Multi-omic sequencing defines the genetic, cellular, and structural risk signatures of severe cutaneous adverse reactions associated with co-trimoxazole

Yueran Li

T-cell-mediated severe cutaneous adverse drug reactions (SCAR) remain the most severe but clinically unpredictable adverse drug reactions. In particular, Stevens-Johnson Syndrome/Toxic epidermal necrolysis (SJS/TEN) is associated with 15-30% mortality and significant long-term morbidities. However, the global consensus for treatment remains supportive care and drug withdrawal only, without a clear understanding of the immunopathogenesis in the affected tissue to develop safe and appropriate targeted pharmacological treatments. Moreover, in recent years, strong human leukocyte antigen (HLA) risk associations have been discovered for some drugs, enabling the implementation of preventative pharmacogenomic screening. However, the HLA risk allele is different for different drugs and populations, many of whom now remain at disparate risk. Further, HLA is not completely predictive, implicating other risk factors and limiting the clinical benefit of preventative screening, for which sequencing-based typing remains the gold standard but is high-cost and prohibitive to clinical uptake. Recently, drug-expanded and risk HLA-restricted T-cell receptors (TCR) have been identified in the skin of patients with SJS/TEN, which are not similarly expanded in the blood. These data propose the future identification of the immunodominant TCR in the tissue will help understand the nature of the immunogenic risk HLA-TCR-presented epitope to improve pre-clinical structural risk screening strategies, but also to act as a biomarker to identify the cytotoxic population driving disease in immunologically diverse skin, towards an understanding of cell-cell and receptor-ligand interactions as appropriate tissue-relevant diagnostic biomarkers and strategies for therapeutic intervention. Thus, here, with a focus on co-trimoxazole, a highly effective antibiotic that is also among the top global drug causes of SCAR including SJS/TEN, I utilize multi-omic analyses to identify new genetic, structural, and cellular risk signatures of SJS/TEN, which will be critical to reduce the incidence, mortality, and morbidity of SJS/TEN in global populations. First, using a geographical population approach and the BioVU control database of genomic data from >94,000 individuals in the US, I identify HLA-B*44:03 as a new risk HLA allele for co-trimoxazole-induced SCAR and SJS/TEN, and provide the first description for the extended risk of the B44 supertype (alleles that share peptide-binding specificities) in predisposing co-trimoxazole SCAR in global populations. Indeed, these data align with the recently reported HLA-B*13:01-restricted predisposition of co-trimoxazole-SCAR in Asian populations, which is a prominent B44 supertype allele in Asia, but is lowly expressed by other global populations. These data now demonstrate a clear rationale for functional studies and the broad implementation of B44 supertype screening to predict and reduce the incidence of co-trimoxazole-induced SCAR across global populations. Second, I provide the first unbiased multi-omic single-cell analyses across the bluster fluid of HLA-B44+ patients with co-trimoxazole-induced SJS/TEN, and the first longitudinal single-cell analyses of SJS/TEN blister fluid. These data now highlight the potential utility of MIF, JAK, STAT, and PKM inhibitors, and LAG3+ cell-depletion therapies as ‘early’ actionable therapeutic targets to dampen the cytotoxic effector response and reduce mortality and morbidity, but also M2 macrophage-polarising treatments to act as ‘late’ therapeutic targets to increase time to re-epithelisation. Importantly, these studies also detail the entire cytotoxic TCR repertoire of CD8+ T-cells in the skin that will be crucial to facilitate peptidome-screening assays such as T-Scan to define the immunogenic peptides that will be crucial to enable safer drug development. Finally, I provide a new, low-cost (< $5 USD) and readily implementable pharmacogenomic PCR-based screening tool with complete specificity and sensitivity for risk HLA-B*35 alleles to reduce the incidence of B*35-associated disease, including diverse drug and chemical-induced reactions, which can be performed at most on-site clinical laboratories. Importantly, an expanding array of more sophisticated bioinformatic tools and reference datasets continue to drive genomic research, and the co-trimoxazole-related single-cell atlas presented in this thesis will continue to expand using carefully curated clinical samples to discern disease-specific pathogenesis for early diagnostics and more tailored treatments. Indeed, my studies now enable an integrated vision for a single atlas of diseased skin at a single-cell resolution that in the future will be able to discern disease severity, progression from resolution, mortality from survival, treatment response (including long-term follow-up after exposure to biologics), and even morbidities associated with SCAR, including investigations for the higher risk of autoimmune sequelae associated with other reactions.

Multi-omic sequencing defines the genetic, cellular, and structural risk signatures of severe cutaneous adverse reactions associated with co-trimoxazole

Files and links (1)

Abstract

Details

Metrics