Objective
Obesity is a significant contributing factor to endometrial cancer risk. We previously demonstrated that estrogen-induced endometrial proliferation is enhanced in the context of hyperinsulinemia and insulin resistance. In this study, we investigate whether pharmacologic agents that modulate insulin sensitivity or normalize insulin levels will diminish the proliferative response to estrogen.
Study Design
Zucker fa/fa obese rats and lean controls were used as models of hyperinsulinemia and insulin resistance. Insulin levels were depleted in ovariectomized rats following treatment with streptozotocin, or modulated by metformin treatment. The number of BrdU-incorporated cells, estrogen-dependent proliferative and antiproliferative gene expression, and activation of mTOR and ERK1/2 MAPK signaling were studied. A rat normal endometrial cell line RENE1 was used to evaluate the direct effects of metformin on endometrial cell proliferation and gene expression in vitro.
Results
Streptozotocin lowered circulating insulin levels in obese rats and decreased the number of BrdU-labeled endometrial cells even in the presence of exogenous estrogen. Treatment with the insulin-sensitizing drug metformin attenuated estrogen-dependent proliferative expression of c-myc and c-fos in the obese rat endometrium compared to untreated controls and was accompanied by inhibition of phosphorylation of the insulin and IGF1 receptors (IRβ/IGF1R) and ERK1/2. In vitro studies indicated metformin inhibited RENE1 proliferation in a dose-dependent manner.
Conclusion
These findings suggest that drugs that modulate insulin sensitivity, such as metformin, hinder estrogen-mediated endometrial proliferation. Therefore, these drugs may be clinically useful for the prevention of endometrial cancer in obese women.
Overweight and obesity not only increase the risk of a variety of chronic illnesses, including cardiovascular disease and type 2 diabetes, but also are known risk factors for a variety of cancer types. Among all cancers, increasing body mass index is most strongly associated with endometrial cancer risk, with >50% of all endometrial cancers attributable to obesity. While hyperestrogenism associated with obesity is a significant contributor to the development of endometrial cancer, other factors, including hyperinsulinemia, contribute to its pathogenesis and progression.
We previously evaluated the effect of obesity-associated insulin resistance and hyperinsulinemia on estrogen-associated endometrial proliferation in a rat model. Specifically, we showed that the expression of the proproliferative genes was increased while the expression of antiproliferative genes was inhibited in the endometrium of estrogen-treated obese, insulin-resistant rats as compared to lean controls. These data suggested that insulin potentiates estrogen-regulated endometrial proliferation in the context of obesity.
To address the effects of insulin modulation as a chemopreventive strategy for endometrial cancer, circulating insulin levels and insulin sensitivity were manipulated in obese female Zucker rats using streptozotocin (STZ) and metformin, respectively, in the presence and absence of estrogen. Like obese human beings, the Zucker rat model develops insulin resistance, hyperinsulinemia, and ultimately, noninsulin-dependent diabetes.
STZ, a glucosamine-nitrosourea compound, has been used to treat cancer of the pancreatic islets of Langerhans in human beings. It is extremely toxic to the beta cells of the pancreas, inhibiting insulin production, and therefore has limited clinical utility. However, this drug can be used to permanently reduce circulating insulin levels in laboratory animals. Metformin, a biguanide drug commonly used to treat type 2 diabetes, has recently been demonstrated to exert chemopreventive and antiproliferative effects for a variety of cancers. Metformin inhibits cell growth both by insulin- and noninsulin-dependent mechanisms. Metformin increases insulin receptor sensitivity, increases insulin uptake, thereby reducing systemic insulin levels. Metformin also inhibits cell proliferation by activating the growth-inhibitory 5″ adenosine monophosphate activated protein kinase (AMPK), which counteracts signaling through both the PI3K/AKT and mitogen-activated protein kinase (MAPK) pathways downstream of the insulin and insulin growth factor 1 (IGF1) receptors.
The overall goal of these studies is to provide preclinical data to determine the ability of insulin-sensitizing drugs to attenuate estrogen-induced endometrial proliferation and serve as chemopreventive agents for endometrial cancer in obese individuals.
Materials and Methods
Cell lines
RENE1, a Sprague-Dawley rat normal endometrial cell line, was purchased from Sigma-Aldrich (St. Louis, MO).
Cell proliferation assay
RENE1 cells were treated with metformin or vehicle for 72 hours and cell proliferation was evaluated using 3- (4,5-dimethylthiazol-2-Yl) -2,5-diphenyltetrazolium bromide assay as previously described.
