Correction: Sirt1 coordinates with ERα to regulate autophagy and adiposity

Sex difference in adiposity has long been recognized but the mechanism remains incompletely understood. Previous studies suggested that adiposity was regulated by autophagy in response to energy status change. Here, we show that the energy sensor Sirt1 mediates sex difference in adiposity by regulating autophagy and adipogenesis in partnership with estrogen receptor α (ERα). Autophagy and adipogenesis were suppressed by Sirt1 activation or overexpression, which was associated with reduced sex difference in adiposity. Mechanistically, Sirt1 deacetylated and activated AKT and STAT3, resulting in suppression of autophagy and adipogenesis via mTOR-ULK1 and p55 cascades. ERα induced Sirt1 expression and inhibited autophagy in adipocytes, while silencing Sirt1 reversed the effects of ERα on autophagy and promoted adipogenesis. Moreover, Sirt1 deacetylated ERα, which constituted a positive feedback loop in the regulation of autophagy and adiposity. Our results revealed a new mechanism of Sirt1 regulating autophagy in adipocytes and shed light on sex difference in adiposity.

Sirt1 was shown to promote or suppress autophagy, partly because of Sirt1 deacetylating autophagy proteins such as Atg5, Atg7, or LC3 [28][29][30] . However, it is unknown whether and how Sirt1 interacts with autophagy in the regulation of adipogenesis and adiposity. In the present study, we investigated the effects of gain and loss of Sirt1 on autophagy and adiposity. We found that the mice with increased expression of Sirt1 showed reduced adiposity, to a greater extent in females than males. Sirt1-induced reduction of adiposity was associated with lower autophagy activity, and the sex difference in adiposity change may be ascribed to the crosstalk between estrogen receptor ERα signaling and Sirt1-autophagy axis. Mechanistically, Sirt1 deacetylated AKT and STAT3, which activated mTOR-ULK1 and STAT3-p55 (p55 subunit of phosphoinositide 3-kinase) signaling pathways, respectively, to mitigate autophagy in adipocytes. ERα and Sirt1 form a positive feedback loop that enhances the effects on autophagy and adiposity. Our study unraveled a new mechanism of Sirt1 regulating adiposity and sex difference through the crosstalk with ERα and autophagy.
The effects of gain of Sirt1 on autophagy and adipogenesis To examine whether the gain of Sirt1 suppresses autophagy in adipocytes, we overexpressed Sirt1 in 3T3L1 cells using knock-in technique. As shown in Fig. 2a, b, Sirt1 knock-in (Sirt1-KI) significantly increased the expression of Sirt1 protein in 3T3L1 cells, which largely suppressed adipogenesis even in the presence of differentiation inducer (DI). The mitigated adipogenesis was associated with reduced autophagy flux activity (Fig.  2b, c). These findings support the notion that autophagy is required for adipogenesis ( Fig. 1) [4][5][6][7][8]33 , and Sirt1 appears to be an autophagy suppressor in adipocytes. In addition, our data, along with the evidence from MEFs, stem cells, and cancer cells [28][29][30] , suggests that Sirt1 may play tissueor cell type-dependent roles in autophagy.

The effects of loss of Sirt1 on autophagy and adipogenesis
To test whether the loss of Sirt1 may increase autophagy activity and adipogenesis, we used RNA interfering Fig. 1 Adipogenesis was associated with downregulation of Sirt1 and activation of autophagy. a Oil red O staining of 3T3L1 cells during differentiation, showing the data of day 0 (preadipocyte), day 6 (differentiating adipocyte), and day 12 (differentiated or mature adipocyte). Scale bar, 50 µm. b Western blotting analysis of Sirt1 and LC3 proteins during 3T3L1 cell differentiation. GAPDH was probed as a loading control. c Autophagy flux was determined by contrasting the rates of removing the substrates LC3-II and p62 by autophagy in the absence and presence of autophagy inhibitors BL (bafilomycin A1 at 0.1 μM and leupeptin at 10 μg/ml). Pre, preadipocyte; Mat, mature adipocyte. *p < 0.05; **p < 0.01. n = 8.

