Purpose: The incidence of bone metastasis in advanced breast cancer (BrCa) exceeds 70%. Bortezomib, a proteasome inhibitor used for the treatment of multiple myeloma, also promotes bone formation. We tested the hypothesis that proteasome inhibitors can ameliorate BrCa osteolytic disease.
Experimental Design: To address the potentially beneficial effect of bortezomib in reducing tumor growth in the skeleton and counteracting bone osteolysis, human MDA-MB-231 BrCa cells were injected into the tibia of mice to model bone tumor growth for in vivo assessment of treatment regimens before and after tumor growth.
Results: Controls exhibited tumor growth, destroying trabecular and cortical bone and invading muscle. Bortezomib treatment initiated following inoculation of tumor cells strikingly reduced tumor growth, restricted tumor cells mainly to the marrow cavity, and almost completely inhibited osteolysis in the bone microenvironment over a 3- to 4-week period as shown by [18F]fluorodeoxyglucose positron emission tomography, micro–computed tomography scanning, radiography, and histology. Thus, proteasome inhibition is effective in killing tumor cells within the bone. Pretreatment with bortezomib for 3 weeks before inoculation of tumor cells was also effective in reducing osteolysis. Our in vitro and in vivo studies indicate that mechanisms by which bortezomib inhibits tumor growth and reduces osteolysis result from inhibited cell proliferation, necrosis, and decreased expression of factors that promote BrCa tumor progression in bone.
Conclusion: These findings provide a basis for a novel strategy to treat patients with BrCa osteolytic lesions, and represent an approach for protecting the entire skeleton from metastatic bone disease. Clin Cancer Res; 16(20); 4978–89. ©2010 AACR.
Breast cancer (BrCa) metastasis to the bone and osteolytic lesions typically caused by these metastases are particularly resistant to pharmacologic therapy. This study shows that the Food and Drug Administration–approved proteasome inhibitor bortezomib strikingly decreases the size of BrCa metastatic tumors and associated osteolytic lesions through multiple mechanisms: by inducing cellular necrosis and apoptosis, inhibiting tumor cell Wnt signaling and matrix degradation, and reducing vascularization. In addition, the proteosome inhibitor provided a significant skeleton-wide bone anabolic effect, despite the presence of a metastatic lesion. Current treatment of metastases with bisphosphonates limits bone resorption but does not rebuild bone volume lost to osteolysis. Our findings also provide evidence for increased capability to treat BrCa osteolytic disease preemptively using bortezomib before tumor cell growth in bone by inhibiting tumor responses to the bone microenvironment and by providing protective anabolic effects on the skeleton.
Metastatic osteolytic disease is prevalent in cancer patients. In advanced breast cancer (BrCa), 70% of women develop osteolytic lesions, resulting in pain, pathologic fracture, and increased morbidity. Dysfunction of the ubiquitin-proteasome system is associated with tumor growth and metastatic disease, providing the rationale for development of proteasome inhibitors as antineoplastic therapies (1, 2). The proteasome is a ubiquitous enzyme complex that plays a critical role in the degradation of proteins involved in cell cycle regulation, apoptosis, and angiogenesis (2, 3). Bortezomib, a selective proteasome inhibitor used to treat multiple myeloma, has a potent anabolic effect on bone (4–9). Bortezomib alters the bone marrow microenvironment by increasing the number and differentiation of resident mesenchymal stem cells into osteoblasts, thereby increasing bone formation rates within 4 weeks in normal mice and resulting in trabecular bone formation in bone loss model (7). A similar enhancement of osteoblast differentiation is found in myeloma patients treated with bortezomib who show sustained increases in circulating osteocalcin, a marker of bone formation (6, 10). Thus, bortezomib treatment represents a novel and clinically feasible approach for increasing bone formation in the setting of the osteolytic bone disease accompanying metastatic breast, prostate, and lung cancers (4, 5, 11).
