PPARγ Deacetylation Suppresses Aging-Associated Atherosclerosis and Hypercholesterolemia

Abstract

Background:

Atherosclerosis is a medical urgency manifesting at the onset of hypercholesterolemia and is associated with aging. Activation of PPARγ counteracts metabolic dysfunction influenced by aging, and its deacetylation displays an atheroprotective property. Despite the marked increase of PPARγ acetylation during aging, it is unknown whether PPARγ acetylation is a pathogenic contributor to aging-associated atherosclerosis.

Methods:

Mice with constitutive deacetylation-mimetic PPARγ mutations on lysine residues K268 and K293 (2KR) in an LDL-receptor knockout (Ldlr^−/−) background (2KR:Ldlr^−/−) were aged for 18 months on a standard laboratory diet to examine the cardiometabolic phenotype, which was confirmed in Western type diet (WTD)-fed 2KR:Ldlr^+/− mice. Whole-liver RNA sequencing and in vitro studies in bone marrow-derived macrophages (BMDMs) were conducted to decipher the mechanism.

Results:

In contrast to severe atherosclerosis in WT:Ldlr^−/− mice, aged 2KR:Ldlr^−/− mice developed little to no plaque, which was underlain by a significantly improved plasma lipid profile, with particular reductions in circulating LDL. The protection from hypercholesterolemia was recapitulated in WTD-fed 2KR:Ldlr^+/− mice. Liver RNA-sequencing analysis revealed suppression of liver inflammation rather than changes in cholesterol metabolism. This anti-inflammatory effect of 2KR was attributed to polarized M2 activation of macrophages. Additionally, the upregulation of core circadian component Bmal1, perceived to be involved in anti-inflammatory immunity, was observed in the liver and BMDMs.

Conclusions:

PPARγ deacetylation in mice prevents the development of aging-associated atherosclerosis and hypercholesterolemia, in association with the anti-inflammatory phenotype of 2KR macrophages.

Graphic Abstract

Background

Atherosclerosis is a predominant contributor to cardiovascular disease (CVD) mortality, presenting with the hallmark of inflamed, cholesterol-rich arterial lesions ¹. Aging is arguably the leading risk factor of atherosclerosis ², with the incidents of first CVD events increasing >20-fold in aged groups (85-95 years) compared to mid-age groups (35-44 years) ³, and an acceleration in atherosclerosis development from the age of 40 ⁴. Various traditional risk factors of atherosclerosis have been attributed to its staggering increase in aging, such as hyperlipidemia, inflammation, hypertension, obesity, diabetes, and smoking ^5,6; however, a better understanding of aging-associated atherosclerosis and identification of novel therapeutic targets are desperately needed to curb this leading cause of death.

Atherosclerosis poses an increased risk of morbidity and mortality in patients with type 2 diabetes (T2D), which worsens with aging ⁷. Several anti-diabetic drugs have been shown to reduce atherosclerosis burden, like PPARγ agonists ⁸. PPARγ belongs to a family of ligand-activated nuclear receptors and is predominately expressed in adipocytes ⁹. Studies have also shown milder PPARγ expression in various tissues, where PPARγ can elicit anti-inflammatory and anti-atherogenic effects, mainly through immune cells like macrophages ^10,11. Agonists of PPARγ, thiazolidinediones (TZDs), are the most potent insulin sensitizers used in the clinic ¹². Interestingly, low-dose TZD treatment has been reported to promote longevity and functional rejuvenations in aged mice, including improved glucose metabolism and insulin sensitivity, reduced inflammation, and prevented tissue atrophy ¹³. TZD treatment also relieves symptoms of anxiety and depression and maintains mitochondrial functionality in various metabolic organs ¹⁴. Patients treated with PPARγ agonist pioglitazone (Pio) presented with decreased mortality than in patients with PPARγ-independent insulin-sensitizing drugs ¹³. In PPARγ-deficient aged mice, the inverse was observed, showing exacerbated subcutaneous fat deposition and metabolic decline ¹⁵. Collectively, a functional decline of PPARγ appears to be actively involved in the development of aging-associated pathologies.

Despite the potent beneficial effects of TZDs, adverse side effects are common, such as weight gain, edema, bone loss, and cardiac hypertrophy ^16,17. Reports have found that PPARγ carries various posttranslational modifications (PTMs), such as acetylation, phosphorylation, SUMOylation, and O-GlcNacylation ^18–21, which specifically regulate PPARγ activity and ultimately the outcome of PPARγ-targeted effects ²². We have previously shown that TZDs can promote PPARγ deacetylation on lysine residues K268 and K293 ²³. Interestingly, PPARγ acetylation is increased during aging, and mice with homozygous K268R/K293R (2KR) substitutions mimicking deacetylation are protected against aging-induced visceral adiposity ¹⁸ and display an anti-atherogenic effect in WTD-fed Ldlr^−/− mice while preventing TZD-induced bone loss and fluid retention ¹⁹. However, the therapeutic implications of manipulating posttranslational modifications (PTMs) of PPARγ have never been studied in the context of aging.

