http://orcid.org/0000-0001-7709-658X
Figures
Abstract
Many susceptibility genes of inflammatory bowel disease (IBD) are associated with ankylosing spondylitis (AS). Fucosyltransferase 2 (FUT2) and FUT3 genes are related to IBD. This study aimed to investigate whether these genes correlated with the susceptibility to AS. Questionnaires of 673 patients with AS, and peripheral blood specimens of the patients and 687 healthy controls were collected. FUT2 and FUT3 genes were genotyped using the SNPscan method. Frequency differences of the genes at different levels, haplotypes, and interactions were analyzed. No frequency differences were found between the cases and the controls in all the genotypes and the alleles of rs1047781, rs1800028, rs1800030, and rs812936. For rs28362459, a significant difference in allele frequencies was observed in the total participants between the groups [χ2 = 7.515, Pcorrected = 0.030; adjusted odds ratio (ORadjusted)G/T = 0.782; 95% confidence interval (CI), 0.650–0.941]. The frequencies of haplotypes TT (rs812936–rs28362459) (χ2 = 5.663, Ppermutation = 0.039) and TG (rs812936–rs28362459) (χ2 = 7.456, Ppermutation = 0.013) in the total participants, and TG (rs812936–rs28362459) in the female subgroup (χ2 = 5.624, Ppermutation = 0.047) showed significant differences between the cases and the controls. No frequency differences at the phenotypic level were found. Two-factor interactions were observed between rs28362459-TG and age, rs28362459-TG and sex, rs28362459 and rs1047781, and the Lewis and secretor status. Rs28362459-G was related to some aggravated symptoms of AS (all Pcorrected < 0.05). These findings indicated that FUT3 polymorphisms were associated with human predisposition to AS at the allele and haplotype level. Rs28362459-G might decrease the susceptibility to AS, but aggravate relevant symptoms.
Citation: Jiang G, Han R, Chen M, Liu R, Wu M, Zhang X, et al. (2020) Associations between fucosyltransferase 3 gene polymorphisms and ankylosing spondylitis: A case–control study of an east Chinese population. PLoS ONE 15(8): e0237219. https://doi.org/10.1371/journal.pone.0237219
Editor: Narasimha Reddy Parine, King Saud University, SAUDI ARABIA
Received: January 4, 2020; Accepted: July 22, 2020; Published: August 7, 2020
Copyright: © 2020 Jiang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files. Protocols are available from https://dx.doi.org/10.17504/protocols.io.bhnjj5cn.
Funding: This study was funded by the National Natural Science Foundation of China (http://www.nsfc.gov.cn/; grant numbers: 81273169, 81573218, and 81773514; recipient: FP), the Scientific Research Foundation of Translational Medicine of Anhui Province (grant number: 2017zhyx03; recipient: FP), and the Academic and Technical Leaders Foundation of Anhui Province (grant number: 2017D140; recipient: FP). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Ankylosing spondylitis (AS) is a serum-negative connective tissue disease characterized by back pain and rigidity. Its etiology involves genetic, environmental, infectious, and immune factors. A number of studies have proved that genetic factors play a significant role in the pathogenesis of the disease. At least 36 susceptibility genes of AS have been found so far [1], among which human leukocyte antigen B27 (HLA-B27) has the closest association with the disease [2, 3], especially HLA-B*2705, although HLA-B*2706 and HLA-B*2709 are not related to AS [4–6]. However, these genes can explain only 24.4% of the genetic predisposition toward AS, with HLA-B*27 accounting for 20.1% and the remaining genes for 4.3%. This indicates that more than 75% of AS genetic susceptibility has yet to be found [1, 2, 7].
Some studies suggested that about 50%–60% of patients with AS also suffered from intestinal inflammation [2, 8], and 5%–10% of patients with AS had clinical evidence of inflammatory bowel disease (IBD), including Crohn's disease (CD) and ulcerative colitis (UC) [9, 10]. Another study found that the first-degree relatives of patients with AS were three times more likely to suffer from IBD compared with the general population [1, 11]. A genome-wide association study revealed that 20 of 31 AS-susceptibility genes overlapped with those of IBD; the same single-nucleotide polymorphisms (SNPs) had the same or similar effects in both diseases [1, 7, 12, 13]. In addition, the genes linked to AS involved various aspects of infection and immunology [14]. Therefore, it is promising to further search for susceptibility genes of AS among infection-related genes, especially relevant genes of IBD.
