Egr1 mediates the effect of insulin on leptin transcription in adipocytes

In mammals, leptin production in adipocytes is up-regulated by feeding and insulin. Although this regulatory connection is central to all physiological effects of leptin, its molecular mechanism remains unknown. Here, we show that the transcription factor early growth response 1, Egr1, is rapidly but transiently induced by insulin in adipose cells both in vitro and in vivo, and its induction is followed by an increase in leptin transcription. ChIP and luciferase assays demonstrate that Egr1 directly binds to and activates the leptin promoter. Interestingly, the lipid droplet protein FSP27 may work as a co-factor for Egr1 in regulating leptin expression. By using siRNA-mediated knockout of Egr1 along with its overexpression in adipocytes, we demonstrate that Egr1 is both necessary and sufficient for the stimulatory effect of insulin on leptin transcription.

Leptin, a 16-kDa product of the ob gene (1), is synthesized predominantly in adipocytes and targets the central nervous system. It has been established as a major metabolic regulator that controls food intake, energy expenditure, neuroendocrine functions, carbohydrate and lipid metabolism, and several other important physiological functions of the mammalian organism (2)(3)(4)(5). Regardless of how leptin exerts its biological activity, it is essential that its production in adipocytes is coupled to nutrient uptake and energy status of the body as circulating leptin levels increase within hours after feeding and decrease shortly after food deprivation (3, 6 -8). Although in humans, this effect may not be as fast and robust as in rodents, all mammals studied thus far demonstrate a direct link between food intake and circulating leptin (3). Based on these results, the predominant hypothesis in the field has been that leptin expression is controlled by nutrients and/or insulin. Indeed, multiple studies have shown that insulin increases leptin production by adipose cells both in vivo and in vitro (3). Although this regulatory connection is central to all proposed mechanisms of leptin action, its mechanism remains unknown. Here, we are showing that insulin/mTORC1-inducible transcription factor Egr1 binds to the leptin promoter and activates leptin expression in 3T3-L1 adipocytes.

Results
Treatment of differentiated 3T3-L1 adipocytes with insulin causes a strong but transient induction of Egr1 (see also Ref. 9) followed by an increase in the leptin mRNA (Fig. 1A). By the same token, a single intraperitoneal insulin injection induces Egr1 and leptin mRNA in epididymal fat pads (Fig. 1B) and in circulating leptin in vivo (Fig. 1C).
A luciferase assay performed in HEK 293T cells demonstrates that induction of Egr1 not only precedes leptin expression in adipocytes, but Egr1 can activate the leptin promoter ( Fig. 2A). Furthermore, a ChIP assay carried out in differentiated 3T3-L1 adipocytes shows that Egr1 directly binds to the leptin promoter in an insulin-sensitive fashion (Fig. 2B).
Previously, we have found that the lipid droplet protein FSP27 (also known as CIDEC) represents a co-repressor of Egr1 (10). In agreement with these results, we now show that FSP27 blocks the effect of Egr1 on the activity of the leptin promoter (Fig. 2C).
To determine whether Egr1 is necessary for the effect of insulin on leptin transcription, we have knocked down Egr1 in 3T3-L1 adipocytes with the help of siRNA. As is shown in Fig.  2D, knockdown of Egr1 not only decreases expression of leptin in basal cells, but also completely blocks up-regulation of leptin mRNA by insulin. Importantly, Egr1 not only is necessary but also is sufficient for the up-regulation of leptin expression in adipocytes; overexpression of Egr1 in these cells with the help of adenoviral infection strongly increases levels of the leptin mRNA (Fig. 2E).
Previously, we have determined that treatment of cultured adipocytes with insulin stimulates expression of both Egr1 mRNA and protein. The MEK inhibitor PD98059 blocks insulin-stimulated increase in the Egr1 mRNA but has only a moderate effect on the Egr1 protein. Expression of the latter is completely suppressed by the mTORC1 inhibitor, PP242 (11), suggesting that the mTORC1-mediated pathway is essential for the activation of Egr1 expression by insulin. Here, we confirm this result and also show that transcription of leptin correlates with the expression of Egr1 in PD98059-and PP242-treated adipocytes (Fig. 3A).
To further prove the role of mTORC1 and Egr1 in the regulation of leptin transcription, we decided to mimic the effect of insulin specifically on the expression of Egr1 by deleting its highly structured 5Ј-UTR using the CRISPR/Cas9 technique (note the loss of the 2.5 kb band in the ⌬5ЈUTR lane in Fig. 3B).

