Ethnic divergence in human leukocyte antigen-linked immunogenicity in inflammatory bowel disease

Abstract

Inflammatory bowel disease (IBD), encompassing both Crohn’s disease and ulcerative colitis, poses significant treatment challenges due to its complex etiology and variable therapeutic responses. Anti-tumor necrosis factor biologic therapies have substantially reduced corticosteroid dependence, hospitalization rates, and the need for surgery in moderate-to-severe IBD, yet their clinical utility is frequently compromised by the development of anti-drug antibodies (ADAs) which reduce drug trough levels, diminish therapeutic response, and accelerate treatment failure. Concomitant immunomodulator co-therapy attenuates ADA formation by approximately 47%, but introduces competing risks that necessitate individualized benefit-risk assessment. While the genetic determinants of immunogenicity have been characterized extensively in European populations – most notably the human leukocyte antigen (HLA)-DQA1*05 allele – a clinically consequential gap persists in non-European ethnic groups. This opinion review critically evaluates the work, which identified HLA class I alleles HLA-C*03:04:01 and HLA-B*15:18:01 as ethnic-specific predictors of ADA formation in Taiwanese IBD patients, challenging the presumed universality of HLA-DQA1*05. We place these findings within the broader context of a shifting global IBD epidemiological burden, newly refined four-digit allele subtype evidence in European and East Asian populations, the mechanistic basis of HLA class I-driven immunogenicity, the complementary roles of immunomodulator co-therapy and therapeutic drug monitoring, and the broader precision pharmacogenomics agenda in IBD. We propose that pre-treatment ethnically calibrated HLA-based stratification, integrated with therapeutic drug monitoring-guided monitoring, represents the next frontier in biologic optimization and immunologic safety in IBD.

Key Words: Inflammatory bowel disease; Anti-drug antibodies; Human leukocyte antigen genotyping; Immunogenicity; Precision medicine; Therapeutic drug monitoring; Crohn’s disease; Ulcerative colitis

Core Tip: This opinion review contextualizes the landmark findings, which identified human leukocyte antigen (HLA)-C*03:04:01 and HLA-B*15:18:01 as ethnic-specific risk alleles for anti-drug antibody formation in Taiwanese inflammatory bowel disease (IBD) patients receiving infliximab and adalimumab, respectively, challenging the universality of HLA-DQA1*05 as a pharmacogenomic predictor of anti-tumor necrosis factor immunogenicity. We integrate these findings within: (1) The shifting global IBD epidemiology; (2) Four-digit allele subtype behavior in European and East Asian populations; (3) Mechanistic pathways including cross-presentation, peptide-groove structural specificity, and killer cell immunoglobulin-like receptor-mediated tolerance regulation; (4) The complementary roles of immunomodulator co-therapy and therapeutic drug monitoring; and (5) The broader precision pharmacogenomics agenda in IBD. Global multi-ethnic collaboration to validate allele-immunogenicity associations is an ethical and clinical imperative.

INTRODUCTION

The management of inflammatory bowel disease (IBD) has undergone a paradigm shift with the advent of biologic therapies, particularly tumor necrosis factor (TNF) inhibitors. Anti-TNF agents have substantially reduced corticosteroid dependence, hospitalization rates, and surgical requirements in patients with moderate-to-severe Crohn’s disease and ulcerative colitis[1,2]. Primary non-response, occurring in 10%-40% of patients, and secondary loss of response (LoR), affecting a further 20%-50% of initial responders, represent the principal pharmacological challenges in long-term IBD management[3-5]. While pharmacodynamic failure – driven by non-TNF-mediated inflammatory pathways[6] – accounts for a proportion of these failures, immune-mediated pharmacokinetic failure, defined by the development of anti-drug antibodies (ADAs) that neutralize and accelerate the clearance of anti-TNF monoclonal antibodies, is the most clinically important driver of LoR in the first years of therapy[2,7-9].

