Alterations of tumor microenvironment by nitric oxide impedes castration-resistant prostate cancer growth

Significance

This study presents insights into the underexplored areas of castration-resistant prostate cancer (CRPC) therapeutics—the role of nitric oxide (NO) in CRPC reduction through its microenvironment. Results of this study provide important information on the tumor reduction capabilities of increased NO levels and its mechanistic aspect and demonstrates the potential long-term efficacy of NO on CRPC. An in-depth understanding of how NO affects the tumor microenvironment will allow development of chemotherapeutics based on NO for a CRPC cure.

Abstract

Immune targeted therapy of nitric oxide (NO) synthases are being considered as a potential frontline therapeutic to treat patients diagnosed with locally advanced and metastatic prostate cancer. However, the role of NO in castration-resistant prostate cancer (CRPC) is controversial because NO can increase in nitrosative stress while simultaneously possessing antiinflammatory properties. Accordingly, we tested the hypothesis that increased NO will lead to tumor suppression of CRPC through tumor microenvironment. S-nitrosoglutathione (GSNO), an NO donor, decreased the tumor burden in murine model of CRPC by targeting tumors in a cell nonautonomous manner. GSNO inhibited both the abundance of antiinflammatory (M2) macrophages and expression of pERK, indicating that tumor-associated macrophages activity is influenced by NO. Additionally, GSNO decreased IL-34, indicating suppression of tumor-associated macrophage differentiation. Cytokine profiling of CRPC tumor grafts exposed to GSNO revealed a significant decrease in expression of G-CSF and M-CSF compared with grafts not exposed to GSNO. We verified the durability of NO on CRPC tumor suppression by using secondary xenograft murine models. This study validates the significance of NO on inhibition of CRPC tumors through tumor microenvironment (TME). These findings may facilitate the development of previously unidentified NO-based therapy for CRPC.

Prostate cancer is the second most frequent cause of cancer-related deaths in men. Men with prostate cancer that has recurred after local therapy usually respond to androgen deprivation therapy (ADT); however, despite this treatment, most patients eventually experience progression of the disease within 2 y, a condition known as castration-resistant prostate cancer (CRPC) (1). In trying to understand the causes of this androgen resistance that develops in CRPC, most research has focused directly on the splice variants of the androgen receptor (ARVs) (2). However, the tumor microenvironment (TME) has been shown to play a major role in tumor progression, yet the response to therapy in other cancer types has been inadequately studied in CRPC (3–5). TME is comprised of a variety of cell types, including immune cells, fibroblasts, pericytes, and tumor-associated macrophages (TAMs) (6). TAMs are recruited to tumors from diverse signaling molecules such as chemokines (CCL-2 and CCL-5) and cytokines (IL-34 and CSF-1) (7). While the exact mechanism is unknown, TAMs have been reported to play a key role in the progression of prostate cancer through the secretion of cytokines, matrix metalloproteinases, and growth factors (8, 9).

A key molecule in the regulation of TME interactions is the ubiquitous nitric oxide (NO) (10–12). We have previously established the importance of NO in the cardiovascular and immune systems and in male secondary hypogonadism (13–20). Others have investigated the tumoricidal implications of NO in therapeutic resistance (21–23), cell survival, proliferation of tumors, inhibition of tumor growth, and reduction in lung metastases (24) in many cancer types (25). Several NO donors, including S-nitrosothiols, organic nitrates, and Metal-NO complexes, have shown impacts on cancer progression (26, 27). S-nitrosothiols such as S-nitroso-N-acetylpenicillamine (SNAP) and S-nitrosoglutathione (GSNO) have also shown promising effects as antineoplastic agents. However, these studies have only focused on certain aspects of NO such as its role in both progrowth and antigrowth effects (28), cellular localization, endogenous expression of the androgen receptor (AR) (29), and AR function inactivation by S-nitrosylation (30, 31). In the present study, we document the effect of NO on tumor suppression by targeting the CRPC TME through TAMs.

Results

Increased NO Levels Affect Testosterone and Luteinizing Hormone.

