Adipoq^rtTA[C57BL/6-Tg(Adipoq-rtTA)2Zvw/J, JAX#:033448], and TRE-Cre[B6.Cg-Tg(tetO-cre)1Jaw/J, JAX#:006234] strains were purchased from Jackson laboratory. Hif1a^loxP/loxP mice were generously provided by Dr. Liwei Xie, Guangdong Academy of Sciences. Mice were maintained in ventilated cages under an inverted 12-h:12-h dark:light cycle with ad libitum access to food and water. All wild-type (WT) and transgenic animals used in this study were on the C57BL/6 background.

Transgene expression was induced with a standard rodent chow diet or chow diet containing 600 mg/kg doxycycline (DOX) (Bio-Serv, S4107). For high-fat diet studies, mice were maintained on a standard high-fat diet (60 kcal% fat, Research Diets, D12492i) or doxycycline-containing high fat diet (600 mg/kg dox, 60% kcal% fat, Bio-Serv, S5867) as indicated. For PX-478 administration, mice were administered intraperitoneally with saline or 5 mg/kg of PX-478 (Cayman Chemical, #10005189) three times weekly for 4 weeks. For AdipoRon administration, mice received intraperitoneal injection with saline or 5 mg/kg of AdipoRon (Cayman Chemical, #15941) twice a week.

The Bruker Minispec mq10 NMR were used to determine the body fat mass and lean mass in conscious mice.

Triglyceride determination kit (Sigma, triglyceride reagent T2449 and free glycerol reagent F6428) were used to measure the serum levels of triglycerides of the mice. Serum adiponectin concentrations were determined using ELISA kit (Sigma-Aldrich, EZMADP-60K). Tissue TNFα and IL6 levels were measured using ELISA kits (BioLegend, 430901; Sigma-Aldrich, RAB0309).

Protein extracts from adipose tissue were prepared by homogenization in NP-40 lysis buffer (Beyotime, P0013F) supplemented with Protease Inhibitor Cocktail (Sigma, P8340). Protein extracts were separated by SDS-PAGE electrophoresis and transferred onto PVDF membrane (Millipore, IPFL00010). After incubation with the indicated primary antibodies at 4°C overnight, the blots were incubated with secondary antibodies and visualized by Pierce ECL Plus Western Blotting Substrate (Thermo Fisher Scientific, 32109).

The primary antibodies and the working concentrations are as following:

HIF1α: 1:1000 dilution; Cell Signaling Technology, 36169.

β-TUBULIN: 1:1000 dilution; Cell Signaling Technology, 2128.

Total RNA was isolated from freshly isolated tissue using the TRIzol reagent (Invitrogen, 15596018). First strand cDNA was reverse-transcribed with M-MLV reverse transcriptase (Invitrogen, 28025-021) and random hexamer primers (Invitrogen, 48190011). Relative expression of mRNAs was determined by quantitative PCR using a TransStart^® Tip Green qPCR SuperMix (TransGen Biotech, AQ132-21) and data was analyzed using the DD-Ct method of gene quantification with Rps18 acting as the reference transcript for mRNA. All qPCR primer sequences are listed as following: Tnfa (Forward 5′-GAAAGGGGATTATGGCTCAGG-3′) (Reverse 5′-TCACTGTCCCAGCATCTTGTG-3′), Il6 (Forward 5′-AAGCCAGA GTCCTTCAGAGAGA-3′) (Reverse 5′-ACTCCTTCTGTGACTCC AGCTT-3′), Saa3 (Forward 5′-TGCCATCATTCTTTGCATCTT GA-3′) (Reverse 5′-CCGTGAACTTCTGAACAGCCT-3′), Ccl2 (Forward 5′-CCACAACCACCTCAAGCACTTC-3′) (Reverse 5′-AA GGCATCACAGTCCGAGTCAC-3′), Fn1 (Forward 5′-GAGAGCA CACCCGTTTTCATC-3′) (Reverse 5′-GGGTCCACATGATGGTG ACTT-3′), Lox (Forward 5′-TCGCTACACAGGACATCATGC-3′) (Reverse 5′-ATGTCCAAACACCAGGTACGG-3′), Adgre1 (Forward 5′-TTGTACGTGCAACTCAGGACT-3′) (Reverse 5′-GATCCCAGA GTGTTGATGCAA-3′), Itgam (Forward 5′-ATGGACGCTGATG GCAATACC-3′) (Reverse 5′-TCCCCATTCACGTCTCCCA-3′), Il1b (Forward 5′-GCAACTGTTCCTGAACTCAACT-3′) (Reverse 5′-ATCTTTTGGGGTCCGTCAACT-3′), and Rps18 (Forward 5′-CATGCAAACCCACGACAGTA-3′) (Reverse 5′-CCTCACGC AGCTTGTTGTCTA-3′).