The landmark PANTS study[10] established that human leukocyte antigen (HLA)-DQA1*05 carriage, present in approximately 40% of Europeans, substantially increased ADA formation rates against both infliximab and adalimumab [hazard ratio (HR) = 1.90; 95%CI: 1.60-2.25], with 92% of HLA-DQA1*05 carriers treated with infliximab monotherapy developing ADAs by year one. Two independent meta-analyses confirmed this association across immune-mediated inflammatory diseases[11,12]. More recently, four-digit subtype-specific analyses[13,14] and East Asian cohort data[15] have refined our understanding, revealing that allele-specific ADA risk differs not only between European subtypes but diverges fundamentally across ethnic groups. Concomitant immunomodulator co-therapy is the principal pharmacological modifier of this risk, reducing ADA frequency by approximately 47% across immune-mediated inflammatory diseases[16], though at the cost of elevated infection and – with thiopurines – lymphoproliferative malignancy risks[17,18].

The study by Weng et al[19], represents a significant and timely contribution. By conducting high-resolution HLA genotyping via next-generation sequencing (NGS) in 95 Taiwanese IBD patients, the authors demonstrate that HLA-DQA1*05 – despite a carriage frequency of 25.3% – fails to predict ADA formation. Instead, two novel HLA class I alleles, HLA-C*03:04:01 and HLA-B*15:18:01, emerge as significant predictors of immunogenicity to infliximab and adalimumab respectively. These findings reframe the immunogenetic landscape of biologic therapy in IBD and challenge the ethnic transferability of European-derived pharmacogenomic models.

This opinion review situates the findings of Weng et al[19] within the context of a shifting global IBD burden[20-22], the mechanistic basis of HLA class I-driven immunogenicity[23-26], the role of immunomodulator co-therapy in attenuating ADA formation[16,27,28], newly delineated allele subtype behavior[13-15], the evolving therapeutic drug monitoring (TDM) paradigm[29-31], and the broader precision pharmacogenomics agenda in IBD[32-35]. We propose a framework for ethnically stratified biologic selection and outline priorities for future multi-ethnic research.

A SHIFTING GLOBAL EPIDEMIOLOGICAL LANDSCAPE

IBD was historically considered a disease of early-industrialized Western nations. Kaplan and Windsor[20] conceptualized its global evolution across four epidemiological stages – emergence, acceleration of incidence, compounding prevalence, and prevalence equilibrium – with most Asian nations currently transitioning from stage two to stage three. An updated analysis incorporating over 500 population-based studies spanning more than a century confirms a dramatic acceleration of IBD burden in East Asia[21]. Global burden of disease analyses document a near-doubling in age-standardized IBD prevalence across Asia between 1990 and 2019[22], driven by Westernization of diet, urbanization, antibiotic exposure, and gut microbiome disruption.

This shift has direct and underappreciated pharmacogenomic implications. The Asian IBD population is rapidly growing and increasingly exposed to anti-TNF biologic therapy, yet the pharmacogenomic models guiding biologic selection were developed almost exclusively in European cohorts. The PANTS study enrolled participants from United Kingdom National Health Service trusts[10]; the meta-analyses of Solitano et al[11] and Rodríguez-Alcolado et al[12] were similarly dominated by European data. If the HLA alleles that drive ADA formation differ between ancestral backgrounds – as Weng et al[19] now demonstrate for Taiwanese patients and Osaka et al[15] for Japanese patients – systematic misclassification of immunogenicity risk follows for the millions of Asian IBD patients increasingly treated with anti-TNF agents. Extending precision pharmacogenomics[32-35] to these populations is not an academic aspiration but a clinical and ethical necessity.

REASSESSING THE ROLE OF HLA-DQA1*05 AND ITS SUBTYPES

The HLA-DQA1*05 allele plays a central role in antigen presentation and immune regulation, predisposing carriers to heightened immunogenic responses against therapeutic proteins. The PANTS study[10] provided compelling evidence that HLA-DQA1*05 carriage significantly increased ADA formation against both infliximab and adalimumab, with a HR of 1.90 (95%CI: 1.60-2.25) and a 92% ADA rate at one year with infliximab monotherapy in carriers. Two independent meta-analyses[11,12] have confirmed this association across immune-mediated inflammatory diseases, and TDM-based prospective cohort data from PANTS[36] established that low drug concentrations at week 14 predicted subsequent ADA-mediated treatment failure, linking the pharmacogenomic finding to a TDM-actionable window.

