PAK3 Exacerbates Cardiac Lipotoxicity via SREBP1c in Obesity Cardiomyopathy Free

The global prevalence of obesity has almost tripled in the past 40 years and, independent of other risk factors, has contributed to the elevated incidence of cardiovascular diseases, especially cardiac dysfunction (1). It has been reported that the pathogenesis of obesity cardiomyopathy is complex and multifaceted (2–4); however, the underlying molecular mechanisms remain unclear.

Fatty acids (FAs) provide an essential energy source for maintaining cardiac contractility. However, the presence of excess FAs in cardiomyocytes causes the accumulation of cytotoxic lipid by-products, including triglycerides (TAGs), diglycerides (DAGs), and ceramides, leading to cardiac lipotoxicity (5,6). In obese individuals, an imbalanced cardiac lipid uptake and FA utilization results in toxic lipid accumulation, albeit higher FA oxidation (FAO), accompanied by oxidative stress, energy deprivation, and apoptosis, ultimately leading to cardiac remodeling and dysfunction (7). A better understanding of the intricate interplay between obesity and consequent cardiac dysfunction at a molecular level is fundamental for developing treatment strategies.

The sterol regulatory element binding protein 1 (SREBP1) serves as a transcription factor in lipid metabolism, orchestrating the transcription of genes that participate in FA and TAG synthesis (8). Initially localized in the endoplasmic reticulum, SREBP1 is transported to the Golgi via SREBP1 cleavage-activating protein (SCAP)-mediated trafficking. In the Golgi, the N terminus of SREBP1 is released and translocated into the nucleus (9) where it targets the promoters of genes responsible for lipogenesis, upregulating intracellular lipid content (10–12). Elevated levels of SREBP1 have been correlated with the progression of cardiac dysfunction (13). Specifically, isoform SREBP1c is increased in failing heart biopsies of patients with metabolic syndrome, which also showed intracellular accumulation of lipids in cardiomyocytes (14). However, the mechanism by which SREBP1 is regulated in the heart is unknown.

P21-activated kinases (PAKs) constitute a family of nonreceptor serine/threonine kinases playing pivotal roles in cell survival, proliferation, and the dynamic processes of cytoskeletal remodeling (15). PAK3 has been implicated in hepatic cell proliferation and migration, which are both central to hepatocellular carcinoma development (16). Our previous research has demonstrated that PAK3 activates mammalian target of rapamycin complex 1 (mTORC1) in the myocardium, thereby restraining cardiac autophagy in response to starvation and neurohormonal stress (17).

The current study explores the role of PAK3 in the development of obesity cardiomyopathy. We observed an increase of PAK3 levels in response to prolonged obesity-induced stress. Cardiac PAK3 overexpression provoked cardiac dysfunction under high-fat diet (HFD)-induced stress, coupled with overload of toxic lipids and oxidative stress. Moreover, SREBP1c is part of the molecular mechanism through which PAK3 exacerbates lipotoxicity in cardiomyocytes.

Detailed methods are available in the Supplementary Materials.

This study was approved by the University Research Ethics Committee, University of Manchester, Manchester, U.K. The animals from which samples were obtained were euthanized in accordance with directive 2010/63/EU of the European Parliament and in compliance with the ethical standards of the University of Manchester and the UK Home Office Animals (Scientific Procedures) Act 1986.

Cardiac-specific PAK3 overexpression was achieved using an adeno-associated virus 9 (AAV9) carrying human Pak3 cDNA driven by the human cardiac troponin T (cTnT) promoter (17). Six-week-old C57BL/6 mice, injected with AAV9-cTnT-Gfp or AAV9-cTnT-Pak3, were fed with either Chow or HFD for 16 weeks, followed by treatment as experimental design.

Mice were killed by a cervical dislocation under terminal anesthesia following echocardiography. Whole hearts were excised immediately from the mice before permanent cessation of circulation. The apex part was snap frozen for molecular assessments, while the other half was embedded for histological analyses.

Fresh human hearts were provided from consented donors through the Maryland Legacy Foundation (the U.S. transplantation network, Novabiosis). All procedures were in agreement with the Declaration of Helsinki and approved by the institutional review boards of the University of Louisville (Louisville, KY) and Baylor College of Medicine (Houston, TX) (18).

