Comparative Proteomic Profiling of Sera from Patients with Refractory Multiple Myeloma Reveals Pathways and Biomarkers Predicting Response to Bortezomib-Based Therapy

Introduction: Identification of drug-resistance biomarkers in multiple myeloma (MM) is a first step to individualization of treatment. Human serum is easily collected and does not require expensive preparation to analyze, therefore it seems the most suitable material to search for biomarkers.

Objective: The goal of this study was to establish serum biomarkers predicting response/resistance to PAD (bortezomib, adriamycin, dexamethasone) chemotherapy in patients with refractory MM.

Material and Methods:Comparative proteome analysis was performed on pretreatment serum samples obtained from 56 proteasome inhibitor-naïve, transplant-eligible patients qualified for PAD chemotherapy due to CTD (cyclophosphamide, thalidomide, dexamethasone) refractory disease. Samples were collected after patients signed informed consent according to Declaration of Helsinki. Subjects were classified into four groups based on the response to PAD: 13 patients achieved complete response (CR group), 10 revealed very good partial response (VGPR group), 17 - partial response (PR group) and 16 - stable disease or progressive disease (SD/PD group). Serum proteins after trypsin digestion were analyzed using Q-Exactive hybrid quadrupole-Orbitrap mass spectrometer coupled to the chromatograph Dionex 3000 Ultimate nanoLC (Thermo Scientific). Obtained data were analyzed using MaxQuant/Perseus software.

Results and discussion: 632 proteins were identified with a false discovery rate of 1% (FDR). A protein was considered to be differentially accumulated if the difference between experimental groups was statistically significant (p< 0.05) and the minimum fold change was ± 1.5. Proteins identified with a minimum of 2 peptides were considered significant. The proteomic signature revealed 52 proteins, which differentiated all analyzed experimental groups (ANOVA significant). The most differential proteins revealed comparison between CR/VGPR vs PD/SD groups. Eleven up-regulated and 35 down-regulated proteins were identified in sera of patients derived from PD/SD (non responders) group compared to CR/VGPR (responders) among them: sulfhydryl oxidase (fold change 2.56), kallistatin (fold change 0.32), gelsolin (fold change 0.50), thyroxine-binding globulin (fold change 0.28), PDZ and LIM domain protein 1 (fold change 2.38), corticosteroid-binding globulin (fold change 0.32), and phosphatidylinositol-glycan-specific phospholipase D (fold change 0.42). Also principal component analysis (PCA) differentiated the CR/VGPR and PD/SD patients. Proteins like thyroxine-binding globulin or corticosteroid-binding globulin seem to be nonspecifically deregulated, however sulfhydryl oxidase may have a causative relation to bad prognosis. Sulfhydryl oxidase plays an important role in growth regulation and has metastatic potential. Overexpression of this enzyme was presented in pancreatic, prostate and breast cancer. Moreover, it was showed that up-regulation of this enzyme predicts a poor prognosis in patients with breast cancer (Katchman et al. 2013).

Conclusions: We identified proteins, some of which may serve as potential prognostic factors predicting response to bortezomib-based regimens in patients with refractory MM. Label-free proteomic analysis is a useful method to screen for proteins differentially expressed in the sera of multiple myeloma patients revealed different response to chemotherapy.