The human fucosyltransferase2 (FUT2) gene encodes α1,2-L-fucosyltransferase, which catalyzes the biosynthesis of type-1 H (Led), a blood group antigen, using a type-1 precursor (Lec) and fucose. Type-1 H is also a precursor of several other blood group antigens, such as type-1 A or type-1 B. An individual having functional FUT2 (Se) is called a secretor, whose fluids and secretions may contain soluble A, B, and H antigens. Otherwise, an individual with two null FUT2 alleles (se) is called a non-secretor, who cannot synthesize these antigens. The human FUT3 gene encodes α1,3/4-L-fucosyltransferase, which catalyzes the synthesis of Lea antigen in non-secretors using Lec and fucose, or catalyzes the synthesis of Leb in secretors using type-1 H and fucose. Therefore, non-secretors with functional FUT3 (Le) exhibit the Le(a+b-) serotype, secretors with Le exhibit the Le(a-b+) serotype, and individuals with two null FUT3 alleles (le) present the Le(a-b-) serotype despite their secretor status [15]. The type-1 A, B, H, and Lewis antigens (such as Leb) are widely distributed in human fluids and secretions and on the surface of genitourinary and gastrointestinal epithelial cells. Hence, they are called histo-blood group antigens (HBGAs).
FUT2 and FUT3 genes are closely related to gut inflammation. The expression of both FUT2 and FUT3 genes needs the participation of host intestinal flora [16–19]. Meanwhile, non-secretors are more susceptible to CD [16, 20]; FUT3 gene polymorphisms and expression in the intestinal tract are associated with the susceptibility to UC [21]. Therefore, it is supposed that FUT2, FUT3, and HBGAs may play certain roles in the pathogenesis of AS and relevant intestinal inflammation. The aim of this study was to explore the associations between FUT2/FUT3 polymorphisms and AS in an east Chinese population.
Materials and methods
Participants
A total of 673 patients with AS were selected at the Clinic of Rheumatology and Immunology, the First Affiliated Hospital of Anhui Medical University from January 2015 to June 2018. All the patients were diagnosed by senior rheumatologists according to the New York Diagnostic Criteria (revised in 1984). Meanwhile, 687 age- and sex-matched healthy controls were recruited from the Health Checkup Center of the hospital. All the participants came from east China, mostly from Anhui Province. They had no underlying diseases such as liver, kidney, and cardiovascular or cerebrovascular diseases. In addition, they had not suffered from infectious diseases, such as acute intestinal inflammation, within six months before sampling; had not undergone surgery within one year; and had not taken non-steroidal anti-inflammatory drugs, glucocorticoids, traditional Chinese medicine, and other anti-rheumatic drugs within three months prior to sampling. Furthermore, none of the healthy controls had family history of rheumatic and autoimmune diseases. All the participants were informed of the study in detail and voluntarily signed a written informed consent prior to their participation. Each of the patients with AS then filled in a questionnaire on personal diet, living habits, and health status. The controls provided their health information orally. Subsequently, the samples of peripheral blood were collected from all the participants with anticoagulant tubes containing ethylene diamine tetraacetic acid (EDTA).
This case–control study fully complied with the Declaration of Helsinki and local laws, and was approved by the Biomedical Ethics Committee of Anhui Medical University (demonstration report number: 20150117). After signing the written informed consent, researchers were forbidden from accessing any identity information of the participants.
DNA preparation and genotyping
The extraction of genomic DNA from the specimens of blood was carried out with a QIAamp DNA Blood Mini Kit (Qiagen, Germany) according to the manufacturer’s protocols. The prepared DNA solutions were cryopreserved in a -80ºC refrigerator until testing. The selection of SNPs for genotyping was accomplished based on their frequencies in the Oriental population and their roles in the determination of HBGAs (Table 1). Genotyping was conducted using a patented method named SNPscan [22], which was based on double-ligation reactions and a multiplex fluorescence polymerase chain reaction (PCR), with a custom-made 48-Plex SNPscan Kit (Genesky, China). The PCR products were subsequently separated and detected with an ABI 3730 XL sequencer (ABI, Carlsbad, CA 92008, USA) using capillary electrophoresis. To meet the technical requirements of the SNPscan test, the complementary bases of rs28362459 and rs812936 were actually tested.
