Impact of time‐restricted feeding on metabolic health and adipose tissue metabolism in aged female mice with high‐fat diet‐induced obesity

All procedures were approved by the University of Minnesota Institutional Animal Care and Use Committee (IACUC 2102A38852) and followed NIH guidelines for laboratory animal care. Animal killing procedure: anaesthetics were not used for animal killing in this study. Instead carbon dioxide (CO₂) inhalation was employed in accordance with protocols approved by the IACUC at the University of Minnesota. Compressed CO₂ gas was supplied from cylinders and regulated using a pressure‐reducing regulator and flow meter. Animals were placed in a euthanasia chamber, and 100% CO₂ was introduced at a fill rate of 30%–70% of the chamber volume per minute to displace ambient air. Unconsciousness was confirmed by the absence of respiration and the presence of corneal opacification. CO₂ flow was maintained for at least 1 min beyond the cessation of respiration to ensure complete euthanasia. After this cardiac puncture was performed for blood collection. When necessary cervical dislocation was used as a secondary method to confirm death and facilitate tissue collection.

Animals were housed at 22°C in a specific pathogen‐free facility at the University of Minnesota. The study included 27 aged (14–18‐month‐old) female C57BL/6 mice (The Jackson Laboratory, Bar Harbor, ME, USA). Twenty mice were fed a HFD for 12 weeks to induce obesity, whereas seven mice remained on an ad libitum normal chow diet (control). After 12 weeks HFD‐fed mice were divided into two groups: 10 continued on ad libitum HFD (HFD‐AL) group and 10 were switched to time‐restricted feeding (HFD‐TRF) group. In addition 11 female mice at 3 months of age were enrolled in a middle‐aged study. These consisted of four chow‐fed controls, four HFD‐AL and three HFD‐TRF. After 12 weeks of HFD feeding these mice reached 6 months of age and were either maintained on ad libitum HFD (HFD‐AL) or switched to time‐restricted feeding (HFD‐TRF) for additional 10 weeks. An experimental timeline for both studies is shown in Fig. 1. The AL groups had constant access to food, and the TRF group had access to food 10 h/day during the active (dark) period. The mice were housed in groups of 3–4 per cage, with water ad libitum and in 12 h light/dark cycles with lights off from 12–24 ZT (8 PM–8 AM). The HFD‐TRF group had food available from 8:30 PM–6:30 AM. The high‐fat diet provided to the mice was a 60% HFD (Bio‐Serv: F3282), and the normal chow was provided by the animal facility (Envigo: 2918). According to the manufacture (https://www.bio‐serv.com/product/HFPellets.html↗) the HFD contains 20.5% protein, 36% fat and 35.7% carbohydrate compared to the normal diet, which contains 20.5% protein, 7.2% fat and 61.6% carbohydrate. Regarding the type of fat the diet contains lard as its primary fat source. Glucose tolerance test (GTT) was conducted on week 6, and insulin tolerance test (ITT) was conducted on week 7. The mice were placed in metabolic cages at week 8. Food intake was measured over three consecutive days. After 10 weeks of the dietary intervention all three groups were killed following a 16 h fast. The mice were killed, and blood was collected through cardiac puncture. Brown adipose tissue, inguinal adipose tissue, epididymal adipose tissue, retroperitoneal fat, kidneys, liver and muscle were collected and weighed. Tissue section from brown adipose tissue and liver was used for histological analysis, whereas small portions of inguinal and epididymal adipose tissue were used for both histological analysis and fat cell sizing, and the remaining tissue was snap‐frozen in liquid nitrogen and stored for later analysis.

Body composition (fat mass and fat‐free mass, both in grams, Echo MRI 3‐in‐1, Echo Medical System) was measured. Mice were individually caged at week 8 of the TRF regimen, and food intake was measured over a 5‐day period using a Biodaq food intake monitoring system (Research Diets Inc., New Brunswick, NJ). The final 2 days of food intake monitoring were averaged to determine total food intake in grams/day. Indirect calorimetry (oxygen consumption (VO₂), carbon dioxide production (VCO₂), heat, respiratory efficiency ratio (RER) and activity) was measured over a 3‐day period using the Oxymax Comprehensive Lab Monitoring System (Columbus Instruments, Columbus, OH, USA). The final light and dark cycles were used to statistically determine differences in energy expenditure between experimental groups.