In addition, myocardial protein extracts and heart sections from consented healthy donors without cardiovascular diseases (female and male, aged 50–70 years, BMI < 24) and donors with obesity (female and male, aged 50–70 years, BMI 30–35) and cardiovascular complications were purchased from Asterand (BioIVT, Burgess Hill, U.K.). Asterand obtained ethical approval and consent following the United Kingdom Human Tissue Authority regulations.

All experimental procedures conducted on human samples were approved by the University of Manchester Research Ethics Committees and in agreement with the Declaration of Helsinki.

The human induced pluripotent stem cell (hiPSC) cell line SEUR7 was established from human dermal fibroblasts by a CytoTune iPS Programming kit from Wellcome Trust Sanger Institute and made available via Public Health England through European Collection of Cell Cultures with a material transfer agreement. Cardiomyocyte differentiation was induced by 4 μmol/L CHIR99021 and 5 μmol/L IWP2. hiPSC–derived cardiomyocytes (hiPSC-CMs) were later cultured in RPMI1640 HEPES Glutamax medium with B27 supplement for ≥21 days prior to further treatment.

Palmitic acids (PAs) and oleic acids (OAs) (500 µM) were conjugated to 0.5% FA-free BSA (7:1 molar ratio of FAs:BSA) (17). The duration of the stimulation on cells and human heart slices is specified for each experiment in RESULTS.

Data depict the mean ± SEM. Statistical analyses were performed using one-way or two-way ANOVA followed by relevant post hoc tests. Student t tests were applied for comparisons between two groups (GraphPad Prism 9) where P values < 0.05 were considered statistically significant.

All data generated and analyzed during this study are available from the corresponding author upon reasonable request.

To investigate the role of PAK3 in obesity cardiomyopathy, we first assessed PAK3 expression in human hearts obtained from individuals with obesity, which showed an increase in PAK3 expression and phosphorylation (Fig. 1A and B). Furthermore, cultured human heart slices, subjected to 16 h of FAs, also exhibited elevated PAK3 levels and phosphorylation (Fig. 1C). Additionally, we observed increased PAK3 in the myocardium from obese mouse models: C57BL/6 mice fed an HFD for 24 weeks and mice with genetically induced obesity (ob/ob, 15 weeks old) (Fig. 1D and E). Collectively, PAK3 is likely involved in the pathogenesis of obesity cardiomyopathy.

To explore the causative link between PAK3 overexpression and the development of obesity cardiomyopathy, mice injected with either AAV9-Pak3 or AAV9-Gfp (as a control group) (Supplementary Fig. 1A) were fed with standard chow diet or HFD for 16 weeks. Cardiac function was evaluated at 4-week intervals using echocardiography (Fig. 2A). Mice fed an HFD displayed metabolic alterations associated with obesity (Supplementary Fig. 1B–E). A notable reduction in cardiac contractile function was observed in the HFD-fed AAV9-Pak3–injected mice, evidenced by a decrease in fractional shortening (FS%) and ejection fraction (EF%) (Fig. 2B and Supplementary Table 1). Additionally, PAK3 overexpression prompted diastolic dysfunction, as shown by a prolonged isovolumic relaxation time (IVRT) as early as week 8 after HFD feeding and a longer IVRT at week 16 compared with AAV9-Gfp–injected mice (Fig. 2C and Supplementary Table 1), suggesting that PAK3 overexpression accelerates the development of diastolic and systolic dysfunction in obesity.

Serum lactate dehydrogenase levels were higher in PAK3-overexpressing HFD-fed mice (Fig. 2D), indicating cardiac tissue damage. Next, the increase in the cross-sectional area of cardiomyocytes was more pronounced in the AAV9-Pak3 mice compared with the corresponding control group (Fig. 2E), while the interstitial fibrotic area in the myocardium was larger (Fig. 2F), and the apoptotic rate was higher in the PAK3-overexpressing myocardium after 16 weeks of HFD feeding (Fig. 2G). These findings indicate that PAK3 overexpression provokes cardiac pathological remodeling, therefore diminishing the ability to withstand the metabolic stress of obesity.