Nonsurgical treatment of bone metastatic lesions includes radiation therapy and bisphosphonates. Bisphosphonates were initially reported to reduce the risk of pathologic fracture and bone pain, although a recent study totaling over 7,000 BrCa patients indicated no reduction in fracture risk compared with placebo or no treatment groups (12–14). In addition, bisphosphonates at doses required for cancer patients can have a significant side effect profile (15, 16), and alternate approaches to protect the skeleton are needed.
Tumor cells in the bone microenvironment overcome the marrow compartment and inhibit the ability of sufficient stromal cells to differentiate into osteoblasts to replace lost bone. Cancer cells secrete factors that induce a vicious cycle of osteoclast activation and growth factor release that promotes tumor survival. Bortezomib contributes to the apoptosis of tumor cells and tumor-activated osteoclasts (17–19); as well, low-dose bortezomib promotes bone anabolic effects in mice (7). Given the aggressive osteolytic disease produced by BrCa cells that metastasize to the bone, we investigated the effectiveness of bortezomib treatment for inhibiting rapid tumor growth within the bone, and the potential for retaining bone volume. We postulated that bortezomib treatment could be an effective therapy by suppressing growth of the metastatic tumor (4, 7–9, 20–22), inhibiting osteoclastogenesis and survival (9, 17, 21, 23–26), and stabilizing osteoprogenitor cells within the marrow (26), or by a combination of all these effects.
Materials and Methods
The metastatic human BrCa MDA-MB-231 (American Type Culture Collection HTB-22) and mouse preosteoblast MC3T3-E1 (American Type Culture Collection CRL-2593) cell lines were used. Cells were cultured in α-MEM containing 10% fetal bovine serum (Invitrogen, Inc.). Cells were maintained at 37°C in a humidified incubator with 5% CO2.
Approval from the Institutional Animal Care and Use Committee was obtained. Six-week-old female severe combined immunodeficient (NCr/SCID) mice were housed in pathogen-free conditions and used for all experiments. Per experiment, groups contained three mice, and all studies except for the bortezomib pretreatment and continuous bortezomib experiment comparison (Fig. 4) were repeated for a total of six to eight mice per group. The repeat studies were terminated at either 6 or 7 weeks. At sacrifice, all mice were analyzed radiographically at weekly intervals followed by either histology, micro–computed tomography (μCT), or positron emission tomography (PET)/single-photon emission CT (SPECT) imaging at sacrifice.
Tibial implantation of MDA-MB-231 cells and bortezomib treatment
Mice were anesthetized with 0.15 mg ketamine/0.015 mg xylazine i.p. per gram of body weight. A medial parapatellar incision was created, and a needle was placed in the intramedullary canal of the tibia by aid of fluoroscopy (XiScan 1000-1, XiTec). MDA-MB-231 cells (1 × 105 in 100 μL of PBS) were slowly injected into the tibia and closed with 5-0 chromic suture (Ethicon, Inc.). Mice were given 0.1 mg/kg buprenorphine subcutaneous (SQ) postoperatively. Bortezomib (Millennium Pharmaceuticals) i.p. injections at a dose of 0.3 mg/kg body weight were begun 24 hours after intratibial injection and continued three times a week, or as indicated in each study.
Following sacrifice, the lower extremities were dissected and then fixed in 4% paraformaldehyde and decalcified in 18% EDTA (pH 8.0). Paraffin sections were cut at 6-μm thickness and stained with routine H&E, TRAP, or Ki-67 immunohistochemistry (27). Photomicrographs were acquired on a Zeiss Axioskop 40 microscope with attached AxioCam HRC and analyzed by AxioVision Rel 4.7 software (Carl Zeiss MicroImaging).
Osteolysis was monitored by serial radiographs using a Faxitron MX-20 X-ray machine unit. We devised a scale to measure the severity of osteolytic lesions based on their appearance by conventional radiography. Blinded radiographs were evaluated by seven different scorers, and significant differences were determined by Student's t test. Osteolytic lesions were scored on a scale of 0 to 5 based on their severity: 0, no visible osteolysis; 5, most severe degree of osteolysis. A grade 1 lesion was small, isolated, and <10% of the width of the bone; grade 2 included single or multifocal small lesions, <33% of the cortical width; grade 3 lesions have a size of 33% to 75% of the cortical width; extensive lesions with >75% of the cortical width are graded as 4; and pathologic fracture with extensive cortical destruction represents a grade 5 lesion.