We have previously shown that mice lacking the S-nitrosoglutathione reductase (GSNOR) gene showed increased nitrosative stress and exhibited secondary hypogonadism (19). Therefore, in this study we examined if increased NO levels are able to suppress testosterone (T) and luteinizing hormone (LH) in control C57BL/6J mice. For these experiments, we administered GSNO (an NO donor) intraperitoneally (10 mg/kg) for 7 d and compared T and LH levels to mice administered with phosphate buffer saline (PBS). We found that T and LH levels were decreased 6× in the mice that received GSNO (Fig. 1A). Further validation was obtained by oral administration of GSNO (10 mg/400 mL bottle for 5 wk) to C57BL/6J mice, as we found that T levels were undetectable in mice that received oral GSNO compared with untreated animals (Fig. 1B). To confirm if increased NO levels are capable of affecting the hypothalamic-pituitary-gonadal axis to regulate LH and T levels, we checked GnRHR expression in the brain (Fig. 1C), and our results confirmed that increased NO levels were capable of reducing the expression of GnRHR in the brain. Together, these results indicated that increased NO levels affect receptors for GnRH, LH, and T through the hypothalamic-pituitary-gonadal axis.

Fig. 1.

Increased levels of NO affect levels of testosterone (T), FSH, and LH. (A) To validate the impact of increased NO levels on T, FSH, and LH, C57/BL6 mice were treated with 10 mg/kg GSNO for 7 d and levels of T, FSH, and LH were checked. (B) Additionally, C57/BL6 mice were kept on the oral dosage of GSNO at 10 mg per cage per week for 5 wk. Brains from i.p. GSNO-treated mice were harvested and analyzed for the presence of GnRHR (C), showing that NO levels affect major pathways of production of hormones (T, LH, FSH) by influencing hypothalamic-pituitary-gonadal axis regulation. (Scale bars: C, first three columns, 750 μm; C, Right, 100 μm.)

NO Reduces Tumor Burden in the Murine Model of CRPC.

We hypothesized that increased NO can lead to CRPC tumor suppression because of its ability to affect the hypothalamic-pituitary-gonadal axis. After castration, 2.5 million 22RV1 cells were xenografted s.c. in each flank of SCID mice. GSNO (10 mg/kg per d i.p.) was administered to half of the mice (experimental group) and an equal volume of PBS to the remaining animals. After 14 d of GSNO treatment, the grafts were harvested (Fig. 2A), and we found that the overall tumor burden was decreased in mice treated with GSNO (high NO levels) (Fig. 2 B–E), significantly affecting the weight of mice (Fig. 2C). We evaluated the NO levels in tumors isolated from mice treated with PBS or GSNO using Griess test and confirmed an inverse association between tumor burden and NO levels (Fig. 2F). Tumors from mice that received GSNO treatment showed areas with less necrosis (Fig. 2G and SI Appendix, Fig. S1). Furthermore, GSNO-treated mice showed a reduced number of Ki-67– positive cells, suggesting a reduced proliferation rate (Fig. 2H).

Fig. 2.

Increased NO levels suppress tumor burden in CRPC mouse model (castrated SCID mice with s.c. xenograft of 22RV1). (A) Experiment with in vivo xenograft. (B) Xenograft tumors isolated from both flanks after 2 wk of treatment (PBS/GSNO) (each cup represents tumors from one mouse). (C) Animal weight. (D) Tumor volume. (E) Tumor weight. GSNO-treated CRPC mouse models have fewer necrotic areas and fewer proliferating cells in tumor grafts. (F) Relative comparison of nitrate levels between control and GSNO-treated mice. (G) H&E staining, showing the sections from tumor xenografts that received PBS (control) vs. treatment with GSNO (10 mg/kg per d i.p.). Grafts from PBS-treated mice have more areas of necrosis. (H) Ki-67 immunostaining of sections showing that proliferation is suppressed in GSNO-treated grafts. (Scale bars: 500 μm.)

NO Suppresses Tumor Burden in a Cell Nonautonomous Manner.