The scientific understanding of HLA-DQA1*05 has been substantially refined through higher-resolution analyses. Three-year PANTS extension data[9] demonstrated that ADAs associated with undetectable infliximab concentrations could be specifically predicted by HLA-DQA1*05 carriage and modified by early concomitant immunomodulator initiation. Hodges et al[13], using four-digit allele resolution in over 25000 United Kingdom IBD BioResource patients, demonstrated distinct subtype-specific behavior: HLA-DQA1*05:01 specifically associated with shortened time-to-LoR to infliximab, while HLA-DQA1*05:05 associated with LoR to both infliximab and adalimumab. Ternette et al[14] confirmed these differential drug-specific associations in the Oxford OASIS cohort, identifying the molecular basis in linkage disequilibrium with co-inherited DQB1 and DRB1 alleles that alter the HLA class II heterodimer peptide-binding groove. The Hodges et al[13] algorithm provides directly actionable guidance: HLA-DQA1*05:01 carriers should receive adalimumab and avoid infliximab, while HLA-DQA1*05:05 carriers can receive either biologic with an immunomodulator (number needed to treat approximately five to prevent one drug-clearing antibody event).

In East Asian populations, the allele landscape diverges strikingly. Osaka et al[15] in 301 Japanese IBD patients identified HLA-DQB1*03:01 rather than HLA-DQA1*05 as the most significant predictor of infliximab discontinuation (HR = 2.03; P = 9.42 × 10^-5), noting that all HLA-DQA1*05:05 alleles were in linkage disequilibrium with HLA-DQB1*03:01. Combined with the finding of Weng et al[19] that HLA-DQA1*05 carries no predictive value in Taiwanese patients, a coherent picture emerges: The European pharmacogenomic signal attributed to HLA-DQA1*05 reflects specific haplotypic architecture of European ancestry rather than universal biology. This discrepancy aligns with real-world evidence from Pascual-Oliver et al[7] and Pau et al[8], who reported absent or inconsistent HLA-DQA1*05 associations in Spanish and Italian cohorts respectively, underscoring the necessity of ethnically stratified pharmacogenomic profiling.

DISCOVERY OF ETHNIC-SPECIFIC HLA CLASS I RISK ALLELES IN TAIWAN

The most striking contribution of Weng et al[19] is the identification of two novel HLA class I alleles as significant predictors of ADA formation. Among infliximab-treated patients, 31.6% of ADA-positive patients carried HLA-C*03:04:01 compared with 0% of ADA-negative patients (P = 0.02). In the adalimumab-treated subgroup, 66.7% of ADA-positive patients carried HLA-B*15:18:01 vs 0% of ADA-negative patients (P = 0.016). These stark frequency differentials, with complete absence of both alleles in ADA-negative patients, provide statistically robust and biologically plausible evidence for allele-specific immunogenicity risk.

HLA-C*03:04:01 has been previously associated with susceptibility to specific autoimmune disorders. In myasthenia gravis, this allele may contribute to aberrant immune recognition and autoantibody production against acetylcholine receptors[36]. In sarcoidosis, HLA-C*03:04:01 has been implicated in modulating granulomatous immune responses in genetically predisposed East Asian individuals[37]. These prior autoimmune associations suggest a broad susceptibility to aberrant immune activation; in the context of repeated biologic exposure, this may manifest as ADA formation against immunogenic epitopes on infliximab’s chimeric molecular regions.

