Silencing steroid receptor coactivator-1 in the nucleus of the solitary tract reduces estrogenic effects on feeding and apolipoprotein A-IV expression

We previously found that 17β-estradiol (E2) stimulates apolipoprotein A-IV (apoA-IV) gene expression in the nucleus of the solitary tract (NTS) of lean ovariectomized (OVX) rodents. Here we report that in the NTS of high-fat diet-induced obese (DIO) rats, the apoA-IV mRNA level is significantly reduced and that the estrogenic effects on apoA-IV gene expression and food intake are impaired. E2 regulates apoA-IV gene expression through its nuclear receptor α (ERα), which requires co-activators, such as steroid receptor coactivator-1 (SRC-1), to facilitate the transcription of targeted genes. Interestingly, SRC-1 gene expression is significantly reduced in DIO OVX rats. SRC-1 is colocalized with apoA-IV in the cells of the NTS and E2 treatment enhances the recruitment of ERα and SRC-1 to the estrogen response element at the apoA-V promoter, implying the participation of SRC-1 in E2's stimulatory effect on apoA-IV gene expression. Using small hairpin RNA (shRNA), which was validated in cultured neuronal cells, we found that SRC-1 gene knockdown specifically in the NTS significantly diminished E2's anorectic action, leading to increased food intake and body weight. More importantly, the stimulatory effect of E2 on apoA-IV gene expression in the NTS was significantly attenuated in SRC-1 knockdown rats. These results collectively demonstrate the critical roles of NTS SRC-1 in mediating E2's actions on food intake and apoA-IV gene expression and suggest that reduced levels of endogenous SRC-1 and apoA-IV expression are responsible for the impaired E2's anorectic action in obese females.

E2 exerts its catabolic action indirectly through enhancing the strength of other signals implicated in the direct control of food intake and energy expenditure (1), including the satiation factor apoA-IV. Ovariectomy (OVX) in female rodents reduces apoA-IV gene expression in the nucleus tractus solitarius (NTS), and this is reversed by cyclic E2 replacement. Additionally, E2 enhances the satiating potency of apoA-IV (2), and more importantly, compared with OVX wildtype controls, OVX apoA-IV knock-out (KO) mice have a blunted reduction of feeding in response to E2, thereby eating significantly more food and gaining more body weight during the period of cyclic E2 replacement (2). These data collectively suggest that increased endogenous apoA-IV signaling may partially mediate E2-induced inhibition of feeding in lean females. Whether or not these effects generalize to high-fat diet (HFD)-induced obesity is unknown.
Obesity is a growing global health problem. One wellestablished risk factor predisposing to obesity is the amount of fat in the diet. Epidemiological studies have identified a significant positive correlation between average dietary fat intake and the incidence of obesity (3,4). Recent data suggest that obese animals have blunted satiety, raising the possibility that defective signaling of apoA-IV and an impaired response of apoA-IV gene expression to E2 may contribute to the etiology of obesity in females. In the present study, we first addressed the feasibility of these hypotheses in an established rat model of HFD-induced obesity (5), and then asked whether the estrogenic effects on apoA-IV gene expression and energy intake are attenuated in the OVX rats after chronic HFD consumption.
Our previous studies demonstrated that E2 regulates apoA-IV gene expression through estrogen receptor ␣ (ER␣) (6). E2-activated ER␣ may recruit steroid receptor coactivator-1 (SRC-1), a nuclear receptor coactivator, to DNA promotion sites to facilitate the transcription of targeted genes (7). Interestingly, SRC-1 KO mice are sensitive to diet-induced obesity (8), and the effects of E2 replacement on body weight, food intake, and energy expenditure are blunted in OVX SRC-1 KO mice (9), suggesting that SRC-1 is required for E2's catabolic actions. We therefore performed in vitro and in vivo studies to determine whether SRC-1 signaling is necessary to mediate E2 effects on apoA-IV gene expression and food intake using lentiviral-mediated small hairpin RNA (shRNA), which was specifically delivered into the NTS.

Chronic consumption of HFD increases food intake and body weight in OVX rats
Adult female rats received bilateral OVX through a midline abdominal incision, as we reported previously (2). One week later, the OVX rats were divided into 3 groups. One group was fed a low-fat diet (LFD) and another was fed a HFD for 10 weeks. Because HFD and LFD rats consumed different amounts of energy each day, a 3rd group of rats, called pair HFD-fed (PHF) rats, received the HFD each day, but in an amount limited to the average daily caloric consumption of the rats fed the LFD ad libitum, as we reported previously (10). Food intake, body weight, and body composition assessed by EchoMRI were recorded weekly. During week 10, blood was collected, and plasma was assayed for leptin and insulin by ELISA kits.
As depicted in Fig. 1A, body weight of the 3 groups of OVX rats differed beginning at 3 weeks. By week 10, HFD-fed rats were significantly heavier than the other two groups (p Ͻ 0.05). As expected, PHF rats weighed significantly less than HFD-fed rats (p Ͻ 0.05), and were not different from the LFD-fed rats. The HFD-fed rats consumed significantly more daily energy (p Ͻ 0.05, Fig. 1B) and had significantly more body weight and fat mass than LFD-fed rats (p Ͻ 0.01, Fig. 1C). Consistent with previous reports (11), HFD-fed OVX rats had elevated plasma leptin and insulin (p Ͻ 0.05), relative to the other two groups (Fig. 1, D and E).

