J Gastroenterol DOI 10.1007/s00535-015-1062-3
ORIGINAL ARTICLE—ALIMENTARY TRACT
Contribution of susceptibility variants at FCGR2A and 13q12 to the risk of relapse among Japanese patients with ulcerative colitis Kouichi Asano • Motohiro Esaki • Junji Umeno • Atsushi Hirano • Yuji Maehata • Tomohiko Moriyama • Shotaro Nakamura • Takayuki Matsumoto • Takanari Kitazono
Received: 15 October 2014 / Accepted: 25 February 2015 Ó Springer Japan 2015
Abstract Background Recent genome-wide association studies have identified nearly 100 susceptibility genes for ulcerative colitis (UC). However, the contribution of susceptibility variants for UC to clinical outcome has scarcely been reported. The aim of this study was to investigate whether UC-associated genetic variants confer a risk of clinical relapse. Methods One hundred and nine consecutive Japanese subjects with quiescent UC were recruited. Four genetic variants of HLA-DRB1*1502, rs6671847 at FCGR2A, rs17085007 at 13q12, and rs2108225 at SLC26A3 were genotyped by Invader assay. The clinical courses were followed after blood sampling, and the risk of relapse according to these genotypes was calculated by Cox proportional hazard model. K. Asano (&) M. Esaki J. Umeno A. Hirano Y. Maehata T. Moriyama T. Kitazono Department of Medicine and Clinical Science, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan e-mail: [email protected]
Results During the mean follow-up period of 35 months (range 1–81 months), 49 of 109 subjects (45 %) relapsed. Carriers of the G allele of rs6671847 showed an increased risk of relapse compared with non-carriers [adjusted hazard ratio (HR), 2.27; 95 % confidence interval (CI), 1.20–4.32; p = 0.01]. Patients with the CT or TT genotypes of rs17085007 also had an increased risk of relapse compared to subjects with the CC genotype (for CT: adjusted HR, 2.16; 95 % CI, 1.10–4.23; p = 0.03; for TT: adjusted HR, 3.25; 95 % CI, 1.18–8.95; p = 0.02). These two risk variants multiplied the risk of relapse by 2.74 times (95 % CI, 1.10–4.23; p = 0.01) in patients with one risk genotype and 5.40 times (95 % CI, 2.06–14.13; p = 0.0006) in patients with both risk genotypes. Conclusions Genetic variants of rs6671847 at FCGR2A and rs17085007 at 13q12 conferred a risk of relapse in patients with UC. Keywords Ulcerative colitis Susceptibility variants FCGR2A Risk of relapse Clinical course
Introduction K. Asano Department of Endoscopic Diagnosis and Therapeutics, Kyushu University Hospital, Fukuoka, Japan A. Hirano Department of Internal Medicine, Yamaguchi Red Cross Hospital, Yamaguchi, Japan S. Nakamura Department of Advanced Medical Initiatives, Faculty of Medical Sciences, Kyushu University, Fukuoka, Japan T. Matsumoto Division of Gastroenterology and Hepatology, Department of Internal Medicine, Iwate Medical University, Morioka, Japan
Ulcerative colitis (UC) is a chronic remittent inflammatory bowel disease [1], and its natural history and the response to treatments vary widely among patients [2]. In clinical practice, about a quarter of patients with UC who begin steroid therapy become steroid dependent [2], and 7–20 % of those patients are obliged to receive surgery 5 years after the initial diagnosis because of their refractory disease or the development of dysplasia or cancer [3]. On the other hand, half of patients with UC have a favorable clinical course even after the initiation of steroid therapy. Further, approximately 20 % of newly diagnosed patients have
123
J Gastroenterol
been estimated to experience a quiescent disease course without requiring medical therapy to maintain remission [4, 5]. Therefore, in order to improve the clinical outcome in UC, it is crucial to identify factors predictive of the natural history and treatment response. Clinical, serological, and genetic markers associated with clinical outcome have been rigorously investigated, and a variety of predictive factors have been proposed so far. Younger age at onset, extent of the disease, smoking cessation, and multiple previous relapses have been shown to be clinical predictors of earlier relapse [6]. Biomarkers, including serum C-reactive protein (CRP), anti-neutrophil cytoplasmic antibody [7], and fecal calprotectin [8] also have been possible candidates, but these predictive factors can vary during the course of the disease, making them less reliable candidates. Therefore, more accurate predictive markers need to be determined in order to implement more effective, personalized treatment. Recent genome-wide association studies (GWAS) and subsequent meta-analyses have identified approximately 100 susceptibility genes for UC [9–11]. These susceptibility genes should help to unravel complex components in the pathogenesis of UC, such as epithelial barrier function, regulation of adaptive immunity, endoplasmic reticulum stress, and host–microbial interactions. In addition, recent genetic discoveries have the potential to change clinical strategies by predicting the natural history and guiding the choice of therapy. In fact, genetic markers are very attractive candidates for predictive factors due to their invariability and low invasiveness [12]. However, little is known about the influence of genetic factors on disease phenotypes and clinical outcomes in UC [13]. We thus aimed to investigate whether UC-associated genetic variants confer a risk of clinical relapse.
Materials and methods Samples Between May 2006 and February 2008, we collected clinical data and whole-blood DNA samples of 129 consecutive Japanese subjects with quiescent UC at Kyushu University Hospital. Twenty subjects were excluded because of their insufficient clinical information, and thus the present study comprised 109 subjects. The diagnosis of UC was made by expert gastroenterologists in accordance with international criteria defined by clinical, endoscopic, and histological features [14]. All patients with quiescent UC fulfilled the following criteria within the 3 months prior to study entry: (1) maintenance of a Seo’s ulcerative colitis activity index (UCAI) of less than 120 [15], (2) no
123
remission induction therapy, including cyclosporine, infliximab, adalimumab, tacrolimus, or cytapheresis therapy, and (3) use of systemic corticosteroid of less than 20 mg/day, which was defined as low-dose corticosteroid. Study follow-up and outcome measurements The medical records were carefully reviewed, and clinical information regarding disease activity, medical treatment, and laboratory data until December 2012 were obtained for the present analyses. During follow-up, each patient received regular clinical evaluation every 1 or 2 months, and disease relapse was determined to have occurred when any remission induction therapy was required to reduce disease activity (UCAI [ 120). In the present study, we assessed colectomy as the primary outcome and disease relapse as the surrogate primary outcome. Genotyping We genotyped four genetic variants (HLA-DRB1*1502, rs6671847 at FCGR2A, rs17085007 at 13q12, and rs2108225 at SLC26A3) that had been shown to have robust associations with UC susceptibility in Japanese [16] as well as in a European population [11]. These variants were successfully genotyped in all samples using an Invader assay. To tag HLA-DRB1*1502, we used a two-marker haplotype—i.e., a G-T haplotype comprising two SNPs, rs3135365 and rs7382794, which corresponds to HLADRB1*1502. This tag was chosen on the basis of the Wellcome Trust Sanger Institute data (http://www.sanger. ac.uk/HGP/Chr6/ng2006-data) [17]. Estimation of haplotype in each individual was conducted using PHASE software [18]. To test how well these tags worked to predict HLA status, we conducted high-resolution (four digits or more) typing of HLA-DRB1 in 191 samples in the previous GWAS [16] using a sequence-based typing method. Sequence data were collected and analyzed with an ABI 3730 DNA Analyzer (Applied Biosystems) and Assign software (Conexio Genomics, Applecross, Western Australia). The results yielded a correct prediction rate of 97.4 % (sensitivity, 98.7 %; specificity, 96.4 %). Statistics Relapse-free probabilities according to various genotypes were estimated using the Kaplan–Meier method, and the probabilities between the two groups were compared using a log-rank test. Comparisons of the risk of relapse between two or more groups and estimation of the hazard ratio (HR) and 95 % confidence interval (CI) with consideration for
J Gastroenterol
multiple confounding factors were carried out using a multivariate Cox proportional hazards model. We confirmed the proportionality of hazards in standard Cox models by assessing with the Schoenfeld goodness of fit test [19]. We compared the risk of each clinical characteristic listed in Table 1 between patients with relapse and those without using age- and sex-adjusted models, and clinical characteristics with a p value of \0.3 were initially chosen as potential risk factors for relapse. Consequently, male sex, age at sampling, disease duration, history of cytomegalovirus (CMV) infection, the duration between last flare-up and baseline, previous corticosteroid use, and current drugs including thiopurines, aminosalicylates, and methotrexate were determined to be confounding factors by the multivariate Cox regression analyses. Age at onset, previous treatment of cytapheresis and current drugs including thiopurine and probiotics were eliminated because
Table 1 Baseline characteristics according to relapse status
several factors were correlated with each other in the present analyses. CMV infection was defined to be positive by the detection of positive CMV antigenemia assay (CMVC7HRP) or by histological confirmation of inclusion body in biopsy specimens under H&E or immunohistochemical staining. All tests were two-sided, and a p value of \0.05 was considered to be statistically significant. For general statistical analysis, we used the R statistical environment version 2.14.1 (cran.r-project.org).
