Adipocyte-specific deletion of the mineralocorticoid receptor improves glucose homeostasis and associates with FGF21-adiponectin signaling in obese male mice

Abstract

Aims

The mineralocorticoid receptor (MR) is a ubiquitous nuclear receptor that is increased during obesity. Fibroblast growth factor 21 (FGF-21) is a hepatokine that enhances glucose metabolism by binding and activating fibroblast growth factor receptors in complex with the beta-klotho co-receptor (β-klotho). This study tested the hypothesis that adipocyte MR contributes to the development of glucose intolerance in obesity settings by impairing FGF-21-mediated glucose handling in adipose tissue.

Materials and Methods

An adipocyte-specific inducible MR knockout (AdipoMRw^KO) mouse model was generated to investigate metabolic effects during obesity. Tamoxifen-induced MR deletion was performed in 14-week-old male and female mice after 9 weeks on high fat diet (HFD). MR floxed (AdipoMR^fl) mice served as control littermates. Body weight, body composition, oral glucose tolerance, insulin tolerance, Homeostatic Model Assessment of Insulin Resistance (HOMA-IR), and adipokines were measured 4 weeks post-treatment.

Results

Body weight and composition were not significantly different before or after MR deletion in mice of either sex. In HFD-fed male AdipoMR^KO mice, glucose tolerance and HOMA-IR were improved, while insulin sensitizing genes including β-klotho were increased in perigonadal white adipose tissue (gWAT) compared with AdipoMR^fl mice. Notably, adipocyte MR deletion increased plasma FGF-21, which was associated with increased FGF-21 and adiponectin protein expression in gWAT. Conversely, MR deletion in HFD-fed female mice did not influence body adiposity, glucose homeostasis, or gWAT gene expression.

Conclusions

Deletion of the adipocyte MR activates the FGF-21/β-klotho/adiponectin axis in obese male mice only. Thus, these data indicate that dysfunctional FGF-21/β-klotho/adiponectin axis is associated with impaired glucose handling in a sex-specific manner.

Introduction

Obesity is a rapidly rising global health epidemic that decreases longevity and quality of life¹. Clinically, obesity is defined by a body mass index (BMI) above 30 kg/m², however metabolically unhealthy obesity is characterized by visceral fat expansion, chronic low-grade inflammation, and impaired glucose homeostasis^2,3. Increased waist circumference is an indicator of abdominal obesity, associated with a higher risk of metabolic syndrome. In mice, perigonadal white adipose tissue (gWAT) represents a predominant visceral fat depot that expands rapidly in diet-induced obesity, providing a translational model to investigate the mechanisms underpinning this process^4,5.

Adipose tissue has increasingly been recognized not only as an energy storage depot but also as a dynamic endocrine organ central to the regulation of glucose and lipid metabolism^6-8. Adipocytes expand and increase in number in response to overnutrition. Adipose tissue responds to metabolic demands by releasing stored lipids (e.g. free fatty acids) used as fuel during periods of starvation⁹, and secreting hormones known as adipokines, which target central and peripheral tissues¹¹. In addition to mature adipocytes, which serve to store lipids in large intracellular droplets, adipose tissue is composed of other cell types, including preadipocytes and immune cells, all contributing to the process of adipose tissue remodeling to adapt to the increasing metabolic demands ⁶. Although adipose tissue accounts for approximately 10-20% of systemic glucose handling, fat expansion is speculated to be a driving factor in the development of insulin resistance¹⁰.

The mineralocorticoid receptor (MR) is a ubiquitous nuclear receptor particularly increased in the visceral adipose tissue of obese patients^12,13. Though classically known for its role in fluid and electrolyte homeostasis in the kidney, the adipocyte MR has gained attention in recent years as a regulator of metabolic and endocrine functions, and mineralocorticoid antagonists are popular in the treatment of obesity-related comorbidities such as resistant hypertension^14,15. Upon binding to mineralocorticoids (i.e. aldosterone) or glucocorticoids (i.e. cortisol), MRs translocate to the nucleus and act as a transcriptional regulator of target genes by binding to hormone response elements upstream of gene promoters¹⁶. Several studies have demonstrated that MR activation is required for the differentiation of preadipocytes into mature adipocytes and is necessary for the intracellular accumulation of lipids^17,18. Pharmacological blockade of the MR has been shown to improve glucose tolerance and insulin sensitivity in rodent models of diet-induced obesity^17,19. Conversely, in genetic models where MR is overexpressed in adipose tissue, fasting insulin and Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) are worsened¹². Moreover, adipocyte MR modulates the secretion of adipokines such as adiponectin and leptin, key endocrine mediators of appetite and glucose homeostasis via central and peripheral mechanisms²⁰.