ACCELERATED COMMUNICATION
This procedure does not significantly change levels of the Egr1 mRNA (Fig. 3C), but it dramatically elevates expression of the Egr1 protein (Fig. 3D). Administration of insulin to the "engineered" cells has only a minimal stimulatory effect on the Egr1 protein levels, suggesting that deletion of the 5Ј-UTR raises its biosynthesis to the maximum. This result supports our previous conclusion that insulin increases translation of the Egr1 mRNA primarily via the mTORC1-4E-BP1/2 regulatory axis (11). As expected, the luciferase assay has demonstrated that the activity of the leptin promoter in these cells is strongly increased (Fig. 3E).

Discussion
In this study, we provide evidence that insulin/mTORC1inducible transcription factor Egr1 mediates the physiologically significant stimulatory effect of insulin on leptin transcription.
In analyzing these data, it is essential to keep in mind that leptin production by adipose cells is regulated at several different levels: transcription, translation, secretion, and autophagic degradation (12,13). Importantly, both transcription (this report) and translation of leptin are positively regulated by insulin via the mTORC1mediated pathway (14 -16) and, therefore, should work in concert to deliver higher leptin amounts in response to nutrients and insulin. As activity of mTORC1 depends not only on insulin levels but also on nutrient and energy availability (17,18), this model provides an additional physiological dimension to the regulation of leptin production. At the same time, mechanisms of regulated leptin secretion and degradation remain virtually unstudied but may substantially contribute to changes in circulating leptin. For example, Lee and Fried (19) have found that up to 50% of newly synthesized leptin molecules are rapidly degraded in lysosomes instead of being secreted.

ACCELERATED COMMUNICATION: Egr1 activates leptin expression
The problem of leptin production has another important aspect. In addition to the short-term connection with food intake, circulating leptin levels are known to be steadily elevated in obesity (3,4). At the cellular level, larger adipocytes contain and secrete more leptin than smaller cells (20,21). Because the adipocyte size is defined primarily by the volume of the central lipid droplet, this phenomenon may show a cell-autonomous connection between the amount of stored energy (i.e. obesity at the molecular level) and leptin expression. The correlation between the size of the adipocyte and the level of leptin production has been recognized for a long time, but its mechanism remains obscure. It is unlikely that a single act of food intake can change the size of the adipocyte in a significant fashion, so there should be another explanation for this phenomenon.
To this end, it has been established that fat storage in adipocytes is also controlled by mTORC1. In particular, inhibition of mTORC1 signaling suppresses early adipogenesis (22)(23)(24)(25)(26) and/or lipogenesis (27)(28)(29)(30)(31)(32). In parallel, mTORC1 inhibits lipolysis and promotes triglyceride storage in adipocytes via the Egr1-mediated transcriptional suppression of the rate-limiting lipolytic enzyme, ATGL (9,31). Importantly, chronic overnutrition and obesity lead to continuous activation of mTORC1 (9,(33)(34)(35), which should promote triglyceride storage and increase the size of adipocytes on one hand and stimulate leptin production on the other. In other words, the size of the adipocyte may serve as an indicator of the cumulative mTORC1 activity that may explain the correlation between the adipocyte size and leptin production.
In addition, we and others have found that lipid droplets (LDs) 2 can directly "talk" to the cell nucleus via LD proteins CIDEA (36), FSP27 (also known as CIDEC) (10), and perilipin 5 (37). According to our findings (10), FSP27 is not only localized on the surface of LDs, but also present in the cell nucleus, where it binds to and regulates transcriptional activity of Egr1. In agreement with these results, we now show that FSP27 blocks the effect of Egr1 on the activity of the leptin promoter (Fig. 2C). Thus, we hypothesize that FSP27 may link the size of LDs (that account for most of the adipocyte volume) to leptin expression by inhibiting Egr1. In other words, we suggest that increasing size of LDs may "pull" FSP27 out of the nucleus, which should lead to the activation of leptin expression via Egr1. Finally, when our manuscript was in preparation, it was reported that leptin may represent a direct transcriptional target of Egr1 in human breast cancer cells treated with TNF␣ (38), which is consistent with our results.