Most chemotherapeutic drugs as well as ionizing radiation promote autophagy in tumor cells (32–35). However, autophagy can have both cell autonomous and cell nonautonomous effects and influence the outcome of therapy (35). Therefore, we evaluated whether the effects of NO are tumor cell autonomous or nonautonomous. We studied the cell proliferation rate after treating the 22Rv1 cells with varying concentrations of GSNO in vitro. Interestingly, we found that in vitro the proliferation of 22Rv1 cells was largely unaffected (Fig. 3A), despite the efficacy of GSNO on tumor burden in vivo (Fig. 2). Because 22Rv1 cell growth is dependent on constitutive signaling of the androgen receptor (AR), we examined the AR signaling markers involved in androgen-induced activation of endogenous AR, such as PSA and TMPRSS2 (29). Both PSA and TMPRSS2 levels were suppressed by escalating doses of GSNO (Fig. 3B). Taken together, changes in AR signaling without changes in cell proliferation in vitro suggest the effects of NO on 22Rv1 cells are likely cell nonautonomous.

Fig. 3.

NO targets CRPC in cell nonautonomous manner. (A) Cell proliferation assay (MTT) using GSNO concentrations of 0, 5, 10, 25, 50, and 100 μM on 22RV1 cells showed no specific inhibition, indicating effects of NO on 22Rv1 cells are likely cell nonautonomous. (B) Relative expression of PSA and TMRPSS2 upon varying concentrations of GSNO ranging from 10, 25, 50, and 100 μM at RNA levels. (C) Twenty-six cytokines (CCL27, CD54, TIMP-1, ACRP30, G-CSF, AR, IL-17A, beta NGF, IL-2 R alpha, CCL28, Axl, CCL7, CCL17, CXCL6, IL-1 F2, M-CSF, TGFBeta 3, CXCL13, BMP-6, CCL23, NT-3, CCL11, CCL1, IL-5, IL-6, and IGFBP4) that were found to be suppressed more than 1.5-fold in tumors isolated from mice that received GSNO treatment compared with control mice that received PBS. (D) Impact of NO on TAM M1 macrophage (iNOS) induced and M2 (CD206 and Arginase-1) at RNA levels. (E–G) Impact of GSNO on M2 (F4/80 and CD206) and M1 macrophage (iNOS) at protein levels. (Scale bars: E, 750 μm; F, 100 μm; G, 75 μm.)

Cytokine Signature Showed That NO Suppresses TAMs Affecting CRPC TME.

The cell nonautonomous effects are related largely to the immune system (35), TME, and subclone heterogeneity (36). Macrophages are known to increase specific circulating cytokines with progressive metastasis (37, 38). Therefore, to study the molecular events leading to NO-induced changes, we evaluated protein expression of 120 cytokines using the tumors from mice that received GSNO versus PBS (SI Appendix, Fig. S2). Among 120 cytokines assayed, there were 26 cytokines (CCL27, CD54, TIMP-1, ACRP30, G-CSF, AR, IL-17A, beta NGF, IL-2 R alpha, CCL28, Axl, CCL7, CCL17, CXCL6, IL-1 F2, M-CSF, TGFBeta 3, CXCL13, BMP-6, CCL23, NT-3, CCL11, CCL1, IL-5, IL-6, and IGFBP4) that were suppressed more than 1.5 times (Fig. 3C). From previously published studies, we determined the role of these cytokines in cancer progression (SI Appendix, Table S2). Among the 26 affected cytokines, the macrophage colony stimulating factor (M-CSF) and granulocyte M-CSF (GM-CSF) play an essential role in the regulation of TAMs (39, 40). TAMs promote tumor progression and are resistant to various chemotherapeutic agents (41, 42) in prostate cancer (43) by differentiating into either cytotoxic (M1) or tumor growth promoting (M2) states (44). Therefore, to study the implications of increased NO levels on TAMs, we evaluated the markers of proinflammatory (M1) and antiinflammatory (M2) macrophages in tumors from mice treated with GSNO versus PBS. We found that following therapy with NO, expression of M2 macrophage markers (F4/80, CD206, Arginase) was suppressed and expression of M1 macrophage marker (iNOS) was increased (Fig. 3 D–G). This indicates that TAMs, a significant component of the antiinflammatory cell (M2) that infiltrates in prostate cancer (44), are suppressed by increased NO levels.

NO Influences TAMs by Targeting Their Activity and Differentiation.

