Obesity
Original Article OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
Fenretinide Prevents Obesity in Aged Female Mice in Association with Increased Retinoid and Estrogen Signaling Kirsty D. Shearer1, Nicola Morrice1,2, Claire Henderson1, Jenny Reekie1, George D. Mcilroy1, Peter J. McCaffery1, Mirela Delibegovic1, and Nimesh Mody1
Objective: The synthetic retinoid fenretinide (FEN) inhibits adiposity in male mice fed a high-fat diet (HFD) in association with alterations in retinoic acid (RA) signaling. Young female mice are protected from obesity via estrogen signaling. We, therefore, investigated whether FEN also influences adiposity in aged female mice differing in parity and whether such effects are mediated by retinoid and estrogen signaling. Methods: Aged nulliparous and parous female mice were maintained on HFD 6 FEN, and adiposity was assessed. Quantitative polymerase chain reaction was performed on white adipose tissue (WAT), liver, and 3T3-L1 adipocytes treated with RA or FEN 6 estrogen. Results: Parous females were more obese than nulliparous mice independent of age. FEN-HFD prevented the HFD-induced increase in adiposity and leptin levels independently of parity. FEN-HFD induced retinoid-responsive genes in WAT and liver. Parous females had reduced expression of hepatic estrogenresponsive genes, but FEN-HFD up-regulated WAT Cyp19a1 and Esr2 in parous mice. Estrogen and RA acted synergistically to increase RA receptor-mediated gene expression in 3T3-L1 adipocytes. FEN increased Cyp19a1 and Esr2, similar to our findings in vivo. Conclusions: The prevention of adiposity by FEN in response to HFD in female mice seems to involve increased retinoid signaling in association with induction of local estrogen production and estrogen signaling in WAT. Obesity (2015) 23, 1655–1662. doi:10.1002/oby.21164
Introduction Vitamin A metabolism leading to retinoic acid (RA) signaling is a nuclear receptor pathway that plays an important role in regulating body fat and glucose homeostasis. Several reports have associated increased vitamin A levels with diabetes (1-3). However, not all studies have confirmed these links, and thus the mechanism(s) by which retinoid metabolism may regulate glucose homeostasis remains unclear. For example, serum retinol-binding protein (RBP)-4 levels are elevated in many, but not all, mouse models and humans with obesity-associated insulin resistance (4). We have recently shown that in male high-fat diet (HFD)-induced obese mice, the synthetic retinoid fenretinide (FEN) can inhibit obesity and normalize eleva-
tions in serum RBP4, insulin, and glucose levels and hepatic insulin resistance (5,6). In 3T3-L1 adipocytes and in vivo, we found that FEN-mediated beneficial effects are associated with alterations in retinoid homeostasis genes (5). In HFD male mice, RA-responsive genes were markedly increased in white adipose tissue (WAT) and liver with short- and long-term FEN (5). In WAT, FEN inhibited the down-regulation of peroxisome proliferator-activated receptor-c (Pparg) and improved insulin sensitivity and levels of adiponectin and RBP4. FEN also inhibited hyperleptinemia in vivo and leptin expression in adipocytes (5). FEN is the most widely studied retinoid in breast cancer chemoprevention trials, because of its selective accumulation in breast tissue
1 Institute of Medical Sciences, College of Life Sciences and Medicine, University of Aberdeen, Foresterhill Health Campus, Aberdeen, Scotland, UK. Correspondence: Nimesh Mody ([email protected]) 2 Centre for Genome-Enabled Biology and Medicine, University of Aberdeen, Aberdeen, Scotland, UK.
Funding agencies: This study was supported by British Heart Foundation Intermediate Basic Research Fellowship grant FS/09/026 (to N. Mody), Research Councils UK Fellowship (to M.D.), European Foundation for the Study of Diabetes (EFSD)/Lilly Programme grant (to M.D. and N. Mody), Biotechnology and Biological Sciences Research Council (BBSRC) studentship (to G.D.M.), University of Aberdeen (UoA) Centre for Genome-Enabled Biology & Medicine (CGEBM) studentship (to N. Morrice), and BBSRC project grant BB/G014272/1 (to P.J.M.). Disclosure: The authors declared no conflict of interest. Author contributions: N. Mody and M.D. conceived the study. All authors participated in performing some of the experiments. K.D.S., N. Morrice, C.H., J.R., G.D.M., and N. Mody analyzed the data. K.D.S. and N. Mody made the figures and wrote the manuscript. P.J.M. and M.D. were involved in editing and proofreading the manuscript. Additional Supporting Information may be found in the online version of this article. Received: 12 November 2014; Accepted: 20 April 2015; Published online 14 July 2015. doi:10.1002/oby.21164
www.obesityjournal.org
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1655
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
and its favorable toxicological profile (7,8). However, the effect of FEN on female obesity has not previously been reported, because all earlier rodent studies were conducted in male models. Young female mice (and humans) are protected from obesity in an estrogen-dependent manner (9-12). Thereby, decreased estrogen levels or global deletion of the estrogen receptor, ERa (Esr1), in mice are both associated with increased adiposity and insulin resistance (13-15). Childbearing and increasing parity are associated with future development of maternal obesity and type 2 diabetes (16,17). Estrogen levels have also been found to be lower in parous women (2 years after delivery/lactation) compared with those measured in nulliparous women (having never given birth) (18). However, very few obesity studies have been performed in parous mice (19). We hypothesized that an interaction between estrogen and RA signaling may play an important role in regulating adiposity and glucose homeostasis. Thus, in this study, we investigated whether FEN could prevent HFD-induced obesity in aged female mice, nulliparous and parous (after pregnancy). In addition, we have examined the molecular mechanisms involved particularly focusing on increased gene expression of metabolically important genes and those involved in estrogen and RA signaling in mouse WAT and liver, and in 3T3-L1 murine adipocytes.
TABLE 1 Parous mice are more obese than nulliparous offspring, independent of a 13-week age difference
Parity status
Age at start of diet (years) Age at end of diet (years) Weight at start of diet (g) Weight at 8 weeks of diet (g) Weight at 21 weeks of diet (g) Weight at end of diet (g) DXA body composition at 8 weeks Lean mass (g) Fat mass (g) Adiposity (% fat) DXA body composition at 21 weeks Lean mass (g) Fat mass (g) Adiposity (% fat)
Nulliparous
Parous
0.5 1 23.3 6 0.8 24.9 6 1.0 25.5 6 1.0 26.5 6 1.2
0.75 1.25 28.9 6 0.7*** 31.4 6 1.2** 31.1 6 1.2**,## 32.8 6 1.0
17.1 6 0.7 4.8 6 0.5 21.4 6 1.2
20.4 6 0.5** 7.5 6 0.9* 26.4 6 2.1*
17.7 6 0.5 5.2 6 0.6 22.2 6 1.5
20.4 6 0.5**,## 7.5 6 0.9*,# 26.4 6 2.1
Data are mean 6 SEM of chow-fed mice only. Data are mean 6 SEM of chow fed mice only. Significance is indicated by *P < 0.05, **P < 0.01, and ***P < 0.001, for parous vs. nulliparous at the same weeks of diet. For comparison of nulliparous at 21 weeks of diet vs. parous at 8 weeks of diet, to take into account the 13-week age difference, significance is indicated by #P < 0.05 and ##P < 0.01. DXA, dual-energy X-ray absorptiometry.
Methods Animals Female C57BL/6J mice (from Charles River, Kent, UK) were put through one cycle of full-term pregnancy at 13 weeks of age (termed parous mice), and their female offspring (nulliparous mice) were group housed in a temperature-controlled facility on a 12-h light-dark cycle with free access to water and food. Nulliparous and parous mice were aged 26 weeks to a stable mid-life body weight (Table 1) and then placed on either standard chow diet, 45% HFD or HFD 6 0.04% (w/w) FEN (5). Animals (n 5 5-6 parous per diet group or n 5 8-9 nulliparous per diet group) were maintained on the diets for 27 weeks during which metabolic procedures were performed. Mice were re-fed for 1 h (following an overnight fast) prior to termination by cardiac exsanguination under CO2-induced anesthesia. Tissues were rapidly dissected, frozen in liquid nitrogen, and stored at 2808C. Perigonadal WAT depot was used in all molecular analyses. All animal procedures were approved by the University of Aberdeen Ethics Review Committee and performed following the UK Home Office project license (PPL60/3951) under the Animals (Scientific Procedures) Act 1986.
Physiology tests Body composition was measured by dual-energy X-ray absorptiometry on a PIXIMUS-Densitometer (GE Medical Systems, Buckinghamshire, UK). Insulin sensitivity was assessed by insulin tolerance testing: mice were fasted (5 h), and blood glucose measured using an Accu-Chek Compact-Plus glucometer (Roche Diagnostics, Basel, Switzerland), following a bolus of insulin (0.85 mU/g, Humulin-R, Eli-Lilly, Indianapolis, IN). Leptin and insulin were measured by ELISA (Crystal Chem, Downers Grove, IL), glucose by glucose-oxidase assay
1656
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
(TR15103; Thermo Scientific, Tewksbury, MA), and RBP4 by quantitative immunoblotting (20).
Gene expression analysis by quantitative polymerase chain reaction Total RNA was extracted from tissue samples or 3T3-L1 monolayers using TRI-Reagent (Sigma, St. Louis, MO) and cDNA synthesized using Tetro-kits (Bioline, London, UK). Quantitative polymerase chain reaction was performed using GoTaq mastermix-SYBR (Promega, Fitchburg, WI), with primers designed for specific gene targets. Standard curves and negative controls for template and reverse transcriptase were run for all genes. The RefFinder tool (Cotton-EST database, http://www.leonxie.com/referencegene.php) was used to rank reference gene stability from each experimental condition investigated. Hprt for mouse tissue or a geometric mean from three reference genes (Ywhaz, Nono, and Actb) for 3T3-L1 cells was determined and used as an internal control standard for samples. Samples were all run on a LightCycler-480 (Roche) and were analyzed using LightCyclerSW1.5 software and Pfaffl calculation (5).