Here we report a strong protective role for PPARγ deacetylation in aging-associated atherosclerosis. Aged Ldlr^−/− mice with PPARγ deacetylation-mimetic mutations (2KR:Ldlr^−/−) fed on a standard laboratory diet developed minimal atherosclerotic plaques, an outcome of lower circulating cholesterol and improved liver health. RNA sequencing of liver tissue and experiments in BMDMs further suggest that PPARγ deacetylation promotes an anti-inflammatory phenotype in macrophages. Hence, targeting PPARγ acetylation may present an opportunity to improve aging-associated conditions like atherosclerosis.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Animal Studies

Mice were bred and housed at 22 ± 1 °C under a 12-hr light /12-hr dark cycle, with access to food and water ad libitum. 2KR:Ldlr^−/− mice (B6.129S7-Ldlrtm1Her/J, JAX, 002207) on a C57BL/6J background (JAX, 000664) were generated as described previously ¹⁹, and the complete replacement of acetylation residues K268 and K293 by acetylation-deficient arginine mutations was confirmed by qPCR (Figure S1A–H). Aged mice were fed on a standard laboratory diet (PicoLab Rodent 5053) and sacrificed at 18 months old. 2KR:Ldlr^+/− mice were fed a western-type diet for 16 weeks containing 42% calories from fat, 15.2% from protein, and 42.7% from carbohydrates, in addition to 0.2% cholesterol (TD.88137; Envigo). Male mice were used in both aging- and diet-induced atherosclerosis studies due to increased susceptibility to atherosclerosis development ²⁴. Plasma profiling assays used include Infinity Triglycerides Liquid Stable Reagent (Thermo Fisher Scientific), HR Series NEFA-HR (Fujifilm Wako), Total Cholesterol E (Fujifilm Wako), and HDL-Cholesterol E (Fujifilm Wako). For glucose tolerance testing (GTT), mice were fasted overnight in cages with fresh bedding and intraperitoneally injected with glucose (2 g/kg BW). Blood glucose was measured with a One Touch Ultra glucometer at 0, 15-, 30-, 60-, 90-, and 120-minutes post-injection. For insulin tolerance testing (ITT), mice were fasted for 4 hr and intraperitoneally injected with insulin (0.7 U/kg BW). Blood glucose was measured at 0, 15-, 30-, 45-, and 60-minutes post-injection. Serum IL-6 was measured using an ELISA (ThermoFisher Scientific, 88-7064-22). All animal protocols used in this study were reviewed and approved by the Columbia University Animal Care and Utilizations Committee.

Fast-Protein Liquid Chromatography

Mouse plasma samples from each group (WT:Ldlr^−/−, 2KR:Ldlr^−/−, WT:Ldlr^+/−, 2KR:Ldlr^+/−) were pooled (n=5), diluted 1:1, and loaded onto a Superdex 200 Increase 10/300GL fast-protein liquid chromatography (FPLC) column (GE Healthcare, 28990944). Fractionated samples were eluted in FPLC buffer (100 mmol/L Tris, 0.4 g/L NaN₃ [pH 7.5]) at a flow speed of 0.3 mL/min. Fresh collections were assayed for lipoprotein profile using Total Cholesterol E (Fujifilm Wako).

Analysis of Mouse Aortic Arch and Root Lesions

Mice were sacrificed with CO₂ euthanasia, proceeded with a left ventricular puncture for blood collection, and followed by perfusion with saline solution. Aortae were cleaned for fat and lymph node removal and dissected for analysis. Aortic arches were photographed, fixed in 10% formalin, stained with Oil Red O, and oriented en face. ImageJ was used to quantify the lesion area. Aortic roots were fixed in 10% formalin, embedded in paraffin, and serial 6 μm sections were obtained on charged frosted slides. Root histology assessments were made with Harris hematoxylin and eosin (H&E) staining, with a total of six slides containing sections 60 μm apart. Plaque area and necrosis were quantified using previously established methods ²⁵. Brightfield images were taken on an Olympus IX71 microscope equipped with a DP73 camera.

Isolation of Bone-Marrow Derived Macrophages

Femur and tibia were collected from 8-week-old mice and immersed in 70% ethanol for 2 minutes and placed in tubes containing serum-free DMEM (1 g/L glucose, Corning, MT10014CM) with small cuts at each end for centrifugation at 8000 RPM for 5 min to collect bone marrow. Pellets were resuspended in a red blood cell lysis buffer and lysed for 10 minutes at room temperature. The reaction was quenched with DMEM and centrifuged for 5 min at 800 g. Pellets were resuspended and aliquoted into 10 cm Petri dishes with DMEM containing 10% FBS, 1% Pen/Strep, and 50 ng/mL macrophage colony stimulating factor (m-CSF, PeproTech). Cells were treated with m-CSF every 48 hours for 6 days to complete monocyte differentiation to macrophages. At day 6, cells were seeded onto 6-well tissue culture-treated plates at equal density for induction. Cells were serum starved overnight and treated with 50 ng/mL lipopolysaccharides (LPS) from Escherichia coli (Sigma, L4524) or 50 ng/mL recombinant interleukin-4 (IL-4, Biolegend, 574304) to elicit M1 or M2 activation, respectively. Cells were collected after 24 hours of induction for immunofluorescence, mRNA analyses, and Western Blotting. Efferocytosis assays were performed as previously described ²⁶.

Liver RNA sequencing and analysis

FastQC (v0.11.8) was used for the quality control of RNA sequencing data ²⁷ and STAR (v2.6.0c) to map the RNA-seq reads data to the mouse genome (GRCm38) ²⁸. Quality control criteria for sequence reads included low quality (average quality score <10) reads < 0.1% and adapter content < 0.1%. Moreover, featureCounts (v1.6.3) was used to calculate the read counts of each gene across the RNA-seq samples ²⁹. Principal component analysis (PCA) was performed by PCAtools (https://github.com/kevinblighe/PCAtools). We used the limma R package (v3.44.3) to conduct differential expression analysis and employed a criterion of log2(Fold Change) > 0.75 and FDR < 0.1 to identify the differentially expressed genes (DEG) between 2KR and WT ³⁰. The functional enrichment analysis of DEG was conducted by DAVID (v6.8) with background adjustment ³¹. Gene Set Enrichment Analysis (GSEA) was performed by GSEA (v4.1.0) ³². Gene expression (log2FPKM) displayed on heatmaps were scaled to a Z-score. The RNA-seq raw data is currently awaiting to be deposited to a publicly accessible database.

