Increase in Genogroup II.4 Norovirus Host Spectrum by CagA-Positive Helicobacter pylori Infection

Transcription

1 MAJOR ARTICLE Increase in Genogroup II.4 Norovirus Host Spectrum by CagA-Positive Helicobacter pylori Infection Nathalie Ruvoën-Clouet, 1,2 Ana Magalhaes, 4 Lara Marcos-Silva, 4,5 Adrien Breiman, 1,3 Ceu Figueiredo, 4,5 Leonor David, 4,5 and Jacques Le Pendu 1 1 Inserm, U892; CNRS, UMR 6299; Nantes University, Nantes 44007, France; 2 Oniris, Ecole Nationale vétérinaire, Agroalimentaire et de l alimentation, and 3 Centre Hospitalier Universitaire de Nantes, Nantes, France; 4 Institute of Molecular Pathology and Immunology and 5 Medical Faculty, University of Porto, Portugal Background. Noroviruses (NoVs) represent a considerable public health burden. Despite their enormous genetic diversity, most outbreaks are due to the single GII.4 genotype, but the reasons for this are poorly understood. NoVs use histo-blood group antigens (HBGAs) as attachment factors. Since HBGAs are present in saliva, binding of strains to saliva is commonly used as a surrogate for recognition of the gut surface by specific strains, although the relationship between saliva and gut tissue expression of HBGAs is not well defined. Methods. The presence of fucosylated HBGAs in saliva and stomach biopsy specimens, as well as that of genogroup I.1 and genogroup II.4 virus-like particles, were compared in a series of 109 donors from Portugal. Results. An overall good concordance between HBGA expression in saliva and stomach surface mucosa was observed. However, unexpected mucosal expression of α(1,2)fucosylated epitopes in nonsecretor individuals was frequently detected, allowing for GII.4 attachment. Although all individuals were infected with Helicobacter pylori, abnormal expression of α(1,2)fucosylated motifs and binding of GII.4 virus-like particles in nonsecretors mucosa were associated with positivity for the H. pylori CagA virulence factor. Conclusions. Infection by CagA-positive H. pylori induces expression of GII.4 attachment factors in nonsecretors mucosa, expanding the host range of these strains and thereby possibly contributing to their epidemiological dominance. Keywords. gastroenteritis; norovirus; Helicobacter pylori; CagA; host-spectrum; histo-blood group antigens; attachment factor; fucosyltransferases. Noroviruses (NoVs) are the most common cause of acute gastroenteritis, affecting all age groups. Although NoVs cause a self-limited disease that generally lasts 1 or 2 days, the very large number of cases for which they are responsible makes these viruses a considerable public health problem [1 3]. In addition, NoV gastroenteritis can be severe and life threatening in subjects with a weakened immune system, such as immunocompromised Received 12 December 2013; accepted 15 January 2014; electronically published 23 January Correspondence: Jacques Le Pendu, PhD, UMR Inserm 892/CNRS 6299, Institut de Recherche en Santé de l Université de Nantes, 8 Quai Moncousu, BP70721, 44007, Nantes, Cedex 1, France (jacques.le-pendu@univ-nantes.fr). The Journal of Infectious Diseases 2014;210: The Author Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please journals.permissions@oup.com. DOI: /infdis/jiu054 individuals, elderly individuals, and young children [4, 5]. NoVs are small, nonenveloped, positive-sense RNA viruses that present a large genetic diversity. They are classified among 5 genogroups, with strains infecting humans classified in genogroups I (GI), GII, and GIV. Genogroups are further subdivided into genotypes and genotypes into clusters. At least 8 GI and 17 GII genotypes, respectively, were recognized in 2006 for these 2 principal genogroups of human viruses [6], although this classification is now incomplete due to the increasing number of available sequences, including recombinants [7]. However, during the past 15 years, NoV epidemiology has been largely dominated by a single genotype [8 11]. Thus, analysis of archival feces samples suggested that, in the 1970s, GII.4 strains might have represented only about 15% of NoV strains in the United States, whereas they are now responsible CagA-Positive H. pylori Increase NoV Host Range JID 2014:210 (15 July) 183

