This study was partly funded by Exelixis Inc., the employer of Frauke Schimmoller and Dana T. Aftab. Eva Corey has received research funding from Exelixis, Inc., via Sponsored Research Agreement with the University of Washington. There are no patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors.
Conceived and designed the experiments: EC RLV FS DTA. Performed the experiments: HMN LGB XZ TSG. Analyzed the data: EC LGB CM RG TSG DTA FS NR. Contributed reagents/materials/analysis tools: DTA FS RG EC RLV CM TSG. Wrote the manuscript: NR EC RLV CM RG DTA FS.
Cabozantinib is an inhibitor of multiple receptor tyrosine kinases, including MET and VEGFR2. In a phase II clinical trial in advanced prostate cancer (PCa), cabozantinib treatment improved bone scans in 68% of evaluable patients. Our studies aimed to determine the expression of cabozantinib targets during PCa progression and to evaluate its efficacy in hormone-sensitive and castration-resistant PCa in preclinical models while delineating its effects on tumor and bone. Using immunohistochemistry and tissue microarrays containing normal prostate, primary PCa, and soft tissue and bone metastases, our data show that levels of MET, P-MET, and VEGFR2 are increasing during PCa progression. Our data also show that the expression of cabozantinib targets are particularly pronounced in bone metastases. To evaluate cabozantinib efficacy on PCa growth in the bone environment and in soft tissues we used androgen-sensitive LuCaP 23.1 and castration-resistant C4-2B PCa tumors.
Metastases remain the major cause of morbidity and mortality in men suffering from advanced prostate cancer (PCa). Despite existing agents that are effective against advanced PCa, survival after development of castration resistance remains very short. Therefore, novel, effective treatments against metastatic and castration-resistant disease are urgently needed.
Cabozantinib is a potent inhibitor of receptor tyrosine kinases, including MET and VEGF receptor 2 (VEGFR2). Other targets inhibited by cabozantinib include AXL, FLT-3, KIT, and RET [
Cabozantinib was recently approved by the FDA for the clinical treatment of progressive, metastatic medullary thyroid cancer. This approval followed early observations of cabozantinib activity against this disease in the initial phase I clinical study [
MET and its ligand, hepatocyte growth factor (HGF), have been implicated in the progression of many cancers. MET signaling promotes cell survival, proliferation, invasion, metastasis, and angiogenesis
VEGFR2 also plays important roles in multiple cancers, including PCa. VEGFR signaling is central in the regulation of angiogenesis and is increased in metastatic CRPC lesions [
We aimed to determine the expression levels of cabozantinib’s primary targets in advanced PCa and assess its effects on subcutaneous PCa tumors and those growing in bone, specifically with regards to tumor burden and bone turnover. Our results show that the targets of cabozantinib are expressed in PCa metastases and that cabozantinib inhibits tumor growth in bone and soft tissue and exhibits beneficial effects on bone in both intact and castrated male mice.
All human tissues were obtained from patients who signed written informed consent and upon University of Washington IRB approval. All animal studies described in this manuscript were approved by and performed in compliance with the University of Washington Institutional Animal Care and Use Committee and NIH guidelines.
Tissue microarrays (TMAs) were used to determine MET, P-MET, and VEGFR2 immunoreactivity in normal prostate (NP) and PCa: 1) UWTMA48: NP and benign prostatic hyperplasia (60 tissues, two cores per tissue) and primary PCa (61 tissues, two cores per tissue); 2) UWTMA52: matched NP and PCa from recurrent and non-recurrent patients (63 recurrent and 64 non-recurrent patients, two cores for each NP and PCa); 3) UWTMA21: PCa metastases from 44 patients (40 bone metastases; 19 liver metastases; 27 LN metastases; and 7 other soft tissue metastases, two cores per tissue); and 4) UWTMA48: 24 LuCaP PCa xenograft models from intact, castrated and docetaxel-treated animals (three cores per tissue). IHC was performed using standard procedures with antigen retrieval [
For each core in each TMA, a staining index was constructed as a weighted combination of the 3-point staining intensities, with weights given by the percentage of tissue staining at each intensity; see
C4-2B cells (Urocor, Inc., Oklahoma City, OK) and MC3T3 cells (ATCC, Manassas, VA) were maintained under standard tissue culture conditions. C4-2B cells were grown in RPMI 1640 and 10 % fetal bovine serum (FBS), and MC3T3 cells were grown in DMEM with 10% FBS. The LuCaP 23.1 PCa xenograft was maintained and serially passaged in CB17 SCID mice [
Intact six-week-old male beige SCID mice (Charles River, Wilmington, MA) were injected with a LuCaP 23.1 single-cell suspension into the right proximal tibiae as published previously [
The study design was the same as for study 1, except that animals were castrated and C4-2B cells were injected into tibiae two weeks post-castration [
C4-2B (2x106, 1:1 with Matrigel) cells were injected subcutaneously into castrated male mice. Animals were randomized into a control group (n=8) or a cabozantinib group (n=12) when the tumor volume reached 100 mm3. 60 mg/kg cabozantinib was administered by oral gavage five times a week for up to nine weeks. Animals were sacrificed when tumors reached 1000 mm3 or when animals were compromised.