Cell culture 3T3-L1 adipocytes were differentiated and maintained as described previously (5).
Statistics Data are presented as mean 6 SEM and were analyzed by one-way ANOVA with Bonferroni post hoc tests or by Student’s t-test where indicated in the figure legends.
www.obesityjournal.org
Original Article
Obesity
OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
Figure 1 Fenretinide prevents HFD-induced obesity in middle-aged, parous female mice. (A) Body weight progression of parous female C57BL/6 mice fed chow (clear circles), HFD (black triangles), or HFD plus 0.04% FEN (FEN-HFD, clear triangles). Data represent the mean 6 SEM. Significance between groups was determined by one-way ANOVA and Bonferroni post hoc test. HFD compared with chow, P < 0.05 from 4 weeks. FEN-HFD compared with HFD, P < 0.05 from 8 weeks. Body composition analysis at 21 weeks of diet: (B) fat mass (first panel), adiposity (second panel), (C) lean mass, and (D) bone mineral density (BMD) (chow, clear bar; HFD, black bar; and FEN-HFD, gray bar). Significant differences (P < 0.05) between HFD 6 FEN compared with chow are marked * and FEN-HFD compared with HFD are marked #. Student’s t-test was used to calculate significant difference for HFD compared with chow for BMD (D).
Results Fenretinide can inhibit HFD-induced obesity in female mice Because young, nulliparous female mice are lean and resistant to obesity, we used middle-aged, parous mice to determine whether FEN-HFD could inhibit HFD-induced obesity in female mice. Parous mice were 6 g heavier than nulliparous mice throughout study (Table 1). Moreover, body composition analysis at 8 and 21 weeks (i.e., 13 weeks apart) confirmed that parous mice were more obese than nulliparous mice independent of the 13-week age difference (Table 1). HFD caused rapid weight gain in parous mice (Figure 1A). FENHFD inhibited HFD-induced obesity, reaching significance from 8 weeks. Comparable results were obtained in the younger, leaner nulliparous mice (Supporting Information Figure S1). HFD markedly increased fat mass and adiposity in parous mice (Figure 1B). FENHFD inhibited this increase in fat mass and adiposity. HFD caused a small increase in lean mass, which was not affected by FEN-HFD (Figure 1C). HFD increased bone mineral density in parous mice, but FEN-HFD inhibited this increase (Figure 1D). HFD caused an approximately four fold increase in serum leptin in parous mice (Figure 2A). FEN-HFD treatment almost completely prevented this increase in circulating leptin level. Correlation analysis with fat mass confirmed that serum leptin in all three diet groups were appropriate for the level of adiposity in each mouse (Figure 2B). Comparable results for body composition and leptin level were obtained in the younger, leaner nulliparous female mice (Supporting Information Figure S1). Thus, FEN-HFD can inhibit HFD-induced obesity and closely associated hyperleptinemia in obese middle-aged, parous female mice and younger, lean nulliparous mice.
www.obesityjournal.org
Both nulliparous and parous mice are protected from obesity-associated insulin resistance FEN-HFD can inhibit HFD-induced obesity and associated hyperglycemia and insulin resistance in male mice (1,5,6). Despite the large increases in adiposity, our HFD treatment (45% kcal from fat predominately from lard) did not impair glucose homeostasis or induce insulin resistance in female C57BL/6J mice. Similar results were obtained in parous (Table 2) and nulliparous mice (Supporting Information Table S1). Thus, in female mice, FEN-HFD primarily inhibited HFD-induced adiposity and associated hyperleptinemia but did not affect insulin sensitivity.
Fenretinide can increase retinoid gene expression in female WAT and liver Next, we examined classical molecular markers of insulin resistance that we found to be down-regulated in WAT of male mice (5). HFD decreased gene expression of Pparg (the target of thiazolidinediones class of anti-diabetes drugs), Adipoq (adiponectin, the adipocytesecreted protein, which is negatively associated with insulin resistance), Pepck (the rate-limiting enzyme of glyceroneogenesis in WAT), Slc2a4 (the insulin responsive glucose transporter-4) and Rbp4 in parous mice (Figure 2C). FEN-HFD did not affect the expression of these genes in comparison with HFD. Similar results were obtained in nulliparous mice (Supporting Information Table S2). RA and FEN act through RA receptors (a, b, and c) to induce gene transcription (5,21,22), Next, we examined classic RA-responsive genes in WAT. FEN-HFD significantly up-regulated Rarb (Figure 3A) in WAT from parous mice and trended to increase Cyp26a1 (Figure 3B). Liver is also a target of FEN action in male mice (5). FEN-HFD significantly induced hepatic Cyp26a1 in parous mice (Figure 3B) and trended to increase Rarb (Figure 3A). Similarly, FEN-HFD induced classic RA-responsive genes in WAT and liver
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1657
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
Figure 2 Fenretinide prevents obesity-induced hyperleptinemia in parous mice but not decreased expression of metabolic genes in WAT. (A) Serum leptin levels at 27 weeks of diet, following overnight fasting and then 1 h re-feeding (chow, clear bar; HFD, black bar; and FEN-HFD, gray bar). (B) Positive linear correlation between serum leptin and fat mass determined from body composition analysis at 21 weeks of diet (Figure 1B). (C) Expression levels of WAT metabolic genes in parous are relative to chow (n 5 5-6 for each). Significant differences determined by one-way ANOVA and Bonferroni post hoc test between HFD and chow are marked *P < 0.05, **P < 0.01, or ***P < 0.001 and FEN-HFD compared with HFD are marked ##P < 0.01. Student’s t-test was used to calculate significant difference for HFD compared with chow for Slc2a4.
of nulliparous mice (Supporting Information Table S2). Thus, FENHFD does not seem to inhibit the down-regulation of metabolic genes in WAT of female mice, but does robustly induce retinoidresponsive gene expression in female WAT and liver.
Parous females have decreased expression of hepatic estrogen-responsive genes To determine whether increased obesity in parous mice (following a first full-term pregnancy) was associated with altered estrogen signaling, we determined the expression of known targets of hepatic estrogen-ERa. Parous mice had decreased levels of hepatic Rara mRNA compared with nulliparous animals (Figure 4A). In addition, parous mice had significantly decreased levels of other known hepatic estrogen-ERa targets, Ppara, insulin-like growth factor-1, Igf1, and X-box binding protein-1, Xbp1, compared with nulliparous mice (Figure 4B-D; references in Supporting Information). To test whether decreased hepatic expression of these estrogenresponsive genes could lead to alterations in expression of down-
1658
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
stream metabolic targets, next, we measured the expression of Acox1 and Cyp7b1 (induced and repressed by PPARa signaling, respectively) and Edem1 and Erdj4 (targets of XBP1 endoplasmic reticulum stress pathway). Parous mice exhibited decreased levels of hepatic Acox1 and Edem1 and had increased hepatic Cyp7b1 compared with nulliparous mice (Supporting Information Figure S2). Parity did not affect hepatic Erdj4. However, in parous mice, HFD significantly increased hepatic Erdj4, indicating induction of ER stress. FEN-HFD prevented this increase in hepatic Erdj4 (Supporting Information Figure S2). Hepatic Esr1 expression was unchanged (not shown). Together, these gene expression alterations suggest that parous females have reduced estrogen signaling and may provide a mechanism for parous mice to develop obesity compared with nulliparous females.
Fenretinide can increase local estrogen signaling in WAT We examined the possibility that in addition to inducing RA signaling, FEN-HFD could increase estrogen signaling in WAT of parous
www.obesityjournal.org
Original Article
Obesity
OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
TABLE 2 Serum parameters associated with insulin sensitivity in parous female mice fed chow, HFD, or fenretinide (FEN)-HFD
Serum levels Glucose (mmol/l) Insulin (ng/ml) RBP4 protein (fold) ITT (blood glucose, mmol/l) Basal 30 min post-insulin bolus 60 min post-insulin bolus
Chow
HFD
FEN-HFD
9.6 6 0.5 2.4 6 0.4 1.0 6 0.08
11.08 6 1.1 2.8 6 0.4 1.1 6 0.09
10.6 6 1.0 nd 0.9 6 0.05
9.3 6 0.4 6.0 6 0.3 4.4 6 0.4
9.9 6 0.2 5.8 6 0.3 5.6 6 0.4
9.8 6 0.2 6.3 6 0.5 5.8 6 0.4
Serum glucose and serum insulin were measured following 5-h food removal at 12 weeks of diet and serum leptin and serum RBP4 at 27 weeks of diet, following overnight fasting and then 1 h re-feeding. Insulin tolerance test (ITT) (0.85 mU/g insulin) was performed following 5-h food removal at 25 weeks of diet in parous mice. Data represent the mean 6 SEM. Significance between groups was determined by one-way ANOVA and Bonferroni post hoc test. Serum insulin for FENHFD not determined (nd).
female mice. HFD decreased expression of Esr1, and FEN-HFD did not alter this down-regulation (Figure 5A). HFD trended to decrease ERb (Esr2) expression, whereas FEN-HFD trended to increase its expression (Figure 5B). Estrogen is synthesized locally by WAT
Figure 4 Decreased levels of estrogen target genes in liver of parous mice. Levels in nulliparous and parous are relative to respective chow group (n 5 4-8 for each group). There were no significant differences with diet in either group of female mice. Significance between nulliparous and parous mice was determined by t-test and marked **P < 0.01 or ***P < 0.001.
