Pioglitazone Inhibits Periprostatic White Adipose Tissue Inflammation in Obese Mice

Obesity is both a risk factor for the development of high-grade prostate cancer and a poor prognostic factor for prostate cancer survivors (1–3). Obesity-related effects on levels of hormones, adipokines, and proinflammatory mediators have been suggested to contribute to the pathogenesis of prostate cancer (4–6). Chronic low-grade inflammation is commonly found in the visceral and subcutaneous white adipose tissue (WAT) of obese humans and mice (7–9). In both mouse models of obesity and obese humans, macrophages infiltrate WAT and surround dead or dying adipocytes forming crown-like structures (CLS; ref. 10). The prostate is contained within a thin connective tissue capsule and is surrounded by adipose tissue. Recently, we showed that WAT inflammation occurred in the periprostatic fat of approximately 50% of men undergoing radical prostatectomy as a treatment for prostate cancer (11). Importantly, periprostatic WAT inflammation, as defined by the presence of CLS, was associated with high-grade prostate cancer, suggesting a potential contributory role (11). Periprostatic WAT inflammation was also associated with systemic effects that have been linked to the pathogenesis of prostate cancer, including reduced levels of adiponectin, an anti-inflammatory adipokine, and insulin resistance (11, 12). It is possible, therefore, that interventions that attenuate periprostatic WAT inflammation will either reduce the risk of prostate cancer or improve the outcomes of prostate cancer patients.

PPARγ, a member of the nuclear receptor family of transcription factors, plays a significant role in regulating lipid and glucose metabolism (13). Ligand-activated PPARγ induces adipocyte differentiation, stimulates mitochondrial biogenesis, and inhibits the production of proinflammatory mediators (14). Pioglitazone, a PPARγ agonist that belongs to a class of drugs called thiazolidinediones, is used to treat type II diabetes. Pioglitazone improves insulin resistance and has a range of anti-inflammatory properties, including being able to induce adiponectin levels and reduce the number of macrophages in adipose tissue (15, 16). Whether pioglitazone can suppress obesity-related periprostatic WAT inflammation is unknown.

In this study, we had two major objectives. The first was to determine whether pioglitazone could suppress periprostatic WAT inflammation in a mouse model of obesity. Our second goal was to elucidate the mechanism underlying the anti-inflammatory effect of pioglitazone. Here, we demonstrate that treatment with pioglitazone attenuated obesity-related periprostatic WAT inflammation. Moreover, we provide evidence that pioglitazone induced adiponectin, which acted, in turn, to suppress the induction of monocyte chemoattractant-1 (MCP-1) in adipose tissue leading to decreased periprostatic WAT inflammation. Taken together, we show for the first time that it is possible to use a pharmacologic strategy to suppress periprostatic WAT inflammation. Our data highlight the importance of targeting the PPARγ–adiponectin–MCP-1 axis as a strategy to reduce periprostatic WAT inflammation and potentially reduce the risk of prostate cancer or improve outcomes for men with prostate cancer.

Pioglitazone was obtained from the NCI Chemical Repository. Mouse recombinant TNFα and ELISA Kits for MCP-1 and adiponectin were obtained from R&D Systems. RNeasy Mini Kits, real-time PCR mouse primers for CD68 (cat. # QT00254051), MCP-1 (cat. # QT00167832), TGFβ (cat. # QT00145250), GAPDH (cat. # QT01658692), adiponectin (cat. # QT01048047), AdipoR1 (cat. # QT00154217), AdipoR2 (cat. # QT00165326), and siRNA for adiponectin were purchased from Qiagen. TNFα primers were obtained from Sigma (17). Nontargeting siRNA, AdipoR2 siRNA, and DharmaFECT4 were obtained from Dharmacon. Murine leukemia virus reverse transcriptase, RNase inhibitor, oligo (dT)₁₆, and Fast SYBR green PCR master mix were obtained from Applied Biosystems. Insulin, IBMX, and dexamethasone were obtained from Sigma.

