CXCR4 inhibitor AMD3100 disrupts the interaction of multiple myeloma cells with the bone marrow microenvironment and enhances their sensitivity to therapy

Abstract

The interaction of multiple myeloma (MM) cells with their microenvironment in the bone marrow (BM) provides a protective environment and resistance to therapeutic agents. We hypothesized that disruption of the interaction of MM cells with their BM milieu would lead to their sensitization to therapeutic agents such as bortezomib, melphalan, doxorubicin, and dexamethasone. We report that the CXCR4 inhibitor AMD3100 induces disruption of the interaction of MM cells with the BM reflected by mobilization of MM cells into the circulation in vivo, with kinetics that differed from that of hematopoietic stem cells. AMD3100 enhanced sensitivity of MM cell to multiple therapeutic agents in vitro by disrupting adhesion of MM cells to bone marrow stromal cells (BMSCs). Moreover, AMD3100 increased mobilization of MM cells to the circulation in vivo, increased the ratio of apoptotic circulating MM cells, and enhanced the tumor reduction induced by bortezomib. Mechanistically, AMD3100 significantly inhibited Akt phosphorylation and enhanced poly(ADP-ribose) polymerase (PARP) cleavage as a result of bortezomib, in the presence of BMSCs in coculture. These experiments provide a proof of concept for the use of agents that disrupt interaction with the microenvironment for enhancement of efficacy of cytotoxic agents in cancer therapy.

Figure 1

AMD3100 overcomes drug resistance to induction of apoptosis by therapeutic agents in MM cell lines induced by BMSCs in vitro. Apoptosis assay measured using Annexin staining by flow cytometry. (A) AMD3100 did not enhance the effect of bortezomib (2.5-5 nM) for 24 hours on MM.1S cells when those were cultured without BMSCs. However, treatment with AMD3100 (50 μM) restored the sensitivity to bortezomib which was reduced as a result of coculture with BMSCs in the MM cell lines MM.1S (B), RPMI 8226 (C), and OPM-2 (D). Cells were treated with bortezomib either alone or in the presence of AMD3100 (50 μM). This effect was not unique to bortezomib, but AMD3100 was shown to increase the sensitivity of MM.1S cells to treatment with dexamethasone 25 nM, melphalan 10 μM, and doxorubicin 150 nM for 24 hours (E). *P = .003; **P < .001; ***P = .010; #P = .027; ##P = .048. Error bars represent SD.

Figure 2

AMD3100 induces disruption of MM migration and adhesion to fibronectin and BMSCs. MM.1S cells pretreated for 2 hours with AMD3100 (50 μM), bortezomib (2.5-5 nM), or their combination. MM.1S cells were used for adhesion to fibronectin (A), adhesion to BMSC (B), or migration (C) assays. AMD3100 induced 50% inhibition of adhesion to both fibronectin and BMSCs. Bortezomib induced a dose-dependent reduction of adhesion, when at concentration of 5 nM it induced 40% and 60% inhibition of adhesion to fibronectin and BMSCs, respectively, compared with control. The combination of AMD3100 and bortezomib did not show an additive effect compared with the effect of AMD3100 alone. In transwell migration assay of MM.1S cells to 30 nM SDF1, AMD3100 inhibited migration by greater than 60% of control. Bortezomib had a mild dose-dependent effect on migration of MM.1S cells, with 5 nM inhibiting migration by 20% compared with control. The combination of AMD3100 and bortezomib showed significant reduction of migration, specifically the combination of AMD3100 and 5 nM bortezomib showed 75% reduction of migration compared with control. *P = .001; **P = .015; ***P = .005; #P = .003; ##P < .001. Error bars represent SD.

Figure 3

The effect of AMD3100 on PARP cleavage and phosphorylation of Akt in MM cells cocultured with BMSCs. (A) MM.1S cells treated with AMD3100 (50 μM), bortezomib (5 nM), or their combination for 6 or 24 hours and expression of CXCR4 compared with nontreated cells were tested by flow cytometry. AMD3100 inhibited CXCR4 expression at 6 hours, but the expression recovered at 24 hours, whereas bortezomib had no effect on CXCR4 expression at either 6 or 24 hours. (B) MM.1S cells were cultured with BMSCs and treated with AMD3100 (50 μM), bortezomib (0, 2.5, and 5 nM), or their combination overnight in PARP cleavage experiments or for 24 hours in pAkt and pS6R experiments. Immunoblotting with anti-PARP antibody showed that AMD3100 alone did not induce PARP cleavage in MM cells, but it increased the PARP cleavage induced by bortezomib especially at low doses. This shows that the increase of the apoptotic effect of bortezomib induced AMD3100. Moreover, immunoblotting for pAkt and pS6R showed that AMD3100 abolished the phosphorylation of Akt and S6R in MM cells in coculture with BMSCs, which was shown to be a resistance mechanism to bortezomib.