The study design was the same as in study 1, except that a lower dose of cabozantinib (30 mg/kg) was used and the treatment lasted up to 15 weeks. Animals were randomized into a control group (n=10) or a cabozantinib group (n=10).
A Scanco vivaCT 40 high-resolution µCT scanner was used to analyze a 0.85 mm section spanning the proximal tibia metaphysis of tumored and non-tumored contralateral tibiae (LuCaP 23.1: n=3–5; C4-2B: n=5–10 per group). Bone volume (BV), tissue volume (TV), trabecular separation (Tb.Sp), trabecular thickness (Tb.Th), and trabecular number (Tb.N) were determined and the BV/TV was calculated.
Longitudinal tumor measurements and PSA serum levels were log-transformed and modeled using linear mixed models conditional on the treatment group with random effects for each animal; see
RNA extraction, cDNA synthesis, and qPCR were performed as described previously [
The effects of cabozantinib on proliferation, mineralization, alkaline phosphatase (ALP) activity, and AR transcriptional activity were assessed
To address whether cabozantinib targets are expressed in PCa, we evaluated the levels of MET, P-MET, and VEGFR2 in tissues representing normal prostate and different stages of PCa progression.
MET and P-MET: Our results showed that MET is present at high levels in the NP and primary PCa cells, but there was only marginal evidence of differences between these tissues (mean staining index NP: 0.96, 95% CI 0.93–0.98; PCa: 0.92, 95% CI 0.88–0.96; P=0.06); see
VEGFR2: Normal prostate epithelial cells and PCa cells both exhibited low VEGFR2 immunoreactivity, and there was marginal evidence of higher VEGFR2 in PCa vs NP (mean staining index NP: 0.03, 95% CI 0.00–0.05; PCa: 0.07, 95% CI 0.03–0.11; P=0.02); see
Because of the reported increases of MET in PCa metastases [
IHC and analyses were performed as described in the Methods section. Graphical profiles illustrating distributions of staining intensity were constructed by calculating simple averages across all non-missing sections in each staining category. In each site, the mean staining index is marked by a filled orange circle and orange bars represent 95% CIs. Representative examples of staining are shown for each protein. A. MET is strongly expressed in both primary and metastatic PCa, though it is significantly increased in BM and decreased in soft tissue metastases vs. primary PCa. B. P-MET levels are higher in BM, LN and other soft tissue metastases, while no alteration was detected in liver metastases when compared to primary PCa. C. VEGFR2 expression is significantly increased across PCa metastatic lesions as compared to primary PCa. Images were taken at 400 x magnification.
P-MET was detected in PCa metastases with strong evidence of higher immunoreactivity in BM compared to primary PCa (mean staining index primary PCa: 0.22, 95% CI 0.17–0.26; BM: 0.37, 95% CI 0.30–0.45; P=0.003); see
VEGFR2 was detected in PCa metastases with very strong evidence of higher immunoreactivity in BM compared to primary PCa (mean staining index primary PCa: 0.10, 95% CI 0.07–0.14; BM: 0.25, 95% CI 0.19–0.31; P=0.0001); see
Since the microenvironment influences the gene expression profiles of tumor cells, we also analyzed the MET expression levels by metastatic site. The levels of MET were significantly higher in BM in comparison to liver, LN, and other soft tissue metastases (mean staining index BM: 0.94, 95% CI 0.91–0.98; mean staining index liver: 0.59, 95% CI 0.52–0.66, P<0.0001; LN: 0.67, 95% CI 0.61–0.73, P<0.0001; other soft: 0.68, 95% CI 0.57–0.83, P<0.0001).