Specimens were scanned using a high-resolution desktop microtomographic imaging system (μCT40, Scanco Medical AG) using an isotropic voxel size of 12 μm, as previously described (28–30). In tumored bone, total bone volume is assessed by μCT by selecting the region of each tibia from the knee joint to the most proximal aspect of the proximal tibial-fibular joint, which ranges from 575 to 650 sections. In the nontumored distal femur, bone was evaluated in the region starting 360 μm proximal to the growth plate and extending 1,800 μm proximally. For bone regions without tumors, we assessed the trabecular bone volume fraction (%), trabecular thickness (μm), trabecular number (mm−1), trabecular separation (μm), connectivity density (1/mm3), and structure model index (SMI). Results were analyzed for significant differences by Student's t test.
PET/CT animal imaging
Bioscan NanoSPECT/CT and Philips Mosaic HP PET cameras were used to collect CT and [18F]fluorodeoxyglucose (FDG) images of mice. Tumored mice were anesthetized and injected i.v. with 100 μCi of 18F-FDG, and PET imaging was done at 30 minutes after administration. All data imaging and analysis were done at the University of Massachusetts Small Animal Imaging Core Facility. After each PET acquisition, the mouse, immobilized on the Minerva bed (Bioscan), was transferred to NanoSPECT/CT camera for the CT acquisition. The CT acquisition was done at standard frame resolution, 45 kVp tube voltage, and 500 ms of exposure time. The CT reconstruction was done using InVivoScope 1.37 software (Bioscan), and a PET/CT fusion image was created.
Real-time reverse transcription-PCR analysis
mRNA levels of metastatic and osteolytic-related genes were analyzed from MDA-MB-231 cells, MC3T3 cells, or BrCa tumor following bortezomib treatment. RNA was isolated using Trizol (Invitrogen) according to the manufacturer's protocol. Oligo(dT) primers were used in conjunction with the SuperScript First-Strand Synthesis System (Invitrogen) to synthesize cDNAs. Primer sequences are as follows: mouse: alkaline phosphatase (AP), TTGTGCGAGAGAAAGAGAGAGA (forward) and GTTTCAGGGCATTTTTCAAGGT (reverse); collagen type 1A (COL1A), CCCAAGGAAAAGAAGCACGTC (forward) and AGGTCAGCTGGATAGCGACATC (reverse); DKK3, TGCACCAGGAAGTTCACAAG (forward) and GGCCCACAGTCTTCATCAAT (reverse); glyceraldehyde-3-phosphate dehydrogenase (GAPDH), AGGTCGGTGTGAACGGATTTG (forward) and TGTAGACCATGTAGTTGAGGTCA (reverse); lymphoid enhanced binding factor 1 (LEF1), AGTGCAGCTATCAACCAGAT (forward) and TTCATAGTATTTGGCCTGCT (reverse); RUNX2, CGGCCCTCCCTGAACTCT (forward) and TGCCTGCCTGGGATCTGTA (reverse); transcription factor 4 (TCF4), CAGAATCCACAGATACAGCA (forward) and CAGCCTTTGAAATCTTCATC (reverse); human: Dickkopf 1 (DKK1), TCATTTCCGAGGAGAAATTGAG (forward) and AACCTTCTTGTCCTTTGGTG (reverse); GAPDH, ATGTTCGTCATGGGTGTGAA (forward) and TGTGGTCATGAGTCCTTCCA (reverse); LEF1, CCCGAGAACATCAAATAAAGTG (forward) and CTCTACAACAAGGGACCCTC (reverse); matrix metalloproteinase-9 (MMP-9), CCTGGAGACCTGAGAACCAATC (forward) and CCACCCGAGTGTAACCATAGC (reverse); receptor activator for NF-κB ligand (RANKL), CACATCAGAGCAGAGAAAGC (forward) and CTTTATGGGAACCAGATGGG (reverse); RUNX2, CGGCCCTCCCTGAACTCT (forward) and TGCCTGCCTGGGGTCTGTA (reverse); and vascular endothelial growth factor (VEGF), TCTTCAAGCCATCCTGTGTG (forward) and GCGAGTCTGTGTTTTTGCAG (reverse). Real-time PCR analysis was done to confirm expression levels by using an ABI machine and PRISM software (Applied Biosystems); significant differences were determined by Student's t test.