Immunohistochemistry

Liver tissue was fixed in 10% formalin overnight, switched to 70% ethanol, and embedded in paraffin. Paraffin sections of 6 microns in thickness were obtained on charged slides. Slides were hydrated through a cascade of Xylenes followed by varying, descending concentrations of ethanol until placed in water. Slides were then placed in chambers containing 10 mM sodium citrate solution and cooked in a pressure cooker for antigen retrieval. After cooling, slides were washed and treated with 3% H₂O₂ for 10 minutes to prevent background staining and washed, and then blocked for 1 hour in blocking solution (1X PBS, 0.1% Tween-20, and 5% normal goat serum), followed by primary incubation with antibodies against F4/80 (Cell Signaling Technology, 70076, RRID: AB_2799771) overnight at 4 °C. The following day, slides were incubated in secondary solution harboring conjugated antibodies with HRP, developed using 3,3´-diaminodbenzidine (DAB, Vector Laboratories, SK-4103), counterstained with hematoxylin, dehydrated, and mounted.

Immunofluorescence

BMDMs were washed in PBS and fixed in 4% paraformaldehyde (PFA) for 20 minutes at room temperature, followed by membrane permeabilization in 0.2% Triton-X in PBS. Cells were blocked in 5% goat serum and incubated with anti-CD206 antibody (abcam, ab64693, RRID: AB_1523910) overnight at 4 °C. 1:400 dilution of fluorescent secondary antibody AF488 (Thermo Fisher Scientific, A32790, RRID: AB_2762833) was used, and subsequently 1:1000 of 4′,6-diamidino-2-phenylindole (DAPI) for nuclear staining. Images were taken at 20X on a Zeiss confocal microscope with an LSM 710 scanning module.

Quantitative Real-Time PCR

RNA was isolated from tissues or cells by using a Tri-Isolate RNA Pure Kit (IBI Scientific, IB47632). An Applied Biosystems High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific) was used to synthesize cDNA from 1 μg total RNA. Quantitative PCR was performed on a Bio-Rad CFX96 Touch Real-Time PCR Detection System using AzuraView GreenFast qPCR Blue Mix (Azura). Relative gene expression was calculated by using the ΔΔCt method, with Cyclophilin A (CpA) as the reference gene in the liver and Hprt in BMDMs.

Western Blotting

Cells were lysed and tissues were homogenized in a Polytron homogenizer immediately after dissection in Western extraction buffer (150 mmol/L NaCl, 10% glycerol, 1% NP-40, 1 mmol/L EDTA, 20 mmol/L NaF, 30 mmol/L sodium pyrophosphate, 0.5% sodium deoxycholate, 0.05% SDS, 25 mmol/L Tris-HCl [pH 7.4]) containing a protease inhibitor cocktail (Roche). The lysate was sonicated and incubated on ice for 30 min before centrifugation. SDS-PAGE and Western blotting were performed and detected with enhanced chemiluminescence (ThermoFisher Scientific, 32106). Antibodies used were anti-BMAL1 (Proteintech, 14268-1-AP, RRID: AB_2878037), anti-HSP90 (Proteintech, 13171-1-AP, RRID: AB_2120924), HRP-conjugated anti-Beta-Actin (Proteintech, HRP-60008, RRID: AB_2819183), and anti-CD206 (abcam, ab64693, RRID: AB_1523910).

Statistical Analysis

Values are presented as the mean ± SEM. Data were assessed for normality using the Shapiro-Wilk tests and D’Agostino & Pearson tests. The unpaired student t-test with the assumption of equal variances was used for statistical analyses of data from two groups. Data that violated equal variance assumption were handled with Welch’s correction. Data that did not pass normality were analyzed using the Mann-Whitney U test. Sample size estimation was based on previous results in comparable studies, and P<0.05 was considered statistically significant. Specific details on how many independent biological samples or mice were included in an experiment are presented in the corresponding figure legends. All data were analyzed in GraphPad Prism software version 9.1.0 (GraphPad Software, San Diego, CA).

Results

PPARγ deacetylation in aged mice significantly dampens plaque development

In humans, plasma total cholesterol over 240 mg/dL is considered hypercholesterolemia. However, in the classic atherogenic model of WTD-fed Ldlr^−/− mice, plasma cholesterol levels often exceed 1000 mg/dL, rendering them with severe atherosclerosis at a young age, departing from the development associated with aging that is observed in humans. To overcome these discrepancies, we kept WT Ldlr^−/− mice on standard laboratory diet feeding until 18-months-old, equivalent to approximately 56 of age in human years. These mice developed advanced lesions at the aortic arch and root (Figures 1A–G) with an average of total cholesterol at ~200 mg/dL (Figure 1H). Hence, in the context of age and hypercholesterolemia, aged standard diet-fed WT Ldlr^−/− mice better mimic the pathogenesis of atherosclerosis in humans.

Figure 1: PPARγ deacetylation suppresses atherosclerosis and hypercholesterolemia in aging.

