HLA Pharmacogenetics and Drug Hypersensitivity

HLA pharmacogenetics and drug hypersensitivity investigates the immunogenetic mechanisms by which specific polymorphic alleles of the Human Leukocyte Antigen (major histocompatibility complex, MHC) locus on chromosome 6p21.3 predispose patients to delayed, T-cell-mediated Type B adverse drug reactions. Unlike metabolic pharmacogenetics variants that alter plasma drug concentrations, HLA Class I (HLA-A, HLA-B, HLA-C) and Class II (HLA-DR, HLA-DQ, HLA-DP) variants encode specialized antigen-presenting glycoproteins whose polymorphic peptide-binding pockets (pockets A through F) interact directly or indirectly with low-molecular-weight pharmaceutical compounds and endogenous peptides.

When a small-molecule drug or its metabolite engages a risk-conferring HLA allotype, it triggers clonal expansion of drug-specific CD8⁺ cytotoxic T lymphocytes (CTLs) or CD4⁺ helper T cells, culminating in life-threatening Severe Cutaneous Adverse Reactions (SCARs)—including Stevens–Johnson Syndrome (SJS), Toxic Epidermal Necrolysis (TEN), and Drug Reaction with Eosinophilia and Systemic Symptoms (DRESS)—as well as immune-mediated Drug-Induced Liver Injury (DILI). Pre-treatment HLA genotyping (such as *HLA-B\*57:01* before abacavir, *HLA-B\*15:02* before carbamazepine, and *HLA-B\*58:01* before allopurinol) is one of the most successful clinical applications of precision medicine, virtually eliminating specific drug-induced SCARs in screened cohorts.

Stevens–Johnson Syndrome (SJS), Toxic Epidermal Necrolysis (TEN), and DRESS

Delayed T-cell-mediated drug hypersensitivity reactions (Gell and Coombs Type IV) typically manifest 4 to 28 days after initiating therapy and encompass three major clinical syndromes:

1. Stevens–Johnson Syndrome (SJS) and Toxic Epidermal Necrolysis (TEN): SJS and TEN represent a severity spectrum of acute mucocutaneous blistering disease characterized by full-thickness epidermal necrosis, hemorrhagic erosions of at least two mucosal surfaces (ocular, oral, and urogenital), and a positive Nikolsky sign (shearing of the epidermis upon lateral mechanical pressure). Classification depends on the percentage of total body surface area (BSA) detachment: SJS involves <10% BSA (mortality ~1–5%), SJS/TEN overlap involves 10% to 30% BSA, and TEN involves >30% BSA (mortality 25% to 35%). Within blister fluid, drug-restricted CD8⁺ CTLs and natural killer (NK) cells induce widespread keratinocyte apoptosis primarily by secreting granulysin (a 15-kDa cationic cytolytic protein discovered by Chung et al. in 2008 as the dominant mediator of epidermal necrosis), alongside perforin, granzyme B, and Fas ligand (sFasL).
2. Drug Reaction with Eosinophilia and Systemic Symptoms (DRESS / DIHS): Presenting 2 to 8 weeks after drug exposure, DRESS features a generalized morbilliform or exfoliative eruption, high fever (>38.5 °C), lymphadenopathy, marked peripheral eosinophilia (>1.5 × 10⁹/L) or atypical lymphocytosis, reactivation of latent human herpesviruses (HHV-6, HHV-7, EBV, or CMV), and visceral organ inflammation (hepatitis, interstitial nephritis, pneumonitis, or myocarditis; mortality ~10%).
3. Abacavir Hypersensitivity Syndrome (AHS) and Immune DILI: Abacavir triggers a distinct systemic syndrome (fever, rash, gastrointestinal distress, and respiratory symptoms within 6 weeks; fatal anaphylactoid shock upon rechallenge), while alleles such as *HLA-B\*57:01* (flucloxacillin) and *HLA-DRB1\*15:01* (amoxicillin–clavulanate) drive cholestatic or hepatocellular DILI.

