Transcriptional control of intestinal cholesterol absorption, adipose energy expenditure and lipid handling by Sortilin

The sorting receptor Sortilin functions in the regulation of glucose and lipid metabolism. Dysfunctional lipid uptake, storage, and metabolism contribute to several major human diseases including atherosclerosis and obesity. Sortilin associates with cardiovascular disease; however, the role of Sortilin in adipose tissue and lipid metabolism remains unclear. Here we show that in the low-density lipoprotein receptor-deficient (Ldlr−/−) atherosclerosis model, Sortilin deficiency (Sort1−/−) in female mice suppresses Niemann-Pick type C1-Like 1 (Npc1l1) mRNA levels, reduces body and white adipose tissue weight, and improves brown adipose tissue function partially via transcriptional downregulation of Krüppel-like factor 4 and Liver X receptor. Female Ldlr−/−Sort1−/− mice on a high-fat/cholesterol diet had elevated plasma Fibroblast growth factor 21 and Adiponectin, an adipokine that when reduced is associated with obesity and cardiovascular disease-related factors. Additionally, Sort1 deficiency suppressed cholesterol absorption in both female mice ex vivo intestinal tissue and human colon Caco-2 cells in a similar manner to treatment with the NPC1L1 inhibitor ezetimibe. Together our findings support a novel role of Sortilin in energy regulation and lipid homeostasis in female mice, which may be a potential therapeutic target for obesity and cardiovascular disease.


Discussion
We report the following novel findings: (1) Sort1 deficiency reduces body and WAT weight, adipocyte lipid droplet size, and increases BAT function in HF/HC-fed female Ldlr −/− mice, (2) Sort1 deficiency increases FGF21 and Adiponectin in HF/HC-fed female Ldlr −/− mice, (3) Sort1 deficiency decreases Npc1l1 mRNA levels and cholesterol absorption in female Ldlr −/− mice and human Caco-2 cells, and 4) Sort1 deficiency decreases Klf4 mRNA and LXR-mediated transcription in female Ldlr −/− mice and human cell lines. Based on our results we present the following working model (Fig. 7): In HF/HC-fed female Ldlr −/− mice, Sort1 deficiency reduces LXR-mediated transcription, possibly in part via reduction of Klf4 mRNA levels in specific cell types, as exhibited by a lack of Klf4 mRNA levels change in intestinal tissue and cells. Transcriptional suppression of LXR leads to elevated FGF21 in BAT. Increased FGF21 increases BAT energy expenditure and induces Adiponectin release in WAT that lowers WAT mass and body weight gain. Suppression of LXR via Sort1 deficiency reduces Npc1l1 mRNA levels that in turn lowers cholesterol absorption. Conversely, reduced NPC1L1 may also act to lower body weight through LXR by suppressing oxysterol production. Together these mechanistic actions lower plasma TC and body weight, and associate Sortilin as a novel regulator of cardiometabolic function in female atherogenic mice.
SCIEntIfIC REPORTS | (2018) 8:9006 | DOI:10.1038/s41598-018-27416-y deficiency reduces body weight in mice; therefore, the mechanisms behind improved cardiovascular pathology and reduced body weight observed in Sort1-deficient Ldlr −/− mice likely involve multiple pathways. The full mechanism of how Sort1 deficiency reduces LXR-mediated transcription in female mice and certain human cell lines requires further investigation beyond the scope of the present study; however, we identified a likely mechanistic link with Sort1 deficiency reducing the LXR-regulator, Klf4 in vitro and in vivo in examined cell types.
