HLA-DQA1*05 Associates with Extensive Ulcerative Colitis at Diagnosis: An Observational Study in Children

The human leukocyte antigen (HLA) allele group HLA-DQA1*05 predisposes to ulcerative colitis (UC) and is associated with the development of antibodies against infliximab in patients with inflammatory bowel disease (IBD). Therefore, we hypothesized that the presence of HLA-DQA1*05 correlates with characteristics of pediatric IBD. Within a multi-center cohort in Poland, the phenotype at diagnosis and worst flare was established and HLA-DQA1*05 status was assessed enabling genotype-phenotype analyses. HLA-DQA1*05 was present in 221 (55.1%) out of 401 children with IBD (UC n = 188, Crohn’s disease n = 213). In UC, the presence of HLA-DQA1*05 was moderately associated with a large extent of colonic inflammation at diagnosis (E4 55% more frequent in HLA-DQA1*05-positive patients, p = 0.012). PUCAI at diagnosis (p = 0.078) and the time from UC diagnosis to the first administration of biologic treatment (p = 0.054) did not differ depending on HLA-DQA1*05 status. The number of days of hospitalization for exacerbation was analyzed in 98 patients for whom sufficient follow-up was available and did not differ depending on HLA-DQA1*05 carriership (p = 0.066). HLA-DQA1*05 carriers with CD were less likely to present with both stenosing and penetrating disease (B2B3, p = 0.048) and to have active disease proximal to the ligament of Treitz (L4a) at the worst flare (p = 0.046). Future research focusing on explaining and preventing anti-TNF immunogenicity should take into account that ADA may develop not only as an isolated reaction to anti-TNF exposure but also as a consequence of intrinsic differences in the early course of UC.


Introduction
Inflammatory bowel diseases (IBD) are life-long conditions affecting up to 0.5-1% of populations in developed countries. Patients typically experience diarrhea and abdominal pain, often requiring treatment with anti-tumor necrosis factor (anti-TNF) biologics. Infliximab has become a key IBD medication globally, with overall spending reaching billions of dollars. However, most patients develop anti-infliximab antibodies that may lead to treatment inefficacy.
The PANTS consortium recently showed that the allele group HLA-DQA1*05 associates with the development of antibodies against infliximab in patients with inflammatory bowel diseases (IBD) [1]. Since HLA-DQA1*05 is carried by 20-40% of Europeans, it is an actionable target providing an accessible foundation for personalized medicine in IBD. Compared with other nascent biomarkers derived from omics research [2], HLA-DQA1*05 assessment is potentially much simpler and cheaper. No randomized trials have yet been conducted but it seems that the presence of HLA-DQA1*05 could prompt more aggressive treatment, e.g., with a combination of infliximab and an immunosuppressant, or a different biologic.
Although the link between HLA-DQA1*05 and anti-drug antibody (ADA) development is statistically significant (rs2097432; HR 1.90, p = 5.88 × 10 −13 ), the mechanisms mediating this effect remain uncertain [3]. Moreover, HLA-DQA1*05 is representative of a group of allelic variants, which is associated with the development of antibodies against infliximab or adalimumab, therefore four-digit HLA analysis could allow for even more detailed predictions [4,5].
Since HLA-DQA1*05 relates to celiac disease (and Grave's disease), it can be argued that some of the pathophysiological mechanisms are shared between celiac autoimmunity and the generation of ADA [6]. Moreover, celiac disease (or Grave's disease) is much less common in IBD patients than ADA development, so it can also be speculated that HLA-DQA1*05 facilitates the formation of ADA in a process dependent on IBD. The link between HLA-DQA1*05 and IBD characteristics can be explored in a setting free from potential confounding by anti-TNF agents, that is, at diagnosis, thus in childhood and adolescence. It was hypothesized that the presence of HLA-DQA1*05 correlates with greater severity of ulcerative colitis (UC) and Crohn's disease (CD) at diagnosis, more rapid onset, and earlier need for immunosuppressive or biological treatment.