2 for 70% 80% of outbreaks worldwide [5]. The reasons for the emergence of such an impressive epidemiological dominance have remained unclear, although 2 main potential nonexclusive mechanisms have been proposed [12]. Changes in catalytic properties of the RNA-dependent RNA polymerase of recent GII.4 strains, including increased activity and error rate, have been suggested to account for a higher replication rate accompanied by a higher ability to evolve immune escape mutants [9]. The second mechanism involves the interaction with attachment factors. GI and GII human strains use carbohydrates of the histo-blood group family (HBGA) as attachment factors [13]. It has been observed that most NoV strains recognize a limited number of these antigens and accordingly infect only subsets of the population, based on the presence of HBGA polymorphisms. The genetic polymorphism of these attachment factors may thus provide herd innate protection, which is akin to herd immunity because individuals not recognized by a given strain fail to transmit the virus, similar to immunized individuals [14, 15]. The large set of HBGA motifs recognized by GII.4 strains allows them to infect a larger spectrum of the human population, compared with other NoV strains, which may provide an advantage in terms of transmission. In addition, pandemic strains reportedly showed increased relative affinity for HBGAs, which might help facilitate escape from blocking antibodies [16]. Since HBGAs are expressed in saliva, binding of virus-like particles (VLPs) to saliva is often used as a surrogate assay to determine ability of a strain to recognize gut surface glycans [17, 18]. However, the relationship between HBGAs expression in saliva and the stomach or small intestine remains somewhat uncertain, because evidence is based on the study of a limited number of unrelated cases. The present study is primarily aimed at comparing the expression of HBGAs and the binding of VLPs from typical GI and GII strains in a large series of both saliva and stomach biopsy specimens from the same individuals. We observed a number of discrepancies between saliva and tissue HBGA expression with relevance for GII.4 attachment. Unexpectedly, we further observed that these discrepancies were associated with infection with CagA-positive Helicobacter pylori strains. METHODS Donors and Samples All samples were obtained following receipt of informed consent from workers at a shipyard in northern Portugal who participated in a gastric pathology survey during 1998 [19]. Stomach biopsy specimens from the antrum, incisura and corpus, blood samples, and saliva samples were obtained from 109 volunteers. The study accorded with the Declaration of Helsinski and received approval from the Comissão de etica do Hospital de S. João (approval no. 47/2003). Biopsy specimens and blood samples were collected in 1998 and saliva was obtained later, in 2004, during a so-called second-look follow-up survey [20]. Reagents Biotin-labeled and peroxidase-labeled UEA-I (Ulex europaeus agglutinin I), specific forhtype2/le y, were obtained from Sigma. Monoclonal antibodies anti-a (clone 3 3A), anti-le a (clone 7-LE), and anti-le b (clone 2 25LE) were obtained from Dr J Bara (CNRS, Villejuif, France). The anti-b (clone ED3) and anti-le a (clone CA3F4) antibodies were obtained from Dr A. Martin (EFS Rennes, France) and H. Clausen (Copenhagen University, Denmark), respectively. The anti-sle x KM93 was from Calbiochem, and the anti-le b BG6 was from Signet Pathology Systems. The secondary peroxidase-labeled anti-mouse or anti-rabbit immunoglobulins were from Vector Laboratories. For staining with the anti-sle x, biotin-labeled anti-mouse immunoglobulins, followed by peroxidase-labeled streptavidin, were used (Dako). Neuraminidase α(2 3) > α(2 6) = α(2 8) from Clostridium perfringens was purchased from Roche Applied Science. The α(1,3/4)-l-fucosidase from Streptomyces species was purchased from Takara Shuzo. The α(1,2)-l-fucosidase from Bifidobacterium bifidum was a kind gift from Dr Takane Katayama (Ishikawa Prefectural University, Japan) [21, 22]. Characterization of the ABO, Secretor, and Lewis Status ABO and Lewis phenotypes were first obtained by standard hemagglutination using specific reagents. Serologic Lewis phenotyping allows determination of the secretor character for Lewis-positive individuals since Le(a b+) individuals have a secretor phenotype and Le(a+b ) individuals are nonsecretors. However, the secretor character remains undefined for Le(a b ) individuals. Saliva samples were used to confirm the ABO, secretor, and Lewis phenotypes and to determine the secretor phenotype of the Lewis-negative individuals, using a set of specific anti-a, anti-b, anti-h, and Lewis reagents as previously described [17]. In brief, boiled and clarified saliva samples were coated onto enzyme-linked immunosorbent assay (ELISA) plates at a 1:1000 dilution and incubated with biotinylated UEA-I lectin, anti-a (clone 3 3A), anti-b (clone ED3), anti- Le a (clone 7-LE), or anti-le b (clone 2 25LE) antibodies at the appropriate dilutions. Binding of UEA-I lectin was detected following a second incubation with peroxidase-conjugated streptavidin. Binding of the primary antibodies was detected following incubation with peroxidase-conjugated anti-mouse immunoglobulins. Reactions were developed with p-nitrophenyl phosphate (Sigma), and the OD was read at 405 nm. The secretor phenotype was confirmed by FUT2 genotyping involving blood samples, as previously described [17, 23]. Assay of VLP Attachment to Saliva VLPs from the GI.1 NV strain and the GII.4 Dijon 1996 strain were produced using a baculovirus expression system and 184 JID 2014:210 (15 July) Ruvoën-Clouet et al