Further tying cardiovascular pathology to energy expenditure and metabolism, we found reduced plasma TC, Npc1l1 mRNA levels, and cholesterol absorption in Sort1-deficient female Ldlr −/− mice and human Caco-2 cells. In agreement with our findings, Npc1l1 deficiency or ezetimibe treatment likely protects against diet-induced obesity in mice via increased energy expenditure 52 . LXR signaling regulates the expression of lipid absorption and efflux genes (Npc1l1 and Abc-gene family members) 32 . LXR agonists increase intestinal cholesterol absorption via Npc1l1 induction 53 . On the other hand, the NPC1L1 inhibitor, ezetimibe, attenuates LXR signaling pathways via reduced cholesterol absorption and oxysterol production 54,55 . Together these studies support the notion that NPC1L1 can both regulate and be regulated by LXR. As such, it is possible that Sort1 deficiency may control body weight in female atherogenic mice, at least in part via regulation of Npc1l1 mRNA levels. and Ldlr −/− Sort1 −/− mice. 10-week-old male and female littermates were fed normal chow (NC) or high-fat/ high cholesterol (HF/HC) (21% fat and 1.25% cholesterol, Research Diets #D12108C, New Brunswick, NJ, USA) diet for 15 weeks. Body weight and food consumption were monitored weekly, and blood was collected from the submandibular vein prior to the start of and at 5, 10, and 15 weeks on the NC and HF/HC diets. Mice were pentobarbital euthanized after 15 weeks on the study diet, and blood, peri-gonadal white adipose tissue (WAT), intrascapular brown adipose tissue (BAT), and jejunum was collected. Plasma total cholesterol (TC), triglycerides (TG), and glucose levels were assessed using Wako Pure Chemical Industries kits (Osaka, Japan; Cholesterol E-test, Triglyceride E-test, Glucose CII-test) according to the manufacturer's protocols. Hepatic and fecal TC and TG levels were measured by chloroform:methanol (2:1) extraction in combination with the Wako TC/TG kits. Plasma Adiponectin and Fibroblast growth factor 21 levels were measured by ELISA according to the manufacture's protocol (R&D systems, Minneapolis, MN). For histology, WAT and BAT were embedded in optimum cutting temperature compound and 7 mm serial sections were cut. Tissue samples were stained with hematoxylin and eosin (H&E). Images were captured with a digital camera (DS-Fi1c, Nikon, Melville, NY, USA) and adipocyte size was quantified using Image J software (National Institutes of Health, Bethesda, MD) as previously described 56  Corp., Lakewood, NJ, USA)) for 60 minutes at 37 °C. The cell suspension was filtered through a 100 μm cell strainer (Corning, Corning, NY, USA) followed by centrifugation (1350 rpm for 10 min). The remaining pellet was suspended with ACK red blood cell lysing buffer (Thermo Fisher Scientific) and centrifuged again. After centrifugation, the cell pellet was re-suspended in DMEM containing 10% fetal bovine serum (FBS, Thermo Fisher Scientific)), 50 μg/mL Gentamicin (Corning), 50 U/mL penicillin, and 50 g/mL streptomycin (Corning), and incubated in 5% CO 2 incubator. After reaching confluency, pre-adipocytes isolated from WAT were treated with adipogenic medium (DMEM containing 10% FBS, gentamicin, penicillin, streptomycin, 25 nM insulin, 1 nM Triiodo-L-thyronine (T3, Sigma Aldrich), 0.5 mM 3-isobutyl-1-methylxanthine (IBMX, Sigma Aldrich), 1 μM dexamethasone (Sigma Aldrich), 1 μM Rosiglitazone (Sigma Aldrich), 500 μM palmitic acid (Sigma Aldrich)). After two days, cells were cultured in insulin medium containing 25 nM insulin, 1 nM T3, 1 μM dexamethasone and 500 μM palmitic acid for 5 to 6 days. Primary pre-adipocytes isolated from BAT were treated with induction medium (DMEM with 10% FBS, gentamicin, penicillin, streptomycin, 200 μM Ascorbic acid (Sigma Aldrich), 25 nM insulin (Sigma Aldrich), 1 nM T3, 0.5 mM IBMX), 1 μM dexamethasone (Sigma Aldrich), 0.25 mM indomethacin (Sigma Aldrich)). After a day in induction media, cells were cultured in DMEM with 10% FBS, gentamicin, penicillin, streptomycin, 200 μM Ascorbic acid, 25 nM insulin, 1 nM T3, 0.25 mM indomethacin, 1 μM CL316243 (Thermo Fisher Scientific) for 5 to 6 days.

Methods
Caco-2, Hek293, and HepG2 cells were obtained from ATCC (Manassas, VA, USA). Caco-2 cells were cultured in DMEM with 10% FBS, 1% penicillin and streptomycin, and 1% Non-Essential Amino Acids Solution (NEAA, Thermo Fisher Scientific). Hek293 cells were cultured in DMEM with 10% FBS, 1% penicillin and streptomycin. HepG2 cells were cultured in Eagle's Minimum Essential Medium (ATCC, Manassas, VA, USA) containing 10% FBS, 1% penicillin and streptomycin. RNA silencing was performed as described previously 22 . Briefly, 50 nM siRNA against Sort1 (L-010620, ONTARGETplus SMART-pool, Thermo Fisher Scientific) and non-targeting siRNA (ON-TARGET Non-Targeting Pool, Thermo Fisher Scientific) were transferred into cells using Dharmafect 4 (Thermo Fisher Scientific) or Lipofectamine RNAiMAX (Thermo Fisher Scientific) according to manufacturer's protocols. RNA interference was performed over a 72-hour period. For LXR agonist experiments, LXR agonist treatment was performed by incubating cells in the final 24 hours of the 72-hour period with 1 μM of the LXR agonist T0901317 in DMSO (synthesized and HPLC purified by Kowa Company LTD, Tokyo, Japan).

