Drug hypersensitivity caused by alteration of the MHC-presented self-peptide repertoire

Abstract

Idiosyncratic adverse drug reactions are unpredictable, dose-independent and potentially life threatening; this makes them a major factor contributing to the cost and uncertainty of drug development. Clinical data suggest that many such reactions involve immune mechanisms, and genetic association studies have identified strong linkages between drug hypersensitivity reactions to several drugs and specific HLA alleles. One of the strongest such genetic associations found has been for the antiviral drug abacavir, which causes severe adverse reactions exclusively in patients expressing the HLA molecular variant B*57:01. Abacavir adverse reactions were recently shown to be driven by drug-specific activation of cytokine-producing, cytotoxic CD8(+) T cells that required HLA-B*57:01 molecules for their function; however, the mechanism by which abacavir induces this pathologic T-cell response remains unclear. Here we show that abacavir can bind within the F pocket of the peptide-binding groove of HLA-B*57:01, thereby altering its specificity. This provides an explanation for HLA-linked idiosyncratic adverse drug reactions, namely that drugs can alter the repertoire of self-peptides presented to T cells, thus causing the equivalent of an alloreactive T-cell response. Indeed, we identified specific self-peptides that are presented only in the presence of abacavir and that were recognized by T cells of hypersensitive patients. The assays that we have established can be applied to test additional compounds with suspected HLA-linked hypersensitivities in vitro. Where successful, these assays could speed up the discovery and mechanistic understanding of HLA-linked hypersensitivities, and guide the development of safer drugs.

Fig. 1.

Schematic presentation of HLA antigen presentation and HLA-linked mechanisms of adverse reactions. (Upper) T-cell receptors monitor the universe of antigens to which an individual is exposed by surveying the ligands presented on an antigen-presenting cell membrane in the context of HLA molecules. The HLA ligands are typically peptides loaded onto the HLA molecule inside the antigen-presenting cells and subsequently exposed on the surface. Different allelic variants of HLA molecules have different binding specificities, resulting in a specific profile of presented ligands. In the example shown, peptide A, but not peptide B, can bind to the HLA molecule. Self-peptides presented to T cells in this manner do not trigger an immune response, because T cells that are self-reactive are negatively selected during thymic development. However, when T cells encounter an unknown ligand (eg, a virus-derived peptide), an immune response is triggered. (Lower) There are three scenarios for HLA-dependent drug-induced modifications that affect the TCR interface: (1) A ligand that is uniquely presented by the HLA allele is modified by the drug; (2) the HLA molecule itself is modified in a region exposed to the TCR; and (3) the binding specificity of the HLA molecule is altered by the presence of the drug, resulting in presentation of novel ligands such as peptide B.

Fig. 2.

The presence of abacavir alters the binding specificity of HLA-B*57:01. (A and B) Combinatorial peptide libraries were tested for binding to HLA-B*57:01 (A) and HLA-B*58:01 (B) in competitive binding assays as described previously (13, 39). Results for libraries with different C-terminal residues are shown for those residues with affinities of 5,000 nM or better, a minimal threshold for binding. Error bars indicate 95% confidence intervals for the mean, and residues marked with an asterisk had significantly different IC₅₀ values in the presence vs. absence of abacavir (P < 0.001, two-tailed Student t test comparing log IC₅₀ values). The most pronounced affinity increases for HLA-B*57:01 in the presence of 1 mg/mL of abacavir were found for peptides with a valine at the C terminus, which increased by more than eightfold, followed by alanine and isoleucine, which increased by fivefold. In contrast, the maximum affinity increase for any peptide library binding to HLA-B*58:01 was less than threefold. (C and D) Individual peptides HSITYLLPV (pep-V) and HSITYLLPW (pep-W) were radiolabeled and tested for binding to HLA-B*57:01 and B*58:01 in increasing doses of abacavir. After washing, no pep-V binding to HLA-B*57:01 was detectable in the absence of abacavir, but strong binding was detected in the presence of abacavir. No significant effect of abacavir was observed for pep-W binding to HLA-B*57:01.

Fig. 3.

Crystal structure of the abacavir–peptide–MHC complex solved to a resolution limit of 2.0 Å reveals intermolecular contacts within the antigen-binding cleft. (A) Cartoon diagram of HLA-B*57:01 in gray. The peptide HSITYLLPV is shown in cyan carbons. Abacavir is shown as spheres, orange for carbon, blue for nitrogen and red for oxygen. (B) Chemical structure of abacavir, with the cyclopropyl moiety shown in green, the purine core in blue, and the hydroxymethyl cyclopentene moiety in red. (C) Abacavir forms H bond interactions (black dashes) with both the peptide and HLA-B*57:01. The residues that distinguish the abacavir-sensitive allele HLA-B*57:01 from abacavir-insensitive HLA-B*57:03 are shown in magenta for carbon, blue for nitrogen, and red for oxygen. (D) Abacavir binding in the F pocket does not alter the peptide conformation compared with other peptide/HLA-B complexes. A cartoon representation of peptide in the crystal structure complexed to abacavir and HLA-B*57:01 is shown in cyan (HSITYLLPV; PDB ID: 3UPR). A 9-mer self peptide (LSSPVTKSF) complexed to HLA-B*57:01 (PDB ID: 2RFX (2) is shown in red, the 8-mer peptide epitope HIV1 Nef 75–82 (VPLRPMTY) bound to HLA-B*35:01 (PDB ID: 1A1N) (40) is shown in pink, a 9-mer EBV peptide (FLRGRAYGL) complexed to HLA-B8 (PDB ID: 1MI5) (41) is shown in green, and the 11-mer EBV peptide HPVGEADYFEY complexed to HLA-B*35:01 (PDB ID: 3MV9) (42) is shown in yellow. The molecular surface of HLA-B*57:01 from 3UPR is shown in gray. The F pocket residues (9) are colored green, and the A pocket is yellow. (E) Experimental electron density corresponding to abacavir in an Fo-Fc difference map contoured at 3.5σ (red mesh) after molecular replacement. Gray mesh depicts the final 2Fo-Fc electron density map of abacavir in the antigen-binding cleft of HLA-B*57:01 (contour level, 1.5σ). H bond interactions between abacavir and HLA-B*57:01 are shown as yellow dashed lines. The residues that distinguish the abacavir-sensitive allele HLA-B*57:01 from abacavir-insensitive HLA-B*57:03 are shown in magenta for carbon, blue for nitrogen, and red for oxygen.

Fig. 4.

T cells from hypersensitive donors respond to specific self-peptides in an abacavir- dependent fashion. PBMCs from five HLA-B*57:01-positive donors with a clinical history of abacavir hypersensitivity were pulsed for 15 min with peptide antigens in the presence or absence of abacavir or with abacavir alone, washed, and then tested by an IFN-γ ELISpot assay (Materials and Methods). The figure shows the calculated mean (± SEM) IFN-γ spots per million input PBMCs. Statistically significant responses compared with the response induced by the abacavir pulse alone were obtained for peptide pool 1 and the individual peptide VTTDIQVKV (paired two-tailed Student t test on the square root of the SFC counts). Incubating PBMCs from these donors with 10 μg/mL of abacavir overnight yielded an average of 200 ± 63 SFCs per million.