Fibroblast growth factor 21 (FGF-21), a hepatokine secreted by the liver under fasting conditions, has also been shown to improve glucose tolerance in the obese state and is being investigated as a clinical target in obesity and metabolic syndrome²¹. FGF-21 acts on the adipocyte via FGF-21 receptor 1 and its obligate co-receptor, β-klotho²², leading to both acute and sustained insulin-sensitizing effects in insulin-resistant ob/ob and diet-induced obese (DIO) rodent models²³. These effects are attributed to both reduced hepatic glucose output and increased adipocyte glucose uptake²³. When insulin binds the insulin receptor (Insr), insulin receptor substrate 1 (Irs1) is phosphorylated, leading to the translocation of glucose transporters such as glucose transporter 4 (Glut4) to the cytosolic membrane, allowing the import of glucose^24,25. β-klotho has been shown to impair the regulation of insulin-induced Glut4 expression and translocation, as well as the insulin signaling cascade²⁶, and is known to be downregulated during HFD feeding²⁷.

Mechanistic studies investigating the role of MR in adipocyte metabolic homeostasis in settings of obesity and metabolic syndrome are limited. Therefore, this study tested the hypothesis that the adipocyte MR contributes to the development of glucose intolerance in the obese state by impairing FGF-21-induced glucose transport expression. We developed an adipocyte-specific MR knockout model (AdipoMR^KO), where male and female mice were subjected to DIO and then treated with tamoxifen to assess the role of adipocyte MR on body weight, body composition, adipose tissue-derived FGF-21, leptin and adiponectin levels, as well as glucose homeostasis and HOMA-IR.

Materials and Methods

Animals

The study, which followed ARRIVE 2.0 guidelines and those of the National Institutes of Health Guide for the Care and Use of Laboratory Animals, was approved and monitored by the Institutional Animal Care and Use Committee (IACUC) at the University of Kentucky. Adipocyte-specific inducible MR knockout mice were generated using a cre-lox system, by crossing mice with floxed MR locus on a C57bl/6 background (Nr3c2^fl/fl; generously provided by Pierre Chambon, Institute of Genetics, Molecular and Cellular Biology (GIE-CERB), Illkirch-Graffenstaden, France). with mice expressing the inducible Cre-recombinase under the adiponectin (Adipoq) promoter (C57BL/6-Tg(Adipoq-cre/ERT2)1Soff/J, Stock No: 025124, Jackson Laboratories, Bar Harbour, ME). Male and female Nr3c2^fl/fl Adipoq^Cre/+ (AdipoMR^KO) were generated for this study, and Nr3c2^fl/fl Adipoq^+/+(AdipoMR^fl) littermates served as controls. Genotypes were determined using real-time PCR with specific probes designed and validated for each gene (Transnetyx, Inc., Cordova, TN).

Experimental design

Mice were weaned to normal rodent chow on postnatal day 21 (Teklad 2918, Inotiv, Madison, WI) and switched to high fat diet (HFD; D12492, 60% Kcal from fat, Research Diets Inc, New Brunswick, NJ) at five weeks of age. Beginning at four weeks of age for a total of 9 weeks, body weight was recorded weekly. Body composition was measured using Echo Magnetic Resonance Imagining (Echo-MRI; Echo Medical System, Houston, TX) at 6, 10, 14, 16, and 18 weeks of age. At week 14, tamoxifen was administered daily for five consecutive days in all groups (Sigma T5648, 40 mg/kg in corn oil, i.p. injection) to activate adipocyte iCre, inducing deletion of the MR gene, Nr3c2. Oral glucose tolerance tests (OGTT) were performed before and after MR deletion, at weeks 14 and 18. In a subset of mice, an insulin tolerance test (ITT) was performed at week 18. After a 3-day recovery from the last OGTT/ITT, mice were fasted overnight and euthanized via cardiac exsanguination under anesthesia (ketamine/xylazine 100/10 mg/kg I.P. injection). Whole blood was collected in EDTA and stored briefly on ice, and plasma was separated by centrifugation at 3000 RPM for 10 minutes at 4°C. Plasma, liver, kidney and adipose tissue were collected and either flash-frozen and stored at −80°C.