3 purified by rate-zonal centrifugation in a sucrose gradient, as described elsewhere [24]. VLPs were incubated on saliva samples coated on ELISA plates as described above. Binding of VLPs was detected using polyclonal rabbit anti-gi.1 and anti-gii.4, respectively, followed by incubation with a peroxidase-labeled anti-rabbit immunoglobulins, as previously described [16, 25]. Immunohistochemical Analysis Stomach biopsy specimens were formalin fixed and paraffin embedded. Histologic sections (4 μm) were deparaffinized and rehydrated. Sections were then incubated with peroxidase-labeled UEA-I lectin at 4 μg/ml, anti-le a (CA3F4) at 10 μg/ml, anti- Le b (BG6) at 10 μg/ml, anti-sle x (KM93) at 10 μg/ml, or VLPs from GI.1 or GII.4 strains at 10 μg/ml each. In the case of primary monoclonal antibodies, binding was detected using secondary peroxidase-labeled anti-mouse immunoglobulin G. VLP binding was detected using anti-gi.1 or anti-gii.4 rabbit antisera, followed by incubation with peroxidase-labeled anti-rabbit immunoglobulin G, and reactions were revealed as previously described. To characterize GII.4 ligand specificity in the case of selected nonsecretor individuals, before incubation with VLP, tissue sections were treated with glycosidases. Screening of the appropriate glycosidases and determination of optimal treatment conditions were performed by ELISA, using saliva from secretor individuals or neoglycoconjugates and then using tissue sections from dog breast carcinoma that express Lewis antigens. Briefly, α(1,2)fucosidase and α(1,3)/α (1,4)fucosidase were first incubated at 10 μg/section and 1 μu/ section, respectively, in 100 mm sodium phosphate ( ph 6.5) for 6 hours at 37 C, followed by a second incubation with fresh enzyme overnight at 37 C. The neuraminidase was used at 0.05 U/section in 50 mm sodium acetate (ph 5.0) in the same conditions as the fucosidases. Control untreated sections were incubated in parallel in the same buffer without glycosidase. Determination of H. pylori Density and CagA Status H. pylori density was semiquantitatively scored according to the modified Sydney system that was also used to score glandular atrophy and intestinal metaplasia [26]. Also, H. pylori caga and vaca genotypes were directly determined in the gastric biopsy specimen from the greater curvature of the antrum by multiplex polymerase chain reaction analysis and reverse hybridization, as previously described [27]. RESULTS Relationship Between HBGAs Expression in Saliva and the Stomach Surface Mucosa A perfect match between secretor saliva phenotypes and FUT2 genotypes was observed for the 109 donors. Eighty-two secretors and 27 nonsecretors (24.7%) were found in the study population. There were 94 Lewis-positive individuals and 15 Lewis negative (13.8%). Among the 94 Lewis positive individuals we observed 2 discrepancies between the serologic phenotypes obtained by classical hemagglutination and by ELISA on saliva. One Le(a+b ) individual was a secretor (FUT2+/ ), and 1 Le (a b+) case was a nonsecretor (FUT2 / ), indicating that the serologic Lewis phenotyping is error prone, as it generated misclassification of the secretor character. Binding of UEA-I that recognizes α(1,2)-linked fucose residues clearly discriminated secretors and nonsecretors at the saliva level (Figure 1A). Since α(1,2)-linked fucose residues are essential for attachment of most NoV strains, we compared saliva expression to the stomach surface mucosa expression of all cases. The surface mucosa of all secretors was strongly or moderately stained by UEA-I, in accordance with the saliva positivity. However, UEA-I binding was also clearly observed in 10 of 27 nonsecretors, indicating that saliva HBGA expression does not always match stomach surface mucosa expression (Figure 1A and Figure 2). Relationship Between VLP Attachment to Saliva Mucins and to the Stomach Surface Mucosa We next compared attachment of VLPs to saliva and gastric surface mucosa. Both GI.1 and GII.4 VLPs could distinguish between secretors and nonsecretors at the saliva level. As previously described [28], saliva samples from B blood group secretors were only weakly recognized by GI.1 VLPs, whereas GII.4 VLPs showed a moderate binding to nonsecretor saliva samples, provided these were Lewis positive. Only the Lewisnegative nonsecretor individuals showed no GII.4 VLP binding to saliva, indicating that the moderate binding to nonsecretor saliva was dependent on the presence of a functional FUT3 allele. At the tissue level, positive staining was obtained with both GI.1 and GII.4 VLPs for all but 1 tested secretor cases. The single exception was a B blood group case not stained by GI.1 VLP (Figure 1B). Among nonsecretors, only 1 of 23 specimens was stained with GI.1 VLP, but 6 of 23 were