Bone morphogenetic protein (BMP) is a pleiotropic growth factor that has been implicated in inflammation and prostate cancer (CaP) progression. We investigated the potential role of BMP-6 in the context of macrophages and castration-resistant prostate cancer. When the androgen-responsive murine (Tramp-C1 and PTENCaP8) and human (LNCaP) CaP cell lines were cocultured with macrophages in the presence of dihydrotestosterone, BMP-6 increased androgen-responsive promoter activity and cell count significantly. Subsequent studies revealed that BMP-6 increased the expression level of androgen receptor (AR) mRNA and protein in CaP cell lines only in the presence of macrophages. Simultaneously, the AR antagonists bicalutamide and MDV3100 partially or completely blocked BMP-6-induced macrophage-mediated androgen hypersensitivity in CaP cells. Abolishing interleukin-6 signaling with neutralizing antibody in CaP/macrophage cocultures inhibited the BMP-6-mediated AR upregulation in CaP cells. Using Tramp-C1 and PTENCaP8 cells with a tetracycline-inducible expression of BMP-6, the induction of BMP-6 in vivo resulted in a significant resistance to castration. However, this resistance was blocked after the removal of macrophages with clodronate liposomes. Taken together, these results show that BMP-6 induces castration resistance by increasing the expression of AR through macrophage-derived interleukin-6.
Prostate cancer is the second leading cause of cancer death in the USA. Despite recent progress, patients with metastatic CaP still have a poor prognosis due to the inevitable emergence of CRPC. Although multiple mechanisms underlying the development of CRPC have been reported, androgen hypersensitivity and altered tumor microenvironment have been implicated.
Bone morphogenetic proteins, with more than 20 subtypes, represent the largest subfamily within the transforming growth factor-β superfamily. In CaP, the expression levels of BMPs increase while those of the cognate receptors decrease with progression of disease.[2, 3] Although the role of BMPs on CaP cells in tissue culture varies with cell types and experimental conditions, the prevailing effect appears to be growth inhibition.[2, 4, 5] As CaP cells produce BMPs, and BMPs generally decrease the proliferation of CaP cells, it is likely that the loss of BMP receptor expression will permit CaP cells to escape the physiologic constraint on cellular proliferation imposed by BMPs and lead to a more aggressive phenotype by altering the tumor microenvironment.
Recently, we reported that BMP-6 induces the expression of IL-6 in macrophages. Tumor-infiltrating macrophages and IL-6 are frequently detected in CaP and have been linked to altered sensitivity to androgens and androgen antagonists.[7-10] Based on these observations, we investigated the effect of BMP-6 on androgen sensitivity in CaP cells in the context of macrophages. Our results indicate that BMP-6 induces castration resistance in CaP cells by modulating TIMs to express IL-6 which, in turn, upregulates the transcription of AR.
Materials and Methods
The LNCaP, THP-1, Tramp-C1, and RAW 264.7 cell lines were purchased from ATCC (Manassas, VA, USA) and maintained in the recommended media. The PTENCaP8 murine CaP cell line, kindly provided by Dr. Hong Wu (UCLA School of Medicine, Los Angeles, CA, USA), was routinely cultured as previously described. To generate cell lines that express BMP-6 in an inducible manner, the lentivirus-based tetracycline inducible system was used (Invitrogen, Carlsbad, CA, USA). When investigating androgen sensitivity, 1% charcoal-stripped FBS was used.
Bone morphogenetic protein-6 and neutralizing antibodies to IL-1α, IL-6, and IL-10 were purchased from R&D Systems (Minneapolis, MN, USA). The small molecule inhibitors U0126, NSC74859, GDC0941, bicalutamide, and MDV3100 were obtained from Selleckchem (Houston, TX, USA). Probasin–luciferase was from Addgene (Cambridge, MA, USA). ClodronateLiposomes.org (Amsterdam, the Netherlands) kindly provided the clodronate- and PBS-loaded liposomes.