through the expression of synthetic enzymes, where the final two steps in this pathway are regulated by Hsd17b1 and cytochrome P450-enzyme aromatase (Cyp19a1). Gene expression of Hsd17b1 was not detectable in WAT. HFD did not affect Cyp19a1 expression; however, FEN-HFD significantly increased Cyp19a1 in parous mice (Figure 5B). The cytokine IL6 has been reported to stimulate aromatase activity in breast tissue, cells, and WAT (23,24). HFD trended to decrease IL6 expression; however, FEN-HFD significantly increased IL6 expression in parous mice (Figure 5). Estrogen has also been reported to induce vascular endothelial growth factor (VEGF) (25). HFD decreased Vegf expression, whereas FEN-HFD significantly increased Vegf expression in parous mice (Figure 5). The effects of FEN-HFD on estrogen signaling were only observed in parous females and not in nulliparous mice (data not shown). Thus, FENHFD seems to increase local estrogen signaling in WAT of parous female mice. Figure 3 Fenretinide can induce classic RA-responsive genes in WAT and liver in parous female mice. Expression levels of (A) Rarb and (B) Cyp26a1 in WAT and liver. Significant differences determined by one-way ANOVA and Bonferroni post hoc test for FEN-HFD compared with chow are marked *P < 0.05 or ***P < 0.001 and FEN-HFD compared with HFD are marked ##P < 0.01 or ###P < 0.001.
www.obesityjournal.org
Fenretinide can directly induce retinoid and estrogen signaling in 3T3-L1 adipocytes In vivo, WAT interacts with and receives signals from many other tissues to produce the responses measured earlier. To test whether
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1659
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
Figure 5 Fenretinide and HFD-induced obesity can independently alter estrogen signaling in WAT. Gene expression levels of estrogen receptors (A) Esr1 and (B) Esr2, estrogen synthetic enzyme (Cyp19a1), and targets of estrogen signaling (IL6 and Vegf) in WAT from parous mice. Data are relative to chow, and significance between groups (n 5 5-6 for each) was determined by one-way ANOVA and Bonferroni post hoc test. Significant differences (P < 0.05) for HFD 6 FEN compared with chow are marked * and FEN-HFD compared with HFD are marked #. Student’s t-test was used to calculate significant difference for HFD compared with chow for Vegf (P 5 0.02).
FEN could directly increase estrogen signaling, 3T3-L1 pre-adipocytes were differentiated into adipocytes for 8 days, followed by treatment with physiological concentrations of E2, RA, and FEN. Both FEN and RA increased the expression of Crbp1 and Rarg (Figure 6A,B). When combined, FEN1E2 or RA1E2 trended to additively increase the expression of Crbp1 (Figure 6A). E2 significantly increased Rarg and, when combined, RA1E2 additively increased the expression of Rarg (Figure 6B). FEN and FEN1E2 increased Esr2 expression similarly (Figure 6C). FEN also increased the expression of Cyp19a1 (Figure 6D). RA did not alter the expression of either Esr2 or Cyp19a1. Together, these results confirm the findings in WAT from parous mice and suggest that one of the beneficial mechanisms of FEN action in females may be through increasing estrogen signaling in WAT via Cyp19a1 and Esr2.
Discussion We recently demonstrated that FEN is a potent modulator of genes involved in retinoid and glucose/lipid homeostasis. These molecular alterations were in association with inhibiting HFD-induced obesity and insulin resistance in male mice (5). Females are partially protected from obesity and insulin resistance via the sex hormone estrogen. However, these prior studies have focused on postmenopausal women or an experimental model of menopause via ovariectomy (912). Our new study clearly demonstrates that FEN can decrease HFD-induced obesity in both nulliparous and parous female mice and that part of the mechanism may be through increased retinoid and estrogen signaling in WAT. Furthermore, this beneficial metabolic effect occurs despite a decrease in the targets of estrogen signaling and increased obesity in parous mice, a physiological relevant model of female obesity (19). A direct target of FEN seems to be leptin, the WAT-secreted protein. FEN inhibited HFD-induced hyperleptinemia in male (5) and female (Figure 2) mice and directly inhibited leptin mRNA by 75%
1660
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
in adipocytes (5). It is known that circulating leptin levels correlate strongly with WAT mass (Figure 2B), and increased levels in obesity are associated with resistance to the anorexic and thermogenic effects of leptin (26). However, in comprehensive studies in male mice (5,6), FEN did not have a measurable effect to decrease food intake or to increase oxygen consumption or Ucp1 expression. FEN treatment has been shown to improve insulin sensitivity and decrease serum leptin levels in overweight premenopausal women (27). However, the parity status of the women was not reported, nor has FEN treatment been examined in postmenopausal women at higher risk of developing obesity and metabolic complications. Our data suggest that future clinical studies should take this into account. We found that parous mice had significantly down-regulated genes that have been shown to be direct targets of estrogen signaling (e.g., Igf1 and Rara). This suggests that similar to ovariectomy-induced obesity in females, decreased circulating estrogen may be the primary mechanism for increased adiposity in parous mice (on chow diet or HFD). The beneficial effects of estrogen on energy and glucose homeostasis are largely mediated by Esr1 (13-15). However, hepatic Esr1 signaling is not required for the effects of E2 treatment on adiposity (28). Interestingly, another target of FEN seems to involve increased local production of estrogen in WAT, via up-regulation of adipose Cyp19a1 and signaling through induction of WAT Esr2 (Figures 4 and 5, WAT and 3T3-L1 adipocytes, respectively). This is in contrast to the liver, which is wholly responsive to circulating levels of estrogen because enzymes involved in estrogen production are expressed at very low levels (if at all). Mice of both sexes lacking Cyp19a1 develop obesity (29). However, any beneficial effects of Esr2 activation are not clear because HFD-fed Esr2-KO mice demonstrate increased obesity, but improved insulin sensitivity and glucose tolerance (30). In contrast, Esr2-selective ligands have been shown to alleviate HFD- and ovariectomy-induced obesity in mice (31). Moreover, FEN-induced induction of Vegf in WAT may be linked to an increase in local estrogen signaling and may contribute to the beneficial effects of FEN treatment because WAT
www.obesityjournal.org
Original Article
Obesity
OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
metabolic homeostasis in insulin-sensitive tissues remains to be elucidated. There were some potential issues with our study design, largely because nulliparous mice were the offspring of the parous mice and, therefore, 13 weeks younger. To minimize the impact of this age difference, all mice were aged for 26 weeks to achieve stable midlife body weight. Parous mice started the study 6 g heavier than nulliparous mice, and nulliparous mice did not gain this difference in weight over the duration of the study. Moreover, body composition analysis confirmed that parous mice were more obese, independent of age (Table 1). Indeed, others have previously demonstrated in mice using age-matched controls that parity leads to increased adiposity independent of maternal age (19). In addition, female mice undergo reproductive senescence much later in life or with major genetic, surgical, or pharmacologic intervention (35,36). Thus overall, our findings are unlikely to be due to the relatively small difference in middle age between nulliparous and parous mice. There is also the possibility that variation in circulating hormone concentrations during stages of the estrous cycle and thereby at the end of the study may have affected gene expression. We did not determine stages of the estrous cycle by visual observation or vaginal cytology, but we assumed that estrous cycles will have been primarily synchronized and suppressed by the presence of male mice in the room and because both parous and nulliparous were housed in the same room (37,38).
Figure 6 Fenretinide increases classic retinoid signaling and estrogen signaling via Esr2 and Cyp19a1 in adipocytes. Differentiated 3T3-L1 adipocytes were utilized to determine gene expression changes in response to FEN or RA 6 E2 treatments. 3T3-L1 cultures were differentiated with MDI alone for 8 days and then subjected to the indicated treatments for an additional 8 days (1 lM FEN, 0.1 lM RA, and 1 nM E2). Relative expression levels of (A,B) retinoid genes or (C,D) estrogen genes observed after treatment (n 5 4 separate experiments on different plates of cells). Significance between groups was determined by two-way ANOVA. There was a significant effect (P < 0.05 overall) of retinoid treatment for both RA and FEN vs. VEH on levels of Crbp1 and Rarg. There was a significant effect of E2 treatment on Rarg levels. Post hoc tests: Holm-Sidak method for pair-wise multiple comparisons determined significance between retinoid treatment groups (RA, FEN, and VEH) within 6 E2 treatment groups are marked *P < 0.05, **P < 0.01, or ***P < 0.001 for comparisons with VEH or #P < 0.05 or ###P < 0.001 for comparison with RA. †† marks significance between 6 E2 treatment for within VEH or RA treatment. For some analyses, Student’s t-test was performed: Rarg was increased by E2 alone (t-test, P 5 0.003, marked ††) and by FEN alone (t-test, P 5 0.02, marked *), and Crbp1 was increased by FEN alone (t-test, P 5 0.03, marked *) compared with VEH without E2.
Although our sample sizes seemed to be sufficient to obtain significant differences in adiposity parameters and retinoid signaling with FEN treatment, many molecular markers of metabolic disturbances were not detected. Interestingly, in our other HFD studies, an alternate source of fat (hydrogenated vegetable fat rich in trans fats) did cause significantly more insulin resistance in female mice (data not shown). This suggests that there is a complex interplay between types of diet/dietary fats consumed, the protective effects of E2 signaling, and alterations in glucose/lipid homeostasis. Childbearing and increasing parity is associated with future development of maternal obesity. Moreover, obesity is a major risk factor for diabetes during pregnancy (gestational diabetes mellitus), which is strongly linked to an increase in maternal and offspring diabetes risk (39,40). There is an urgent need for novel therapies that are efficient and non-toxic to treat obesity and insulin resistance. Recent studies have shown that retinoid homeostasis and estrogen signaling are potential therapeutic targets for preventing or reducing obesity. Here, we present evidence that FEN is a modulator of estrogen and retinoid homeostasis genes in association with prevention of female postpregnancy and HFD-induced obesity.O
overexpression of VEGF protects against diet-induced obesity and insulin resistance (25,32).