Adipose tissue was obtained from inguinal and epididymal WAT of mice fed control or HFD with or without TRF. Tissue samples (25–30 mg) were immediately fixed in 12 ml of the 2% osmium tetroxide solution in collidine buffer in a Wheaton vial (SPI‐Chem no. 986704) and incubated in a water‐bath at 37°C for 48 h, as described previously (Hirsch & Knittle, 1970; McLaughlin et al., 2007). Collidine buffer was prepared from a 4°C stock collidine buffer of 0.2 M 2,4,6‐trimethyl‐pyridine (C‐0505; Sigma Chemical Co) dissolved in distilled water. Subsequently the contents of the vial were washed out using a 25 µm filter to catch the fixed cells, and the container was rinsed three more times. After a final rinse using 0.9% saline, the material was transferred into a 250 µm filter using a squirt bottle filled with 0.9% saline. A gentle rub on the 250 µm mesh was applied to crush any large chunks and then rinsed again with 0.9% saline. The procedure was repeated to collect the cells. The end volume did not exceed more than one conical tube (50 mL). Samples were analysed with an optical particle analyse (model: Microtrac SIA), which is a particle characterization tool. The Bluewave can measure particles in the size range 50 nm to 2800 µm. Each sample was measured in duplicate. Figures were generated using GraphPad Prism version 9.5.1 for Windows (GraphPad Software, San Diego, CA, USA).

GTT was conducted at week 6 of the TRF regimen. The mice were fasted for 16 h. Body weight was recorded. Bolus of glucose (0.75 g/kg body weight) was prepared from a filter‐sterilized d‐glucose solution (300 mg/ml in saline) and administered intraperitoneally at 2.5 µl per gram body weight. Fasting blood glucose was measured from the tail vein prior to glucose intraperitoneal injection. Blood glucose was measured at 15, 30, 45, 60, 90, 120 min intervals. ITT was conducted at week 7 of the TRF regimen. The mice were fasted for 4 h prior to injection. Body weight was recorded. A bolus of insulin (0.75 U/kg body weight) was prepared from a diluted stock solution of regular human insulin (Humulin R; Lilly NDC 0002‐8215‐01). Fasting blood glucose was measured prior to insulin intraperitoneal injection. Blood glucose was measured at 15, 30, 45, 60, 90, 120 min intervals.

Tissues were fixed in 10% neutral buffered formalin (VWR International, LLC, Radnor, PA, USA) for 2–3 days, dehydrated in 70% ethanol solutions for 3–5 days and processed for embedding in paraffin. Tissue samples were haematoxylin and eosin (H&E) stained using a standard protocol at the University of Minnesota Histology Core. Briefly after deparaffinization and rehydration tissues were sectioned with 5–6 µm thickness and stained in haematoxylin for 1 min and rinsed with distilled water. After haematoxylin staining tissues were counterstained with eosin solution for 1 min followed by dehydration through 95% and 100% EtOH and xylene clearance. At last the tissue sections were mounted with resinous mounting medium. Images were then captured using a Leica DM IL microscope.

Total RNA was extracted from frozen tissue using TRIZOL reagent (Invitro, Carlsbad, CA, USA) and treated with DNAase to remove genomic DNA prior to cDNA synthesis using Superscript II reverse transcription kit (Invitrogen). Real‐time quantitative PCR was conducted using FastStart Universal SYBR Green Master (Rox) (Roche) with a QuantStudioTM 3 Real‐Time PCR System (Applied Biosystem, Foster City, CA, USA). The ∆∆Ct method was used to calculate mRNA expression. For quantification TATA‐box binding protein (Tbp) mRNA served as an endogenous control within inguinal and brown adipose tissue. The primer sequences for amplifying the target genes are shown in Table 1.

Forward (F) and reverse (R) primer sequences for each gene are listed along with gene name and gene symbol. Primers were validated for specificity using melt curve analysis.

Serum triglyceride level was determined using Triglycerides Enzymatic Assay Kit (Stanbio Laboratory, Boerne, TX, USA; #2100430). Serum‐free fatty acids and β‐hydroxybutyrate levels were determined using a Free Fatty Acid Quantitation Kit (Sigma‐Aldrich, # MAK044) and a β‐Hydroxybutyrate Assay Kit (Sigma‐Aldrich, # MAK041) according to the manufacturers’ instructions.

Results are expressed as mean ± SD. Data from aged and middle‐aged female mice were analysed using Student's t test when comparing two groups, and one‐way ANOVA when comparing more than two groups, using GraphPad Prism (version Prism 9.4.1). Statistical significance was defined as a P‐value < 0.05.