Total RNA was extracted and reverse transcribed using the ImProm-II reverse transcription system (Promega, Madison, WI, USA). Quantitative PCR conditions were: 95°C for 15 s and 60°C for 1 min for 40 cycles using an Applied Biosystems (Foster City, CA, USA) StepOnePlus Q-PCR machine. Primer sequences are shown in Table S1.
Transfection and luciferase activity assay
Cells were transfected with 1 μg/mL Lipofectamine 2000 (Invitrogen) and 1 μg/mL plasmid. Luciferase activity assay was measured out using the Dual-Luciferase Reporter Assay kit (Promega). The human androgen responsive reporter containing PSA-Lux was generated using PCR.
After harvesting cells, nuclear protein was purified using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, Rockford, IL, USA). After electrophoresis, protein was transferred to a nitrocellulose membrane (Bio-Rad, Hercules, CA, USA) and analyzed with an enhanced chemiluminescence kit (Thermo Scientific). The primary antibodies anti-Ki67 (Abcam, Cambridge, MA, USA) and anti-androgen receptor (Santa Cruz Biotechnology, Dallas, TX, USA) were used.
Intracellular testosterone concentrations were measured using the EIA kit (Cayman Chemical, Ann Arbor, MI, USA) according to the manufacturer's protocol.
The animal protocol was approved by the Institutional Animal Care and Use Committee (University of Medicine and Dentistry of New Jersey, http://rwjms.umdnj.edu/research/orsp/ra/iacuc.html). Mice were obtained from Jackson Laboratory (Bar Harbor, ME, USA) and injected s.c. with 2 × 106 Tramp-C1/BMP-6 cells (C57BL/6 mice) and PTENCaP8/BMP-6 cells (CB6F1/J hybrid mice). Seven days later all animals were randomized into designated groups of five mice each and castrated surgically. Where indicated, mice were given doxycycline (2 mg/mL doxycycline in 5% sucrose) orally in drinking water. In addition, animals in the designated group received either 0.2 mL clodronate liposomes i.p. (final dose of 25 mg/kg) every 4 days, or equal volume of PBS liposomes. At the end of the experiment all animals were killed and palpable tumors were harvested and analyzed.
For all analyses, Student's t-test was carried out. A P-value of <0.05 was considered statistically significant.
Bone morphogenetic protein-6 increases androgen sensitivity in CaP cell lines
Initially, the androgen-sensitive murine CaP cell lines Tramp-C1 and PTENCaP8 were transiently transfected with the androgen-responsive reporter plasmid PB-Lux. Following the transfection, cells were cocultured with the murine macrophage cell line RAW 264.7 and treated with 10 ng/mL BMP-6 and 10 nM DHT for 48 h. The results revealed that BMP-6 raised luciferase activity levels 1.8-fold in Tramp-C1 and 2.6-fold in PTENCaP8 cells in the presence of RAW 264.7 (Fig. 1a). Simultaneously, BMP-6 also increased the CaP cell count by approximately 40–50% under identical experimental conditions (Fig. 1b). Interestingly, BMP-6 alone also moderately increased probasin promoter activity and cell count.
Next, a similar experiment was carried out using the LNCaP human CaP cell line and THP-1 monocytes. Because PB-Lux reporter is only active in murine cell lines, the human equivalent PSA-Lux was used. The results corroborated the findings of the murine CaP cell lines in that 10 ng/mL BMP-6 increased the luciferase activity level in LNCaP by 2.9-fold when LNCaP/THP-1 coculture was treated with 10 nM DHT (Fig. S1).