Aged mice were fed on a standard laboratory diet sacrificed at 18 months. A. Images of plaque-enriched ascending aortic arches (top 3 panels: WT:Ldlr^−/−, bottom 3 panels: 2KR:Ldlr^−/−). B. Oil Red O staining of aortic arches. C. Quantification of aortic arch lesion area. D. Images of H&E-stained aortic roots taken at 4X and 20X, dashed line indicates intimal lesion area. E. Quantification of total aortic root plaque area, F. and percent necrotic tissue in the intima. G. Body weight measurements of mice after 18 months on a standard laboratory diet. H. Total circulating cholesterol in plasma of fasting and refed mice. I. FPLC measurement of lipoprotein distribution. J. Plasma non-HDL cholesterol (n=12 mice per group). K. Plasma triglycerides. L. Plasma non-esterified fatty acids (NEFA). Data represent mean ± SEM, with n = 15 mice per group unless stated otherwise. All datapoints were assessed for normal distribution and equal variances. Two-tailed Student’s t-tests were used for statistical analyses between groups WT:Ldlr^−/− and 2KR:Ldlr^−/−. The following data did not pass normality and were analyzed using the Mann-Whitney U test: NEFA fasting (L).

PPARγ deacetylation confers a mild atheroprotective effect at the aortic arch but not at the root region in an Ldlr^−/− background on WTD feeding, without influencing cholesterol levels maximized at >2500 mg/dL ¹⁹. In contrast, aged 2KR:Ldlr^−/− mice on standard laboratory diet feeding showed a striking 80% abrogation of aortic arch plaque deposition compared to WT:Ldlr^−/− control mice (Figures 1A–C). Additionally, a remarkable reduction of the aortic root plaque area was observed, with over a 4-fold decrease in quantified leaflets (Figures 1D–E). Of the deposited plaque in aged 2KR:Ldlr^−/− mice, the necrotic core area was reduced by 40% (Figure 1F), indicating an improvement in plaque microenvironment. Overall, PPARγ deacetylation in aging reveals a potent protection against atherosclerosis.

Suppression of hypercholesterolemia underlies the protection against atherosclerosis development in aged 2KR:Ldlr^−/− mice

The anti-atherogenic outcome observed in aged 2KR:Ldlr^−/− mice was coupled with a significant decrease in total body weight (Figure 1G). Next, we examined the lipid profile to see any correlation with the reduction in lesion deposition. There were appreciable decreases in total circulating cholesterol, both in the fasting and refed state (Figure 1H). FPLC fractionation of lipoproteins in aged 2KR:Ldlr^−/− mice showed prominent decreases in LDL fractions compared to control mice, with milder changes in VLDL and HDL (Figure 1I). Colorimetric assessments of HDL-cholesterol further confirmed a decrease in non-HDL cholesterol in these mice (Figure 1J). Besides cholesterol, other lipids such as triglycerides (TG) and non-esterified fatty acids (NEFA) in circulation are also associated with atherosclerosis susceptibility and severity ^33,34. Interestingly, the decreases in plasma TG and NEFA levels were more modest (Figures 1K–L), indicating a cholesterol-specific phenotype upon PPARγ deacetylation. Moreover, aged 2KR:Ldlr^−/− mice had non-significant changes in fasting plasma insulin levels (Figure S2A), a diagnostic parameter of insulin resistance, suggesting comparable insulin sensitivity in both groups. Taken together, the extensive reduction in circulating LDL appropriately underlies the inhibition of plaque development in aged 2KR:Ldlr^−/− mice.

The cholesterol-lowering effect of PPARγ deacetylation is independent of diet and LDLR presence

The complete ablation of Ldlr has a strong impact on altering lipoprotein uptake and metabolism, which may blunt some changes caused by PPARγ deacetylation. To overcome this possible artifact, we crossed 2KR mice with Ldlr^+/− mice (2KR:Ldlr^+/−) conferring a heterozygous deletion of the Ldl receptor to maintain relatively normal lipoprotein uptake. Upon WTD feeding for 16 weeks, a period sufficient to induce severe atherosclerosis in homozygous Ldlr^−/− mice, the wildtype heterozygous Ldlr^+/− mice (WT:Ldlr^+/−) failed to develop plaque (Figure 2A) while displaying very minute hypercholesterolemia at approximately the same level of 200 mg/dL observed in aged standard diet-fed Ldlr^−/− mice (Figure 2B). Even without developing any plaque, 2KR:Ldlr^+/− mice persisted with markedly reducing circulating cholesterol levels in both the fasting and refed states, particularly in LDL and milder in VLDL fractions (Figures 2B–C). Indeed, this reduction was exclusively tied to non-HDL cholesterol (Figures 2C–D). The improvement in lipoprotein profile was independent of body weight (Figure 2E), nor of plasma TG and NEFA levels (Figures 2F–G). Moreover, glucose tolerance was minimally improved, with no changes in insulin sensitivity (Figure S2B–C). Overall, the WTD-fed LDLR haploinsufficiency model recapitulates the LDL cholesterol-lowering outcome through PPARγ deacetylation that we see in aged mice, further dissociating it from LDLR presence and metabolic parameters.

Figure 2: Western diet-fed 2KR:Ldlr^+/− mice exhibit a similar decrease in circulating cholesterol.

Male 2KR:Ldlr^+/− and WT:Ldlr^+/− controls were fed on WTD for 16 weeks. A. Images of ascending aortic arch. B. Body weight. C. Total circulating cholesterol in plasma of fasting and refed mice. D. FPLC measurement of lipoprotein distribution. E. Plasma non-HDL cholesterol (n=11 mice per group). F. Plasma triglycerides. G. Plasma non-esterified fatty acids (NEFA). Data represent mean ± SEM, with n = 18 (WT:Ldlr^+/−) and 20 (2KR:Ldlr^+/−) mice per group unless stated otherwise. All datapoints were assessed for normal distribution and equal variances. Two-tailed Student’s t-tests were used for statistical analyses between groups WT:Ldlr^+/− and 2KR:Ldlr^+/−. The following data did not pass normality and were analyzed using the Mann-Whitney U test: Total fasting Cholesterol (B), refed TG (F), and refed NEFA (G).

