Inhibition of hypoxia-inducible factor-1 function enhances the sensitivity of multiple myeloma cells to melphalan Free

Despite new therapeutic approaches, multiple myeloma remains incurable. Myeloma cell intrinsic factors and microenvironmental stimuli contribute to the unfavorable biology of the disease (1). Cytogenetic abnormalities, including immunoglobulin H translocations, confer a growth advantage to myeloma cells through deregulated expression of cyclin D family proteins. Several cytokines and chemokines also support the growth and survival of myeloma cells, which ultimately leads to chemotherapy resistance. The biological functions of these intrinsic factors and microenvironmental stimuli are mediated by a series of transcription factors, including those of the nuclear factor κB, signal transducers and activators of transcription 3, and Forkhead transcription factor families (1). These transcription factors regulate the expression of anti-apoptotic proteins and enhance cytokine production, which in turn induces pathologic angiogenesis, osteoblast activation, and the growth of myeloma cells in an autocrine fashion.

Hypoxia-inducible factor-1 (HIF-1) is a transcription factor that serves as a master regulator of cellular responses to hypoxia (2) and regulates dozens of genes required for adaptation to hypoxia. HIF-1 is composed of two subunits, α-subunit and β-subunit. Although the expression of hypoxia-inducible factor-1 β (HIF-1β)-subunit is constitutive, HIF-1α protein levels are regulated in response to oxygen tension. Under normoxic conditions, HIF-1α is degraded by the proteasome, whereas HIF-1 remains inactive. In hypoxia, HIF-1α is stabilized and forms a complex with HIF-1β that allows HIF-1 to function as a transcription factor. Thus, HIF-1 is activated only during hypoxia under normal physiologic conditions. In contrast, HIF-1 is frequently activated in cancer cells, including under normoxic conditions (3), by oncogene products, impaired activity of tumor suppressor genes (4), or the accumulation of metabolic glucose intermediates (5, 6). Constitutively activated HIF-1 enhances neovascularization, increases glucose metabolism, and induces the expression of anti-apoptotic proteins in cancer tissues. HIF-1 is thought to be one of the most important molecular targets in the treatment of cancer (7), and a variety of chemical inhibitors for HIF-1 have been introduced to date (8).

In contrast to solid tumors, only a limited number of studies on the role of HIF-1 in hematologic malignancies have been published to date. HIF-1 is activated by cytokine stimuli in leukemic cells under normoxic conditions (9). HIF-1 regulates vascular endothelial growth factor production in leukemic cells, which supports cell survival in an intracellular autocrine fashion (9). HIF-1 has important functions in multiple myeloma. First, numerous important genes for the progression of multiple myeloma [i.e., vascular endothelial growth factor, stromal cell-derived factor-1, and myeloid cell leukemia sequence-1 (MCL-1)] are putative targets of HIF-1 (3). Second, several transcription factors that play important roles in the pathophysiology of multiple myeloma regulate HIF-1 expression and function. signal transducers and activators of transcription 3 protects against HIF-1 degradation (10, 11). Several studies have shown that nuclear factor κB regulates HIF-1α transcription (12–14). Proteins from the Forkhead transcription factor family, which are often inactivated in myeloma cells, negatively regulate HIF-1 function (15). In addition to these basic findings, recent studies have proposed that HIF-1 may be a molecular target of certain new generation antimyeloma agents, including bortezomib (16) and heat shock protein 90 inhibitors (17). However, the biological significance of HIF-1 in multiple myeloma is under investigation, and to date, no study has investigated a new HIF-1–based strategy for multiple myeloma treatment.

We investigated the molecular function of HIF-1 in multiple myeloma cells using established myeloma cell lines and primary CD138⁺ myeloma cells derived from patients. HIF-1 is constitutively activated in these cells, even under normoxic conditions, and insulin-like growth factor-1 (IGF-1), one of the most crucial growth factors for myeloma cells, was clearly shown to enhance HIF-1 levels. In addition, blocking HIF-1 function was found to heighten the sensitivity of myeloma cells to melphalan. Our observations suggest that inhibiting HIF-1 function may be a new and effective therapeutic approach for multiple myeloma.

