Mature adipocytes in bone marrow protect myeloma cells against chemotherapy through autophagy activation

Abstract

A major problem in patients with multiple myeloma is chemotherapy resistance, which develops in myeloma cells upon interaction with bone marrow stromal cells. However, few studies have determined the role of bone marrow adipocytes, a major component of stromal cells in the bone marrow, in myeloma chemotherapy resistance. We reveal that mature human adipocytes activate autophagy and upregulate the expression of autophagic proteins, thereby suppressing chemotherapy-induced caspase cleavage and apoptosis in myeloma cells. We found that adipocytes secreted known and novel adipokines, such as leptin and adipsin. The addition of these adipokines enhanced the expression of autophagic proteins and reduced apoptosis in myeloma cells. In vivo studies further demonstrated the importance of bone marrow-derived adipocytes in the reduced response of myeloma cells to chemotherapy. Our findings suggest that adipocytes, adipocyte-secreted adipokines, and adipocyte-activated autophagy are novel targets for combatting chemotherapy resistance and enhancing treatment efficacy in myeloma patients.

Keywords: adipocytes; apoptosis; autophagy; chemotherapy resistance; multiple myeloma.

Figure 1. Characterization of cultured, BM-derived mature human adipocytes

MSCs were derived from the BM mononuclear cells of healthy human fetal bones. (A) Flow cytometry analysis shows the immunophenotype of MSCs, which express the MSC-specific surface markers CD44, CD90, and CD166 but not the hematopoietic, stem cell, or endothelial protein markers CD14, CD34, and CD45. Mature adipocytes were generated from MSCs in a 2-week culture in adipocyte medium. (B) Shown is a mature human MSC-derived adipocyte at high magnification (x60) (indicated by a blue arrow); it contained large amounts of lipid droplets (indicated by a black arrow) in the cytoplasm, as stained with Oil red O. The green arrow points to the nucleus of the adipocyte, and a yellow arrow shows an undifferentiated MSC. Flow cytometry analysis shows the level of (C) BODIPY-stained and (D) aP2 protein (a marker of mature adipocytes)-stained adipocytes. The mature adipocytes were further sorted with the antibody against aP2 to obtain a pure adipocyte population (data not shown). Results of five independent experiments are shown.

Figure 2. Co-culture with adipocytes protects MM cells from melphalan-induced apoptosis in vitro

(A) The representative histograms show the percentages of apoptotic ARP-1 cells in 24-hour co-cultures, without or with adipocytes at a ratio of 5:1 in medium with melphalan (0, 12.5, 25, or 50 μM). Summarized data are shown in (B) for U266 cells and (C) for ARP-1 cells treated with melphalan or bortezomib. Also shown are the percentages of apoptotic U266 (D) and ARP-1 (E) cells, cultured alone (no adipocytes) or with mature adipocytes at a ratio of 20:1, 5:1, or 1:1 for 24 hours. Apoptotic cells were detected using an Annexin V-binding assay. Mature adipocytes were generated from the MSCs derived from the BM mononuclear cells of three healthy human fetal bones (A-E) Similar experiments were done co-culture of U266 cells (F) or ARP-1 cells (G) with mature adipocytes isolated from bone marrow aspirate of three MM patients at a ratio of 5:1 for 24 hours, as well as with adipocytes isolated from adult BM aspirates (aADs; H) or from WAT cells (WAT; I). Results of five independent experiments are shown. **P < 0.01.