DNA preparation and genotyping were completed at Department of Blood Transfusion, the First Affiliated Hospital of USTC. For quality control, 4% of randomly chosen DNA samples were genotyped repeatedly at Genesky Biotechnologies Inc., Shanghai, China.
According to the principle of blind test, all the samples were randomly coded with a test number before DNA extraction. No participant’s information for any specimens was available until the data analysis.
Inference of phenotype
According to the Blood Group Mutation Database (archived now) [23] and the blood typing data of local patients, a sample with any variant of rs1047781-TT, rs1800028-TT, or rs1800030-AA was classified as non-secretor (se), otherwise as secretor (Se). Similarly, an individual with rs28362459-GG or rs812936-CC was sorted as Lewis negative (le), otherwise as Lewis positive (Le). Lewis serotypes were inferred as follows: Lewis-negative individuals, both secretors and non-secretors, were classified as Le(a-b-). Lewis-positive secretors were sorted as Le(a-b+), and Lewis-positive non-secretors as Le(a+b-).
Data analyses
The questionnaire and the genotyping data of all the participants were collected, input into a database, and proofread twice. The frequencies of each genotype, allele, secretor status, Lewis status, serotype, and haplotype were compared between the cases and the controls. Potential interactions among these factors, age, and sex were also analyzed. The effects of FUT2/FUT3 polymorphisms on the Bath ankylosing spondylitis disease activity index (BASDAI) and the Bath ankylosing spondylitis functional index (BASFI), which were usually used to measure the disease severity of AS, were assessed specially.
Statistical analyses
Quantitative variables with normal distribution were reported as mean ± standard deviation (SD); otherwise, as median (interquartile range, IQR). The frequency differences in sex, genotype, allele, and phenotype between the cases and the controls were all analyzed using the chi-square test. The difference in age was analyzed using the independent samples t test. The associations between FUT2/FUT3 and clinical indexes of AS were analyzed using the Mann–Whitney U test for two independent samples and the Kruskal–Wallis test for K independent samples. If there was no statistical difference in a variable between the treated and untreated patients, the two parts of data would be combined. The cases with missing values were excluded test-by-test according to the results of the Missing Value Analysis. Two-factor interactions on a multiplicative scale were assessed employing the binary logistic regression. All these analyses were completed using SPSS Statistics 23.0 (IBM, USA).
The analyses of two-factor interaction on an additive scale were carried out adopting the indexes proposed by Rothman and Hosmer [24, 25]. The estimations of relative excess risk due to interaction (RERI), attributable proportion due to interaction (AP), and synergy index (S) were accomplished using the Excel sheet written by Andersson et al [26]. If the confidence interval (CI) of RERI and AP did not contain 0, and the CI of S did not include 1, it was regarded as having additive interaction between the two factors.
Multifactor interaction analyses were performed using Multifactor Dimensionality Reduction (MDR) 3.0.2. Haplotype analyses and Hardy–Weinberg equilibrium tests were conducted using Haploview 4.2. P values of haplotype analyses were corrected for multiple testing bias using permutation test. Sample size estimation and power analyses were conducted using PASS 11.0.7. P values of other multiple comparisons were all corrected adopting the false discovery rate (FDR) method using R 3.6.2. P or corrected P (if applicable) <0.05 was regarded as statistically significant.
Results
Demographic and clinical characteristics of participants
A total of 925 patients with AS were enlisted and 673 of them (546 males and 127 females) agreed to join this study eventually. Meanwhile, 687 (560 males and 127 females) of 892 eligible healthy controls consented to participate (control/case ≈ 1.021). The male/female ratio (4.30 vs 4.41, χ2 = 0.856, P = 0.890) and the age [26.0 (11.0) vs 27.0 (10.0) years, Z = -1.118, P = 0.264] of the cases and controls were comparable (S1 Table). Furthermore, the genotype frequencies of all the five SNPs included in this study conformed to Hardy–Weinberg equilibrium (all P > 0.05). Raw data of genotyping outcomes were partly provided in S1 Fig, and main lifestyles and clinical characteristics of the patients with AS were listed in S1 Table.