Human MM cell lines KMM-1, U266, and RPMI8226 were obtained from Japanese Collection of Research Bioresources and cultured in RPMI-1640 media (Sigma Chemical) containing 10% FCS, 25 U/mL penicillin, and 25 ng/mL streptomycin (GIBICO). For culture under hypoxic conditions, we used a hypoxic chamber (Juji Field).

IGF-1 and interleukin 6 were purchased from PeproTech EC. LY294002 was from EMD Biosciences, Inc. U0126 was from Cell Signaling Technology. Echinomycin, melphalan, and dexamethasone were sourced from Sigma.

Whole cell lysates or nuclear and cytosolic fractions of lysates were prepared as described elsewhere (9). Nuclear extracts were prepared as described previously (18). Lysate proteins were size fractionated by SDS-PAGE and then electroblotted onto polyvinylidene difluoride membranes. The blots were probed with the following antibodies: anti–HIF-1α, anti–BCL-X, and anti–BCL-2 (BD Transduction Laboratories); anti–phospho-AKT (Ser 473), anti-AKT, and anti–β-actin (Cell Signaling Technology); anti–MCL-1, antisurvivin, and anti-general transcription factor IIH (TFIIH; Santa Cruz); anti–HIF-1α (Novus Biologicals). The blots were visualized using a chemiluminescence detection kit (LumiGLO, Cell Signaling).

Total cellular RNA was isolated using an RNeasy Mini Kit (Qiagen). Reverse transcription was done using an M-MLV reverse transcriptase kit (Invitrogen), according to the manufacturer's instructions. Quantitative analysis of survivin mRNA was done using predeveloped TaqMan assay reagents (Applied Biosystems). 18S RNA was used as an internal control for real-time RT-PCR.

The pSuper HIF-1α/siRNA vector and empty pSuper vector were generous gifts from Dr. Daniel Chung (Massachusetts General Hospital, Harvard Medical School; ref. 19). To generate stable transfectants, KMM-1 cells were transfected and selected with 0.5 μg/mL puromycin (Sigma) for 14 d, and clones were selected and screened for HIF-1α silencing by immunoblot analysis.

Cellular apoptosis was determined using an Annexin V–Azami-Green Apoptosis detection kit (Medical and Biological Laboratories), according to the manufacturer's instructions. For primary myeloma cells, apoptotic cells were detected with the Hoechst staining method because of the limited number of cells.

This study was approved by the institutional review board of the University of Yamanashi. After informed consent, primary multiple myeloma cells were purified from bone marrow aspirates with anti-CD138 magnetic beads (Miltenyi Biotech) in accordance with the manufacturer's protocol. These cells were further analyzed using Wright-Giemsa staining, and a purity of better than 90% was confirmed. Clinical features are summarized in Table 1. CD19-positive B lymphocytes were collected from peripheral blood of healthy volunteers using anti-CD19 magnetic beads (Miltenyi Biotech) and used as normal B lymphocytes.

Immunofluorescence microscopic analysis for HIF-1α was done in accordance with previous reports (9). Images were obtained using an Olympus BX50 microscope (Olympus) and analyzed using the VB-7010 system (KEYENCE).

A luciferase reporter plasmid vector containing hypoxia-responsive elements (20) was a generous gift from Dr. Shinae Kondo (University of Kyoto). The reporter plasmid was introduced into the myeloma cell lines by lipofection with the pRL-TKLuc internal control plasmid (Promega). Forty-eight hours after transfection, the cells were harvested and subjected to dual luciferase assay, according to the manufacturer's instructions (Promega).

Statistical analysis was done using the Student's t test; P < 0.05 was considered statistically significant.