Male wild-type C57BL/6J and MCP-1 knockout mice (B6.129S4-Ccl2^tm1Rol/J, The Jackson Laboratory) were used. Mice received either low-fat (LF) diet (12450Bi; 10 kcal% fat), high-fat (HF) diet (D12492i; 60 kcal% fat), or HF diet containing 0.006% or 0.06% w/w pioglitazone (Research Diets). Previously, 0.05% w/w pioglitazone was found to activate PPARγ in adipose tissue (18). Following sacrifice, periprostatic fat tissue was snap-frozen in liquid nitrogen and stored at −80°C for molecular analyses or formalin fixed for histologic analyses. The animal protocol was approved by the Institutional Animal Care and Use Committee at Weill Cornell Medicine (New York, NY).

Adipocyte diameter was quantified as described previously (19).

3T3-L1 preadipocytes (Zenbio) were seeded 5,000 cells/cm² in cell culture plates and maintained in preadipocyte medium (Zenbio). All experiments were carried out within one year (2015–2016) of purchasing the cell line. Independent authentication of the purchased cell line was not carried out. Once the cells were confluent, cells were incubated an additional 48 hours before initiating differentiation. The cells were then switched to differentiation medium containing 10% FBS, 50 U/mL penicillin, 50 μg/mL streptomycin, 500 μmol/L IBMX, 50 μmol/L dexamethasone, and 2.5 μg/mL insulin for 3 days. Three days after the initiation of differentiation, cells were switched to maintenance medium containing 10% FBS, 50 U/mL penicillin, 50 μg/mL streptomycin, 50 μmol/L dexamethasone, and 2.5 μg/mL insulin and maintained until 7 days post differentiation. All treatments were performed on cells that were maintained in serum-free medium overnight.

siRNAs were transfected into 3T3-L1 adipocytes using DharmaFECT4 (20). Cells transfected with nontargeting siRNA were used as negative controls. Seventy-two hours after transfection, the cells were analyzed.

Total RNA was isolated from periprostatic fat and cells using the RNeasy Mini Kit. RNA was reversed transcribed using murine leukemia virus reverse transcriptase and oligo (dT)₁₆ primer, and the resulting cDNA was then used for amplification. GAPDH was used as an endogenous normalization control. Real-time PCR was performed using 2× Fast SYBR green PCR master mix on a 7500 Fast Real-time PCR system (Applied Biosystems). Relative fold induction was determined using the ddC_T (relative quantification) analysis protocol.

MCP-1 and adiponectin protein levels in cell culture medium were quantified using ELISA Kits from R&D Systems.

In the mouse experiments, endpoints of interest include mouse weights over time or at the time of sacrifice, mouse caloric consumption over time, and various biomarker measurements obtained from mouse periprostatic WAT, including number of CLS/cm², average adipocyte diameter, and relative mRNA expression level. Differences in the endpoints across multiple groups were examined using ANOVA or the nonparametric Kruskal–Wallis test, where appropriate. Differences in the endpoints between two groups were examined using t test or the nonparametric Wilcoxon rank sum test, where appropriate. For data measured longitudinally over time, such as mouse weight and caloric consumption, differences in the average value of the endpoint over time were examined using the mixed effects linear regression model, which takes into account both the between and within mouse variations. All statistical tests were two-sided and P < 0.05 is considered statistically significant.

Mice were fed either a HF or LF diet to determine whether obesity caused periprostatic WAT inflammation as defined by the presence of CLS (Fig. 1A and B). Body weight, calorie consumption, and periprostatic WAT inflammation were determined (Fig. 1C–E). After feeding these diets for 12 weeks, the HF diet–fed mice weighed approximately 40% more than the mice that received the LF diet. This increase in weight was paralleled by the development of periprostatic WAT inflammation (Fig. 1C and E). To determine whether pioglitazone inhibited obesity-associated periprostatic WAT inflammation, obese HF diet–fed mice were either continued on HF diet alone or treated with HF diet containing two doses (0.006%, 0.06% w/w) of pioglitazone (Fig. 1A). Interestingly, treatment with either dose of pioglitazone led to a marked increase in calorie consumption without causing significant weight gain (Fig. 1C and D). Notably, treatment with pioglitazone led to a dose-dependent decrease in periprostatic WAT inflammation (Fig. 1E).