Associations between FUT2/FUT3 polymorphisms and human susceptibility to AS
The genotype frequencies of the five SNPs included in this study showed no differences between the cases and the controls, both in the total participants and in either sex subgroup (all Pcorrected > 0.05) (Table 2). At the allele level, however, a significant frequency difference of rs28362459 was observed between the patients with AS and the healthy controls among the total participants [χ2 = 7.515, Pcorrected = 0.030; adjusted odds ratio (ORadjusted)G/T = 0.782; 95% CI, 0.650–0.941; Power = 0.8] (Table 2, S2 Fig). Rs28362459-G, which usually leads to a Lewis-negative phenotype, showed a lower frequency in the patients with AS than in the healthy controls (20.3% vs 24.7%). No differences in allele frequency were found in the remaining SNPs among the total participants (all Pcorrected > 0.05). In the male and female subgroups, all the five SNPs showed no significant differences in allele frequency between the cases and the controls (all Pcorrected > 0.05). These suggested that the polymorphisms of rs28362459 were probably associated with human susceptibility to AS, and the allele rs28362459-G might be a protective factor for AS.
Haplotype analyses of the total participants revealed significant differences between the cases and the controls in the frequencies of haplotype TT (rs812936–rs28362459) (χ2 = 5.663, Ppermutation = 0.039; Power = 0.7; D' = 0.931) and TG (rs812936–rs28362459) (χ2 = 7.456, Ppermutation = 0.013; Power = 0.8; D' = 0.931). A significant difference in the frequencies of haplotype TG (rs812936-rs28362459) was also found between the two groups in the female subgroup (χ2 = 5.624, Ppermutation = 0.047; Power = 0.7; D' = 0.916) (Table 3). These results further suggested a link between FUT3 polymorphisms and the susceptibility to AS.
In the analyses of two-factor interaction, multiplicative interactions were found in the total participants between rs28362459-TG and age (ORadjusted = 0.071; 95% CI, 0.005–0.980; Pcorrected = 0.048), and rs28362459-TG and sex (ORadjusted = 0.549; 95% CI, 0.305–0.989; Pcorrected = 0.046). In the male subgroup, a multiplicative interaction was observed between rs28362459-TG and age (ORadjusted = 0.032; 95% CI, 0.002–0.616; Pcorrected = 0.022). In the female subgroup, no multiplicative interactions were found (all Pcorrected > 0.05). Moreover, additive interactions were observed between rs1047781 and rs28362459 in the total participants (RERI, 0.235–3.167; AP, 0.226–0.762; S, 1.109–9.792), the male subgroup (RERI, 0.669–3.282; AP, 0.327–0.727; S, 1.246–10.173), and the female subgroup (RERI, 0.334–1.499; AP, 0.181–0.565; S, 1.001–7.251) (Fig 1A, 1C and 1E). No multifactor interactions were found in the total participants and either subgroup (all P > 0.05).
(A, C, and E) Interactions on an additive scale between rs1047781and rs28362459 for total participants, male subgroup, and female subgroup, respectively. (B, D, and F) Interactions on an additive scale between Lewis status and secretor status for the total participants, the male subgroup, and the female subgroup, respectively.
Association between FUT2/FUT3 phenotypes and human susceptibility to AS
At the phenotypic level, no frequency differences in the secretor status were found between the cases and the controls, whether in the total participants or either sex subgroup (all Pcorrected > 0.05). Similarly, no differences were observed between the patients and the controls in terms of the frequencies of the Lewis status (all Pcorrected > 0.05) and the Lewis serotype (all Pcorrected > 0.05) (Table 4).
No multiplicative interactions were found at phenotypic level in all the factors (all Pcorrected > 0.05). Additive interactions were found between the Lewis status and the secretor status in the total participants (RERI, 0.747–2.338; AP, 0.247–0.590; S, 1.225–4.513), the male subgroup (RERI, 0.840–2.706; AP, 0.272–0.610; S, 1.285–4.567), and the female subgroup (RERI, 0.396–1.614; AP, 0.147–0.528; S, 1.031–4.009) (Fig 1B, 1D and 1F). No multifactor interactions were found at phenotypic level (all P > 0.05). These findings suggested that the phenotypes of FUT2 and FUT3, and their co-determined Lewis serotypes, might not correlate with the predisposition to AS.