Ethical considerations The ethical committees of Kyushu University approved this study, and written informed consent was obtained from all subjects.
Characteristics
Non-relapse N = 60
Relapse N = 49
P**
Male sex
22 (36.7)
24 (49)
0.08
Age at onseta (years)
28 (8–69)
29 (13–69)
0.14
Age at samplinga
36 (14–60)
35 (15–74)
0.29
Disease durationa (years)
4.5 (0.5–36)
4.0 (0.5–45)
Disease type Proctitis
11 (18.3)
3 (6.1)
Left-sided
16 (26.7)
15 (30.6)
Extensive
31 (51.7)
29 (40.8)
Segmental
2 (3.3)
2 (4.1)
Family history
7 (11.7)
2 (4.1)
Smoking status Never smoker
0.14 0.72
0.61 0.91
49 (81.7)
40 (81.3)
Former smoker
5 (8.3)
4 (8.2)
Current smoker
6 (10.0)
5 (10.4)
C-reactive protein [0.1 mg/dl
22 (36.7)
20 (40.8)
0.58
Time since last flare-upa (months)
26.1 (3–252)
5.4 (3–110)
0.0004
History of CMV infection
8 (13.3)
8 (16.3)
0.12
39 (65.0)
44 (89.8)
0.02
Previous therapy Corticosteroids
Time since last flare-up was defined as the duration between the last flare-up and baseline. Low-dose corticosteroid was defined as less than 20 mg/day CMV cytomegalovirus ** p values were calculated by Cox proportional hazard model a
Data are summarized as median and range (min–max)
Cyclosporine
5 (8.3)
5 (10.2)
0.32
Tacrolimus
1 (1.7)
1 (2.0)
0.54
Cytapheresis
14 (23.3)
14 (28.6)
0.15
13 (21.7)
16 (33.3)
0.13
Current therapy Thiopurines Aminosalicylates
49 (81.7)
47 (95.9)
0.02
Low dose corticosteroids Methotrexate
14 (23.3) 0 (0)
17 (34.7) 5 (10.4)
0.03 0.0002
Topical therapies
5 (8.3)
4 (8.3)
0.98
Probiotics
22 (36.7)
22 (45.8)
0.03
123
J Gastroenterol
Results Clinical characteristics according to relapse status Among our 109 subjects, only five patients underwent colectomy during the mean follow-up period of 71.8 months, thus failing to conduct statistical analysis concerning genetic association with colectomy. We subsequently assessed a possible association of genetic factors with the disease relapse. During the mean follow-up period of 35.1 (1–81) months, the disease relapse was determined to have occurred in 49 patients (45 %). The relapse rate at 2 years after study entry was estimated to be 32 % (95 % CI, 22–41 %) by the Kaplan–Meier method. Baseline characteristics according to the relapse status are shown in Table 1. In regard to previous treatment, 86 patients (76 %) received systemic corticosteroids, but no patient received anti-TNF agents. When analyzing the risk of relapse according to baseline characteristics, the time since last flare-up (the duration between the last flare-up and baseline) was significantly shorter in patients without relapse than in those with relapse (p = 0004). As for previous therapies, systemic corticosteroids were more frequently applied to patients with relapse than in those without relapse (p = 0.02). Similarly, current use of aminosalicylates (p = 0.02), low-dose corticosteroids (p = 0.03), methotrexate (p = 0.0002), and probiotics (p = 0.03) were more frequent in patients with relapse than in those without relapse. The proportion of males was slightly higher in patients with relapse than in those without relapse, but the difference did not reach the level of statistical significance (p = 0.08). Other characteristics were not different between patients with relapse and those without. Risk of relapse according to genotype In the present study, the minor allele frequencies (MAF) of HLA-DRB1*1502, rs6671847, rs17085007, and rs2108225 were 30.7, 11.9, 30.3, and 28.4 %, respectively (Table 2). These frequencies were equivalent to those of UC cases in our previous GWAS. First, we assessed relapse-free probability stratified by the four genetic variants (Fig. 1). The relapse-free rate was lower in carriers of the G allele of rs6671847 than in nonTable 2 Genotype and allele frequency of variants
SNP
Chr
carriers, but there was no subject with a homozygous minor allele of rs6671847. The relapse-free rate was lower in carriers of the T allele of rs17085007 than in non-carriers (p = 0.02), while no difference was observed in relapsefree rate between the CT and TT genotypes of rs17085007. On the other hand, no difference in relapse-free rate was observed when stratified by genotypes of either HLADRB1*1502 or rs2108225 at SLC26A3 (Fig. 2). Next, we performed Cox proportional analyses to assess the contribution of genetic variants to the disease relapse (Table 3). Consequently, carriage of the G allele of rs6671847 was found to be significantly associated with an increased risk of relapse compared with non-carriage in a sex- and age-adjusted model [hazard ratio (HR), 1.92; 95 % confidence interval (CI), 1.04–3.54; p = 0.04]. This association remained significant even after adjustment for multiple confounding factors (multivariate adjusted HR, 2.27; 95 % CI, 1.20–4.32; p = 0.01). The CT and TT genotypes of rs17085007 were associated with an increased risk of relapse compared to subjects with the CC genotype in a multivariate-adjusted model (for CT: adjusted HR, 2.16; 95 % CI, 1.10–4.23; p = 0.03; for TT: adjusted HR, 3.25; 95 % CI, 1.18–8.95; p = 0.02). No difference was observed in the effect size between the CT and TT genotypes. Together with the results of the Kaplan–Meier survival analysis, these results indicate, in our samples, the risk genotype for rs17085007 was carriage of the T allele. On the other hand, HLA-DRB1*1502 and rs2108225 at the SLC26A3 gene were not associated with disease relapse. Combined effects of genetic variants on the risk of relapse Since each genetic variant had a different effect on disease susceptibility, we assessed possible multiplicative effects of the two genetic variants of rs6671847 and rs17085007. The relapse-free rate (Fig. 1) was lower in carriers having at least one risk variant than in non-carriers (p = 0.01). As shown in Table 4, multivariate analyses demonstrated that the risk of relapse in patients with one of the two risk genotypes increased by 2.74 times (95 % CI, 1.10–4.23; p = 0.01) compared to that in patients without either risk genotype. Moreover, the value in patients with both risk genotypes increased by 5.40 times (95 % CI, 2.06–14.13;
Gene
Minor allele
UC risk allele
Genotypea 0 (%)
1 (%)
MAF (%) 2 (%)
MAF minor allele frequency a
Genotype; ‘‘0’’ is homozygous major alleles, and genotype ‘‘1’’ is heterozygous, and ‘‘2’’ is homozygous minor alleles
123
HLA-DRB1*1502
6
HLA-DRB1
*1502
*1502
45.9
46.8
7.3
30.7
rs6671847
1
FCGR2A
G
A
76.1
23.9
0
11.9
rs17085007 rs2108225
13 9
– SLC26A3
T A
T G
49.5 48.6
40.4 45.9
10.1 5.5
30.3 28.4
J Gastroenterol
Fig. 1 Relapse-free probability stratified by genotypes of four markers using the Kaplan–Meier method. a HLA-DRB1*1502, b rs6671847 at FCGR2A, c rs17085007 at 13q12, d rs2108225 at SLC26A3. The bold characters in each genotype indicate risk alleles for developing UC
p = 0.0006) compared to that in patients without either risk genotype.
