Blister fluid as a cellular input for ex vivo diagnostics in drug-induced severe cutaneous adverse reactions improves sensitivity and explores immunopathogenesis

Abstract

Background: Drug-induced severe cutaneous adverse reactions (SCARs) are presumed T-cell-mediated hypersensitivities associated with significant morbidity and mortality. Traditional in vivo testing methods, such as patch or intradermal testing, are limited by a lack of standardization and poor sensitivity. Modern approaches to testing include measurement of IFN-γ release from patient PBMCs stimulated with the suspected causative drug.

Objective: We sought to improve ex vivo diagnostics for drug-induced SCARs by comparing enzyme-linked immunospot (ELISpot) sensitivities and flow cytometry-based intracellular cytokine staining and determination of the cellular composition of separate samples (PBMCs or blister fluid cells [BFCs]) from the same donor.

Methods: ELISpot and flow cytometry analyses of IFN-γ release were performed on donor-matched PBMC and BFC samples from 4 patients with SCARs with distinct drug hypersensitivity.

Results: Immune responses to suspected drugs were detected in both the PBMC and BFC samples of 2 donors (donor patient 1 in response to ceftriaxone and case patient 4 in response to oxypurinol), with BFCs eliciting stronger responses. For the other 2 donors, only BFC samples showed a response to meloxicam (case patient 2) or sulfamethoxazole and its 4-nitro metabolite (case patient 3). Consistently, flow cytometry revealed a greater proportion of IFN-γ-secreting cells in the BFCs than in the PBMCs. The BFCs from case patient 3 were also enriched for memory, activation, and/or tissue recruitment markers over the PBMCs.

Conclusion: Analysis of BFC samples for drug hypersensitivity diagnostics offers a higher sensitivity for detecting positive responses than does analysis of PBMC samples. This is consistent with recruitment (and enrichment) of cytokine-secreting cells with a memory/activated phenotype into blisters.

Keywords: BFC; PBMC; Severe cutaneous drug reactions; ex vivo assays.

Fig 1

IFN-γ release ELISpot assay release for PBMCs and BFCs. The data are for cryogenically stored PBMC and BFC samples from case patients 1 to 4 (see Table E1 in the Online Repository at www.jaci-global.org). A positive result is defined by at least 50 spot-forming units (SFU) per million cells (green dotted line). The maximum doses for each drug were shown to not elicit responses and cell death on a healthy control sample when analyzed by using flow cytometry (7-AAD staining) or lactate dehydrogenase (LDH) viability assay (see Fig E4 in the Online Repository at www.jaci-global.org). Pen G, Penicillin G; SMX, sulfamethoxazole; TMP, trimethoprim.

Fig 2

Lymphocyte composition of blood and blister samples. Donor-matched BFCs and PBMCs were analyzed by flow cytometry. A, Graphs show the percentages of total IFN-γ–positive and CD3⁺ lymphocytes (to the left of the red line) among the total live lymphocytes (gated as per Fig E1, i). T cells (CD3⁺) (gated after exclusion of CD14 [monocytes] and CD19 [B cells] as per Fig E1, ii) were subsequently analyzed for CD4 and CD8 coreceptors (CD4^–/CD8^– cells are indicated as double-negative [DN]), CD45RO (memory), CD69 (activation), CD69 and CD103 coexpression (egress/tissue residency/memory), γδ T-cell receptor (TCR), binding to MR1 5-OP-RU tetramers^, (mucosal-associated invariant T [MAIT] cells), or expression of the NK receptor CD56 (NK-like T cells) (to the right of the red line). B, Graphs show proportions of CD4, CD8, and CD4/CD8 DN T cells, γδ T cells, MAIT cells, and CD56⁺ T cells among IFN-γ–secreting cells, gated as per Fig E2.