The levels of P-MET were weakly or non-significantly different in BM in comparison to liver, LN, and other soft tissue metastases (mean staining index BM: 0.37, 95% CI 0.32–0.44; liver: 0.28, 95% CI 0.13–0.36, P=0.10; LN: 0.46, 95% CI 0.34–0.53, P=0.05; soft other: 51%, 95% CI 0.27–0.69); see
The VEGFR2 levels were significantly different in BM in comparison to LN metastases (mean staining index BM: 0.25, 95% CI 0.21-0.31; LN: 0.40, 95% CI 0.30-0.45, P<0.0001) but not the liver or other soft tissue metastases.
Crosstalk between AR, MET and VEGFR2 signaling has been reported in CRPC (
To determine the expression of selected cabozantinib targets in PCa, we evaluated levels of MET, murine VEGFR2, AXL, KIT, and RET mRNA in 24 different LuCaP PCa xenografts that closely model the heterogeneity of PCa in humans [
To gain a better understanding of cabozantinib’s potential effects in patients with advanced CRPC who are on ADT and/or treated with docetaxel, we examined the levels of MET, P-MET, and VEGFR2 in LuCaP tumors from intact, castrated, and docetaxel-treated animals. Our analyses show moderate evidence that MET and P-MET expression levels are negatively correlated across tumor types (R=0.29; P=0.02), and marginal evidence that the mean MET and P-MET staining indices are 5–6% higher in tumors after docetaxel treatment than in tumors from intact animals (P=0.08 and P=0.05, respectively). Our analyses did not reveal any evidence that expression levels for any other pair of proteins are correlated across or within tumor types (all P>0.14), that mean MET and P-MET staining indices are different between tumors harvested from intact and castrate animals, or that mean MET and P-MET staining indices differ significantly between LuCaP xenografts that display a high and low response to castration; see
To increase our understanding of cabozantinib’s effects on PCa bone metastases, we examined its effects on serum PSA, body weight, and bone turnover in models of PCa growth in the bone. Similarly, we assessed changes in serum PSA, tumor volume, and body weight in response to treatment in animals bearing subcutaneous tumors.
To evaluate the efficacy of cabozantinib on growth of PCa in bone, we treated intact or castrate animals bearing intratibial LuCaP 23.1 or C4-2B tumors, respectively. We selected these two models because LuCaP 23.1 elicits a pronounced osteoblastic reaction and C4-2B elicits a mixed osteoblastic/osteolytic response. Furthermore, LuCaP 23.1 represents androgen-sensitive PCa while C4-2B represents castration-resistant disease. Our qPCR results show that MET, VEGFR2m, KIT, RET and AXL are expressed in LuCaP 23.1. In C4-2B tumors we detected VEGFR2m, AXL and RET, very low levels of MET and no signal for KIT (results are shown in
A. Levels of cabozantinib receptors in LuCaP 23.1 and C4-2B subcutaneous tumors. qPCR was used on RNA isolated from subcutaneous tumors to determine expression levels of MET, VEGFR2m, AXL, RET and KIT. To calibrate the signal we used four-fold dilution of LNCaP cDNA (RET), PC-3 cDNA (MET, AXL, KIT) and LuCaP 23.1 (VEGFR2m). Selection of the calibrator cDNA was based on signal for each specific message. Signal was normalized to housekeeping gene RPL13a. Our results indicate that LuCaP 23.1 tumors express all of the cabozantinib targets and C4-2B tumors express VEGFR2m, AXL and RET, low levels of MET, and no KIT. In these qPCR experiments, we measured relative levels of the target transcripts and not their actual numbers, therefore we cannot compare expression levels of the different targets to each other, and comment whether RET, which gave the higher signal, might be expressed at higher copy number vs MET, and therefore is more important in these models.