Recent studies have suggested the importance of phospho-ERK1/2 levels with respect to combinations of lactate and hypoxia that eventually affect the fate of TAMs (45–48). Therefore, to verify whether TAM activity is sensitive to increased levels of NO, we checked the levels of pERK in tumors from mice which received GSNO treatment and compared them with tumors from PBS-treated mice using immunostaining and Western blot. We found that levels of pERK were suppressed upon GSNO treatment (Fig. 4 A and B and SI Appendix, Fig. S5D). However, total levels of ERK protein remained similar in all of the conditions (Fig. 4A), In addition, we determined the impact of increased levels of NO on other factors which are markers of TAM activity, such as VEGF (angiogenic marker), androgen receptor (AR), and androgen receptor splice variant 7 (AR-V7) (a critical determinant of resistance development in CRPCs) (49, 50), and found their levels to be consistently suppressed upon treatment of tumors with GSNO (Fig. 4 A and C and SI Appendix, Fig. S5 A–C). Together, the results suggest a strong effect of NO levels on regulation of TAM activity.

Fig. 4.

Impact of NO on TAM activity and differentiation. To validate the impact of increased levels of NO on TAM activity, the expression of pERK, ERK, AR-V7, and VEGF was checked in tumor grafts isolated from animals treated with PBS or GSNO. (A) Western blot confirmed reduced levels of AR-V7 as well as pERK upon increased NO. Immunostaining confirmed a reduced number of cells staining positive for pERK (B) and VEGF (C) upon increased NO levels. (Scale bars: B, 500 μm; C, 100 μm.) (D) To validate the impact of increased levels of NO on TAM differentiation, we evaluated expression of IL-34 in 22RV1 cells treated with 10, 25, 50, and 100 μM GSNO. (E) To establish the significance of suppression of IL-34–CSF1R interaction, CSF1R was blocked by GW2580 at 0.5, 10, 25, and 50 μM, followed by treating 22RV1 cells with GSNO. Inhibiting CSF1R abrogated GSNO suppressive impacts on PSA and TMPRSS2 levels.

Another aspect of TAM regulation that we focused on is TAM differentiation. Previous studies showed that tumor-derived factors like IL-34 and M-CSF educate macrophages to become the alternatively activated M2 type to promote angiogenesis, tissue remodeling, and immune suppression (13, 24). For differentiation of TAM, the binding of cytokines (IL34, CSF) to the CSF1 receptor (CSF1R) is critical (7). Enrichment analysis suggested the significance of IL-34 and CSF1R in resistance development in CRPC patients (Fig. 4D and SI Appendix, Fig. S3 A–E). Therefore, we explored the implications of increased NO levels on 22RV1 cells in the presence or absence of GW2580 (CSF1R inhibitor) at varying concentrations (0, 0.5, 10, 25, and 50 μM, respectively). Upon the suppression of IL-34–CSF1R interaction, expression of PSA and TMRPSS2 was abrogated (Fig. 4E). Taken together, decreased levels of pERK, AR in tumors, and abrogated levels of PSA and TMRPSS2 upon CSF1R inhibition strongly supports the regulatory role of NO on TAM activity and differentiation.

Short Half-Life of NO Does Not Affect Its Long-Term Impact on TAMs.

We evaluated the efficacy of increased NO levels in the reduction of CRPC tumor considering the short half-life of NO. For this, the CRPC cells (5 million cells per mouse) that were isolated from animals in the two treatment groups (PBS and GSNO) were xenografted s.c. into castrated SCID mice (n = 4 per group) (secondary xenograft). Tumors were then allowed to grow for a period of 4 wk (Fig. 5A). The results revealed that the overall tumor burden was significantly suppressed (Fig. 5B), with a reduction in the number of Ki-67–positive cells in animals that received cells from GSNO-treated mice (SI Appendix, Fig. S4). In addition, cells staining positive for M2 macrophages markers (F4/80, CD206) were decreased in animals that received cells from GSNO-treated mice (Fig. 5 C and D). Taken together, the effect of GSNO in secondary xenografts indicates the potential long-term therapeutic implications of increased NO levels on CRPC.

Fig. 5.

Long-term implications of NO on CRPC. (A) To validate the efficacy of NO on CRPC, we used secondary xenograft models in which castrated SCID mice were xenografted s.c. with 22RV1 cells followed by GSNO treatment in half of them for 4 wk. Cells from the two groups (control and experimental) of mice were harvested and reinjected into another set of castrated SCID mice, which were maintained for 4 wk but not treated (secondary xenograft). (B) The latter group showed a significant decrease in tumor burden. Mice that received cells from GSNO-treated animals showed a lower percentage of cells staining positive for M2 macrophages like F4/80 (C) and CD206 (D). (Scale bars: 100 μm.)