Annexin V binding and fluorescence-activated cell sorting analysis
Bortezomib-treated MDA-MB-231 cells were stained with Annexin V–FITC and propidium iodide (PI) using the Annexin V–FITC Apoptosis Detection kit (BD Biosciences). Stained live cells were submitted to the Flow Cytometry Core Lab at the University of Massachusetts for fluorescence-activated cell sorting (FACS) analysis (FACSCalibur, BD Biosciences). All fluorescent parameters were acquired in logarithmic amplification, and forward and side scatter parameters were acquired in linear. Data were analyzed using CellQuest software.
Bortezomib slows the growth of osteolytic lesions caused by BrCa tumors in the bone microenvironment
Given the anabolic effects of bortezomib on bone in multiple myeloma, we investigated if bortezomib treatment could ameliorate the osteolysis caused by BrCa cells in bone. Mice that received intratibial inoculations of MDA-MB-231 BrCa cells were treated with 0.3 mg/kg body weight bortezomib administered three times a week for 7 weeks. This dose and schedule represents the highest bortezomib dose used in a study in which the serum osteocalcin level, a marker of bone formation, was found to increase in response to bortezomib treatment (7). We observed this dose to be nontoxic throughout the duration of the study (n > 24 mice).
Radiographs in Fig. 1A (three representative tibias) show osteolytic lesions in controls at day 21 that are inhibited by bortezomib treatment. By 35 days, small focal areas of missing trabeculae were observed with bortezomib treatment. With continued bortezomib (to day 49 in Fig. 1A), there was evidence of a slow rate of progression of bone lysis in the majority (∼70%) of the bortezomib-treated mice in multiple independent experiments. At termination of studies, control mice had very large tumors causing significant morbidity, whereas the bortezomib groups consistently had greatly reduced bone loss in the region of the tumor cell inoculation when compared with control-treated mice. These conclusions are in part based on a semiquantitative evaluation of the radiographs based on a scoring system (see Materials and Methods; Fig. 1B). An aggressive increase in osteolysis was observed in controls after day 28 when tumor destroyed the cortical bone, invaded the surrounding tissue, and contributed to osteolysis from the periosteal side of bone as well as from the medullary cavity. In the bortezomib group, a clear inhibition of osteolytic disease is delayed until 4 to 5 weeks. After this point, low levels of osteolytic activity occurred, indicated by the appearance of lesions at the end of the study (Fig. 1A). A similar late appearance of osteolytic lesions was observed in different experiments (data not shown).
The tumor-containing tibias were further analyzed by ex vivo μCT in three mice. Figure 1C shows tibias of control mice with extensive osteolytic disease eroding through the cortex, whereas the tibias of the bortezomib-treated mice had minimal evidence of cortical erosion and mild osteolysis. Quantitative analysis by μCT shows that the tumor-bearing tibias of the bortezomib-treated mice have >2-fold higher bone volume than the tibias of control mice (Fig. 1D). Taken together, these findings show that the bortezomib-treated mice had a striking inhibition of osteolytic disease.