PCR for MR genomic DNA

DNA was extracted from gWAT, liver and kidney (AS1610, Promega Maxwell^® RSC Tissue Minikit) for determination of tissue-specific gene excision. 20 ng of DNA was amplified with forward and reverse primers surrounding the floxed region of MR, and an additional reverse primer within the floxed region: 5’-CCACTTGTATCGGCAATACAGTTTAGTGTC-3’; 5’CACATTGCATGGGGACAACTGACTTC-3’; 5’-CTGTGATGCGCTCGGAAACGG-3’. The products were separated using gel electrophoresis (2% agarose gel) and compared against a standard ladder (N0559S, New England BioLabs, Ipswich, Massachusetts).

Stromal vascular fraction (SVF) and mature adipocyte (MAT) isolation

Stromal vascular cells were isolated by collagenase digestion using at least 1 gram of adipose tissue from each mouse as previously reported²⁸. After 3 washes with PBS, the MAT supernatant was separated from the pellet (SVF) and both were flash-frozen and stored at −80°C for the subsequent RNA isolation.

Metabolic testing

OGTT was performed after 15-hour overnight fast by administration of oral glucose gavage (2g/kg) as previously reported²⁹. Blood glucose was measured by tail nick prior to gavage, then 15-, 30-, 60-, and 120-minutes post-gavage using Accu-check Aviva Plus glucometer and test strips (Roche diagnostics). HOMA-IR was calculated in the fasting state using the following formula: insulin (μIU/mL) x glucose (mmol/mL)/22.5^30,31. ITT was performed after 6-hour fast during the light phase by intraperitoneal injection of insulin (Humalin-R, Eli-Lilly and Company, Indianapolis, IN, 0.1 U/kg). The area under the curve (AUC) for OGTT and ITT was calculated via the trapezoidal method.

Plasma and explant measurements

Plasma aldosterone, leptin, FGF-21, insulin, and adiponectin were measured via commercial ELISA kits (Cayman cat # 501090, Cayman cat #10007609, Bio-techne cat # MF2100, Crystal Chem cat # 900808, Bio-techne cat# MRP300, respectively) following the manufacturer’s specifications. For tissue explants, ~100 mg of gWAT tissue was incubated in 100 μL, or ~40mg of liver was incubated in 200 μL DMEM+2% FFA-free BSA, in a sterile incubator for 1 hour at 37°C. The explant media was flash frozen and stored at −80°C before used to perform ELISAs. Values were normalized per mg of tissue.

qRT-PCR

RNA was extracted using the Aurum Total RNA Mini Kit (BioRad laboratories, Hercules, CA, #7326820). 1 ug of RNA was converted to cDNA using iScript gDNA Clear cDNA Synthesis Kit (BioRad #1725035). qRT-PCR was performed by amplifying cDNA for 40 cycles using either iTaq Universal SYBR Green Supermix (BioRad #1725121) or 2x TaqMan Advanced Fast Master Mix (Applied Biosystems^™ 4444557). Validation of MR deletion was confirmed via SYBR Green PCR assays using primers as follows: 18s, 5’-AGTCGGCATCGTTTATGGTC-3’, 5’-CGAAGCATTTGCCAAGAAT-3’; Nr3c2, 5’-GAAGAGCCCCTCTGTTTGCAG-3’, 5’-TCCTTGAGTGATGGGACTGTG-3’. Additional gene expression was measured using the following TaqMan assays: HpRT (housekeeper), Mm030024075_m1; Insulin receptor substrate 1 (Irs1), Mm01278327_m1; Insulin receptor (InsR), Mm01211875_m1; Glucose transporter 4 (Glut4), Mm00436615_m1; β-klotho (Klb), Mm00473122_m1; Leptin (Lep), Mm00434759_m1; Adiponectin (Adipoq), Mm04933656; Fibroblast growth factor 21 (Fgf21), Mm07297622_g1; Peroxisome proliferator-activated receptor gamma (Pparg), Mm00440940_m1, Phosphoenolpyruvate caboxykinase (Pck1), Mm01247058_m1; and Glucose-6-Phosphatase (G6pc), Mm04207416_m1.