Correlations between FUT2/FUT3 polymorphisms and BASFI/BASDAI
To explore the links between the gene polymorphisms and the clinical manifestations of AS, association analyses were conducted at the gene and phenotypic levels. No associations were found between the gene polymorphisms and the BASFI/BASDAI (all Pcorrected > 0.05). However, at the item level, the results revealed that rs1800030-A was related to an alleviated morning stiffness (the sixth item of the BASDAI, BASDAI 6; BASDAI 6GA: BASDAI 6GG = 0.0: 1.0; Z = -2.575, Pcorrected = 0.045; Power = 1.0), while rs28362459-G was related to a severer limitation with regard to standing without support (the sixth question of the BASFI, BASFI 6) among patients with AS (BASFI 6GG: BASFI 6GT: BASFI 6TT = 1.0: 0.0: 0.0; χ2 = 10.089, Pcorrected = 0.030; Power = 0.8). Lewis-positive (Le) patients suffered less functional loss compared with Lewis-negative (le) patients in terms of BASFI 6 (Z = -2.319, Pcorrected = 0.040; Power = 0.3) and bending forward from the waist (BASFI 2; Z = -2.333, Pcorrected = 0.040; Power = 0.6) (Table 5). In line with this, both Le(a+b-) and Le(a-b+) groups had less functional limitations in BASFI 6 (χ2 = 7.220, Pcorrected = 0.027; Power = 0.3) and BASFI 2 (χ2 = 6.419, Pcorrected = 0.040; Power = 0.5) compared with the Le(a-b-) patients. In addition, the Le(a+b-) and Le(a-b+) patients also suffered less than the Le(a-b-) ones in regard to reaching up to a high shelf (BASFI 3; χ2 = 6.848, Pcorrected = 0.033; Power = 0.5) and climbing 12–15 steps independently (BASFI 7; χ2 = 6.800, Pcorrected = 0.033; Power = 0.3) (Table 5). No significant differences were observed between the remaining factors and clinical variables (all Pcorrected > 0.05). These results indicated that the polymorphisms of FUT2 and FUT3 were likely to be associated with the symptoms of AS, and rs28362459-G might aggravate relevant symptoms.
Discussion
In view of a large degree of overlap between the susceptibility genes of IBD and AS [1, 7, 12, 13], and a strong correlation between FUT2/FUT3 and IBD [16, 20, 21], it was hypothesized that a link existed between the two genes and human susceptibility to AS. The results of this study suggested that FUT3 polymorphisms were associated with the susceptibility to AS, and rs28362459-G might be a protective factor for AS. These findings suggested that some pathogenic factors might attach to Lewis antigens in the development of AS, and the lack of these antigens might block the process. This was consistent with the report that the adhesion of Helicobacter pylori to gastric mucosa correlated closely with Leb in the host body, and the binding of Escherichia coli to the epithelial cells of intestinal mucosa needed a glycosylation structure [16, 18]. This might also be similar to the situation in which infection rates of norovirus, rotavirus, and Helicobacter pylori were commonly much lower in non-secretors than in secretors [20, 27, 28].
The results of the haplotype analyses revealed significant differences between the cases and the controls in the frequencies of haplotypes TT (rs812936–rs28362459) and TG (rs812936–rs28362459). Both the SNPs located in FUT3 gene, indicating associations between the gene and AS at the haplotype level. These were consistent with the results at the allele level. Although no reports on the roles of these haplotypes can be found so far, the following points may provide a rational explanation for the results: It is clear the haplotype TG (rs812936–rs28362459) is one of the determinants of Lewis-negative status; and it is supposed that the haplotype TT (rs812936–rs28362459) may influence the structure and quantity of Lewis antigens.