Discussion The present study demonstrated genetic associations between variants at the FCGR2A and 13q loci and clinical relapse in UC. Along with the disease susceptibility, possible genetic associations with clinical phenotypes, disease course, or treatment response have been investigated in
UC. These studies found that HLA-DRB1*0103 was associated with extensive colitis, severe disease and requirement of colectomy, in addition to disease susceptibility [20–22]. In addition, MDR1 polymorphism, which has been widely studied in the context of steroid therapy, was suggested to be associated with refractoriness to corticosteroids [23]. This polymorphism has also been shown to be associated with a lack of response to cyclosporine in patients with steroid-refractory UC [24]. Haritunians et al. [13] recently performed an unbiased GWAS in which they compared candidate SNPs between refractory UC patients
123
J Gastroenterol
pathogenesis of UC [26]. Thus, replication studies in various ethnic populations and comprehensive analyses considering environment factors are needed, since no previous study has assessed the predictive values of genetic variants from the point of view of disease relapse in UC. Predictive markers that influence on the disease phenotype can vary by the difference in the setting of clinical outcome [27]. In the present study, we initially investigated a possible association of genetic factors with colectomy; however, only five patients underwent colectomy during the mean follow-up period of 71.8 months. Considering that all of the patients were carriers of at least one risk variant (rs6671847 or rs17085007), these variants can be possible predictors of intractable clinical course requiring colectomy, whereas we could not draw any conclusion owing to the small number of colectomy cases. We thus investigated possible effect of genetic factors on the disease relapse. It has previously been reported that clinical factors including serum CRP levels at baseline [28], duration from last flare-up [6], smoking status [29], and history of CMV infection [30] were associated with unfavorable clinical outcomes in UC, and thus these have been considered as potential risk factors for disease relapse. Of these clinical markers, the duration from last flare-up was the only factor significantly associated with disease relapse in the present study. The discrepant results for the other markers may have been due to differences in the clinical characteristics of the study subjects or in the study design between the present and previous studies. Indeed, the clinical characteristics, including sex, age at onset, disease duration, and extent of disease in our subjects were slightly different from those of the previous studies [6, 28, 29]. However, the relapse rate at 2 years of follow-up was calculated to be 32 % (95 % CI, 22–41 %) in the present study, which was roughly equivalent to the value of a Europe-wide
Fig. 2 Relapse-free probability stratified by number of risk genotypes using the Kaplan–Meier method. The relapse-free rate was lower in carriers having at least one risk variant than in non-carriers (p = 0.01)
requiring colectomy and controls with favorable response to medical treatments. Their results indicated that TNFSF15 (TL1A), IL-10 (1q32.1), and IL-12B (5q33.3) were relevant to the disease course, but not HLA alleles or polymorphisms at the FCGR2A, 13q, or SLC26A3 loci. In contrast, TNFSF15, IL-10, and IL-12B were not associated with relapse in our samples (TNFSF15, p = 0.11; IL10, p = 0.25; IL12B, p = 17, respectively). This fact implies that the genetic contribution of UC-associated variants to the disease course varies among ethnicities [25]. In addition, accumulative evidence suggest that environment factors, especially intestinal microflora, also can play important roles in disease relapse as well as in the Table 3 Cox proportional hazards analysis of the effect the four genetic variants on the risk of relapse
Variables
rs6671847
HR hazard ratio, CI confidence interval a
Adjusted for sex, age at sampling, disease duration, history of cytomegalovirus infection, time since last flareup, previous corticosteroids use, current drugs including thiopurines, aminosalicylates, methotrexate
123
n
Event
rs17085007
rs2108225
24
Multivariate analysisa
Univariate analysis HR
HLA-DRB1*1502 Bold characters are risk alleles for developing UC in the previous Japanese genome-wide association study [1]
Genotype
95 % CI
p
Ref
HR
95 % CI
p
00
50
Ref
01
51
21
0.66
0.37–1.19
0.17
0.68
0.36–1.29
0.24
11 AA
8 83
4 34
1.31 Ref
0.45–3.80
0.62
1.65 Ref
0.52–5.23
0.39
AG
26
15
1.75
0.95–3.21
0.07
2.27
1.20–4.32
0.01
GG
0
NA
NA
NA
CC
54
19
Ref
Ref
CT
44
24
1.92
1.05–3.50
0.03
2.16
1.10–4.23
0.03
TT
11
6
1.84
0.73–4.63
0.19
3.25
1.18–8.95
0.02
GG
53
27
Ref
AG
50
19
0.74
0.41–1.33
0.31
0.92
0.49–1.73
0.79
AA
6
3
0.94
0.29–3.10
0.92
1.41
0.40–4.98
0.59
Ref
J Gastroenterol Table 4 Cox proportional hazard analysis of combined effects of rs6671847 and rs17085007 on the risk of relapse Number of risk variant
n
Event
Multivariate analysisa
Univariate analysis HR
95 % CI
p
HR
95 % CI
p
0
42
12
Ref
1
53
29
2.48
1.26–4.87
0.01
2.74
Ref 1.27–5.91
0.01
2
14
8
2.81
1.15–6.87
0.02
5.40
2.06–14.13
0.0006
p for trend
0.006
p for trend
0.0003
HR hazard ratio, CI confidence interval a
Adjusted for sex, age at sampling, disease duration, history of cytomegalovirus infection, time since last flare-up, previous corticosteroids use, current drugs including thiopurines, aminosalicylates, methotrexate
population-based study (40 % at 2 years) [31]. It thus can be assumed that the clinical background of our subjects was not particularly biased when compared to those of the previous investigations. When discussing disease relapse in UC, medical treatments should also be considered. In the present study, current use of aminosalicylates, methotrexate, and probiotics, as well as previous use of systemic corticosteroids were associated with increased risk of disease relapse. On the other hand, no significant association was found between treatment by thiopurines and disease relapse. It may be possible that patients treated with these medications have more severe disease, which would result in a higher rate of disease relapse [31]. However, it seems more plausible that the peculiar association between the medical treatments and disease relapse was caused by inappropriate application of maintenance treatment to aggressive disease. Indeed, thiopurines have been considered to have more beneficial maintenance effects than 5-ASA, especially in refractory UC [32]. Considering the inverse association of medical treatment with disease relapse, the insignificant association of thiopurines with disease relapse might have been the result of their modest but beneficial maintenance effect in patients with severe or relapsing disease. FCGR2A is implicated to be a susceptibility gene for several autoimmune diseases. Although the risk allele for severe or relapsing disease is assumed to be the same as that for disease susceptibility [33], the carriage of the G allele at rs6671847 of the FCGR2A gene, which is considered to be protective for developing UC, was associated with a higher risk for disease relapse in the present study. For this paradoxical result, two possible explanations can be postulated. First, this polymorphism may exert one set of biological activities for disease onset and a different set for the rest of the disease course [34]. Second, it is known that this polymorphism also affects the response to medical treatment, thereby modifying the risk for disease relapse. Actually, rs6671847 is in strong linkage disequilibrium with FCGR2A-131H/R that can cause functional difference. The H131 amino acid has higher binding affinity for
immune complexes than R131, resulting in a hyperactivation of immune cells [35]. This hyperactivation includes a higher capacity for antigen presentation to T cells and activation of B cells, as well as the overproduction of cytokines, resulting in perpetual inflammation of the colorectal mucosa. In addition, FCGR2A-131H/R has been reported to be associated with response to infliximab among patients with rheumatoid arthritis (RA) [36, 37]. In these reports, patients with FCGR2A-R131, low affinity FCGR2A allele, have been shown to have better response to IFX. In the present study, since no patients received antiTNF therapy during the follow-up periods, we could not evaluate the effect FCGR2A-131H/R or rs6671847 on the response to anti-TNF therapy. However, it is possible that carriers of the A allele of rs6671847 at the FCGR2A gene (UC risk allele) are more sensitive to maintenance therapies including aminosalicylates and immunomodulators. Further examination to explore possible differences in the response to treatment according to FCGR2A genotypes is necessary. Rs17085007 is located in the non-coding regions on chromosome 13q12.13. The gene nearest to this region is USP12 (encoding ubiquitin-specific peptidase-12), while its function remains unclear. However, possible contribution of rs17085007 to disease relapse in UC can be postulated as shown in CD-associated polymorphisms at 5p13.1. That is to say, while this variant also lies in the non-coding region, the variant has been thought to be a cisregulatory element that modulates the expression level of prostaglandin E receptor 4 (PTGER4) approximately 270 kb away from the variant [38]. It is thus possible that the rs17085007 locus contains regulatory sequences. In the present study, several limitations should be noted. These include the limited number of study samples, the assessment only during specific periods of the disease course, and the retrospective nature of the study itself, any of which could cause a selection bias. Statistically, while the present study showed genetic contributions of the FCGR2A gene and 13q12 to disease relapse in UC, 95 % confidence intervals of the risk variants were rather broadly
123
J Gastroenterol
calculated in the present study, indicating the necessity of a larger study sample size. Moreover, the sensitivity and specificity of risk variants for disease relapse were comparatively low, and the area under the receiver operating characteristic curve was 0.63 even in the combined effects model using rs6671847 and rs17085007. Therefore, these genetic variations alone would seem to be insufficient for predicting the disease course in UC. However, since the risk variants had a multiplicative effect in predicting disease relapse in the present study, future investigations combining analyses of genetic variants and clinical and biological markers could yield clinically useful indices for improving the treatment strategies in UC. In conclusion, genetic variants of rs6671847 at FCGR2A and rs17085007 at 13q12 confer risk of clinical relapse in Japanese patients with UC. These findings may help identify patients who exhibit a relapsing disease course and thus require intensive maintenance treatments. Conflict of interest The authors declare that they have no competing financial interests.