Statistical analysis

Body weight and body composition at each time point was analyzed by a 2-way RM ANOVA. At the endpoint, differences due to genotype were analyzed by unpaired Student’s t test for each sex using a threshold of p = 0.05. Results are expressed as means ± SEM. Gene expression was compared using ΔΔCT² and analyzed using Student’s unpaired t test. Correlations were analyzed using the F-test and the F-statistic is reported at a threshold of p=0.05. Number of mice per group, and the report of excluded mice showing a mean >2xSD from the group are reported in supplementary table 1. Data from mice that died after the tamoxifen treatment was removed from the study.

Results

Deletion of MR is adipose tissue-specific in male and female mice fed a HFD.

Deletion of MR from genomic DNA was confirmed using a three-primer system, in which the forward primer corresponds to a DNA region directly upstream of the first loxP site, the first reverse primer is located within the floxed region, and the second reverse primer is located directly downstream of the second loxP site. Using this design, the intact MR gene was amplified by the forward and first reverse primer during PCR and yielded a product ~350 base pairs in length. The non-functional MR gene was amplified by the forward and second reverse primer, yielding a product of ~400 base pairs (Figure 1A). When the PCR products from adipose tissue, kidney and liver of AdipoMR^fl and AdipoMR^KO mice were run through gel electrophoresis, the 400 base pair excised-gene band was only present in adipose tissue of AdipoMR^KO mice for both males and females (Figure 1B). Despite MR expression in whole adipose tissue (gWAT) was greater in females compared to males (Supplementary Fig. 1), the deletion reduced MR expression in both sexes, mRNA expression of the MR gene is reduced by approximately 75% in male AdipoMR^KO adipose tissue (gWAT) compared to AdipoMR^fl (Figure 1C) and approximately 85% in females (Figure 1D), while gene expression in liver and renal cortex was similar between groups.

Figure 1: The MR is deleted in a tissue-specific manner in both male and female obese AdipoMR^KO mice and does not affect body weight or body composition:

A) Schematic diagram of the primer locations used to verify that mineralocorticoid receptor (MR) deletion is occurring. In cells expressing MR, the forward primer and reverse 1 primer will produce a polymerase chain reaction (PCR) product of ~350 base pairs (bp); B) In mature adipocytes in which exons 5 and 6 of the MR gene have been excised, the forward primer and reverse 2 primer will produce a PCR product of ~400 bp; Agarose gel of PCR products from adipose tissue, liver and kidney AdipoMR^fl and AdipoMR^KO male and female mice. In HFD-fed male AdipoMR^fl and AdipoMR^KO mice, C) mRNA expression of MR in adipose tissue, liver and kidney, and plasma aldosterone. In HFD-fed female AdipoMR^fl and AdipoMR^KO mice, D) mRNA expression of MR in adipose tissue, liver and kidney, and plasma aldosterone. Weekly body weight across the duration of the study, percentage of fat mass and lean mass of E) male AdipoMR^fl and AdipoMR^KO mice and f) female AdipoMR^fl and AdipoMR^KO mice. Results are expressed as means ± SEM and were analyzed by unpaired Student’s t test. n = 5-10 animals/group. Each timepoint was analyzed by unpaired Student’s t test and was found insignificant at a threshold of p = 0.05.

AdipoMR^fl=mineralocorticoid receptor (MR) floxed mice; AdipoMR^KO= MR knock out mice.

In a subset of mice, we determined the cell type-specific MR deletion. MR mRNA expression in SVF was similar between AdipoMR^fl and AdipoMR^KO; however, reduced MR mRNA expression was found only in the MAT fraction from AdipoMR^KO mice (Supplementary Fig. 2). Additionally, we verified that MR deletion did not affect circulating aldosterone levels (Figure 1C-D).

Deletion of the adipocyte MR does not affect body weight or body composition of mice fed a HFD.

Weekly body weights and the percentage of fat and lean mass were not different between genotypes across the duration of the study for male (Figure 1E) or female (Figure 1F) mice. As expected, body weight and percent fat mass slightly decreased in all mice following tamoxifen administration; however, body composition recovery was achieved before the end of the study. At the end of the study, there were no differences in the masses of liver, kidney or various fat depots, including gWAT, subcutaneous (scWAT), perirenal (prWAT), and brown adipose tissue (BAT) between AdipoMR^KO and AdipoMR^fl groups in the tissues collected at the endpoint (Supplementary table 2).