Patients and Setting
Patients for this cross-sectional study were recruited between 2016-2019 from seven Polish pediatric gastroenterology centers: Poznan University of Medical Sciences (Poznań, Poland), The Children's Memorial Health Institute (Warsaw, Poland), Medical University of Warsaw (Warsaw, Poland), Wroclaw Medical University (Wrocław, Poland), Medical University of Silesia (Zabrze and Katowice, Poland), and Nicolaus Copernicus University (Bydgoszcz, Poland). The diagnosis of IBD was established using standard diagnostic criteria [7,8]. The inclusion criteria involved diagnosis of CD or UC as well as age ≥ 3 years and ≤18 years. Patients in life-threatening, severe general condition were excluded from the study. All the patients' parents or guardians provided informed written consent and the study was approved by the Bioethical Committee of Poznan University of Medical Sciences (960/15 with the associated amendments) and adheres to the revised Declaration of Helsinki.

Disease Severity Evaluation
Disease activity was assessed using the Pediatric Ulcerative Colitis Activity Index (PUCAI) or Pediatric Crohn's Disease Activity Index (PCDAI) at diagnosis and worst flare [9]. The worst flare was defined with the highest PUCAI or PCDAI result in the medical history. Serum albumin and C-reactive protein (CRP) concentrations at diagnosis and worst flare were obtained from medical records. Data were collected regarding systemic steroid intake with the total number of courses, immunosuppressive treatment with the age at first intake, biological therapy with the time at first infusion, and operative treatment with the age at first surgery. Localization and behavior of the disease were defined by the Paris Classification at diagnosis and worst flare [10] but these data were missing for some hospitalizations. Based on medical records, the total number of hospitalizations triggered by exacerbation was counted, with exacerbation defined by PUCAI ≥ 65 or PCDAI ≥ 50. The number of days spent in hospital for exacerbations was summarized for patients with a disease duration of at least one year. These exacerbation-related metrics were then divided by the number of years since diagnosis to obtain an individual estimate of severity.

Genotyping
DNA was isolated from whole blood using Blood Mini (A&A Biotechnology, Gdynia, Poland). The primers proposed by Profaizer et al. [11] were used: forward 5 CAC GTC GCC TCT TAT GGT GTA; reverse 5 GAC TGT TCA AGT TAT GTT TTA GG. Reactions were prepared as follows: 1 µL of each specific primer (10 mM), 1 µL of water, 5 µL of FastStart Universal SYBR Green Master Mix (Sigma Aldrich, St. Louis, MO, USA) and 2 µL of genomic DNA for a total of 10 µL per reaction. The genotyping was performed using the Bio-Rad CFX96 thermocycler (Hercules, CA, USA) and the following thermo-profile: 95 • C for 5 min, 35 cycles of 10 s at 95 • C, followed by 30 s at 50 • C, 10 s at 72 • C with signal acquisition. The melting profile started with 5 min at 72 • C, then the temperature was increased from 65-95 • C for PCR product verification. Melting peaks representing the characteristic Tm of each particular DNA product were analyzed by CFX96 BioRad software. Inconclusive melting profiles, along with index positive and negative results, were sequenced (Sanger) to resolve uncertainties and serve as controls, respectively. As Sazonovs et al. demonstrated the relevance of HLA-DQA1*05 for ADA development in the dominant model, this work focused on the presence of HLA-DQA1*05, regardless of hetero-or homozygosity [1].

Statistical Analysis
The data were analyzed using Statistica 13.3 (TIBCO, Palo Alto, CA, USA) and R (R Software Foundation, Vienna, Austria). Z-scores for anthropometric parameters were calculated using reference values for the local population. The data are presented as mean ± mean or percentage, unless specified otherwise. The Mann-Whitney U test was used to compare continuous variables. The Fisher's exact test was used for the comparison of binary parameters. A forward stepwise regression was built to compensate for confounding. The significance threshold was set at 0.05.

Cohort Characteristics
HLA-DQA1*05 status was assessed in 401 children with IBD, including 188 with UC and 213 with CD ( Table 1). The typical patient with IBD was diagnosed at 7-16 years and followed up for 3 years on average. Children with UC were younger and had better nutritional status at diagnosis compared to CD, as well as exhibiting lower CRP levels both at diagnosis and worst flare (Table 1). Most IBD patients received steroids and immunosuppressants. However, UC was more frequently treated with systemic steroids while azathioprine was used more often in CD. A greater fraction of CD patients received biological therapy or needed to undergo IBD-related surgery. Most children (55.1%) carried the HLA-DQA1*05 allele group.