Numerous studies have demonstrated a correlation between adipocyte size and the presence and severity of WAT inflammation (10, 19, 21). Consistent with the increase in body weight, adipocytes were larger in the periprostatic WAT of HF diet compared with LF diet–fed mice (Fig. 2A). Although pioglitazone reduced periprostatic WAT inflammation, it did not affect adipocyte size (Fig. 2A). Next, we characterized a series of molecular endpoints in periprostatic WAT. Consistent with the findings for CLS in periprostatic WAT (Fig. 1E), the increase in expression of the macrophage marker CD68, induced by HF diet feeding, was attenuated by treatment with 0.06% (w/w) pioglitazone (Fig. 2B). Next, we measured levels of MCP-1, a chemokine that is important for recruitment of blood monocytes into adipose tissue where they undergo differentiation and become macrophages (22, 23). As shown in Fig. 2C, HF diet led to an increase in levels of MCP-1, an effect that was suppressed by 0.06% (w/w) pioglitazone. Similarly, periprostatic WAT inflammation was associated with increased levels of TNFα and TGFβ; these changes were attenuated by treatment with 0.06% (w/w) pioglitazone (Fig. 2D and E).

Having established that pioglitazone suppressed HF diet–induced periprostatic WAT inflammation, we next carried out a study to determine the time course of this anti-inflammatory effect. The experimental schema is shown in Fig. 3A. Body weight, calorie consumption, and periprostatic WAT inflammation were assessed over time (Fig. 3B–D). Consistent with the findings in Fig. 1, treatment with HF diet for 12 weeks led to weight gain and periprostatic WAT inflammation (Fig. 3B and D). Mice were then either continued on HF or LF diets or treated with HF diet containing 0.06% (w/w) pioglitazone for 2, 4, or 9 weeks. Here, treatment with pioglitazone for 4 or 9 weeks caused hyperphagia (Fig. 3C) in the absence of inducing weight gain (Fig. 3B). Treatment with pioglitazone led to reduced number of CLS/cm² in WAT after 4 and 9 weeks of treatment (Fig. 3D).

As mentioned above, MCP-1 plays a significant role in regulating the accumulation of macrophages in adipose tissue. The inhibitory effects of pioglitazone on HF diet–mediated induction of periprostatic WAT inflammation (Fig. 1E) were paralleled by changes in MCP-1 expression (Fig. 2C). Hence, we tested whether the observed suppression of MCP-1 expression was required for the anti-inflammatory effects of pioglitazone. To address this question, we compared the anti-inflammatory efficacy of pioglitazone in wild-type versus MCP-1^−/− mice (Fig. 4A). Body weight and periprostatic WAT inflammation were determined (Fig. 4B and C). Treatment with pioglitazone did not affect the weights of either the wild-type or the MCP-1^−/− mice (Fig. 4B). Consistent with findings in Figs. 1E and 3D, pioglitazone markedly suppressed periprostatic WAT inflammation in wild-type mice (Fig. 4C). Importantly, this anti-inflammatory effect of pioglitazone was abrogated in MCP-1^−/− mice (Fig. 4C). Taken together, these data indicate that the anti-inflammatory effect of pioglitazone is mediated by its ability to suppress HF diet–mediated induction of MCP-1.

Pioglitazone is a known inducer of adiponectin in adipose tissue (24, 25). Moreover, adiponectin can inhibit MCP-1 expression (26). Hence, experiments were next carried out to determine whether the suppressive effects of pioglitazone on MCP-1 expression were mediated by adiponectin. As shown in Fig. 5A, treatment with pioglitazone led to a dose-dependent increase in adiponectin levels in mice fed an HF diet. To determine whether pioglitazone-mediated increases in adiponectin suppressed the induction of MCP-1, a series of in vitro experiments were carried out using 3T3-L1 adipocytes. As shown in Fig. 5B, pioglitazone caused dose-dependent induction of adiponectin. Levels of TNFα, a known inducer of MCP-1 (27), are elevated in inflamed periprostatic WAT (Fig. 2D). Interestingly, pioglitazone caused dose-dependent suppression of TNFα-mediated induction of MCP-1 (Fig. 5C). Moreover, treatment of adipocytes with TNFα inhibited adiponectin mRNA levels, an effect that was attenuated by pioglitazone (Fig. 5D). To confirm that pioglitazone-mediated suppression of MCP-1 induction is mediated by adiponectin, siRNA to adiponectin was used. Adipocytes transfected with siRNA to adiponectin produced approximately 70% less adiponectin protein (Fig. 5E). Importantly, silencing adiponectin both enhanced TNFα-mediated induction of MCP-1 and attenuated the ability of pioglitazone to suppress the induction of MCP-1 mRNA (Fig. 5F). Collectively, these results indicate that pioglitazone-mediated increases in adiponectin suppress the induction of MCP-1.