Sirt1 reduces sex difference in adiposity
To examine how the Sirt1-autophagy axis affects adiposity in vivo, we measured the fat mass of S1tg mice and control mice (Fig. 5a, b). S1tg mice exhibited significantly lower fat mass compared with the control mice, in line with Sirt1 suppressing autophagy and adipogenesis in vitro (Figs. 1-3) and Sirt1 activation preventing adipose expansion in mice 16,17 . Intriguingly, the sex difference in S1tg mice was attenuated drastically in comparison to control mice (Fig. 5c). Western blotting analysis revealed that in control mice, females had a higher expression of Sirt1 than in males, while the difference was largely abolished by Sirt1 overexpression in S1tg mice (Fig. 5d). These findings suggested the Sirt1 might crosstalk with sex hormone signaling, particularly estrogen receptor ERα signaling, because ERα has been shown to regulate autophagy and adiposity 7,48,49 . Indeed, treatment of 3T3L1 cells with estradiol (E2) markedly induced Sirt1 expression (Fig. 5e), while knockout of ERα attenuated Sirt1 expression (Fig. 5f), underscoring Sirt1 as a downstream target of ERα. When adipocytes were treated with E2, it increased phosphorylation of mTOR and ULK1 (Fig.  5g, h), the signaling pathway known to mitigate the initiation of autophagy (Figs. 1-4) 7,34 . In line with autophagy required for adipogenesis, E2 treatment substantially inhibited adipocyte differentiation (Fig. 5i). However, knockdown of Sirt1 (Sirt1-KD) abolished E2induced effects on the mTOR-ULK1 cascade and adipogenesis, further confirming that E2/ERα signaling regulates autophagy and adipogenesis by inducing Sirt1 expression ( Fig. 5g-i). Therefore, the reduced sex difference in adiposity in S1tg mice may be explained at least in part by the fact that Sirt1 overexpression overrides the effects of E2/ERα.

Sirt1 deacetylates ERα and reduces adiposity
Given the nature of Sirt1 as a deacetylase, we asked the question whether adipose ERα is deacetylated by Sirt1. Intriguingly, overexpression of Sirt1 significantly reduced acetylation level of ERα in S1tg mice in comparison to the control mice (Fig. 6a). Total ERα protein level was higher in S1tg mice than in the control mice, but the difference was not statistically significant (Fig. 6a). In mature 3T3L1  Fig. 6b. Total ERα protein level was reduced in mature 3T3L1 adipocytes, consistent with the notion that suppression of ERα signaling promotes adipogenesis 7 . However, when 3T3L1 cells were treated with nicotinamide mononucleotide (NMN, 100 µM), a Sirt1 activator that promotes biosynthesis of NAD +15,50,51 , it significantly reduced acetylation level of ERα (Fig. 6b). These in vitro and in vivo data confirm that ERα is deacetylated by Sirt1, suggesting that Sirt1 forms a positive feedback loop and promotes E2/ERα signaling cascade in the regulation of adiposity. In line with this, Sirt1 overexpression in males reduced adiposity by 0.7% (Fig. 6c); however, Sirt1 overexpression in females (known to have stronger E2/ERα signaling 52,53 ) resulted in greater reduction of adiposity, i.e., by 2.3% (p < 0.001 vs. 0.7% in the males). Thus, our data support the model where E2/ ERα signaling induces Sirt1-mediated suppression of autophagy and adipogenesis, and Sirt1 forms a positive feedback loop to enhance E2/ERα signaling in the regulation of adiposity (Fig. 6d).
promoter of Sirt1 gene and increasing its transcription 56 . Overexpression of Sirt1 phenocopied the effects of E2/ ERα signaling, during which mTOR-ULK1 cascade was activated and dampened autophagy and adipogenesis (Figs. 2, 4 and 5) 7 . Consistently, silencing Sirt1 abolished the effects of E2/ERα signaling on autophagy and adipogenesis (Figs. 3 and 5). In line with E2/ERα signal differing in males and females 52,53 , Sirt1 expression was higher in females vs. males in the control mice (Fig. 5d). However, the sex difference in Sirt1 expression was normalized in S1tg mice, and the sex differences in adiposity was significantly reduced in comparison to control mice (p < 0.001, Fig. 5c).
Our study is the first to document the role of Sirt1 in regulating autophagy in adipose tissue. Sirt1 appeared to act on the initiation of autophagy via inhibitory phosphorylation of ULK1 by mTOR. Notably, Sirt1 overexpression resulted in marginal changes in phosphorylation of mTOR and ULK1 in skeletal muscle (data not shown), suggesting that Sirt1 plays tissue-or cell type-dependent roles in autophagy. Indeed, Sirt1 induces autophagy by deacetylating and activating Atg5, Atg7, or LC3, but this function was absent in certain cell types [28][29][30]39 . For instance, deacetylation of LC3 and Atg7 leads to activation of autophagy in spermatids; nevertheless, knockout of Sirt1 attenuated autophagy and spermiogenesis in germ cells but not in steroidogenic cells 39 . It is, therefore, of interest for future studies to explore the molecular and cellular factors that determine the cell typedependent roles for Sirt1 in the regulation of autophagy.