RNA-sequencing analysis highlights reduced inflammation in aged 2KR:Ldlr^−/− mouse livers

Cholesterol biosynthesis occurs in the liver, as is the transport of LDL from the liver to the circulation for its uptake and deposition in arteries ³⁵. Furthermore, the metabolic, lipogenic, and inflammatory microenvironments of the liver can influence atherosclerosis outcome ^36,37. To understand the persistent suppression of hypercholesterolemia by PPARγ deacetylation, we performed an RNA-seq analysis of the livers of aged 2KR:Ldlr^−/− mice and their controls. Principal component analysis (PCA) of sequenced reads uniquely mapped 2KR:Ldlr^−/− mice in clusters separate from their WT controls (Figure 3A). There were 98 upregulated and 341 downregulated genes in the aged 2KR mouse livers compared to WT controls (Figure 3B). Functional enrichment by DAVID revealed only one biological process (BP) – the acute-phase response, to be upregulated (Figure 3C), whereas the downregulated genes were strongly associated with immune BPs (Figure 3D). Chronic states of inflammation, like in atherosclerosis, can lead to a steady low-level induction of an acute phase response predominantly in the liver. This response involves a set of proteins that can contribute to instances of tissue damage, inflammation, and stress ^38–40. Gene set enrichment analysis (GSEA) further demonstrated that acute-phase response, Golgi organization, protein folding, and fatty acid beta oxidation were among the top upregulated pathways. On the other hand, the top downregulated gene sets in 2KR mice were mainly involved in the inflammatory response, with genes exhibiting pro-inflammatory properties (Figure 3E). Heatmaps were generated to highlight significantly changed genes of high relevance (Figure 3F, Figure S3). The upregulated ones by PPARγ deacetylation were grouped into three categories: lipoprotein metabolism, fatty acid oxidation, and anti-inflammatory immunity. Of the genes downregulated, the largest sets included genes involved in pro-inflammatory signaling, as well as lipid and lipoprotein metabolism. These findings imply PPARγ deacetylation as a repressor of liver inflammation in association with its anti-atherogenic effect observed in aged mice.

Figure 3: Liver RNA sequencing reveals the repression of inflammation by PPARγ deacetylation.

A. Principal component analysis (PCA) of the RNA sequencing of aged liver samples. PC: Principal Component. B. The number of the upregulated and downregulated genes between WT:Ldlr^−/− and 2KR:Ldlr^−/− mice. Functional enrichment analysis of upregulated C. and downregulated D. genes. The WT1-4 and 2KR1-4 columns are the expression profile of the genes related to the enrichment BPs (Biological Processes). The −log10 (DEG_FDR) column indicates the significant level of DEG (differentially expressed gene) between aged 2KR and WT mouse livers. The other columns represent the association between gene and the enrichment BP. The red color (TRUE) indicates the gene is associated with the corresponding BP, while the grey color (FALSE) represents the gene is not. E. Top 15 positive and negative enrichment BPs from Gene Set Enrichment Analysis (GSEA). The red color indicates the positive (upregulated) BPs, while the blue color depicts the negative (downregulated). The points represent the gene associated with the corresponding BP. F. Heatmap of upregulated and downregulated genes in aged 2KR livers corresponding to their respective functions, Lipoprotein metabolism, pro- and anti-inflammatory responses, and fatty acid oxidation. Gene expression (log2FPKM) scaled to Z-score. Data represent n = 4 mice per group.

Improved liver health by PPARγ deacetylation

We then assessed metabolic function in the liver to understand the decrease in atherosclerosis burden with 2KR mutations. As fatty livers can contribute to CVD risk ⁴¹, aged 2KR:Ldlr^−/− mice displayed less hepatic steatosis by histological analysis (Figure 4A) without significantly reducing liver mass (Figure 4B). Regardless of the robust changes in circulating LDL, the genes involved in cholesterol synthesis and export pathways remained largely unchanged (Figure S4), validating our RNA-seq analysis. Given that an anti-inflammatory response can promote cholesterol efflux and preclude LDL uptake ⁴², we hypothesized that the changes in cholesterol could be a result of a chronic immune response influencing a protective environment in the liver. First, a decrease in F4/80 immunostaining, a marker for pan-macrophage activation, was observed in the liver (Figure 4C), as well as a modest reduction in serum IL-6 (Figure 4D). Interestingly, the expression of pro-inflammatory genes in aged 2KR:Ldlr^−/− mouse livers were mostly unaltered except for a trend towards a decrease in liver Tnfa (Figure 4E). Instead, anti-inflammatory markers found in M2-polarized macrophages were significantly upregulated, including Stat6, Mrc1 (encoding CD206), Arg1, Fizz1, and Il10 (Figure 4E). Hence, the low cholesterol phenotype in aged 2KR:Ldlr^−/− mice is independent of cholesterol biosynthesis but is associated with decreases in liver steatosis and inflammation.

Figure 4: PPARγ deacetylation decreases liver steatosis and inflammation in aged- and WTD-fed mice.