stained with GII.4 VLP. In the latter case, there was no relationship between the level of saliva binding and the presence of tissue staining. Since various NoV strains recognize Lewis-type epitopes, we examined the relationship between the presence of Le a and Le b antigens in saliva and at the stomach surface. As shown in Figure 3, Lewis-positive individuals express variable levels of either Le a or Le b antigens in saliva, and there is a clear relationship between expression of these 2 antigens. Secretor individuals express preferentially Le b, but can also express significant levels of Le a in saliva. In nonsecretors, Le a is present but very little Le b can be detected. Since the anti-le b that we used is known to show some cross-reactivity with Le a, the limited Le b reactivity in saliva of nonsecretors may actually be due to that cross-reactivity [29]. At the tissue level, most secretors expressed only Le b (Lea b+). However, both Le a and Le b were detected in 9 of 80 secretors tested. Consistent with their secretor CagA-Positive H. pylori Increase NoV Host Range JID 2014:210 (15 July) 185

4 as defined by serologic analysis and saliva phenotyping, surprisingly expressed Le b at the tissue level. Concerning the nonsecretors, Le a expression was found as expected in all Lewis-positive cases in the stomach surface mucosa. However, 19 of 27 cases were also stained by anti-le b.bothgi.1and GII.4 VLPs can attach to Le b. Yet, the GI.1 VLP did not bind to any of these 19 cases, and there was no relationship between staining by anti-le b and the GII.4 VLP, suggesting that anti-le b staining was mainly due to cross-reactivity with the Le a epitope. Figure 1. Binding of the α(1,2)fucose-specificlectinulex europaeus agglutinin I (UEA-I) (A), virus-like particles (VLPs) from the NV genogroup I.1 (GI.1) strain (B), andvlpfromthedijon96gii4strain(c) to saliva samples and the surface stomach mucosa of secretors and nonsecretors. Binding to saliva was performed by enzyme-linked immunosorbent assay, and ODs are shown for each individuals, ranked by signal intensity. Tissue binding was semiquantitatively evaluated as strongly positive (++), positive (+), and negative ( ). The numbers of individuals in each subgroup are indicated for the stomach surface staining. The ABO phenotypes of secretors are indicated in light grey boxes below the graphs (B and C). In panel C, the Lewis-negative individuals among nonsecretors are shown by a dark grey box below the graph. Secretors have the 428G/G or 428G/A genotype, whereas all nonsecretors have the G428A/A genotype. A dashed line separates the 2 subgroups. status, none expressed Le a only (Lea+b ), and consistent with their Lewis status neither Le a nor Le b were detected in 11 of the 12 Lewis-negative individuals. Only 1 Lewis-negative individual, Unexpected Relationship Between Infection by CagA-Positive H. pylori and Abnormal HBGAs and GII.4 VLP Binding to the Surface Mucosa of Nonsecretors It has been shown that GII.4 strains have acquired the ability to recognize the sialyl-lewis x motif (SLe x )[16, 30]. The significance of this observation has remained obscure because SLe x is considered a tumor-associated antigen that is not expressed, or only weakly expressed, on the healthy gut epithelium [31]. Interestingly, it was previously reported that SLe x expression could be induced by inflammation and H. pylori infection [32, 33]. Since all cases of our study were infected with H. pylori, we examined SLe x expression at the stomach surface of a subset of cases to examine its possible contribution to GII.4 strains binding to the tissue of nonsecretors. As shown in Figures 2B and 4A, SLe x expression ranged from strong to negative among secretors, without relationship to H. pylori density (data not shown), whereas all nonsecretors strongly expressed SLe x at the stomach mucosa surface. This is likely due to the absence of competition in nonsecretors between the FUT2 enzyme and the sialyltransferase involved SLe x synthesis. Nevertheless, we observed a clear association between the unexpected UEA-I staining of nonsecretor cases at the tissue level and the binding of GII.4 to the stomach surface mucosa (Figure 4B). UEA-I tissue reactivity or GII.4 binding in nonsecretors were not related to H. pylori density, but a relationship was found with the presence of the H. pylori virulence factor CagA. CagA positivity was detected in 55 of 105 tested individuals (52%) in our series. Yet, all nonsecretors who were UEA-I positive and GII.4 binders at the tissue level were infected by CagApositive strains (Figure 4C and 4D). Since the presence of the CagA toxin is associated with H. pylori virulence, we looked for possible relationships between UEA-I or GII.4 VLP staining and H. pylori related pathological signs. We observed strong associations between UEA-I positivity among nonsecretors and the presence of both gastric glandular atrophy and intestinal metaplasia (Figure 4E and 4F). Likewise, there were significant associations between the binding of GII.4 VLP to the surface mucosa of nonsecretors and either villus atrophy or the presence of intestinal metaplasia (Figure 4G and 4H). Glycosidase treatment of the tissue sections was then performed to ascertain the specificity of GII.4 and UEA-I binding to the stomach surface mucosa of nonsecretors infected with 186 JID 2014:210 (15 July) Ruvoën-Clouet et al