Western blot
The effect of metformin on cell-signaling pathways was evaluated by Western blot analysis. RENE1 cells were plated in 6-well plates at 2 × 10 5 /well. Following 24 hours, cells were treated by metformin (5 mmol/L in culture medium) for 72 hours. Cells were lysed in protein extraction reagent (ThermoScientific, Rockford, IL). Equal amounts of protein for each treatment group were resolved by sodium dodecyl sulfate poly acrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes, and probed for phospho-AMPK (T172), phospho-ERK1/2 (Thr202/Tyr204), phospho-S6 protein (Cell Signaling Technology Inc, Danvers, MA), or β-actin (Sigma-Aldrich) followed by horse radish peroxide-conjugated secondary antibody, per manufacturer’s instructions. When necessary, polyvinylidene fluoride membranes were stripped (62.5 mmol/L Tris-HCl, PH 6.8, 2% SDS, and 100 mmol/L 2-mercaptoethanol) at 50°C for 30 minutes), washed twice in tris buffered saline/tween 20, then reprobed.
Animals care and use
All animal experiments were conducted in compliance with federal guidelines and approved by the institutional animal care and use committee. Mature (5 weeks old) female Zucker fa/fa rats and their lean littermates were purchased from Harlan Laboratories (Indianapolis, IN). After 1 week of acclimation, animals were ovariectomized, and held for 5 days to clear endogenous ovarian hormones.
STZ treatment
Obese and lean animals were randomized to 4 treatment groups (n = 6 per group): vehicle, estradiol (Innovative Research of America, Sarasota, FL), STZ (Sigma-Aldrich), and estradiol plus STZ. Estradiol (10 μg/kg body weight/d, for 2 weeks) was administrated by subcutaneously implanting the timed-releasing estradiol pellets 2 weeks before the end of the experiment, while in control group animals were implanted with the placebo pellets. Seven days after estradiol administration, animals were injected intraperitoneally with either vehicle (citrate buffer) or STZ (45 mg/kg). Seven days after STZ administration, animals were sacrificed for tissue collection.
Metformin treatment
Obese and lean rats were randomized to 3 treatment groups (17-26 animals per group): vehicle alone, estradiol, and estradiol plus metformin. Metformin (300 mg/kg body weight/d in 1% methyl-cellulose solution) was administrated by daily oral gavage for 3 weeks. Control animals received vehicle alone. Estradiol (40 μg/kg body weight/d, for 3 days) was administrated intraperitoneally for the last 3 days of the experiment. Control animals received saline alone. Animals were sacrificed and uteri were collected for histochemical evaluation and RNA isolation.
Plasma glucose level and insulin level detection
Three to 5 rats from each treatment group were fasted overnight, and were subjected to an oral glucose tolerance test. Plasma glucose concentrations were tested with a blood glucose monitoring system (Ascensia Contour; Bayer Health Care, New York, NY). Insulin levels were by enzyme-linked immunosorbent assay (insulin ultrasensitive EIA kit; ALPCO Diagnostics, Salem, NH).
Immunohistochemistry
All rats were injected intraperitoneally with 5-bromo-2′-deoxyuridine (BrdU) at a dose of 100 mg/kg body weight 90 minutes before sacrifice. Fresh uterine tissues were collected and fixed in 10% neutral-buffered formalin, and processed for paraffin embedding. BrdU immunostaining was performed using BrdU in situ detection kit (BD Biosciences, San Diego, CA). The slides were counterstained with Mayer hematoxylin for 1 minute. The total number of BrdU-stained nuclei per 200 endometrial cells was counted in 10 randomly selected fields (200×).
Immunohistochemical analysis of rat uterine tissue was performed using Ki67 (BD Biosciences), phospho-IGF1R (Tyr1131)/insulin receptor β (Tyr1146), phospho-S6 ribosomal protein (Ser235/236), phospho-ERK1/2 (Thr202/Tyr204), phospho-acetyl-CoA carboxylase (Ser79) (ACC), and cleaved caspase-3 (Asp175) (Cell Signaling Technology Inc) per manufacturers’ instructions. The sections were counterstained with Mayer hematoxylin. The average number of positively Ki67 or caspase-3 stained cells in 5-10 high-power microscopic fields were counted per slide, and calculated as: 200 × (numbers of stained endometrial cell/total endometrial cells). For all other markers, staining was scored based on intensity as negative or weak (0 or 1+) vs positive or strong (2+ or 3+).