Since obesity cardiomyopathy correlates with increased lipid accumulation and oxidative stress, we examined intracellular lipid homeostasis and oxidative stress markers. Transmission electron microscopy (TEM) revealed an increase in lipid droplets (LDs) (Fig. 3A), and immunofluorescent staining showed increased perilipin 2 (PLIN2) (Fig. 3B) in PAK3-overexpressing mouse hearts upon diet stress for 16 weeks. Genes involved in lipid metabolism were altered in the fact of PAK3 overexpression; genes responsible for FA uptake and synthesis were upregulated, such as Cd36, Fabp3, Acsl1, Acaca, Acacb, Dgat2; but genes responsible for FAO, such as Cpt1b, Cpt2, Acadl and Acadm, were lower in the fact of PAK3 overexpression (Fig. 3C). Moreover, lipidomic analyses revealed that FAs and toxic DAGs were significantly increased (Fig. 3D), indicating lipotoxicity in the PAK3-overexpressing hearts following HFD feeding.

Next, pronounced oxidative stress was also observed when cardiac PAK3 was overexpressed, as indicated by the increased intensity of lipid peroxidation marker 4-hydroxynonenal (4-HNE) (Fig. 3E). The antioxidative response genes, including Sod2, Gpx1, and Sirt3, were lower in the PAK3-overexpressing myocardium from mice with HFD feeding for 16 weeks (Fig. 3F). An upregulation of the LD markers, PLIN2 and PLIN5, and a downregulation of the antioxidant SOD2 were observed in PAK3-overexpressing obese hearts, accompanied by an increase in the DNA damage marker cleaved poly(ADP-ribose) polymerase (PARP) (Fig. 3G). Altogether, the data imply that increased cardiac PAK3 leads to abnormalities in the expression of genes involved in lipid homeostasis, accelerating the development of myocardial lipotoxicity in obesity.

Consistently, phosphorylation and expression of endogenous PAK3 were increased in cardiomyocytes stressed with PA and OA for 8 h (Supplementary Fig. 2). Moreover, when infected with Ad-Pak3-T421E, an adenovirus carrying constitutively active PAK3 (Supplementary Fig. 3A), followed by a short-term FA stimulation (4 h), we observed cardiomyocyte hypertrophy in comparison with control cells (Supplementary Fig. 3B). Lipid accumulation, reactive oxygen species (ROS) production, and apoptotic cells were increased in the active PAK3-overexpressing cardiomyocytes after short-term stimulation of FAs (Fig. 4A–C). These were corroborated by increased PLIN2, PLIN5, and cleaved PARP, and decreased SOD2 in active PAK3-overexpressing cells (Supplementary Fig. 3C). Notably, FAO was lower in active PAK3-overexpressing cells after FA exposure (Fig. 4D), in addition to higher intracellular DAG content (Fig. 4E) and reduced ATP from the cells (Fig. 4F).

Opposite results were obtained when cardiomyocytes were infected with Ad-Pak3-T421A, an adenovirus overexpressing the kinase-dead PAK3 upon longer-term FA stress (8 h) (Fig. 4G–I and Supplementary Fig. 4). Moreover, FAO was comparable in normal or inactive PAK3-expressing cells under stress (Fig. 4J), whereas kinase-dead PAK3 decreased DAGs and preserved ATP (Fig. 4K and L). Thus, these observations reinforce the in vivo findings that PAK3 impairs cardiac lipid homeostasis, and further suggest that its activation is key to the detrimental effects on cardiomyocyte function.

Since we discovered that the genes involved in lipid homeostasis were affected by PAK3, we first tested whether PAK3-induced lipotoxicity arises through transcriptional regulation of lipogenic genes. SREBP1c is a master transcription factor governing the expression of lipogenesis-associated genes (19). We found that nuclear SREBP1c was increased in the failing human and mouse hearts under obesity (Supplementary Fig. 5), and in cardiomyocytes overexpressing active PAK3 under FA stimulation (Fig. 5A). We confirmed that, even though cytosolic full-length SREBP1c level was not affected, the cleaved SREBP1c as the maturated format was increased in the nuclei of cells because of PAK3 overexpression (Fig. 5B). Consistently, the dominant expression of SREBP1c in nuclei was also observed in PAK3-overexpressing obese myocardium (Fig. 5C).