References 1. Abraham C, Cho JH. Inflammatory bowel disease. N Engl J Med. 2009;361:2066–78. 2. Faubion WA Jr, Loftus EV Jr, Harmsen WS, et al. The natural history of corticosteroid therapy for inflammatory bowel disease: a population-based study. Gastroenterology. 2001;121:255–60. 3. Langholz E, Munkholm P, Davidsen M, et al. Changes in extent of ulcerative colitis: a study on the course and prognostic factors. Scand J Gastroenterol. 1996;31:260–6. 4. Solberg IC, Vatn MH, Hoie O, et al. Clinical course in Crohn’s disease: results of a Norwegian population-based ten-year followup study. Clin Gastroenterol Hepatol. 2007;5:1430–8. 5. Jess T, Riis L, Vind I, et al. Changes in clinical characteristics, course, and prognosis of inflammatory bowel disease during the last 5 decades: a population-based study from Copenhagen, Denmark. Inflamm Bowel Dis. 2007;13:481–9. 6. Bitton A, Peppercorn MA, Antonioli DA, et al. Clinical, biological, and histologic parameters as predictors of relapse in ulcerative colitis. Gastroenterology. 2001;120:13–20. 7. Ferrante M, Vermeire S, Katsanos KH, et al. Predictors of early response to infliximab in patients with ulcerative colitis. Inflamm Bowel Dis. 2007;13:123–8. 8. Turner D, Leach ST, Mack D, et al. Faecal calprotectin, lactoferrin, M2-pyruvate kinase and S100A12 in severe ulcerative colitis: a prospective multicentre comparison of predicting outcomes and monitoring response. Gut. 2010;59:1207–12. 9. Anderson CA, Boucher G, Lees CW, et al. Meta-analysis identifies 29 additional ulcerative colitis risk loci, increasing the number of confirmed associations to 47. Nat Genet. 2011;43:246–52. 10. Arimura Y, Isshiki H, Onodera K, et al. Characteristics of Japanese inflammatory bowel disease susceptibility loci. J Gastroenterol. 2014;49:1217–30. 11. Jostins L, Ripke S, Weersma RK, et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature. 2012;491:119–24.
123
12. Beaugerie L, Sokol H. Clinical, serological and genetic predictors of inflammatory bowel disease course. World J Gastroenterol. 2012;18:3806–13. 13. Haritunians T, Taylor KD, Targan SR, et al. Genetic predictors of medically refractory ulcerative colitis. Inflamm Bowel Dis. 2010;16:1830–40. 14. Schachter H, Kirsner JB. Definitions of inflammatory bowel disease of unknown etiology. Gastroenterology. 1975;68:591–600. 15. Seo M, Okada M, Maeda K, Oh K. Correlation between endoscopic severity and the clinical activity index in ulcerative colitis. Am J Gastroenterol. 1998;93:2124–9. 16. Asano K, Matsushita T, Umeno J, et al. A genome-wide association study identifies three new susceptibility loci for ulcerative colitis in the Japanese population. Nat Genet. 2009;41:1325–9. 17. de Bakker PI, McVean G, Sabeti PC, et al. A high-resolution HLA and SNP haplotype map for disease association studies in the extended human MHC. Nat Genet. 2006;38:1166–72. 18. Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet. 2001;68:978–89. 19. Zhou B, Fine J, Laird G. Goodness-of-fit test for proportional subdistribution hazards model. Stat Med. 2013;32:3804–11. 20. Ahmad T, Armuzzi A, Neville M, et al. The contribution of human leucocyte antigen complex genes to disease phenotype in ulcerative colitis. Tissue Antigens. 2003;62:527–35. 21. Lappalainen M, Halme L, Turunen U, et al. Association of IL23R, TNFRSF1A, and HLA-DRB1*0103 allele variants with inflammatory bowel disease phenotypes in the Finnish population. Inflamm Bowel Dis. 2008;14:1118–24. 22. Annese V, Piepoli A, Latiano A, et al. HLA-DRB1 alleles may influence disease phenotype in patients with inflammatory bowel disease: a critical reappraisal with review of the literature. Dis Colon Rectum. 2005;48:57–64 (discussion 64-55). 23. Potocnik U, Ferkolj I, Glavac D, Dean M. Polymorphisms in multidrug resistance 1 (MDR1) gene are associated with refractory Crohn disease and ulcerative colitis. Genes Immun. 2004;5:530–9. 24. Daniel F, Loriot MA, Seksik P, et al. Multidrug resistance gene-1 polymorphisms and resistance to cyclosporine A in patients with steroid resistant ulcerative colitis. Inflamm Bowel Dis. 2007;13:19–23. 25. Ng SC, Tsoi KK, Kamm MA, et al. Genetics of inflammatory bowel disease in Asia: systematic review and meta-analysis. Inflamm Bowel Dis. 2012;18:1164–76. 26. Kostic AD, Xavier RJ, Gevers D. The microbiome in inflammatory bowel disease: current status and the future ahead. Gastroenterology. 2014;146:1489–99. 27. Billiet T, Ferrante M, Van Assche G. The use of prognostic factors in inflammatory bowel diseases. Curr Gastroenterol Rep. 2014;16:416. 28. Henriksen M, Jahnsen J, Lygren I, et al. C-reactive protein: a predictive factor and marker of inflammation in inflammatory bowel disease. Results from a prospective population-based study. Gut. 2008;57:1518–23. 29. Rosenfeld G, Bressler B. The truth about cigarette smoking and the risk of inflammatory bowel disease. Am J Gastroenterol. 2012;107:1407–8. 30. Maher MM, Nassar MI. Acute cytomegalovirus infection is a risk factor in refractory and complicated inflammatory bowel disease. Dig Dis Sci. 2009;54:2456–62. 31. Hoie O, Wolters F, Riis L, et al. Ulcerative colitis: patient characteristics may predict 10-year disease recurrence in a European-wide population-based cohort. Am J Gastroenterol. 2007;102:1692–701. 32. Dignass A, Eliakim R, Magro F, et al. Second European evidence-based consensus on the diagnosis and management of
J Gastroenterol
33.
34.
35. 36.
ulcerative colitis part 1: definitions and diagnosis. J Crohns Colitis. 2012;6:965–90. Weizman AV, Silverberg MS. Have genomic discoveries in inflammatory bowel disease translated into clinical progress? Curr Gastroenterol Rep. 2012;14:139–45. Charbonneau B, Wang AH, Maurer MJ, et al. CXCR5 polymorphisms in non-Hodgkin lymphoma risk and prognosis. Cancer Immunol Immunother. 2013;62:1475–84. Takai T. Roles of Fc receptors in autoimmunity. Nat Rev Immunol. 2002;2:580–92. Canete JD, Suarez B, Hernandez MV, et al. Influence of variants of Fc gamma receptors IIA and IIIA on the American College of Rheumatology and European League Against Rheumatism
responses to anti-tumour necrosis factor alpha therapy in rheumatoid arthritis. Ann Rheum Dis. 2009;68:1547–52. 37. Montes A, Perez-Pampin E, Narvaez J, et al. Association of FCGR2A with the response to infliximab treatment of patients with rheumatoid arthritis. Pharmacogenet Genomics. 2014;24: 238–45. 38. Libioulle C, Louis E, Hansoul S, et al. Novel Crohn disease locus identified by genome-wide association maps to a gene desert on 5p13.1 and modulates expression of PTGER4. PLoS Genet. 2007;3:e58.
