Résumé
Multiple myeloma (MM) is a malignant plasma cell disorder characterized by a wide diversity from the molecular level up to the treatment response. Diversity is supported by cytogenetic abnormalities, especially by non random hyperdiploidy or translocations of IgH locus with recurrent partners. This classification based on cytogenetic aberrations has been confirmed at the molecular level leading to the molecular classification of patients (Zhan et al, Blood 2006 108:2020). This classification defines 7 groups of patients, 5 being linked to cytogenetic abnormalities, CD-1 (cyclin D1/D3–1), CD-2 (cyclin D1/D3–2), HY (hyperdiploid), MS (MMSET), MF (MAF) and 2 to functional characteristics, LB (low bone disease) and PR (proliferation). In the present work, we have addressed the study of mRNA expression profile (global expression profile, GEP) of a very large panel of human myeloma cell lines (HMCL) (n=37) in order to define whether HMCL were or not representative of patient’s diversity. Unsupervised hierarchic clustering of 37 HMCL identified 5 distinct classes (groups A to E). SAM analysis (with 1000 permutations) revealed a set of 213 genes with a FDR ≤ 1% and a ratio ≥ 2. Group A (n=8) gathered HMCL with t(11,14) translocation and genes expressed by CD-1 and CD-2 groups of patients. HMCL of groups B (n=6) and C (n=7) were both IL-6 dependent HMCL exclusively. They were mainly characterized by expression of more than 10 genes coding for cancer testis antigens (GAGE, MAGE, SSX). Group C had also an IL6 signature (IL6, SOCS3) found in one group of HY patients reflecting the presence of several HY HMCL within this group (CHNG et al, Cancer Res 2007, 67/2982). Both groups B and C were related to the PR group. Group D (n=8) had a Maf (Maf, Integrin-beta7) and group E (n=8) a WHSC1 gene signature and gathered HMCL with t(14;16) or t(4;14) translocation, respectively. GSEA analysis revealed very similar results: group A presented a gene signature linked to glycerophospholipid metabolism. Group B overexpressed genes indicative of a plasmablastic signature (Tarte et al., Blood2003,102:592). Group C was characterized by a gene signature linked to the IL-6 pathway and cancer testis antigens, group D by a gene signature relative to the MF group and to the TACIhigh patient’s group. Group E presented a gene signature relative to the MS group. Moreover, in this set of 213 genes, we identified new genes with strong prognosic value for newly-diagnosed patients treated with high dose chemotherapy (both EFS and OS). These data show that HMCL have kept molecular signature of primary myeloma cells in good agreement with their genomic abnormalities (mainly 14q32 translocation). Interestingly, two groups of HMCL matched with PR group suggesting that PR group could be related to two different pathways of proliferation. This GEP analysis showed that heterogeneity of HMCL is representative of patient’s diversity. For the first time, this study establishes at the molecular level the relevance of HMCL to study myeloma and should help to identify target genes in the context of heterogeneity. This large panel of HMCL should be useful to study mechanisms of drug action as well as predict drug efficiency with respect to diversity.