Discussion

Our study reveals an essential role for NO in CRPC tumor suppression. We found that increased levels of NO, which are associated with lowering LH and T under physiological conditions, lead to inhibition of prostate tumor growth in a cell nonautonomous manner. In addition, we established that increased levels of NO down-regulate the translational activity of AR, leading to a decreased expression of AR-V7 and pERK levels in tumors. In addition, we showed that NO levels have the potential to influence differentiation of TAM by affecting ligands like IL-34 and the receptors to which they bind (CSF1R). NO levels contribute to a decrease in the M2 macrophages (F4/80, CD206) and the induction of M1 macrophages (iNOS). Finally, we studied whether the short half-life of NO affects its efficacy on tumor macrophages. The results show that tumor burden was significantly suppressed in mice that received cells from mice treated with GSNO. In addition, the M2 macrophage marker was decreased in the tumor grafts that acquired cells from GSNO-treated mice, thus supporting the potential long-term therapeutic implications of increased NO levels on CRPC.

Several studies have demonstrated the tumoricidal effects of NO in other cancer types. Schleiffer et al. (51) showed that NO likely plays a role in neoplastic changes in a rat model of colon cancer. This work revealed that preneoplastic changes were promoted in the colon by decreasing the release of NO through the inhibition of iNOS (51). A separate study (24) demonstrated the role of NO in renal cancer cells and the ability of NO to inhibit tumor growth, as authors induced NO production by transfection of tumor cells with iNOS. Their data indicate that the high levels of NO are associated with inhibition of tumor growth and a reduction in lung metastases (24). The tumoricidal effects of NO are understudied in prostate cancer. Usually, tumor growth in therapeutics of prostate cancer is controlled through steroid management by blocking the AR or by decreasing circulating androgens. However, this method has been limited due to the progression of the disease to a castration-resistant prostate cancer in which steroid manipulation becomes ineffective (52). Previous studies have demonstrated that even modest increases in AR and ARV7 expression may contribute to the development of resistance and progression of disease (14702632) (24909511). Therefore, increased NO capability to decrease AR expression (SI Appendix, Fig. S5) may promote the susceptibility of PCa tumor cells to current therapies. Recently, it has been reported that elevated NO levels are able to inhibit tumor growth in androgen-dependent as well as CRPC cell lines such as LNCAP (30). However, the in vivo regulatory mechanisms behind the impact of increased NO levels on CRPC tumors have largely been unknown. Our study reveals that the TME could be the target of NO for suppression of tumorigenesis in CRPC cell lines like 22RV1.

There are several strengths and limitations in our study. The study reports four major findings: (i) This investigation demonstrates that NO levels increased by GSNO are capable of arresting tumor burden in highly tumorigenic 22RV1 CRPC xenografts; (ii) the work shows that NO deffects on CRPC are cell nonautonomous effects and are targeting the TME; (iii) increased NO affects both activity and differentiation of TAMs in CRPC; and (iv) the study validates the efficacy and durability of NO using a secondary xenograft mouse model. The limitations of our study are as follows: (i) The exact mechanisms by which NO-affected TAM differentiation remain unexplored; (ii) increased NO levels are capable of affecting hypogonadism and lead to CRPC tumor suppression, therefore leading to the strong possibility that the tumors inhibition could be profound in noncastrate mice. However, this assumption still needs to be explored; (iii) the dose-dependent effect of NO on CRPC is not validated; and (iv) we have used 22Rv1 cells as a model cell line to investigate the implications of NO; however, use of cells from human xenograft models of CRPC which were beyond the scope of the present study, but are ongoing in our laboratory, may further confirm and extend the findings discovered in this current study.

In conclusion, this study shows the regulatory role of NO in the TME of CRPC. To better understand the role of NO in the therapeutics of CRPC, a further in-depth evaluation of other NO donors and their effects on the TME of CRPC as well as a long-term follow-up to determine how the loss of function of NO could revert the TAMs and affect CRPC is currently being addressed.

Materials and Methods

Animal experiments were carried out in compliance with the Institutional Animal Care and Use Committee of University of Miami. Molecular analyses were performed using standard procedures. A more detailed description and additional data are provided in SI Appendix, Materials and Methods.