Previous studies have demonstrated that the mTORC1 pathway controls maturation of SREBP1c processing via activation of ribosomal protein S6 kinase β-1 (S6K1) (20,21). We had observed that PAK3 hyperactivates mTORC1 in cardiomyocytes (17). Here, we show not only that constitutively active PAK3 increased phosphorylation levels of mTOR and S6K1, but also that kinase-dead PAK3 hindered their activation (Supplementary Fig. 6). PAK3 activation of mTORC1 also negatively regulated autophagy, evidenced by lower ratio of LC3II and LC3I and higher p62 (Supplementary Fig. 6). Consistently, augmented phosphorylation of mTOR and S6K1 were observed in PAK3-overexpressing myocardium from mice fed with HFD diet (Fig. 5D). Of note, a specific S6K1 inhibitor, PF-4708671 (22), blocked nuclear SREBP1c in cardiomyocytes, albeit with PAK3 overexpression (Supplementary Fig. 7A and B). Accordingly, the key genes for lipid synthesis transcriptional regulated by SREBP1c were reduced by PF-4708671 in active PAK3-overexpressed cells (Supplementary Fig. 7C). As a result, lipid accumulation and oxidative stress were both attenuated (Supplementary Fig. 7D and E). These data indicate that PAK3 potentiates mTOR and S6K1 pathway processing SREBP1c in cardiomyocytes upon FAs.

On a further attempt to determine the treatment potential, AAV9-Pak3 injected mice were fed an HFD for 12 weeks, followed by treatment with PF-4708671 for 4 weeks (Fig. 5E). Cardiac dysfunction in PAK3-overexpressing HFD-fed mice was mitigated (Fig. 5F and G and Supplementary Table 2). Neutral lipid amount was reduced by the treatment (Fig. 5H), as was oxidative stress (Fig. 5I). More importantly, nuclear SREBP1c was reduced by PF-4708671 (Fig. 5J); consequently, its targets were downregulated (Fig. 5K). These data provide the functional evidence indicating that PAK3 contributes to cardiac lipotoxicity via SREBP1c in an S6K1-dependent manner.

To demonstrate that PAK3 disrupts intracellular lipid homeostasis through regulation of SREBP1c, we attenuated SREBP1c levels using a small molecule, betulin, under PAK3 overexpression. Betulin directly binds to SCAP to inhibit the cleavage and maturation of SREBP1 (23,24). The efficacy of betulin in inhibiting SREBP1c was seen in cardiomyocytes (Supplementary Fig. 8). PAK3-promoted SREBP1c accumulation in the nucleus was ameliorated by betulin (Fig. 6A and B). Notably, in active PAK3-overexpressing cells, the key genes regulated by SREBP1c were reduced by betulin (Fig. 6C). Following short-term FA stimulation, PAK3-provoked LD deposition was mitigated by betulin (Fig. 6D and Supplementary Fig. 9A). Additionally, cell hypertrophy was attenuated (Supplementary Fig. 9B), and both ROS production and cell death were mitigated by betulin (Fig. 6E and F). The above data support that increased PAK3 exerts its deleterious effects on cardiac lipid homeostasis via SREBP1c under metabolic stress.

We next tested whether the administration of betulin could exert effects on mitigating PAK3-induced cardiac dysfunction in response to HFD. To do so, AAV9-Pak3–injected mice were fed an HFD for 8 weeks, followed by botulin treatment for 8 weeks (Fig. 7A). Cardiac dysfunction in PAK3-overexpressing HFD-fed mice was attenuated by betulin (Fig. 7B and C and Supplementary Table 3), even though it did not affect body weight (Supplementary Fig. 10A). More importantly, myocardial lipid content was reduced by the treatment, despite PAK3 overexpression in the obese heart (Fig. 7D and E), accompanied by fewer toxic lipids (Fig. 7F). Consequently, cardiac remodeling, oxidative stress, and apoptotic cells in PAK3-overexpressing myocardium were also alleviated (Supplementary Fig. 10B–F). As anticipated, PAK3-exacerated increment of nuclear SREBP1c in the myocardium was blunted by betulin, although it did not affect the expression and phosphorylation of PAK3 (Fig. 7G). SREBP1c targets were therefore downregulated (Fig. 7H). The data indicate that betulin-mediated blockage of nuclear SREBP1c mitigates lipogenesis, restraining PAK3-triggered cardiac lipotoxicity and attenuating the development of obesity cardiomyopathy.