Initially, we investigated the expression of HIF-1α by Western blotting in the three kinds of multiple myeloma-derived cell lines: KMM-1, U266, and RPMI8226. As shown in Fig. 1A (lanes 1, 4, and 7), HIF-1α protein was expressed in all three myeloma cell lines under basal culture conditions. By contrast, normoxic HIF-1α expression was not observed in normal bone marrow mononuclear cells (data not shown). Because serum starvation may enhance HIF-1α expression (21), we compared expression of the protein in the presence or absence of serum and found that serum starvation did not enhance HIF-1α expression (Supplementary Fig. S1). We also analyzed HIF-1α expression in primary multiple myeloma cells isolated from bone marrow aspirates from two myeloma patients (patients 1 and 2 in Table 1) using immunofluorescence microscopy. CD19⁺ B lymphocytes were isolated from healthy volunteers and used as the normal control. Primary myeloma cells expressed HIF-1α protein under basal conditions (Fig. 1B). In contrast, expression of the protein was not recorded in normal B lymphocytes under normoxic conditions, whereas culture of these cells under hypoxic conditions drastically induced expression of the protein. These results suggest that constitutive expression of HIF-1α is specific for multiple myeloma cells.

In addition to constitutive activation, it is well known that growth factors for cancer cells may enhance HIF-1 function in several cancer cell types. To confirm this in our system, we evaluated whether interleukin 6 and IGF-1, both known to be potent growth factors for myeloma cells, enhanced HIF-1α expression in these cells. As shown in Fig. 1A (lanes 3, 6, and 9), IGF-1 significantly induced HIF-1α expression in all myeloma cell lines; by contrast, interleukin 6 exhibited only modest effects on HIF-1α expression levels (Fig. 1A, lanes 2, 5, and 8). In accordance with the results from the established myeloma cell lines, HIF-1α protein was drastically up-regulated after culturing the primary myeloma cells with IGF-1 (Fig. 1B). We also found that IGF-1 failed to induce HIF-1α in normal B lymphocytes (Fig. 1B).

To confirm that IGF-1 not only induces the expression of HIF-1α but also activates HIF-1, we did a reporter assay using a reporter plasmid that contained hypoxia-responsive elements. As shown in Fig. 1C, treatment with IGF-1 significantly increased the transcriptional activity of HIF-1.

We explored the molecular mechanisms that induce HIF-1α expression in myeloma cells. Signal transduction molecules, including AKT and MAPK, play important roles in HIF-1α expression under normoxic conditions (22, 23). Both signal transduction molecules play important functions in the growth and survival of multiple myeloma cells (24, 25). We focused our attention on these two signal transduction pathways. We found that both pathways are constitutively activated in multiple myeloma cell lines and that IGF-1 drastically enhanced the phosphorylation of these molecules (Fig. 2A). Based on these observations, we treated the cells with a chemical inhibitor for each signal transduction pathway: LY294002 to inhibit PI3-K and U0126 to inhibit MAPK. Both LY294002 and U0126 reduced the basal levels of HIF-1α expression in all three myeloma cell lines (Fig. 2B). These inhibitors also prevented IGF-1–induced HIF-1α accumulation in the myeloma cell lines (Fig. 2C). These results suggest that constitutive activation of PI3-K and MAPK in myeloma cells is required for basal HIF-1α expression in myeloma cells and that significant activation of these pathways by IGF-1 can lead to enhanced HIF-1α expression in myeloma cell lines.

We explored the effects of inhibiting HIF-1 functions on the growth and survival of multiple myeloma cells. We established stable HIF-1α knockdown KMM-1 subclones (1 and 12) using siRNA against HIF-1α. In these clones, baseline HIF-1α expression was strongly suppressed (Fig. 3A, comparing lane 1 to lanes 3 and 5); however, expression of the β subunit of HIF-1 was not affected in these cells, indicating that the siRNA specifically blocked HIF-1α expression. In addition, IGF-1 did not induce HIF-1α elevation in the HIF-1α–silencing clones (Fig. 3A, comparing lane 2 to lanes 4 and 6). These results suggest that the siRNA against HIF-1α functions efficiently. Initially, we compared growth of the clones with the growth of the parental cells. Although HIF-1α knockdown did not affect the growth or survival of KMM-1 cells under basal culture conditions (data not shown), these HIF-1α–silencing clones exhibited higher sensitivity to melphalan, one of the most common antimyeloma drugs, at 10 and 25 μmol/L compared with the parental KMM-1 cells (Fig. 3B). These results are consistent with several lines of evidence that the inhibition of HIF-1 functions may sensitize cancer cells to chemotherapeutic agents (26, 27).