5 Figure 2. Examples of typical expression of histo-blood group antigen motifs and virus-like particle (VLP) attachment in the stomach surface mucosa. A, Expression of α(1,2)fucosylated structures stained by Ulex europaeus agglutinin I (UEA-I), expression of Lewis a and Lewis b antigens, and binding of genogroup I.1 (GI.I) and GII.4 VLPs to the mucosa of an O phenotype, secretor, Lewis-negative individual (O, Sec, Le ); an A phenotype, secretor, Lewis-positive individual (A, Sec, Le+); and a nonsecretor, Lewis-positive individual (nonsec, Le+). B, Expression of the sialyl-lewis x epitope in the mucosa of secretors presenting with negative staining (A), moderate staining (B), and strong staining (C) and of a nonsecretor with strong staining (D). The staining to some mucus at the lower right corner in panel A is probably nonspecific. CagA-positive H. pylori (Table 1 and Figure 5). α(1,2)fucosidase treatment almost completely removed both UEA-I and GII.4 staining in 3 of the 4 cases assayed. In the fourth case, only a slight reduction in staining was obtained. Interestingly, that case (V394) presented with intestinal metaplasia. Similarly, treatment with either α(1,3/4)fucosidase or sialidase led to a near Figure 3. Relationship between the presence of Lewis a and Lewis b antigens in saliva and at the stomach surface. ODs obtained in the saliva assay using both anti-lewis a and anti-lewis b antibodies are plotted for each individual, and the results of the regression analysis are shown for both secretors (left) and nonsecretors (right). The dashed lines indicate the threshold for positivity (>3 times background). The corresponding stomach surface staining results are shown below the graphs, with the numbers of individuals in each subtype. CagA-Positive H. pylori Increase NoV Host Range JID 2014:210 (15 July) 187

6 complete loss of GII.4 binding to the same 3 cases but hardly affected binding to case V394. A combined α(1,2)fucosidase and α (1,3)/α(1,4)fucosidase was required to completely remove GII.4 binding to V394 tissue. In that case, all α(1,2)fucosylated structures may have been α(1,3/4)-fucosylated too. Since the presence of the α(1,3/4)-linked fucoses prevents action of the α(1,2)fucosidase, their prior removal was required to subsequently remove the α(1,2)-linked fucose that served as a ligand to the VLPs. When limited amounts of α(1,2)fucosylated structures are available, GII.4 binding may be attained by a combination with SLe x binding. Yet, when larger amounts of difucosylated structures (Le b or Le y ) are present, the contribution of SLe x is no longer required, and GII.4 VLP binding may become resistant to sialidase treatment. Regardless, these results indicate that a combination of α(1,2)fucosylated, α(1,3)/α(1,4)fucosylated, and sialylated motifs are involved in the binding of GII.4 VLPs to the surface mucosa of nonsecretors infected with CagA-positive H. pylori strains. DISCUSSION Overall, a rather good concordance was found between expression of HBGAs in saliva and the stomach surface mucosa. This concordance was reflected in the binding of either GI.1 or GII.4 VLPs. Of note, this was not true in the glandular compartment, where HBGAs are expressed regardless of the secretor character and where Le a and Le b antigens are absent. Accordingly, in the glands, staining was observed with both types of VLPs in all individuals (not shown). Since GI.1 and GII.4 infect secretors either exclusively or preferentially [34 36], it is clear that HBGA expression in the glandular compartment is irrelevant for infection. It is not clear whether the stomach can be a point of entry of NoVs, but HBGA expression in the stomach and the duodenum are similar [37, 38]. Since small intestinal biopsy specimens are difficult to obtain, we concentrated our attention on the stomach mucosa. Figure 4. Relationships between Helicobacter pylori associated glycans of the surface stomach mucosa and genogroup II.4 (GII.4) virus-like particle attachment. Tissue labeling was semiquantitatively evaluated as strongly positive (++), positive (+), and negative ( ). A, Expression of sialyl-lewis x (SLe x ) in secretors and nonsecretors. B, Relationship between the presence of α(1,2)-linked fucose residues detected by Ulex europaeus agglutinin I (UEA-I) and GII.4 VLP binding among nonsecretors. C, Relationship between UEA-I staining and infection by CagA-expressing Helicobacter pylori strains among nonsecretors. D, Relationship between GII.4 VLP binding and infection by CagA-expressing H. pylori strains among nonsecretors. E, Relationship between glandular atrophy and UEA-I staining among nonsecretors. F, Relationship between intestinal metaplasia (IM) and UEA-I staining among nonsecretors. G, Relationship glandular atrophy and GII.4 VLP binding among nonsecretors. H, Relationship between IM and GII.4 VLP binding among nonsecretors. 188 JID 2014:210 (15 July) Ruvoën-Clouet et al