Glucose homeostasis is improved only in male AdipoMR^KO male mice fed a HFD.

Before MR deletion, OGTT was similar between Adipoq^+/+ and Adipoq^Cre/+ mice fed a HFD for both male (Figure 2A&B) and female mice (Figure 2D&E). Four weeks after the deletion, OGTT was improved in AdipoMR^KO male mice compared to AdipoMR^fl mice. The AUC calculated from the OGTT was reduced in AdipoMR^KO male mice (Figure 2B). Furthermore, the change in AUC before and after MR deletion was significantly reduced in AdipoMR^KO male mice compared to controls (Figure 2C). Unlike males, AUC for OGTT, HOMA-IR were similar between female AdipoMR^KO and AdipoMR^fl mice (Figure 2D-F). Fasting blood glucose was not influenced by MR deletion in either sex (Figure 2A&2D, inset).

Figure 2: Glucose tolerance is improved in male AdipoMR^KO mice fed a HFD, but not in females:

In HFD-fed male AdipoMR^fl and AdipoMR^KO mice, A) blood glucose after challenge with an oral gavage of glucose (OGTT) prior to MR deletion at 14 weeks of age and after deletion at 18 weeks, and B) accompanying area under the curve (AUC) calculated from glucose levels over 120 minutes. C) Change in AUC from 14 weeks to 18 week of age. In HFD-fed female AdipoMR^fl and AdipoMR^KO mice, D) blood glucose during OGTT before and after MR deletion, and E) accompanying AUC. F) Change in AUC from 14 weeks of age to 18 week of age. Male and female fasting blood glucose in inset A) and D), respectively. Results are expressed as means ± SEM and were analyzed by unpaired Student’s t test. n = 5-10 animals/group, α= 0.05.

AdipoMR^fl=mineralocorticoid receptor (MR) floxed mice; AdipoMR^KO= MR knock out mice.

Fasting insulin was decreased in male AdipoMR^KO mice (Figure 3A), which drives an improvement in HOMA-IR, a measure of insulin sensitivity (Figure 3B). However, direct insulin sensitivity measured by ITT was not different (Figure 3C). In addition, we found that the liver expression of gluconeogenic genes, Pck1 and G6pc, was similar between AdipoMR^KO and AdipoMR^fl mice (Supplementary Table 3). Nevertheless, the metabolic improvement displayed by AdipoMR^KO male mice was accompanied by the upregulation of genes related to insulin-mediated glucose handling in gWAT, including Irs1, Glut4, and InsR., as well as Klb (Figure 3D).

Figure 3. Insulin homeostasis is improved in male AdipoMR^KO mice fed HFD:

In HFD-fed male AdipoMR^fl and AdipoMR^KO mice, A) plasma insulin, B) Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) index, and C) blood glucose during insulin tolerance test (ITT) and area under the curve (AUC, inset). D) Expression of Insulin receptor (InsR), Insulin receptor substrate 1 (Irs1), Glucose transporter 4 (Glut4) and β-klotho (Klb) in gWAT. In HFD-fed female AdipoMR^fl and AdipoMR^KO mice, E) Plasma insulin, f) HOMA-IR, and g) blood glucose during ITT and AUC inset. F) InsR, Irs1, Glut4 and Klb expression in gWAT. Plasma results are visualized on a violin plot with the median represented by solid black bar and interquartile range indicated by dashed lines within the violin and were analyzed by unpaired Student’s t test. n = 5-12 animals/group.

AdipoMR^fl=mineralocorticoid receptor (MR) floxed mice; AdipoMR^KO= MR knock out mice.

Unlike male mice, deletion of adipocyte MR did not influence insulin levels or gWAT gene expression in female mice (Figure 3E-H).

Plasma FGF-21 is increased in male mice fed a HFD with adipocyte MR deletion.

Male AdipoMR^KO mice showed increased circulating levels of plasma FGF-21 (Figure 4A), while plasma adipokines implicated in the FGF-21 signaling pathway, adiponectin and leptin, were not different from AdipoMR^fl mice (Figure 4B and 4C, respectively). Notably, the positive correlation between gWAT mass and circulating levels of FGF-21, typically reported clinically in obesity and metabolic disease, was overturned by adipocyte MR deletion (Figure 4D).