Classical immunology establishes that the αβ T-cell receptor (TCR) recognizes 8- to 11-mer peptides bound to MHC Class I or 13- to 25-mer peptides bound to MHC Class II, whereas synthetic drugs (<1,000 Da) are too small to be immunogenic alone. Three distinct structural mechanisms explain how small molecules overcome this barrier:

1. The Hapten / Prohapten Model (Covalent Modification)

Formulated by Karl Landsteiner, the hapten concept applies to chemically reactive drugs (such as β-lactam penicillins and cephalosporins, whose strained four-membered β-lactam ring undergoes nucleophilic attack by the ε-amino group of lysine residues on host serum or cellular proteins) or prohaptens (chemically inert parent drugs such as sulfamethoxazole, which CYP2C9 oxidizes into a reactive hydroxylamine and nitroso-sulfamethoxazole metabolite that covalently modifies cysteine sulfhydryl groups). Antigen-presenting cells internalize these drug-modified proteins, process them through the immunoproteasome, and load the resulting covalently haptenated peptides onto MHC grooves via TAP transporters to prime neo-antigen-specific T cells.

2. The p-i Concept (Pharmacological Interaction with Immune Receptors)

Proposed by Werner Pichler, the p-i concept demonstrates that chemically inert drugs lacking covalent reactivity can bind non-covalently and reversibly (via hydrogen bonds, salt bridges, and van der Waals interactions) directly to either the surface of an exposed HLA-peptide complex (p-i HLA) or the variable complementarity-determining regions (CDRs) of the T-cell receptor (p-i TCR). Because p-i binding bypasses intracellular metabolism, proteasomal processing, and TAP transport, aldehyde-fixed antigen-presenting cells pulsed with the parent drug can trigger immediate Ca²⁺ influx in T-cell clones within seconds.

Canonical examples of the p-i HLA model include:

• Carbamazepine and *HLA-B\*15:02*: Aromatic antiepileptic drugs (carbamazepine, oxcarbazepine, phenytoin, lamotrigine) bind non-covalently near the B- and F-pockets of HLA-B\*15:02 (stabilized by Arg62, Asn63, Ile95, and Leu156), where the drug-HLA complex is recognized by a restricted public CD8⁺ TCR clonotype (Vβ-11-ISGSY).
• Allopurinol / Oxypurinol and *HLA-B\*58:01*: Allopurinol and its long-lived active xanthine oxidase inhibitor metabolite oxypurinol bind non-covalently within the peptide-binding groove of HLA-B\*58:01 (making specific contacts with Arg97), triggering polyclonal CD8⁺ T-cell activation in SJS/TEN and DRESS.

3. The Altered Self-Peptide Repertoire Model (Abacavir and HLA-B*57:01)

Solved simultaneously by X-ray crystallography in 2012 by Illing et al. and Ostrov et al., the altered self-peptide repertoire model explains the specificity of abacavir (a guanosine analog HIV-1 nucleoside reverse transcriptase inhibitor) for HLA-B\*57:01. Abacavir binds non-covalently deep at the bottom of the antigen-binding cleft across pockets C, D, E, and specifically the F-pocket of HLA-B\*57:01, forming specific hydrogen bonds with two polymorphic floor residues: Asp114 and Ser116 (even a single amino acid change in closely related allotypes such as HLA-B\*57:03 [Asn114/Tyr116], HLA-B\*57:02, or HLA-B\*58:01 abolishes abacavir binding).

Under physiological conditions without abacavir, the F-pocket of HLA-B\*57:01 prefers endogenous self-peptides terminating at their C-terminus (PΩ anchor position, P9) with bulky aromatic amino acids (tryptophan [Trp] or phenylalanine [Phe]). When abacavir occupies the bottom of the F-pocket in the endoplasmic reticulum, its cyclopropyl-aminopurine scaffold sterically protrudes upward, altering the shape and chemistry of the cleft so that bulky aromatic PΩ residues can no longer fit. Instead, HLA-B\*57:01 loads a new repertoire of endogenous self-peptides terminating in small aliphatic residues (isoleucine [Ile], leucine [Leu], or valine [Val]). Mass spectrometry shows that 20% to 25% of peptides presented on the cell surface in the presence of abacavir are novel self-peptides to which the patient's T-cell repertoire was never negatively selected in the thymus—provoking a broad polyclonal CD8⁺ T-cell response mimicking acute allogeneic graft rejection.

The landmark double-blind PREDICT-1 randomized clinical trial (Mallal et al., NEJM 2008; N = 1,956 HIV patients) demonstrated that prospective *HLA-B\*57:01* screening reduced immunologically patch-test-confirmed abacavir hypersensitivity from 2.7% to 0.0%—yielding a Negative Predictive Value (NPV) of 100% and a Positive Predictive Value (PPV) of 47.9%. Similarly, nationwide implementation of pre-treatment *HLA-B\*15:02* screening in Taiwan (Chen et al., NEJM 2011) eliminated carbamazepine-induced SJS/TEN across 4,877 screened patients.

