CD32B Expression Reflects Phenotypic and Functional Heterogeneity in Multiple Myeloma (MM)

Intraclonal heterogeneity (IH) in MM may provide a basis for the selection of resistance to therapy that occurs over time. Model systems to study IH in MM have not been defined so we focused on the expression of CD32B, the low-affinity inhibitory Fcγ-receptor that is a key participant in normal plasma cell apoptosis and a potential target for monoclonal antibody therapy. At diagnosis, CD138+ MM marrow cells are largely CD32B+ (n=72; median 94%, range 1–100) while at relapse the CD32B+ MM subpopulation is significantly smaller in both bone marrow (n=25, median 83%, (10–100), P = 0.02) and extramedullary sites (n=8, median 17% (0–78); P << 0.01). Among human MM cell lines only two of ten tested expressed cell-surface CD32B. We hypothesize that the CD32B phenotype of MM subpopulations contributes to the functional heterogeneity that is critical to the selection of resistance, and sought a model system to study CD32B-related IH. The KMS-12BM cell line (BM), a human MM line from marrow, has CD32B+ and CD32B− subpopulations as MM marrows do, while the KMS-12PE cell line (PE), from the pleural fluid of the same patient, is completely CD32B negative, as is often the case in extramedullary MM. Both BM and PE cells share t(11;14), del 17p and other cytogenetic abnormalities, and both contain intact CD32B genes (FCGR2B). We used 3 populations, BM CD32B+, BM CD32B− and PE (CD32B−), as well as sorted patient MM cells, to study the IH associated with CD32B. RNAi studies demonstrated that the expression of FCGR2B in BM cells is dependent on the Ets family transcription factor PU.1, while PE cells do not express PU.1. RT-PCR of patient MM cells sorted into paired CD32B+ and CD32B− fractions showed PU.1 and FCGR2B (CD32B) expression only in CD138+/CD32B+ cells. Cell-tracking studies with CFSE showed that BM CD32B+ cells gave rise to BM CD32B− cells but that the reverse did not occur. Cell-culture studies showed that BM CD32B+ cells have significantly lower apoptotic rates than BM CD32B− cells, while the latter display greater dependence on insulin for survival and on IGF-1 and MIP-1α for proliferation. In order to better understand the differences between CD32B+ and CD32B− populations, we performed transcript profiling. BM CD32B+ cells have significantly higher expression of the activin inhibin, while BM CD32B− cells overexpress insulin-response genes such as the kinase Akt2. Compared to BM cells, PE cells overexpress germ-cell transcription factors and inflammasome-related genes, while BM cells overexpress tetraspanins (CD9, CD37, CD53) and insulin-response genes. Of interest, there are also significant differences in the expression of BMP and Wnt-related genes (but not Hh or Notch genes). BM cells overexpress BMPR1A, FZD2 and Wnt5A while PE cells overexpress CRIM1, FZD1 and LEF1. These data indicate that functional IH exists in MM, that the KMS-12 cell lines provide a useful model for its study, and that MM may originate in a CD32B+ cell that begets CD32B− heirs. We continue to study PU.1 and the generation of CD32B− cells by CD32B+ cells; to validate transcript data using sorted patient specimens; and to assess with RNAi the contributions of BMP, Wnt and insulin-response genes to the survival and proliferation of CD32B+ and CD32B− cells, seeking novel therapeutic targets on which selection of resistance may depend.