123
ORIGINAL ARTICLE—ALIMENTARY TRACT
Contribution of susceptibility variants at FCGR2A and 13q12 to the risk of relapse among Japanese patients with ulcerative colitis Kouichi Asano • Motohiro Esaki • Junji Umeno • Atsushi Hirano • Yuji Maehata • Tomohiko Moriyama • Shotaro Nakamura • Takayuki Matsumoto • Takanari Kitazono
Received: 15 October 2014 / Accepted: 25 February 2015 Ó Springer Japan 2015
Abstract Background Recent genome-wide association studies have identified nearly 100 susceptibility genes for ulcerative colitis (UC). However, the contribution of susceptibility variants for UC to clinical outcome has scarcely been reported. The aim of this study was to investigate whether UC-associated genetic variants confer a risk of clinical relapse. Methods One hundred and nine consecutive Japanese subjects with quiescent UC were recruited. Four genetic variants of HLA-DRB1*1502, rs6671847 at FCGR2A, rs17085007 at 13q12, and rs2108225 at SLC26A3 were genotyped by Invader assay. The clinical courses were followed after blood sampling, and the risk of relapse according to these genotypes was calculated by Cox proportional hazard model. K. Asano (&) M. Esaki J. Umeno A. Hirano Y. Maehata T. Moriyama T. Kitazono Department of Medicine and Clinical Science, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan e-mail: [email protected]
Results During the mean follow-up period of 35 months (range 1–81 months), 49 of 109 subjects (45 %) relapsed. Carriers of the G allele of rs6671847 showed an increased risk of relapse compared with non-carriers [adjusted hazard ratio (HR), 2.27; 95 % confidence interval (CI), 1.20–4.32; p = 0.01]. Patients with the CT or TT genotypes of rs17085007 also had an increased risk of relapse compared to subjects with the CC genotype (for CT: adjusted HR, 2.16; 95 % CI, 1.10–4.23; p = 0.03; for TT: adjusted HR, 3.25; 95 % CI, 1.18–8.95; p = 0.02). These two risk variants multiplied the risk of relapse by 2.74 times (95 % CI, 1.10–4.23; p = 0.01) in patients with one risk genotype and 5.40 times (95 % CI, 2.06–14.13; p = 0.0006) in patients with both risk genotypes. Conclusions Genetic variants of rs6671847 at FCGR2A and rs17085007 at 13q12 conferred a risk of relapse in patients with UC. Keywords Ulcerative colitis Susceptibility variants FCGR2A Risk of relapse Clinical course
Introduction K. Asano Department of Endoscopic Diagnosis and Therapeutics, Kyushu University Hospital, Fukuoka, Japan A. Hirano Department of Internal Medicine, Yamaguchi Red Cross Hospital, Yamaguchi, Japan S. Nakamura Department of Advanced Medical Initiatives, Faculty of Medical Sciences, Kyushu University, Fukuoka, Japan T. Matsumoto Division of Gastroenterology and Hepatology, Department of Internal Medicine, Iwate Medical University, Morioka, Japan
Ulcerative colitis (UC) is a chronic remittent inflammatory bowel disease [1], and its natural history and the response to treatments vary widely among patients [2]. In clinical practice, about a quarter of patients with UC who begin steroid therapy become steroid dependent [2], and 7–20 % of those patients are obliged to receive surgery 5 years after the initial diagnosis because of their refractory disease or the development of dysplasia or cancer [3]. On the other hand, half of patients with UC have a favorable clinical course even after the initiation of steroid therapy. Further, approximately 20 % of newly diagnosed patients have
123
J Gastroenterol
been estimated to experience a quiescent disease course without requiring medical therapy to maintain remission [4, 5]. Therefore, in order to improve the clinical outcome in UC, it is crucial to identify factors predictive of the natural history and treatment response. Clinical, serological, and genetic markers associated with clinical outcome have been rigorously investigated, and a variety of predictive factors have been proposed so far. Younger age at onset, extent of the disease, smoking cessation, and multiple previous relapses have been shown to be clinical predictors of earlier relapse [6]. Biomarkers, including serum C-reactive protein (CRP), anti-neutrophil cytoplasmic antibody [7], and fecal calprotectin [8] also have been possible candidates, but these predictive factors can vary during the course of the disease, making them less reliable candidates. Therefore, more accurate predictive markers need to be determined in order to implement more effective, personalized treatment. Recent genome-wide association studies (GWAS) and subsequent meta-analyses have identified approximately 100 susceptibility genes for UC [9–11]. These susceptibility genes should help to unravel complex components in the pathogenesis of UC, such as epithelial barrier function, regulation of adaptive immunity, endoplasmic reticulum stress, and host–microbial interactions. In addition, recent genetic discoveries have the potential to change clinical strategies by predicting the natural history and guiding the choice of therapy. In fact, genetic markers are very attractive candidates for predictive factors due to their invariability and low invasiveness [12]. However, little is known about the influence of genetic factors on disease phenotypes and clinical outcomes in UC [13]. We thus aimed to investigate whether UC-associated genetic variants confer a risk of clinical relapse.
Materials and methods Samples Between May 2006 and February 2008, we collected clinical data and whole-blood DNA samples of 129 consecutive Japanese subjects with quiescent UC at Kyushu University Hospital. Twenty subjects were excluded because of their insufficient clinical information, and thus the present study comprised 109 subjects. The diagnosis of UC was made by expert gastroenterologists in accordance with international criteria defined by clinical, endoscopic, and histological features [14]. All patients with quiescent UC fulfilled the following criteria within the 3 months prior to study entry: (1) maintenance of a Seo’s ulcerative colitis activity index (UCAI) of less than 120 [15], (2) no
123
remission induction therapy, including cyclosporine, infliximab, adalimumab, tacrolimus, or cytapheresis therapy, and (3) use of systemic corticosteroid of less than 20 mg/day, which was defined as low-dose corticosteroid. Study follow-up and outcome measurements The medical records were carefully reviewed, and clinical information regarding disease activity, medical treatment, and laboratory data until December 2012 were obtained for the present analyses. During follow-up, each patient received regular clinical evaluation every 1 or 2 months, and disease relapse was determined to have occurred when any remission induction therapy was required to reduce disease activity (UCAI [ 120). In the present study, we assessed colectomy as the primary outcome and disease relapse as the surrogate primary outcome. Genotyping We genotyped four genetic variants (HLA-DRB1*1502, rs6671847 at FCGR2A, rs17085007 at 13q12, and rs2108225 at SLC26A3) that had been shown to have robust associations with UC susceptibility in Japanese [16] as well as in a European population [11]. These variants were successfully genotyped in all samples using an Invader assay. To tag HLA-DRB1*1502, we used a two-marker haplotype—i.e., a G-T haplotype comprising two SNPs, rs3135365 and rs7382794, which corresponds to HLADRB1*1502. This tag was chosen on the basis of the Wellcome Trust Sanger Institute data (http://www.sanger. ac.uk/HGP/Chr6/ng2006-data) [17]. Estimation of haplotype in each individual was conducted using PHASE software [18]. To test how well these tags worked to predict HLA status, we conducted high-resolution (four digits or more) typing of HLA-DRB1 in 191 samples in the previous GWAS [16] using a sequence-based typing method. Sequence data were collected and analyzed with an ABI 3730 DNA Analyzer (Applied Biosystems) and Assign software (Conexio Genomics, Applecross, Western Australia). The results yielded a correct prediction rate of 97.4 % (sensitivity, 98.7 %; specificity, 96.4 %). Statistics Relapse-free probabilities according to various genotypes were estimated using the Kaplan–Meier method, and the probabilities between the two groups were compared using a log-rank test. Comparisons of the risk of relapse between two or more groups and estimation of the hazard ratio (HR) and 95 % confidence interval (CI) with consideration for
J Gastroenterol
multiple confounding factors were carried out using a multivariate Cox proportional hazards model. We confirmed the proportionality of hazards in standard Cox models by assessing with the Schoenfeld goodness of fit test [19]. We compared the risk of each clinical characteristic listed in Table 1 between patients with relapse and those without using age- and sex-adjusted models, and clinical characteristics with a p value of \0.3 were initially chosen as potential risk factors for relapse. Consequently, male sex, age at sampling, disease duration, history of cytomegalovirus (CMV) infection, the duration between last flare-up and baseline, previous corticosteroid use, and current drugs including thiopurines, aminosalicylates, and methotrexate were determined to be confounding factors by the multivariate Cox regression analyses. Age at onset, previous treatment of cytapheresis and current drugs including thiopurine and probiotics were eliminated because
Table 1 Baseline characteristics according to relapse status
several factors were correlated with each other in the present analyses. CMV infection was defined to be positive by the detection of positive CMV antigenemia assay (CMVC7HRP) or by histological confirmation of inclusion body in biopsy specimens under H&E or immunohistochemical staining. All tests were two-sided, and a p value of \0.05 was considered to be statistically significant. For general statistical analysis, we used the R statistical environment version 2.14.1 (cran.r-project.org).