HLA-B*15:18:01 has been linked to adverse drug reactions and therapeutic outcomes in hematological conditions. It has been associated with drug-induced liver injury through hypersensitivity-mediated HLA class I antigen-presentation pathways[38]. Its carriage also appears to shape the immune milieu governing bone marrow failure and recovery following immunosuppressive therapy in aplastic anemia[39]. The breadth of these associations – spanning drug hypersensitivity, autoimmunity, and immune-mediated tissue destruction – suggests that HLA-B*15:18:01 occupies a consequential position in the immune tolerance-susceptibility axis, one that in the context of adalimumab exposure may predispose to high-titer ADA development.

MECHANISTIC BASIS OF HLA CLASS I-DRIVEN IMMUNOGENICITY

HLA class I molecules primarily present endogenously derived 8-10 amino acid peptides to CD8+ cytotoxic T lymphocytes (CTLs) – a function traditionally considered less directly relevant to ADA formation than HLA class II-mediated CD4+ T helper activation, which provides cognate help to B cells for immunoglobulin class switching and high-affinity antibody production. However, three mechanistic pathways provide biologically plausible explanations for HLA class I allele-driven ADA formation.

First, cross-presentation enables antigen-presenting cells to load exogenously derived therapeutic protein fragments onto HLA class I molecules via non-classical endosomal pathways[23]. CD8+ T cells primed through cross-presentation can provide non-cognate help to B cells in germinal centers, facilitating immunoglobulin class switching integral to high-affinity ADA production[25]. Anti-drug CD8+ T cell responses to biotherapeutics have now been documented across multiple biologic drug classes[25]. In silico deep learning models trained to predict HLA class I epitope presentation[40] could be adapted to model biologic drug-derived peptide immunogenicity for specific alleles, providing a tractable experimental first step.

Second, specific structural features of HLA-C*03:04:01 and HLA-B*15:18:01 peptide-binding grooves – defined by critical residues at positions 9, 63, 67, 70, and 99 – may create particular affinities for immunogenic epitopes derived from infliximab and adalimumab[23]. Allele-specific presentation of drug-derived peptides not subject to central thymic tolerance could trigger T cell-dependent ADA responses absent in non-carriers[26]. The IPD-IMGT/HLA Database, now cataloguing over 35000 alleles[23], provides the reference framework for structural peptide-binding analyses that can test this hypothesis directly.

Third, HLA-B*15:18:01 may modulate peripheral immune tolerance through differential interaction with inhibitory killer cell immunoglobulin-like receptors (KIRs) expressed on natural killer cells and CTLs[24]. Allelic variants that reduce KIR-mediated inhibitory signalling may lower the threshold for CTL activation and impair regulatory T cell-mediated peripheral tolerance to therapeutic proteins[24]. Each of these pathways is amenable to experimental validation through structural peptide-binding assays, HLA-peptide tetramer studies, and in vitro T cell activation assays using patient-derived lymphocytes – work urgently needed to substantiate the associations reported by Weng et al[19].

METHODOLOGICAL RIGOR AND STUDY DESIGN CONSIDERATIONS

The study of Weng et al[19] is commendable for its methodological rigor. Conducted across three major Taiwanese medical centers, it enrolled 95 IBD patients who underwent high-resolution HLA genotyping via NGS – a methodology that provides allele-level resolution across all classical HLA loci simultaneously and represents the current standard for pharmacogenomic profiling[41,42]. NGS-based HLA typing resolves the phase ambiguities that confound Sanger sequencing, enables cost-effective high-throughput analysis, and permits comprehensive four-digit resolution across HLA-A, HLA-B, HLA-C, HLA-DRB1, HLA-DQA1, HLA-DQB1, and HLA-DPB1 in a single assay[41,42]. The rapid expansion of the IPD-IMGT/HLA allele catalogue[23] – driven predominantly by NGS-based novel allele discovery – underscores the growing clinical relevance of this methodology for ethnically diverse populations.