Tissues were dissected in PBS (Beyotime, ST448) and fixed in 4% paraformaldehyde (Beyotime, P0099) overnight. Paraffin embedding, sectioning, and H&E staining, were performed by the Wuhan Servicebio Technology Co., Ltd. The following antibodies and concentrations were used for indirect immunofluorescence assays: guinea pig anti-perilipin 1:500 (Fitzgerald 20R-PP004); mouse anti-F4/80 1:500 (Cell Signaling Technology, 70076S); goat anti-rabbit Alexa 488 1:200 (Invitrogen, A-11008); goat anti-guinea pig Alexa 647 1:200 (Invitrogen, A21450). Briefly, after dewaxing and rehydration, antigen retrieval of the sections was performed by microwave oven with buffer of citric acid (Beyotime, P0086) for 14 min. Afterward, sections were washed in water and PBS for 5 min, three times and then blocked with 10% normal goat serum (Sigma-Aldrich, G9023) in PBS for at least 30 min at room temperature. Primary antibodies were diluted in PBS containing 10% normal goat serum and sections were incubated with the primary antibodies overnight at 4°C. Following three times of wash with PBS, sections were incubated with secondary antibodies diluted in PBS containing 10% normal goat serum for 2 h at room temperature. Nuclei were counterstained with ProLong™ Gold Antifade Mountant with DAPI (Invitrogen, P36931). Bright-field and fluorescent images were acquired using ECHO RVL-100-G microscope. Confocal micrographs were captured using Olympus SpinSR10 Ixplore system at the Live Cell Imaging Core at Institut Pasteur of Shanghai.

All data were expressed as the mean ± SEM. Statistical analyses were performed using GraphPad Prism7.0 (GraphPad Software, Inc., La Jolla, CA, USA). A two-tailed unpaired Student’s t-test was performed to assess the statistical significance between two independent groups. A p value of 0.05 was set as the significance threshold. The predicted numbers of independent replicates per group were estimated based on independent preliminary studies and experience.

To evaluate the effects of HFD on muscle health in the setting of sarcopenic obesity, standard chow diet-fed 6-week old C57BL/6 female mice were ovariectomized (OVX) or sham-operated. Two weeks after the operations, mice were divided into four groups: sham + chow, sham + HFD, OVX + chow, and OVX + HFD, and the animals were kept on indicated diets for another 8 weeks (Figure 1A). As expected, OVX and HFD significantly increased body weight on female mice (Figure 1B), and HFD feeding led to additional body weight gain of OVX mice (Figure 1B). Of note, 10 weeks after the operations, OVX led to ∼10% loss of lean mass in female mice, while HFD did not show additional effects on muscle wasting in OVX mice (Figure 1C). H&E staining of gastrocnemius muscle sections confirmed the reduction in cross-sectional area (CSA) in OVX mice, but did not indicate that HFD caused apparent morphological difference compared to chow-fed mice subjected to the same operations (Figures 1F, G). However, despite of the minimal impact of 8-week HFD feeding on muscle wasting in OVX mice, qPCR measurements of macrophage marker genes and expression of pro-inflammatory cytokine genes in gastrocnemius muscle revealed heightened immune cell activation and inflammation levels in HFD-fed OVX mice (Figures 1D, E). Hence, HFD per se did not exacerbate muscle wasting in OVX sarcopenic obesity model, but the muscle inflammation was markedly augmented in OVX fed with HFD.

