In-silico discovery of the hapten-based interaction between idiosyncratic hepatotoxic ingredients of Polygonum multiflorum and HLA-B*35:01

Abstract Background The previous research has confirmed the existence of idiosyncratic drug-induced liver injury (IDILI) caused by Polygonum multiflorum (PM-IDILI), and demonstrated that PM-IDILI is an immune-mediated injury, with HLA-B*35:01 identified as a genetic susceptibility marker. Additionally, emodin-8- O-β -D-glucoside (EG) and 2,3,5,4′-tetrahydroxystilbene-2- O-β -D-glucoside have been proposed as potential contributory ingredients in the pathogenesis of PM-IDILI. However, the precise mechanisms through which these susceptible factors contribute to the development of PM-IDILI remain unclear. Objectives This study aims to explore the molecular characteristics of HLA-B*35:01 that contribute to PM-DILI and to propose a mechanistic hypothesis based on our previous research on PM-induced protein adducts. Methods Key differences between HLA-B*35:01 and general Chinese HLA-B alleles were identified by comparing protein sequences, peptide binding motifs, and protein structures. Molecular docking was employed to assess whether PM-induced haptenated peptides can be presented by HLA-B*35:01 and other related alleles. Additionally, a simplified dipeptide model was used to evaluate the binding affinity of HLA-B*35:01 to EG-haptenated peptides. Results Our findings revealed significant differences in the residues of the B and F peptide binding pockets of HLA-B*35:01 compared to general Chinese HLA-B alleles. Further analysis suggested that the F pocket of HLA-B*35:01 was capable of binding EG-cysteine adducts and might be a key feature in the PM-IDILI pathogenesis. Peptide docking using DINC and molecular dynamics simulations indicated that HLA-B*35:01 could form stable complexes with EG-haptenated peptides. Molecular dynamics simulations also highlighted the critical roles of both the B and F pockets in peptide binding. Specifically, the F pocket binds the EG-modified residue in haptenated peptides, while the B pocket, despite lacking shared features among PM-IDILI patients, may indirectly influence the incidence of PM-IDILI by filtering haptenated peptides. The binding affinity of HLA-B*35:01 to EG-modified cysteine residues was experimentally validated through a dipeptide-based assay, confirming that HLA-B*35:01 could bind EG-haptenated peptides. Conclusions This study identified the unique B and F binding pockets of HLA-B*35:01 as key factors in PM-IDILI pathogenesis and demonstrated that HLA-B*35:01 could bind EG-haptenated peptides. These findings suggest that PM-IDILI may be a hapten-based drug hypersensitivity reaction driven by EG, providing a theoretical framework for further research aimed at elucidating the molecular mechanisms underlying PM-IDILI.