Date of Award

January 2024

Document Type

Thesis

Degree Name

Medical Doctor (MD)

Department

Medicine

First Advisor

William Damsky

Abstract

Targeted systemic therapy selection in psoriasis and eczema does not account for patient specific, treatment-relevant immunologic heterogeneity. The goal of this project was to develop a rapid, non-invasive method of obtaining and analyzing epidermal protein biomarkers and evaluate its ability to molecularly subclassify psoriasis and eczema and identify biomarkers that correlate with treatment response. Our first aim was to optimize and validate a method of noninvasive, full-thickness epidermal sampling and subsequent protein detection. Our second aim was to identify protein biomarkers that could reliably discriminate between psoriasis and eczema, and to characterize intra-disease proteomic heterogeneity within psoriasis and eczema, respectively. Our third aim was to identify protein biomarkers in eczema that correlated with lack of response to dupilumab. We enrolled patients with psoriasis (n=44 samples) and eczema (n=31 samples); patients underwent noninvasive sampling of lesional skin and protein levels were measured using high-throughput protein assays. Differentially expressed proteins were compared between psoriasis and eczema samples, and supervised and unsupervised machine learning approached were applied to protein data within each diagnosis to molecularly subclassify patients. The sampling approach was found to be painless, nonscarring, and enabled rapid turnaround from time of sample collection to data output. We used this approach to accurately differentiate psoriasis and eczema using a limited set of proteins and to identify cases of eczema/psoriasis overlap with non-canonical molecular profiles; specifically, we identified a subset of low NOS2 expressing psoriasis patients with molecular profiles enriched for Type 2 immunity biomarkers including IL-4, CCL17, and CCL22. Additionally, we identified patient-specific cytokine biomarkers in eczema, including IFN-γ, CXCL9, IL-9, and CXCL14, that were correlated with nonresponse to IL-13 blockade. These results suggest that our approach has potential to realize the goal of personalized medicine in inflammatory skin disease.

Comments

This thesis is restricted to Yale network users only. It will be made publicly available on 06/01/2027

Share

COinS