Figure 4

AMD3100 enhances tumor reduction induced by bortezomib in vivo. Mice were treated with AMD3100 (5 mg/kg, daily), bortezomib (0.5 mg/mL, twice weekly), or their combination and were compared with the control untreated group (A). Tumor growth was determined by bioluminescence imaging. Tumor growth in the AMD3100-treated group was similar to the control group, whereas the bortezomib-treated group showed reduction in tumor progression compared with control (P = .041), and the mice treated with the combination of AMD3100 and bortezomib showed significant tumor reduction compared with control (P = .001) and bortezomib alone (P = .021). Each data point represents 3 to 5 mice. Error bars represent SD. (B) Representative bioluminescence images of each treatment group with time. (C) Immunohistochemistry of specimens taken from BM, liver, and spleen. Top panel shows induction of apoptosis of MM cells in the BM detected by TUNEL assay. AMD3100 did not induce apoptosis compared with control, whereas the bortezomib-treated group showed low levels of apoptosis in the BM, and the combination of AMD3100 and bortezomib showed significant induction of apoptosis. The bottom 3 panels represent tumor spread in the BM, liver, and spleen by staining with anti–human CD138, showing that AMD3100 had a minimal effect on reducing the number of plasma cells present in the BM, liver, and spleen. However, bortezomib and more significantly the combination of bortezomib and AMD3100 showed reduction of tumor burden in the BM, liver, and spleen. Images were taken at 40× magnification with Leica DMLB microscope. (D) Quantification of the number of CD138⁺ cells in the BM, liver, and spleen and of TUNEL⁺ cells in the BM. Statistically significant differences were found between numbers of CD138⁺ and TUNEL⁺ cells in the bortezomib-treated group and the bortezomib + AMD3100–treated group. *P < .001; **P = .002; #P = .021; ##P < .016. Error bars represent SD.

Figure 5

AMD3100 mobilizes MM cells to the circulation and sensitizes them to bortezomib-induced apoptosis in vivo. (A) Correlation of the number of viable circulating cells with tumor burden in the bone marrow. The number of circulating cells was in direct linear correlation with the tumor burden. Error bars represent SD. (B) Comparison of the slope of the linear correlations, which represents the amount of the mobilized MM cells per 1 unit of tumor burden. AMD3100 showed intensive mobilization of MM cells. Bortezomib did not show a significant decrease compared with the control group. However, the combination of bortezomib and AMD3100 reduces significantly the number of circulating cells compared with AMD3100 alone. (C) The ratio of apoptotic/total MM cells in the circulation detected by in vivo real-time flow cytometry. AMD3100 did not induce apoptosis compared with the control. Bortezomib showed a dose-dependent induction of apoptosis, which was enhanced by daily treatment with AMD3100. Error bars represent SD. (D) Representative in vivo confocal images of BM niches of control and combination of AMD3100 and bortezomib–treated mice, showing viable MM cells (green), apoptotic MM cells (violet), and blood vessels (red). These images show that apoptotic MM cells were present in the circulation of mice treated with the combination of AMD3100 and bortezomib and absent in the control mice. *P = .046.

Figure 6

AMD3100 mobilizes MM cells and CD34⁺ HSCs in different kinetics and does not induce cytotoxicity of hematopoietic progenitor cells. (A) Mobilization of MM.1S, MM patient sample CD138⁺, and MM patient sample CD34⁺ cells after daily treatment of AMD3100 (5 mg/kg) detected by in vivo flow cytometry. AMD3100 induced similar and continuous mobilization of MM.1S and patient sample CD138⁺ cells at days 2, 3, and 4, whereas CD34 mobilization occurred on day 1 and decreased at days 2 and 3. (B) Colony-forming assays show that treatment with AMD3100 (50 μM), bortezomib (5 nM), or their combination had no significant cytotoxic effect on colony-forming units of erythroid (E), granulocyte-macrophage (MG), macrophage (M), and granulocyte-erythrocyte-monocyte-megakaryocyte (GEMM) origin. *P = .002. Error bars represent SD.