Bortezomib reduces the size of BrCa tumors
The volume of intratibial BrCa tumors was measured in vivo in bortezomib- and control-treated mice by injecting the mice with [18F]FDG and visualizing the tumor using PET imaging. Figure 2A shows representative PET images that identify a larger tumor volume and metabolic activity in controls compared with bortezomib-treated mice. A 2-fold increase in tumor volume in the control was confirmed by histologic examination of tibias with tumors from bortezomib and control mice (Fig. 2B and C, left). In controls, there is aggressive lytic disease destroying trabecular and cortical bone with tumor growth invading muscle. In striking contrast, tumor growth in the bortezomib-treated mice was initially restricted to the medullary cavity until a cortical break occurred, allowing slow tumor growth in muscle (Fig. 2B). The bottom panels of Fig. 2B illustrate this point. Viable tumor cells are found in muscle (panel 1) of bortezomib-treated mice, whereas areas of necrotic cells were evident within the medullary cavity (panel 2). Solid tumor tissue outside the bone exhibited similar cell morphology between bortezomib and controls (panels 3 and 4). The fraction of actively growing cells within the tumor was examined by Ki-67 detection and found to be far less in bortezomib-treated tumor than in control tumor (Fig. 2C, right; P < 0.05). The modest decrease in tumor growth fraction (% Ki-67+ cells) compared with tumor size can be accounted for by stimulated tumor growth in the bortezomib-treated mice after invasion into muscle.
The effect of tumor growth on bone osteolysis in controls is visualized by minimal detection of osteoclasts by TRAP staining, as there is little bone remaining to be resorbed in controls (Fig. 2D, panel 1). Only small remnants of bone remained with TRAP-positive cells (panels 2 and 3 at ×10 and ×40 magnifications of osteoclasts, respectively). The bortezomib group exhibited osteoclast activity on cortical surfaces, where tumor growth was expanding along the periosteal side as well as on the endosteal surface. These findings suggest that bortezomib treatment reduces osteolytic disease by killing off tumor cells and that remaining tumor cells can still locally secrete osteoclast-activating factors and can survive as a solid tumor.
Bortezomib promotes increased bone formation in the distal femur of the nontumor-bearing limb
BrCa cells secrete many factors that cause local bone resorption and also circulate systemically to affect the entire skeleton. To assess the influence of bortezomib on bone away from the area of the BrCa tumor, μCT analysis was done and various parameters of bone growth were measured in the distal femur of bortezomib-treated and control mice. Figure 3 shows that there is increased trabecular bone formation in femurs of mice treated with bortezomib compared with control as evidenced by significant improvement in the parameters of bone growth, including bone volume fraction, trabecular thickness, connectivity density, and SMI (reflecting a more plate-like architecture in the bortezomib-treated bone). The trabecular number and spacing showed greater bone formation in bortezomib-treated mice, but did not reach significance. These results are entirely consistent with the previously described bone anabolic property of bortezomib (7) and occur in the presence of metastatic tumor growth.
Treatment with bortezomib before tumor cell inoculation protects bone from osteolysis
To assess clinical relevance, a study was designed to establish if bortezomib given before tumor metastasis decreased tumor growth and osteolytic disease and would protect the bone from subsequent osteolysis. Therefore, we examined bortezomib treatment of mice for a period of time before the intratibial inoculation of MDA-MB-231 cells. The rationale was 2-fold. First, the anabolic effect that bortezomib treatment has on bone might render the bone less susceptible to subsequent osteolysis. Second, this has translational relevance, as bortezomib treatment might be used prophylactically to decrease the likelihood of BrCa patients developing osteolytic lesions. The study consisted of three groups: mice that were not treated before or following tumor cell inoculation, mice that were treated with bortezomib (0.3 mg/kg body weight i.p. three times a week) both before and after tumor inoculation, and mice that were treated with bortezomib only before but not after injection of tumor cells (bortezomib pretreatment; Fig. 4A). All three mice in each group exhibited consistent results. Radiographic analysis of the tibias of mice pretreated with bortezomib 3 weeks before time of inoculation with BrCa cells and mice treated continuously with bortezomib both exhibited smaller osteolytic lesions than untreated mice (Fig. 4B). The radiographic grading of the tibias showed a statistically significant lower osteolytic lesion grade in the bortezomib continuous and bortezomib pretreatment groups compared with control mice with tumors (P < 0.0001) and a more effective inhibition of osteolysis in the continuously treated bortezomib group (P < 0.05; Fig. 4C).