Bone morphogenetic protein-6 induces AR expression in CaP cells in the presence of macrophages
As reported mechanisms of enhanced androgen sensitivity in CaP cells include AR upregulation and intracrine androgen synthesis, we next measured the AR expression and cellular testosterone levels. In Tramp-C1 and PTENCaP8 cells, BMP-6 increased AR mRNA and protein when incubated with macrophages (Fig. 2a, left and right panels, respectively). The slight increase in AR in prostate cancer cell lines in the presence of macrophages (without BMP-6) likely reflects the previously reported low basal expression of BMP-6 in prostate cancer cell lines. Simultaneously, decrease in p21 protein (Fig. 2a, right panel) and increased nuclear translocation of AR were detected (Fig. 2b). However, there was no effect on the cellular testosterone levels (Fig. 2c). When bicalutamide and MDV3100 were used to block AR, BMP-6 no longer increased the probasin promoter activity in Tramp-C1 and PTENCaP8 when RAW 264.7 cells were present (Fig. 2d).
When LNCaP was examined, BMP-6 again increased AR mRNA and induced AR nuclear translocation in the presence of THP-1 cells (Fig. S2a,b, respectively). More interestingly, there was an increased intracrine androgen synthesis in LNCaP/THP-1 coculture when treated with BMP-6 (Fig. S2c). Subsequent analysis with Q-PCR showed that the key enzymes in androgen synthesis, CYP11A1, HSD17β3, and AKRC1, were all upregulated in LNCaP cells when THP-1 cells along with BMP-6 were present (Fig. S2d). In contrast, these three enzymes were not detected in Tramp-C1 or PTENCaP8 cells (data not shown).
Bone morphogenetic protein-6-induced AR expression in CaP cells when cocultured with macrophages is mediated through IL-6
Next, the macrophage-derived factor induced by BMP-6 that mediates AR upregulation in CaP cells was investigated using a cytokine array. In addition to the previously reported IL-6, additional cytokines with increased expression levels in RAW 264.7 after BMP-6 treatment included IL-1α and IL-10 (data not shown). When CaP/macrophage cocultures were incubated with blocking antibodies to these three cytokines, only the IL-6 neutralizing antibody abrogated AR upregulation by BMP-6 in Tramp-C1, PTENCaP8, and LNCaP (Fig. 3a, Fig. S3a). When CaP/macrophage cocultures were incubated with IL-6 neutralizing antibody or LDN-193189, a small molecule inhibitor of BMP signaling, BMP-6 no longer induced the proliferation of Tramp-C1, PTEN CaP8, or LNCaP cells (Fig. 3b, Fig. S3b).
There are at least three different IL-6 signaling pathways (PI3K, STAT3, and ERK), so we next used the small molecule inhibitors U0126 (ERK), NSC74859 (STAT3), and GDC0941 (PI3K) to identify the pathway involved in AR upregulation in the current experimental conditions. Quantitative PCR showed that blocking PI3K (GDC0941) completely inhibited the induction of AR expression by BMP-6 in Tramp-C1/RAW 264.7 and PTENCaP8/RAW264.7 cocultures (Fig. 3c); interestingly, STAT3 (NSC74859) also partially blocked AR upregulation in Tramp-C1 and PTENCaP8 murine cells but not the LNCaP human CaP cell line (Fig. 3c, Fig. S3c). We have previously reported NE differentiation of CaP cells involving the BMP-6/macrophages/IL-6 loop, we investigated whether the PI3K signaling pathway also mediates NE differentiation. The results showed that blocking STAT3 but not PI3K abolished BMP-6-induced NE differentiation (indicated by parathyroid hormone-related protein expression) in all three CaP/macrophage cocultures (Fig. 3d, Fig. S3d).