ApoA-IV mRNA level is significantly reduced in the NTS of DIO OVX rats
To determine whether the apoA-IV mRNA level and the response to E2 are altered after chronic HFD consumption, half the rats (n ϭ 8) in each group above then received subcutaneous injection of E2 (2 g) in 100 l of vehicle every 4th day (mimicking the ovarian cycle of female rats) for 2 cycles, and the other half received vehicle (100 l of sesame oil) for the same period. The dose of 2 g of E2 was chosen because it produced plasma E2 levels similar to peak levels occurring during the ovarian cycle in intact rats (12). We selected 2 cycles of E2 treatment because, within such a short period, E2 reduces food intake, but no change of body weight in LFD-fed rats, as we reported previously (2,12). The injections occurred at 10 a.m. (lights off from 4 p.m. to 4 a.m.). The rats were fasted for 4 h before being sacrificed at the onset of dark on the 2nd day after the final injections, based on the time when E2 exerts its most robust anorexigenic effects in OVX rats (12,13). The brains were dissected, and the NTS was micropunched for determining apoA-IV mRNA levels by quantitative real-time PCR (qPCR) (2,14).
Relative to the LFD-fed group, HFD-fed rats receiving vehicle (oil) had a significant reduction of apoA-IV mRNA level. E2 treatment stimulated apoA-IV gene expression in LFD-fed, but not in HFD-fed, rats ( Fig. 2A). Importantly, E2 treatment significantly increased apoA-IV gene expression in PHF-fed rats ( Fig. 2A), implying that it is the extra calories consumed, but not the HFD itself, which is responsible for the reduction of apoA-IV gene expression.

DIO OVX rats have a diminished response in food intake to acute cyclic E2 treatment
To determine whether the reduced apoA-IV expression in the NTS is associated with altered energy intake in the DIO OVX rats, new cohorts of OVX rats were fed HFD and LFD, respectively, for 10 weeks. They were then treated with E2 or oil for 2 cycles as described above. On the 2nd day after the last injections, the rats were fasted for 4 h before dark and their 24-h food intake and body weight were measured. Compared with oil-treated controls, two cycles of E2 treatment significantly suppressed food intake in the LFD-OVX, but not in the DIO OVX rats (Fig. 2B), suggesting an impaired response in food intake to the acute E2 treatment in DIO rats. There was no significant difference in body weight between E2-and oiltreated rats either on LFD or HFD (data not shown).

Chronic HFD consumption significantly reduces SRC-1 gene expression in the NTS of OVX rats
To elucidate the underlying molecular mechanisms, we assayed SRC-1 as a possible candidate mediator. Vehicletreated HFD-fed rats had significantly lower SRC-1 mRNA than LFD-fed controls. Interestingly, cyclic E2 treatment did not significantly affect SRC-1 gene expression in either LFD-fed or HFD-fed rats (Fig. 2C). No significant difference was found in SRC-1 mRNA levels between LFD-fed rats and PHF-fed rats with either oil or E2 treatment (Fig. 2C).

Colocalization of SRC-1 with apoA-IV in the NTS
We next asked if SRC-1 is colocalizated with apoA-IV in the NTS. Four OVX rats were deeply anesthetized and perfused, as we have reported previously (2,15). Coronal brainstem sections were cut through the NTS at 30 m and dual-labeling immunohistochemistry was conducted with anti-SRC-1 and apoA-IV antibodies to localize SRC-1 and apoA-IV expression. Both antibodies were validated previously (9,15,16). SRC-1 signals were identified as red fluorescence (Fig. 3A), and apoA-IV signals were recognized as green fluorescence (Fig. 3B). The immunohistochemical staining revealed that SRC-1 is colocalized with apoA-IV NTS cells (Fig. 3C), providing an anatomical Figure 1. Alteration of metabolic parameters in DIO OVX rats. HFD-fed OVX rats had significantly increased body weight (A), daily energy intake (B), and fat mass (C), compared with LFD-fed OVX controls. Chronic HFD consumption also significantly increased plasma leptin and insulin levels (D and E). Data represent mean Ϯ S.E., n ϭ 8. *, p Ͻ 0.05; **, p Ͻ 0.01, compared with LFD-fed rats.

Estrogen up-regulates central apoA-IV via SRC-1
basis for the participation of SRC-1 in E2's regulatory effect on apoA-IV gene expression.