Figure 4. Deletion of adipocyte MR increases plasma FGF-21 and inverts the correlation between FGF-21 and gWAT only in male mice:

In HFD-fed male AdipoMR^fl and AdipoMR^KO mice, A) Plasma levels of fibroblast growth factor 21 (FGF-21), B) adiponectin and C) leptin, and D) correlation between plasma FGF-21 and gWAT mass. In HFD-fed female AdipoMR^fl and AdipoMR^KO mice E) Plasma levels FGF-21, F) adiponectin and G) leptin in female AdipoMR^fl and AdipoMR^KO mice, and H) correlation between plasma FGF-21 and gWAT mass in female AdipoMR^fl and AdipoMR^KO mice. Results are visualized on a violin plot with the median represented by solid black bar and interquartile range indicated by dashed lines within the violin and were analyzed by unpaired Student’s t test. n = 6-12 animals/group. For B) and D), correlation significance was determined using F-test and are expressed as the F-statistics.

AdipoMR^fl=mineralocorticoid receptor (MR) floxed mice; AdipoMR^KO= MR knock out mice.

Conversely, female AdipoMR^KO and AdipoMR^fl mice showed similar plasma FGF-21, adiponectin, and leptin, and deletion of the adipocyte MR did not improve the relationship between plasma FGF-21 and gWAT mass (Figure 4E-H).

Deletion of the adipocyte MR is associated with local activation of the FGF-21-Adiponectin axis in adipose tissue of male mice fed HFD

To determine the source of increased circulating FGF-21 levels, we measured the contributions from liver and gWAT. In male AdipoMR^KO mice, FGF-21 in liver explant media was not statistically different from AdipoMR^fl (Figure 5A); however, MR deletion increased gWAT FGF-21 and adiponectin, but not leptin, in explant media compared to tissues with intact adipocyte MR (Figure 5B-D). Accordingly, MR deletion also increased the gene expression of Fgf21 and Adipoq, but not Lep, in gWAT (Supplementary Table 2).

Figure 5. Deletion of adipocyte MR activates the local FGF-21-adiponectin axis in adipose tissue:

In HFD-fed male AdipoMR^fl and AdipoMR^KO mice, A) Fibroblast growth factor 21 (FGF-21) from liver tissue explants, B) FGF-21 from gWAT tissue explants, C) leptin from gWAT tissue explants, and D) adiponectin from gWAT tissue explants. In HFD-fed female AdipoMR^fl and AdipoMR^KO mice, E) FGF-21 from liver tissue explants, F) FGF-21 from gWAT tissue explants, G) leptin from gWAT tissue explants, and H) adiponectin from gWAT tissue explants. Results are visualized on a violin plot with the median represented by solid black bar and interquartile range indicated by dashed lines within the violin and were analyzed by unpaired Student’s t test. n = 4-12 animals/group.

AdipoMR^fl=mineralocorticoid receptor (MR) floxed mice; AdipoMR^KO= MR knock out mice.

Once again in female mice, deletion of adipocyte MR did not influence the levels of liver- or gWAT-derived FGF-21 and adiponectin production (Figure 5E-H and Table S2).

Discussion

This study is the first to show that adipocyte-specific deletion of the MR can improve glucose homeostasis in a sex-specific manner, independent of changes in fat mass. Male mice fed a HFD displayed improved glucose homeostasis associated with the activation of the FGF-21-adiponectin axis in adipose tissue. In addition, adipocyte-specific MR deletion reduced fasting insulin and HOMA-IR, as well as upregulated the insulin receptor, insulin substrate receptor, glucose transporter 4, and β-klotho, a co-receptor required for functional FGF-21 signaling. Furthermore, mice undergoing MR deletion displayed increased gWAT-derived FGF-21 and adiponectin, which could play an important role in mediating the local insulin-sensitizing actions of FGF-21. Notably, these metabolic improvements in AdipoMR^KO mice were present in male counterparts only. This data suggest that adipocyte MR deletion contributes to metabolic improvements in a sex-specific manner.