To investigate the beneficial effects of TRF on metabolic health in aged mice we placed mice at the age of 14–18 months on a HFD for 12 weeks followed by either HFD‐AL or HFD‐TRF for an additional 10 weeks. During the TRF period the HFD‐AL group showed no significant differences in weight gain compared to the control group, whereas the HFD‐TRF group exhibited significantly reduced weight gain relative to both the control and HFD‐AL groups (Fig. 2A). Specifically at weeks 2, 4, 6 and 7 body weight change in the HFD‐TRF group was significantly lower than in the HFD‐AL group (Fig. 2A). Both average body weight and total fat mass were significantly higher in HFD‐AL mice compared to controls (Fig. 2B and C). TRF partially reversed this HFD‐induced increase, though levels remained significantly elevated compared to the control group (Fig. 2B and C). Lean mass did not differ among groups (Fig. 2D) in terms of adipose tissue weight (as a percentage of body weight); the HFD‐AL group showed significantly increased inguinal WAT (Ing‐WAT) and gonadal WAT (Gon‐WAT) compared to controls, with no changes in BAT or perirenal WAt (P‐WAT) (Fig. 2E). The HFD‐TRF group showed significant increases in BAT, Ing‐WAT and Gon‐WAT compared to controls, with no change in P‐WAT (Fig. 2E). However there were no significant differences in adipose tissue as a percentage of body weight between the HFD‐AL and HFD‐TRF groups, except for a trend towards a decrease in Ing‐WAT in the HFD‐TRF group (Fig. 2E). Organ weights were also affected. HFD‐AL mice had significantly increased weights of kidney, liver and heart compared to controls (Fig. 2F). TRF significantly reduced heart weight and showed a trend towards lower liver and kidney weights compared to HFD‐AL mice (Fig. 2F). Histological analysis of the liver revealed marked lipid accumulation in the HFD‐AL group, indicating significant ectopic fat deposition within the liver (Fig. 2G). HFD‐TRF mice showed a noticeable reduction in hepatic lipid accumulation compared to HFD‐AL mice, although levels remained elevated relative to controls (Fig. 2G).

Overall food intake was comparable among groups, with slightly increased intake observed in both the HFD‐AL and HFD‐TRF groups (Fig. 3A). During the feeding window the HFD‐TRF group consumed significantly more food in both 0–5 h and 5–10 h intervals compared to the control and HFD‐AL groups (Fig. 3B). In contrast during the inactive phase when HFD‐TRF mice had no food access, the control and HFD‐AL mice consumed roughly 30% of their daily intake (Fig. 3B). Meal frequency analysis revealed a significant increase in the HFD‐TRF group compared to the control group during both 0–5 h and 5–10 h periods (the designated feeding windows for TRF) (Fig. 3C). Meal frequency in the HFD‐TRF group was also significantly higher than the HFD‐AL group during the 0–5 h period but not during 5–10 h (Fig. 3C). During the inactive period HFD‐AL mice exhibited higher meal frequency compared to controls (Fig. 3C). No significant differences in meal size were observed among groups, although a downward trend was noted in the HFD‐AL group (Fig. 3D). As shown in Fig. 3E meal duration (time per meal) was significantly reduced in both HFD‐AL and HFD‐TRF groups compared to controls, likely due to the higher caloric density of the HFD.

Indirect calorimetry analysis showed that both VO₂ and VCO₂ were significantly reduced during both the light and dark cycles in the HFD‐AL and HFD‐TRF groups compared to controls (Fig. 4A and B). However TRF significantly increased VO₂ and VCO₂ relative to the HFD‐AL group, although values remained lower than those in the control group (Fig. 4A and B). The RER followed a similar pattern during the light cycle (0–12 ZT), with the HFD‐TRF group showing an even greater reduction compared to the HFD‐AL group. During the dark phase (12–24 ZT) both HFD‐AL and HFD‐TRF groups exhibited reduced RER compared to controls, with no significant difference between the two groups (Fig. 4C). Activity levels during the light cycle significantly decreased in both HFD‐AL and HFD‐TRF groups of mice compared to controls (Fig. 4D). During the dark cycle when mice are typically more active, the HFD‐AL group showed a reduction in activity relative to controls, although not significantly. This reduction was partially reversed by TRF (Fig. 4D). Heat production increased in both HFD‐AL and HFD‐TRF groups compared to controls during both the light and dark cycles. Notably heat production in the HFD‐TRF group was lower than that in the HFD‐AL group, tending back towards the control level (Fig. 4E). These results suggest that TRF can partially counteract the HFD‐induced reduction in energy expenditure in aged female mice.