E2 enhances the recruitment of ER␣ and SRC-1 to the estrogen response element (ERE) at the apoA-IV promoter
Although SRC-1 has been reported to form a protein complex with ER␣ in the hypothalamus in an in vitro study (17), it is unknown whether ER␣ and SRC-1 interact with the ERE located at the 5Ј-upstream region of the apoA-IV promoter in the NTS. We therefore examined this possibility using a sequential biotinylated double-stranded DNA pulldown assay and immunoblot approach, as described previously (18 -20). Fourteen OVX rats were divided into two groups (n ϭ 7), one receiving E2 (2 g, subcutaneously) and the other receiving vehicle (sesame oil, 100 l, subcutaneously) for 2 cycles. On the 2nd day after the last injections, the rats were sacrificed at the onset of dark and the NTS was micropunched.
Two 5Ј-biotinylated double-stranded 51-mer DNAs that represent apoA-IV promoter sequence at 5645 to 5696 were synthesized, one containing an ERE, and the other containing a mutated ERE (6). These double-stranded DNAs were incubated with protein extracts from the NTS for 2 h at 4°C with constant rotation. The protein-DNA complexes were then immobilized to Streptavidin Dynabeads and separated with a Dynal magnet, according to the manufacturer's instructions (18 -20). The beads were washed, resuspended, and boiled in SDS sample buffer for Western blot (6). As depicted in Fig. 4, E2 treatment significantly enhanced the recruitment of ER␣ and SRC-1 to the ERE located at the apoA-IV promoter, compared with vehicle treatment. No ER␣ or SRC-1 proteins were detected from the protein extract of the NTS incubated with the biotinylated DNA containing a mutated ERE (Data not shown).

shRNA inhibits SRC-1 expression in cultured neuronal cells
To manipulate SRC-1 expression, we used short-hairpin RNA (shRNA), a powerful tool for gene knockdown. From the Viral Vector Core at the University of South Carolina, we obtained three different lentiviral particles with packaged shR-NAs that were predicted to target SRC-1 mRNA, as well as a control lentiviral particle with an empty vector. These particles were transduced into immortalized neuronal cells (CLU213), which express ER␣, SRC-1, and apoA-IV genes (6). After 48 h, the transduced cells were selectively grown in a medium containing puromycin (5 g/ml) for an additional 2 weeks. The cells were treated with E2 (10 nM) for an addi- A, under oil treatment, the apoA-IV mRNA level was significantly reduced in the NTS of HFD-fed OVX rats, compared with LFD-fed OVX rats. E2 treatment increased apoA-IV gene expression in both LFD-fed and PHF-fed OVX rats, but not in HFD-fed OVX rats. B, comparison of energy intake between LFD-and HFD-fed OVX rats after 2 cycles of E2 or oil treatment. E2 treatment significantly suppressed food intake in LFD-fed OVX, but not in HFD-fed OVX rats, compared with that in oil-treated controls. C, HFD-fed OVX rats with oil treatment also had significantly reduced SRC-1 mRNA levels, compared with LFD-fed and oil-treated OVX rats. E2 treatment did not significantly affect SRC-1 mRNA levels in either HFD-and LFD-fed OVX rats. No significant difference in SRC-1 gene expression was found between LFD-fed and PHF-fed OVX rats. Data represent mean Ϯ S.E., n ϭ 7-8. *, p Ͻ 0.05, compared with LFD-fed OVX rats with oil treatment; #, p Ͻ 0.05, compared with E2-treated LFD-fed rats; and &, p Ͻ 0.05, compared with oil-treated PHF-fed OVX rats.  . E2 treatment increases the recruitment of ER␣ and SRC1 to the ERE of the apoA-IV promoter. OVX rats received E2 or vehicle treatment for 2 cycles. The NTS was micropunched. Two biotinylated double-stranded DNAs containing the ERE and mutant ERE, respectively, of the apoA-IV promoter were incubated with protein extracts from the NTS. The protein-DNA complexes were then immobilized to Streptavidin Dynabeads and separated with a Dynal magnet. The beads were washed, resuspended, and boiled in SDS sample buffer. ER␣ and SRC-1 proteins were detected by immunoblotting. Mean Ϯ S.E., n ϭ 6 -7. *, p Ͻ 0.05, compared with oiltreated OVX rats.

Estrogen up-regulates central apoA-IV via SRC-1
tional 24 h, and then lysed for measuring SRC-1 and apoA-IV mRNA and protein levels by qPCR and Western blotting, respectively.
Among the three SRC-1 shRNAs, the one with the V3LMM_453913 lentiviral vector was selected because of its highest efficiency in the knockdown of SRC-1 expression. Compared with control shRNA (named shCTL), this SRC-1 shRNA (named shSRC-1) significantly reduced SRC-1 mRNA levels in the neuronal cells (by 75.8%, p Ͻ 0.01, Fig. 5B), consistent with the changes in SRC-1 protein level (69.6% reduction, p Ͻ 0.01, Fig. 5E). No significant difference was found for SRC-1 mRNA or protein levels between shCTL-transduced cells and untreated neuronal cells.
To determine whether E2's regulation of apoA-IV expression requires SRC-1, we further measured apoA-IV mRNA and protein expression in lentiviral shRNAs-transduced neuronal cells by qPCR and Western blotting, respectively. ApoA-IV mRNA and protein levels were reduced by 65.2 and 59.2%, respectively (p Ͻ 0.01), compared with those in shCRL-transduced cells (Fig. 5, C and E). No significant difference was found in apoA-IV mRNA or protein levels between shCTL-transduced cells and untreated neuronal cells. These results indicate that SRC-1 is, at least partially, required for E2's stimulation of apoA-IV expression.