Mice
The S1tg and control mice were bred and housed as described previously 7,20 . The ERαKO (ERα −/− ) and control mice (ERα +/+ ) were obtained by breeding heterozygous males to females 7 . All the mice were housed in plastic cages on a 12-h light-dark photocycle and with free access to water and regular chow diet 7,59 . At the age Fig. 6 Sirt1-induced deacetylation of ERα and reduction of adiposity in mice. a Effects of Sirt1 overexpression on ERα acetylation in mice. Upper, acetylated ERα was analyzed by immunoprecipitation (IP) with an anti-acetylated lysine (anti-acLys) antibody that pulled down acetylated proteins, followed by immunoblot (IB) with an anti-ERα antibody. Normal rabbit IgG (non-specific IgG) was used as negative control. Middle, whole lysates (WL) were analyzed by IB with anti-ERα and anti-GAPDH antibodies. Lower, densitometric analyses were used to quantify acetylated and total ERα in control (Ctrl) and Sirt1 transgenic (S1tg) mice. *p < 0.05; n = 8. b Effects of Sirt1 activator nicotinamide mononucleotide (NMN, 100 µM) on ERα acetylation in 3T3L1 cells. Left-upper, acetylated ERα was analyzed by IP with an anti-ERα antibody that pulled down ERα protein, followed by IB with an anti-acetylated lysine (anti-acLys) antibody. Normal rabbit IgG (non-specific IgG) was used as the negative control. Left-lower, whole lysates (WL) were analyzed by IB with anti-ERα and anti-GAPDH antibodies. Right, densitometric analyses were used to quantify acetylated and total ERα in 3T3L1 cells. DI, differentiation inducer. *p < 0.05 and **p < 0.01 by comparing DI(−) NMN(−) with DI(+) NMN(−). # p < 0.05 and ## p < 0.01 by comparing DI(+) NMN(−) with DI(+) NMN(+). n = 8. c Sirt1-induced reduction of adiposity was greater in female than in male mice. ***p < 0.001; n = 6-10. d A schematic view of Sirt1 coordinating with ERα to regulate autophagy and adipogenesis (adiposity). of 6-8 weeks, the S1tg and control mice were weighed, and fat mass was measured with a Bruker Minispec LF90 NMR Analyzer (Bruker Optics, Billerica, MA, USA), and then euthanized for tissue collection. ERαKO and control mice were sacrificed for tissue collection at the age of 12-16 weeks. The inguinal WAT pads were collected and weighed quickly and snap freezing in liquid nitrogen. Power calculation suggested that 6 mice per group were required to obtain statistically meaningful data. Mice were randomly grouped for each genotype. Animal use procedures followed the National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committees at University of Florida and Virginia Tech.

Cell transfection
After cells reached 80-90% confluent, overexpression (Sirt1-KI) or knockdown (Sirt1-KD) of Sirt1 was achieved by incubating the cells with adenoviruses carrying GFP (control) and Sirt1 coding sequences, or adenoviruses carrying GFP (GCATCAAGGTGAACTTCAAGA, control) and Sirt1 (GCACCGATCCTCGAACAATTC) shRNA sequences, at 100 multiplicity of infection 61 . After 2 days, cell differentiation was induced as described above and cells were analyzed or harvested on day 9 or day 12 as indicated.

Oil red O staining
The Oil Red O working solution was freshly prepared by mixing 0.35% stock solution with dH 2 O (6:4) and filtered, and the staining was conducted as described 6,33,60 . After the media were removed, the cells were washed once with cold phosphate buffered saline and fixed in 4% formaldehyde at room temperature for 10 min. The cells were then washed with dH 2 O and air dried. Oil Red O working solution was added to start the staining at room temperature for 30 min. The stained cells were washed with dH 2 O for 4 times before the images were captured with a Nikon ECLIPSE Ti Inverted Microscope (Melville, NY, USA).

Autophagy flux assay
Preadipocytes and mature adipocytes were treated with bafilomycin A1 (inhibitor of autophagosome acidification, at 0.1 μM) plus leupeptin (the inhibitor of lysosomal proteases, at 10 μg/ml) for 4 h. The cells were then harvested to prepare cell lysates as previously described 6,33,60 . The turnover of LC3-II or p62 protein, i.e., the autophagic removal of the substrate, was measured by Western blotting and image analysis to assess autophagy flux 6,31,62 .

Statistical analysis
Measurements were duplicated or triplicated, with 6-10 mice included in each group. Data were presented as mean ± SD. Unless the use of female mice were specified, the animal studies were conducted on males. Differences between groups and treatments were validated by oneway analysis of variance or a two-sided t-test. A value of p < 0.05 was considered statistically significant.