Beta-Hydroxybutyrate May Protect the Aging Liver Like Exercise

SPF‐grade male C57BL/6J mice were procured from Gempharmatech (Jiangsu Jicui Yaokang Biotechnology Co. Ltd., China). The animals were categorized into two age groups: 6‐week‐old mice, housed until 4 months of age, constituted the young control (YC) group, whereas 6‐month‐old mice were kept until 15 months of age and subsequently randomized into four subgroups: old control (OC), old exercise (OE), old intervention with placebo (OIPA), and old intervention with β‐HB (OIPB). The mice were kept at ambient temperature (23°C ± 2°C) with a humidity of 50% ± 5%, a 12:12 h light–dark cycle, and unrestricted access to potable water and normal laboratory diet. Experimental interventions began at 15 months of age for the older cohorts, whereas the YC group received concurrent analyses to establish baseline comparisons. All animal research received approval from the Animal Care and Use Committee of Shanghai University of Sport (Ethics no: 102772023DW022).

The treadmill exercise program consists of 1 week of familiarization, followed by a 16‐week moderate‐intensity aerobic regimen, comprising 5 days per week of 1 h of continuous treadmill running at 15 m/min, with no slope (Donato et al. 2005). Throughout the treadmill training, we continuously monitored to ensure that all mice were running and not resting on the treadmill. This would ensure that all mice underwent an equivalent level of exercise training. The YC and OC groups remained inactive under the same housing conditions.

The OIPB group got intraperitoneal injections of β‐HB (200 mg/kg body weight), while the OIPA group was administered an equivalent volume of saline (Tan et al. 2022). Injections were given for 5 days per week for a duration of 16 weeks. β‐HB (Catalog no. H6501; Sigma, USA) was solubilized and diluted in sterile saline prior to application.

To evaluate the alterations in blood β‐HB levels induced by exercise, mice underwent a 1‐h treadmill session at a speed of 15 m/min. Blood was obtained from the tail vein before injection and at intervals of 15 min, 30 min, 1, 3, and 24 h post‐exercise, and the concentration of β‐HB was assessed using an Abbott ketone meter (FreeStyle Optium, YZB/UK 5991‐2012; Abbott, UK) (Roberts et al. 2017).

AML12 mouse hepatocyte cells were cultured in DMEM/F‐12 with high glucose, 10% FBS, insulin‐transferrin‐selenium, 0.1 μM dexamethasone, and 1% penicillin–streptomycin. Raw264.7, mouse leukemia monocyte macrophage cells, were cultured in DMEM containing high glucose, 10% FBS, and 1% penicillin–streptomycin. AML12 and Raw264.7 were cultivated under usual conditions of 37°C, 95% air, and 5% CO₂. A 1:3 ratio was employed for cell passage, utilizing cells in the exponential development phase for experimentation.

To develop a Dgal‐induced cellular aging model, cells were exposed to Dgal at several concentrations (50, 75, 100, 150, and 200 mM) for 24 h, and cytotoxicity was evaluated to identify the best concentration. β‐HB treatments (2.5, 5, 10, and 20 mM) were administered to assess dose–response effects and determine the best protective concentration. The experimental design for evaluating the STING signaling pathway contained the following conditions: control (Con), Dgal (100 mM, Dgal), Dgal + β‐HB (100 mM + 5 mM, DgalBHB), Dgal + DMXAA (150 μM, STING agonist), Dgal + β‐HB + c‐176 (1 μM, STING antagonist), and Dgal + β‐HB + DMXAA, utilizing Raw264.7 cells. Dgal and BHB were concurrently introduced into the cell culture medium at the specified concentration. DMXAA or c‐176 should be included in the culture medium 1 h prior, if required. Cells underwent treatment for 24 h prior to subsequent examination.