Overexpression of BMP-6 induces castration resistance in CaP cells
The current results collectively suggest that BMP-6 induces androgen hypersensitivity in CaP cells through macrophages. To test whether this effect of BMP-6 translates into castration resistance in vivo, Tramp-C1 and PTENCaP8 cells expressing BMP-6 (Tramp-C1/BMP6 and PTENCaP8/BMP6, respectively) in a tetracycline-inducible manner were established by screening more than 20 clones each. A representative figure showing the induction of BMP-6 expression by tetracycline is shown in Figure S4. Subsequently, Tramp-C1/BMP6 was inoculated s.c. into wild-type B6. As PTENCaP8 cells were derived by crossing B6 and BALB/c mice, PTENCaP8/BMP6 was injected into CB6F1/J hybrid mice. When tumors became palpable, animals in the designated group were surgically castrated and BMP-6 was induced with doxycycline. The results showed that the induction of BMP-6 rendered both Tramp-C1 and PTENCaP8 cells resistant to castration (Fig. 4a). In these tissues, Q-PCR confirmed that both AR and IL-6 increased with BMP-6 induction (doxy+) following castration (Fig. S4b). When macrophages were removed with clodronate liposomes (Figs S5,S6), this BMP-6-mediated castration resistance was completely abolished (Fig. 4b).
In the present study, we showed that BMP-6 increases androgen sensitivity in CaP cells in the presence of macrophages by stimulating macrophages to produce IL-6 which, in turn, upregulates the expression of AR. Intracellularly, the upregulation of AR involves the PI3K pathway. Studies in immune intact mice revealed that the overexpression of BMP-6 leads to CRPC through macrophages. Taken together, these results suggest that tumor-derived BMP-6 may render CaP cells resistant to castration by stimulating TIMs to produce IL-6.
Although BMPs have been implicated in carcinogenesis, their precise roles remain unclear. In the context of CaP, frequent overexpression of BMPs accompanied by the loss of expression of BMP receptors has been reported.[2, 3] As the direct effect of BMPs on CaP cells includes inhibition of cellular proliferation and androgen signaling,[2, 16] the loss of BMP receptors may result in the break of a negative feedback loop that ultimately leads to overexpression of BMPs. Such a hypothesis then suggests that BMPs may have a paracrine-mediated oncogenic activity. In this regard, results of the present study have shown that the overexpression of BMP-6 in CaP leads to the emergence of CRPC in mice, which is mediated by TIMs.
Although the mechanism underlying the emergence of CRPC is heterogeneous and multifactorial, one overarching mechanism involves alteration in the tumor microenvironment. For example, deranged inflammatory response and altered immune effector cells have been detected and lymphotoxin B secreted by B cells has been linked to the development of CRPC. Results of the present study suggest that TIMs also may contribute to the emergence of CRPC. Specifically, when BMP-6 was overexpressed, CaP cells became hypersensitive to androgens and resistant to castration when macrophages were present. These results, taken together with our recent observation that BMP-6 regulates macrophages and a previously published report that indicated increase expression levels of BMP-6 in human CaP tissues, suggest that tumor-derived BMP-6 may be a potent regulator of TIMs in CaP.
Macrophages are frequently detected in many solid tumors, including prostate.[7, 19] Yet, the precise effect of these macrophages found in tumors has been controversial. For example, macrophage activation increases cytotoxicity. In contrast, TIMs have been linked to enhanced angiogenesis, invasion, and poor prognosis. In the present study, TIMs were pro-tumorigenic. Specifically, the overexpression of BMP-6 by CaP cells permitted hypersensitivity to DHT only in the presence of macrophages in tissue culture. However, as the precise phenotype of TIMs is unclear, it is uncertain whether RAW 264.7 and THP-1 are appropriate models to test the biology of TIMs. For this reason, however, an in vivo study was carried out using CaP cell lines with a tet-inducible BMP-6 expression. The results showed that BMP-6 secreted by CaP cells induced castration resistance, likely by permitting CaP cells to respond to low levels of androgens. This BMP-6-mediated castration resistance was completely blocked when macrophages were removed. As the oncogenic effect of macrophages in the present experimental context was driven entirely by tumor-derived BMP-6, it is likely that the decision by macrophages to become pro- or antitumorigenic is dictated by the factors secreted by tumor cells.
