668 An integrated model of alopecia areata biomarkers highlights TH1 and TH2 pathways

Abstract

Alopecia areata (AA) is a common nonscarring hair loss disorder with a lifetime risk of 2%. AA is characterized by TH1/IFN-skewing, with additional TH2 and IL-23 activation. Severity of alopecia tool (SALT) was recently correlated with lesional, nonlesional scalp, and serum AA biomarkers. To integrate serum and scalp biomarkers, we used OLINK high-throughput proteomics to evaluate 363 inflammatory and cardiovascular proteins in serum, and 52 lesional and nonlesional scalp biomarkers in moderate-to-severe AA (>30% scalp involvement; mean age 43.5 years; mean SALT 59.57). SALT significantly correlated with 35 serum, 6 lesional, and 5 nonlesional markers including lesional TH1 (IFNg [r=.66, P<.05], CXCL10 [r=.54; P<.1]) and nonlesional (IFNg [r=.51, P<.1], CXCL10 [r=.51; P<.1]) biomarkers as well as TH2 lesional (IL5 [r=.77, P<.01], IL10 [r=.62, P<.05]), nonlesional (IL10, r=.73, P<.01), and serum (IL5, r=.59, P<.05) biomarkers. SALT also significantly correlated with lesional IL12/23p40 (r=.52, P<.1) and nonlesional IL23p19 (r=.58, P<.05). Multivariate U statistics largely improved SALT correlations compared to univariate analyses, with top models including TH1(IFNg, CXCL10) and TH2 (IL5, IL10) markers. Top correlations of SALT with lesional or nonlesional markers were comparable (r=.78, P<.01). The integration of lesional and nonlesional markers further improved correlations (r=.84, P<.001) as did the integration of overlapping blood and skin markers. Integrating serum and skin markers (LAIR2, ADM, KYNU, TNFRSF12A) further improved SALT correlations (r=.90, P<.001). An integrated tissue and blood biomarker approach provides clues to the systemic nature of AA. Such an integrated biomarker strategy should also be attempted in future therapeutic trials for AA, to better track disease reversal.