Patients with metastatic BrCa have a potential for continuously developing new bony lesions. Therefore, we addressed if the anabolic effect of bortezomib treatment on bones that did not contain tumor (shown in Fig. 3) would provide protection to the skeleton in a patient with advanced BrCa, and if this effect would remain over time despite the presence of tumor at a remote site. To determine if treatment with bortezomib may protect the skeleton from subsequent osteolysis, μCT analysis of the femurs of the three groups of mice described above (Fig. 4A) was done examining both femurs of each group. Several significant changes are observed in trabecular bone structure between the control and the two bortezomib groups and between the left femur and the right femur (Fig. 5). Bortezomib treatment supports increased trabecular bone formation in both tumor-bearing (right) and contralateral (left) limbs reflected in all parameters of trabecular bone, in both bortezomib continuous and pretreatment groups, compared with control (Fig. 5, continuous versus control; P < 0.05 to P < 0.005). Both pretreatment and continuous bortezomib groups exhibited a significant improvement in bone formation (∼60% increase in bone volume) compared with control. However, it is noteworthy that the femurs from the tumor-affected right leg of both bortezomib and control mice consistently have values that indicate less bone when compared with the left femurs (Fig. 5, right versus left). This suggests that mice develop disuse osteopenia in the right femur as a result of decreased weight bearing on the tumored leg. The bone anabolic effects of bortezomib were not attenuated during the subsequent tumor growth period. Mice receiving a 21-day course of bortezomib before treatment (before inoculation of tumor cells) did not have significantly different bone parameters compared with the mice receiving continuous treatment (Fig. 5, continuous versus pretreatment; P > 0.05). This effect is significant in that a pulsed dose of bortezomib before inoculation of cancer cells was sufficient to confer an anabolic effect on bone and provide protection from osteolysis. We conclude, from the increased bone volume following 3 weeks of pretreatment in the nontumored femur and the reduced tibial osteolysis after inoculation during the following 25 days of tumor growth, that bortezomib could contribute to protecting the skeleton in advanced BrCa.
Bortezomib decreases survival of MDA-MB-231 BrCa cells and affects gene expression
To gain insight into the mechanisms of bortezomib on the cellular activities of BrCa cells, cell growth and mRNA levels of metastatic and osteolytic genes were examined. MDA-MB-231 proliferating cells were treated in vitro with varying doses of bortezomib, ranging from 0 to 50 nmol/L, to determine the effects on cell proliferation and survival. Initial cell count studies done after 24 hours showed that adherent cell counts decreased beginning at 10 nmol/L bortezomib (21% of control), with a 62% and 74% cell loss at 20 and 50 nmol/L bortezomib, respectively (data not shown). The loss of cell viability by bortezomib was the result mainly of necrosis of the cells as determined by an Annexin V binding assay done on proliferating cells treated with varying doses of bortezomib (Fig. 6A). FACS analysis of the cells (represented in Fig. 6B) showed that the percentage of viable cells (neither stained with PI nor bound to Annexin V; Fig. 6A, solid line, diamond; Fig. 6B, bottom-left quadrant) decreased steadily to 60% at a dose of 20 nmol/L bortezomib and then remained constant to a dose of 50 nmol/L bortezomib. The percentage of cells undergoing apoptosis (bound to Annexin V, but not stained with PI; Fig. 6A, dashed line, open square; Fig. 6B, bottom-right quadrant) did not exceed 10%. The amount of cells undergoing necrosis (bound to Annexin V and stained with PI; Fig. 6A, solid line, circle; Fig. 6B, upper-right quadrant) was greater than that of apoptotic cells and remained relatively constant from bortezomib doses of 10 to 50 nmol/L. These results indicate that both necrosis and apoptosis contribute to death of MDA-MB-231 cells by bortezomib treatment, but also suggest that the MDA-MB-231 cell line may have a small population of bortezomib-resistant cells.
We determined the contribution of bortezomib in preventing osteolytic disease by analyzing dose-dependent expression of marker genes associated with bone resorption and formation in MDA-MB-231 cells treated 24 hours with bortezomib (Fig. 6C). Expression of genes promoting tumor growth (RUNX2, MMP-9, and VEGF) decreased at increased bortezomib concentrations, whereas GAPDH, an internal marker of cellular RNA levels, remained constant. In response to bortezomib, we found a steady increase in expression of the Wnt antagonist DKK1 (31), with a concomitant decrease in the canonical Wnt transcription factor LEF1. This finding indicated that bortezomib is inhibiting tumor cell growth mediated by the Wnt pathway, as well as reducing the levels of genes in MDA-MB-231 cells related to osteolytic activity in bone (32).