This study discovered a potential susceptibility gene for AS, providing fresh data for exploring the etiology of AS. Meanwhile, this was the first report of a link between a blood group gene and AS, leading to new developments in the functional study of human blood groups. Although Shinebaum reported associations between ABO antigens of non-secretors and host susceptibility to spondyloarthropathies [29], the results proved to be false later [30]. Another study also suggested that the secretor status did not correlate with AS [31].
No significant differences were found between the cases and the controls in the frequencies of secretor status, Lewis status, and Lewis serotype. However, interactions on an additive scale were found between the Lewis status and the secretor status. The interactions might reflect the fact that the two factors jointly determined the type of Lewis HBGAs, which were likely the real molecules participating in the pathogenesis of AS. Moreover, the interaction between the FUT3 polymorphisms and age was possibly related to the influence of age on the expression of Lewis antigens, and the influence might also be associated with the maturity of human gut microbiota [18, 19].
The possible reasons for not finding correlations between FUT3 and AS at the phenotypic level were as follows. First, both FUT2 and FUT3 directly encode fucosyltransferase, which may have other biological functions besides determining the human secretor status and Lewis antigens. These functions may play varied roles in the development of AS. Second, the synthesis of the Lewis HBGAs complex, such as ALeb, is also influenced by FUT1 and ABO genes, besides being controlled by FUT3. Finally, the Lewis serotypes in this study were inferred only from genotyping results, with no verifications using serological and salivary tests due to specimen limitations.
This study found multiple associations between rs28362459 and the clinical manifestations of AS at the genetic and phenotypic levels. Generally speaking, rs28362459-G, Lewis-negative status, and Le(a-b-) serotype might correlate with several aggravated disease indexes. These demonstrated to some extent that disease activity, functional impairments, and imaging changes in AS were also highly heritable [1]. As mentioned earlier, rs28362459-G might undermine the development of AS, but it was also associated with some severer clinical manifestations. These seemingly inconsistent results indicated that FUT3 might play complicated roles in the pathogenesis of AS. A probable cause for the results was that rs28362459-G might bring about multiple changes in human gut microbiota, and further affect host immune functions. This could be similar to the situation in which the intestinal flora in a non-secretor’s colon changed significantly in both structure and function among patients with CD [32, 33].
As with other case–control studies, this study also had some shortcomings. Correlations were found only between the genes and AS, but relevant mechanisms were not explored. This study involved only FUT2 and FUT3 genes, but did not include FUT1 and ABO genes, which were also related to HBGAs. Therefore, the results of this study had certain one-sidedness. Moreover, due to the limited sample size, the power of analysis was affected to some extent for some subgroups. Due to the limited quality of questionnaire data, some analyses could not be carried out, such as the correlations between FUT3 and relevant laboratory variables. Due to the limited size of untreated cases, the combination of treated and untreated cases, without statistical difference in nonparametric tests, was used in the analyses of associations between the gene polymorphisms and disease indexes. This might bring some bias.
To verify the findings of this study and further uncover the underlying mechanisms, researchers can investigate the compositional and functional characteristics of intestinal flora among patients with AS in the future. Furthermore, they can establish proper animal models of AS and explore the roles of HBGAs by blocking the expression of FUT2 or FUT3. The findings may help in deepening the understanding of the associations between the FUT3 gene and AS, and also in developing new treatment strategy for AS.
Conclusions
FUT3 gene polymorphisms correlated with human susceptibility to AS, both at the allele and haplotype level. Rs28362459-G might decrease the susceptibility, but enhance part of disease indexes of AS. This study highlighted the probability of wide associations between the blood group antigens and human autoimmune diseases.
Supporting information
S1 Checklist. STROBE statement—checklist of items that should be included in reports of observational studies.
https://doi.org/10.1371/journal.pone.0237219.s001
(DOCX)
S1 Fig. Genotyping results of all the analyzed SNPs.
https://doi.org/10.1371/journal.pone.0237219.s002
(PDF)
S2 Fig. Genotype frequencies (left) and allele frequencies (right) of rs28362459 between patients with AS and healthy controls.
https://doi.org/10.1371/journal.pone.0237219.s003
(PDF)
S1 Table. Demographic and clinical characteristics of total participants.
https://doi.org/10.1371/journal.pone.0237219.s004
(PDF)
Acknowledgments
The authors are grateful to all the participants for their involvement and contributions to this study. They also thank the physicians at the Department of Rheumatology and Immunology, the First Affiliated Hospital of Anhui Medical University, for their help.
