Investigation of the CXCL12-CXCR4/CXCR7 Axis in Multiple Myeloma (MM): Adhesion Molecules As Novel Targets of the Potent Proteasome Inhibitor Carfilzomib

Introduction: The interaction of malignant plasma cells (PC) with the bone marrow (BM) microenvironment (BMM) is fundamental for multiple myeloma (MM) pathogenesis, with BMM playing a key role in disease development, progression and patients’ symptoms. The CXCL12/CXCR4 axis is pivotal in this cross-talk. By activating CXCR4, CXCL12 drives MM cells to the protective BM and facilitates environment-mediated drug resistance. Whether or not CXCR7, the second receptor for CXCL12, is expressed by malignant PC has not been determined yet. In AML and CLL, the phosphorylation status of CXCR4 proved important: phosphorylation at Ser339 by PIM-1 kinase promoted re-externalization of CXCR4 and its reactivation to CXCL12 stimulation (Grundler et al. JEM 2009; Decker et al. Mol Cancer Therapeutics 2014).

Methods: Human MM cell lines (MMCLs) U266, L363, NCI-H929 and primary BM specimens (patient samples n=9, median BM-infiltration: 30%; range: 15-90%) were used. MMCLs and primary MM patient specimens were assessed for CXCR4, CXCR7 and CD138 surface (s) expression via flow cytometry. For CXCR4 analysis, the conformation independent antibody clone 4G10 was used for both flow and image cytometry experiments. The MMCLs were cultured alone or with the stroma cell line M2-10B4, and treated with the novel proteasome inhibitor carfilzomib (20, 50, 100nM), either alone or combined with the CXCR4 antagonist AMD3100 (50µM) or anti-CXCL12 "Spiegelmer" NOX-A12 (100nM). Cell viability was assessed by PI/Annexin staining. CXCR4 downstream pathways and PIM-1 kinase expression were analyzed by Western blot.

Results: All MMCLs and primary MM cells expressed sCXCR7. sCXCR7 expression was represented as mean (% of positive live cells) +/- SD: U266 (89.1 +/-7.1), L363 (82.1 +/-7.5) and NCI-H929 (82.4 +/-8.2) were largely CXCR7 positive. The sCXCR7 expression of unsorted primary BM cells was 21.7 +/- 10.5 (n=9), being slightly higher than their sCXCR4 expression: 14.6 +/- 7.5, (n=7). When analyzing only the sCD138 positively stained population, the malignant PC showed an even higher level of marker expression: sCXCR7: 78.4 +/- 26.6 (n=8) and sCXCR4: 65.5 +/- 39.1 (n=7).

Within our in vitro coculture model (Udi J,..Engelhardt. Br J Haematol 2013; Zlei,...Engelhardt. Exp Hematol 2007; Schüler,..Engelhardt. Expert Opin Biol Ther 2013), CXCL12 stimulation of U266 led to CXCR4 internalization, CXCR4 phosphorylation at Ser339 and actin polymerization. The latter was abrogated by AMD3100 and NOX-A12, proving the functionality of CXCL12-CXCR4 in our MM model. In U266, CXCL12 did not induce CXCR7 internalization. Carfilzomib was rigorously tested in 3 different conditions: it proved cytotoxic for U266 and L363 with pulse (t=1h), continuous treatment (t=3d) and to a lesser extent in coculture (t=3d). Neither AMD3100 nor NOX-A12 significantly reversed the stroma protection. Apart from its cytotoxic effects, carfilzomib (t=3d) reduced sCD138 in U266 or L363 in coculture. As a pulse treatment (t=1h) in U266 monoculture, carfilzomib reduced sCD138 and sCXCR4 without affecting sCXCR7. A 1h pulse of carfilzomib had no effect on pERK/ERK (n=4), nor did it change CXCR4 protein levels (n=2). However, pCXCR4 (Ser339) and its phosphorylating kinase PIM-1 were significantly reduced by carfilzomib in a dose dependent fashion (n=4).

Conclusion: CXCR7 is widely expressed on malignant PC and its contribution to the CXCL12-CXCR4 axis in MM is worthy of future study. Carfilzomib shows strong cytotoxic activity in monoculture, which cannot be completely reversed by stroma protection. Carfilzomib appears to downregulate sCXCR4 by reduction of pCXCR4 via decreased PIM-1 protein levels. These data could explain its potency and why the CXCR4/CXCL12-inhibitors AMD3100 and NOX-A12 do not induce additive cytotoxicity. Moreover, and to the best of our knowledge unknown to date, these results suggest a link between CXCR4 and PIM-1 in MM. Therefore, analyzing the (off-) target effects of anti-myeloma substances on the MM BMM interaction seems to be a promising approach to discover optimal drug combinations and new therapeutic targets.