Supplementary Thesis data.

This is the supplementary data for Mullan et al. Monash thesis (2018-2022). 

Abstract

The unpredictable Type IV drug hypersensitivity reactions (DHRs) are T cell mediated reactions that range in severity from the mild maculopapular exanthema (MPE) to severe Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN). Unlike mild MPE reactions, SJS/TEN has associated morbidities and mortality in both acute and resolved disease. Anti-seizure medications (ASM), especially carbamazepine (CBZ), are among the most common causes of DHRs. The HLA-B*15:02 allele is strongly associated with increased risk of CBZ-SJS/TEN in East Asians but not in other populations (Odds ratio[OR]>2000), while another HLA allele, HLA-A*31:01, confers increased risk of CBZ-DHRs in European (OR: 25.9) and Japanese (OR: 33.9) populations. However, neither of these risk HLA allotypes are sufficient to cause CBZ-DHRs, which suggests other factors contribute to the DHR. To identify other risk factors, I used a multi-omics approach that involved two investigative modalities. The first was genomics, where I identified common variants from the whole genome of 116 ASM-SJS/TEN cases and 84 ASM tolerant controls. This analysis identified nine genome wide variants significantly associated with ASM-SJS/TEN. Using an interaction analysis to qualify the relative risk of the novel risk variant by HLA-B*15:02 carrier status, I identified two variants predicted to lower the relative risk of ASM-SJS/TEN, providing a potential explanation for ASM tolerance in some HLA-B*15:02 carriers. Additionally, the analysis suggested ASM-SJS/TEN had complex inheritance and indicated that many variants were contributing a small fraction of risk, and that some of these variants may not have reached genome wide significance (p-value <5e-08). Therefore, I interrogated areas of variant enrichment that were also predicted to have altered gene expression. This identified several novel biomarkers including CD9, CD24 and ABCA1. Utilising transcriptomics as the second omics’ investigation, I examined the differential expression analysis of both PBMCs and T cells from CBZ-SJS/TEN (resolved cases) and CBZ-MPE (active and resolved cases) relative to CBZ-tolerant controls. Intriguingly CD9, CD24 and ABCA1 were all dysregulated in CBZ- SJS/TEN compared to tolerant controls and reinforced their promise as biomarkers. Interrogation of the transcriptomics data also identified differences between CBZ-SJS/TEN and CBZ-MPE samples. Interestingly, there was an overlap of proinflammatory cytokine expression, including IL-6, in both the transcriptomic and follow-up cytokine secretion immunoassays, suggested that a subset of CBZ-MPE resembles a Type IVc reaction, which is in line with the previous classification. Against expectation, transcriptional and cytokine secretion signatures observed for the other subsets of the other subsets CBZ-MPE implicated a Type IVa hypersensitivity, which may explain why the mild MPE rash will not progress to SJS/TEN. Finally, differential expression of gamma-delta (γδ) T cell receptor (TCR) transcripts was observed, with upregulation of TRDV1, TRDV2, TRGV9, TRGV2 in CBZ-SJS/TEN. This suggested γδ T cell may have a role in CBZ-SJS/TEN pathogenesis. Therefore, I sought to determine functionally if the Vδ1+ or Vδ2+ γδ T cell subsets were CBZ responsive. Firstly, I demonstrated preferential proliferation and CBZ specific activation of the Vδ1+ T cells. Interrogating the bulk Vδ1+ TCR repertoire, I identified several CBZ expanded clonotypes and constructs encoding five representative TCRs were transduced into the SKW3 reporter T cell line for further functional interrogation. Of note, one SKW3 transfectant expressing a Vδ1γ2+ TCR isolated from an individual that lack either HLA-A*31:01 or HLA-B*15:02 risk allele had a modest activation in the presence of CBZ and a keratinocyte cell line. This suggested direct CBZ presentation to the γδTCR through an unknown mechanism. Lastly, utilising the limited clinical DHR samples available, I confirmed CBZ can preferentially activate Vδ1+ T cells. Overall, ‘omics informed functional experimentation identified Vδ1+ T cells could have a role in the initiation and/or progression of CBZ-SJS/TEN, and follow-up utilising additional clinical samples is warranted. Additionally, I identified an opportunity to develop bespoke webtools for differential gene expression data (ggVolcanoR) and T cell receptor repertoire analysis (TCR_Explore). Overall this project utilised a multi-omics approach that identified novel biomarkers to aid in DHR classification, or as treatment targets. Importantly, the transcriptomics approach with T cell interrogations implicated a role for non-classical Vδ1+ T cells in CBZ-SJS/TEN pathogenesis.