Prompted by the observations that inactivated-PAK3 impeded LD deposition, we next examined the potential of PAK3 as a therapeutic target to prevent cardiac lipotoxicity. First, siPak3 was used to silence PAK3 in rat cardiomyocytes (Supplementary Fig. 11A). The rise in LD accumulation following exposure to prolonged FAs was ameliorated by PAK3 knockdown (Supplementary Fig. 11A–C). Furthermore, prolonged stimulation-induced oxidative stress was mitigated by PAK3 knockdown (Supplementary Fig. 11C–E). In addition, elevated DAGs and decreased ATP from the cells were prevented by PAK3 knockdown (Supplementary Fig. 11F and G). Altogether, our findings demonstrate that suppression of PAK3 ameliorates lipotoxicity.

Finally, we sought human-relevant evidence for PAK3-associated obesity cardiomyopathy and the potential cardioprotective effects of targeting PAK3 under FA stress. Also, PAK3 was increased in hiPSC-Cs by long-term stimulation of FAs (Supplementary Fig. 12A). Next, active PAK3 promoted a significant increase in LD accumulation (Fig. 8A), accompanied by increased oxidative stress and apoptosis (Fig. 8B and C and Supplementary Fig. 12B). Conversely, the prolonged FA stress-increased toxic outcomes in hiPSC-CMs were blunted by kinase-dead PAK3 (Fig. 8D–F and Supplementary Fig. 13). Consistently, knockdown of PAK3 led to a significant reduction in the level of LDs in response to prolonged FA stress (Supplementary Fig. 14A and Fig. 8G), as well as oxidative stress and apoptotic event (Fig. 8H and I and Supplementary Fig. 14B). In conclusion, these human-relevant data demonstrate that PAK3 contributes to lipid overload in cardiomyocytes in response to FA stimulation, while suppression of PAK3 can prevent cardiac lipotoxicity.

In the current study, we have demonstrated a causal link between increased cardiac PAK3-S6K1 regulation of SREBP1c and obesity cardiomyopathy, as well as the beneficial effects of inhibiting this pathway on cardiac function in obesity.

Excess body fat dramatically increases the risk of developing heart failure (25). Obesity cardiomyopathy refers to obesity-induced morphological, functional, and metabolic abnormalities in the myocardium, leading to cardiac dysfunction. For instance, ventricular pathological remodeling due to obesity is a risk factor for sudden cardiac death (26). In our study, we also observed cardiac pathological remodeling in a preclinical model of diet-induced obesity, associated with cardiac dysfunction in response to prolonged stress.

One of the most striking findings in our study was that PAK3 is a causative factor in the pathogenesis of cardiac dysfunction in obesity. Cardiac PAK3 was consistently increased in various models of obesity, indicating its involvement in obesity cardiomyopathy. But more importantly, PAK3-overexpressing heart manifested cardiac lipid overload, dysfunction, and remodeling after only 16 weeks of HFD, a time point at which PAK3 levels and cardiac function remain unchanged in wild-type mice, suggesting that PAK3 promotes the development of diet-induced cardiomyopathy. In particular, the impacts of PAK3 on ventricular geometry and function were not observed with chow diet, which is likely explained by the fact that overexpressed PAK3 showed an increase of its phosphorylation only upon diet stress for 16 weeks; however, it needs further study to investigate the upstream of PAK3 activation in response to prolonged obese stress.