After the culture of RAW264.7 in standard medium (Con), Dgal (100 mM, Dgal), and Dgal + BHB (100 mM + 5 mM, DgalBHB) for 24 h, the cell supernatants were collected and centrifuged at 500 g for 10 min at a low temperature of 4°C. The conditioned medium (CM) was acquired following filtration through a 0.22 μm pore filter. Raw264.7‐CM was combined with double the volume of standard complete medium of AML12 to produce CM‐Con, CM‐Dgal, and CM‐DgalBHB, which were subsequently utilized to culture AML12 cells for 24 h.

Serum parameters were assessed with ELISA kits in accordance with the manufacturer's guidelines. The Alanine aminotransferase (ALT) Test Kit (C009‐2‐1), Aspartate aminotransferase (AST) Test Kit (C010‐2‐1), Triglyceride (TG) Assay Kit (A110‐1‐1), Total Cholesterol (TC) Assay Kit (A111‐1‐1), Low‐density lipoprotein cholesterol (LDL‐C) determination reagent (A113‐1‐1), and High‐density lipoprotein cholesterol (HDL‐C) determination kit (A112‐1‐1) were acquired from JianCheng Bioengineering Institute.

The tissues were fixed in 4% paraformaldehyde and sectioned (5 μm) after being paraffin embedded. H&E staining and immunofluorescence were performed as described previously (Olympus microscope) (Zheng et al. 2023). For immunofluorescence, the cells were initially fixed in 4% paraformaldehyde for 10 min, thereafter rinsed three times with PBS, and then permeabilized with 0.2% Triton X‐100 for 5 min. The cells were washed with PBS, blocked with 10% BAS/PBS for 1 h, then incubated with primary antibodies at 4°C overnight. After three rinses with PBS, secondary antibodies were applied at room temperature for 1 h. The cells were subsequently rinsed thrice with PBS and mounted utilizing an anti‐fade mounting solution containing DAPI. The immunofluorescence‐stained cells were examined using a Leica microscope, and fluorescence intensity was assessed with ImageJ software.

Fresh liver tissue was wrapped in optimal cutting temperature compound, frozen, and sectioned (8 μm). Oil Red O staining was performed as described previously (Olympus microscope) (Wang et al. 2022). AML12 cells were subjected to three washes with cold PBS and subsequently fixed for 20 min using 4% formaldehyde. Oil Red O (Sigma; 0.5% in isopropanol) was diluted with water in a 3:2 ratio, filtered through a 0.45‐μm filter, and incubated with the fixed cells for 30 min at room temperature. The cells were rinsed with water, and the stained lipid droplets within the cells were analyzed using light microscopy and captured in photographs.

According to the manufacturer's instructions, total RNA was extracted from liver tissues (20–30 mg) or cell samples utilizing Trizol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA), followed by quantification of the RNA concentration using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, San Jose, CA, USA). RNA was reverse transcribed with a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) to synthesize cDNA for qPCR using SYBR Green PCR Master Mix (Vazyme Biotech Co. Ltd., Nanjing, China) in StepOnePlus (Applied Biosystems, Carlsbad, CA, USA). The mRNA expression content of target genes was standardized to the expression of β‐actin or 18 s, and was calculated with themethod. The mitochondrial RNA and cytosolic RNA were isolated by Tissue Mitochondria Isolation Kit (Beyotime, C3606). 2 − ∆ ∆ C T

Tissues and cells were harvested, prepared, and Western blotting was performed as described previously (Guo et al. 2024). The quantification of western blotting results was performed utilizing ImageJ. The BCA assay was employed for protein quantification. Proteins were denatured by combining them with 5× loading buffer. Western blotting was conducted by separating proteins at 90 V for 30 min, followed by 120 V for 1 h, then transferring them onto PVDF membranes at 110 V for 1.5 h. Incubation of primary and secondary antibodies was conducted on membranes. Protein signals were detected using chemiluminescence, and the images were processed with ImageJ software.

Cells were allocated into 96‐well plates at an approximate concentration of 1 × 10⁴ cells per well, with each well supplemented with 100 μL of complete media. Following a 24‐h incubation at 37°C in 5% CO₂, the growth medium was replaced with either fresh complete medium or medium augmented with test chemicals at specified quantities. The cells were incubated for an additional 24 h, after which the medium was replaced with 100 μL of CCK‐8 solution diluted to one‐tenth of its original concentration. Thereafter, the plate underwent a 2‐h incubation, and the optical density at 450 nm was measured using a spectrophotometer.