Several studies have suggested that IGF-1 plays a key role in drug resistance in multiple myeloma cells and that the inhibition of IGF-1 signals may sensitize myeloma cells to melphalan in a mouse model (28–31). In addition, we found that IGF-1 significantly protects against melphalan-induced apoptosis in parental KMM-1 cells (Fig. 3C). Considering that IGF-1 drastically enhances HIF-1α levels in myeloma cells, we postulated that IGF-1 may exert its protective effects on melphalan-induced apoptosis through HIF-1 activation. To this end, we treated HIF-1α–silencing cells with melphalan in the presence or absence of IGF-1. As expected, IGF-1 failed to protect the HIF-1α knockdown cells from melphalan-induced apoptosis (Fig. 3C), suggesting that HIF-1α up-regulation is a prerequisite for the anti-apoptotic effect of IGF-1 in myeloma cells.

Considering that HIF-1 inhibition may be a new strategy to increase the efficacy of melphalan and to overcome IGF-1–induced drug resistance in myeloma cells, we decided to explore the effects of echinomycin, which blocks the DNA binding activity of HIF-1 (32). First, inhibition of HIF-1 function by echinomycin was confirmed by a reporter assay with a vector containing hypoxia-responsive elements. We found that 1 nmol/L of echinomycin was sufficient to block HIF-1 activity in KMM-1 and RPMI8226 cells (Fig. 4A). By contrast, 10 nmol/L of the inhibitor was required to block HIF-1 function in U266 cells (Fig. 4A). Although it has been reported elsewhere that echinomycin enhances HIF-1α levels under normoxic conditions in the human HepG2 hepatoma cell line (33), we failed to observe HIF-1α elevation in the cell lines used in this study (data not shown). Based on these observations, we used the inhibitor at 1 nmol/L for KMM-1 and RPMI8226 cells and at 10 nmol/L for the U266 cells throughout the study. We treated each myeloma cell line with echinomycin and with various concentrations of melphalan for 24 hours, and we monitored the ratio of apoptotic cells by using Annexin V staining. Consistent with siRNA silencing of the KMM-1 cells, treatment with echinomycin alone did not induce apoptosis in myeloma cell lines (Fig. 4B). However, echinomycin significantly enhanced melphalan-induced apoptosis in these cells (Fig. 4B). Similar results were obtained using a caspase-3 activation assay. Echinomycin alone did not significantly increase the level of caspase-3 cleavage in myeloma cells during the observation period. However, it drastically enhanced melphalan-induced caspase-3 cleavage (Fig. 4C). Importantly, the apoptosis-enhancing effects of echinomycin were not observed in HIF-1α–silencing clones (Supplementary Fig. S2). This result indicates that the inhibitor exerts its apoptosis-enhancing effects specifically through inhibition of HIF-1 activity. We also examined whether the inhibitor inhibited the anti-apoptotic effects of IGF-1 on melphalan-treated cells. As illustrated in Fig. 4D, IGF-1 pretreatment significantly reduced melphalan-induced apoptosis. However, adding echinomycin completely canceled the IGF-1 anti-apoptotic effect against melphalan-induced apoptosis in these cell lines. These results suggest that HIF-1 is an attractive agent to increase the antimyeloma efficacy of melphalan. Because we know that IGF-1 also decreases the response of myeloma cells to other cytotoxic chemotherapeutic reagents, that is, doxorubicin (31), we investigated whether the inhibitor also enhanced the response of myeloma cells to doxorubicin. We found that echinomycin also enhanced apoptosis after doxorubicin treatment and abrogated the protective effects of IGF-1 (Supplementary Fig. S3). In contrast, the inhibitor did not show these effects when treated with dexamethasone (data not shown).

To test the relevance of our findings, we used CD138⁺ primary myeloma cells harvested from six patients at diagnosis (patients 3–8 in Table 1). The cells were cultured under several different conditions for 24 hours, and apoptotic cells were identified by Hoechst 33342 staining (Fig. 5). Consistent with the results from myeloma cell lines, 1 nmol/L echinomycin did not promote apoptosis in primary myeloma cells. However, melphalan induced apoptosis at significant levels in all purified myeloma cells. Furthermore, the addition of echinomycin enhanced the cytotoxic effect of melphalan against primary myeloma cells (Fig. 5). We also analyzed whether treatment with echinomycin enhanced the toxicity of melphalan to normal B lymphocytes. Consistent with the finding that HIF-1 is not activated in normal B lymphocytes (Fig. 1B), the inhibitor did not show apoptosis-enhancing effects on these cells. Together, these findings indicate that our hypothesis that the inhibition of HIF-1 function enhances the sensitivity of myeloma cells to melphalan may also apply to primary myeloma cells.