Diet-induced obesity is associated with remarkable AT expansion and remodeling (1, 26). The manner by which AT remodels determines AT metabolic health and profoundly affects non-adipose organs (e.g., liver and skeletal muscle) and systemic energy homeostasis (20, 21). In obesity, pathological AT remodeling featured by increased levels of inflammation, fibrosis and adipocyte hypertrophy is driven by a variety of key regulatory pathways including HIF1α signaling pathway (22–25). Chemical inhibition of HIF1α by PX-478 is effective to improve AT metabolic health in obese animal models (24, 25). In agreement, PX-478 treatment protected HFD-fed OVX mice against pathological AT remodeling (Supplementary Figure 1). C57BL/6 female mice were kept on chow-diet till 6 weeks of age prior to sham or OVX operations. Two weeks after the operations, mice were switched to HFD feeding for another 8 weeks during which the animals were administrated by vehicle or PX-478 (Supplementary Figure 1A). PX-478 treatment did not lead to difference in body weight or lean mass between sham-operated and OVX mice (Supplementary Figures 1B, C). As expected, PX-478 treatment significantly reduced the inflammation- and fibrosis-related genes in AT isolated from obese OVX mice (Supplementary Figure 1D). Of note, PX-478 administration decreased inflammation gene expression in gastrocnemius muscle of obese OVX mice, despite of the lack of apparent difference of muscle morphology and CSA (Supplementary Figures 1E–G). These data implied a potential link between improved AT metabolic remodeling which benefits muscle health in the setting of sarcopenic obesity.

To determine that healthy AT remodeling improves muscle inflammation in HFD-fed OVX mice, we employed an adipocyte-specific HIF1α loss-of-function genetic model. Mature adipocyte HIF1α activation is a potent regulator of fibrogenic and pro-inflammatory gene program (22, 24). Genetically modified mice in which Hif1a is constitutively inactivated in Adipoq-expressing mature adipocytes are protected from unhealthy AT remodeling in diet-induced obesity (22, 23). First, we detected slightly increased HIF1α protein levels in retroperitoneal AT isolated from OVX mice, and greater increase of HIF1α abundance was observed in HFD-fed sham and OVX female mice (Figure 2A). To assess whether improved AT metabolic health caused by adipocyte HIF1α inactivation would have beneficial effects on muscle health in the setting of sarcopenic obesity, we generated a mouse model in which adipocyte Hif1a was inactivated in a doxycycline (DOX)-inducible manner (Figure 2B). This genetic mouse model consists of a transgene that expresses the reverse tetracycline transactivator (rtTA) under the control of a 5.4 kb promoter sequence from adipoq locus, another transgene in which Cre recombinase is expressed from a promoter containing Tet-response element (TRE-Cre), and two “floxed” (loxP-flanked) Hif1a alleles (Adipoq^rtTA; TRE-Cre; Hif1a^loxP/loxP mice, herein denoted as Hif1a-iAKO mice) (Figure 2B). Female Hif1a-iAKO mice and littermate controls (Adipoq^rtTA; Hif1a^loxP/loxP mice) were kept on chow diet until 6 weeks of age prior to OVX or sham operations. 2 weeks after the operations, mice were switched to DOX (600 mg/kg)-containing HFD diet for another 8 weeks (Figure 2C). Immunoblot analysis was used to determine the inactivation of HIF1α in retroperitoneal AT in Hif1a-iAKO mice following the DOX treatment (Figure 2D). Control and Hif1a-iAKO mice subjected to OVX gained extra body weight to a comparable extend after DOX-HFD feeding (Figure 2E). Immunofluorescence staining of AT sections isolated from Hif1a-iAKO mice showed reduction of crown-like structure (CLS) accumulation compared to control mice after HFD-feeding, and Hif1a deletion in adipocytes induced the decreased levels of IL6 in retroperitoneal AT, which are both indicative of improved AT immune cell infiltration and inflammation (Figures 2F, G). Indeed, the expression levels of inflammation-related genes (Tnf, Il6, Saa3, and Ccl2) were significantly decreased in ovariectomized Hif1a-iAKO mice after HFD-feeding (Figure 2H). These data jointly proved that adipocyte Hif1a inactivation led to improved AT inflammation in OVX mice fed with HFD.