To evaluate the mechanism for the anabolic effect of bortezomib on bone formation, we examined a preosteoblast cell line (MC3T3) analogous to osteoprogenitor cells within the marrow for its responsiveness to the same dose range that affects BrCa cells (Fig. 6D). The high doses of 20 and 50 nmol/L were toxic to both cell types, osteoblasts and BrCa cells, resulting in loss of expression of bone formation markers (COL1 and AP in the osteoblasts). A striking difference between BrCa and MC3T3 cells was the opposite effects on the Wnt pathway, which was inhibited in the tumor cells but was stimulated in osteoblasts at lower doses for bone formation. This finding is consistent with the anabolic effect of bortezomib in multiple myeloma (33). We next determined if the same effects occurred in response to the in vivo dose of bortezomib. Tumor tissue excised from the tibia showed a striking inhibition of MMP-9 with a modest increase in DKK, consistent with the in vitro effect of bortezomib in BrCa cells. Additionally, in the tumor tissue, RANKL, promoting bone resorption, was stimulated. This finding reflects the histologic analyses of the bortezomib-treated group (see Fig. 2D) exhibiting osteoclastic activity.
We conclude from these studies that metastasis-associated osteolytic disease is reduced in vivo by bortezomib through multiple mechanisms in the bone microenvironment: by inducing cellular necrosis, decreasing the tumor response to Wnt signaling, reducing expression of a key factor in tumor vascularization, and decreasing the RUNX2 and MMP-9 osteolytic cascade. Together, these modifications in tumor cell activity by bortezomib contribute to inhibiting tumor growth to the bone microenvironment.
Here, we established that the proteasome inhibitor bortezomib effectively suppresses BrCa tumor growth within bone and stimulates new bone formation in the presence of metastatic disease. Antitumor growth effects by bortezomib occur in the bone microenvironment, where the vicious cycle of tumor growth and osteolytic disease is activated in response to BrCa cells. We show that the anti-osteolytic effects of bortezomib are primarily due to significantly decreased tumor size as evidenced by histology, radiographic monitoring, and μCT quantitation of bone volume. Lastly, we have defined mechanisms contributing to the inhibition of tumor growth in the bone microenvironment and osteolytic disease by bortezomib that include (a) the sensitivity of highly metastatic BrCa cells to necrosis, (b) reduction in expression of metastatic and tumor growth–related genes, and (c) promotion of bone formation throughout the skeleton. These effects of bortezomib provide a beneficial antitumor and bone anabolic effect.
In our continuously treated study, we find a striking inhibition of osteolytic disease that continued through 4 weeks after inoculation. However, at 5 weeks, evidence of the onset of osteolysis was identified in the bortezomib group, which slowly progressed until the study was terminated when tumor size became unbearable in the control group (7 or 8 weeks in repeat study). The Ki-67 assays suggest that surviving tumor cells exposed to bortezomib result in a delayed onset of osteolysis. As our bortezomib dose did not cause toxic effects in mice, perhaps a higher dose would be more effective in killing all BrCa cells, as indicated from our in vitro studies. Two mechanisms are contributing to reduced osteolytic disease: the killing of tumor cells by bortezomib and bortezomib inhibition of osteoclast activity through induction of apoptosis, which has been identified in several studies (17–19, 23–26). BrCa cells are known to produce many different factors that induce osteolytic disease. Notably, at sacrifice after 7 weeks, we observe TRAP-positive osteoclasts at the tumor-bone interface in bortezomib-treated mice and also found RANKL to be expressed in the tumor tissue. Our in vitro studies analyzing bortezomib dose effects suggest that a small population of MDA-MB-231 cells survive at high doses.