B. Linear models of PSA growth show that cabozantinib decreases PSA levels in both androgen-sensitive LuCaP 23.1 and castration-resistant C4-2B models. C. BrdU staining of tibiae shows that cabozantinib decreases proliferation of androgen-sensitive and castration-resistant tumor cells in bone, 2-sided t-test was used to determine the significance of the differences. D & E. AR and PSA immunoreactivity is present in control tumors. Cabozantinib treatment resulted in decreases in AR and PSA immunoreactivity in both models (LuCaP 23.1 (D) and C4-2B (E)). C4-2B cells express lower levels of PSA in comparison to LuCaP 23.1, and the staining is weaker in these tumor. Furthermore, large necrotic areas of tumor are present in the treated tibiae (marked by red asterics). F. 60 mg/kg cabozantinib is well tolerated up to 4 weeks in androgen-sensitive LuCaP 23.1 animals. After this period significant BW decreases vs. control were detected (up to 17%), but because of variation and number of animals, these decreases did not reach significance. 60 mg/kg cabozantinib is well tolerated up to 5 weeks in the castration-resistant C4-2B model, with a 12% significant decrease at week 6. Significance was determined by comparing enrollment BW to BW at each week using 2-sided t-test. Mean ± SEM of the groups is plotted.
We performed a detailed analysis of cabozantinib’s effects on the bone/tumor microenvironment by µCT. We have chosen this type of analysis instead of histomorphometry analysis because µCT evaluates the whole tibia in 3D while histomorphometry analyses are done usually on a single 2D longitudal section of the tibia. The analysis of trabecular bone showed that LuCaP 23.1 growth results in significant increases in bone volume (tumored tibiae: 0.42 ± 0.06 (mean ± SEM); normal tibiae: 0.09 ± 0.01; 5-fold increase in BV/TV, P=0.02). These increases were attenuated by cabozantinib, resulting in a 52% decrease in BV/TV as compared to the control LuCaP 23.1 tibiae. This decrease was reflected in altered trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp); see
TUMORED TIBIAE |
|||||||||
---|---|---|---|---|---|---|---|---|---|
0.42 ± 0.06 | 0.71 ± 0.05 | 0.08 ± 0.02 | |||||||
0.20 ± 0.04 | 0.19 ± 0.08 | 0.12 ±0.05 | |||||||
6.30 ± 0.28 | 6.06 ± 0.83 | 3.28 ± 0.71 | |||||||
3.93 ± 0.39 | 2.05 ± 0.53 | 3.61 ± 1.35 | |||||||
|
|
|
|||||||
0.09 ± 0.01 | 0.17 ± 0.02 | 0.05 ± 0.00 | |||||||
0.09 ± 0.01 | 0.11 ± 0.01 | 0.06 ± 0.01 | |||||||
0.14 ± 0.01 | 0.12 ± 0.03 | 0.32 ± 0.09 | |||||||
0.25 ± 0.03 | 0.33 ± 0.06 | 0.31 ± 0.10 | |||||||
|
|
|
|||||||
NORMAL TIBIAE | |||||||||
0.088 ± 0.014 | 0.055 ± 0.007 | ||||||||
0.164 ± 0.037 | 0.075 ± 0.009 | ||||||||
3.185 ± 0.382 | 2.820 ± 0.316 | ||||||||
4.424 ± 0.194 | 3.221 ± 0.221 | ||||||||
|
|
||||||||
Tb.Th | 0.051± 0.003 | 0.046± 0.004 | |||||||
0.0618 ± 0.007 | 0.047 ± 0.005 | ||||||||
|
|
||||||||
0.333± 0.037 | 0.365 ± 0.043 | ||||||||
0.217 ± 0.015 | 0.319 ± 0.017 | ||||||||
|
|
BV/TV: Bone volume in tissue volume; Tb.N: number of trabeculae; Tb.Th.: trabecular separation; Tb.Sp. trabecular separation
A. LuCaP 23.1 and C4-2B cell growth in tibiae causes large increases in trabecular bone volume. µCT images show that cabozantinib alleviates the bone response to both tumors. In LuCaP 23.1 tumored tibiae cabozantinib caused decreases in BV, while increases in BV were detected in C4-2B tumored tibiae of cabozantinib-treated animals vs control-tumored tibiae. The overall effects are combination of abolishment of tumor effects on the bone as well as cabozantinib effects on normal bone. Details of the effects are provided in Table 1. B. Analysis of non-tumored contralateral tibiae of the experimental animals shows that treatment with cabozantinib results in increased bone volume in both intact and castrated male mice. C.