7 Table 1. Effect of Glycosidase Treatment on Genogroup II.4 (GII.4) and Ulex europaeus agglutinin I (UEA-I) Binding to Stomach Tissue Sections of 4 Nonsecretors Selected for Abnormal Expression of UEA-I Donor Control a α(2)f α(3/4)f α(2)f and α(3/4)f Sialidase Sialidase and α(3/4)f GII.4 VLPs V / V / +/ +/ V / V / ND UEA-I V / ++ ND ND ++ V / ++ ND ++ V / ++ ND ND V ND ND ND ND ND Data denote findings of semiquantitative analysis of tissue labeling. Because of the limited number of available tissue sections from biopsy specimens, not all enzyme treatments could be performed for each individual. Abbreviations: ND, not done; VLP, virus-like particle; α(2)f, α(1,2)fucosidase; α(3/4)f, α(1,3)/α(1,4)fucosidase;, negative; +/, weakly positive; +, positive; ++, strongly positive. a No enzyme treatment. UEA-I recognizes exclusively α(1,2)-linked fucose residues, representing the ideal tool for secretor phenotyping in saliva. Nevertheless, at the surface mucosa level, unexpected UEA-I staining was observed in 10 of 27 nonsecretors, even though all nonsecretors carried 2 G428A FUT2 null alleles. The presence of such α(1,2)-fucosylated motifs on the surface mucosa in the absence of a functional FUT2 allele was associated with infection by CagA-positive H. pylori. Since humans possess only 2 enzymes able to catalyze the transfer of a fucose residue in α(1,2) linkage [39], in these individuals, FUT1, the other enzyme, is most likely responsible for the synthesis of the UEA-I binding sites in CagApositive H. pylori infected FUT2-deficient individuals, although this remains to be formally demonstrated. GI.1 VLPs attached to the surface mucosa of a single nonsecretor individual, which turned out to be infected by CagA-positive H. pylori and UEA- I positive. GII.4 VLPs attached to the mucosa of 6 of 23 nonsecretors. This could be related to weak binding to the saliva of nonsecretor Lewis-positive individuals, in accordance with the ability of GII.4 to recognize the α(3/4)-linked fucose in the socalled Lewis binding mode. Nevertheless, at the tissue level no association was found with the expression of Lewis antigens. Instead an association was found with the UEA-I positivity linked to CagA-positive H. pylori infection, indicating that the mere presence of Lewis antigens was not sufficient to support GII.4 VLP attachment. Yet, treatment of tissue sections with either an α(1,2)fucosidase, an α(1,3/4)fucosidase, or both inhibited GII.4 VLP binding, showing that both the α(1,2)fucosylated antigens and Lewis antigens were involved in GII.4 attachment to the stomach surface mucosa of nonsecretors. Approximately half of adults worldwide are infected with H. pylori, although the incidence greatly varies across populations. In most developed countries, the prevalence has significantly decreased over the past several decades. It now ranges from 25% to 50% [40, 41]. Yet, it is still high, often >90%, in many countries [40]. In the early 1990s, when this Portuguese study population was sampled, the incidence of H. pylori infection was around 90%, and all individuals in the present study were infected [42]. CagA is a cytotoxin delivered into the cytoplasm of target cells through the type IV secretion system. It interacts with various proteins involved in cell signaling, thereby causing cellular dysfunctions. Its presence is strongly associated with the development of glandular atrophy, intestinal metaplasia, severe inflammation, and gastric cancer [43]. Over 90% of H. pylori strains isolated in Eastern Asia express CagA, whereas large fractions of strains isolated in Europe and the USA are CagA negative [43, 44]. In our cohort, only 52% of cases were CagA positive, and a recent analysis indicated that as few as 19% of Dutch women infected with H. pylori carried CagA-positive strains [41]. We observed marked associations between the abnormal UEA-I positivity in the surface mucosa from nonsecretors and glandular atrophy, intestinal metaplasia, as well as CagA seropositivity. This strongly suggests that CagA-positive H. pylori strains induce expression of FUT1 in the surface gastric epithelial cells, generating binding sites for GII.4 strains, even in the absence of a functional FUT2 allele. Because infection with CagA-positive H. pylori is strongly associated with the presence of intestinal metaplasia, intestinal metaplasia may serve as a point of entry for the virus. Individuals who are not susceptible to infection with a given NoV strain are similar to immunized individuals since they do not support transmission of the virus. The polymorphism of CagA-Positive H. pylori Increase NoV Host Range JID 2014:210 (15 July) 189