HLA-DQA1*05 Associates with Extensive Colitis at Diagnosis
In UC, the presence of HLA-DQA1*05 was associated with pancolitis at diagnosis (E4, which was 55% more frequent in HLA-DQA1*05-positive patients, p = 0.012; Table 2). However, pancolitis was also associated with the male gender (p = 0.038). In forward stepwise regression analysis of data from UC patients, HLA-DQA1*05 but not gender was weakly associated with Paris E4 status at diagnosis (R2 = 0.03, p model = 0.0077; β = 0.1939, 95% CI 0.0520-0.3358). There was no other evidence of a more severe UC course at diagnosis in children positive for HLA-DQA1*05. The time from UC diagnosis to first administration of immunosuppressive (p = 0.133) or biologic treatment (p = 0.054) did not differ depending on HLA-DQA1*05 status. HLA-DQA1*05 patients also did not have lower mass, height (p = 0.060) or BMI at diagnosis, PUCAI at diagnosis (p = 0.078) or worst flare and had CRP levels similar to HLA-DQA1*05-negative peers both at diagnosis and worst flare. The number of days of hospitalization for exacerbation was analyzed in 98 patients with UC for whom sufficient follow-up was available and did not differ depending on HLA-DQA1*05 status (p = 0.066). Four patients with UC received immunosuppressants prior to diagnosis. All were HLA-DQA1*05 positive and half of them developed pancolitis at diagnosis. HLA-DQA1*05 carriers with CD were less likely to present with both stenosing and penetrating disease (B2B3, p = 0.048) and have active disease proximal to the ligament of Treitz (L4a) at worst flare (p = 0.046; Table 3). Nominally more frequent perianal involvement was not significantly associated with HLA-DQA1*05 (13.4% vs. 5.0%, p = 0.054) or age at first CD surgery (13.2 ± 2.2 years vs. 14.6 ± 3.3 years, p = 0.051). No other relationships were found between HLA-DQA1*05 and the clinical characteristics of CD.

Discussion
This study explored the clinical correlation of carrying the HLA-DQA1*05 allele group in 401 children and adolescents with IBD focusing on disease severity, age at onset, and the time to initiation of immunosuppressive or biologic treatment. HLA-DQA1*05 positively associated with the broad extent of colonic inflammation at diagnosis of UC. Therefore, HLA-DQA1*05 seems relevant for UC beyond ADA generation. It may be speculated that more extensive colitis at diagnosis, which is associated with HLA-DQA1*05, predisposes to ADA development.

Gender and the Relationship between HLA-DQA1*05 and Colitis Extent
The observation that males were over 40% more frequently HLA-DQA1*05-positive than females leads to the question of whether the main finding can be confounded by gender. As indicated by regression analysis, the link between HLA-DQA1*05 and the extent of colitis proved stronger than with gender. Moreover, the male gender is a known factor for a severe course of UC, therefore the larger percentage of men in the E4 group is unsurprising [12]. However, the interaction between gender and HLA-DQA1*05 was not excluded in this study, so one could speculate that there is an interaction between gender and HLA-DQA1*05 in IBD development.

Celiac Disease and the Relationship between HLA-DQA1*05 and the Extent of Colitis
Five patients with UC had celiac disease. Of these five patients four carried HLA-DQA1*05. Pancolitis (E4) was found in only one of these children suggesting that the relationship between HLA-DQA1*05 and more severe colitis at diagnosis is not dependent on factors directly predisposing to celiac disease. However, any link between celiac disease and the occurrence of ADA would need to be explored separately.