Adiponectin binds to adiponectin receptor R1 (AdipoR1) and R2 (AdipoR2) and regulates gene expression in adipocytes (28). We next investigated whether pioglitazone modulated the expression of the adiponectin receptors. Feeding a HF diet for 20 weeks was associated with decreased expression of AdipoR2 in periprostatic WAT (Fig. 6A). This effect was reversed by treatment with pioglitazone, which led to dose-dependent induction of AdipoR2. In contrast, levels of AdipoR1 in periprostatic WAT were unaffected by either the HF diet or pioglitazone. To extend upon these findings, additional experiments were carried out in 3T3-L1 cells. Consistent with the findings in periprostatic WAT (Fig. 6A), treatment of 3T3-L1 cells with pioglitazone caused dose-dependent induction of AdipoR2 levels (Fig. 6B). Similar to the effects on adiponectin, TNFα-mediated downregulation of AdipoR2 was attenuated by treatment with pioglitazone (Fig. 6C). To confirm that pioglitazone-mediated suppression of MCP-1 induction is mediated by AdipoR2, siRNA was used. Levels of AdipoR2 in 3T3-L1 cells were suppressed more than 80% by the siRNA to AdipoR2 (Fig. 6D). Notably, silencing AdipoR2 both enhanced TNFα-mediated induction of MCP-1 and attenuated the ability of pioglitazone to suppress the induction of MCP-1 (Fig. 6E). Taken together, our study demonstrates that treatment with pioglitazone suppresses obesity-related periprostatic WAT inflammation by inducing adiponectin, which acts, in turn, to block the induction of MCP-1 resulting in reduced macrophage accumulation (Fig. 6F).

WAT inflammation is associated with the development, grade, or progression of several solid tumors, including prostate cancer (11, 19, 29–31). Here, we demonstrate that diet-induced obesity led to periprostatic WAT inflammation in mice consistent with recent findings in men (11). Importantly, pioglitazone, an antidiabetic drug, protected against HF diet–induced periprostatic WAT inflammation. Similar anti-inflammatory effects of thiazolidinediones, including pioglitazone, have been demonstrated in other adipose depots in both mice and humans (16, 32–34).

We carried out a series of studies to explore the mechanism underlying the anti-inflammatory effects of pioglitazone. Treatment with pioglitazone led to a dose-dependent decrease in the expression of MCP-1 in the periprostatic fat of obese mice (Fig. 2C). Because MCP-1 plays a key role in the recruitment of blood monocytes into adipose tissue where they differentiate and become macrophages (22, 23), it seemed likely that the observed decrease in MCP-1 levels could account for the reduction in periprostatic adipose inflammation. This possibility was supported by our observation that pioglitazone failed to suppress periprostatic adipose inflammation in obese MCP-1 knockout mice (Fig. 4C). The fact that the ability of pioglitazone to suppress periprostatic adipose inflammation was dependent on its effect on MCP-1 is consistent with prior reports that thiazolidinediones can suppress levels of MCP-1 (16, 34). In men, periprostatic adipose inflammation is associated with reduced circulating levels of adiponectin, an anti-inflammatory adipokine (11). Pioglitazone is known to induce adiponectin in both humans and mice (24, 35–38). Adiponectin has been reported to suppress MCP-1 levels (26). Hence, we next determined whether pioglitazone induced adiponectin or its receptor AdipoR2, leading to reduced levels of MCP-1. Pioglitazone caused dose-dependent induction of adiponectin and AdipoR2 in both periprostatic fat and in cultured adipocytes (Figs. 5A and B and 6A and B). Moreover, pioglitazone blocked TNFα-mediated induction of MCP-1, an effect that was attenuated when either adiponectin or AdipoR2 was silenced (Figs. 5F and 6E). Collectively, these results indicate that treatment with pioglitazone reduces obesity-related periprostatic WAT inflammation by inducing adiponectin, which acted, in turn, to suppress MCP-1 resulting in less macrophage accumulation. In preclinical studies, agonists of the adiponectin receptors can suppress inflammation (39). Whether such agents attenuate obesity-related periprostatic adipose inflammation or inhibit the development or progression of prostate cancer warrants further investigation.