Lentivirus-mediated knockdown of SRC-1 expression in the NTS
To determine the impact of SRC-1 knockdown specifically in the NTS on estrogenic actions on feeding and apoA-IV gene expression, OVX rats received bilateral intra-NTS injections of shSRC-1 or shCTL. After a 7-day recovery, the rats with each lentiviral treatment were divided into two subgroups, one receiving E2 (2 g, subcutaneously), and the other receiving oil (100 l, subcutaneously) every 4th day for 10 cycles. Body weight and food intake were monitored on the 2nd day after each cycle of E2 or oil treatment. On the 2nd day after the 9th cycle of E2 or oil treatment, the rats were monitored for energy expenditure, as we reported previously (16,21).
On the 2nd day after the final E2 or oil treatment, 4-h fasted rats were sacrificed and the brains quickly dissected. The brainstem containing the NTS, from Ϫ11.8 to Ϫ14.1 mm posterior to bregma (22), was sectioned at consecutive intervals of 15 m (for examining GFP expression in the NTS) and then 200 m (for the NTS micropunch). All micropunches were snap-frozen in dry ice, prepared for RNA extraction, and the SRC-1 mRNA level was measured by qPCR.
Animals with correct green fluorescent protein (GFP) location and successful SRC-1 gene knockdown were included in the data analyses. Fig. 6A is a schematic drawing that depicts the location of the NTS of rat brain. Fig. 6B is a representative micrograph documenting the expression of GFP of the lentiviral shNRA in the NTS as determined under fluorescence microscopy. SRC-1 mRNA levels were significantly reduced (by 47.8%) in NTS micropunches of OVX rats receiving shSRC-1, relative to their controls receiving shCTL (Fig. 6C). E2 treat-  . SRC-1 gene knockdown specifically in the NTS of OVX rats. A, a schematic drawing for depicting the location in the NTS of rat brain. B, a representative micrograph documenting tGFP expression of the lentiviral shNRA in the NTS. C, SRC-1 mRNA levels were significantly reduced in the NTS of rats receiving shSRC-1, relative to that of control rats receiving shCTL. E2 treatment did not significantly affect SRC-1 mRNA levels in rats receiving either shSRC-1 or shCTL. D, SRC-1 knockdown also significantly reduced the apoA-IV mRNA level in the NTS, compared with that in rats receiving shCTL. E2 treatment stimulated apoA-IV gene expression only in OVX rats receiving shCTL, but not in SRC-1 knockdown OVX rats. Mean Ϯ S.E. n ϭ 6 -7, *, p Ͻ 0.05, compared with oil-treated and shCTL-transduced rats, and #, p Ͻ 0.05, compared with E2-treated and shCTL-transduced rats.

SRC-1 knockdown attenuates E2's effect on apoA-IV gene expression in the NTS
If SRC-1 in the NTS underlies E2's stimulatory effect on apoA-IV gene expression through enhancing the efficiency of ER␣-activated by E2, gene knockdown of SRC-1 specifically in the NTS of OVX rats should cause them to be less responsive to E2, compared with their controls. Consistent with this hypothesis, the apoA-IV mRNA level was significantly reduced in SRC-1 knockdown rats relative to that in shCTL-transduced rats (p Ͻ 0.05, Fig. 6D). E2 treatment stimulated apoA-IV gene expression in shCTL-transduced (p Ͻ 0.05), but not in shSRC-1-transduced, rats (Fig. 6D).

E2's anorectic action is blunted in rats with SRC-1 knockdown in the NTS
It has been reported that a major site for E2 to suppress food intake is the NTS (23)(24)(25)(26)(27). Consistent with that, in the OVX rats receiving shCTL, daily food intake was significantly reduced after cyclic treatment of E2, when compared with oil-treated controls (Fig. 7A). However, this anorectic action of E2 was diminished in SRC-1 knockdown OVX rats; i.e. no difference in food intake between E2 and the oil-treated group was found at any time point (Fig. 7B). When comparing food intake on the 2nd day after the 8th cycle of E2 or oil treatment (Fig. 7C), a significant difference between SRC-1 knockdown rats and the rats receiving shCTL was found only after E2 treatment (solid bars), but not after oil treatment (open bars).
Consistent with the changes in food intake, oil-treated OVX rats receiving shCTL had markedly increased body weight, whereas body weight gain was attenuated after cyclic E2 treatment (Fig. 7D). Although the growth curve of body weight in SRC-1 knockdown rats was comparable with that in control rats receiving shCTL under oil treatment, cyclic E2 replacement caused much less body weight reduction in SRC-1 knock-down OVX rats (Fig. 7E). On the 2nd day after the 8th E2 treatment, the SRC-1 knockdown rats were significantly heavier than that in control rats receiving shCTL (solid bars) (Fig. 7F).

SRC-1 knockdown in the NTS does not affect energy expenditure
Considering that the E2-induced reduction in body weight gain is partly due to an increase of energy expenditure, which occurs mainly through E2's action at the hypothalamus (28), we also measured energy expenditure. Consistent with previous reports (29,30). However, no significant difference occurred between shSRC-1-and shCTL-transduced OVX rats under either oil or E2 treatments (Fig. 8, A-D). Additionally, no difference in respiratory quotient occurred among the four groups (data not shown). These observations indicate that SRC-1 knockdown in the NTS specifically affects the estrogenic effect on food intake, but not on energy expenditure, leading to a reduction in body weight.

