et al. Interleukin-1 polymorphisms associated with increased risk of gastric cancer. Nature 2000;404:398 – 402. 7. El-Omar EM, Rabkin CS, Gammon MD, et al. Increased risk of noncardiac gastric cancer associated with proinflammatory cytokine gene polymorphisms. Gastroenterology 2003;124:1193– 1201. 8. El-Omar EM. The importance of interleukin 1beta in Helicobacter pylori associated disease. Gut 2001;48:743–747. 9. Waghray M, Zavros Y, Saqui-Salces M, et al. Interleukin-1beta promotes gastric atrophy through suppression of Sonic Hedgehog. Gastroenterology 2010;138:562–572. 10. van Dop WA, van den Brink GR. Sonic hedgehog: a link between inflammation, gastric atrophy, and acid suppression? Gastroenterology 2010;138:426 – 429. 11. Queiroz DM, Guerra JB, Rocha GA, et al. IL1B and IL1RN polymorphic genes and Helicobacter pylori cagA strains decrease the risk of reflux esophagitis. Gastroenterology 2004;127:73–79. 12. Ando T, El-Omar EM, Goto Y, et al. Interleukin 1B proinflammatory genotypes protect against gastro-oesophageal reflux disease through induction of corpus atrophy. Gut 2006;55:158 –164. 13. Whiteman DC, Parmar P, Fahey P, et al. Association of Helicobacter pylori infection with reduced risk for esophageal cancers is independent of environmental and genetic modifiers. Gastroenterology. 2010;139:73– 83. 14. Naylor GM, Gotoda T, Dixon M, et al. Why does Japan have a high incidence of gastric cancer? Comparison of gastritis between UK and Japanese patients. Gut 2006;55:1545–1552.
Reprint requests Address requests for reprints to: Professor Emad M El-Omar, Division of Applied Medicine, Institute of Medical Sciences, Aberdeen University, Foresterhill, Aberdeen AB25 2ZD, Scotland, UK. E-mail: [email protected]; Phone: ⴙ44-(0)1224-553021; Fax: ⴙ44-(0)1224-555766. Conflicts of interest The author declares no conflicts. © 2010 by the AGA Institute 0016-5085/$36.00 doi:10.1053/j.gastro.2010.05.022
19