RNA isolation and real-time quantitative reverse transcription-polymerase chain reaction
Total RNA was extracted from frozen endometrial tissue using Tri-reagent (as described previously). For each transcript, specific polymerase chain reaction (PCR) primer pairs and a dual fluorochrome-labeled hybridization probe (hydrolysis probe) were designed using Primer Express (Applied Biosystems, Carlsbad, CA) or Beacon Designer (Premier Biosoft International, Palo Alto, CA) (Appendix; Supplemental Table ). All real-time quantitative reverse transcription PCR reactions were set up using liquid-handling robotics. Samples, controls, and 5-log standard curves were run on 384-well plates using an Applied Biosystems 7900 quantitative PCR instrument under the following conditions: 95°C for 2 minutes followed by 40 cycles of 95°C for 12 seconds and 60°C for 30 seconds. Data were analyzed using SDS version 2.4 software (Life Technologies, Grand Island, NY) post-run using auto baseline and manual threshold settings and was normalized to 18SrRNA levels.
Statistical analysis
Statistical analyses were performed using statistical software (SAS, version 9.1; SAS Institute Inc, Cary, NC; and STATA/SE, version 10.1; Santa Corp LP, College Station, TX). For BrdU, Ki67, caspase-3, and real-time quantitative reverse transcription PCR test, 1-way analysis of variance was used to compare treatment groups. Tests were made using log-transformed measurements. For other immunohistochemical tests, Fisher exact tests were used in place of logistic regression models. A significance level of .05 was used to judge statistical significance.
Results
Direct effects of metformin on endometrial cell growth in vitro
We examined the direct effects of metformin on endometrial cell proliferation and gene expression in vitro, using the normal rat endometrial cell line, RENE1. This in vitro evaluation also permitted the direct analysis of several concentrations of metformin on endometrial cell proliferation by 3- (4,5-dimethylthiazol-2-Yl) -2,5-diphenyltetrazolium bromide. RENE1 proliferation was inhibited in a dose-dependent manner after 3 days of metformin ( P < .001) ( Figure 1 , A).
The effect of metformin on growth-promoting and inhibitory pathways were evaluated by Western blot using activation-specific antibodies ( Figure 1 , B). Metformin inhibited phosphorylation of pERK1/2 and S6R protein, while promoting AMPK phosphorylation. Overall, these studies suggest that metformin can inhibit endometrial proliferation, in part as a consequence of its ability to directly modulate proproliferative and antiproliferative pathways.
Proliferative effect of estrogen under low insulin conditions
We confirmed the effect of STZ in lowering serum insulin levels using an oral glucose tolerance test ( Supplemental Figure 1 , A). Low-dose β-toxin STZ treatment decreased obese rat serum insulin level ( P = .0107 vs obese control) at all time points after glucose challenge, but showed no effect in lean rats ( P = .9519). STZ administration significantly increased serum glucose level in both lean ( P < .0001) and obese ( P < .0001) rats.
BrdU incorporation and Ki67 immunohistochemical staining confirmed the proliferative effects of estrogen under low insulin conditions ( Figure 2 ). Estradiol treatment increased BrdU incorporation in both lean (48.8 ± 23.8 vs 0.3 ± 0.5) and obese (111.1 ± 37.7 vs 1.7 ± 1.2) endometrium. The number of estrogen-induced, BrdU-labeled endometrial cells was 2.3-fold higher in obese animals as compared to that observed in lean rats (111.1 ± 37.7 vs 48.8 ± 23.8; P < .001). STZ treatment decreased BrdU incorporation in both estrogen-treated lean (34.1 ± 23.2 vs 48.8 ± 23.8) and obese (14.0 ± 10.1 vs 111.1 ± 37.7) rat endometrium. In obese rat endometrium, the proliferative effect of estrogen was antagonized by STZ treatment. BrdU incorporation was significantly decreased in obese rats treated with estradiol plus STZ when compared with rats treated with estrogen alone ( P < .0001). Ki67 staining validates these findings (data not shown), and supports the observation that a reduction in circulating insulin blunts the effects of proliferative effects of estrogen in the endometrium.
Effect of metformin therapy on rat endometrial proliferation
Metformin decreased serum glucose levels. At 45 minutes following a glucose challenge, glucose and insulin levels were significantly higher in obese rats compared with lean rats ( P = .0176). Treatment with metformin decreased serum glucose in obese rats as compared with the nontreated group ( Supplemental Figure 2 ), however metformin did not significantly decrease circulating insulin levels in this obese animal model during the 3-week treatment period. This is perhaps not surprising, as metformin has been shown to decrease gluconeogenesis in the liver, with no demonstrated impact on insulin synthesis by the pancreas. Instead, metformin has been shown to increase insulin sensitivity and uptake, which contributes to a modest decrease in circulating insulin levels after prolonged use. Indeed, a reduction in circulating insulin was observed in mice fed a high-fat diet, following 8-10 weeks of metformin therapy. Levels observed in metformin-treated vs untreated animals mice approached but did not reach statistical significance, as reflected by C-peptide levels, a surrogate marker for insulin.