We tried to identify downstream molecules that are directly involved in the anti-apoptotic effects of HIF-1 in myeloma cells. Under physiologic and pathologic conditions, HIF-1 regulates numerous anti-apoptotic genes, including Bcl-2 family proteins (e.g., MCL-1) and survivin, a member of the inhibitor of apoptosis protein family (34). In addition, these proteins are known to be essential for myeloma cell survival. Thus, we focused our attention on these proteins as downstream effectors. As shown in Fig. 6A, echinomycin significantly suppressed the expression of survivin, whereas levels of MCL-1, BCL-2, and BCL-xL proteins were not affected in the three myeloma cell lines that we tested. Consistent with the down-regulation of survivin at the protein level, echinomycin suppressed survivin expression at the mRNA level (Fig. 6B), suggesting that survivin is transcriptionally regulated by HIF-1. To confirm this notion, we examined the expression of survivin in stable HIF-1α–silencing KMM-1–derived clones (1 and 12). In agreement with our previous results, basal levels of survivin were decreased in the HIF-1α–silencing clones (Fig. 6C, comparing lane 1 to lanes 3 and 5), whereas expression levels of MCL-1, BCL-2, and BCL-xL were not altered. IGF-1–induced survivin expression was abrogated in stable HIF-1α–silencing clones (Fig. 6C, lanes 2, 4, and 6). In addition, IGF-1 treatment clearly enhanced survivin expression, and echinomycin significantly blocked IGF-1–dependent survivin expression (Fig. 6D).

HIF-1 was originally identified as a master transcription factor that regulates cellular hypoxic responses (2). In addition to these primary functions, evidence suggests that this transcription factor plays a central role in cancer biology (3, 7). Recent studies have suggested that HIF-1 is also involved in the pathophysiology of multiple myeloma. Using a myeloma animal model, Asosingh et al. (35) showed that oxygen levels in the bone marrow microenvironment regulate microvessel density in bone marrow, a significant characteristic of the disease. They found that HIF-1 contributes to this process by enhancing vascular endothelial growth factor production in myeloma cells. Colla et al. (36) reported that reduced expression of the tumor-suppressor gene known as ING-4 (inhibitor of growth family member 4) was correlated with enhanced angiogenesis in myeloma cells. Interestingly, they also showed that inhibition of ING-4 by siRNA induced HIF-1 activation and the subsequent production of angiogenic cytokines in myeloma cells. Although these studies also reported that HIF-1 is expressed in established murine and human myeloma cell lines, our study provides the first evidence that HIF-1α accumulates in primary CD138⁺ myeloma cells under normoxic conditions.

