Cellular indexing of transcriptomes and epitopes (CITE-Seq) in hidradenitis suppurativa identifies dysregulated cell types in peripheral blood and facilitates diagnosis via machine learning

TL;DR

This study uses CITE-Seq to identify immune cell changes in hidradenitis suppurativa patients and shows machine learning can help diagnose the condition.

Contribution

The study introduces CITE-Seq to uncover immune cell dysregulation in HS and applies machine learning for diagnostic potential.

Findings

HS patients show increased CD14+ and CD16+ monocytes, cDC2, plasmablasts, and CD4+ T cells compared to controls.

Inflammatory markers like TNF, IL1B, and NF-κB are upregulated in monocytes from HS patients.

Machine learning models can identify key immune markers for HS diagnosis and therapeutic development.

Abstract

Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition characterized by painful nodules, abscesses, and scarring, predominantly affecting intertriginous regions and it is often underdiagnosed. This study aimed to utilize single cell RNA and cell-surface protein sequencing (CITE-Seq) to delineate the immune composition of circulating cells in Hidradenitis suppurativa (HS) peripheral blood compared to healthy controls. CITE-Seq was used to analyze the gene and protein expression profiles of peripheral blood mononuclear cells (PBMCs) from 9 HS and 29 healthy controls. The study identified significant differences cell composition between HS patients and healthy controls, including increased proportions of CD14+ and CD16+ monocytes, cDC2, plasmablasts, and proliferating CD4+ T cells in HS patients. Differential expression analysis revealed upregulation of inflammatory markers such as TNF, IL1B, and NF-κB in monocytes, as well as chemokines and cell adhesion molecules involved in immune cell recruitment and tissue infiltration. Pathway enrichment analysis highlighted the involvement of IL-17, IL-26 and TNF signaling pathways in HS pathogenesis. Machine learning identified key markers for diagnostics and therapeutic development. The findings also support the potential for machine learning models to aid in the diagnosis of HS based on immune cell markers. These insights may inform future therapeutic strategies targeting specific immune pathways in HS.

Figures