Melphalan Affects Genes Critical for Myeloma Survival, Homing, and Response to Cytokines and Chemokines

Introduction: We have previously used global gene expression (GEP)-based clinical pharmacogenomics of dexamethasone, thalidomide, pomalidomide and bortezomib (Bor) to understand their mechanism of actions and how this impacts their clinical efficacy in multiple myeloma (MM) patients.  High dose therapy with melphalan (Mel) followed by autologous stem cell transplantation is a major treatment regimen for MM and consequently assessing the pharmacogenomics of Mel is clinically relevant. Importantly not all patients benefit from Mel exposure and, therefore, it remains important to understand the molecular mechanism of its actions and how they underlie treatment resistance.

Materials and methods: 252 newly diagnosed MM patients randomized to Total Therapy (TT)4 (GEP70 low-risk, N=210) and TT5 (GEP70 high-risk, n=42) clinical trials received single administration  of Bor (1.0mg/m2) followed after 48 hrs latter by a single administration of Mel (10mg/m2).  BM aspirates were obtained at baseline, 48 hrs post-Bor and 48 hrs Post-Mel. Purified CD138-selected MM cells underwent GEP analysis using the Affymetrix U133plus2 microarrays and the results generated were analyzed using Gene Set Enrichment Analysis (GSEA) and Ingenuity. The effect of Mel on expression of cell surface receptors in MM cell lines was validated by flow cytometry.

Results: Expression of 176 probe sets was changed 48 hours after a single Bor administration at FDR < 0.01, and expression of 5117 additional probe sets was further changed after Mel. Expression of 6309 probe sets was changed when comparing post-Mel to baseline, of these 108 were also changed post-Bor and 4043 overlapped with changes observed between post-Mel and post-Bor. By utilizing GSEA and Ingenuity we identified the top pathways associated with Mel activity including activation of p53, nitrogen metabolism and metabolism of xenobiotics by p450, whereas Bor was associated with proteasome activation. Bor and Mel both downregulated pathways related to cell cycle and DNA replication and damage response. Top listed genes differentially expressed between baseline and post-Mel and/or post-Bor and post-Mel and reportedly linked to MM pathogenesis include underexpression of IRF4, ASPM, MYC and NEK2, and upregulation of TNFSF10 (TRAIL), MDM2, BAX and KLF9. Among the top upregulated genes by Mel were also set of 7 cell surface receptors (MERTK, CXCR4, OGFRL1, INSR, TGFBR2, S1PR1, IL1R2) and 5 cytokines (AREG, TNFSF8, BDNF, IGF1, TNFSF15). We initially focused our analysis on MERTK and CXCR4 which have been previously implicated in MM and whose expression was upregulated by 2.5 (FDR < 4.6x10^-39) and 1.8 (FDR < 3.9x10^-36) folds, respectively.  Increased expression of MERTK and CXCR4was associated with shorter progression-free survival (PFS) and over survival (OS) in our Total Therapy trials, including analysis restricted to GEP70 high-risk patients in TT3 and TT5.

Moreover, GEP of paired MM cell samples obtained from focal lesion and random BM sites of the same patient (n=170) showed reduced CXCR4 expression in MM cells residing within focal lesions (FDR < 6.1x10^-5), further implicating intra-patient heterogeneity of CXCR4 expression in distinct BM niches with response to Mel. To validate direct effect of Mel on these factors at the protein level, flow cytometry analysis for MERTK and CXCR4 was performed on MM cell lines (n=3) following treatment with Mel (5-10µM) or Bor (2.5-5nM) for 72 hrs. Mel but not Bor increased mean fluorescent intensity of MERTK and CXCR4 by 3.1±0.5 (p < 0.05) and 5.1±0.5 (p<0.04) folds, respectively, and the percentage of MERTK- and CXCR4-expressing MM cells by 11.9±2.5 (p < 0.04) and 4.1±0.1 (p < 0.002) folds, respectively,  compared to control vehicle-treated cells.

Conclusions: The upregulation of cell surface receptors and growth factors by Mel suggests that exposure to the drug promotes the retention of MM cells within the BM and their reliance on the BM microenvironment.  This pharmacogenomics approach is useful as it can identify previously unrecognized biomarkers and pathways associated with Mel activity and drug resistance in MM patients.