Acknowledgments Rara is an estrogen-induced gene, and its expression in breast tumors has been shown to correlate with ER expression (33). More recent studies have implicated Rara as an essential component of the ER complex that can cooperate for effective transcriptional activity in breast cancer cells (34). Our data also suggest an interaction between estrogen and FEN/retinoid signaling in adipocytes. The details of this interaction and its potential importance in maintaining
www.obesityjournal.org
The authors thank J.R. Speakman (UoA) for invaluable discussion regarding energy expenditure, T. Martin (Johnson & Johnson, NJ) and U. Thurneer (Cilag AG, Switzerland) for fenretinide, to use completely without restriction or obligation, and undergraduate student K. Towle (UoA) for assisting in experiments. C 2015 The Obesity Society V
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1661
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
References 1. Yang Q, Graham TE, Mody N, et al. Serum retinol binding protein 4 contributes to insulin resistance in obesity and type 2 diabetes. Nature 2005;436:356-362. 2. Graham TE, Yang Q, Bluher M, et al. Retinol-binding protein 4 and insulin resistance in lean, obese, and diabetic subjects. N Engl J Med 2006;354:2552-2563. 3. Gavi S, Stuart LM, Kelly P, et al. Retinol-binding protein 4 is associated with insulin resistance and body fat distribution in nonobese subjects without type 2 diabetes. J Clin Endocrinol Metab 2007;92:1886-1890. 4. McTernan PG, Kumar S. Editorial: retinol binding protein 4 and pathogenesis of diabetes. J Clin Endocrinol Metab 2007;92:2430-2432. 5. Mcilroy GD, Delibegovic M, Owen C, et al. Fenretinide treatment prevents dietinduced obesity in association with major alterations in retinoid homeostatic gene expression in adipose, liver, and hypothalamus. Diabetes 2013;62:825-836. 6. Preitner F, Mody N, Graham TE, Peroni OD, Kahn BB. Long-term fenretinide treatment prevents high-fat diet-induced obesity, insulin resistance, and hepatic steatosis. Am J Physiol Endocrinol Metab 2009;297:E1420-E1429. 7. Rotmensz N, De Palo G, Formelli F, et al. Long-term tolerability of fenretinide (4HPR) in breast cancer patients. Eur J Cancer 1991;27:1127-1131. 8. Zanardi S, Serrano D, Argusti A, Barile M, Puntoni M, Decensi A. Clinical trials with retinoids for breast cancer chemoprevention. Endocr Relat Cancer 2006;13:51-68. 9. Carr MC. The emergence of the metabolic syndrome with menopause. J Clin Endocrinol Metab 2003;88:2404-2411. 10. Kanaya AM, Herrington D, Vittinghoff E, et al. Glycemic effects of postmenopausal hormone therapy: the heart and estrogen/progestin replacement study. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 2003;138:1-9. 11. Margolis KL, Bonds DE, Rodabough RJ, et al. Effect of oestrogen plus progestin on the incidence of diabetes in postmenopausal women: results from the women’s health initiative hormone trial. Diabetologia 2004;47:1175-1187. 12. Wagner JD, Thomas MJ, Williams JK, Zhang L, Greaves KA, Cefalu WT. Insulin sensitivity and cardiovascular risk factors in ovariectomized monkeys with estradiol alone or combined with nomegestrol acetate. J Clin Endocrinol Metab 1998;83:896901. 13. Heine PA, Taylor JA, Iwamoto GA, Lubahn DB, Cooke PS. Increased adipose tissue in male and female estrogen receptor-alpha knockout mice. Proc Natl Acad Sci U S A 2000;97:12729-12734. 14. Ohlsson C, Hellberg N, Parini P, et al. Obesity and disturbed lipoprotein profile in estrogen receptor-alpha-deficient male mice. Biochem Biophys Res Commun 2000; 278:640-645. 15. Ribas V, Nguyen MT, Henstridge DC, et al. Impaired oxidative metabolism and inflammation are associated with insulin resistance in ERalpha-deficient mice. Am J Physiol Endocrinol Metab 2010;298:E304-E319. 16. Gunderson EP, Jacobs DR Jr, Chiang V, et al. Childbearing is associated with higher incidence of the metabolic syndrome among women of reproductive age controlling for measurements before pregnancy: the CARDIA study. Am J Obstet Gynecol 2009;201:177.e1-177.e9. 17. Hajiahmadi M, Shafi H, Delavar MA. Impact of parity on obesity: a cross-sectional study in Iranian women. Med Princ Pract 2015;24:70-74. 18. Bernstein L, Pike MC, Ross RK, Judd HL, Brown JB, Henderson BE. Estrogen and sex hormone-binding globulin levels in nulliparous and parous women. J Natl Cancer Inst 1985;74:741-745. 19. Rebholz SL, Jones T, Burke KT, et al. Multiparity leads to obesity and inflammation in mothers and obesity in male offspring. Am J Physiol Endocrinol Metab 2012;302:E449-E457.
1662
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
20. Mody N, Graham TE, Tsuji Y, Yang Q, Kahn BB. Decreased clearance of serum retinol-binding protein and elevated levels of transthyretin in insulin-resistant ob/ob mice. Am J Physiol Endocrinol Metab 2008;294:E785-E793. 21. Fanjul AN, Delia D, Pierotti MA, Rideout D, Yu JQ, Pfahl M. 4-Hydroxyphenyl retinamide is a highly selective activator of retinoid receptors. J Biol Chem 1996; 271:22441-22446. 22. Sabichi AL, Hendricks DT, Bober MA, Birrer MJ. Retinoic acid receptor beta expression and growth inhibition of gynecologic cancer cells by the synthetic retinoid N-(4-hydroxyphenyl) retinamide. J Natl Cancer Inst 1998;90:597-605. 23. Reed MJ, Coldham NG, Patel SR, Ghilchik MW, James VH. Interleukin-1 and interleukin-6 in breast cyst fluid: their role in regulating aromatase activity in breast cancer cells. J Endocrinol 1992;132:R5-R8. 24. Zhao Y, Nichols JE, Bulun SE, Mendelson CR, Simpson ER. Aromatase p450 gene expression in human adipose tissue. Role of a jak/STAT pathway in regulation of the adipose-specific promoter. J Biol Chem 1995;270:16449-16457. 25. Koos RD. Mini review: putting physiology back into estrogens’ mechanism of action. Endocrinology 2011;152:4481-4488. 26. Considine RV, Sinha MK, Heiman ML, et al. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 1996;334:292295. 27. Johansson H, Gandini S, Guerrieri-Gonzaga A, et al. Effect of fenretinide and lowdose tamoxifen on insulin sensitivity in premenopausal women at high risk for breast cancer. Cancer Res 2008;68:9512-9518. 28. Zhu L, Brown WC, Cai Q, et al. Estrogen treatment after ovariectomy protects against fatty liver and may improve pathway-selective insulin resistance. Diabetes 2013;62:424-434. 29. Jones ME, Thorburn AW, Britt KL, et al. Aromatase-deficient (ArKO) mice have a phenotype of increased adiposity. Proc Natl Acad Sci U S A 2000;97:12735-12740. 30. Foryst-Ludwig A, Clemenz M, Hohmann S, et al. Metabolic actions of estrogen receptor beta (ERbeta) are mediated by a negative cross-talk with PPARgamma. PLoS Genet 2008;4:e1000108. 31. Yepuru M, Eswaraka J, Kearbey JD, et al. Estrogen receptor-{beta}-selective ligands alleviate high-fat diet- and ovariectomy-induced obesity in mice. J Biol Chem 2010;285:31292-31303. 32. Elias I, Franckhauser S, Ferre T, et al. Adipose tissue overexpression of vascular endothelial growth factor protects against diet-induced obesity and insulin resistance. Diabetes 2012;61:1801-1813. 33. Roman SD, Ormandy CJ, Manning DL, et al. Estradiol induction of retinoic acid receptors in human breast cancer cells. Cancer Res 1993;53:5940-5945. 34. Ross-Innes CS, Stark R, Holmes KA, et al. Cooperative interaction between retinoic acid receptor-alpha and estrogen receptor in breast cancer. Genes Dev 2010;24:171-182. 35. Danilovich N, Babu PS, Xing W, Gerdes M, Krishnamurthy H, Sairam MR. Estrogen deficiency, obesity, and skeletal abnormalities in follicle-stimulating hormone receptor knockout (FORKO) female mice. Endocrinology 2000;141:42954308. 36. Danilovich N, Ram Sairam M. Recent female mouse models displaying advanced reproductive aging. Exp Gerontol 2006;41:117-122. 37. Whitten WK. Modification of the oestrous cycle of the mouse by external stimuli associated with the male. J Endocrinol 1956;13:399-404. 38. Whitten WK. Occurrence of anoestrus in mice caged in groups. J Endocrinol 1959; 18:102-107. 39. Chu SY, Callaghan WM, Kim SY, et al. Maternal obesity and risk of gestational diabetes mellitus. Diabetes Care 2007;30:2070-2076. 40. Lawlor DA, Relton C, Sattar N, Nelson SM. Maternal adiposity-a determinant of perinatal and offspring outcomes? Nat Rev Endocrinol 2012;8:679-688.