ADA levels were measured using ELISA-based assays, and TDM was integrated into the study protocol, enabling direct correlation between ADA titers and serum drug trough concentrations. The authors reported statistically significant inverse correlations (Pearson coefficients -0.446 for infliximab and -0.455 for adalimumab), mechanistically consistent with prior multinational evidence from Moots et al[43] and Nencini et al[44], both of whom documented that ADA formation accelerates drug clearance and significantly diminishes clinical efficacy. Crucially, these negative correlation coefficients are numerically similar to those in European multinational studies[43,44], confirming that the pharmacokinetic consequences of immunogenicity are ethnicity-independent and that reactive TDM algorithms validated in European cohorts can be transferred to East Asian patients without modification of the monitoring framework.

Several limitations warrant acknowledgment. The modest sample size constrains statistical power for rare-allele testing. Near-universal immunomodulator use may have attenuated allele-specific effect sizes; prior evidence demonstrates that combination therapy reduces ADA frequency by approximately 47%[16] and thiopurines specifically prevent ADA development in infliximab-treated patients[27,28]. The free ADA ELISA methodology detects only unbound circulating antibodies, potentially underestimating total immunogenicity burden compared with drug-tolerant assays[31,45], a recognized limitation that may result in systematic underestimation of pharmacokinetic failure driven by immune complex formation.

THE ROLE OF CONCOMITANT IMMUNOMODULATORS IN THE CONTEXT OF HLA RISK

The role of immunomodulator co-therapy in mitigating ADA formation intersects directly with the pharmacogenomic findings of Weng et al[19]. The meta-analysis by Garcês et al[16] demonstrated that concomitant methotrexate or thiopurine therapy reduced ADA frequency by 47% [risk ratio (RR) = 0.53; 95%CI: 0.42-0.67] and substantially attenuated the negative impact of ADAs on drug response rates. Thiopurines contribute to prevention of ADA development in patients treated with infliximab through multiple mechanisms, including reduction of T cell help to B cells and modulation of regulatory T cell function[27,28]. Low-dose oral methotrexate represents a pharmacokinetically equivalent alternative to thiopurines for the specific indication of anti-TNF immunogenicity prevention[27], with a more favorable long-term safety profile in men and older patients where thiopurine-associated lymphoproliferative risk is highest.

Papamichael et al[28] demonstrated that reactive addition of an immunomodulator in patients with immunogenic LoR was associated with ADA elimination and regain of clinical remission in 55% of patients who received combination therapy compared with only 15% of those receiving anti-TNF dose intensification alone (P = 0.032). This evidence directly supports the management of Taiwanese patients carrying HLA-C*03:04:01 or HLA-B*15:18:01 who continue on anti-TNF therapy: If infliximab must be continued in HLA-C*03:04:01 carriers, infliximab-immunomodulator combination therapy should be prioritized given the elevated a priori ADA risk. However, Dai et al[17] and international guidelines[46,47] emphasize that anti-TNF combination therapy carries elevated risks of serious infection and lymphoproliferative malignancy with thiopurines, requiring individualized benefit-risk assessment rather than blanket co-prescribing. HLA-based pre-treatment stratification[13,32,35] provides a framework for targeting combination therapy toward those with genuine pharmacogenomic necessity.

TDM AS AN ESSENTIAL COMPLEMENT TO HLA PHARMACOGENOMICS

TDM in IBD has evolved from reactive utilization at clinical failure to proactive application during maintenance therapy. Reactive TDM – measuring drug levels and ADA titers at the time of LoR – is the accepted standard of care for rationalizing pharmacokinetic failure management, endorsed in international guidelines including the European Crohn’s and Colitis Organisation[47] and British Society of Gastroenterology[48]. When reactive TDM identifies low trough levels with detectable high-titer ADAs, it implicates immunogenic failure and directs management toward biologic class switching; low trough levels without ADAs support dose escalation[45,49].