Results of the present study yielded new insights concerning the mechanism of androgen hypersensitivity and NE differentiation in the context of BMP-6/macrophages/IL-6. First, the dynamic interplay between CaP cells and macrophages involving the BMP-6/IL-6 loop results in increased androgen sensitivity through the upregulation of AR. Interleukin-6 has long been implicated in prostate carcinogenesis, especially in the context of NE differentiation and androgen sensitivity. The precise effect of IL-6, however, varies with the experimental conditions. For example, in the LNCaP human CaP cell line, IL-6 induces NE differentiation and decreases androgen sensitivity acutely. However, chronic exposure to IL-6 resulted in upregulation of AR and intracrine androgen synthesis and loss of NE differentiation.[8, 9] As the AR expression level increased in the current experiments, the coculture of macrophages/CaP in the presence of BMP-6 likely mimics the chronic exposure IL-6 model. Second, the IL-6 signaling pathways involved in AR upregulation and NE differentiation in the present experimental context are different. Indeed, using pathway-specific small molecule inhibitors, results of the present study showed that macrophage-derived IL-6 induces AR upregulation through PI3K and NE differentiation by STAT3. This is consistent with a previous report that STAT3 induces NE differentiation in CaP cells. These observations show that AR upregulation and NE differentiation are not interdependent and that these two pathways can be uncoupled. More importantly, these results collectively suggest that NE differentiation and AR expression are not mutually exclusive in the context of macrophage-derived IL-6, and that AR upregulation and NE differentiation may be investigated separately.
Our data also suggest a provocative hypothesis concerning NE differentiation and AR expression status in CaP cells. Classically, NE cells associated with CaP cells are associated with aggressive features and do not express AR. Previously, we have reported that the BMP-6/IL-6 loop involving CaP–macrophage interaction induces NE differentiation. In the present study, we have shown that identical experimental conditions with BMP-6 in CaP–macrophage coculture induce upregulation of AR through IL-6. Collectively, these results suggest that NE differentiation and AR upregulation occur at the same time in the context of macrophage-derived IL-6 and point to the potential existence of NE-like CaP cells that express AR. Additional studies are underway to verify this concept.
It should be pointed out, however, that the mechanism of castration resistance induced by BMP-6 in the presence of macrophages may not be limited to AR upregulation. Indeed, when LNCaP–macrophage coculture was treated with BMP-6, intracellular testosterone levels increased. Coincidentally, induction of key enzymes involved in androgen synthesis was detected. As IL-6 has been reported to enhance intracrine androgen production, it is likely that IL-6 is the mediator of increased testosterone levels in LNCaP in the experiments described herein. However, the effect of BMP-6 on androgen production in CaP/macrophages was not observed in murine CaP cell lines Tramp-C1 and PTENCaP8. These observations suggest that murine CaP cell lines are not adequate models for investigating intracrine androgen synthesis. Currently, studies using additional human cell lines are in progress to determine whether the link between BMP-6 and intracellular androgen production is a unique feature of LNCaP cells.
Finally, we have previously reported that BMP-6 inhibits the proliferation of prostate cancer cell lines when cultured in normal serum. In the present study, BMP-6 moderately increased Tramp-C1 and PTEN CaP8 proliferation when supplemented with charcoal-stripped serum. These observations suggest that the presence of hormones alters the ultimate effect of BMPs in prostate cancer cell lines. Such a hypothesis is consistent with the report that BMP-2 inhibits LNCaP cells in the presence of androgen analog R1881 and stimulates in the absence of androgens. At the present time, the precise underlying mechanism for the serum-dependent effect of BMP-6 is unclear. However, the downregulation in p21 suggests that the effect of BMPs on protein expression varies with culture conditions. Future studies will focus on identifying factors involved in this process.
In conclusion, the present study showed that TIMs potentially contribute to the emergence of CRPC. Mechanistically, this interaction between CaP and TIMs in the context of androgen sensitivity involved the BMP-6/IL-6 loop. Thus, targeting BMP-6 or macrophages may be a novel therapeutic approach in men with CRPC.
This study was funded by the Marion & Norman Tanzman Charitable Foundation, Mr Malcolm Wernik and Mr Jeffries Shein.