Why is the PPV of HLA risk alleles less than 100% (ranging from 47.9% for *HLA-B\*57:01*–abacavir, ~3% to 5.6% for *HLA-B\*15:02*–carbamazepine, ~2% to 3% for *HLA-B\*58:01*–allopurinol, down to 0.12% for *HLA-B\*57:01*–flucloxacillin DILI)? Possessing the risk HLA allele is necessary but not sufficient: clinical disease additionally requires the stochastic presence of matching drug-reactive TCR clonotypes in the patient's T-cell repertoire, impaired regulatory T-cell (T_reg) or PD-1/CTLA-4 checkpoint suppression, endoplasmic reticulum aminopeptidase (ERAP1) trimming polymorphism, or reduced systemic drug clearance (such as chronic kidney disease elevating plasma oxypurinol concentrations in allopurinol patients).

How does the abacavir altered self-peptide repertoire model differ from the hapten and p-i models?

In the hapten model, a reactive drug binds covalently to host proteins that are processed into drug-conjugated peptides. In the p-i model, an inert drug binds non-covalently to the outer surface of an already-formed HLA-peptide complex or TCR to stimulate T cells directly. In the altered self-peptide repertoire model, abacavir binds non-covalently deep inside the F-pocket of HLA-B\*57:01 during endoplasmic reticulum assembly, sterically shifting the C-terminal anchor preference from Trp/Phe to Ile/Leu so that 20–25% of presented host peptides are novel self-peptides that trigger a polyclonal alloreactive-like T-cell response.

Why is HLA-B*15:02 screening prioritized in Asian populations while HLA-A*31:01 is relevant across Europeans and Japanese?

The allele frequency of *HLA-B\*15:02* is high in Han Chinese, Thai, Malaysian, Filipino, and Vietnamese populations (8% to 15%) and moderate in South Asians (2% to 4%), where it specifically drives carbamazepine-induced SJS and TEN (odds ratio >100), but is virtually absent (<0.1%) in Europeans and Japanese. Conversely, *HLA-A\*31:01* is present at 2% to 10% in Europeans, Japanese, Koreans, and Indigenous Americans, and predisposes carriers to the broader spectrum of carbamazepine hypersensitivity (DRESS, maculopapular exanthema, and SJS/TEN).

Why do cross-reactive aromatic antiepileptic drugs need to be avoided in HLA-B*15:02 carriers?

Carbamazepine, oxcarbazepine, phenytoin, fosphenytoin, and lamotrigine share a conserved tricyclic or aromatic carboxamide/hydantoin ring structure that docks into the same B/F-pocket region of HLA-B\*15:02. Consequently, an *HLA-B\*15:02*-positive patient (or a patient with prior carbamazepine SJS/TEN) faces a high risk of cross-reactive SJS/TEN if switched to oxcarbazepine or phenytoin; non-aromatic antiepileptics such as levetiracetam, valproic acid, or topiramate should be prescribed instead.

1. Mallal, S., Phillips, E., Carosi, G., Molina, J. M., Workman, C., Tomazic, J., et al. (2008). "HLA-B*5701 screening for hypersensitivity to abacavir." New England Journal of Medicine, 358(6), 568–579. https://doi.org/10.1056/NEJMoa0706135
2. Illing, P. T., Vivian, J. P., Dudek, N. L., Kostenko, L., Chen, Z., Bharadwaj, M., et al. (2012). "Immune self-reactivity triggered by drug-modified HLA-peptide repertoire." Nature, 486(7404), 554–558. https://doi.org/10.1038/nature11147
3. Chen, P., Lin, J. J., Lu, C. S., Ong, C. T., Hsieh, P. F., Yang, C. C., et al. (2011). "Carbamazepine-induced toxic effects and *HLA-B*1502* screening in Taiwan." New England Journal of Medicine, 364(12), 1126–1133. https://doi.org/10.1056/NEJMoa1009717
4. Phillips, E. J., Sukasem, C., Whirl-Carrillo, M., Müller, D. J., Dunnenberger, H. M., Chantratita, W., et al. (2018). "Clinical Pharmacogenetics Implementation Consortium guideline for HLA genotype and use of carbamazepine and oxcarbazepine: 2017 update." Clinical Pharmacology & Therapeutics, 103(4), 574–581. https://doi.org/10.1002/cpt.1004