Intestinal cholesterol absorption.
Caco-2 cells were seeded in DMEM containing 10% FBS, 1% penicillin and streptomycin, and 1% NEAA, at 4 × 10 5 cells per well in a 12-well plate. Between 15 and 21 days after confluency, differentiated cells were washed with PBS twice and pre-incubated in serum-free DMEM for 16 hours. After pre-incubation, cells were washed with transportation buffer (140 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl 2 , 0.8 mM MgSO 4 , 5 mM Glucose, 25 mM Tris (pH 7.5)) and incubated in DMEM with 5% lipoprotein-deficient serum (Sigma Aldrich) supplemented with ezetimibe (Selleck Chemicals, Houston, TX, USA) or dimethyl sulphoxide (DMSO, control) for 1 hour. After 1 hour, cells were incubated with lipid micelles (5 mM Taurocholate (Sigma Aldrich), 0.5 mM oleic acid (Sigma Aldrich), 0.04 mM phosphatidylcholine (Sigma Aldrich), 0.16 mM lysophosphatidylcholine (Sigma Aldrich), 0.3 mM mono-olein (Sigma Aldrich)), and 10 μM NBD-cholesterol (Thermo Fisher Scientific) for 4 hours. After incubation, cells were washed with PBS containing 5 mM taurocholate twice and lysed using 0.2 N NaOH with 1% SDS. Cell lysates were measured for fluorescence at excitation 465 nm and emission 535 nm, and values were normalized by protein concentration using bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific). For ex vivo cholesterol absorption assay, jejunum was harvested from Ldlr −/− Sort1 +/+ or Ldlr −/− Sort1 −/− mice. Isolated tissues were cut to expose the surface using flat sections, and then cultured on 12-well plates with serum-free DMEM for 16 hours. After pre-incubation, tissues were washed with transfer buffer, and incubated with DMEM containing 5% lipoprotein-deficient serum with or without ezetimibe for 1 hour. Lipid micelles, including NBD-cholesterol were added to the culture medium and incubated for 4 hours. Tissues were homogenized using a bead-based homogenizer and NBD-cholesterol fluorescence was measured with fluorescent values normalized to total protein content.
SCIEntIfIC REPORTS | (2018) 8:9006 | DOI:10.1038/s41598-018-27416-y RNA and protein analysis. Total RNA was extracted from tissue and cells using Trizol reagent (Invitrogen, Waltham, MA, USA) according to the manufacturer's instructions. cDNA was synthesized using a qScript cDNA Synthesis Kit (Quanta, Beverly, MA, USA). Quantitative real-time PCR was performed with commercially obtained Taqman probes (Thermo Fisher) on a 7900HT fast real-time PCR system (Applied Biosystems, Carlsbad, CA, USA). mRNA relative expression was determined by normalization to Gapdh using the delta-delta CT method. Cell protein was isolated and processed in a similar manner to which we have previously reported 57 . Briefly cells were washed with PBS and scraped from plates in RIPA buffer (Thermo Fisher Scientific) containing protease and phosphatase inhibitors. Protein content was determined by BCA assay, and 20 μg of total protein lysate was processed by SDS-PAGE and transferred onto nitrocellulose membranes using the iblot 2 apparatus (Thermo Fisher Scientific). Human cell line Sortilin was detected using an anti-Sortilin antibody (R&D Systems #AF3154) and GAPDH (Santa Cruz, #sc-47724, Dallas, TX, USA) was used as a loading control. Adipose tissue protein was isolated in RIPA buffer (Thermo Fisher Scientific) containing protease and phosphatase inhibitors, in a similar manner to which we have previously reported 37 . Briefly tissues were homogenized using a plastic pestle and run through a syringe prior to centrifuging at 5000 g for 5 minutes to pellet insoluble tissue. Supernatant protein content was quantified by BCA assay and 100 μg of total protein lysate was processed by SDS-PAGE. Mouse adipose Sortilin was detected with an anti-Sortilin antibody (Abcam #ab16640, Cambridge, UK) and GAPDH was used as a loading control.
Graphing and statistical analysis. Data are presented as mean ± SEM; n indicates the number of samples or independent experiments performed. Statistical analyses were performed using GraphPad Prism Version 5 (Prism Software, Inc., La Jolla, CA, USA). For comparison between two groups, t-test were performed using Prism. For comparison among multiple groups, one-way ANOVA followed by post hoc tests were performed using Prism. P values less than 0.05 were considered significant. Heat maps were generated by quantitative PCR obtained mRNA levels using Qlucore software (Lund, Sweden). The working model was generated using Microsoft PowerPoint and the Motifolio illustration tool kit (Ellicott City, MD). Data availability. All data generated or analyzed during this study are included in this article.