Dominik Dytfeld, MD, PhD1*, Malathi Kandarpa, PhD1*, John R Strahler, PhD2*, Dattatreya Mellacheruvu3*, Suchitra Subramani1*, Stephanie J Kraftson1*, Lambert Ngoka, PhD4*, Alexey Nesvizhskii, PhD3*, Arun Sreekumar, PhD4*, Paul G Richardson, MD5 and Andrzej J Jakubowiak, MD, PhD1
1University of Michigan Comprehensive Cancer Center, Ann Arbor, MI
2Department of Biological Chemistry, University of Michigan, Ann Arbor, MI
3Department of Pathology, University of Michigan, Ann Arbor, MI
4Medical College of Georgia, Augusta, GA
5Medical Oncology, Dana-Farber Cancer Institute, Boston, MA
Despite an increased number of new treatments, multiple myeloma (MM) remains mostly incurable. There is emerging evidence that achieving complete or near complete response (CR/nCR), or at least a 90% reduction of the disease (≥VGPR), in response to initial treatment when MM is most sensitive to chemotherapy is associated with improved progression-free survival (PFS) and possibly overall survival (OS). However, even with the most active regimens, a majority of patients (pts) with newly diagnosed MM achieve less than CR/nCR to initial therapy. The objective of this study is to establish predictors of response and drug resistance by applying proteomic profiling of MM. Here we present the analysis of differential proteomic profiling of baseline plasma cells (PCs) from pts with MM predicting achievement of CR/nCR after completion of a course of first line treatment with lenalidomide (Revlimid®), bortezomib (Velcade®), and dexamethasone (RVD) regimen.
Methods: After obtaining informed consent from pts, we performed quantitative proteomic analysis of PCs isolated from bone marrow of 16 pts with previously untreated MM enrolled at the University of Michigan site in the Phase II portion of the multi-site frontline RVD clinical trial. Eight of the analyzed pts achieved CR/nCR, while the remaining had a lesser response (6 VGPR, 2 PR i.e ≥50% but < 90% reduction of disease). We used two independent proteomic platforms: iTRAQ (Isobaric Tags for Relative and Absolute Quantitation) technique in 8-plex variant, as well as a label-free approach (LF) based on spectra counting. PCs were acquired from bone marrow aspiration and, thereafter, were enriched with a RosetteSep negative selection kit (StemCell Technologies). In iTRAQ experiments, proteins were processed with reagents according to the manufacturer's protocol (Applied Biosystems) followed by SCX fractionation and LC-MS/MS analysis (4800 Plus MALDI TOF/TOF). Peptides from the MM1S cell line were used as an internal reference. The data were analyzed using ProteinPilot software. For LF analysis, proteins were pre-fractionated before trypsin digestion on NuPage Bis-Tris-Gel and subsequently run on LC-ESI-MS/MS on a linear trap mass spectrometer (LTQ Orbitrap). Database search was carried out using X!Tandem followed by Trans-proteomic Pipeline (TPP). A 1.5-fold difference in expression in both platforms was used as a cut-off value.
Results: A total of 926 proteins were identified with high confidence (FDR<1%) in iTRAQ experiments and 1197 in LF. Comparison of proteomic alterations in CR/nCR vs. a lesser response using a quantitative approach revealed the up-regulation of 205 proteins while 85 proteins showed reduced expression. Similarly, using LF analysis, 273 proteins were up-regulated while 32 proteins were down-regulated. Overall, combining data from the two platforms revealed the alteration of a total of 70 (51 up-regulated and 19 down-regulated) proteins in patients showing CR/nCR vs. a lesser response. This therapeutic response signature was analyzed for its biological nuance using GeneGo (www.portal.genego.com). The top 3 concepts discovered in this analysis described the perturbation in c-Myc regulated proteins, which confirms our previous observations (Dytfeld et al. ASMS, 2009). Interestingly, one of these enriched networks of 50 proteins consisted of 26 proteins that were derived from our therapeutic response signature. Included within this network were canonical pathways involving vascular endothelial growth factor receptor 2, p53 binding protein homolog mdm2, p53, nucleoside-diphosphate kinase A/B and Bcl2.
Conclusions: We analyzed proteomic characteristics of patient-derived MM cells using two independent proteomics platforms. As a proof of concept, analysis of PCs obtained from pts enrolled in the frontline RVD trial shows differential expression of 70 proteins in patients who achieved CR/nCR versus those with a lesser response. Consistent with our prior observations, differentially regulated proteins are involved in the c-Myc pathway, which has an established critical role in pathogenesis of MM. Validation studies of candidate proteins are in progress.
This study was supported by a grant from the Multiple Myeloma Research Foundation.