Abstract
A number of serious adverse drug reactions are caused by T cells. An association with HLA alleles has been identified with certain reactions, which makes it difficult to develop standardized preclinical tests to predict chemical liability. We have recently developed a T cell priming assay using the drug metabolite nitroso sulfamethoxazole (SMX-NO). We now report on reproducibility of the assay, establishment of a biobank of PBMC from 1000 HLA-typed volunteers, and generation of antigen-specific responses to a panel of compounds. Forty T cell priming assays were performed with SMX-NO; 5 gave weak responses (1.5-1.9) and 34 showed good (SI 2.0-3.9) or strong responses (SI > 4.0) using readouts for proliferation and cytokine release. Thus, SMX-NO can be used as a model reagent for in vitro T cell activation. Good to strong responses were also generated to haptenic compounds (amoxicillin, piperacillin and Bandrowski's base) that are not associated with an HLA risk allele. Furthermore, responses were detected to carbamazepine (in HLA-B*15:02 donors), flucloxacillin (in 1 HLA-B*57:01 donor) and oxypurinol (in HLA-B*58:01 donors), which are associated with HLA-class I-restricted forms of hypersensitivity. In contrast, naïve T cell priming to ximelagatran, lumiracoxib, and lapatinib (HLA-class II-restricted forms of hypersensitivity) yielded negative results. Abacavir, which activates memory T cells in patients, did not activate naïve T cells from HLA-B*57:01 donors. This work shows that the priming assay can be used to assess primary T cell responses to drugs and to study mechanisms T cell priming for drugs that display HLA class I restriction. Additional studies are required to investigate HLA-class II-restricted reactions.