To understand how TRF influences adipose tissue plasticity adipocyte size distribution was assessed in Ing‐WAT and Gon‐WAT. In Ing‐WAT HFD feeding led to a marked increase in the 150–200 µm size adipocyte population compared to controls (Fig. 5A). TRF significantly reduced the 75–100 µm adipocyte population and increased the 150–200 µm population compared to controls (Fig. 5A). Additionally TRF significantly increased the percentage of 25–50 µm adipocytes compared to HFD‐AL mice (Fig. 5A). Overall the HFD‐AL group exhibited a significant increase in average adipocyte diameter compared to both the control and HFD‐TRF groups (Fig. 5B). Although TRF partially reversed this enlargement, adipocytes in the HFD‐TRF group remained significantly larger than those in the control group (Fig. 5B). These morphological changes were confirmed by H&E staining of Ing‐WAT sections (Fig. 5C).

In contrast more pronounced changes were observed in the Gon‐WAT depot. In the HFD‐AL group the 75–100 and 100–150 µm adipocyte populations were significantly decreased compared to the control group, whereas the larger adipocyte population (150–200 µm) was increased (Fig. 5D). TRF significantly increased the percentage of both 25–50 and 150–200 µm adipocyte populations compared to the control and HFD‐AL groups, while reducing the 75–150 µm populations (Fig. 5D). In terms of average adipocyte diameter in Gon‐WAT both the HFD‐AL and HFD‐TRF groups showed significant increases compared to the control group (Fig. 5E). H&E staining supported these findings, showing enlarged adipocytes in the HFD‐AL group relative to the control group, with partial reversal observed in the HFD‐TRF group (Fig. 5C). These data suggest that TRF promotes a more heterogenous adipocyte population in Gon‐WAT, with a mixture of smaller and larger adipocytes, whereas HFD‐AL feeding results in a more homogeneous population dominated by larger adipocytes (Fig. 5D).

To assess the impact of TRF on metabolic health we evaluated glucose and lipid homeostasis, along with insulin sensitivity. GTT and ITT revealed distinct metabolic responses among the groups. The HFD‐AL group exhibited a modest impairment in glucose tolerance compared to the control group, as indicated by elevated blood glucose levels during the GTT (Fig. 6A). Insulin sensitivity was significantly reduced in the HFD‐AL group, showing higher blood glucose levels during the ITT (Fig. 6B). However TRF did not improve glucose tolerance but showed a trend towards enhanced insulin sensitivity in aged female mice on HFD (Fig. 6B). Fasting blood glucose levels at the time of killing were significantly elevated in the HFD‐AL group compared to the control group (Fig. 6C). TRF partially reversed this increase, though not to control levels (Fig. 6C). Serum free fatty acids (FFAs) and triglycerides trended upward in the HFD‐TRF group relative to the control group (Fig. 6D and E), but no significant differences were observed between HFD‐AL and HFD‐TRF groups. Additionally serum β‐hydroxybutyrate levels did not differ significantly among the three experimental groups (Fig. 6F).

We next examined the effects of TRF on BAT metabolism, focusing on the expression of key genes involved in mitochondrial function, lipid metabolism and thermogenesis. The histological analysis showed increased lipid droplet size in BAT from the HFD‐AL group, which was significantly reversed by TRF (Fig. 7A). Gene expression analysis showed no significant changes in thermogenic genes such as Ucp1, Cidea and Tfam across the groups (Fig. 7B). However genes involved in mitochondrial biogenesis and oxidative metabolism were significantly altered. Specifically Erra, Atp5b and Cpt1 were significantly upregulated by HFD‐AL, whereas TRF reversed the expression of Erra and Atp5b and downregulated Pgc1a expression (Fig. 7D). HFD‐AL also reduced the expression of lipogenic and lipolytic genes, including Scd1, Elovl5 and Atgl (Fig. 7E). TRF partially or fully restored the expression of these genes, suggesting that TRF can reverse HFD‐induced disruption of lipogenesis‐lipolysis futile cycle in BAT. Interestingly TRF also restored the expression of circadian rhythm‐related genes altered by HFD. Bmal1 gene expression in BAT was suppressed by HFD relative to controls but was restored to control levels by TRF (Fig. 7C). In contrast Per1 expression was elevated by HFD and reduced back to control levels with TRF (Fig. 7C). These findings suggest that TRF can counteract HFD‐induced impairments in mitochondrial metabolism, lipid cycling and circadian regulation in BAT of aged female mice.