www.obesityjournal.org
Original Article OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
Fenretinide Prevents Obesity in Aged Female Mice in Association with Increased Retinoid and Estrogen Signaling Kirsty D. Shearer1, Nicola Morrice1,2, Claire Henderson1, Jenny Reekie1, George D. Mcilroy1, Peter J. McCaffery1, Mirela Delibegovic1, and Nimesh Mody1
Objective: The synthetic retinoid fenretinide (FEN) inhibits adiposity in male mice fed a high-fat diet (HFD) in association with alterations in retinoic acid (RA) signaling. Young female mice are protected from obesity via estrogen signaling. We, therefore, investigated whether FEN also influences adiposity in aged female mice differing in parity and whether such effects are mediated by retinoid and estrogen signaling. Methods: Aged nulliparous and parous female mice were maintained on HFD 6 FEN, and adiposity was assessed. Quantitative polymerase chain reaction was performed on white adipose tissue (WAT), liver, and 3T3-L1 adipocytes treated with RA or FEN 6 estrogen. Results: Parous females were more obese than nulliparous mice independent of age. FEN-HFD prevented the HFD-induced increase in adiposity and leptin levels independently of parity. FEN-HFD induced retinoid-responsive genes in WAT and liver. Parous females had reduced expression of hepatic estrogenresponsive genes, but FEN-HFD up-regulated WAT Cyp19a1 and Esr2 in parous mice. Estrogen and RA acted synergistically to increase RA receptor-mediated gene expression in 3T3-L1 adipocytes. FEN increased Cyp19a1 and Esr2, similar to our findings in vivo. Conclusions: The prevention of adiposity by FEN in response to HFD in female mice seems to involve increased retinoid signaling in association with induction of local estrogen production and estrogen signaling in WAT. Obesity (2015) 23, 1655–1662. doi:10.1002/oby.21164
Introduction Vitamin A metabolism leading to retinoic acid (RA) signaling is a nuclear receptor pathway that plays an important role in regulating body fat and glucose homeostasis. Several reports have associated increased vitamin A levels with diabetes (1-3). However, not all studies have confirmed these links, and thus the mechanism(s) by which retinoid metabolism may regulate glucose homeostasis remains unclear. For example, serum retinol-binding protein (RBP)-4 levels are elevated in many, but not all, mouse models and humans with obesity-associated insulin resistance (4). We have recently shown that in male high-fat diet (HFD)-induced obese mice, the synthetic retinoid fenretinide (FEN) can inhibit obesity and normalize eleva-
tions in serum RBP4, insulin, and glucose levels and hepatic insulin resistance (5,6). In 3T3-L1 adipocytes and in vivo, we found that FEN-mediated beneficial effects are associated with alterations in retinoid homeostasis genes (5). In HFD male mice, RA-responsive genes were markedly increased in white adipose tissue (WAT) and liver with short- and long-term FEN (5). In WAT, FEN inhibited the down-regulation of peroxisome proliferator-activated receptor-c (Pparg) and improved insulin sensitivity and levels of adiponectin and RBP4. FEN also inhibited hyperleptinemia in vivo and leptin expression in adipocytes (5). FEN is the most widely studied retinoid in breast cancer chemoprevention trials, because of its selective accumulation in breast tissue
1 Institute of Medical Sciences, College of Life Sciences and Medicine, University of Aberdeen, Foresterhill Health Campus, Aberdeen, Scotland, UK. Correspondence: Nimesh Mody ([email protected]) 2 Centre for Genome-Enabled Biology and Medicine, University of Aberdeen, Aberdeen, Scotland, UK.
Funding agencies: This study was supported by British Heart Foundation Intermediate Basic Research Fellowship grant FS/09/026 (to N. Mody), Research Councils UK Fellowship (to M.D.), European Foundation for the Study of Diabetes (EFSD)/Lilly Programme grant (to M.D. and N. Mody), Biotechnology and Biological Sciences Research Council (BBSRC) studentship (to G.D.M.), University of Aberdeen (UoA) Centre for Genome-Enabled Biology & Medicine (CGEBM) studentship (to N. Morrice), and BBSRC project grant BB/G014272/1 (to P.J.M.). Disclosure: The authors declared no conflict of interest. Author contributions: N. Mody and M.D. conceived the study. All authors participated in performing some of the experiments. K.D.S., N. Morrice, C.H., J.R., G.D.M., and N. Mody analyzed the data. K.D.S. and N. Mody made the figures and wrote the manuscript. P.J.M. and M.D. were involved in editing and proofreading the manuscript. Additional Supporting Information may be found in the online version of this article. Received: 12 November 2014; Accepted: 20 April 2015; Published online 14 July 2015. doi:10.1002/oby.21164
www.obesityjournal.org
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1655
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
and its favorable toxicological profile (7,8). However, the effect of FEN on female obesity has not previously been reported, because all earlier rodent studies were conducted in male models. Young female mice (and humans) are protected from obesity in an estrogen-dependent manner (9-12). Thereby, decreased estrogen levels or global deletion of the estrogen receptor, ERa (Esr1), in mice are both associated with increased adiposity and insulin resistance (13-15). Childbearing and increasing parity are associated with future development of maternal obesity and type 2 diabetes (16,17). Estrogen levels have also been found to be lower in parous women (2 years after delivery/lactation) compared with those measured in nulliparous women (having never given birth) (18). However, very few obesity studies have been performed in parous mice (19). We hypothesized that an interaction between estrogen and RA signaling may play an important role in regulating adiposity and glucose homeostasis. Thus, in this study, we investigated whether FEN could prevent HFD-induced obesity in aged female mice, nulliparous and parous (after pregnancy). In addition, we have examined the molecular mechanisms involved particularly focusing on increased gene expression of metabolically important genes and those involved in estrogen and RA signaling in mouse WAT and liver, and in 3T3-L1 murine adipocytes.
TABLE 1 Parous mice are more obese than nulliparous offspring, independent of a 13-week age difference
Parity status
Age at start of diet (years) Age at end of diet (years) Weight at start of diet (g) Weight at 8 weeks of diet (g) Weight at 21 weeks of diet (g) Weight at end of diet (g) DXA body composition at 8 weeks Lean mass (g) Fat mass (g) Adiposity (% fat) DXA body composition at 21 weeks Lean mass (g) Fat mass (g) Adiposity (% fat)
Nulliparous
Parous
0.5 1 23.3 6 0.8 24.9 6 1.0 25.5 6 1.0 26.5 6 1.2
0.75 1.25 28.9 6 0.7*** 31.4 6 1.2** 31.1 6 1.2**,## 32.8 6 1.0
17.1 6 0.7 4.8 6 0.5 21.4 6 1.2
20.4 6 0.5** 7.5 6 0.9* 26.4 6 2.1*
17.7 6 0.5 5.2 6 0.6 22.2 6 1.5
20.4 6 0.5**,## 7.5 6 0.9*,# 26.4 6 2.1
Data are mean 6 SEM of chow-fed mice only. Data are mean 6 SEM of chow fed mice only. Significance is indicated by *P < 0.05, **P < 0.01, and ***P < 0.001, for parous vs. nulliparous at the same weeks of diet. For comparison of nulliparous at 21 weeks of diet vs. parous at 8 weeks of diet, to take into account the 13-week age difference, significance is indicated by #P < 0.05 and ##P < 0.01. DXA, dual-energy X-ray absorptiometry.
Methods Animals Female C57BL/6J mice (from Charles River, Kent, UK) were put through one cycle of full-term pregnancy at 13 weeks of age (termed parous mice), and their female offspring (nulliparous mice) were group housed in a temperature-controlled facility on a 12-h light-dark cycle with free access to water and food. Nulliparous and parous mice were aged 26 weeks to a stable mid-life body weight (Table 1) and then placed on either standard chow diet, 45% HFD or HFD 6 0.04% (w/w) FEN (5). Animals (n 5 5-6 parous per diet group or n 5 8-9 nulliparous per diet group) were maintained on the diets for 27 weeks during which metabolic procedures were performed. Mice were re-fed for 1 h (following an overnight fast) prior to termination by cardiac exsanguination under CO2-induced anesthesia. Tissues were rapidly dissected, frozen in liquid nitrogen, and stored at 2808C. Perigonadal WAT depot was used in all molecular analyses. All animal procedures were approved by the University of Aberdeen Ethics Review Committee and performed following the UK Home Office project license (PPL60/3951) under the Animals (Scientific Procedures) Act 1986.
Physiology tests Body composition was measured by dual-energy X-ray absorptiometry on a PIXIMUS-Densitometer (GE Medical Systems, Buckinghamshire, UK). Insulin sensitivity was assessed by insulin tolerance testing: mice were fasted (5 h), and blood glucose measured using an Accu-Chek Compact-Plus glucometer (Roche Diagnostics, Basel, Switzerland), following a bolus of insulin (0.85 mU/g, Humulin-R, Eli-Lilly, Indianapolis, IN). Leptin and insulin were measured by ELISA (Crystal Chem, Downers Grove, IL), glucose by glucose-oxidase assay
1656
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
(TR15103; Thermo Scientific, Tewksbury, MA), and RBP4 by quantitative immunoblotting (20).
Gene expression analysis by quantitative polymerase chain reaction Total RNA was extracted from tissue samples or 3T3-L1 monolayers using TRI-Reagent (Sigma, St. Louis, MO) and cDNA synthesized using Tetro-kits (Bioline, London, UK). Quantitative polymerase chain reaction was performed using GoTaq mastermix-SYBR (Promega, Fitchburg, WI), with primers designed for specific gene targets. Standard curves and negative controls for template and reverse transcriptase were run for all genes. The RefFinder tool (Cotton-EST database, http://www.leonxie.com/referencegene.php) was used to rank reference gene stability from each experimental condition investigated. Hprt for mouse tissue or a geometric mean from three reference genes (Ywhaz, Nono, and Actb) for 3T3-L1 cells was determined and used as an internal control standard for samples. Samples were all run on a LightCycler-480 (Roche) and were analyzed using LightCyclerSW1.5 software and Pfaffl calculation (5).