In addition to the baseline activation of HIF-1, IGF-1 clearly enhances HIF-1α protein levels in established myeloma cell lines and in primary CD138⁺ myeloma cells. These results suggest that HIF-1α lies downstream of the IGF-1 signaling pathway in myeloma cells. This is supported by the previous report that inhibition of the IGF-1 receptor suppresses HIF-1 activity in the human myeloma cell line MM-1S (31). IGF-1 plays important roles in the pathophysiology of multiple myeloma; it induces growth not only of established multiple myeloma cells but also of primary myeloma cells from patients in vitro (24, 25). This cytokine confers multiple drug resistance on myeloma cells (25, 30). Importantly, clinical studies have revealed that increases in serum IGF-1 levels may be correlated with poor outcomes of multiple myeloma (37). IGF-1 is also known to activate HIF-1 in several cancer cell lines. Fukuda et al. reported that exposing colon cancer cells to IGF-1 leads to activation of HIF-1 and results in vascular endothelial growth factor secretion (22). They proposed a model whereby IGF-1 regulates protein synthesis of the α-subunit of HIF-1 through the activation of the PI3-K/AKT and MAPK pathways (22). Elevation of HIF-1α by IGF-1 was also reported in retinal epithelial cells, and PI3-K and MAPK were also shown to be involved in this process (23). In accordance with these observations, our study found that inhibitors for PI3-K and MAPK can suppress IGF-1–induced HIF-1α elevation. In solid tumors, HIF-1 is believed to be one of the most important molecular targets, and numerous reagents that target HIF-1 have been introduced in this field (8). In addition to solid tumors, our results suggest that HIF-1 may be an attractive molecular target in multiple myeloma. Stable silencing of HIF-1α by siRNA caused increased sensitivity to melphalan and loss of the IGF-1 protective effects against apoptosis. Moreover, we used a pharmacologic inhibitor, echinomycin, to block HIF-1 functions. The inhibitory function of this reagent is confirmed by in vitro and in vivo studies (38). In agreement with studies using siRNA against HIF-1α, echinomycin clearly sensitized myeloma cells to conventional chemotherapeutic agents and abrogated the protective effect of IGF-1 on melphalan-induced apoptosis. Inhibition of HIF-1 also enhanced the efficiency of doxorubicin, another key chemotherapeutic reagent for multiple myeloma (Supplementary Fig. S3). In addition to strategies that directly block HIF-1 function (i.e., siRNA, specific pharmacologic inhibitors), recent studies highlight the way in which several new-generation antimyeloma reagents already used in clinical settings exhibit anti–HIF-1 activity. Bortezomib is one of the most important new-generation agents in the treatment of myeloma (39). Although the precise molecular mechanism remains unclear, bortezomib inhibits the transcriptional activity of HIF-1 in cervical cancer cells (40), breast cancer cells (41), and myeloma cells (16). Inhibitors of heat shock protein 90 are also effective in controlling multiple myeloma (17). Heat shock protein 90 is a molecular chaperone that stabilizes numerous proteins. Interestingly, Issacs et al. (42) reported that heat shock protein 90 regulates HIF-1α stability. Our preliminary observations also suggest that 17-(dimethylaminoethylamino)-17-demethoxygeldanamycin, a heat shock protein 90 inhibitor that exhibits antimyeloma effects (43), clearly reduces basal and IGF-induced HIF-1α expression levels in myeloma cell lines. Although the pathway remains speculative at this stage, HIF-1 may be an important molecular target of heat shock protein 90 inhibitors in multiple myeloma. Bortezomib or heat shock protein 90 inhibitors may enhance the cytotoxicity of conventional chemotherapeutic agents, including melphalan through the inhibition of HIF-1 transcriptional activity.

Finally, we also explored the mechanism by which HIF-1 inhibition sensitizes myeloma cells to melphalan. Although HIF-1 regulates dozens of genes, we chose to focus on the expression of anti-apoptotic proteins that are known to be important for myeloma cell survival. We found that echinomycin or siRNA against HIF-1α specifically blocked both basal and IGF-1–induced survivin levels. Survivin is widely expressed in many types of established myeloma cell lines and contributes to cell survival (28, 44). Romagnoli et al. (44) showed that siRNA-mediated inhibition of survivin expression sensitizes myeloma cells to conventional antimyeloma agents, including melphalan. Furthermore, Peng et al. found that HIF-1 regulates survivin expression in breast cancer cells (34). Although we did not show direct evidence that survivin mediates a HIF-1 survival effect in myeloma cells, survivin may serve as a downstream molecular target of this HIF-1 inhibition strategy. If so, the combination of antimyeloma agents and a survivin inhibitor may be a promising therapeutic strategy for multiple myeloma.

In summary, this is the first study to show the relevance of HIF-1 in multiple myeloma cells. Inhibition of HIF-1 function enhanced the antimyeloma cytotoxicity of melphalan and also abrogated the survival effects of IGF-1 against melphalan-induced apoptosis in myeloma cells. Our results suggest that the inhibition of HIF-1 function can be used as an attractive therapeutic approach to overcome the unfavorable biology of myeloma cells and improve patient outcomes in multiple myeloma.

No potential conflicts of interest were disclosed.

We thank Mirai Iwaya, Mariko Ozawa, Hidemi Yamazoe, and Naoko Enya for their technical assistance.