Ethical considerations The ethical committees of Kyushu University approved this study, and written informed consent was obtained from all subjects.
Characteristics
Non-relapse N = 60
Relapse N = 49
P**
Male sex
22 (36.7)
24 (49)
0.08
Age at onseta (years)
28 (8–69)
29 (13–69)
0.14
Age at samplinga
36 (14–60)
35 (15–74)
0.29
Disease durationa (years)
4.5 (0.5–36)
4.0 (0.5–45)
Disease type Proctitis
11 (18.3)
3 (6.1)
Left-sided
16 (26.7)
15 (30.6)
Extensive
31 (51.7)
29 (40.8)
Segmental
2 (3.3)
2 (4.1)
Family history
7 (11.7)
2 (4.1)
Smoking status Never smoker
0.14 0.72
0.61 0.91
49 (81.7)
40 (81.3)
Former smoker
5 (8.3)
4 (8.2)
Current smoker
6 (10.0)
5 (10.4)
C-reactive protein [0.1 mg/dl
22 (36.7)
20 (40.8)
0.58
Time since last flare-upa (months)
26.1 (3–252)
5.4 (3–110)
0.0004
History of CMV infection
8 (13.3)
8 (16.3)
0.12
39 (65.0)
44 (89.8)
0.02
Previous therapy Corticosteroids
Time since last flare-up was defined as the duration between the last flare-up and baseline. Low-dose corticosteroid was defined as less than 20 mg/day CMV cytomegalovirus ** p values were calculated by Cox proportional hazard model a
Data are summarized as median and range (min–max)
Cyclosporine
5 (8.3)
5 (10.2)
0.32
Tacrolimus
1 (1.7)
1 (2.0)
0.54
Cytapheresis
14 (23.3)
14 (28.6)
0.15
13 (21.7)
16 (33.3)
0.13
Current therapy Thiopurines Aminosalicylates
49 (81.7)
47 (95.9)
0.02
Low dose corticosteroids Methotrexate
14 (23.3) 0 (0)
17 (34.7) 5 (10.4)
0.03 0.0002
Topical therapies
5 (8.3)
4 (8.3)
0.98
Probiotics
22 (36.7)
22 (45.8)
0.03
123
J Gastroenterol
Results Clinical characteristics according to relapse status Among our 109 subjects, only five patients underwent colectomy during the mean follow-up period of 71.8 months, thus failing to conduct statistical analysis concerning genetic association with colectomy. We subsequently assessed a possible association of genetic factors with the disease relapse. During the mean follow-up period of 35.1 (1–81) months, the disease relapse was determined to have occurred in 49 patients (45 %). The relapse rate at 2 years after study entry was estimated to be 32 % (95 % CI, 22–41 %) by the Kaplan–Meier method. Baseline characteristics according to the relapse status are shown in Table 1. In regard to previous treatment, 86 patients (76 %) received systemic corticosteroids, but no patient received anti-TNF agents. When analyzing the risk of relapse according to baseline characteristics, the time since last flare-up (the duration between the last flare-up and baseline) was significantly shorter in patients without relapse than in those with relapse (p = 0004). As for previous therapies, systemic corticosteroids were more frequently applied to patients with relapse than in those without relapse (p = 0.02). Similarly, current use of aminosalicylates (p = 0.02), low-dose corticosteroids (p = 0.03), methotrexate (p = 0.0002), and probiotics (p = 0.03) were more frequent in patients with relapse than in those without relapse. The proportion of males was slightly higher in patients with relapse than in those without relapse, but the difference did not reach the level of statistical significance (p = 0.08). Other characteristics were not different between patients with relapse and those without. Risk of relapse according to genotype In the present study, the minor allele frequencies (MAF) of HLA-DRB1*1502, rs6671847, rs17085007, and rs2108225 were 30.7, 11.9, 30.3, and 28.4 %, respectively (Table 2). These frequencies were equivalent to those of UC cases in our previous GWAS. First, we assessed relapse-free probability stratified by the four genetic variants (Fig. 1). The relapse-free rate was lower in carriers of the G allele of rs6671847 than in nonTable 2 Genotype and allele frequency of variants
SNP
Chr
carriers, but there was no subject with a homozygous minor allele of rs6671847. The relapse-free rate was lower in carriers of the T allele of rs17085007 than in non-carriers (p = 0.02), while no difference was observed in relapsefree rate between the CT and TT genotypes of rs17085007. On the other hand, no difference in relapse-free rate was observed when stratified by genotypes of either HLADRB1*1502 or rs2108225 at SLC26A3 (Fig. 2). Next, we performed Cox proportional analyses to assess the contribution of genetic variants to the disease relapse (Table 3). Consequently, carriage of the G allele of rs6671847 was found to be significantly associated with an increased risk of relapse compared with non-carriage in a sex- and age-adjusted model [hazard ratio (HR), 1.92; 95 % confidence interval (CI), 1.04–3.54; p = 0.04]. This association remained significant even after adjustment for multiple confounding factors (multivariate adjusted HR, 2.27; 95 % CI, 1.20–4.32; p = 0.01). The CT and TT genotypes of rs17085007 were associated with an increased risk of relapse compared to subjects with the CC genotype in a multivariate-adjusted model (for CT: adjusted HR, 2.16; 95 % CI, 1.10–4.23; p = 0.03; for TT: adjusted HR, 3.25; 95 % CI, 1.18–8.95; p = 0.02). No difference was observed in the effect size between the CT and TT genotypes. Together with the results of the Kaplan–Meier survival analysis, these results indicate, in our samples, the risk genotype for rs17085007 was carriage of the T allele. On the other hand, HLA-DRB1*1502 and rs2108225 at the SLC26A3 gene were not associated with disease relapse. Combined effects of genetic variants on the risk of relapse Since each genetic variant had a different effect on disease susceptibility, we assessed possible multiplicative effects of the two genetic variants of rs6671847 and rs17085007. The relapse-free rate (Fig. 1) was lower in carriers having at least one risk variant than in non-carriers (p = 0.01). As shown in Table 4, multivariate analyses demonstrated that the risk of relapse in patients with one of the two risk genotypes increased by 2.74 times (95 % CI, 1.10–4.23; p = 0.01) compared to that in patients without either risk genotype. Moreover, the value in patients with both risk genotypes increased by 5.40 times (95 % CI, 2.06–14.13;
Gene
Minor allele
UC risk allele
Genotypea 0 (%)
1 (%)
MAF (%) 2 (%)
MAF minor allele frequency a
Genotype; ‘‘0’’ is homozygous major alleles, and genotype ‘‘1’’ is heterozygous, and ‘‘2’’ is homozygous minor alleles
123
HLA-DRB1*1502
6
HLA-DRB1
*1502
*1502
45.9
46.8
7.3
30.7
rs6671847
1
FCGR2A
G
A
76.1
23.9
0
11.9
rs17085007 rs2108225
13 9
– SLC26A3
T A
T G
49.5 48.6
40.4 45.9
10.1 5.5
30.3 28.4
J Gastroenterol
Fig. 1 Relapse-free probability stratified by genotypes of four markers using the Kaplan–Meier method. a HLA-DRB1*1502, b rs6671847 at FCGR2A, c rs17085007 at 13q12, d rs2108225 at SLC26A3. The bold characters in each genotype indicate risk alleles for developing UC
p = 0.0006) compared to that in patients without either risk genotype.