References
- 1. Brown MA, Kenna T, Wordsworth BP. Genetics of ankylosing spondylitis—insights into pathogenesis. Nat Rev Rheumatol. 2016; 12(2):81–91. pmid:26439405
- 2. Tsui FW, Tsui HW, Akram A, Haroon N, Inman RD. The genetic basis of ankylosing spondylitis: new insights into disease pathogenesis. Appl Clin Genet. 2014; 7:105–115. pmid:24971029
- 3. Mathieu A, Paladini F, Vacca A, Cauli A, Fiorillo MT, Sorrentino R. The interplay between the geographic distribution of HLA-B27 alleles and their role in infectious and autoimmune diseases: a unifying hypothesis. Autoimmun Rev. 2009; 8(5):420–425. pmid:19185064
- 4. Mathieu A, Cauli A, Fiorillo MT, Sorrentino R. HLA-B27 and ankylosing spondylitis geographic distribution as the result of a genetic selection induced by malaria endemic? A review supporting the hypothesis. Autoimmun Rev. 2008; 7(5):398–403. pmid:18486928
- 5. Lopez-Larrea C, Sujirachato K, Mehra NK, Chiewsilp P, Isarangkura D, Kanga U, et al. HLA-B27 subtypes in Asian patients with ankylosing spondylitis. Evidence for new associations. Tissue Antigens. 1995; 45(3):169–176. pmid:7761976
- 6. D'Amato M, Fiorillo MT, Galeazzi M, Martinetti M, Amoroso A, Sorrentino R. Frequency of the new HLA-B*2709 allele in ankylosing spondylitis patients and healthy individuals. Dis Markers. 1995; 12(3):215–217. pmid:8590548
- 7. International Genetics of Ankylosing Spondylitis C, Cortes A, Hadler J, Pointon JP, Robinson PC, Karaderi T, et al. Identification of multiple risk variants for ankylosing spondylitis through high-density genotyping of immune-related loci. Nat Genet. 2013; 45(7):730–738. pmid:23749187
- 8. Van Praet L, Van den Bosch FE, Jacques P, Carron P, Jans L, Colman R, et al. Microscopic gut inflammation in axial spondyloarthritis: a multiparametric predictive model. Ann Rheum Dis. 2013; 72(3):414–417. pmid:23139267
- 9. Costello ME, Ciccia F, Willner D, Warrington N, Robinson PC, Gardiner B, et al. Brief Report: Intestinal Dysbiosis in Ankylosing Spondylitis. Arthritis Rheumatol. 2015; 67(3):686–691. pmid:25417597
- 10. Bremander A, Petersson IF, Bergman S, Englund M. Population based estimates of common comorbidities and cardiovascular disease in ankylosing spondylitis. Arthritis Care Res (Hoboken). 2011; 63(4):550–556.
- 11. Thjodleifsson B, Geirsson AJ, Bjornsson S, Bjarnason I. A common genetic background for inflammatory bowel disease and ankylosing spondylitis: a genealogic study in Iceland. Arthritis Rheum. 2007; 56(8):2633–2639. pmid:17665420
- 12. Jostins L, Ripke S, Weersma RK, Duerr RH, McGovern DP, Hui KY, et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature. 2012; 491(7422):119–124. pmid:23128233
- 13. Parkes M, Cortes A, van Heel DA, Brown MA. Genetic insights into common pathways and complex relationships among immune-mediated diseases. Nat Rev Genet. 2013; 14(9):661–673. pmid:23917628
- 14. O'Rielly DD, Uddin M, Rahman P. Ankylosing spondylitis: beyond genome-wide association studies. Curr Opin Rheumatol. 2016; 28(4):337–345. pmid:27224740
- 15. Grubb R. Observations on the human group system Lewis. Acta Pathol Microbiol Scand. 1951; 28(1):61–81. pmid:14818817
- 16. Maroni L, van de Graaf SF, Hohenester SD, Oude Elferink RP, Beuers U. Fucosyltransferase 2: a genetic risk factor for primary sclerosing cholangitis and Crohn's disease—a comprehensive review. Clin Rev Allergy Immunol. 2015; 48(2–3):182–191. pmid:24828903