In the obese state, FGF-21 has actions in the liver and pancreas but does not affect insulin sensitivity in skeletal muscle. In the liver, FGF-21 improves glucose handling similarly to the fat, in addition to improving insulin clearance, which could be a critical factor to reduce plasma insulin levels^35,36. In the pancreas, FGF-21 can induce insulin secretion in the diabetic primary islet cells, but this effect is not observed in healthy cells³⁷. Though FGF-21 is responsible for improvements in metabolic function and inflammation, including enhanced insulin signaling and lipolysis, it is paradoxically increased in obesity^23,32. This has led to the controversial hypothesis of obesity as an FGF-21-resistant state^32-34. However, in addition to further increased FGF-21, our male AdipoMR^KO mice displayed increased expression of β-klotho. Accordingly, MR deletion in obesity settings abrogated the positive correlation between FGF-21 and adiposity displayed by AdipoMR^fl mice. This data supports the idea that adipocyte MR deletion promotes beneficial metabolic actions that could be mediated by FGF-21.

Hyperleptinemia mediates central nervous system-dependent production of liver FGF-21^38,39, which has been shown to stimulate the production of adiponectin, a metabolically protective adipokine. Adiponectin improves insulin sensitivity through reduced inflammation, adenosine monophosphate-activated protein kinase (AMPK) activation, reactive oxygen species generation and peroxisome proliferator-activated receptor-alpha (PPARα) activation^40-42. Studies have demonstrated that adiponectin is a key regulator of the insulin-sensitizing actions of FGF-21^40,43. In our model, both FGF-21 and adiponectin were increased in the media explant from gWAT, but not in circulation, indicating an autocrine/paracrine effect on glucose handling limited to the adipose tissue. Remarkably, as we did not find reduced gluconeogenic gene expression in the liver, decreased fasting insulin without differences in ITT after MR deletion suggests that increases in FGF-21 displayed by mice with adipocyte MR deletion could affect insulin homeostasis in other tissues.

Of note, metabolic improvements associated with adipocyte MR deletion were absent in female mice. In fact, glucose tolerance and HOMA-IR in female mice, regardless of genotype, were closer to the AdipoMR^KO male mice than AdipoMR^fl male mice. This could be taken as an indication that female mice did not develop as severe of a metabolic syndrome-like phenotype as male mice; however, this study did not utilize low-fat diet controls, which would allow to determine the extent of metabolic improvement in each sex. It is known that female rodents have an inherent resistance to diet-induced obesity, including lower insulin resistance⁴⁴. Thus, female mice might require a longer exposure to HFD in order to develop a distinct metabolic phenotype that could be corrected by deletion of the adipocyte MR.

Plasma levels of aldosterone were not affected by MR deletion, confirming that our mice did not develop any compensatory responses to the genetic manipulation that could affect the outcomes of the study. Overactivation of MR by aldosterone in cells other than the adipocyte has the potential to confound studies of obesity, a complex systemic disease. For example, aldosterone mediates water and electrolyte homeostasis in the kidneys, inducing hypertension⁴⁵. MR activation in immune cells can potentiate chronic systemic inflammation present in the obese state and exacerbate hypertension and cardiovascular disease ^46,47. Therefore, it is reasonable to conclude that increases in aldosterone are not a confounding variable in our model.

Previous studies have investigated the effects of adipocyte MR deletion on body composition and glucose handling, reporting only outcomes in male mice ^48,49. A study by Feraco et al. ⁴⁸ used an MR flox mouse line generated by Dr. Berger ⁵⁰, in which exon 3 of the Nr3c2 gene is floxed, as opposed to exons 5 and 6 in the genetic mouse model used in our study ⁵¹. This flox was crossed with the Offerman Adipo^cre mouse line ⁵², whereas our study used the Rosen Adipo^cre. Additionally, Ferguson et al.⁴⁹ reported the outcomes of an adipocyte-specific MR knockout, also using the Berger MR flox; however, this line was crossed with the same Adipo^cre model used in our study. While all of these genetic lines are generated on a C57bl/6J background, genetic drift and founder effects have the potential to impact phenotypic outcomes between studies.