In previous studies, adipocyte hypertrophy has been associated with WAT inflammation (10, 19, 21). It has been presumed that hypoxia, endoplasmic stress, or both stimulate the inflammatory response (40, 41). Consistent with findings in human periprostatic fat (11), we found that periprostatic adipocyte hypertrophy was associated with WAT inflammation in obese mice. Behavioral or pharmacologic interventions that result in reduced caloric intake lead to decreased adipocyte size and improved WAT inflammation (17, 42). In this study, pioglitazone stimulated caloric intake but did not induce weight gain, presumably because this agent induces mitochondrial biogenesis and thereby energy expenditure (43). Notably, treatment with pioglitazone reduced periprostatic WAT inflammation in the absence of a reduction in body weight or periprostatic adipocyte size. Recently, pioglitazone was reported to increase capillary density in association with decreasing adipose tissue inflammation, which would be anticipated to increase local oxygen tension (44). If similar effects occur in the periprostatic fat, this could explain the observed decrease in periprostatic WAT inflammation in the absence of a reduction in adipocyte size.

Inflamed periprostatic fat is associated with both high-grade prostate cancer and insulin resistance in prostate cancer patients (11). Elevated insulin levels have been suggested to play a role in prostate cancer growth and mortality (12, 45, 46). Our preclinical study indicates that pioglitazone, an insulin sensitizer, causes a reduction in periprostatic inflammation after relatively short-term treatment. Pioglitazone can also suppress levels of the hormone resistin, which can stimulate the proliferation of prostate cancer cells (47, 48). A large meta-analysis comparing thiazolidinedione use and cancer incidence suggested that pioglitazone may reduce the risk of prostate cancer (49). However, another recent study suggested that use of pioglitazone was associated with a small increase in the risk of prostate cancer (50). Hence, it is unclear from the epidemiologic evidence whether pioglitazone will be beneficial to either prostate cancer patients or men at high risk for prostate cancer. These studies failed to select for men with periprostatic adipose inflammation, the group that may be most likely to benefit from this type of agent. Given this uncertainty, a window-of-opportunity clinical trial should be considered where pioglitazone would be given to men in the preoperative period, allowing biological endpoints to be evaluated in both the prostate and periprostatic fat obtained at the time of surgery. Taken together, the current results raise the possibility that targeting the PPARγ–adiponectin–MCP-1 axis will decrease periprostatic adipose inflammation and thereby reduce the risk of high-grade prostate cancer or improve the outcomes of men with prostate cancer.

No potential conflicts of interest were disclosed.

Conception and design: K. Subbaramaiah, A.J. Dannenberg

Development of methodology: M. Miyazawa, D.D. Giri, A.J. Dannenberg

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Miyazawa, K. Subbaramaiah, P. Bhardwaj, D.J. Falcone, D.D. Giri, A.J. Dannenberg

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Miyazawa, X.K. Zhou, H. Wang, D.D. Giri

Writing, review, and/or revision of the manuscript: M. Miyazawa, K. Subbaramaiah, P. Bhardwaj, X.K. Zhou, D.D. Giri, A.J. Dannenberg

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P. Bhardwaj, X.K. Zhou, A.J. Dannenberg

Study supervision: K. Subbaramaiah, A.J. Dannenberg

This work was supported by the Prostate Cancer Foundation (Movember-Challenge Award, A.J. Dannenberg), Tokyo Association for Clinical Surgery (M. Miyazawa), Patricia and William Kleh (A.J. Dannenberg), and NIH/NCIP50 CA092629 (A.J. Dannenberg).

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.