8 Figure 5. Effect of glycosidase treatments on genogroup II.4 (GII.4; A) and Ulex europaeus agglutinin I (UEA-I; B) binding to stomach (cases V128 and V422) or intestinal metaplasia (case V394) tissue sections of nonsecretors who express binding sites on their surface mucosa. Brown staining at the surface of epithelial cells indicates binding, whereas the blue hematoxylin counterstaining shows histologic features. Staining in the absence of enzyme treatment (control), following α(1,2)fucosidase treatment (α2f), α(1,3)/α(1,4)fucosidase treatment (α3/4f) for case V128, combined α2f and α3/4f treatments for cases V422 and V394, or sialidase treatment (Sia). HBGAs therefore restricts the circulation of NoVs in a manner akin to herd immunity. For this reason, we have called this phenomenon herd innate protection [15]. GII.4 strains that have the ability to recognize the broadest range of HBGAs are at an advantage in this respect because it gives them a high host range, providing an increased ability to circulate as compared to other strains. We have seen here that the host range of GII.4 strains can be further expanded owing to another common gut pathogen responsible for chronic infections. In populations where most individuals, with the exception of young children, are infected with H. pylori strains expressing the CagA virulence factor, restriction of transmission of GII.4 strains by the genetic polymorphisms of FUT2 and, possibly, FUT3 can be bypassed. Therefore, our observations suggest that CagApositive H. pylori infection contributed to generate and maintain the epidemiological dominance of GII.4 strains once they acquired the ability to bind to a large array of HBGA motifs, including A, B, H, Lewis y, Lewis b, and sialyl-lewis x. Reducing the incidence of H. pylori infection, particularly in countries where it is high and mainly involves CagA-positive strains, could therefore indirectly contribute to a significant reduction in the incidence of NoV infection, as well. Notes Acknowledgments. We thank Dr Takane Katayama (Graduate School of Biostudies, Kyoto University), for his generous gift of α(1,2)fucosidase; and the Cellular and Tissular Imaging core facility of the Nantes University (MicroPiCell). Financial support. This work was supported by the ARMINA project from the Région des Pays de la Loire, France, and by Project NORTE (grant FEDER ) cofinanced by Programa Operacional Regional do Norte, under Quadro de Referência Estratégico Nacional and Fundo Europeu de Desenvolvimento Regional. Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. References 1. Glass RI, Parashar UD, Estes MK. Norovirus gastroenteritis. N Engl J Med 2009; 361: Patel MM, Hall AJ, Vinje J, Parashar UD. Noroviruses: a comprehensive review. J Clin Virol 2009; 44: Hall A, Lopman B, Payne DC, et al. Norovirus Disease in the United States. Emerg Infect Dis 2013; 19: Bok K, Green KY. Norovirus gastroenteritis in immunocompromised patients. N Engl J Med 2012; 367: Boon D, Mahar JE, Abente EJ, et al. Comparative evolution of GII.3 and GII.4 norovirus over a 31-year period. J Virol 2011; 85: JID 2014:210 (15 July) Ruvoën-Clouet et al

9 6. Zheng DP, Ando T, Fankhauser RL, Beard RS, Glass RI, Monroe SS. Norovirus classification and proposed strain nomenclature. Virology 2006; 346: Kroneman A, Vega E, Vennema H, et al. Proposal for a unified norovirus nomenclature and genotyping. Arch Virol 2013; 158: Bok K, Abente EJ, Realpe-Quintero M, et al. Evolutionary dynamics of GII.4 noroviruses over a 34-year period. J Virol 2009; 83: Bull RA, Eden JS, Rawlinson WD, White PA. Rapid evolution of pandemic noroviruses of the GII.4 lineage. PLoS Pathog 2010; 6:e Siebenga JJ, Vennema H, Renckens B, et al. Epochal evolution of GGII.4 norovirus capsid proteins from 1995 to J Virol 2007; 81: Siebenga JJ, Vennema H, Zheng DP, et al. Norovirus illness is a global problem: emergence and spread of norovirus GII.4 variants, J Infect Dis 2009; 200: Bull RA, White PA. Mechanisms of GII.4 norovirus evolution. Trends Microbiol 2011; 19: Tan M, Jiang X. Norovirus-host interaction: Multi-selections by human histo-blood group antigens. Trends Microbiol 2011; 19: Lindesmith L, Donaldson EF, Lobue AD, et al. Mechanisms of GII.4 norovirus persistence in human populations. PLoS Med 2008; 5:e Ruvöen-Clouet N, Belliot G, Le Pendu J. Noroviruses and histo-blood groups: the impact of common host genetic polymorphisms on virus transmission and evolution. Rev Med Virol 2013; 23: de Rougemont A, Ruvoën-Clouet N, Simon B, et al. Qualitative and quantitative analysis of the binding of GII.4 norovirus variants onto human blood group antigens. J Virol 2011; 85: Marionneau S, Airaud F, Bovin NV, Le Pendu J, Ruvoën-Clouet N. Influence of the combined ABO, FUT2 and FUT3 polymorphism on susceptibility to Norwalk virus attachment. J Infect Dis 2005; 192: Huang P, Farkas T, Zhong W, et al. Norovirus and histo-blood group antigens: demonstration of a wide spectrum of strain specificities and classification of two major binding groups among multiple binding patterns. J Virol 2005; 79: Peleteiro B, Lunet N, Figueiredo C, Carneiro F, David L, Barros H. Smoking, Helicobacter pylori virulence, and type of intestinal metaplasia in Portuguese males. Cancer Epidemiol Biomarkers Prev 2007; 16: Barros R, Peleteiro B, Almeida R, et al. Relevance of high virulence Helicobacter pylori strains and futility of CDX2 expression for predicting intestinal metaplasia after eradication of infection. Scand J Gastroenterol 2010; 45: Nagae M, Tsuchiya A, Katayama T, Yamamoto K, Wakatsuki S, Kato R. Structural basis of the catalytic reaction mechanism of novel 1,2-α-lfucosidase from Bifidobacterium bifidum. J Biol Chem 2007; 282: Katayama T, Sakuma A, Kimura T, et al. Molecular cloning and characterization of Bifidobacterium bifidum 1,2-α-L-Fucosidase (AfcA), a novel inverting glycosidase(glycoside hydrolase family 95). J Bacteriol 2004; 186: Serpa J, Mendes N, Reis CA, et al. Two new FUT2 (fucosyltransferase 2 gene) missense polymorphisms, 739G>A and 839T>C, are partly responsible for non-secretor status in a Caucasian population from Northern Portugal. Biochem J 2004; 383: Jiang X, Wang M, Graham DY, Estes MK. Expresion, self-assembly, and antigenicity of the Norwalk virus capsid protein. J Virol 1992; 66: Marionneau S, Ruvöen-Clouet N, Le Moullac-Vaidye B, et al. Norwalk virus binds to histo-blood group antigens on gastro-duodenal epithelial cells of secretor individuals. Gastroenterology 2002; 122: Dixon MF, Genta RM, Yardley JH, Correa P, the participants in the International Workshop of the Histopathology of Gastritis H, Classification and grading of gastritis. The updated Sydney system. Am J Surg Pathol 1996; 20: van Doorn LJ, Figueiredo C, Rossau R, et al. Typing of Helicobacter pylori vaca gene and detection of caga gene by PCR and reverse hybridization. J Clin Microbiol 1998; 36: Huang PW, Farkas T, Marionneau S, et al. Norwalk-like viruses bind to ABO, Lewis and secretor histo-blood group antigens but different strains bind to distinct antigens. J Infect Dis 2003; 188: Bara J, Gautier R, Le Pendu J, Oriol R. Immunochemical characterization of mucins. Polypeptide (M1) and polysaccharide (A and Le b ) antigens. Biochem J 1988; 254: Rydell GE, Nilsson J, Rodriguez-Diaz J, et al. Human noroviruses recognize sialyl-lewis x neoglycoprotein. Glycobiology 2009; 19: Le Pendu J, Marionneau S, Cailleau-Thomas A, Rocher J, Le Moullac- Vaydie B, Clément M. ABH and Lewis histo-blood group antigens in cancer. APMIS 2001; 109: Mahdavi J, Sonden B, Hurtig M, et al. Helicobacter pylori SabA adhesin in persistent infection and chronic inflammation. Science 2002; 297: Marcos NT, Magalhaes A, Ferreira B, et al. Helicobacter pylori induces beta3gnt5 in human gastric cell lines, modulating expression of the SabA ligand sialyl-lewis x. J Clin Invest 2008; 118: Hutson AM, Airaud F, Le Pendu J, Estes MK, Atmar RL. Norwalk virus infection associates with secretor status genotyped from sera. J Med Virol 2005; 77: Lindesmith L, Moe CL, Marionneau S, et al. Human susceptibility and resistance to Norwalk virus infection. Nat Med 2003; 9: Frenck R, Bernstein DI, Xia M, et al. Predicting susceptibility to norovirus GII.4 by use of a challenge model involving humans. J Infect Dis 2012; 206: Mollicone R, Bara J, Le Pendu J, Oriol R. Immunohistologic pattern of Type 1 (Le a,le b ) and type 2 (X, Y, H) blood group-related antigens in the human pyloric and duodenal mucosae. Lab invest 1985; 53: Le Pendu J, Le Cabellec M, Bara J. Immunohistological analysis of antibodies against ABH and other glycoconjugates in normal human pyloric and duodenal mucosae. Transfus Clin Biol 1997; 1: Oriol R, Candelier JJ, Mollicone R. Molecular genetics of H. Vox Sang 2000; 78: Parkin DM. The global health burden of infection-associated cancers in the year Int J Cancer 2006; 118: den Hollander WJ, Holster IL, den Hoed CM, et al. Ethnicity is a strong predictor for Helicobacter pylori infection in young women in a multiethnic European city. J Gastroenterol Hepatol 2013; 28: Bastos J, Peleteiro B, Barros R, et al. Sociodemographic determinants of prevalence and incidence of Helicobacter pylori infection in Portuguese adults. Helicobacter 2013; 18: Hatakeyama M. Helicobacter pylori CagA a bacterial intruder conspiring gastric carcinogenesis. Int J Cancer 2006; 119: Suzuki R, Shiota S, Yamaoka Y. Molecular epidemiology, population genetics, and pathogenic role of Helicobacter pylori. Infect Genet Evol 2012; 12: CagA-Positive H. pylori Increase NoV Host Range JID 2014:210 (15 July) 191