Cell culture 3T3-L1 adipocytes were differentiated and maintained as described previously (5).
Statistics Data are presented as mean 6 SEM and were analyzed by one-way ANOVA with Bonferroni post hoc tests or by Student’s t-test where indicated in the figure legends.
www.obesityjournal.org
Original Article
Obesity
OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
Figure 1 Fenretinide prevents HFD-induced obesity in middle-aged, parous female mice. (A) Body weight progression of parous female C57BL/6 mice fed chow (clear circles), HFD (black triangles), or HFD plus 0.04% FEN (FEN-HFD, clear triangles). Data represent the mean 6 SEM. Significance between groups was determined by one-way ANOVA and Bonferroni post hoc test. HFD compared with chow, P < 0.05 from 4 weeks. FEN-HFD compared with HFD, P < 0.05 from 8 weeks. Body composition analysis at 21 weeks of diet: (B) fat mass (first panel), adiposity (second panel), (C) lean mass, and (D) bone mineral density (BMD) (chow, clear bar; HFD, black bar; and FEN-HFD, gray bar). Significant differences (P < 0.05) between HFD 6 FEN compared with chow are marked * and FEN-HFD compared with HFD are marked #. Student’s t-test was used to calculate significant difference for HFD compared with chow for BMD (D).
Results Fenretinide can inhibit HFD-induced obesity in female mice Because young, nulliparous female mice are lean and resistant to obesity, we used middle-aged, parous mice to determine whether FEN-HFD could inhibit HFD-induced obesity in female mice. Parous mice were 6 g heavier than nulliparous mice throughout study (Table 1). Moreover, body composition analysis at 8 and 21 weeks (i.e., 13 weeks apart) confirmed that parous mice were more obese than nulliparous mice independent of the 13-week age difference (Table 1). HFD caused rapid weight gain in parous mice (Figure 1A). FENHFD inhibited HFD-induced obesity, reaching significance from 8 weeks. Comparable results were obtained in the younger, leaner nulliparous mice (Supporting Information Figure S1). HFD markedly increased fat mass and adiposity in parous mice (Figure 1B). FENHFD inhibited this increase in fat mass and adiposity. HFD caused a small increase in lean mass, which was not affected by FEN-HFD (Figure 1C). HFD increased bone mineral density in parous mice, but FEN-HFD inhibited this increase (Figure 1D). HFD caused an approximately four fold increase in serum leptin in parous mice (Figure 2A). FEN-HFD treatment almost completely prevented this increase in circulating leptin level. Correlation analysis with fat mass confirmed that serum leptin in all three diet groups were appropriate for the level of adiposity in each mouse (Figure 2B). Comparable results for body composition and leptin level were obtained in the younger, leaner nulliparous female mice (Supporting Information Figure S1). Thus, FEN-HFD can inhibit HFD-induced obesity and closely associated hyperleptinemia in obese middle-aged, parous female mice and younger, lean nulliparous mice.
www.obesityjournal.org
Both nulliparous and parous mice are protected from obesity-associated insulin resistance FEN-HFD can inhibit HFD-induced obesity and associated hyperglycemia and insulin resistance in male mice (1,5,6). Despite the large increases in adiposity, our HFD treatment (45% kcal from fat predominately from lard) did not impair glucose homeostasis or induce insulin resistance in female C57BL/6J mice. Similar results were obtained in parous (Table 2) and nulliparous mice (Supporting Information Table S1). Thus, in female mice, FEN-HFD primarily inhibited HFD-induced adiposity and associated hyperleptinemia but did not affect insulin sensitivity.
Fenretinide can increase retinoid gene expression in female WAT and liver Next, we examined classical molecular markers of insulin resistance that we found to be down-regulated in WAT of male mice (5). HFD decreased gene expression of Pparg (the target of thiazolidinediones class of anti-diabetes drugs), Adipoq (adiponectin, the adipocytesecreted protein, which is negatively associated with insulin resistance), Pepck (the rate-limiting enzyme of glyceroneogenesis in WAT), Slc2a4 (the insulin responsive glucose transporter-4) and Rbp4 in parous mice (Figure 2C). FEN-HFD did not affect the expression of these genes in comparison with HFD. Similar results were obtained in nulliparous mice (Supporting Information Table S2). RA and FEN act through RA receptors (a, b, and c) to induce gene transcription (5,21,22), Next, we examined classic RA-responsive genes in WAT. FEN-HFD significantly up-regulated Rarb (Figure 3A) in WAT from parous mice and trended to increase Cyp26a1 (Figure 3B). Liver is also a target of FEN action in male mice (5). FEN-HFD significantly induced hepatic Cyp26a1 in parous mice (Figure 3B) and trended to increase Rarb (Figure 3A). Similarly, FEN-HFD induced classic RA-responsive genes in WAT and liver
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1657
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
Figure 2 Fenretinide prevents obesity-induced hyperleptinemia in parous mice but not decreased expression of metabolic genes in WAT. (A) Serum leptin levels at 27 weeks of diet, following overnight fasting and then 1 h re-feeding (chow, clear bar; HFD, black bar; and FEN-HFD, gray bar). (B) Positive linear correlation between serum leptin and fat mass determined from body composition analysis at 21 weeks of diet (Figure 1B). (C) Expression levels of WAT metabolic genes in parous are relative to chow (n 5 5-6 for each). Significant differences determined by one-way ANOVA and Bonferroni post hoc test between HFD and chow are marked *P < 0.05, **P < 0.01, or ***P < 0.001 and FEN-HFD compared with HFD are marked ##P < 0.01. Student’s t-test was used to calculate significant difference for HFD compared with chow for Slc2a4.
of nulliparous mice (Supporting Information Table S2). Thus, FENHFD does not seem to inhibit the down-regulation of metabolic genes in WAT of female mice, but does robustly induce retinoidresponsive gene expression in female WAT and liver.
Parous females have decreased expression of hepatic estrogen-responsive genes To determine whether increased obesity in parous mice (following a first full-term pregnancy) was associated with altered estrogen signaling, we determined the expression of known targets of hepatic estrogen-ERa. Parous mice had decreased levels of hepatic Rara mRNA compared with nulliparous animals (Figure 4A). In addition, parous mice had significantly decreased levels of other known hepatic estrogen-ERa targets, Ppara, insulin-like growth factor-1, Igf1, and X-box binding protein-1, Xbp1, compared with nulliparous mice (Figure 4B-D; references in Supporting Information). To test whether decreased hepatic expression of these estrogenresponsive genes could lead to alterations in expression of down-
1658
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
stream metabolic targets, next, we measured the expression of Acox1 and Cyp7b1 (induced and repressed by PPARa signaling, respectively) and Edem1 and Erdj4 (targets of XBP1 endoplasmic reticulum stress pathway). Parous mice exhibited decreased levels of hepatic Acox1 and Edem1 and had increased hepatic Cyp7b1 compared with nulliparous mice (Supporting Information Figure S2). Parity did not affect hepatic Erdj4. However, in parous mice, HFD significantly increased hepatic Erdj4, indicating induction of ER stress. FEN-HFD prevented this increase in hepatic Erdj4 (Supporting Information Figure S2). Hepatic Esr1 expression was unchanged (not shown). Together, these gene expression alterations suggest that parous females have reduced estrogen signaling and may provide a mechanism for parous mice to develop obesity compared with nulliparous females.
Fenretinide can increase local estrogen signaling in WAT We examined the possibility that in addition to inducing RA signaling, FEN-HFD could increase estrogen signaling in WAT of parous
www.obesityjournal.org
Original Article
Obesity
OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
TABLE 2 Serum parameters associated with insulin sensitivity in parous female mice fed chow, HFD, or fenretinide (FEN)-HFD
Serum levels Glucose (mmol/l) Insulin (ng/ml) RBP4 protein (fold) ITT (blood glucose, mmol/l) Basal 30 min post-insulin bolus 60 min post-insulin bolus
Chow
HFD
FEN-HFD
9.6 6 0.5 2.4 6 0.4 1.0 6 0.08
11.08 6 1.1 2.8 6 0.4 1.1 6 0.09
10.6 6 1.0 nd 0.9 6 0.05
9.3 6 0.4 6.0 6 0.3 4.4 6 0.4
9.9 6 0.2 5.8 6 0.3 5.6 6 0.4
9.8 6 0.2 6.3 6 0.5 5.8 6 0.4
Serum glucose and serum insulin were measured following 5-h food removal at 12 weeks of diet and serum leptin and serum RBP4 at 27 weeks of diet, following overnight fasting and then 1 h re-feeding. Insulin tolerance test (ITT) (0.85 mU/g insulin) was performed following 5-h food removal at 25 weeks of diet in parous mice. Data represent the mean 6 SEM. Significance between groups was determined by one-way ANOVA and Bonferroni post hoc test. Serum insulin for FENHFD not determined (nd).
female mice. HFD decreased expression of Esr1, and FEN-HFD did not alter this down-regulation (Figure 5A). HFD trended to decrease ERb (Esr2) expression, whereas FEN-HFD trended to increase its expression (Figure 5B). Estrogen is synthesized locally by WAT
Figure 4 Decreased levels of estrogen target genes in liver of parous mice. Levels in nulliparous and parous are relative to respective chow group (n 5 4-8 for each group). There were no significant differences with diet in either group of female mice. Significance between nulliparous and parous mice was determined by t-test and marked **P < 0.01 or ***P < 0.001.