Discussion The present study demonstrated genetic associations between variants at the FCGR2A and 13q loci and clinical relapse in UC. Along with the disease susceptibility, possible genetic associations with clinical phenotypes, disease course, or treatment response have been investigated in
UC. These studies found that HLA-DRB1*0103 was associated with extensive colitis, severe disease and requirement of colectomy, in addition to disease susceptibility [20–22]. In addition, MDR1 polymorphism, which has been widely studied in the context of steroid therapy, was suggested to be associated with refractoriness to corticosteroids [23]. This polymorphism has also been shown to be associated with a lack of response to cyclosporine in patients with steroid-refractory UC [24]. Haritunians et al. [13] recently performed an unbiased GWAS in which they compared candidate SNPs between refractory UC patients
123
J Gastroenterol
pathogenesis of UC [26]. Thus, replication studies in various ethnic populations and comprehensive analyses considering environment factors are needed, since no previous study has assessed the predictive values of genetic variants from the point of view of disease relapse in UC. Predictive markers that influence on the disease phenotype can vary by the difference in the setting of clinical outcome [27]. In the present study, we initially investigated a possible association of genetic factors with colectomy; however, only five patients underwent colectomy during the mean follow-up period of 71.8 months. Considering that all of the patients were carriers of at least one risk variant (rs6671847 or rs17085007), these variants can be possible predictors of intractable clinical course requiring colectomy, whereas we could not draw any conclusion owing to the small number of colectomy cases. We thus investigated possible effect of genetic factors on the disease relapse. It has previously been reported that clinical factors including serum CRP levels at baseline [28], duration from last flare-up [6], smoking status [29], and history of CMV infection [30] were associated with unfavorable clinical outcomes in UC, and thus these have been considered as potential risk factors for disease relapse. Of these clinical markers, the duration from last flare-up was the only factor significantly associated with disease relapse in the present study. The discrepant results for the other markers may have been due to differences in the clinical characteristics of the study subjects or in the study design between the present and previous studies. Indeed, the clinical characteristics, including sex, age at onset, disease duration, and extent of disease in our subjects were slightly different from those of the previous studies [6, 28, 29]. However, the relapse rate at 2 years of follow-up was calculated to be 32 % (95 % CI, 22–41 %) in the present study, which was roughly equivalent to the value of a Europe-wide
Fig. 2 Relapse-free probability stratified by number of risk genotypes using the Kaplan–Meier method. The relapse-free rate was lower in carriers having at least one risk variant than in non-carriers (p = 0.01)
requiring colectomy and controls with favorable response to medical treatments. Their results indicated that TNFSF15 (TL1A), IL-10 (1q32.1), and IL-12B (5q33.3) were relevant to the disease course, but not HLA alleles or polymorphisms at the FCGR2A, 13q, or SLC26A3 loci. In contrast, TNFSF15, IL-10, and IL-12B were not associated with relapse in our samples (TNFSF15, p = 0.11; IL10, p = 0.25; IL12B, p = 17, respectively). This fact implies that the genetic contribution of UC-associated variants to the disease course varies among ethnicities [25]. In addition, accumulative evidence suggest that environment factors, especially intestinal microflora, also can play important roles in disease relapse as well as in the Table 3 Cox proportional hazards analysis of the effect the four genetic variants on the risk of relapse
Variables
rs6671847
HR hazard ratio, CI confidence interval a
Adjusted for sex, age at sampling, disease duration, history of cytomegalovirus infection, time since last flareup, previous corticosteroids use, current drugs including thiopurines, aminosalicylates, methotrexate
123
n
Event
rs17085007
rs2108225
24
Multivariate analysisa
Univariate analysis HR
HLA-DRB1*1502 Bold characters are risk alleles for developing UC in the previous Japanese genome-wide association study [1]
Genotype
95 % CI
p
Ref
HR
95 % CI
p
00
50
Ref
01
51
21
0.66
0.37–1.19
0.17
0.68
0.36–1.29
0.24
11 AA
8 83
4 34
1.31 Ref
0.45–3.80
0.62
1.65 Ref
0.52–5.23
0.39
AG
26
15
1.75
0.95–3.21
0.07
2.27
1.20–4.32
0.01
GG
0
NA
NA
NA
CC
54
19
Ref
Ref
CT
44
24
1.92
1.05–3.50
0.03
2.16
1.10–4.23
0.03
TT
11
6
1.84
0.73–4.63
0.19
3.25
1.18–8.95
0.02
GG
53
27
Ref
AG
50
19
0.74
0.41–1.33
0.31
0.92
0.49–1.73
0.79
AA
6
3
0.94
0.29–3.10
0.92
1.41
0.40–4.98
0.59
Ref
J Gastroenterol Table 4 Cox proportional hazard analysis of combined effects of rs6671847 and rs17085007 on the risk of relapse Number of risk variant
n
Event
Multivariate analysisa
Univariate analysis HR
95 % CI
p
HR
95 % CI
p
0
42
12
Ref
1
53
29
2.48
1.26–4.87
0.01
2.74
Ref 1.27–5.91
0.01
2
14
8
2.81
1.15–6.87
0.02
5.40
2.06–14.13
0.0006
p for trend
0.006
p for trend
0.0003
HR hazard ratio, CI confidence interval a
Adjusted for sex, age at sampling, disease duration, history of cytomegalovirus infection, time since last flare-up, previous corticosteroids use, current drugs including thiopurines, aminosalicylates, methotrexate
population-based study (40 % at 2 years) [31]. It thus can be assumed that the clinical background of our subjects was not particularly biased when compared to those of the previous investigations. When discussing disease relapse in UC, medical treatments should also be considered. In the present study, current use of aminosalicylates, methotrexate, and probiotics, as well as previous use of systemic corticosteroids were associated with increased risk of disease relapse. On the other hand, no significant association was found between treatment by thiopurines and disease relapse. It may be possible that patients treated with these medications have more severe disease, which would result in a higher rate of disease relapse [31]. However, it seems more plausible that the peculiar association between the medical treatments and disease relapse was caused by inappropriate application of maintenance treatment to aggressive disease. Indeed, thiopurines have been considered to have more beneficial maintenance effects than 5-ASA, especially in refractory UC [32]. Considering the inverse association of medical treatment with disease relapse, the insignificant association of thiopurines with disease relapse might have been the result of their modest but beneficial maintenance effect in patients with severe or relapsing disease. FCGR2A is implicated to be a susceptibility gene for several autoimmune diseases. Although the risk allele for severe or relapsing disease is assumed to be the same as that for disease susceptibility [33], the carriage of the G allele at rs6671847 of the FCGR2A gene, which is considered to be protective for developing UC, was associated with a higher risk for disease relapse in the present study. For this paradoxical result, two possible explanations can be postulated. First, this polymorphism may exert one set of biological activities for disease onset and a different set for the rest of the disease course [34]. Second, it is known that this polymorphism also affects the response to medical treatment, thereby modifying the risk for disease relapse. Actually, rs6671847 is in strong linkage disequilibrium with FCGR2A-131H/R that can cause functional difference. The H131 amino acid has higher binding affinity for
immune complexes than R131, resulting in a hyperactivation of immune cells [35]. This hyperactivation includes a higher capacity for antigen presentation to T cells and activation of B cells, as well as the overproduction of cytokines, resulting in perpetual inflammation of the colorectal mucosa. In addition, FCGR2A-131H/R has been reported to be associated with response to infliximab among patients with rheumatoid arthritis (RA) [36, 37]. In these reports, patients with FCGR2A-R131, low affinity FCGR2A allele, have been shown to have better response to IFX. In the present study, since no patients received antiTNF therapy during the follow-up periods, we could not evaluate the effect FCGR2A-131H/R or rs6671847 on the response to anti-TNF therapy. However, it is possible that carriers of the A allele of rs6671847 at the FCGR2A gene (UC risk allele) are more sensitive to maintenance therapies including aminosalicylates and immunomodulators. Further examination to explore possible differences in the response to treatment according to FCGR2A genotypes is necessary. Rs17085007 is located in the non-coding regions on chromosome 13q12.13. The gene nearest to this region is USP12 (encoding ubiquitin-specific peptidase-12), while its function remains unclear. However, possible contribution of rs17085007 to disease relapse in UC can be postulated as shown in CD-associated polymorphisms at 5p13.1. That is to say, while this variant also lies in the non-coding region, the variant has been thought to be a cisregulatory element that modulates the expression level of prostaglandin E receptor 4 (PTGER4) approximately 270 kb away from the variant [38]. It is thus possible that the rs17085007 locus contains regulatory sequences. In the present study, several limitations should be noted. These include the limited number of study samples, the assessment only during specific periods of the disease course, and the retrospective nature of the study itself, any of which could cause a selection bias. Statistically, while the present study showed genetic contributions of the FCGR2A gene and 13q12 to disease relapse in UC, 95 % confidence intervals of the risk variants were rather broadly
123
J Gastroenterol
calculated in the present study, indicating the necessity of a larger study sample size. Moreover, the sensitivity and specificity of risk variants for disease relapse were comparatively low, and the area under the receiver operating characteristic curve was 0.63 even in the combined effects model using rs6671847 and rs17085007. Therefore, these genetic variations alone would seem to be insufficient for predicting the disease course in UC. However, since the risk variants had a multiplicative effect in predicting disease relapse in the present study, future investigations combining analyses of genetic variants and clinical and biological markers could yield clinically useful indices for improving the treatment strategies in UC. In conclusion, genetic variants of rs6671847 at FCGR2A and rs17085007 at 13q12 confer risk of clinical relapse in Japanese patients with UC. These findings may help identify patients who exhibit a relapsing disease course and thus require intensive maintenance treatments. Conflict of interest The authors declare that they have no competing financial interests.