We examined the effect of metformin on the expression of genes associated with estrogen-mediated endometrial proliferation. In the normal physiologic state, estrogen induces both growth-stimulatory (c-myc, c-fos) and growth-inhibitory (RALDH2 and secreted frizzled-related protein 4 [SFRP4]) pathways. The result is controlled, balanced endometrial growth. We have already shown that estradiol treatment augments transcription of the proproliferative gene c-myc in the obese rat endometrium as compared to the lean rat endometrium. Conversely, the growth-inhibitory genes, RALDH2 and SFRP4, whose transcription is induced by estrogen in the endometrium of lean rats, are attenuated in obese rats. In this study, we further demonstrate the induction of c-fos transcription in estrogenized obese rat endometrium compared to lean controls (0.04 ± 0.017 vs 0.025 ± 0.010; P < .025) ( Figure 3 , A). We anticipate these transcriptional changes reflect the changes in insulin and IGF1 levels associated with obesity.
To address the effect of metformin on proliferation via estrogen-induced gene expression, we compared the messenger RNA level of c-myc, c-fos, SFRP4, and RALDH2 transcripts in metformin- and vehicle-treated rat endometrium. Metformin treatment significantly decreased transcript levels for both c-myc (0.011 ± 0.003 vs 0.029 ± 0.014; P < .001) and c-fos (0.024 ± 0.016 vs 0.040 ± 0.017; P < .001) in the estrogenized obese rat endometrium, as compared to untreated obese animals. No significant effect was observed in lean rat endometrium ( Figure 3 , A). Interestingly, expression of the antiproliferative RALDH2 and SFRP4 genes in estrogenized obese rat endometrium were not significantly affected by metformin ( Figure 3 , A). Overall, these data suggest that metformin treatment attenuates the transcription of a subset of estrogen-induced proproliferative genes, but does not significantly promote the expression of estrogen-induced, growth-inhibitory genes in the endometrium of obese rats.
The effect of metformin on endometrial cell proliferation was evaluated by both BrdU and Ki67 staining. Three days of treatment with estradiol vs control treatment induced endometrial proliferation in both lean (13.48 ± 10.5 vs 0.1 ± 0.4) and obese (22.3 ± 17.2 vs 1.6 ± 2.1) rats ( Figure 3 , B). Significant endometrial proliferation was observed in obese animals as compared to lean animals, in response to estrogen (22.3 ± 17.2 vs 13.4 ± 10.5; P = .056). Metformin therapy did not significantly alter estrogen-mediated endometrial proliferation when compared to controls in both lean (11.3 ± 6.9 vs 13.4 ± 10.5) and obese (17.6 ± 4.7 vs 22.3 ± 17.2; data not shown) rats.
While metformin inhibits the transcription of growth-promoting genes c-myc and c-fos in the endometrium of obese, estrogen-treated rats, the levels of the growth-inhibitory genes were seemingly unaffected within the time frame of this experiment. Furthermore, given the lack of short-term effects resulting from a 3-week course of metformin on circulating insulin levels, we hypothesize that the overall effect on endometrial proliferation as measured by Ki67 and BrdU incorporation is not yet fully apparent. As reflected by the trend of reduced BrdU incorporation in obese, estrogen-treated rats following treatment with metformin ( P = .056), we expect the antiproliferative effects of metformin on endometrial tissue may become more pronounced over time.
Effect of metformin on endometrial cell apoptosis
To address the possibility that metformin may induce apoptosis, rather than inhibit proliferation in the obese rat endometrium, we tested endometrial cell apoptosis by caspase-3 staining. Metformin treatment did not produce a significant increase in caspase-3 staining in obese rat endometrium when compared with untreated obese rat endometrium ( Supplemental Figure 3 ).
Effect of metformin on insulin/IGF signaling
Hyperinsulinemia in the obese rat can contribute to elevated IGFI levels and activation of the IGF-insulin receptor (IR). The effect of metformin on IGFI and insulin signaling in rat endometrial tissue was determined by immunohistochemical staining for phospho-IGF1 receptor (Tyr-1131)/IR (Tyr-1146). These sites represent one of the early sites of IGF1R and IR autophosphorylation, which is required for full receptor tyrosine kinase activation.
Metformin treatment significantly inhibited IGF1R/IRβ activation in obese rat endometrium. Phospho-IGF1R/IRβ staining was significantly weaker in obese rat treated with metformin as compared to those treated with estrogen alone (31% vs 92%, 4/13 vs 12/13 positive samples; P < .025) ( Figure 4 , A). These findings suggest that metformin may regulate IGF1R/IR activity by modulating receptor autophosphorylation.