HLA-DQA1*05 Effect in UC, but Not CD
Sazonovs et al. initially reported the link between HLA-DQA1*05 in CD, later confirming the association in both CD and UC within a validation sub-study, revealing similar hazard ratios for both diseases (HR = 2.26 in CD; HR = 2.02 in UC) [5]. However, our study identified a relationship between disease severity and HLA-DQA1*05 only in UC, despite a greater number of patients with CD (over 200). This calls for hypotheses explaining the impact of HLA-DQA1*05 on UC but not CD severity at diagnosis. Molecular dissection of UC vs. CD is challenging because of the many similarities that these entities share in omics-based research. Genome-wide association studies found that the single nucleotide polymorphism rs6927022 located immediately downstream from HLA-DQA1 (1 kbp) strongly correlates with the development of UC but not CD [13]. In some analyses, rs6927022 was considered to be the genetic loci most strongly associated with UC [14].
It is also interesting that HLA-DQA1*05:01 remains in linkage disequilibrium with HLA-DRB1*03:01 (D' = 0.75 according to LDpair; European populations), which is associated with an increased risk of Crohn's disease in a study focusing on four-digit HLA typing by microarrays in over 30,000 IBD patients [15]. Yet, whereas HLA-DQA1*05 in numerous cases should predispose to disease occurrence and ADA formation in both UC and CD, UC alone predisposes to more severe disease.
Other UC-specific genetic predispositions include WNT2, a gene involved in epithelial proliferation and the modulation of inflammatory responses. Furthermore, some CDspecific genetic predispositions are related to bacterial antigen sensing (TLR4, NOD2), autophagy (ATG16L1), the killing of bacteria (NCF4), and the interferon response (IFNGR2). Despite the above relationships and several other differences, the HLA locus containing HLA-DQA1 and HLA-DQB1 seems to be key to the observed lack of a relationship with severity at onset in CD.

Strengths and Limitations of the Current Analysis
Apart from focusing on HLA-DQA1*05 without four-digit HLA allele genotyping, this study is limited by the lack of long-term data on the response to infliximab or ADA development, precluding validation of the PANTS study. Also, even though the dominant model has been established for the influence of HLA-DQA1*05 on ADA development, investigation of the additive model could also be of value. Nevertheless, the results presented in this work reveal an important relationship between HLA-DQA1*05 and UC extent at diagnosis. This was accomplished in a large cohort with rich phenotype data, providing a new dimension to our understanding of the relationship between HLA-DQA1*05 and anti-TNF immunogenicity.

The Cohort Captures a Transition Period in Infliximab Use in Polish Children
This study describes the clinical characteristics and treatment modalities of a large pediatric IBD cohort from Poland, covering a transition period in the availability of biologics for pediatric IBD. Initially, the access to therapeutic programs was rather limited, requiring high PCDAI scores for eligibility but the introduction of biosimilars gradually increased the availability of anti-TNF agents for this cohort. Regardless of the improved availability, the patients within the cohort were still treated using a step-up strategy, and this transition is unlikely to confound analyses of time from diagnosis to treatment with anti-TNF agents.
Another noteworthy aspect of the cohort is that patients with CD had poorer nutritional status and stronger inflammation at diagnosis compared to UC and that almost 80% were exposed to immunosuppressants relatively early.

New Research Directions in UC
The relationship between HLA-DQA1*05 and greater PUCAI (mean 46 vs. 40, p = 0.078) as well as lower height Z-score at UC diagnosis (mean −0.06 vs. 0.18, p = 0.060) would require additional investigation in even larger cohorts. Similarly, such research could be helpful to determine links between HLA-DQA1*05 and the shorter time from diagnosis to first biologic (mean 16.3 vs. 27.6 months, p = 0.054) and the greater number of days of hospitalization for exacerbation per year in UC (mean 9.22 vs. 5.45 days, p = 0.066). Our study shows that if such effects exist, they are likely subtle. As already discussed, there were no similar results in CD. Future studies could focus on targeting links between HLA genetics and CD location at diagnosis and worst flare, with and without the use of anti-TNF agents, to potentially uncover disease subtypes. Analysis of HLA-DQA1*05 allele homo-and heterozygosity in the context of polygenic risk scores could help elucidate the interactions between gender and HLA-DQA1*05-related pancolitis in UC.

Conclusions
In summary, HLA-DQA1*05 correlates with a greater extent of colonic inflammation at diagnosis in children with UC. Future research focusing on explaining and preventing anti-TNF immunogenicity should take into account that ADA may develop not only as an isolated reaction to anti-TNF exposure but as a consequence of intrinsic differences in the early course of UC.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data will be available from the corresponding author at reasonable request.