Cells were allocated into 6‐well plates at an approximate concentration of 2 × 10⁵ cells per well and stained with the senescence β‐galactosidase (SA‐β‐Gal) Staining Kit (Beyotime, C0602) in accordance with the manufacturer's guidelines.

To detect the level of free mtDNA in the cytosol of Raw264.7 cells, cells were collected following experimental treatments. Mitochondria were isolated using the Cell Mitochondria Isolation Kit (Beyotime, C3601) to obtain the cytosolic fraction. Subsequently, gDNA was extracted from the cytosolic fraction using the TIANamp Genomic DNA Kit (TIANGEN, DP304). Finally, the DNA levels of the mitochondrial gene Nd1 and the nuclear gene Cftr were determined by quantitative PCR (qPCR). The relative expression level of Nd1 was calculated using the ΔC_t method with Cftr as the reference gene to assess the enrichment of mtDNA in the cytosol.

To assess the apoptosis of AML12 cells, the cells were stained utilizing the Annexin V‐FITC Apoptosis Detection Kit (Beyotime, C1062) according to the manufacturer's guidelines (Zhao et al. 2020). Annexin V‐FITC, which emits green fluorescence, identified early apoptotic cells, while PI, exhibiting red fluorescence, indicated necrotic or late apoptotic cells with damaged membranes.

The publicly accessible microarray dataset GSE133815 was analyzed to assess STING (TMEM173) expression levels in human liver tissues. This dataset contains transcriptomic profiles from 11 young individuals (21–45 years) and 9 aged individuals (61–80 years). Raw data files were retrieved from the NCBI GEO repository (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE133815↗) for comparative analysis of STING expression between age groups.

Data analysis was carried out with SPSS 22.0 and visualized with GraphPad Prism 8.0.2. Independent sample t‐tests were employed for comparing two groups. To compare multiple groups, one‐way ANOVA was conducted, followed by Bonferroni corrections. All data are expressed as mean ± standard deviation (SD). Statistical significance was defined as a p‐value under 0.05.

In order to investigate the dynamic changes in blood β‐HB levels induced by exercise, our study administered a single bout of exercise intervention (15 m/min, 1 h) to aged mice and measured their blood β‐HB levels before exercise and within 24 h post‐exercise. The results indicated that aerobic exercise substantially increased blood β‐HB levels, which returned to the baseline after 24 h (Figure 1a).

This study found that old mice, in contrast to young mice, demonstrated a large increase in body weight and body fat percentage, along with a substantial decrease in lean mass proportion. Aerobic exercise significantly reduced body weight growth and altered body composition in aged mice, notably by significantly reducing the rise in body fat percentage (Figure 1b,c). Subsequent examination of the four blood lipid parameters indicated that aerobic exercise markedly mitigated the heightened levels of serum TG, TC, LDL‐C, and HDL‐C in elderly mice (Figure 1d–g). Moreover, aerobic exercise substantially reduced the elevated blood activity of AST and ALT enzymes in aged mice, suggesting its ability to mitigate aging‐related liver damage (Figure 1h,i). To assess the impact of aerobic exercise on hepatic steatosis in aged mice, this study further measured hepatic TG levels and H&E staining and Oil Red O staining on liver tissues. The findings indicated that aerobic exercise markedly inhibited hepatic lipid deposition and the increase in TG levels in aged mice (Figure 1j,k). Examination of lipid metabolism‐associated genes demonstrated that aerobic exercise markedly decreased the hepatic mRNA expression levels of fatty acid synthesis proteins (Scd1, Fas) and fatty acid transport proteins (Cd36, Fatp5) (Figure 1l). Furthermore, aerobic exercise markedly reduced the hepatic mRNA expression levels of pro‐inflammatory cytokines (Il‐1β, Tgf‐β, Mcp1, Ccl5), PANoptosis‐related molecules (Aim2, Zbp1, Pyrin, Gsdmd, Caspase3, Caspase8, Mlkl, Ripk1, Ripk3), and senescence indicators (p16, p21, p53) in aged mice (Figure 1m–o). The protein level of p16 detected by Western blot was consistent with the qPCR results (Figure 1p,q). These results indicate that aerobic exercise collectively decreased body weight and body fat percentage in aged mice, enhanced hepatic lipid deposition, mitigated inflammatory responses and PANoptosis, consequently providing hepatoprotective effects.