Animals
12-16-week-old male C57BL6J mice were obtained from Charles River Laboratories (Wilmington, MA) and acclimatized for 2 weeks with a 12-h light/day cycle. Prior to experiments, animals were housed in complete darkness (free-running) for 3 days with access to food and water ad libitum. On the fourth day, animals were fasted for 6 h and injected with insulin (2 units/kg) intraperitoneally. All experimental procedures were approved by the Boston University School of Medicine Animal Care and Use Committee.

Cell culture and fractionation
Culturing and differentiation of 3T3-L1 cells were performed as described previously (9). Differentiated 3T3-L1 adipocytes were washed twice with cold PBS and lysed with 200 l of cold radioimmune precipitation assay buffer (Millipore, Burlington, MA) supplemented with Halt TM Protease and Phosphatase Inhibitor Mixture (Thermo Fisher Scientific) for 10 min on ice. Cell lysates were then centrifuged at 14,000 ϫ g for 15 min, and protein concentration was determined using the Pierce BCA Protein Assay kit (Thermo Fisher Scientific) on a Biotek Synergy TM HT microplate reader (Biotek, Winooski, VT).

Fractionation of mouse adipose tissue
Freshly excised adipose tissue (ϳ100 mg) were collected in prechilled microcentrifuge tubes containing 1 scoop (measuring apparatus provided by Next Advance) of sterile 1.0-mm diameter RNase-free zirconium oxide beads and 200 l of extraction buffer (40 mM HEPES, 120 mM NaCl, and 1 mM EDTA). Tissue was homogenized in a bullet blender homogenizer and for 5 min each at the maximum intensity. Lysates were then centrifuged at 1,000 ϫ g for 5 min at 4°C to allow triglycerides to solidify at the top. Nonidet P-40 (1% final) was added to liquid phase for 10 min at 4°C on a shaker, and samples were centrifuged at 10,000 ϫ g for 10 min at 4°C.

Adenoviral infection of 3T3-L1 adipocytes
The adenoviral vector containing human Egr1 cDNA (AdEgr1) was kindly provided by E. Hofer (Medical University of Vienna, Vienna, Austria). GFP containing adenovirus was kindly provided by Andy Greenberg (Tufts Medical Center, Boston, MA). Differentiated 3T3-L1 adipocytes in 24-well plates were serum-starved in Opti-MEM medium for 2 h. AdEGR1 or AdGFP were then added to cells at a multiplicity of infection of 4,000. After 24 h, the medium was replaced with DMEM supplemented with 10% FBS, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. After 48 h, cells were harvested for protein and RNA analysis.

siRNA-mediated knockdown of Egr1
Suspension transfections were performed according to Kilroy et al. (39). Briefly, 3T3-L1 cells were differentiated until day 4. Cells were washed gently with PBS, trypsinized with 1 ml of 0.25% trypsin-EDTA for 3-5 min, resuspended in 10% FBS/ DMEM (5 ml), and placed immediately into 24-well plates containing either Egr1 siRNA or scrambled siRNA. To prepare siRNA transfection mix, scrambled and Egr1 siRNA (Dharmacon, Lafayette, CO) were mixed with Opti-MEM medium at a final concentration of 100 nM and allowed to sit for 5 min at room temperature. DharmaFECT Duo (Dharmacon, Lafayette, CO) was added to the previous mixture (1.4 l/cm 2 surface area of transfection). The mixture was allowed to sit for 20 min at room temperature, and then 200 l was added to each well of a 24-well plate. Cells were incubated for 24 h and thereafter replenished with fresh 10% FBS/DMEM and maintained for an additional 48 h before being harvested for the analysis of protein and RNA.

