Pathogenesis of carbamazepine-induced Stevens-Johnson Syndrome : Identification of the drug-modified peptides bind to the HLA-B*1502

Abstract

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are related life-threatening cutaneous adverse reactions most often caused by medication. Carbamazepine (CBZ), a commonly prescribed anticonvulsant, is the number one culprit drug associated with SJS in Taiwan, accounts for 25% of all drug-induced SJS. Susceptibility to drug induced idiosyncratic reactions is thought to be genetically determined and immune-mediated. Previous studies suggest that the pathogenesis of the severe cutaneous adverse drug reactions involves MHC-restricted presentation of a drug or its metabolites for T-cell activation. However, the specific MHC molecules involved are not known until our study of carbamazepine-induced SJS/TEN in which we identified HLA-B*1502 as the MHC molecule. To further investigate the pathogenesis mechanism, we hypothesized that CBZ or its metabolites bind to endogenous proteins through processing and complexed with HLA-B*1502 or directly bind to the HLA-B*1502-bound peptides, which are then recognized by T-lymphocytes. The specific aims of my thesis are: 1. Establishment of the soluble HLA-B*1502-producing stable clones. 2. Identification of the peptides bind to the HLA-B*1502. 3. Identification of the drug-modified peptides involved in CBZ-SJS. Using soluble HLA-B*1502 molecule as a bait, I identified over 100 peptides bind to the HLA-B*1502. An unusual and an unique feature of these peptides were that many of these endogenous bound peptides (up to 20.9 %) contained polyproline, ranged from 4 to 9 continuous proline residues, but typically 6, 7 or 8 proline residues in the peptides. There were a total of 38 different proline-rich peptides been identified. Among these peptides, the existence of continuous proline residues in the center was a consistent finding; amino acid residues at the N- or C- terminal of peptides were variable. Interestingly, in the presence of CBZ, these proline-rich peptides decreased dramatically. We hypothesized that CBZ may bind to the polyproline peptides covalently, which modify them to become unnatural peptides thus unable to be recognized as the proline-rich peptides under analysis of LC/MS/MS and SEQUEST program. Further functional studies will be needed to demonstrate that these CBZ-modified peptides are indeed involved in the induction of T cell activation in both drug and peptide-specific manners. This is the first study in identification of the MHC-bound peptides associated with adverse drug reactions. The discovery opened a door to understanding the complete mechanism of this life-threatening condition.