Aged WT:Ldlr^−/− and 2KR:Ldlr^−/− mice fed on a standard laboratory diet were sacrificed at 18 months. A. Representative images of liver histology at 40X magnification. B. Liver mass (n=15). C. Immunohistochemical staining of F4/80 in the liver imaged at 40X, and D. measurement of plasma IL-6 levels using an ELISA (n=5). E. Gene expression profile in the liver using quantitative real-time PCR (Q-PCR) of pro- and anti-inflammatory genes (n=15). Heterozygous WT:Ldlr^+/− and 2KR:Ldlr^+/− mice were sacrificed after 16 weeks of WTD feeding. F. Representative images of liver histology at 40X magnification. G. Liver mass in grams (n=18, 20). H. Gene expression profile in the liver using Q-PCR of pro- (n=12) and anti-inflammatory genes (n=6), I. Immunohistochemical staining of F4/80 in the liver imaged at 40X, and J. measurement of plasma IL-6 levels using an ELISA (n=16). K. Gene expression profile in the liver using Q-PCR of fibrotic gene markers (n=12). Data represent mean ± SEM. All datapoints were assessed for normal distribution and equal variances. Two-tailed Student’s t-tests were used for statistical analyses between groups WT:Ldlr^−/− and 2KR:Ldlr^−/−, or WT:Ldlr^+/− and 2KR:Ldlr^+/−. The following data did not pass normality and were analyzed using the Mann-Whitney U test: Liver mass (B, G), Mcp1 (E, H), Il6 (E), Il10 (E), Mrc1 (H), Timp1 (K), Col1a1 (K).

The improvements in liver health were recapitulated in WTD-fed 2KR:Ldlr^+/− mice, which show a similar low cholesterol profile to our aging model. By challenging the outcomes of WTD feeding, liver steatosis was remarkably prevented in 2KR:Ldlr^+/− mice, resulting in lighter livers (Figure 4F–G). Again, these mice exhibited no changes in gene expression of markers involved in cholesterol biosynthesis and transport (not shown). In profiling other advanced changes in hepatic pathology associated with WTD feeding, 2KR:Ldlr^+/− mice presented with significant decreases in pro-inflammatory markers Tnfa and Mcp1, whereas the differences in anti-inflammatory gene expression were negligible (Figure 4H). This was further confirmed with a decrease in F4/80 staining in the liver, and a significant reduction in total serum IL-6 (Figures 4I–J). Furthermore, there appeared to be protection from liver fibrosis in WTD-fed 2KR:Ldlr^+/− mice, as indicated by their strong repression of fibrotic genes Timp1, Spp1, Col1a1, and Col3a1 (Figure 4K). Overall, PPARγ deacetylation contributes to a positive remodeling of the liver microenvironment, which can ultimately lead to improvements in aging- and diet-induced atherosclerosis.

PPARγ deacetylation induces M2-like behavior in macrophages

2KR mutations in standard diet-fed aged mouse livers presented with an upregulated anti-inflammatory phenotype, and this protective effect was switched to a suppression of the pro-inflammatory response in WTD feeding, which presumably induces advanced inflammation and steatosis. Considering that macrophages are most impacted in this phenomenon, we hypothesized that PPARγ deacetylation executes an anti-inflammatory function in macrophages. To test this, we isolated bone marrow-derived macrophages (BMDMs) from WT and 2KR mice to treat with recombinant IL-4 for the induction of macrophage polarization. Immunofluorescence of the M2 marker CD206 in BMDMs showed similar and robust expression in both groups of IL-4-treated cells; but, surprisingly, non-polarized “M0-like” 2KR BMDMs presented with abundant CD206-positive cells, suggesting a basal M2 phenotype (Figure 5A). Western blots further validated the increase in CD206 protein expression in 2KR BMDMs, displaying M2-like behavior in non-treated cells compared to WT (Figure 5B). Gene expression of anti-inflammatory marker Arg1 was upregulated in non-treated 2KR BMDMs, with other relevant genes showing modest but insignificant changes (Figure 5C). Interestingly, Arg1 has been shown to be a direct target gene of PPARγ ⁴³. Conversely, no changes were observed in pro-inflammatory gene expression at the basal state (Figure 5C) or when stimulated with LPS for M1 activation (not shown). However, the upregulation of M2 markers in 2KR macrophages was saturated when treated with IL-4 (Figure 5D), probably owing to the prominent effects of IL-4 on inducing M2-like polarization ⁴⁴. Collectively, the in vitro data suggest that deacetylation of PPARγ primes macrophages to display M2-like properties, possibly explaining the suppression of aging-associated atherosclerosis.

Figure 5: 2KR macrophages display anti-inflammatory M2-like behaviors.

A. Representative immunofluorescent images of BMDMs stained for CD206, and DAPI for nuclei. Images were taken at 20X magnification. WT and 2KR BMDMs were treated with IL-4 for M2 polarization or left untreated for M0 basal activity. B. Western blot of CD206 protein expression in M0, M1 (LPS-treated), and M2-activated BMDMs; β-actin was used as a loading control. C. Gene expression profile of pro- and anti-inflammatory markers in basal M0-like BMDMs using Q-PCR, D. and of anti-inflammatory genes in M2-polarized BMDMs. Data represent mean ± SEM, with n of 3 (WT) and 4 (2KR) biological replicates unless stated otherwise. All datapoints were assessed for normal distribution and equal variances. Two-tailed Student’s t-tests were used for statistical analyses between groups WT and 2KR. The following data did not pass normality and were analyzed using the Mann-Whitney U test: IL-10 (C, D).