Discussion
In the present studies, we found that the apoA-IV mRNA level is significantly reduced in the NTS of HFD-induced obese (DIO) OVX rats. Additionally, the effects on apoA-IV gene expression and the reduction of food intake induced by cyclic E2 replacement, which mimics what occurs during the ovarian cycle of female rats (2), was significantly blunted in those DIO rats with OVX. To elucidate the mechanisms underlying the impaired E2 effect on apoA-IV expression after chronic HFD consumption, we found that SRC-1, a co-activator of the ER␣, is colocalized with apoA-IV in NTS cells and that cyclic E2 treatment enhanced the recruitment of ER␣ and SRC-1 to the ERE of the apoA-IV promoter. More importantly, the SRC-1 mRNA level was significantly diminished in the NTS of DIO OVX rats. Using shRNA technology, SRC-1 knockdown significantly reduced the stimulatory effect of E2 on apoA-IV expression in cultured neuronal cells. In OVX rats with shRNA-induced SRC-1 knockdown specifically in the NTS, the effects of cyclic The comparison of daily food intake (C) and body weight (F) was conducted on the 2nd day after the 8th cyclic of E2 or oil treatment in OVX rats. Data are presented as mean Ϯ S.D., n ϭ 6 -7/group. *, p Ͻ 0.05; **, p Ͻ 0.01, compared with oil-treated OVX rats at the same time points; #, p Ͻ 0.05, compared with E2-treated OVX rats receiving shCTL.