FAs, as the primary source of fuel for the heart, can be esterified to form TAGs for storage; however, TAG storage capacity within the cardiomyocytes is minimal (27). In obesity, FA oversupply coupled with a mismatch between cardiac FA uptake and utilization results in accumulation of toxic lipid species, such as TAGs and DAGs (28). Despite an increase of FAO, it is insufficient to prevent myocardial lipid overload in obesity cardiomyopathy (29). On the other hand, lower FAO may impair energy provision and cardiac performance. The failing heart has 20–40% less ATP content than a healthy heart (28). Thus, abnormal FAO along with excessive lipids can cause oxidative stress, pathological myocardial remodeling, and subsequent heart failure (30). PAK3 impaired cardiac lipid homeostasis under metabolic stress, as indicated by the significant toxic lipid overload and oxidative stress in cardiomyocytes. FAO was less by overexpression of active PAK3 under short-term FAs stress; this could explain why ATP production was lower. In addition to more ROS induced by PAK3 activation, intracellular lipotoxicity can lead to defects in ATP production, which is likely due to impaired activity of rate-limiting enzymes in the tricarboxylic acid cycle, a reduced mitochondrial oxidative capacity, or increased ROS levels (31). On the other hand, kinase-dead PAK3 gave rise to a similar FAO phenomenon compared with control cells under longer stress, while ATP amount was higher. This indicates that blockage of PAK3 does prevent oxidative stress and energy deprivation, but not through modulation of FAO. Although it is admitted that a longer-term stress of FAs promotes lipotoxicity along with apoptosis through multiple pathways, either knockdown or suppression of PAK3 prevented the vicious effects in cardiomyocytes under our experimental durations, giving it the potential to become a promising therapeutic target in obesity cardiomyopathy. However, the treatment outcomes in in vivo models upon various durations of stress and the mechanisms underlying PAK3 reduction of ATP production in obesity cardiomyopathy require more in-depth studies.

SREBP1 is a master transcription factor governing the expression of critical lipogenesis genes (19,32). SREBP1c contributes to cardiac dysfunction by promoting FA uptake and lipid accumulation within cardiomyocytes (14,33). Consistent with SREBP1c-induced lipotoxicity, we determined that PAK3 overexpression led to an alteration of lipid gene profiles. In addition, PAK3 inhibited antioxidant systems under obese stress; however, it cannot be concluded whether the reduced antioxidant genes are due to SREBP1c activation, are due to PAK3 modulation of other transcription factors upregulating antioxidant genes, or are an indirect consequence. Nevertheless, PAK3 is likely a nodal modulator for the vicious cycle of FA-induced oxidative stress in cardiomyocytes. Given that activated PAK3 altered the expression of genes for lipid uptake, synthesis, and antioxidative stress in response to FA stress, it is theorized that PAK3 exacerbates lipid overload in obesity cardiomyopathy through multilevel molecular mechanisms. Of note, PAK3 overexpression itself does not activate the SREBP signaling pathway under normal diet, which is consistent with the impacts of PAK3 on cardiac phenotypes. This is likely because PAK3 activation is one of the key causative factors to trigger heart failure upon diet stress, along with activation of other signaling pathways. This study has provided proof-of-concept evidence that PAK3 is required for promoting heart failure in obesity, prompting further investigation of the assistant pathways involved in this procedure.

Once transported into the Golgi apparatus, SREBP1 is subjected to proteolytic cleavage by proteases to generate its activated form (34). Previous studies have presented that the mTORC1 signals enhance the maturation of SREBP1c via S6K1 (20,21). S6K1 canonically governs protein expression and turnover; we hypothesize that S6K1 may control site-1 protease at Golgi for the activation of SREBP1c. Furthermore, mTOR signaling can modulate coat protein complex II–dependent SREBP1 trafficking to Golgi (35). PAK3 phosphorylated mTOR-S6K1, involved in facilitating SREBP1c processing and its nuclear translocation in cardiomyocytes. More convincingly, the S6K1 specific inhibitor PF-4708671 prevented SREBP1c nuclear distribution in PAK3-overexpressing cells, upregulation of its key targets, lipid overload, and cardiac dysfunction. Both in vitro and in vivo functional observations supported that PAK3 triggers lipotoxicity at least partially via activation of S6K1.