The evidence base for proactive TDM is more nuanced. A 2022 meta-analysis of nine randomized controlled trials by Nguyen et al[29] involving 1405 IBD patients found no statistically significant benefit of proactive TDM over conventional management in preventing failure to maintain clinical remission (38% vs 42%; RR = 0.96; 95%CI: 0.81-1.13). Conversely, the meta-analysis by Sethi et al[30] incorporating 26 studies concluded that proactive TDM was associated with significantly lower risk of treatment failure compared to standard of care (RR = 0.64; 95%CI: 0.48-0.85). Papamichael et al[31] argue that this discrepancy reflects patient heterogeneity: Proactive TDM confers greatest benefit in pharmacokinetically high-risk patients – precisely those identified by HLA genotyping – who have not yet optimized drug exposure. This ‘pharmacogenomic-TDM integration’ model enables TDM to function as the pharmacokinetic instrument and HLA genotyping as the patient-selection filter, each substantially amplifying the clinical utility of the other.

The TDM data from Weng et al[19] reinforce this framework. The inverse correlation coefficients between ADA titers and drug trough levels observed in Taiwan are numerically similar to those in European multinational studies[43,44], confirming that reactive TDM algorithms validated in European populations apply to East Asian patients without framework modification. Steenholdt et al[50] demonstrated in a randomized controlled trial that individualized TDM-guided therapy was more cost-effective than empirical dose intensification, providing health-economic justification for this integrated approach. Table 1 provides a summary of human leukocyte antigen alleles associated with anti-drug antibody formation across ethnic groups in inflammatory bowel disease.

Table 1 Summary of human leukocyte antigen alleles associated with anti-drug antibody formation across ethnic groups in inflammatory bowel disease.

HLA allele	Class	Ethnic group	Biologic	Key finding/ADA risk	Ref.
HLA-DQA1*05 (group)	II	European	IFX, ADA	HR = 1.90 (95%CI: 1.60-2.25); 92% ADA rate at 1 year with IFX monotherapy in carriers	Sazonovs et al[10]
HLA-DQA1*05:01	II	European	IFX	Specifically shortens time-to-LoR to infliximab; actionable by switching to adalimumab monotherapy	Hodges et al[13]; Ternette et al[14]
HLA-DQA1*05:05	II	European	IFX, ADA	Shortens time-to-LoR to both IFX and ADA; mitigated by concomitant immunomodulator (number needed to treat approximately 5 to prevent one drug-clearing ADA)	Hodges et al[13], Ternette et al[14]
HLA-DQB1*03:01	II	Japanese	IFX	HR = 2.03 (P = 9.42 × 10-5) for IFX discontinuation; subsumes DQA1*05:05 information in Japanese patients	Osaka et al[15]
HLA-C*03:04:01	I	Taiwanese	IFX	31.6% of ADA-positive vs 0% of ADA-negative patients (P = 0.02); prior autoimmune associations	Weng et al[19]
HLA-B*15:18:01	I	Taiwanese	ADA	66.7% of ADA-positive vs 0% of ADA-negative patients (P = 0.016); linked to Drug-Induced Liver Injury and aplastic anaemia outcomes	Weng et al[19]
HLA-DQA1*05 (group)	II	Taiwanese (25.3% allele frequency)	IFX, ADA	No significant association with ADA formation despite comparable carriage frequency; fails to predict in this cohort	Weng et al[19]; Pascual-Oliver et al[7]

ADA: Anti-drug antibodies; HLA: Human leukocyte antigen; HR: Hazard ratio; IFX: Infliximab; LoR: Loss of response.

IMPLICATIONS FOR PERSONALIZED BIOLOGIC THERAPY

The identification of ethnic-specific HLA risk alleles carries profound implications for the personalization of IBD biologic therapy. Pre-treatment HLA screening integrated with TDM-guided protocols could transform biologic optimization from reactive trial-and-error to a predictive, precision-driven strategy (Figure 1)[19]. In practical terms, Taiwanese patients carrying HLA-C*03:04:01 might be directed toward adalimumab, ustekinumab, or vedolizumab over infliximab as first-line biologic therapy[51-54]. Ustekinumab, as an anti-interleukin-12/23 agent mechanistically distinct from anti-TNF biologics, and vedolizumab, as a gut-selective integrin inhibitor, offer alternatives that circumvent the class-specific ADA pathways[51,52]. Patients carrying HLA-B*15:18:01 should be directed away from adalimumab, preserving this agent for future lines of therapy.