To evaluate the effects of cabozantinib on non-tumored bone, we analyzed the contralateral non-tumored tibiae of the treated and control animals by µCT. Our analyses show that cabozantinib treatment increased BV/TV in intact and castrated male mice; see
We also monitored the effects of this treatment on body weight to evaluate its tolerability. A 60 mg/kg dose of cabozantinib in animals bearing LuCaP 23.1 tumors was well tolerated for 4 weeks with a maximum body weight decrease of only 5.4%. Larger decreases were detectable at week 5 and 6 (14% and 22%, respectively), but none of these differences reached statistical significance. Similar results were observed in animals with C4-2B tumors, with decreases of 5.5% at 4 and 5 weeks, and 2.7% at week 6, which were also not statistically significant; see
Because of the above described decreases in body weight after prolonged treatment, we evaluated whether a lower dose of cabozantinib maintains the tumor inhibitory effects in the bone while alleviating the loss in body weight. For this study, we used LuCaP 23.1 in intact male mice. Our results demonstrate that this lower dose inhibits tumor progression as demonstrated by decreases in serum PSA ; see
The lower dose was well tolerated for 6 weeks of treatment with no significant body weight loss. Prolonged treatment (7–15 weeks) caused 7–11% body weight decreases which were, however, statistically significant (P=0.0002-0.04). Significance of the changes was determined by comparing enrollment BW to BW at each week using 2-sided t-test. Mean ± SEM is plotted. B. Cabozantinib treatment (60 mg/kg) inhibited tumor growth of subcutaneous castration-resistant C4-2B tumors as determined by TV and serum PSA levels. This treatment also significantly increases survival as determined by log-rank test. C4-2B tumor growth causes decreases of BW in the experimental animals (4–13%, P= 0.0001-0.003), and the cabozantinib treatment prevents this effect resulting in no significant BW loss. Significance of the changes was determined by comparing enrollment BW to BW at each week using 2-sided t-test. Mean ± SEM of the group is plotted.
To evaluate the efficacy of cabozantinib in soft tissue metastases, we analyzed tumor progression in a subcutaneous C4-2B model. Cabozantinib treatment resulted in a significant decrease in tumor volume (TuV) and serum PSA in this model. TuV increased by 30% per week in the control group while decreasing by 14% per week in the cabozantinib group (P<0.0001); PSA increased by 71% per week in the control group while decreasing by 28% per week in the cabozantinib group (P<0.0001); see
In contrast to animals with tumors in bone, the animals bearing subcutaneous tumors treated with 60 mg/kg cabozantinib did not show decreases in body weight; in fact, treated animals had higher body weights in comparison to control animals (3–5 weeks, increases 5.5–22%, P<0.005); see
Given the pleiotropic effect observed on bone in non-tumored and tumored bone in intact and castrated animals, we evaluated the effects of cabozantinib on osteoblasts and tumor cells directly. In concordance with the
Recent years have seen the development of a variety of new drugs for the treatment of advanced metastatic CRPC that provide modest yet significant improvements in overall patient survival and symptom management. Beyond chemotherapeutics, inhibitors of androgen signaling (e.g. abiraterone and MDV3100), as well as new treatments to minimize skeletal related events (e.g. zoledronic acid and denosumab) have been modestly successful in managing patient symptoms in late stage disease. However, sustained suppression of CRPC, particularly in bone, will likely require new drugs and/or drug combinations that target both the tumor and its interplay with the bone microenvironment. In this context, drugs targeting angiogenic factors and tyrosine kinases have been of particular interest, as both have been shown to influence tumor biology and bone turnover [