Mitochondrial dysfunction is a fundamental characteristic of aging. This work examined the role of mitochondrial homeostasis in the mechanism by which aerobic exercise reduces hepatic lipid accumulation in aged mice by evaluating the expression levels of markers associated with mitochondrial dynamics by immunofluorescence, qPCR, and Western blotting. Results from immunofluorescence (Figure 2a–d) and Western blotting (Figure 2e–g) indicated that aerobic exercise markedly reduced the protein expression levels of MFN2 and DRP1 in the livers of elderly mice. Aerobic exercise altered the mRNA expression levels of hepatic mitochondrial fusion protein (Mfn2), fission proteins (Drp1, Fis1), and biogenesis protein (Tfam), indicating that it can rectify aberrant mitochondrial dynamics in the livers of aged mice (Figure 2h–k).

Aberrant mitochondrial dynamics resulting in mtDNA release is a direct cause of various cellular damage. Cytosolic double‐stranded DNA (dsDNA) could originate from exogenous viral DNA, the release of nuclear DNA, or the release of mtDNA, with mtDNA being more readily released into the cytosol than nuclear dsDNA. This study employed double immunofluorescence staining to detect the expression and distribution of dsDNA in mouse livers. The findings indicated a substantial presence of free dsDNA in the cytoplasm of hepatocytes from elderly mice; aerobic exercise markedly diminished this aberrant release of dsDNA (Figure 2l).

The cGAS‐STING signaling pathway is recognized as the principal mechanism by which cells detect and respond to cytosolic free dsDNA. Our study found that both the gene (Figure 2m,n) and protein (Figure 2e,o,p) expression levels of cGAS and STING were significantly upregulated in the livers of aged mice, and more importantly, the levels of phosphorylated STING (p‐STING)—a direct indicator of pathway activation—and the p‐STING/STING ratio (Figure 2q,r) were also significantly increased. This confirms the activation of the pathway during aging. Notably, aerobic exercise effectively suppressed the overactivation of the cGAS‐STING signaling pathway, including significantly reducing p‐STING levels and its ratio. To investigate the expression changes of STING (also known as TMEM173) in human liver aging, we analyzed a gene expression profile database (GSE133815) comprising liver samples from 11 young controls and 9 aged individuals. The results demonstrated that STING expression in the liver was significantly upregulated in the aged population compared to the young controls (Figure 2s). This suggests that STING upregulation represents a conserved molecular alteration during liver aging. In liver tissue, STING is primarily expressed in macrophages. Immunofluorescence results indicated that STING colocalized with the macrophage marker Mac‐2, and its expression level was similar to the Western blot results (Figure 2t,u).

STING is closely associated with macrophage polarization, a process that significantly contributes to chronic inflammation during aging. To investigate the regulatory effect of aerobic exercise on hepatic macrophage polarization in aged mice, our study examined the regulatory influence of aerobic exercise on hepatic macrophage polarization in aged mice by measuring the protein expression levels of CD11C and CD206 through Western blotting (Figure 2e,v,w) and immunofluorescence (Figure 2x–a2), as well as evaluating the mRNA expression levels of M1 macrophage markers (Cd11c, Tnf‐α) and M2 macrophage markers (Cd206, Cd163) using qPCR (Figure 2b2–e2). The results indicated that the gene and protein expression levels of M1 markers were upregulated, while those of M2 markers were downregulated in the livers of aged mice; aerobic exercise significantly inhibited the polarization of hepatic macrophages towards the M1 pro‐inflammatory phenotype in aged mice.