Estrogen up-regulates central apoA-IV via SRC-1
E2 replacement on the reduction of food intake and body weight, as well as the stimulation of apoA-IV gene expression, were significantly attenuated. These findings collectively demonstrate that SRC-1 has an important role in mediating E2's actions on feeding as well as upon apoA-IV gene expression.
Obesity is a major health problem worldwide. Recent data suggest that obese animals have blunted satiety, raising the possibility that defective signaling of apoA-IV and impaired response of apoA-IV gene expression to E2 may contribute to the etiology of obesity in females. Consistent with this hypothesis, we found that apoA-IV gene expression in the NTS was significantly decreased in HFD-fed OVX rats, compared with that in either LFD-or HFP-fed OVX rats. More importantly, the stimulatory effect of cyclic E2 replacement on apoA-IV gene expression was significantly attenuated.
E2 has potent suppressive effects on food intake and body weight (26). Normal intact female rats eat different amounts of food across their 4-day ovarian cycles. Specifically, the females eat the least during estrus, which occurs right after the preovulatory rise in E2 secretion, and the most during diestrus when E2 levels are lower (12). Disruption of ovarian cycling in rats through OVX leads to increased food intake and body weight gain. However, a cyclic regimen of E2 replacement at a physiologic dose abolishes OVX-induced hyperphagia and normalizes body weight to the levels of gonadal intact rats (2).
E2 acts within the brain to decrease meal size (1), and compelling evidence indicates that the NTS is the critical brain site for E2's anorexigenic action (23,25,27). ER␣ in the dorsal hindbrain is predominately localized in the NTS (31). Administration of a low dose of E2 onto the surface of the hindbrain directly over the caudal NTS significantly decreased food intake and activated ER␣-containing neurons in the NTS, without changing the plasma E2 level (27). These observations support our decision to focus on the NTS as the target area to study the impact of chronic HFD consumption on estrogenic actions on feeding and apoA-IV gene expression.
It has been well-accepted that E2 acts as an indirect controller of food intake; i.e. it inhibits feeding, at least in part, by increasing the potency of satiation factors involved in meal termination (32). Our previous studies have provided evidence that apoA-IV is such a factor. In rats and mice, apoA-IV reduces food intake in a dose-dependent manner after central administration. Blocking its endogenous action with an apoA-IV-specific antibody increases meal size, implying that endogenous apoA-IV plays a role in the tonic inhibition of food intake (33,34). To determine whether apoA-IV gene and protein expression within the NTS are modulated under normal physiological conditions, we compared both apoA-IV mRNA and protein levels between male and female rats (6) and found that apoA-IV mRNA and protein levels in the NTS of female rats at both diestrous and estrous stages of the ovarian cycle were significantly higher than those in males. The apoA-IV levels in female rats at estrus were significantly higher than that at diestrus, consistent with the preceding increase in circulating E2 levels (35) and higher ER␣ gene expression during the estrus phase (36).
Given the important role of E2 in the regulation of energy homeostasis in the females, OVX should significantly increase food intake and body weight, and suppress apoA-IV gene expression in the NTS of female rats. Furthermore, cyclic replacement of E2 should be able to normalize these changes in food intake, body weight, and apoA-IV mRNA levels in OVX animals. Consistent with these hypotheses, OVX-induced alterations, including increased food intake and body weight, and decreased apoA-IV gene expression in the NTS, were all reversed by cyclic E2 treatment (2).
In our previous studies, we demonstrated that E2 enhances the anorectic potency of apoA-IV in OVX rats. Intra-4th ven- Estrogen up-regulates central apoA-IV via SRC-1 tricular administration of low doses of apoA-IV reduced food intake more in OVX rats treated with E2 than that in vehicletreated controls (2). Consistent with this, OVX apoA-IV KO mice had a reduced feeding response to E2, eating significantly more food and gaining more body weight than OVX wildtype mice during the period of cyclic E2 replacement (2). These data indicate that increased signaling by endogenous apoA-IV may partially mediate E2-induced inhibition of feeding.
E2 exerts its physiological function through its receptors, and there is compelling evidence that E2's effect on energy balance is primarily mediated by ER␣. ER␣ KO mice exhibit an obesity phenotype, whereas ER␤ KO mice remain lean (37). Deletion of ER␣ in mice blocks the antiobesity effects of E2 replacement (38). Additionally, administration of an ER␣-selective agonist (PPT), but not an ER␤-selective agonist (DPN), promotes hypophagia in OVX rats (13). Given that apoA-IV is involved in E2's anorexigenic action, we therefore hypothesized that E2 stimulates apoA-IV gene expression through ER␣. This hypothesis was supported by the observation that apoA-IV is colocalized with ER␣ in the NTS of female rats (2). Using cultured primary neuronal cells isolated from embryonic brainstems of rats, we found that both E2 and PPT (ER␣ agonist), but not DPN (ER␤ agonist), significantly stimulated apoA-IV gene expression, compared with vehicle controls. When the neuronal cells were treated with E2 conjugated with bovine serum albumin (E2-BSA), which activates membrane-associated ERs, but does not enter cells due to its large size and charge properties (2), no significant difference in apoA-IV gene expression was found between E2-BSA and vehicle-treated cells, implying that a nuclear mechanism with ER␣ is involved in E2's effect on apoA-IV gene expression (2).
E2 binds to ER␣, forms dimers, and relocalizes the ER␣ to the nucleus, resulting in subsequent recruitment of the E2-ER␣ complex to a palindromic sequence, known as ERE, to regulate target gene expression (39). In our previous studies, using a chromatin immunoprecipitation (ChIP) assay, we demonstrated that an ERE located at the upstream region of rat apoA-IV promoter is able to recruit ER␣ in the NTS of OVX rats, and this recruitment was significantly increased after E2 treatment (6). Furthermore, we confirmed that this ERE is necessary for the response to E2 treatment in the apoA-IV promoter with luciferase reporter assay (6).
SRC-1 is a transcriptional coregulatory protein that contains several nuclear receptor interacting domains and has an intrinsic histone acetyltransferase activity. SRC-1 is recruited to DNA promotion sites by ligand-activated nuclear receptors, such as E2-activated ER␣. SRC-1, in turn, acylates histones, which makes downstream DNA more accessible to transcription. Therefore, SRC-1 promotes ER␣'s up-regulation of target gene expression (7,40). Previous studies have demonstrated that SRC-1 KO mice are sensitive to diet-induced obesity (8). More importantly, the estrogenic effects on body weight and food intake are blunted in OVX SRC-1 KO mice (9), implying that SRC-1 is required for E2's actions on catabolic metabolism. Although these findings from SRC-1 KO mice are interesting, it remains unknown at which brain area(s) SRC-1 mediates E2's catabolic action. In the present study, we found that SRC-1 was abundantly expressed and colocalized with apoA-IV in the NTS. Furthermore, the recruitment of ER␣ and SRC-1 to the ERE of the apoA-IV promoter was enhanced by the cyclic replacement of E2. These findings provided neuroanatomical and biochemical basis that SRC-1 in the NTS could be involved in effects of E2-ER␣ signaling on energy intake and apoA-IV gene expression.
RNA interference is a cellular mechanism to inhibit expression of a target gene in a highly specific manner (41). shRNA expression has the advantage of providing a continuous source of silencing molecules and is a potent experimental tool for long-term gene silencing (42). With a newly generated shRNA for SRC-1, we validated its efficiency in cultured neuronal cells. Compared with shCTL, SRC-1 shRNA (shSRC-1) significantly reduced SRC-1 mRNA and protein expression. Interestingly, the SRC-1 knockdown also significantly reduced apoA-IV mRNA and protein levels in E2-treated neuronal cells, indicating that SRC-1 is, at least partially, required for E2's stimulation on apoA-IV gene expression.
We further determined the functional involvement of SRC-1 in E2's effects on feeding and apoA-IV gene expression by directly transducing shSRC-1 or shCTL into the NTS of OVX rats. In control OVX rats receiving shCTL, cyclic E2 treatment significantly reduced daily food intake, compared with oil treatment. However, this anorectic action of E2 was diminished in SRC-1 knockdown OVX rats because no significant difference in food intake between E2-and oil-treated rats was found at any time point. Consistent with the changes in food intake, oiltreated shCTL-transduced OVX rats had markedly increased body weight, and this body weight gain was reduced after cyclic E2 treatment. Although the increase of body weight in SRC-1 knockdown rats was comparable with that in control rats receiving shCTL under oil treatment, cyclic E2 replacement caused much less reduction of body weight in SRC-1 knockdown OVX rats, compared with shCTL rats. By the end of the experiment, SRC-1 knockdown rats were significantly heavier than shCTL-control rats. Interestingly, no significant difference in energy expenditure was found between shSRC-1 knockdown and shCTL-control rats under either oil or E2 treatment, indicating that SRC-1 knockdown in the NTS specifically attenuates the estrogenic effect on energy intake, but not on energy expenditure. The augmented food consumption is responsible for the increase of body weight.
It is noteworthy that knockdown of SRC-1 also reduced basal expression of apoA-IV expression in the NTS, indicating that ER␣ may not be the only receptor interacting with SRC-1. This could have any of several possible explanations. First, whereas ovariectomy significantly reduces E2 levels, the OVX rats are not absolutely devoid of E2 in the circulation (35), because estrogens can be synthesized in non-reproductive tissue, including liver, heart, muscle, and adipose tissue (43). The low level of endogenous E2 may still be able to stimulate apoA-IV gene expression to a certain degree through the interaction with SRC-1. Therefore, SRC-1 knockdown will attenuate endogenous E2's effect in OVX rats, leading to reduced basal expression of apoA-IV. Second, our previous studies have found that central apoA-IV gene expression is up-regulated by leptin through the JAK/STAT3 signaling pathway (15). The transcriptional activation by STAT3 has been reported to Estrogen up-regulates central apoA-IV via SRC-1 require the recruitment of SRC-1, which functions as a coactivator of STAT3 protein (44). SRC-1 knockdown may therefore impair the effect of endogenous leptin signaling on apoA-IV gene expression, leading to reduced basal levels of apoA-IV. Third, SRC-1 may serve as a transcription factor for other receptors or signaling pathways, for example, SRC-1 could interact with the p50 subunit and coactivate nuclear factor B-mediated transactivation (45). Although these questions are beyond the scope of the current studies, determining whether SRC-1 mediates the actions of leptin and/or cytokines on apoA-IV gene expression in diet-induced obese animals will be investigated in the future.
In summary, we have demonstrated that SRC-1 in the NTS is a physiological mediator of E2's anorectic action, and provided insight into the underlying molecular mechanisms as to where and how E2 inhibits food intake without influencing energy expenditure. It is known that women, after entering menopause, are at increased risk for developing the metabolic syndrome. Hormone replacement therapy can curb postmenopausal weight gain but also increase the risk of heart disease and breast cancer (46). One solution to this dilemma would be to target the downstream genes of E2 signaling that are involved in energy balance without causing unwanted estrogenic side effects. Thus, identifying altered key components, such as SRC-1, which mediate E2's effect on apoA-IV gene expression in OVX rats, will reveal novel pharmacological targets for reducing post-menopausal obesity risk without deleterious side effects.