Perhaps of greater significance, in both of these previous studies, tamoxifen treatment was administered in young mice between 6-8 weeks of age, prior to HFD exposure, whereas our study was designed to investigate the effects of MR deletion in mature adult mice already exhibiting metabolic dysfunction. Therefore, we administered tamoxifen using similar dosing procedures in mice undergoing a chronic HFD feeding. Additionally, both the Feraco and Ferguson studies used diets with ~45% kcal from fat, while our study utilized a more aggressive model of diet-induced obesity using a 60% kcal/fat diet. Studies of metabolic disturbances in diet-induced obesity frequently use the 60% HFD to rapidly progress the onset of weight gain and metabolic dysfunction in order to model a more progressed disease state in a more efficient timeframe⁵³. Therefore, the magnitude of metabolic disease achieved in 13-16 weeks of 60% HFD feeding will greater than in mice fed 45% HFD⁵⁴. The lower fat content of the diets used in previous studies may contribute to the lack of noticeable effects in glucose-handling. Additionally, deletion of the adipocyte MR in 6-8-week-old mice represents a different physiological context both at the age of deletion (prior to diet-induced obesity), and at the age of metabolic assessment (several months after the intervention). One potential explanation for the outcomes we report in our study could be that deleting the gene during a state of metabolic dysfunction can benefit glucose homeostasis, whereas deleting the gene prior to the onset of obesity does not prevent the development of glucose intolerance. Further, Feraco et al found that the AdipoMR^KO mice had a blunted adipogenic response to HFD feeding, however in our model, body weight and body composition were not different between groups during the 4 weeks post-deletion. Of note, tamoxifen treatment significantly decreased body weight in all experimental animals during the first two weeks; however, mice recovered from their weight loss at the end-point of the study, not showing differences between genotypes. Whether feeding AdipoMR^KO mice a HFD for a longer period of time may reduce body adiposity could be further investigated. However, these studies would require additional validation showing that deletion of the adipocyte MR persists for longer periods of time in obese adult mice. Because the cre line we used is specific to the mature adipocyte, and because obesity induces increased adipocyte differentiation from preadipocytes⁵⁵, we speculate that a longer period of time on HFD after deletion may increase the proportion of MR-expressing adipocytes in the tissue.

Of note, our study is the first to report differences in glucose homeostasis not previously observed in male mice, though methods of assessment were unique between all three studies. Feraco et al. and Ferguson et al. assessed glucose tolerance by intraperitoneal (IP) injection of glucose, and in one case after only 4 hours of fasting⁴⁹, whereas our mice were challenged with an oral gavage of glucose after overnight fasting. The present study utilized overnight fasting, a more severe fast which induces a catabolic state and depletes glycogen stores. While overnight fasting is widely used in metabolic testing, more recent guidelines recommend shorter fasting periods between 2-4 hours to prevent weight loss and impaired counterregulatory response^56,57. Additionally, oral administration of glucose provides a more physiologically-relevant assessment of glucose tolerance, as IP injection bypasses the enteroendocrine regulation of incretin hormones, including GLP-1, ghrelin and gastrin, which play important roles in glucose homeostasis⁵⁸. Additionally, the Feraco and Ferguson studies used a lower dose of glucose to challenge their mice, 1g/kg compared to the 2g/kg used in the current study. It is also documented that administration of tamoxifen itself is can influence glucose tolerance in a sex- and strain-specific manner⁵⁹. In our hands, we observed an overall effect of tamoxifen improving glucose tolerance in female mice, which was unaffected by genotype, as shown by delta AUC. Male AdipoMR^fl mice also demonstrated a modest reduction in glucose tolerance after tamoxifen treatment, however only the AdipoMR^KO male mice experienced significant improvements. In this sense, it is worth considering that while total AUC in OGTT highlights overall differences in glucose handling, this significance is lost when baseline glucose is accounted for in the calculation of AUC, referred to as integrated AUC (iAUC). The lack of difference in iAUC between obese male AdipoMR^fl and AdipoMR^KO mice, while total AUC remains significant, alludes to the subtlety of the differences in glucose tolerance in our model. While MR antagonism is known to improve glucose tolerance^60,61; our findings suggest that systemic MR antagonism, and possibly off-target receptor effects also contribute to this regulation.

In conclusion, our data indicate that adipocyte MR deletion could facilitate the upregulation of FGF-21/β-klotho/adiponectin axis, contributing to increase the expression of insulin sensitizing genes expression in adipose tissue, and improving glucose homeostasis in male mice fed a HFD. Thus, we have identified a potential sex-specific MR-dependent pathway contributing with the worsening of metabolic function in setting of obesity.