Skin proteomic analysis of immune activation associated with regression of melanoma metastases induced by diphencyprone

Abstract

To the Editor: Diphencyprone (DPCP), a hapten that causes delayed-type hypersensitivity reactions, has been used to treat warts, alopecia areata, and cutaneous metastases in melanoma patients.1Damian D.L. Saw R.P. Thompson J.F. Topical immunotherapy with diphencyprone for in transit and cutaneously metastatic melanoma.J Surg Oncol. 2014; 109: 308-313https://doi.org/10.1002/jso.23506Crossref PubMed Scopus (51) Google Scholar Previously, our group performed transcriptomic analysis of skin biopsies from melanoma metastases treated with topical DPCP, revealing significant increases in Th1-related genes but not in genes of other major T-cell subsets.2Gulati N. Garcet S. Fuentes-Duculan J. et al.Molecular profiling of immune activation associated with regression of melanoma metastases induced by diphencyprone.J Invest Dermatol. 2016; 136: 2101-2103https://doi.org/10.1016/j.jid.20 16.06.611Crossref PubMed Scopus (0) Google Scholar While gene expression profiling has long been used as a proxy for protein expression, correlation between messenger RNA (mRNA) and protein levels is inconsistent due to factors such as post-transcriptional regulation. Therefore, to more precisely ascertain the mechanisms involved in immune-mediated tumor regression induced by DPCP, we performed proteomic analysis of skin biopsies from 5 patients with cutaneous melanoma metastases treated with DPCP twice weekly for 7 to 14 weeks (Supplemental Material and Methods and Supplementary Table I, available via Mendeley at https://doi.org/10.17632/f5ttw4mygy.1). In this study, we assessed 96 proteins using the Olink immuno-oncology panel for the 5 patients who underwent molecular profiling. All patients had expected skin inflammation and at least partial regression of skin metastases, without systemic side effects. There were 10 proteins (interleukin 6 [IL-6], IL-12, IL-18, TIE2, HGF, PDGFB, TNFSF14, PD-L2, MMP7, MMP12) significantly upregulated (P < .05) in post-treatment versus pretreatment metastases (Fig 1), including biomarkers associated with Th1 response (IL-12), innate immunity (IL-6), and immune checkpoints (PD-L2) (P < .05). All of these proteins had progressively increased expression in cutaneous metastases with repeated DPCP applications, and similar increases were observed in nonlesional skin of these patients following a single DPCP application. Prior work has shown that administration of IL-18 in mice increases activated T cells and natural killer cells that contribute to effective antimelanoma immunity,3Zhou T. Damsky W. Weizman O.E. et al.IL-18BP is a secreted immune checkpoint and barrier to IL-18 immunotherapy.Nature. 2020; 583: 609-614https://doi.org/10.1038/s41586-020-2422-6Crossref PubMed Scopus (93) Google Scholar so the significant upregulation of this protein in our study may contribute to the successful skin metastasis regression observed in our patients. TNFSF14 also likely plays an active role in immune-mediated regression, as in vivo studies have demonstrated an association between the expression of this protein in melanoma metastasis lesions and proliferation of T cells.4Mortarini R. Scarito A. Nonaka D. et al.Constitutive expression and costimulatory function of LIGHT/TNFSF14 on human melanoma cells and melanoma-derived microvesicles.Cancer Res. 15 2005; 65: 3428-3436https://doi.org/10.1158/0008-5472.CAN-04-3239Crossref PubMed Scopus (35) Google Scholar Additionally, there were 40 proteins significantly upregulated in pretreatment metastases versus untreated nonlesional skin (P < .05) (Fig 1), suggesting baseline immune activation in melanoma metastases. Upon DPCP treatment, changes in nonlesional skin positively correlated with those in skin metastases (Supplementary Fig 1, available via Mendeley at https://doi.org/10.17632/f5ttw4mygy.1; Spearman correlation = 0.61, P < .0001). We performed a correlation analysis between the proteomic data and the previously published mRNA data,2Gulati N. Garcet S. Fuentes-Duculan J. et al.Molecular profiling of immune activation associated with regression of melanoma metastases induced by diphencyprone.J Invest Dermatol. 2016; 136: 2101-2103https://doi.org/10.1016/j.jid.20 16.06.611Crossref PubMed Scopus (0) Google Scholar which both were completed on the same tissue samples, to assess the relationship between these 2 “omic” approaches (Fig 2). Correlation was poor (r = −0.1, P = .4012), including for proteins that were significantly upregulated by proteomics (IL-6 and TNFSF14) (both P < .05). Gene expression from mRNA is variably correlated with protein expression, the latter being influenced by many regulatory processes prior to translation. Another explanation for the discrepancy observed between mRNA and protein expression includes the shorter half-life of mRNA molecules compared to proteins.5Mauger D.M. Cabral B.J. Presnyak V. et al.mRNA structure regulates protein expression through changes in functional half-life.Proc Natl Acad Sci U S A. 2019; 116: 24075-24083https://doi.org/10.1073/pnas.1908052116Crossref PubMed Scopus (152) Google Scholar Despite our limited sample size, we were able to observe statistically significant protein expression changes, and this study is, to our knowledge, the first to assess the proteomic signature of successful immune-mediated regression of cutaneous melanoma metastases induced by a topical immunomodulator. Given the lack of correlation between mRNA and protein expression, along with the relatively greater feasibility of measuring protein compared to gene expression levels, our study demonstrates that proteomic analysis can give unique and potentially complementary insights into immune reactions induced by topical immunotherapy. None disclosed.