Cabozantinib inhibits multiple receptor tyrosine kinases (RTKs) that play key roles in tumor and bone biology as well as in angiogenesis. Therefore, this drug has the potential to affect the malignant tumor cells and the tumor microenvironment. In a phase II randomized discontinuation trial, cabozantinib has shown encouraging results. Sixty-eight percent of patients showed improved bone scans, including complete resolution of lesions in 12 percent of these patients, and regression of soft tissue lesions in 72 percent of evaluable patients. Even though 5% of patients showed an objective response, and 75% of patients showed stable disease at 12 weeks, these results are still encouraging given the heterogeneous nature of PCa and the lack of non-palliative treatments for bone metastases. Furthermore, in 31 patients with stable disease that were randomized to placebo and cabozantinib in the randomized stage, cabozantinib treatment resulted in a median progression-free survival of 23.9 weeks (95% CI 10.7 to 62.4 weeks) compared to 5.9 weeks (95% CI 5.4 to 6.6 weeks, hazard ratio 0.12, P<0.001) with placebo. These results demonstrate the promising clinical activity of cabozantinib in advanced PCa. Little direct evidence to date has shown cabozantinib’s inhibitory effects on the tumor and bone response to tumors, as well as its effects on normal bone
To determine the target rationale for cabozantinib in advanced PCa, we first set out to examine the expression of MET and VEGFR2 and levels of P-MET in PCa metastases. Our results highlight MET and VEGFR2 as important targets in PCa metastases, especially in bone, as levels of MET, P-MET and VEGFR2 were significantly upregulated in PCa BM as compared to primary PCa. Furthermore, both of the receptors are also expressed in soft tissue metastases, indicating that MET and VEGFR2 play important roles in the progression of PCa in multiple metastatic settings. In addition, the expression profiles of all cabozantinib targets tested (MET, VEGFR2, AXL, RET, and KIT) in LuCaP PCa xenografts demonstrate that all of these targets are expressed in advanced PCa. The expression pattern in the xenografts is heterogeneous, suggesting that cabozantinib may be broadly applicable in PCa treatment. In concordance with our results with the LuCaP models, AXL, RET and KIT were also detected in advanced PCa in other studies [
In our studies, we focused on advanced metastatic disease but our results also provided additional information about primary disease. Our analyses did not reveal significant differences between MET and P-MET levels in NP epithelium vs. primary PCa, but indicated increases in VEGFR2 in primary PCa. In addition, there was no association between levels of MET, P-MET and VEGFR2 with disease recurrence. Similarly, no differences in MET levels were detected between NP and PCa in another report [
The AR is important in primary PCa as well as in advanced CRPC. Interestingly, AR activity has been reported to suppress MET expression in preclinical models [
Despite no apparent association between AR and MET in clinical specimens, evaluation of the xenografts showed that neuroendocrine LuCaP PCa models express higher levels of all cabozantinib targets in comparison to adenocarcinoma models. This finding indicates that this rare but aggressive subtype of PCa, which is not dependent on AR signaling, may be sensitive to cabozantinib or other agents targeting a similar profile of RTKs. Killing or inhibiting the growth of neuroendocrine cells might also provide an additional benefit in the treatment of adenocarcinoma, since it has been reported that in adenocarcinoma the presence of the rare neuroendocrine cells supports tumor progression. Efficacy of simultaneous inhibition of VEGFR2 and MET was also shown in pancreatic neuroendocrine tumors, with decreased tumor growth and reduction in invasion and metastasis [
In the phase II randomized discontinuation trial, cabozantinib treatment resulted in resolution of lesions on bone scan and regression of soft tissue lesions. However, the effects of cabozantinib on PCa tumors in soft tissues or in the bone microenvironment have not yet been delineated. Therefore, we focused on evaluating the effects of cabozantinib on osteoblastic and mixed osteoblastic/osteolytic PCa in bone and subcutaneous tumors in a preclinical setting. Our data clearly show that cabozantinib treatment kills tumor cells in the bone and in subcutaneous tumors as demonstrated by large necrotic areas in treated tumors, tumor regression, and lower PSA in animals treated with cabozantinib vs. control animals. Interestingly, there are remaining foci of healthy tumor cells adjacent to the necrotic areas, suggesting involvement of angiogenesis and also indicating that the treatment did not constitute a cure, and possibly continuous treatment might be needed to control the disease progression. Notably this was true for androgen-sensitive as well as castration-resistant tumors.