through the expression of synthetic enzymes, where the final two steps in this pathway are regulated by Hsd17b1 and cytochrome P450-enzyme aromatase (Cyp19a1). Gene expression of Hsd17b1 was not detectable in WAT. HFD did not affect Cyp19a1 expression; however, FEN-HFD significantly increased Cyp19a1 in parous mice (Figure 5B). The cytokine IL6 has been reported to stimulate aromatase activity in breast tissue, cells, and WAT (23,24). HFD trended to decrease IL6 expression; however, FEN-HFD significantly increased IL6 expression in parous mice (Figure 5). Estrogen has also been reported to induce vascular endothelial growth factor (VEGF) (25). HFD decreased Vegf expression, whereas FEN-HFD significantly increased Vegf expression in parous mice (Figure 5). The effects of FEN-HFD on estrogen signaling were only observed in parous females and not in nulliparous mice (data not shown). Thus, FENHFD seems to increase local estrogen signaling in WAT of parous female mice. Figure 3 Fenretinide can induce classic RA-responsive genes in WAT and liver in parous female mice. Expression levels of (A) Rarb and (B) Cyp26a1 in WAT and liver. Significant differences determined by one-way ANOVA and Bonferroni post hoc test for FEN-HFD compared with chow are marked *P < 0.05 or ***P < 0.001 and FEN-HFD compared with HFD are marked ##P < 0.01 or ###P < 0.001.
www.obesityjournal.org
Fenretinide can directly induce retinoid and estrogen signaling in 3T3-L1 adipocytes In vivo, WAT interacts with and receives signals from many other tissues to produce the responses measured earlier. To test whether
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1659
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
Figure 5 Fenretinide and HFD-induced obesity can independently alter estrogen signaling in WAT. Gene expression levels of estrogen receptors (A) Esr1 and (B) Esr2, estrogen synthetic enzyme (Cyp19a1), and targets of estrogen signaling (IL6 and Vegf) in WAT from parous mice. Data are relative to chow, and significance between groups (n 5 5-6 for each) was determined by one-way ANOVA and Bonferroni post hoc test. Significant differences (P < 0.05) for HFD 6 FEN compared with chow are marked * and FEN-HFD compared with HFD are marked #. Student’s t-test was used to calculate significant difference for HFD compared with chow for Vegf (P 5 0.02).
FEN could directly increase estrogen signaling, 3T3-L1 pre-adipocytes were differentiated into adipocytes for 8 days, followed by treatment with physiological concentrations of E2, RA, and FEN. Both FEN and RA increased the expression of Crbp1 and Rarg (Figure 6A,B). When combined, FEN1E2 or RA1E2 trended to additively increase the expression of Crbp1 (Figure 6A). E2 significantly increased Rarg and, when combined, RA1E2 additively increased the expression of Rarg (Figure 6B). FEN and FEN1E2 increased Esr2 expression similarly (Figure 6C). FEN also increased the expression of Cyp19a1 (Figure 6D). RA did not alter the expression of either Esr2 or Cyp19a1. Together, these results confirm the findings in WAT from parous mice and suggest that one of the beneficial mechanisms of FEN action in females may be through increasing estrogen signaling in WAT via Cyp19a1 and Esr2.
Discussion We recently demonstrated that FEN is a potent modulator of genes involved in retinoid and glucose/lipid homeostasis. These molecular alterations were in association with inhibiting HFD-induced obesity and insulin resistance in male mice (5). Females are partially protected from obesity and insulin resistance via the sex hormone estrogen. However, these prior studies have focused on postmenopausal women or an experimental model of menopause via ovariectomy (912). Our new study clearly demonstrates that FEN can decrease HFD-induced obesity in both nulliparous and parous female mice and that part of the mechanism may be through increased retinoid and estrogen signaling in WAT. Furthermore, this beneficial metabolic effect occurs despite a decrease in the targets of estrogen signaling and increased obesity in parous mice, a physiological relevant model of female obesity (19). A direct target of FEN seems to be leptin, the WAT-secreted protein. FEN inhibited HFD-induced hyperleptinemia in male (5) and female (Figure 2) mice and directly inhibited leptin mRNA by 75%
1660
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
in adipocytes (5). It is known that circulating leptin levels correlate strongly with WAT mass (Figure 2B), and increased levels in obesity are associated with resistance to the anorexic and thermogenic effects of leptin (26). However, in comprehensive studies in male mice (5,6), FEN did not have a measurable effect to decrease food intake or to increase oxygen consumption or Ucp1 expression. FEN treatment has been shown to improve insulin sensitivity and decrease serum leptin levels in overweight premenopausal women (27). However, the parity status of the women was not reported, nor has FEN treatment been examined in postmenopausal women at higher risk of developing obesity and metabolic complications. Our data suggest that future clinical studies should take this into account. We found that parous mice had significantly down-regulated genes that have been shown to be direct targets of estrogen signaling (e.g., Igf1 and Rara). This suggests that similar to ovariectomy-induced obesity in females, decreased circulating estrogen may be the primary mechanism for increased adiposity in parous mice (on chow diet or HFD). The beneficial effects of estrogen on energy and glucose homeostasis are largely mediated by Esr1 (13-15). However, hepatic Esr1 signaling is not required for the effects of E2 treatment on adiposity (28). Interestingly, another target of FEN seems to involve increased local production of estrogen in WAT, via up-regulation of adipose Cyp19a1 and signaling through induction of WAT Esr2 (Figures 4 and 5, WAT and 3T3-L1 adipocytes, respectively). This is in contrast to the liver, which is wholly responsive to circulating levels of estrogen because enzymes involved in estrogen production are expressed at very low levels (if at all). Mice of both sexes lacking Cyp19a1 develop obesity (29). However, any beneficial effects of Esr2 activation are not clear because HFD-fed Esr2-KO mice demonstrate increased obesity, but improved insulin sensitivity and glucose tolerance (30). In contrast, Esr2-selective ligands have been shown to alleviate HFD- and ovariectomy-induced obesity in mice (31). Moreover, FEN-induced induction of Vegf in WAT may be linked to an increase in local estrogen signaling and may contribute to the beneficial effects of FEN treatment because WAT
www.obesityjournal.org
Original Article
Obesity
OBESITY BIOLOGY AND INTEGRATED PHYSIOLOGY
metabolic homeostasis in insulin-sensitive tissues remains to be elucidated. There were some potential issues with our study design, largely because nulliparous mice were the offspring of the parous mice and, therefore, 13 weeks younger. To minimize the impact of this age difference, all mice were aged for 26 weeks to achieve stable midlife body weight. Parous mice started the study 6 g heavier than nulliparous mice, and nulliparous mice did not gain this difference in weight over the duration of the study. Moreover, body composition analysis confirmed that parous mice were more obese, independent of age (Table 1). Indeed, others have previously demonstrated in mice using age-matched controls that parity leads to increased adiposity independent of maternal age (19). In addition, female mice undergo reproductive senescence much later in life or with major genetic, surgical, or pharmacologic intervention (35,36). Thus overall, our findings are unlikely to be due to the relatively small difference in middle age between nulliparous and parous mice. There is also the possibility that variation in circulating hormone concentrations during stages of the estrous cycle and thereby at the end of the study may have affected gene expression. We did not determine stages of the estrous cycle by visual observation or vaginal cytology, but we assumed that estrous cycles will have been primarily synchronized and suppressed by the presence of male mice in the room and because both parous and nulliparous were housed in the same room (37,38).
Figure 6 Fenretinide increases classic retinoid signaling and estrogen signaling via Esr2 and Cyp19a1 in adipocytes. Differentiated 3T3-L1 adipocytes were utilized to determine gene expression changes in response to FEN or RA 6 E2 treatments. 3T3-L1 cultures were differentiated with MDI alone for 8 days and then subjected to the indicated treatments for an additional 8 days (1 lM FEN, 0.1 lM RA, and 1 nM E2). Relative expression levels of (A,B) retinoid genes or (C,D) estrogen genes observed after treatment (n 5 4 separate experiments on different plates of cells). Significance between groups was determined by two-way ANOVA. There was a significant effect (P < 0.05 overall) of retinoid treatment for both RA and FEN vs. VEH on levels of Crbp1 and Rarg. There was a significant effect of E2 treatment on Rarg levels. Post hoc tests: Holm-Sidak method for pair-wise multiple comparisons determined significance between retinoid treatment groups (RA, FEN, and VEH) within 6 E2 treatment groups are marked *P < 0.05, **P < 0.01, or ***P < 0.001 for comparisons with VEH or #P < 0.05 or ###P < 0.001 for comparison with RA. †† marks significance between 6 E2 treatment for within VEH or RA treatment. For some analyses, Student’s t-test was performed: Rarg was increased by E2 alone (t-test, P 5 0.003, marked ††) and by FEN alone (t-test, P 5 0.02, marked *), and Crbp1 was increased by FEN alone (t-test, P 5 0.03, marked *) compared with VEH without E2.
Although our sample sizes seemed to be sufficient to obtain significant differences in adiposity parameters and retinoid signaling with FEN treatment, many molecular markers of metabolic disturbances were not detected. Interestingly, in our other HFD studies, an alternate source of fat (hydrogenated vegetable fat rich in trans fats) did cause significantly more insulin resistance in female mice (data not shown). This suggests that there is a complex interplay between types of diet/dietary fats consumed, the protective effects of E2 signaling, and alterations in glucose/lipid homeostasis. Childbearing and increasing parity is associated with future development of maternal obesity. Moreover, obesity is a major risk factor for diabetes during pregnancy (gestational diabetes mellitus), which is strongly linked to an increase in maternal and offspring diabetes risk (39,40). There is an urgent need for novel therapies that are efficient and non-toxic to treat obesity and insulin resistance. Recent studies have shown that retinoid homeostasis and estrogen signaling are potential therapeutic targets for preventing or reducing obesity. Here, we present evidence that FEN is a modulator of estrogen and retinoid homeostasis genes in association with prevention of female postpregnancy and HFD-induced obesity.O
overexpression of VEGF protects against diet-induced obesity and insulin resistance (25,32).