Additionally, various SREBP1c posttranslational modifications impact its expression and activation, although these mechanisms remain ambiguous in the cardiomyocytes. AMP-activated protein kinase (AMPK) directly phosphorylates nuclear SREBP1c at the serine site (Ser372), suppressing its activity (36). We previously detected inhibition of AMPK by active PAK3 (17); PAK3 could be indirectly upregulating SREBP1c activity via defeated AMPK. Furthermore, glycogen synthase kinase 3 β (GSK3β) phosphorylates SREBP1c at Ser73 in liver, inducing its degradation by ubiquitylation (37). Increased mTORC1 downregulates the kinase activity of GSK3β (38). As such, whether increased SREBP1c activity in PAK3-overexpressing cardiomyocytes upon FAs is through modulating AMPK or GSK3β warrants further investigation.

Lipolysis releases FAs from LDs to generate energy (39). Autophagy and lysosome-mediated lipolysis prompt degradation of intracellular lipid stores (40). It has been reported that mTORC1 prevents lipophagy by blocking autophagy in adipocytes and hepatocytes (41,42). Given that phosphorylation of mTOR was enhanced by constitutively active PAK3, and autophagy was negatively regulated under FA stress, it is possible that FA-induced lipid overload may also be enacted via PAK3 suppression of lipophagy, which needs further investigation.

Betulin is a pentacyclic triterpene natural product isolated from the bark of birch trees, which can reduce the biosynthesis of FAs in liver and adipose tissue by either inhibiting SREBP1 cleavage and activation or enhancing AMPK-mediated suppression of SREBP (23,24,43). Here, for the first time, we have demonstrated the effects of betulin treatment on cardiac function in obesity. Betulin shows antidiabetic effects (44), although we did not observe improved systemic metabolism in the treated mice, which was likely because of the low dose of treatment. However, betulin ameliorated PAK3-triggered LD accumulation and toxic lipid overload, consequently preserving cardiac function in response to stress from a high-calorie diet and highlighting its sensitive function in preventing cardiac lipotoxicity even with a low dose.

In summary, this research has illustrated that cardiac PAK3 interferes with lipid homeostasis in the obese myocardium through modulation of SREBP1c. Furthermore, we have provided evidence demonstrating that inhibition of PAK3-S6K1-SREBP1c mitigates lipotoxicity and confers cardiac protection in the setting of obesity.

Acknowledgments. The authors are grateful to several facilities at The University of Manchester for providing valuable technical assistance throughout this project. Firstly, we thank Steven Marsden and Peter March in the Bioimaging Facility for microscopy training. Secondly, we thank Aleksandr Mironov and Samantha Forbes at the Electron Microscopy Core Facility for TEM technical training and support. We also acknowledge George Taylor and David Knight of the Mass Spectrometry Facility for professional advice, processing samples, conducting experiments, and performing lipidomic data analyses.

Funding. This work was supported by the British Heart Foundation (FS/15/16/31477, FS/18/73/33973, FS/19/70/34650, PG/19/66/34600, FS/PhD/22/29307, PG/22/10904, and PG/PG/22/11075 to W.L.; BHF Accelerator award AA/18/4/34221 to University of Manchester; and FS/18/4/33310 to C.P.) and a professorship from the German Centre for Cardiovascular Research (81Z0700201 to O.J.M.). T.M.A.M. is supported by National Institutes of Health (NIH) grants R01HL147921 and P30GM127607, U.S. Department of Defense grant W81XWH-20-1-0419, and American Heart Association grant 16SDG29950012. The authors also acknowledge NIH grant F32HL149140 (R.R.E.A.). This work was also supported by the NIHR Manchester Biomedical Research Centre (NIHR203308).

Duality of Interest. T.M.A.M holds equities at Tenaya Therapeutics. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. W.L. conceptualized the idea for the project detailed in this manuscript. X.C., A.R.-V., and Z.Z. contributed significantly to experimental design and data acquisition, analysis, and interpretation, with assistance from C.R., O.F., S.R.G., R.R., J.Z., N.K., and H.M.-D. W.L., N.H.o.A., J.M.M., R.R.E.A., Q.O., and T.M.A.M. cultured and treated human heart slices. S.S.H., D.F., and O.J.M. prepared AAV9-cTnT-Pak3, monitored the project, and reviewed the manuscript. C.P. provided technical support for TEM. X.Z. and T.W. provided technical support regarding hiPSC experiments. This manuscript was primarily drafted by W.L. with contributions from X.C., Z.Z., and C.R. Finally, M.K.R. provided mentorship and contributed to the manuscript. W.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.