- 17. Nanthakumar NN, Dai D, Newburg DS, Walker WA. The role of indigenous microflora in the development of murine intestinal fucosyl- and sialyltransferases. FASEB J. 2003; 17(1):44–46. pmid:12475916
- 18. Cooling L. Blood Groups in Infection and Host Susceptibility. Clin Microbiol Rev. 2015; 28(3):801–870. pmid:26085552
- 19. Mackie RI, Sghir A, Gaskins HR. Developmental microbial ecology of the neonatal gastrointestinal tract. Am J Clin Nutr. 1999; 69(5):1035S–1045S. pmid:10232646
- 20. McGovern DP, Jones MR, Taylor KD, Marciante K, Yan X, Dubinsky M, et al. Fucosyltransferase 2 (FUT2) non-secretor status is associated with Crohn's disease. Hum Mol Genet. 2010; 19(17):3468–3476. pmid:20570966
- 21. Hu D, Zhang D, Zheng S, Guo M, Lin X, Jiang Y. Association of Ulcerative Colitis with FUT2 and FUT3 Polymorphisms in Patients from Southeast China. PLoS One. 2016; 11(1):e0146557. pmid:26766790
- 22. Jin G, Zhu M, Yin R, Shen W, Liu J, Sun J, et al. Low-frequency coding variants at 6p21.33 and 20q11.21 are associated with lung cancer risk in Chinese populations. Am J Hum Genet. 2015; 96(5):832–840. pmid:25937444
- 23.
National Center for Biotechnology Information (NCBI). Blood Group Mutation Database [Internet]. Archived in 2017. Available from: http://ftp.ncbi.nlm.nih.gov/pub/mhc/rbc/.
- 24.
Rothman KJ. Epidemiology: an introduction. New York: Oxford University Press; 2002.
- 25. Hosmer DW, Lemeshow S. Confidence interval estimation of interaction. Epidemiology. 1992; 3(5):452–456. pmid:1391139
- 26. Andersson T, Alfredsson L, Källberg H, Zdravkovic S, Ahlbom A. Calculating measures of biological interaction. Eur J Epidemiol. 2005; 20(7):575–579. pmid:16119429
- 27. Carlsson B, Kindberg E, Buesa J, Rydell GE, Lidon MF, Montava R, et al. The G428A nonsense mutation in FUT2 provides strong but not absolute protection against symptomatic GII.4 Norovirus infection. PLoS One. 2009; 4(5):e5593. pmid:19440360
- 28. Gunaydin G, Nordgren J, Sharma S, Hammarstrom L. Association of elevated rotavirus-specific antibody titers with HBGA secretor status in Swedish individuals: The FUT2 gene as a putative susceptibility determinant for infection. Virus Res. 2016; 211:64–68. pmid:26454189
- 29. Shinebaum R, Blackwell CC, Forster PJ, Hurst NP, Weir DM, Nuki G. Non-secretion of ABO blood group antigens as a host susceptibility factor in the spondyloarthropathies. Br Med J (Clin Res Ed). 1987; 294(6566):208–210.
- 30. Smith GW, James V, Mackenzie DA, Stewart J, Blackwell CC, Elton RA, et al. Ankylosing spondylitis and secretor status: a re-evaluation. Br J Rheumatol. 1997; 36(7):778–780. pmid:9255113
- 31. Pal A, Hill M, Wordsworth P, Brown M. Secretor status and ankylosing spondylitis. J Rheumatol. 1998; 25(2):318–319. pmid:9489826
- 32. Rausch P, Rehman A, Kunzel S, Hasler R, Ott SJ, Schreiber S, et al. Colonic mucosa-associated microbiota is influenced by an interaction of Crohn disease and FUT2 (Secretor) genotype. Proc Natl Acad Sci USA. 2011; 108(47):19030–19035. pmid:22068912
- 33. Tong M, McHardy I, Ruegger P, Goudarzi M, Kashyap PC, Haritunians T, et al. Reprograming of gut microbiome energy metabolism by the FUT2 Crohn's disease risk polymorphism. ISME J. 2014; 8(11):2193–2206. pmid:24781901