Acknowledgments Rara is an estrogen-induced gene, and its expression in breast tumors has been shown to correlate with ER expression (33). More recent studies have implicated Rara as an essential component of the ER complex that can cooperate for effective transcriptional activity in breast cancer cells (34). Our data also suggest an interaction between estrogen and FEN/retinoid signaling in adipocytes. The details of this interaction and its potential importance in maintaining
www.obesityjournal.org
The authors thank J.R. Speakman (UoA) for invaluable discussion regarding energy expenditure, T. Martin (Johnson & Johnson, NJ) and U. Thurneer (Cilag AG, Switzerland) for fenretinide, to use completely without restriction or obligation, and undergraduate student K. Towle (UoA) for assisting in experiments. C 2015 The Obesity Society V
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
1661
Obesity
Fenretinide Prevents Obesity in Female Mice Shearer et al.
References 1. Yang Q, Graham TE, Mody N, et al. Serum retinol binding protein 4 contributes to insulin resistance in obesity and type 2 diabetes. Nature 2005;436:356-362. 2. Graham TE, Yang Q, Bluher M, et al. Retinol-binding protein 4 and insulin resistance in lean, obese, and diabetic subjects. N Engl J Med 2006;354:2552-2563. 3. Gavi S, Stuart LM, Kelly P, et al. Retinol-binding protein 4 is associated with insulin resistance and body fat distribution in nonobese subjects without type 2 diabetes. J Clin Endocrinol Metab 2007;92:1886-1890. 4. McTernan PG, Kumar S. Editorial: retinol binding protein 4 and pathogenesis of diabetes. J Clin Endocrinol Metab 2007;92:2430-2432. 5. Mcilroy GD, Delibegovic M, Owen C, et al. Fenretinide treatment prevents dietinduced obesity in association with major alterations in retinoid homeostatic gene expression in adipose, liver, and hypothalamus. Diabetes 2013;62:825-836. 6. Preitner F, Mody N, Graham TE, Peroni OD, Kahn BB. Long-term fenretinide treatment prevents high-fat diet-induced obesity, insulin resistance, and hepatic steatosis. Am J Physiol Endocrinol Metab 2009;297:E1420-E1429. 7. Rotmensz N, De Palo G, Formelli F, et al. Long-term tolerability of fenretinide (4HPR) in breast cancer patients. Eur J Cancer 1991;27:1127-1131. 8. Zanardi S, Serrano D, Argusti A, Barile M, Puntoni M, Decensi A. Clinical trials with retinoids for breast cancer chemoprevention. Endocr Relat Cancer 2006;13:51-68. 9. Carr MC. The emergence of the metabolic syndrome with menopause. J Clin Endocrinol Metab 2003;88:2404-2411. 10. Kanaya AM, Herrington D, Vittinghoff E, et al. Glycemic effects of postmenopausal hormone therapy: the heart and estrogen/progestin replacement study. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 2003;138:1-9. 11. Margolis KL, Bonds DE, Rodabough RJ, et al. Effect of oestrogen plus progestin on the incidence of diabetes in postmenopausal women: results from the women’s health initiative hormone trial. Diabetologia 2004;47:1175-1187. 12. Wagner JD, Thomas MJ, Williams JK, Zhang L, Greaves KA, Cefalu WT. Insulin sensitivity and cardiovascular risk factors in ovariectomized monkeys with estradiol alone or combined with nomegestrol acetate. J Clin Endocrinol Metab 1998;83:896901. 13. Heine PA, Taylor JA, Iwamoto GA, Lubahn DB, Cooke PS. Increased adipose tissue in male and female estrogen receptor-alpha knockout mice. Proc Natl Acad Sci U S A 2000;97:12729-12734. 14. Ohlsson C, Hellberg N, Parini P, et al. Obesity and disturbed lipoprotein profile in estrogen receptor-alpha-deficient male mice. Biochem Biophys Res Commun 2000; 278:640-645. 15. Ribas V, Nguyen MT, Henstridge DC, et al. Impaired oxidative metabolism and inflammation are associated with insulin resistance in ERalpha-deficient mice. Am J Physiol Endocrinol Metab 2010;298:E304-E319. 16. Gunderson EP, Jacobs DR Jr, Chiang V, et al. Childbearing is associated with higher incidence of the metabolic syndrome among women of reproductive age controlling for measurements before pregnancy: the CARDIA study. Am J Obstet Gynecol 2009;201:177.e1-177.e9. 17. Hajiahmadi M, Shafi H, Delavar MA. Impact of parity on obesity: a cross-sectional study in Iranian women. Med Princ Pract 2015;24:70-74. 18. Bernstein L, Pike MC, Ross RK, Judd HL, Brown JB, Henderson BE. Estrogen and sex hormone-binding globulin levels in nulliparous and parous women. J Natl Cancer Inst 1985;74:741-745. 19. Rebholz SL, Jones T, Burke KT, et al. Multiparity leads to obesity and inflammation in mothers and obesity in male offspring. Am J Physiol Endocrinol Metab 2012;302:E449-E457.
1662
Obesity | VOLUME 23 | NUMBER 8 | AUGUST 2015
20. Mody N, Graham TE, Tsuji Y, Yang Q, Kahn BB. Decreased clearance of serum retinol-binding protein and elevated levels of transthyretin in insulin-resistant ob/ob mice. Am J Physiol Endocrinol Metab 2008;294:E785-E793. 21. Fanjul AN, Delia D, Pierotti MA, Rideout D, Yu JQ, Pfahl M. 4-Hydroxyphenyl retinamide is a highly selective activator of retinoid receptors. J Biol Chem 1996; 271:22441-22446. 22. Sabichi AL, Hendricks DT, Bober MA, Birrer MJ. Retinoic acid receptor beta expression and growth inhibition of gynecologic cancer cells by the synthetic retinoid N-(4-hydroxyphenyl) retinamide. J Natl Cancer Inst 1998;90:597-605. 23. Reed MJ, Coldham NG, Patel SR, Ghilchik MW, James VH. Interleukin-1 and interleukin-6 in breast cyst fluid: their role in regulating aromatase activity in breast cancer cells. J Endocrinol 1992;132:R5-R8. 24. Zhao Y, Nichols JE, Bulun SE, Mendelson CR, Simpson ER. Aromatase p450 gene expression in human adipose tissue. Role of a jak/STAT pathway in regulation of the adipose-specific promoter. J Biol Chem 1995;270:16449-16457. 25. Koos RD. Mini review: putting physiology back into estrogens’ mechanism of action. Endocrinology 2011;152:4481-4488. 26. Considine RV, Sinha MK, Heiman ML, et al. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 1996;334:292295. 27. Johansson H, Gandini S, Guerrieri-Gonzaga A, et al. Effect of fenretinide and lowdose tamoxifen on insulin sensitivity in premenopausal women at high risk for breast cancer. Cancer Res 2008;68:9512-9518. 28. Zhu L, Brown WC, Cai Q, et al. Estrogen treatment after ovariectomy protects against fatty liver and may improve pathway-selective insulin resistance. Diabetes 2013;62:424-434. 29. Jones ME, Thorburn AW, Britt KL, et al. Aromatase-deficient (ArKO) mice have a phenotype of increased adiposity. Proc Natl Acad Sci U S A 2000;97:12735-12740. 30. Foryst-Ludwig A, Clemenz M, Hohmann S, et al. Metabolic actions of estrogen receptor beta (ERbeta) are mediated by a negative cross-talk with PPARgamma. PLoS Genet 2008;4:e1000108. 31. Yepuru M, Eswaraka J, Kearbey JD, et al. Estrogen receptor-{beta}-selective ligands alleviate high-fat diet- and ovariectomy-induced obesity in mice. J Biol Chem 2010;285:31292-31303. 32. Elias I, Franckhauser S, Ferre T, et al. Adipose tissue overexpression of vascular endothelial growth factor protects against diet-induced obesity and insulin resistance. Diabetes 2012;61:1801-1813. 33. Roman SD, Ormandy CJ, Manning DL, et al. Estradiol induction of retinoic acid receptors in human breast cancer cells. Cancer Res 1993;53:5940-5945. 34. Ross-Innes CS, Stark R, Holmes KA, et al. Cooperative interaction between retinoic acid receptor-alpha and estrogen receptor in breast cancer. Genes Dev 2010;24:171-182. 35. Danilovich N, Babu PS, Xing W, Gerdes M, Krishnamurthy H, Sairam MR. Estrogen deficiency, obesity, and skeletal abnormalities in follicle-stimulating hormone receptor knockout (FORKO) female mice. Endocrinology 2000;141:42954308. 36. Danilovich N, Ram Sairam M. Recent female mouse models displaying advanced reproductive aging. Exp Gerontol 2006;41:117-122. 37. Whitten WK. Modification of the oestrous cycle of the mouse by external stimuli associated with the male. J Endocrinol 1956;13:399-404. 38. Whitten WK. Occurrence of anoestrus in mice caged in groups. J Endocrinol 1959; 18:102-107. 39. Chu SY, Callaghan WM, Kim SY, et al. Maternal obesity and risk of gestational diabetes mellitus. Diabetes Care 2007;30:2070-2076. 40. Lawlor DA, Relton C, Sattar N, Nelson SM. Maternal adiposity-a determinant of perinatal and offspring outcomes? Nat Rev Endocrinol 2012;8:679-688.
www.obesityjournal.org