It is important to note that in the clinical situation cabozantinib treatment does not result in consistent decreases of PSA. In many patients PSA levels did not correlate with the reduction in tumor burden that resulted from cabozantinib treatment. Since the observed regression in soft tissue lesions suggests that cabozantinib may have an effect on tumors that is independent of bone environment, this finding is surprising. As such, it is necessary to delineate whether cabozantinib acts on the tumor and/or the microenvironment, and whether this efficacy is a function of androgen sensitivity. Our results agree with the clinical observations. Even though the PSA levels were lower in the treated group vs. the control group, the PSA levels were not lower in the majority of animals bearing intratibial tumors after treatment when compared to pre-treatment levels. Taken together, our data indicate that cabozantinib is efficacious in models of androgen-sensitive PCa as well as in CRPC, and show that cabozantinib efficacy does not solely depend on the bone microenvironment. These data are consistent with cabozantinib’s clinically beneficial observations in patients both with and without bony metastases.
One of our objectives was to investigate the mechanisms of cabozantinib effects on the tumor. Our analysis showed the presence of large necrotic areas in the tumors in bone and significant decreases in the volume of subcutaneous tumors. To demonstrate that the observed inhibitory effects of cabozantinib are due to inhibition of VEGFR2 and MET signaling one would like to demonstrate “on target effects”
The 60 mg/kg dose was well-tolerated in our preclinical models for four weeks of administration when tumors were growing in the bone; but body weight loss was observed with longer treatment. Interestingly the loss of body weight was more pronounced in animals bearing LuCaP 23.1 tumors than in those with C4-2B tumors. A lower dose of cabozantinib (30 mg/kg) was well-tolerated up to six weeks, and the animals lost less weight compared to those that received a higher dose. Therefore, we conclude that cabozantinib treatment is well-tolerated for 4-6 weeks with some negative effects on body weight after prolonged treatment of intra-tibial tumors.
Our results show that cabozantinib affects not just the tumor, but also the bone response to the tumor. We used two models that cause bone formation when growing in the bone environment: LuCaP 23.1 and C4-2B. Interestingly, we observed opposing effects of cabozantinib on bone volume in these models; large decreases in bone volume in the LuCaP 23.1 tumored tibiae and small increases in bone volume in C4-2B tumored tibiae. We hypothesize that the differences are due to the different magnitude of new bone formation associated with the two different tumors. LuCaP 23.1 is highly osteoblastic, causing 5-fold increases in bone volume, and we hypothesize that the decreases in BV in the LuCaP 23.1 model after cabozantinib treatement are due to cabozantinib’s tumor inhibitory activity (less tumor=less new bone). In comparison to LuCaP 23.1, C4-2B tumors are osteoblastic but cause only ~1.5 fold increase in BV and also induce a significant osteolytic reaction [
Because VEGFR2 and MET signaling are important in bone biology, it is clear that cabozantinib treatment may not only affect tumored bone, but potentially normal bone as well. Our results show that cabozantinib treatment resulted in increased BV in intact as well as castrated animals. Patients with CRPC are typically on ADT, which causes osteopenia and osteoporosis. Therefore, cabozantinib therapy might provide additional benefits to patients with advanced PCa who are on ADT. The systemic effect of cabozantinib on the skeleton might lead to decreased number of and time to skeletal related events in patients.
In conclusion, we demonstrate that cabozantinib targets are expressed in advanced PCa, indicating that this treatment has the potential for substantial efficacy in the heterogeneous PCa population. Angiogenesis inhibition has not yielded significant advances in the treatment of CRPC to date, and a growing body of evidence suggests that signaling through MET may be a compensatory mechanism by which tumor cells escape anti-angiogenic therapy. Thus, since cabozantinib targets VEGFR2, MET and a number of other RTKs simultaneously, it represents an attractive, new opportunity in anti-angiogenic CRPC treatment. Furthermore, our results show that cabozantinib inhibits tumors in two different PCa: an androgen-dependent osteoblastic model and a castration-resistant mixed osteoblastic/osteolytic model. Furthermore, cabozantinib also affects tumors growing in the bone and subcutaneous tumors, again indicating the potentially high clinical impact of this treatment.
(TIF)
(TIF)
(TIF)
(DOCX)
(DOCX)
The authors would like to thank Michiyo Dalos and Jessica Olson for assistance with animal studies, Douglas Laird for helpful discussions, and Peter Kirk and I-Chuan Hsiao for additional technical assistance.