ChIP
ChIP was performed with the EZ-ChIP kit from Millipore (Burlington, MA). Briefly, 3T3-L1 adipocytes were cross-linked with 18.5% formaldehyde for 10 min and then quenched with 0.125 M glycine for 5 min at 37°C. Cells were then lysed in SDS lysis buffer. Chromatin fragments were prepared by sonication in the Bioruptor Sonicator using 1.5 ml of Bioruptor Plus TPX microtubes (Diagenode, Denville, NJ) on ice for 15 cycles (30 s with 30-s intervals) at the high-intensity setting. Novus Egr1 antibody or nonspecific mouse IgG (10 g) was added to sonicated samples and incubated overnight at 4°C with rotation. Protein G-agarose (60 l) was added to samples and incubated for 1 h at 4°C with rotation. Samples were centrifuged at 3,000 ϫ g for 1 min to pellet the beads and washed before removing the supernatant fraction. Protein-DNA complexes were eluted, treated with RNase A and proteinase K, and transferred to DNA purification spin columns to be eluted and stored in Ϫ20°C.
Purified DNA eluates were analyzed by qPCR according to Cardamone et al. (40), using the primers 5Ј-TCCCGCCCCA-CCGCTAGCGAGCTC-3Ј (forward) and 5Ј-GAGCTCGCTA-GCGGTGGGGCGGGA-3Ј (reverse) to detect the Egr1 binding site. Dilutions of total lysates (1:10, 1:100, and 1:1,000) were amplified by PCR to form a standard reference curve. Numbers from eluates relative to the total lysate were used to calculate the slope and intercept for the standard reference curve. Ct values of samples were input in the formula, relative input ϭ 10 (Ct Ϫ intercept)ϫ10/slope .

Construction of leptin promoter reporter
Leptin core promoter spanning Ϫ350/ϩ27 was amplified from mouse genomic DNA by PCR with the primers 5Ј-TAA-GCAGCTAGCCTGTAGCCTCTTGCTCCCTGCG-3Ј (forward) and 5Ј-TGCTTAAAGCTTACCTTGCAGCTGCTGG-AGCAGGGATCC-3Ј (reverse). Underlined sequences contain restriction enzyme sites NheI for forward and HindIII for reverse primer. Upstream of the restriction enzyme sites are additional base pairs to reduce GC content. pGL4.1 luciferase vector was purchased from Promega (Madison, WI). Amplified PCR segments were purified using a PCR purification kit (Qiagen, Hilden, Germany). Products were ligated into the NheI and HindIII sites of pGL4.1 vector ligated following an overnight ligation method, using the Quick Ligation kit (New England Biolabs, Ipswich, MA), yielding the leptin promoter reporter (Ϫ350/ϩ27). Ligated plasmid was then transformed into DH5␣ supercompetent cells (New England Biolabs, Ipswich, MA). Successfully transformed colonies were extracted for plasmid DNA and verified by sequencing.

Leptin ELISA
Mouse blood samples (200 l) were incubated at room temperature for 25 min before being centrifuged at 2,000 ϫ g for 10 min at 4°C. Serum was then collected, and circulating leptin was measured using the mouse leptin ELISA kit (Thermo Fisher Scientific). Samples were read at 450-nm absorbance by a Biotek Synergy TM HT microplate reader and were fitted by a standard curve to determine leptin concentration.
The lentiCRISPR v2 plasmid (52961; Addgene) was used to express single guide RNAs and the Cas9 protein along with the lentiviral envelope plasmid pCMV-VSV-G and the packaging plasmid psPAX2. 3T3-L1 cells transduced with lentiviruses were cultured under selection with puromycin (Invivogen, San Diego, CA) (5 g/ml) for 4 days. Single colonies of stably transfected cells were grown in 10% FBS/DMEM containing puromycin (5 g/ml). The genomic region surrounding the single guide RNA target site was amplified by PCR using the primers 5Ј-TGTTAACCTTGACGTCCCCG-3Ј and 5Ј-GGGTGTCC-CTAGACATCCCT-3Ј and sequenced.

Gel electrophoresis and immunoblotting
Proteins were separated in either 8 or 12% SDS-polyacrylamide gels, and Western blotting was performed as described previously (9). Protein bands were detected with Immobilon Western Chemiluminescent horseradish peroxidase substrate (Millipore) using a Bio-Rad image station.