Animals
Adult female Long-Evans rats (Harlan, Indianapolis, IN) were individually housed in a temperature-controlled vivarium on a 12-h light/dark cycle. Rodent chow (Teklad; Harlan) and water were provided ad libitum except where noted. All animal procedures were conducted in accordance with the National Institutes of Health Guidelines for the Care and Use of Animals and approved by the University of Cincinnati Institutional Animal Care and Use Committee.

Materials and diets
E2 and sesame oil (as vehicle) were purchased from Sigma. Antiserum against rat apoA-IV was raised from goat and characterized as described previously (47). Mouse monoclonal antibody against ER␣ (catalog number PIMA513304), Dynabeads TM M-280 Streptavidin (catalog number 11205D), and poly(dI-dC) (catalog number 20148E) were purchased from Thermo Fisher Scientific Inc. (Waltham, MA). Rabbit monoclonal antibody against SRC-1 (catalog number 2191) was purchased from Cell Signaling (Danvers, MA). Rat hypothalamic neuronal cells (CLU213) were obtained from Cedarlane Laboratories (Burlington, NC). Cell culture medium was from Invitrogen. All other chemicals were purchased from Sigma. Two pelleted semi-purified (AIN-93M), nutritionally complete experimental HFD and LFD diets were prepared by Dyets (Bethlehem, PA), as reported previously (5).

Double-labeling immunohistochemistry
The rats were transcardially perfused with 0.9% saline, followed by 4% sodium phosphate-buffered paraformaldehyde. Brains were postfixed for an additional 4 h, stored in 15% and then 30% sucrose solution at 4°C. Free-floating coronal brainstem sections (30 m) were cut through the NTS on a microtome (Leica, Buffalo Grove, IL) (2). After washing, the sections were blocked with 5% normal donkey serum in PBS containing 0.3% Triton X-100 for 1 h. Then, sections were incubated with rabbit monoclonal antibody against SRC-1 (1:1000) and goat anti-apoA-IV antibody (1:400) overnight at 4°C. Secondary antibodies were used as appropriate, including 1:500 diluted donkey anti-rabbit (for SRC-1) or donkey anti-goat (for apoA-IV) IgG conjugated to fluorescein, Alexa Fluor 594, or Alexa Fluor 488 (Molecular Probes, Inc., Eugene, OR), respectively, for 1 h at room temperature. Fluorescence images were taken using a IX83 fluorescence microscope with cellSens-Life Science Imaging Software (Olympus, Waltham, MA). Omission of the primary antibodies as well as substituting the primary antibody with normal rabbit serum or normal goat serum were used to determine the specificity of the antibodies, as we have done previously (15).