Early studies found mixed results in ongoing clinical trials evaluating the effect of bortezomib on osteolytic lesions caused by solid tumors (34–37). However, in one other study using the intratibial model of prostate cancer, it was suggested that osteoclast activity was diminished (38), similar to what has been shown in multiple myeloma patients (39). In our studies, the net effect of systemic bortezomib is that proteasome inhibition is effective in preventing BrCa tumor growth in bone and greatly diminishes the osteolytic disease. Thus, just as bortezomib is used for multiple myeloma cells that have increased proteasome activity (40) inducing osteolytic bone disease [as does BrCa (18)], we find that bortezomib has the similar anti-osteolytic effects in BrCa tumors.
We identified mechanisms that are directly attributed to bortezomib, modifying the properties of BrCa cells in vivo to facilitate an anabolic effect in the bone microenvironment in the presence of a tumor. Analysis of the tumor tissue revealed a reduction in the metastasis-related genes RUNX2, VEGF, and MMP-9 at high doses of bortezomib. This profile reflects altered cellular properties, reduced tumor vascularization, migration, matrix destruction, and osteolytic disease (27, 30, 41–44). Another mechanism of bortezomib inhibiting tumor growth is by elevation of DKK1, an inhibitor of the Wnt signaling pathway, and a concomitant decrease in LEF1, the transcriptional mediator of Wnt signaling, thereby decreasing tumor growth potential. In multiple myeloma patients, serum levels of DKK are very low, contributing to activated Wnt signaling (21, 22). BrCa tumorigenesis and bone metastasis are linked to the Wnt signaling pathway, having downregulated DKK and dysregulation of β-catenin linked to progression and prognosis (31, 45–47). These beneficial effects of bortezomib are also observed in the tumor tissue by the changes in expressed genes (reduced LEF1 and MMP-9). Notably, osteoblasts have the opposite response at low bortezomib doses, exhibiting increased LEF1/TCF4, AP, and COL1 and reflecting bone formation activity. In multiple myeloma patients, bortezomib stimulates osteoblast activity (22, 33), consistent with the significant anabolic effect of bortezomib in the nontumor-bearing limbs of mice with bone tumors.
Our studies are significant, as the anabolic effects of bortezomib on BrCa-free bone in the setting of metastatic bone disease were shown by a 1.5-fold increase in bone volume. Furthermore, we showed that the effects of a 3-week pretreatment allowed accrual of bone throughout the skeleton from the analyses of the nontumor-bearing femurs. This suggests a preemptive, protective effect to the bone, thereby diminishing the effect of tumor-mediated osteolysis. Thus, we find an additional benefit to the anabolic bone effect, as pretreatment with bortezomib has a repressive effect on subsequent tumor growth. The clinical implications of these results are considerable. Proteasome inhibitors could provide a protective effect on the skeleton if bortezomib were to be combined with other treatments for BrCa before metastasis.
In summary, proteasome inhibitors may treat solid tumors and the osteolytic bone disease that accompanies metastasis by myriad effects that include decreasing tumor cell proliferation and survival, inhibiting bone-destructive pathways and enzymes and vascularity of the tumors, reducing osteoclast number and function (17, 24, 25, 39), and increasing osteoblast differentiation and bone formation (7, 9, 22, 33). Our studies show inhibitory effects of bortezomib on solid tumor (BrCa) growth in bone and prevention of the early-onset osteolytic disease. We conclude that proteasome inhibitors have multifactorial beneficial effects in prevention of BrCa growth in bone and induced osteolysis.
Disclosure of Potential Conflicts of Interest
D.C. Ayers: commercial research grant, Zimmer, Inc.; other commercial research support, Musculoskeletal Transplant Foundation (MTF).
Grant Support: Studies reported were in part supported by NIH grants P01CA082834, S10RR023540, and F32AR055030. Core resources supported by the Diabetes Endocrinology Research Center grant DK32520 were also used. J.B. Lian is a member of the University of Massachusetts Diabetes and Endocrinology Resource Center (DK32520). The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received December 15, 2009.
- Revision received August 19, 2010.
- Accepted August 25, 2010.