References 1. Abraham C, Cho JH. Inflammatory bowel disease. N Engl J Med. 2009;361:2066–78. 2. Faubion WA Jr, Loftus EV Jr, Harmsen WS, et al. The natural history of corticosteroid therapy for inflammatory bowel disease: a population-based study. Gastroenterology. 2001;121:255–60. 3. Langholz E, Munkholm P, Davidsen M, et al. Changes in extent of ulcerative colitis: a study on the course and prognostic factors. Scand J Gastroenterol. 1996;31:260–6. 4. Solberg IC, Vatn MH, Hoie O, et al. Clinical course in Crohn’s disease: results of a Norwegian population-based ten-year followup study. Clin Gastroenterol Hepatol. 2007;5:1430–8. 5. Jess T, Riis L, Vind I, et al. Changes in clinical characteristics, course, and prognosis of inflammatory bowel disease during the last 5 decades: a population-based study from Copenhagen, Denmark. Inflamm Bowel Dis. 2007;13:481–9. 6. Bitton A, Peppercorn MA, Antonioli DA, et al. Clinical, biological, and histologic parameters as predictors of relapse in ulcerative colitis. Gastroenterology. 2001;120:13–20. 7. Ferrante M, Vermeire S, Katsanos KH, et al. Predictors of early response to infliximab in patients with ulcerative colitis. Inflamm Bowel Dis. 2007;13:123–8. 8. Turner D, Leach ST, Mack D, et al. Faecal calprotectin, lactoferrin, M2-pyruvate kinase and S100A12 in severe ulcerative colitis: a prospective multicentre comparison of predicting outcomes and monitoring response. Gut. 2010;59:1207–12. 9. Anderson CA, Boucher G, Lees CW, et al. Meta-analysis identifies 29 additional ulcerative colitis risk loci, increasing the number of confirmed associations to 47. Nat Genet. 2011;43:246–52. 10. Arimura Y, Isshiki H, Onodera K, et al. Characteristics of Japanese inflammatory bowel disease susceptibility loci. J Gastroenterol. 2014;49:1217–30. 11. Jostins L, Ripke S, Weersma RK, et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature. 2012;491:119–24.
123
12. Beaugerie L, Sokol H. Clinical, serological and genetic predictors of inflammatory bowel disease course. World J Gastroenterol. 2012;18:3806–13. 13. Haritunians T, Taylor KD, Targan SR, et al. Genetic predictors of medically refractory ulcerative colitis. Inflamm Bowel Dis. 2010;16:1830–40. 14. Schachter H, Kirsner JB. Definitions of inflammatory bowel disease of unknown etiology. Gastroenterology. 1975;68:591–600. 15. Seo M, Okada M, Maeda K, Oh K. Correlation between endoscopic severity and the clinical activity index in ulcerative colitis. Am J Gastroenterol. 1998;93:2124–9. 16. Asano K, Matsushita T, Umeno J, et al. A genome-wide association study identifies three new susceptibility loci for ulcerative colitis in the Japanese population. Nat Genet. 2009;41:1325–9. 17. de Bakker PI, McVean G, Sabeti PC, et al. A high-resolution HLA and SNP haplotype map for disease association studies in the extended human MHC. Nat Genet. 2006;38:1166–72. 18. Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet. 2001;68:978–89. 19. Zhou B, Fine J, Laird G. Goodness-of-fit test for proportional subdistribution hazards model. Stat Med. 2013;32:3804–11. 20. Ahmad T, Armuzzi A, Neville M, et al. The contribution of human leucocyte antigen complex genes to disease phenotype in ulcerative colitis. Tissue Antigens. 2003;62:527–35. 21. Lappalainen M, Halme L, Turunen U, et al. Association of IL23R, TNFRSF1A, and HLA-DRB1*0103 allele variants with inflammatory bowel disease phenotypes in the Finnish population. Inflamm Bowel Dis. 2008;14:1118–24. 22. Annese V, Piepoli A, Latiano A, et al. HLA-DRB1 alleles may influence disease phenotype in patients with inflammatory bowel disease: a critical reappraisal with review of the literature. Dis Colon Rectum. 2005;48:57–64 (discussion 64-55). 23. Potocnik U, Ferkolj I, Glavac D, Dean M. Polymorphisms in multidrug resistance 1 (MDR1) gene are associated with refractory Crohn disease and ulcerative colitis. Genes Immun. 2004;5:530–9. 24. Daniel F, Loriot MA, Seksik P, et al. Multidrug resistance gene-1 polymorphisms and resistance to cyclosporine A in patients with steroid resistant ulcerative colitis. Inflamm Bowel Dis. 2007;13:19–23. 25. Ng SC, Tsoi KK, Kamm MA, et al. Genetics of inflammatory bowel disease in Asia: systematic review and meta-analysis. Inflamm Bowel Dis. 2012;18:1164–76. 26. Kostic AD, Xavier RJ, Gevers D. The microbiome in inflammatory bowel disease: current status and the future ahead. Gastroenterology. 2014;146:1489–99. 27. Billiet T, Ferrante M, Van Assche G. The use of prognostic factors in inflammatory bowel diseases. Curr Gastroenterol Rep. 2014;16:416. 28. Henriksen M, Jahnsen J, Lygren I, et al. C-reactive protein: a predictive factor and marker of inflammation in inflammatory bowel disease. Results from a prospective population-based study. Gut. 2008;57:1518–23. 29. Rosenfeld G, Bressler B. The truth about cigarette smoking and the risk of inflammatory bowel disease. Am J Gastroenterol. 2012;107:1407–8. 30. Maher MM, Nassar MI. Acute cytomegalovirus infection is a risk factor in refractory and complicated inflammatory bowel disease. Dig Dis Sci. 2009;54:2456–62. 31. Hoie O, Wolters F, Riis L, et al. Ulcerative colitis: patient characteristics may predict 10-year disease recurrence in a European-wide population-based cohort. Am J Gastroenterol. 2007;102:1692–701. 32. Dignass A, Eliakim R, Magro F, et al. Second European evidence-based consensus on the diagnosis and management of
J Gastroenterol
33.
34.
35. 36.
ulcerative colitis part 1: definitions and diagnosis. J Crohns Colitis. 2012;6:965–90. Weizman AV, Silverberg MS. Have genomic discoveries in inflammatory bowel disease translated into clinical progress? Curr Gastroenterol Rep. 2012;14:139–45. Charbonneau B, Wang AH, Maurer MJ, et al. CXCR5 polymorphisms in non-Hodgkin lymphoma risk and prognosis. Cancer Immunol Immunother. 2013;62:1475–84. Takai T. Roles of Fc receptors in autoimmunity. Nat Rev Immunol. 2002;2:580–92. Canete JD, Suarez B, Hernandez MV, et al. Influence of variants of Fc gamma receptors IIA and IIIA on the American College of Rheumatology and European League Against Rheumatism
responses to anti-tumour necrosis factor alpha therapy in rheumatoid arthritis. Ann Rheum Dis. 2009;68:1547–52. 37. Montes A, Perez-Pampin E, Narvaez J, et al. Association of FCGR2A with the response to infliximab treatment of patients with rheumatoid arthritis. Pharmacogenet Genomics. 2014;24: 238–45. 38. Libioulle C, Louis E, Hansoul S, et al. Novel Crohn disease locus identified by genome-wide association maps to a gene desert on 5p13.1 and modulates expression of PTGER4. PLoS Genet. 2007;3:e58.
123