DNA pulldown assays and Western blotting
The 5Ј-biotinylated double-stranded DNA (5Ј-TCAGAG-CAAGACGAACAAGGGGCAGCATGACCCAGTTCCCAT-CAGCTCTGC-3Ј) corresponding to position 5645 to 5696 at the 5Ј-upstream region of the apoA-IV promoter and containing the ERE (6) were synthesized by Integrated DNA Technologies (Coralville, IA). The biotinylated DNA containing a mutated ERE (5Ј-TCAGAGCAAGACGAACAAGGTTGATG-CAGCTATCAGTTCCCATCAGCTCTGC-3Ј), in which the conserved nucleotides of the ERE are replaced (6), as indicated by italic characters, was also synthesized.
All lentiviral particles were prepared by the Viral Vector Core at the University of South Carolina (48 -50) using an approach described previously (48) by transiently transducting 107 293T cells with 15 g of vector cassette, 10 g of packaging cassette, and 5 g of envelope cassette. All vectors were pseudotyped with the vesicular stomatitis virus-glycoprotein envelope cassette. Titers of the vectors were determined by p24 gag enzyme-linked immunosorbent assay. The absence of replication-competent retroviruses was determined by three independent methods: tat transfer assay, vector rescue assay, and p24 gag enzyme-linked immunosorbent assay, as described previously (51). For in vitro and in vivo studies, titers of 1 ϫ 10 7 infection units (IU)/ml and 5 ϫ 10 9 IU/ml were used, respectively. All experimental procedures were approved by the University of Cincinnati Institutional BioSafety Committee.

Neuronal cell culture and transduction
Because no brainstem neuronal cell line is available commercially, we used rat hypothalamic cells (CLU213), which express ER␣, SRC-1, and apoA-IV. Those neuronal cells were seeded in HyClone Dulbecco's modified Eagle's medium/high glucose media (with L-glutamine and L-glucose; Thermo Scientific) and 10% heat-inactivated fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA). Cells were treated with trypsin (Invitrogen) and transferred to 6-well tissue-cultured plate such that they reached 70% confluence overnight. Medium was then removed and replaced with fresh medium with one of the four different lentiviral particles (final concentration 5 ϫ 10 5 IU/ml). After 48 h, the transduced cells were selectively grown in medium containing puromycin (5 g/ml) for an additional 2 weeks. The neuronal cells transduced with lentiviral particles were then treated with E2 (10 nM) for an additional 24 h, and lysed for measuring SRC-1 and apoA-IV mRNA and protein levels by qPCR and Western blotting, respectively (2).

Stereotaxic surgery in OVX rats
Overnight-fasted OVX rats were divided into two weightmatched groups (n ϭ 14/group). One group received bilateral intra-NTS injections with shSRC-1 (5 ϫ 10 9 g/ml, 0.5 l/site), and the other group received shCTL as control. The coordinates for the NTS were anteroposterior from bregma Ϫ13.6 mm, mediolateral Ϯ0.7 mm, and dorsoventral Ϫ8.0 mm, based on the atlas of Paxinos and Watson and previous reports (22,36). To reduce potential tissue damage, each NTS injection was made with a Hamilton syringe (5 l, model: 75-RN) via an Intergrated Stereotaxic Injector (iSi TM ) system (Stoelting Co., Wood Dale, IL) at a rate of 0.1 l/min for 5 min with a 10-min pause between sides (52).

Cyclic E2 or oil treatment in the OVX rats transduced with lentiviral particles
The OVX rats receiving shSRC-1 or shCTL injections were divided into two weight-matched groups on day 7 after surgery. One group was injected with E2 (2 g/100 l if sesame oil, subcutaneously, every 4th day) for 10 cycles, and the other group was injected with the same volume of sesame oil for the same period. Injections were done between 9 and 10 a.m. Food intake and body weight were monitored on the 2nd day after each cycle of E2 or oil treatment.

Indirect calorimetry
A continuous monitoring system (Oxymax Equal Flow System, Columbus OH) was used to determine energy expenditure during the 9th cycle of E2 or oil treatment. Rats from each group were placed in the system for 48 h. The first 24 h were considered adaptation and the data from the next 24 h were analyzed. Data for indirect calorimetry analysis were sampled every 30 min (16).

qPCR for apoA-IV mRNA measurement
Total RNA was extracted from the cultured neuronal cells or the tissues from rat NTS using a PureLink RNA Mini Kit (Ambion-Life Technologies). RNA was quantified using a NanoDrop 2000 (Thermo Scientific). Then, 250 ng of RNA from each sample was used for reverse transcription to cDNA with a Transcriptor First Strand cDNA Synthesis Kit (Roche Molecular Diagnostics). ApoA-IV mRNA levels were quantified by qPCR using TaqMan Fast Advanced Master Mix with Taq-Man Gene Expression Assay for apoA-IV with a StepOne TM Plus device (Thermo Scientific). 45 S mRNA levels from each sample were used as internal controls to normalize the mRNA levels, as described previously (53).

Measurement of plasma samples
Blood samples were taken from the tail vein, and plasma leptin and insulin were measured with ELISA kits (Millipore) (16).

Statistical analysis
All data are presented as mean Ϯ S.E. Data were analyzed using parametric statistics (SigmaPlot version 12.0). In vitro experiments were carried out in triplicate and performed on 3-4 separate occasions. Differences among more than two groups were determined using one-way or two-way repeated measures analysis of variance followed by Student-Newman-Keuls test for comparison between treatments. p values less than 0.05 were considered statistically significant.