Figure 3. Addition of adipocyte-conditioned media (CM) inhibits MM cell apoptosis

(A) Annexin V-binding assay shows the percentages of apoptotic U266 and ARP-1 cells co-cultured with adipocytes together (cell-cell), co-cultured with adipocytes but separated by transwell inserts (Transwell), or cultured in medium containing adipocyte-CM in the presence of the drug melphalan (Mel; 25 μM). annexin V-binding assay shows the percentages of apoptotic cells in (B) the MM cell lines ARP-1, ARK, CAG, MM.1S, RPMI8266, and U266; and (C) primary MM cells isolated from the BM aspirates of six MM patients (Pt1 to Pt6) cultured in medium alone, adipocyte-CM, or melphalan (Mel; 25 μM), or a combination (adipocyte-CM+Mel); and (D) the percentages of apoptotic U266 and ARP-1 cells cultured with or without (medium) adipocyte-CM and containing melphalan (Mel), bortezomib (BTZ), dexamethasone (DEX), or doxorubicin (Dox). Cells were cultured for 24 hours. Purified mature adipocytes were generated from the MSCs derived from five healthy human fetal bones. Results of five independent experiments are shown. *P ≤ 0.05; **P ≤ 0.01. Western blot analysis shows the reduced levels of cleaved caspase-9 (c-Cas9), caspase-3 (c-Cas3), and PARP and unchanged levels of caspase-8 and β-actin in U266 and ARP-1 cells that had been cultured in medium containing adipocyte-CM or (E) melphalan (Mel) or (F) bortezomib (BTZ), alone or in combination. Results of three independent experiments are shown.

Figure 4. Addition of adipocyte- conditioned media (CM) activates autophagy in MM cells

(A) Representative images of MDC-stained ARP-1 and U266 cells in medium, with or without adipocyte-CM, at high magnification (×60). (B) MDC-positive cells were counted under the microscope, and the percentages of MDC-positive cells were plotted. (C) Western blot analysis shows the enhanced expression of autophagy proteins Atg3, Atg5, and LC3-I/II, but not Beclin-1, in U266 and ARP-1 cells after 24 hours of culture in medium, with or without adipocyte-CM (AD-CM). Moreover, (D) Western blot analysis shows the reduced expression of autophagy proteins LC3-I/II in U266 and ARP-1 cells cultured in medium with or without adipocyte-CM and treated with or without autophagy inhibitors 3-methyladenine (3-MA; 1 mM) or chloroquine (CQ; 50 μM). In addition, Western blot analysis (lower panels) shows the levels of cleaved (c-) PARP, and β-actin, and an annexin V-binding assay (upper panels) shows the percentages of apoptosis in U266 (E, G) and ARP-1 cells (F, H) in 24-hour cultures treated with (E, F) melphalan (Mel) or (G, H) bortezomib (BTZ), in the presence or absence of adipocyte-CM. Autophagic protein Atg5 was knocked down in ARP-1 and U266 cells by using lentivirus-carried Atg5 shRNAs. (I) Western blot analysis (upper panels) shows the reduced Atg5 expression and annexin V-binding assay (lower panels) shows the enhanced apoptosis in Atg5 shRNA-MM cells as compared with non-target control MM cells. Addition of AD-CM to the cultures at a ratio of 1:5 is shown. β-actin levels served as an internal loading control. Results of three independent experiments are shown. **P ≤ 0.01.

Figure 5. Adipocytes produce adipokines that activate autophagy

(A) Representative images of arrays show the profile of adipokine expression in the supernatants of cultured adipocytes. Medium and the supernatants from cultured MSCs or fibroblasts served as controls. We performed quantitative real-time PCR to determine the adipocyte-expressed (B) and ELISA to determine the adipocyte-secreted adipsin (C) and leptin (D), and the expression and secretion in MSCs served as controls. (E) Western blot analysis shows the increased expression of the autophagy proteins Atg3 and LC3-I/II and unchanged β-actin in U266 cells in 24-hour culture in medium with adipocyte-conditioned media (AD-CM) or the adipokines adipsin (100 ng/ml) or leptin (25 ng/ml). Neutralizing antibodies against adipsin (2 μg/ml) or leptin (2 μg/ml), alone or in combination, were added to U266 cells in medium with adipocyte-CM. Addition of isotype IgG served as control. (F) Western blot analysis shows the reduced expression of the autophagy proteins Atg3 and LC3-I/II in U266 cells treated with the neutralizing antibodies. Antibody combination had a synergistic reduction in the autophagy protein expression as compared with treatment alone. (G) Annexin V-binding assay showed the percentage of apoptotic U266 cells in cultures with melphalan, and with or without adipsin (100 ng/ml) or leptin (25 ng/ml). (H) Annexin V-binding assay shows the percentage of apoptotic U266 cells in cultures with adipocyte-CM, with or without neutralizing antibodies in the presence or absence of melphalan. Representative results from three experiments are shown. **P ≤ 0.01.

