P38 Integrating proteomics and genomics to investigate the mechanisms of paradoxical eczema developing in patients with psoriasis treated with biologics: a UK multicentre, longitudinal, cohort study

Abstract Few studies have explored the immunology and genetic risk of paradoxical eczema occurring in patients with psoriasis treated with biologics. This study aims to describe the systemic inflammatory signature of paradoxical eczema using proteomics and explore whether it is genetically mediated. Participants were adults treated with biologics for plaque psoriasis, who were corecruited to the British Association of Dermatologists Biologics and Immunomodulators Register (BADBIR) and Biomarkers of Systemic Treatment Outcomes in Psoriasis (BSTOP) studies. To identify differentially expressed proteins and enriched gene sets, we used the Olink Target 96 Inflammation panel on 256 serum samples from BSTOP for 71 cases with paradoxical eczema and 75 controls. Paradoxical eczema was defined as eczema, atopic eczema or atopic dermatitis occurring during biologic exposure. Case samples across three time points [T1: prebiologic (n = 42); T2, postbiologic (n = 45); T3, postparadoxical eczema (n = 41)] were matched 1 : 1 with control samples. Case samples at T2 and T3 were taken during exposure to the same biologic that caused paradoxical eczema. Genes contributing to proteomic signatures were selected, and single-nucleotide polymorphisms (SNPs) that were nonsynonymous or associated with significant expression quantitative trait loci were used to create polygenic risk scores (PRS) in a genotyped cohort of 88 paradoxical eczema cases and 3124 psoriasis controls. Signal transducing adapter molecule 1-binding protein expression was lower in cases at T1 than in controls (log-fold change −0.44, adjusted P = 0.022); no other proteins achieved statistical significance at equivalent time points. Eleven gene sets, including cytokine and chemokine pathways, were enriched in cases at T2 and 10 at T3. Of the 39 proteins contributing to enriched gene sets, 21 are elevated in the sera of patients with atopic dermatitis. A PRS including 38 SNPs linked to enriched gene sets was associated with paradoxical eczema (adjusted P = 0.046). We subdivided the SNPs into those linked to specific gene sets, such as chemokine, cytokine and interleukin (IL)-10 signalling, to generate 10 partitioned PRS. Seven were associated with paradoxical eczema (adjusted P < 0.05). These SNPs do not map to known atopic loci. The paradoxical eczema systemic inflammatory proteome trends toward atopic dermatitis at a gene set level and is detectable before the onset of the phenotype. This signature may be mediated genetically by SNPs associated with these gene sets. Further studies are needed to explore the role of specific pathways (such as IL-10 signalling) in paradoxical eczema and define the inflammatory profile in the skin.