Plasma proteomic analyses and treatment predictive biomarker candidates in melanoma patients receiving immune checkpoint blockade or targeted therapy.

Abstract

9574 Background: The introduction of immune checkpoint inhibitors (ICIs) or therapies targeting the MAPK-pathway (MAPKis) has significantly improved clinical outcomes in metastatic melanoma patients. Still, a large proportion of the patients become resistant to therapy and there is a need for treatment predictive biomarkers. The aim of this study was to analyze the treatment predictive biomarkers based on the plasma proteome of patients with metastatic melanoma treated with ICIs or MAPKis. Methods: We analyzed serial plasma samples from 48 patients with metastatic melanoma collected; 24 patients were treated with ICIs and with MAPKis, respcetively. A non-biased, high-resolution isoelectric focusing of peptides-liquid chromatography-mass spectrometry (HiRIEFLC-MS/MS)-based method, and with proximity ligation assays (PEA) targeting 92 immuno-oncology-related proteins were used.We analyzed the change in protein levels during treatment with a paired t-test, and their association with progression free survival (PFS) with Cox proportional hazards models. Results: HiRIEFLC-MS/MS detected 1,835 proteins.We detected statistically-significant log2-fold-changes in 109 protein levels out of 1,160 proteins tested (not corrected for multiple testing). PDCD-1 had the highest log2-fold change (FC = 1.27) after treatment (p = 0.02). After stratifying for treatment type, PDCD-1 levels increased in patients treated with ICIs (FC = 2.13, p= 0.0008), but not in MAPKis-treated patients. PEA analyses confirmed this observation. The PEA panel showed association between 44 proteins and shorter PFS (pcoefficient <0.05, pLRT<0.05, qLRT<0.05), among them: LGALS1, CSF1, VEGFA, CASP8, CCL2, TNFSF14, ANGPT2, IL10, IL6, and ADGRG1. Of these, increase in plasma levels during treatment of LGALS1, CCL2 and ADGRG1 were associated with longer PFS. HiRIEF LC-MS/MS detected 69 proteins associated with PFS (pcoefficient< 0.05, pLRT< 0.05, qLRT < 0.05). Conclusions: By using HiRIEFLC-MS/MS, we could detect putative treatment predictive proteins in plasma from patients with metastatic melanoma treated with ICIs or MAPKis. Our findings require further validation.