Figure 6. Adipocytes activate autophagy via the Stat3 signaling pathway

Western blot analysis shows (A) the enhanced phosphorylated (p) and unchanged non-phosphorylated levels of Stat3 and Jak1 in U266 and ARP-1 cells cultured in medium with or without adipocyte-conditioned media (CM) for 15, 30, and 60 minutes (min), (B) the levels of pStat3 in U266 and ARP-1 cells 24 hours after addition of leptin (25 ng/ml) or adipsin (100 ng/ml), and (C) the levels of pStat3 in 24-hour culture of ARP-1 cells with adipocyte-CM and control IgG or the neutralizing antibodies again adipsin (2 μg/ml) or leptin (2 μg/ml), alone or both. Shown are (D) Western blot analysis results for levels of the autophagy proteins Atg3 and LC3-I/II and unchanged β-actin and annexin V-binding assay results for the percentages of apoptotic ARP-1 cells treated with melphalan (E) or bortezomib (F) in 24-hour culture in medium, without or with adipocyte-CM, Stat3 inhibitor, or both (Stat3 Inh; 10 μM and 20 μM). **P ≤ 0.01. Representative results from three experiments are shown.

Figure 7. Adipocytes inhibit MM cell apoptosis and induce MM drug resistance in vivo

SCID mouse femurs were injected with MM cells ARP-1 or MM.1S or co-injected with MM cells and GFP-labeled mature human adipocytes, and the mice were treated with or without melphalan (Mel). After treatment, BM cells from SCID mouse femurs were flushed out and labeled with anti-human CD138 antibody. The percentages of human CD138⁺ cells are shown in (A). The cells were further sorted by anti-human CD138-coated magnetic beads. An aliquot of CD138⁺ cells was examined with the annexin V-binding assay for apoptosis, and the percentages of apoptotic cells are shown in (B). An aliquot of CD138⁺ cells was stained with the MDC assay for autophagy, and the percentages of MDC⁺ cells are shown in (C). An aliquot of CD138⁺ cells isolated from the BM aspirates of ARP-1 tumor-bearing mice was examined with the Western blot analysis and the levels of phosphorylated (p) Stat3 are shown in (D). In addition, the aliquot of CD138⁻ cells was labeled with anti-GFP antibody to detect the injected GFP-carried adipocytes (E) or with anti-human aP2 antibodies to detect injected adipocytes (F), and counted by flow cytometry. (G) Concentration of human leptin or human adipsin in the serum of mouse injected with or without human adipocytes as measured by ELISA. (H) Representative images of immunohistochemical staining show the expression of perilipin (an adipocyte marker) and CD138 (a MM marker) in the BM of SCID mice bearing ARP-1 cells with mature adipocytes at week 2 or week 4 post cell injection. (I) Representative images of radiography show the osteolytic bone lesions in the femurs of the ARP-1 tumor-bearing mice four weeks post cell injection. The image from the mice without MM cells served as control. Red arrows indicate osteolytic lesions. In addition, MM ARP-1 cells with or without mature adipocytes (ADs) were subcutaneously injected into SCID mice. 4 weeks after the cell injection, mice were intraperitoneally injected with 50 μg/mouse of melphalan (Mel) or PBS (control), twice a week for 3 weeks. (J) Shown are tumor volume changes in the mice after the treatment. The results represent average values from three independent experiments of five mice per group. **P < 0.01.