PPARγ deacetylation upregulates Bmal1 in macrophages

In understanding the anti-inflammatory phenotype in 2KR mouse livers, we found that Bmal1, a master regulator of circadian rhythm, stood out as a top upregulated target in the RNA-seq analysis (Figure 3F). Studies have gauged the connection between Bmal1 and anti-inflammatory signaling ^45,46, which led us to propose that Bmal1 could be a mediator of M2-like priming in 2KR macrophages. The significant increase of Bmal1 in aged 2KR:Ldlr^−/− mouse livers was confirmed by Q-PCR and western blotting analyses (Figures 6A–B). Its expression was consistently upregulated in WTD-fed heterozygous 2KR:Ldlr^+/− cohorts (Figure 6C), confirming the independence of diet and LDLR presence for this phenotype. Additional insights elucidated alterations in liver circadian genes by PPARγ deacetylation (Figure 6D), with significant increases in Rorα, encoding a receptor that activates Bmal1, and Cry1, which is known to be transcriptionally activated by Bmal1 ^47,48. To determine whether the regulation of Bmal1 by PPARγ deacetylation occurs in macrophages, we measured Bmal1 expression in BMDMs (Figure 6E). Bmal1 was induced upon polarizing WT BMDMs by IL-4 treatment, suggesting pro-M2 properties. Interestingly, 2KR macrophages conveyed higher Bmal1 expression at the basal M0 state, while it was not further induced with M2 polarization like in WT cells (Figure 6E), likely due to the blunted effects IL-4 has on activating M2 genes in macrophages (Figures 5A and 5D). Overall, these findings suggest that the anti-inflammatory phenotype in 2KR:Ldlr^−/− mice could be underlined by the upregulation of Bmal1.

Figure 6: Circadian rhythm regulator Bmal1 is upregulated upon PPARγ deacetylation.

A. Relative mRNA expression of Bmal1 in aged standard laboratory diet-fed mouse livers using Q-PCR (n=15). B. Western blot to detect Bmal1 protein in the liver, with quantification of bands as arbitrary units (a.u.) C. mRNA expression of Bmal1 in in WT:Ldlr^+/− and 2KR:Ldlr^+/− western diet-fed mouse livers (n=6). D. Gene expression profile of circadian rhythm-related genes in the liver using Q-PCR (n=11-13). E. mRNA expression of Bmal1 in WT and 2KR-isolated BMDMs (n=3, 4). Data represent mean ± SEM, with n = 8-15 mice per group, and n of 3 (WT) and 4 (2KR) biological replicates for BMDM experiments. All datapoints were assessed for normal distribution and equal variances. Two-tailed Student’s t-tests were used for statistical analyses between groups WT:Ldlr^−/− and 2KR:Ldlr^−/−, or WT:Ldlr^+/− and 2KR:Ldlr^+/−. The following data did not pass normality and were analyzed using the Mann-Whitney U test: Clock (D), Cry1 (D).

Discussion

Here we describe the protective implications of PPARγ deacetylation in aging-associated atherosclerosis. Aged 2KR:Ldlr^−/− mice develop little to no atherosclerotic lesions underlain by reduced hypercholesterolemia, which is independent of diet, age, and LDLR presence, being reproduced in WTD-fed 2KR:Ldlr^+/− mice. Furthermore, our study unraveled the ability of 2KR macrophages to present with anti-inflammatory behaviors, possibly explaining the profound improvement in atherosclerosis outcome. Transcriptomic analysis of liver and in vitro studies in BMDMs identified Bmal1 as a potential mediator of PPARγ deacetylation to convey the anti-inflammation phenotype and thus, protect against aging-associated atherosclerosis.

Studies in human and animal models of aging have proposed an anti-aging role of PPARγ. Genome-wide association studies show genetic variations of PPARγ to be implicated with longevity ⁵¹. In the brain, PPARγ can be found in regions such as the hippocampus, where it regulates stress and aging in a protective manner ¹⁴. Conversely, low PPARγ expression has been documented in senescence-accelerated prone mice (SAMP1) ⁵². Moreover, PPARγ agonists like TZDs can delay or even reverse the aging process and its accompanying complications. In one study, long-term treatment of aged mice with low doses of rosiglitazone (Rosi) improved glucose metabolism, reduced inflammation, and decreased tissue atrophy ¹³. PPARγ agonists can also alleviate responses to psychological stress, preventing the onset of cognitive and mood disorders ¹⁴. In humans, pioglitazone (Pio) treatment improved survival and promoted longevity in comparison to patients treated with non-TZD insulin sensitizers ¹³. We have detected a marked increase of PPARγ acetylation during aging, and here we demonstrated that prevention of PPARγ acetylation by the 2KR mutant improved aging-associated atherosclerosis. Therefore, PPARγ acetylation is a potential pathogenic factor of aging complications and could serve as a target for anti-aging therapies. In line with this prospect, PPARγ acetylation can promote one hallmark of aging - cellular senescence, which can be reversed by its deacetylase SirT1 ⁵³.