Figure 1 Ethnic-specific human leukocyte antigen risk alleles and anti-drug antibody formation in inflammatory bowel disease. This infographic illustrates how anti-drug antibody formation differs by ethnicity in patients with inflammatory bowel disease. While human leukocyte antigen (HLA)-DQA1*05 is a known risk allele in European populations, Weng et al[19] identified HLA-C*03:04:01 and HLA-B*15:18:01 as novel predictors of anti-drug antibody development in Taiwanese patients treated with infliximab and adalimumab, respectively. The graphic emphasizes the role of high-resolution human leukocyte antigen genotyping and therapeutic drug monitoring in guiding personalized biologic therapy. ADAs: Anti-drug antibodies; HLA: Human leukocyte antigen; IBD: Inflammatory bowel disease; TNF: Tumor necrosis factor.

For European patients, the four-digit subtype modeling of Hodges et al[13] provides an immediately actionable prescribing algorithm applicable to approximately 40% of patients. For Japanese patients, Osaka et al[15] provide an analogous algorithm anchored in HLA-DQB1*03:01. Together with the Weng et al[19] Taiwanese data, these converging findings constitute the first multi-ethnic pharmacogenomic framework for anti-TNF biologic selection in IBD and require NGS-based four-digit HLA genotyping[41,42] for implementation. Table 2 synthesizes this proposed clinical decision framework[6,9,10,13,15,19,28,51].

Table 2 Proposed ethnically stratified clinical decision framework integrating human leukocyte antigen genotype and therapeutic drug monitoring for biologic therapy in inflammatory bowel disease.

Clinical scenario	HLA genotype	TDM strategy	Recommended biologic approach	Ref.
European patient, anti-TNF naïve	HLA-DQA1*05:01 carrier	Proactive TDM post-induction; monitor trough levels	Adalimumab monotherapy; avoid infliximab entirely	Hodges et al[13]; Chanchlani et al[9]
European patient, anti-TNF naïve	HLA-DQA1*05:05 carrier	Proactive TDM; monitor ADA titers from induction	IFX or ADA with concomitant immunomodulator; number needed to treat approximately 5 to prevent drug-clearing ADA	Hodges et al[13]; Papamichael et al[28]
Japanese patient, anti-TNF naïve	HLA-DQB1*03:01 carrier	Proactive TDM from induction; closely monitor IFX persistence	IFX + concomitant immunomodulator; consider alternative class if monotherapy required	Osaka et al[15]
Taiwanese patient, naïve to IFX	HLA-C*03:04:01 carrier	Reactive TDM with early ADA titer surveillance	Prefer adalimumab, ustekinumab, or vedolizumab over infliximab	Weng et al[19]; Ashraf et al[51]
Taiwanese patient, naïve to ADA	HLA-B*15:18:01 carrier	Reactive TDM; early switch protocol if ADA detected	Avoid adalimumab; use IFX + immunomodulator or ustekinumab	Weng et al[19]; Atreya and Neurath[6]
Any patient with secondary loss of response	Any relevant allele (or unknown)	Reactive TDM (drug trough + ADA titers)	Dose-optimise for pharmacokinetic failure; switch biologic class for high-titer immunological failure	Chanchlani et al[9]; Sazonovs et al[10]

ADA: Adalimumab; HLA: Human leukocyte antigen; IFX: Infliximab; TDM: Therapeutic drug monitoring; TNF: Tumor necrosis factor.