In summary, in the livers of aged mice, a disruption in mitochondrial dynamics equilibrium results in elevated mtDNA release. The increased cytosolic free dsDNA induces the cGAS‐STING signaling pathway, promoting macrophage polarization towards the M1 pro‐inflammatory phenotype. Aerobic exercise can restore mitochondrial dynamics balance, inhibit cGAS‐STING pathway activation, and reverse the pro‐inflammatory phenotypic transformation of macrophages.

This work examined whether the exercise‐induced increase of β‐HB exerts its protective benefits on the aging liver by evaluating the direct impact of β‐HB on Dgal‐induced damage in AML12 mice hepatocytes. Initially, AML12 cells were subjected to several amounts of Dgal (50, 75, 100, 150, and 200 mM) or β‐HB (2.5, 5, 10, and 20 mM). Cytotoxicity assays indicated that elevated doses of Dgal were detrimental to AML12 cells (Figure 3a); therefore, 100 mM Dgal was chosen for subsequent treatments. Conversely, β‐HB at concentrations ranging from 2.5 mM to 20 mM exhibited no cytotoxicity (Figure 3b) and even enhanced cell viability; consequently, 5 mM β‐HB was chosen for the experiments. qPCR analysis demonstrated that treatment with 100 mM Dgal markedly increased the mRNA expression levels of senescence markers p16, p21, and p53 in AML12 cells (Figure 3c). SA‐β‐gal staining and Oil Red O staining further validated that Dgal effectively triggered cellular senescence and lipid deposition in AML12 cells (Figure 3d). Furthermore, Dgal treatment elevated the mRNA expression levels of pro‐inflammatory cytokines Il‐1β and Tgf‐β, as well as Sting, although no significant alteration was noted in Cgas mRNA levels (Figure 3e–h). However, co‐treatment with 5 mM β‐HB failed to reverse Dgal‐induced senescence, lipid deposition, or inflammatory responses in AML12 cells. Collectively, these findings indicate that the hepatoprotective effects of β‐HB may not be mediated through its direct action on hepatocytes.

Numerous published studies show that macrophages, such as Kupffer cells and infiltrating macrophages, are the main source of STING expression and activation in the livers of humans and mice (Yang et al. 2024; Luo et al. 2018; Thomsen et al. 2016). Additionally, our investigation revealed that the majority of STING expression is found in macrophages throughout the entire liver tissue.

This investigation aims to ascertain if macrophage senescence correlates with STING pathway activation, M1 polarization, and inflammation. First, Raw264.7 cells were treated with different concentrations of Dgal (50, 75, 100, 150, and 200 mM) to establish a senescence model. CCK‐8 assays confirmed that Dgal doses of 20 mM or below were non‐cytotoxic (Figure 4a). qPCR analysis demonstrated that the mRNA expression levels of Cgas and Sting escalated with increasing Dgal concentrations (Figure 4b,c). Simultaneously, the mRNA expression of the M1 marker Cd11c escalated with concentration, but that of the M2 marker Cd206 diminished (Figure 4d,e). Dgal administration elicited a dose‐dependent elevation of mRNA expression for several pro‐inflammatory cytokines (Il‐1β, Tgf‐β, Mcp1, and Ccl5) (Figure 4f–i). These results indicate that Dgal‐induced macrophage senescence correlates with the increase of the cGAS‐STING signaling pathway, M1 polarization, and inflammatory activity.

Subsequently, to investigate the effect of β‐HB, Raw264.7 cells were exposed to various concentrations of β‐HB (2.5, 5, 10, and 20 mM). CCK‐8 assays indicated that β‐HB concentrations ranging from 2.5 to 20 mM had no cytotoxic effect (Figure 4j). Under the influence of 100 mM Dgal stimulation, β‐HB administration substantially and dose‐dependently reduced the mRNA expression of Cgas and Sting (Figure 4k,l). Simultaneously, β‐HB downregulated the mRNA level of the M1 marker Cd11c and upregulated that of the M2 marker Cd206 (Figure 4m,n), suggesting that β‐HB can inhibit the cGAS‐STING signaling pathway and reverse M1 polarization. β‐HB also dose‐dependently reduced the mRNA expression levels of pro‐inflammatory cytokines (Figure 4o–r), further validating its anti‐inflammatory effect.