PPARγ is predominantly expressed in adipocytes, but is also well-documented to be expressed in monocytes and macrophages to influence differentiation, classical activation, and alternative polarization ^10,54,55. Ultimately, these PPARγ-mediated processes are involved in lipid metabolism ⁵⁶ and cytokine production in these cells ⁵⁷. More specifically, high PPARγ in immune cells correlates with a dampened release of pro-inflammatory cytokines IL-6 and IL-1β, and in PPARγ-deficient macrophages, the opposite is observed ⁵⁸. PPARγ agonists like TZDs exhibit an anti-inflammatory function in macrophages ^59,60. In atherosclerosis, PPARγ can influence cholesterol uptake in macrophages and thus limit foam cell formation ⁶¹, and protect mice against vascular dysfunction ^62,63. The polarization of macrophages into an M2-like anti-inflammatory state has been shown to be regulated by PPARγ, driving several components of the anti-inflammatory response ⁶⁴. Examples include the transcription factor STAT6 ⁶⁵, Arginase-1, encoded by Arg1), which is a direct target gene of PPARγ ⁴³, and the mannose receptor CD206. Additional experiments described native monocytes to be primed to an anti-inflammatory state by PPARγ ⁶⁶, complementing our findings in untreated basal macrophages that present with an anti-inflammatory phenotype. Moreover, macrophage-specific deletions of PPARγ can delay the maturation of macrophages and impair their activation to an alternative state ⁶⁷. In our study, we showed transcriptional changes in Arg1 and increased protein expression of CD206 upon PPARγ deacetylation. Hence, the polarized activation of macrophages involves not only the expression of PPARγ, but its deacetylation too. It is thus a plausible suggestion that the increase of PPARγ acetylation in aging promotes inflammatory activation of macrophages to favor atherosclerosis, which is indeed a chronic inflammatory disease.

Bridging things together, circadian rhythm plays important roles in cardiovascular function, T2D, obesity, and is impacted with aging ⁶⁸. Bmal1 is a core component of the biological clock, and the influences of Bmal1 on different cell types and in various pathologies are noteworthy. Bmal1 ablation in mice present with increased inflammation ⁶⁹, vascular endothelial dysfunction, and atherogenesis ⁷⁰ and affect lipoproteins ⁷¹. In the liver, hepatocyte-specific Bmal1 knockout mice confer an alteration in metabolic and lipid profile, promoting hyperlipidemia and enhancing atherosclerosis ^72,73, whereas an increase in Bmal1 attenuated liver steatosis ⁷⁴, and promoted the upregulation of anti-inflammatory markers Arg-1 and IL-10 ⁷⁵. Gain-of-function studies in mouse endothelial cells and observations in human carotid plaques show that Bmal1 can inhibit atherosclerosis and promote plaque stability by suppressing oxidized-LDL (oxLDL) and ROS accumulation ⁷⁶, and in one study, treatment with recombinant Bmal1 decreased total cholesterol in mice ⁷⁷. In humans, Bmal1 is higher in stable plaques than in plaques more vulnerable to rupture ^78,79, and in our findings, plaque necrosis is significantly reduced. Necrotic tissue contributes to plaque instability in humans, and although plaque rupture does not occur in mice ⁸⁰, it is a notable observation that holds relevance to human atherosclerosis. Alterations in circadian rhythm were also demonstrated in human plaque-derived vascular smooth muscle cells (VSMCs), where VSMC-specific disruptions in Bmal1 expression contributed to worsened carotid atherosclerotic lesions ^79,81, and increased monocyte transmigration ⁷⁹. Though the effects of Bmal1 deletion in myeloid cells on atherosclerosis outcome are controversial ^82,83, it has been well established as a metabolic sensor and regulator of the inflammatory response in macrophages ^84–87. Bmal1 is a direct downstream target of PPARγ ⁵⁰, and TZDs stimulate Bmal1 and thus impact circadian activity ⁸⁸. Thereby, identifying a possible link between PPARγ deacetylation and Bmal1 could provide an explanation of the alleviated inflammation and atherosclerosis outcomes in aged 2KR mice.

Taken together, we show a reversal of the pathogenic manifestations in aging-associated atherosclerosis upon PPARγ deacetylation through rectifying hypercholesterolemia and promoting an anti-inflammatory response. The present study is limited in its mechanistic approaches. Besides improving hypercholesterolemia and inflammation, there are other mechanisms that may account for the anti-atherosclerosis function of PPARγ deacetylation, such as endothelial protection and energy catabolism ¹⁹. Nevertheless, to our knowledge, this is the first study regarding PPARγ PTMs in macrophages in aging-associated atherosclerosis. Therapeutic applications for targeting PPARγ in aging- and also diabetes-induced atherosclerosis are limited due to the downsides accompanied with drugs like TZDs ⁸⁹. However, given the uncoupling of TZD’s metabolic benefits from adverse side effects by PPARγ deacetylation, our findings provide a framework for considering new approaches of targeting PPARγ acetylation to lower the risk of cardiometabolic diseases in aging with improved safety.

Highlights.

• PPARγ deacetylation strongly suppresses aging-associated atherosclerosis.

• PPARγ deacetylation dampens circulating LDL.

• PPARγ deacetylation exhibits an anti-inflammatory function and favors M2 polarization of macrophages.

• Bmal1 is upregulated in the liver and in macrophages upon PPARγ deacetylation.

Non-standard Abbreviations and Acronyms

2KR

PPARγ deacetylation on lysine residues K268 and K293

ARG1

arginase 1

BMAL1

brain and muscle ARNT-like 1

BMDM

bone marrow-derived macrophages

CD

cluster of differentiation

CRY1

cryptochrome circadian regulator 1

FIZZ1

found in inflammatory zone 1

HDL

high-density lipoprotein

IFN

interferon

IL

interleukin

LDL

low-density lipoprotein

LDLR

LDL receptor

MRC1

mannose receptor c-type 1 (CD206)

Pio

pioglitazone

PPARγ

peroxisome proliferator-activated receptor γ

PTM

post-translational modification

RORα

RAR-related orphan receptor alpha (RORα)

Rosi

rosiglitazone

SIRT1

sirtuin 1

STAT6

signal transducer and activator of transcription 6

TNF-α

tumor necrosis factor-α

Trog

troglitazone

TZDs

thiazolidines

VLDL

very low-density lipoprotein