Many genetic models have provided evidence that healthy AT remodeling, independent of decrease in adiposity, benefits non-adipose metabolic organs in the contexts of limited release of deleterious pro-inflammatory factors from dysfunctional AT, and the reduction of circulating lipid levels (27–34). In OVX Hif1a-iAKO mice, the serum levels of TNFα and triglycerides (TG) were significantly lower compared to those in control mice after HFD feeding (Figures 3A, B). These metabolic benefits did not improve lean mass loss in HFD-fed OVX Hif1a-iAKO mice (Figure 3C), but led to marked reduction of muscle inflammation reflected by the decreased expression of macrophage markers and inflammation-related genes in gastrocnemius muscle isolated from ovariectomized control and Hif1a-iAKO mice after HFD-feeding (Figures 3D, E). In alignment, the tissue levels of pro-inflammatory cytokine TNFα and IL6 were also significantly suppressed in muscle samples from ovariectomized control and Hif1a-iAKO mice after HFD-feeding (Figures 3F, G). Of note, these effects were largely absent in sham-operated Hif1a-iAKO mice compared with their control counterparts (Figures 3D–G), implying that the protective effects produced by improved AT metabolic remodeling on muscle inflammation are selectively more robust in the setting of greater HIF1α activation (Figure 2A).

In addition to its effects on AT production of deleterious factors, improved AT remodeling also leads to elevated levels of metabolically beneficial factors (2, 15). The best example is the increased levels of adiponectin (APN), an anti-inflammatory and insulin-sensitizing adipokine, in genetic models with improved AT metabolic health (17, 27, 31), and in metabolically healthy obese individuals whose AT exhibits features of healthy remodeling (35–37). In agreement, our OVX Hif1a-iAKO mice preserved similar serum APN levels compared to those in sham control mice (Figure 4A), and mRNA level of Adipoq in fractionated adipocytes were upregulated in both Hif1a-iAKO groups (Figure 4B), which prompted us to speculate that the increased circulating APN could contribute to the protection against muscle inflammation and APN chemical agonist could be effective in alleviating muscle conditions in this setting of sarcopenic obesity. Therefore, we next set to determine if the administration of adiponectin receptor agonist AdipoRon (38) is effective to recapitulate, at least in part, the protective effects on skeletal muscle in obese OVX mice. Therefore, a cohort of C57BL/6 female mice were kept on chow-diet till 6 weeks of age prior to sham or OVX operations. Two weeks after the operations, mice were switched to HFD feeding for another 8 weeks during which the animals were administrated by vehicle or AdipoRon (5 mg/kg) twice a week (Figure 4C). During this course, the administration of AdipoRon did not show obvious effects on body weight gain or lean mass loss in OVX mice (Figures 4D, E). But, the qPCR analysis of macrophage marker and inflammation-related genes in gastrocnemius muscle showed that AdipoRon administration was effective to ameliorate muscle inflammation in OVX mice after HFD feeding (Figures 4F, G), which is associated with reduced levels of pro-inflammatory cytokines TNFα and IL6 in muscle samples isolated from AdipoRon treated OVX mice after HFD feeding (Figures 4H, I). Hence, pharmacological adiponectin receptor activation effectively improved muscle inflammation in OVX mice fed with HFD.