We next examined the expression of genes involved in adipogenesis, lipid metabolism, senescence, inflammation and circadian rhythm in Gon‐WAT. Expression of adipogenic genes, including Pparg and its downstream targets Lpl and Glut4, was significantly increased by TRF compared to both the control and HFD‐AL groups (Fig. 8A). Srebp‐1c, a key transcription factor for lipogenesis, was significantly upregulated by HFD and further increased by TRF compared to the control (Fig. 8B). Although Fasn expression was suppressed by HFD, TRF significantly elevated its expression (Fig. 8B). Regarding lipolytic genes Atgl expression was significantly downregulated by HFD relative to the control (Fig. 8C), but TRF restored it towards control levels (Fig. 8C). Hsl expression did not show significant changes (Fig. 8C).

We also assessed circadian rhythm gene expression in Gon‐WAT. Both Clock and Bmal1 were significantly decreased in the HDF‐AL group compared to the control (Fig. 8D), and TRF did not restore their expression (Fig. 8D). Per1 expression showed an increasing trend with HFD, which was significantly reduced by TRF (Fig. 8D). For genes associated with senescence, SASP and inflammation p16 and p21 (cellular senescence markers) were significantly upregulated by HFD (Fig. 8E), and TRF failed to reverse these changes (Fig. 8E). The pro‐inflammatory cytokines Tnfa and Mcp1 were elevated in the HFD‐AL group (Fig. 8E), but TRF partially mitigated these increases (Fig. 8E). Overall these findings indicate that TRF can modulate gene expression in Gon‐WAT by upregulating adipogenic genes and partially reversing mitigating HFD‐induced changes in lipogenesis, lipolysis and inflammation. However TRF had limited effects on senescence markers and circadian rhythm genes, highlighting the complex, depot‐specific interactions between diet, metabolic health and circadian regulation.

To explore how TRF affects metabolic adaptations in Ing‐WAT we analysed the expression of genes involved in adipogenesis, lipid metabolism, mitochondrial function, senescence, inflammation and circadian rhythm. Pparg and its target gene Glut4 was significantly upregulated by HFD, whereas Lpl showed no significant change (Fig. 9A). TRF significantly increased Lpl expression beyond control levels and showed a trend towards increased Glut4 (Fig. 9A), indicating a partial reversal of HFD‐induced adipogenic impairment. Lipogenic gene Fasn and lipolytic genes Atgl and Hsl were significantly downregulated by HFD (Fig. 9B). TRF reversed the suppression of Fasn and Atgl but had no significant effect on Hsl expression (Fig. 9B). For mitochondrial function Atp5b and Cpt1 were significantly upregulated by TRF compared to both the control and HFD‐AL groups (Fig. 9C), suggesting enhanced mitochondrial oxidation and ATP production in Ing‐WAT. However TRF had no reversal effect on the HFD‐induced downregulation of Pgc1a gene (Fig. 9C).

The assessment of circadian rhythm genes revealed that the expression of Clock and Per1 genes was decreased by HFD (Fig. 9D), whereas Bmal1 was significantly upregulated (Fig. 9D). TRF did not significantly alter these expression patterns (Fig. 9D). Collectively these findings suggest that TRF improves lipogenesis‐lipolysis cycling and enhances mitochondrial oxidative capacity in Ing‐WAT, although it has limited effects on circadian gene expression in aged female mice.

To determine whether ageing impacts the effectiveness of TRF we also investigated the effects of TRF in middle‐aged female mice. As shown in Fig. 10 the HFD‐AL group exhibited significant weight gain in middle‐aged female mice (Fig. 10A and B), whereas body weight in aged HFD‐fed females remained stable throughout the TRF regimen (Fig. 2A). Consequently TRF induced more pronounced weight changes in middle‐aged female mice (Fig. 10A) compared to their HFD‐AL counterparts in aged female mice (Fig. 2A). Furthermore TRF led to a greater reduction in adipose tissue mass, with significant decreases observed in Ing‐WAT, Gon‐WAT and perirenal WAT (Fig. 10C). However in contrast to aged female mice TRF did not significantly affect organ weights in middle‐aged females (Fig. 10D). These differences may reflect age‐related variations in metabolic adaptations, which could influence responsiveness to dietary interventions such as HFD and TRF.

The datasets generated and analysed during the current study are available in Supporting Information. Additional data and analysis are available from the corresponding author on reasonable request.