To investigate the potential protective effects of β‐HB on Raw264.7 cells against Dgal‐induced pathological alterations, the cells were categorized into three groups: Control (Con), Dgal‐treated (Dgal, 100 mM), and Dgal plus β‐HB‐treated (DgalBHB, 100 mM Dgal + 5 mM β‐HB). The positive rate of SA‐β‐gal staining (Figure 4s), the protein level of p16 (Figure 4t,u), and the mRNA expression levels of senescence markers (p21, p53) (Figure 4v,w) indicated that β‐HB markedly suppressed Dgal‐induced cellular senescence. The immunofluorescence results indicated that β‐HB greatly inhibited Dgal‐induced cytoplasmic dsDNA levels (Figure 4x). mRNA levels and ratios of the mtDNA marker Nd1 and the nuclear DNA marker Cftr in the cytoplasm (post mitochondrial extraction) were quantified by qPCR (Figure 4y). β‐HB mostly blocked the release of mtDNA into the cytoplasm.

The qPCR (Figure 4z,a2) and Western Blot (Figure 4z,b2–e2) results confirmed that β‐HB significantly inhibited the Dgal‐induced upregulation of the cGAS‐STING signaling pathway at both the mRNA and total protein levels, as well as the level of p‐STING. In addition, β‐HB inhibited the expression of M1 polarization markers while promoting the expression of M2 polarization markers (Figure 4f2–k2). Furthermore, β‐HB markedly suppressed Dgal‐induced mRNA expression of pro‐inflammatory cytokines (Figure 4l2), again confirming its anti‐inflammatory effect.

This study aimed to determine if the protective effect of β‐HB is mediated by the suppression of the STING signaling pathway, utilizing the STING‐specific agonist DMXAA or the inhibitor C‐176 in a Dgal‐induced senescence paradigm with Raw264.7 cells. First, non‐cytotoxic working concentrations of DMXAA (150 μM) and C‐176 (1 μM) for Raw264.7 cells were determined using the CCK‐8 assay and used in subsequent experiments (Figure 5a,b). qPCR results (Figure 5c–j) indicated that DMXAA treatment (Group D vs. Group B) significantly enhanced Dgal‐induced mRNA expression of Cgas and Sting, while also increasing the expression of M1 polarization markers and pro‐inflammatory cytokines. This suggests that STING‐mediated macrophage M1 polarization and inflammatory responses play a role in Dgal‐induced senescence in Raw264.7 cells. Moreover, STING activation by DMXAA diminished the protective effect of β‐HB (Group F vs. Group C), but β‐HB mitigated the adverse effects caused by DMXAA (Group F vs. Group D). The data indicate that β‐HB confers its protective effect through the inhibition of the STING pathway. Additionally, treatment with the STING inhibitor C‐176 (Group E vs. Group C) further suppressed the mRNA expression of the M1 marker Cd11c and the pro‐inflammatory cytokines Tgf‐β and Ccl5, replicating the inhibitory effects of β‐HB on M1 polarization and inflammation.

This section demonstrates that pharmacological modulation of STING pathway activity in Raw264.7 cells reveals that STING pathway activation is a crucial mechanism driving Dgal‐induced macrophage M1 polarization and inflammatory responses. The protective effect of β‐HB on Raw264.7 cells is achieved by inhibiting the STING pathway.