Precision pharmacogenomics in IBD must also extend beyond anti-TNF HLA profiling. TPMT and NUDT15 genotyping – the most established pharmacogenomic biomarkers in IBD clinical practice[32,33] – identify patients at risk of thiopurine-induced myelosuppression and inform dose selection before therapy initiation. Integrating HLA-based anti-TNF immunogenicity risk assessment alongside these existing panels represents a natural evolution toward a multi-marker precision medicine model. Vieujean and Louis[34] articulated this integrated framework in detail, and the Challenges in IBD Research 2024 precision medicine working group[35] identified HLA-DQA1*05 testing as a near-term pharmacogenomic translation priority – a recommendation that must now explicitly extend to the ethnic-specific alleles identified by Weng et al[19] and Osaka et al[15]. The multi-omics pharmacogenomics framework recently outlined by Khoshnam Rad et al[55], integrating genomics, proteomics, and metabolomics with AI-driven predictive modeling, provides the roadmap for scaling this approach to population level.

LIMITATIONS AND FUTURE DIRECTIONS

Despite its landmark contribution, Weng et al[19] acknowledge several limitations that temper immediate clinical translation. The modest sample size, particularly in the adalimumab subgroup, raises type I error risk in rare-allele testing. Near-universal immunomodulator use may have attenuated allele-specific effect sizes, given the established approximately 47% reduction in ADA frequency with combination therapy[16] and the specific thiopurine-mediated prevention of infliximab immunogenicity[27,28]. Free ADA ELISA methodology may underestimate total immunogenicity burden compared with drug-tolerant assays[31,45]. Independent replication in larger prospective Taiwanese and broader East Asian cohorts – stratified by immunomodulator use and employing drug-tolerant ADA assays – is essential before these alleles can be incorporated into clinical prescribing algorithms. The BSG’s position statement on TDM[48] and European Crohn’s and Colitis Organisation guidelines[47] both emphasize the importance of assay standardization in multicenter studies, an additional methodological consideration for future replication.

Several research priorities emerge. First, multi-center prospective studies across Chinese, Japanese, Korean, and Southeast Asian IBD cohorts are needed to determine population-specific allele frequency and predictive performance. The Osaka et al[15] finding that HLA-DQB1*03:01 dominates in Japanese patients illustrates that pan-Asian generalization is inappropriate. Trans-ethnic genome-wide association studies[56] in adequately powered non-European IBD cohorts will be required to comprehensively map the pharmacogenomic immunogenicity landscape. Second, structural peptide-binding studies facilitated by the IPD-IMGT/HLA reference database[23], HLA-peptide tetramer assays, cross-presentation pathway characterization, and assessment of KIR-mediated tolerance mechanisms[24] represent tractable experimental priorities. Third, health technology assessments evaluating the cost-effectiveness of pre-treatment HLA screening are urgently needed, building on the demonstrated cost-effectiveness of TDM-guided individualized therapy[50]. Fourth, AI-driven multi-omics predictive models[32,55] incorporating multi-locus HLA genotypes, disease phenotype, pharmacokinetic parameters, and biomarker profiles hold transformative potential for generating individualized immunogenicity risk scores.

CONCLUSION

Weng et al[19] have published a timely and impactful study that challenges prevailing assumptions about HLA-associated immunogenicity in IBD. By identifying novel ethnic-specific risk alleles in a Taiwanese cohort, they demonstrate that the concept of a universal pharmacogenomic predictor of anti-TNF immunogenicity – developed in European patients and applied globally without ethnic calibration – is untenable. This conclusion is reinforced by parallel findings in Japanese patients, the four-digit subtype specificity established in European patients, and convergent real-world evidence from non-European settings.

The paradigm of precision biologic therapy in IBD – anchored in ethnogenetic diversity, immunomodulator co-therapy directed by pharmacogenomic necessity, TDM[57], and translational pharmacogenomics – demands urgent expansion beyond its current Eurocentric boundaries. As the global burden of IBD continues to rise across Asia, integrating high-resolution NGS-based HLA genotyping into pre-treatment clinical algorithms, with ethnically calibrated allele panels and TDM-guided monitoring, transforms biologic selection from a reactive process into a predictive, precision-driven strategy. Realizing this vision will require dedicated multi-ethnic biobanking, standardized NGS-HLA typing protocols, harmonized TDM assays[58-60], and international pharmacogenomic consortia that ensure the evidence base for precision biologic therapy reflects the full ancestral diversity of the patients it is designed to serve.