Given that β‐HB supplementation did not directly reverse Dgal‐induced lipid deposition in AML12 hepatocytes, we hypothesized that β‐HB might indirectly regulate hepatocyte lipid deposition by protecting macrophages. To evaluate this idea, we collected CM from Raw264.7 subjected to three treatments: Con (Control), Dgal (100 mM), DgalBHB (100 mM + 5 mM). The CM was then mixed with twice its volume of complete AML12 hepatocyte medium and used to culture AML12 cells (corresponding groups: CM‐Con, CM‐Dgal, and CM‐DgalBHB) (Figure 6a). Results showed that CM from senescent Raw264.7 cells (CM‐Dgal) significantly increased the positive area of SA‐β‐GAL staining, Oil Red O staining, and Bodipy fluorescent staining (a lipid marker) in AML12 cells (Figure 6b–d). CM‐Dgal significantly increased the mRNA expression levels of senescence markers and pro‐inflammatory cytokines in AML12 cells, along with the mRNA levels of genes associated with lipid synthesis (Pparγ, Fas, Acc1, and Scd1) and lipid transport (Cd36, Fatp5) (Figure 6e–g). Nonetheless, the protective influence of β‐HB on senescent Raw264.7 cells (CM‐DgalBHB) markedly mitigated the adverse effects of CM on AML12 cells. In the mechanism exploration, the detection of the protein levels of the PANoptosis marker ZBP1 (Figure 6h,i), the Annexin V‐FITC/PI apoptosis assay assessing apoptosis and late necrosis (Figure 6j), and the qPCR analysis of mRNA levels of PANoptosis markers (Figure 6k) revealed that D‐galactose‐induced senescent Raw264.7 cells enhanced PANoptosis in AML12 cells. Conversely, β‐HB, by protecting macrophages from Dgal‐induced damage, significantly reduced the level of PANoptosis in hepatocytes.

In summary, β‐HB indirectly inhibited hepatocyte lipid deposition and PANoptosis by inhibiting the STING pathway in macrophages, attenuating macrophage M1 polarization and inflammatory responses.

To validate the role of β‐HB in vivo and investigate whether exogenous β‐HB could recapitulate the protective effects of exercise on the liver in aged mice, we administered β‐HB to aged mice via intraperitoneal injection (OIPB, 200 mg/kg BW). As expected, the exogenous β‐HB reached its peak more rapidly than exercise‐induced ketosis due to distinct pharmacokinetics. Nevertheless, both interventions produced a highly congruent blood ketone profile in terms of peak concentration and temporal duration (Figure 7a). Exogenous β‐HB supplementation had no significant effect on mouse body weight but significantly improved body composition, manifested as reduced fat mass and increased lean mass (Figure 7b,c). Moreover, β‐HB supplementation markedly reduced the levels of four serum lipid indicators and the enzymatic activities of liver damage markers (Figure 7d–i). Hepatic TG level detection and histological staining both showed (Figure 7j,k) that β‐HB supplementation significantly alleviated lipid deposition in the livers of aged mice, an effect closely associated with its regulation of the mRNA expression of lipid metabolism‐related proteins (Figure 7l). β‐HB supplementation also downregulated the mRNA expression levels of hepatic pro‐inflammatory cytokines, PANoptosis markers, and senescence markers (Figure 7m–o). Moreover, the attenuation of senescence was substantiated by a marked reduction in the protein level of p16 (Figure 7p,y). In summary, exogenous β‐HB supplementation had substantial protective effects on the livers of aged mice, principally evidenced by diminished hepatic lipid deposition, attenuated inflammatory responses, and reduced levels of PANoptosis. These effects were similar to those observed with exercise intervention.

In mechanistic exploration, detection of mitochondrial dynamics‐related parameters indicated that β‐HB supplementation significantly improved mitochondrial function (Figure 7q–a2). The mitochondrial protective effect diminished the release of cytosolic dsDNA (Figure 7b2), consequently reducing the overactivated cGAS‐STING pathway in the livers of aged mice (Figure 7y,e2–j2). Given that previous in vitro experiments confirmed that STING activation promotes macrophage M1 polarization, this study further evaluated the impact of β‐HB on macrophage polarization in the livers of aged mice. qPCR, Western blotting, and immunofluorescence data consistently indicated (Figure 7y,k2–t2) that exogenous β‐hydroxybutyrate supplementation inhibited the M1 polarization state and facilitated the transition towards an M2 phenotype in hepatic macrophages of aged mice.