Metabolic diseases are a group of chronic conditions caused by metabolic abnormalities. These conditions include type 2 diabetes mellitus (T2DM), obesity, and non-alcoholic fatty liver disease (NAFLD). Over the past two decades, the prevalence and mortality rates of metabolic diseases have increased significantly, imposing a substantial disease burden, particularly among males [1] . The prevalence of NAFLD, for example, is strongly associated with metabolic syndrome, and the risk of progression to cirrhosis and hepatocellular carcinoma is also increasing year by year [1] [2] . Meanwhile, the COVID-19 pandemic has further exposed the health vulnerabilities of patients with metabolic diseases, with significantly higher rates of severe illness and mortality after infection in this population, highlighting the complex interaction between metabolic disorders and immune imbalance [3] . Metabolic diseases pose a significant threat to global public health. Despite the emergence of novel therapies, such as gut flora modulation and epigenetically targeted treatments, gut flora modulation represents a novel therapeutic avenue. Microbial metabolites, such as short-chain fatty acids (e.g., butyrate), can directly signal to host intestinal and peripheral cells (including β-cells), activating autophagy pathways (e.g., via AMPK or HDAC inhibition) and thereby influencing metabolic health [4] . The global disease burden remains uncontrolled [1] , and further breakthroughs are urgently needed from pathological mechanisms to clinical interventions [5] [6] .

Autophagy is a highly conserved intracellular “cleaner” system that selectively removes damaged mitochondria, misfolded proteins, and excess lipid droplets via the lysosomal pathway, playing a central role in the regulation of metabolic homeostasis. Studies have shown that in metabolic diseases such as NAFLD, autophagy dysfunction forms a vicious circle with lipotoxicity, mitochondrial damage and inflammatory response. For example, in T2DM, dysregulation of autophagy, which maintains insulin secretion by removing aberrant protein aggregates (mainly composed of misfolded islet amyloid polypeptides (IAPP)) from β-cells, results in islet amyloid deposition and impaired glucose tolerance [7] [8] . In obesity models, autophagy deficiency in adipose tissue suppresses lipolysis and thermogenesis, exacerbating white adipose tissue expansion and systemic inflammation [9] [10] . In addition, metabolic abnormalities in the skeletal and cardiovascular systems are also closely related to autophagy imbalance, e.g., defective autophagy in osteoblasts triggered osteoporosis through impaired mitochondrial clearance [11] , whereas dysregulated autophagy in cardiomyocytes exacerbates ischemia-reperfusion injury [12] . All of the above findings suggest that autophagy defects are important drivers of metabolic disorders and highlight the potential of targeting autophagic pathways for multi-organ therapy. Therefore, in-depth investigation of the relationship between autophagy regulation and the development of metabolic diseases will be of great significance for the prevention and treatment of metabolic diseases.

In this article, we summarize the core strategies for regulating autophagy in the treatment of metabolic diseases (including lifestyle intervention, pharmacological modulation, and the application of natural compounds), analyze the scientific basis and clinical limitations of the existing therapies, and explore the future direction of tissue-specific drug development, precision medicine application, and interdisciplinary technology integration, with the aim of providing theoretical basis and practical reference for the development of safer and more effective therapeutic options for metabolic diseases.

Decreased activity of autophagy, as one of the core mechanisms for cellular removal of damaged components and maintenance of energy homeostasis, is closely associated with metabolic disorders [13] [14] . Caloric Restriction (CR), as a non-pharmacological intervention to activate the autophagy pathway through a multidimensional signaling network, has become one of the important strategies to ameliorate metabolic diseases [15] [16] . Studies have shown that CR can directly upregulate the expression of autophagy-related proteins (e.g., LC3, Beclin1) and improves mitochondrial function by inhibiting the mammalian target of rapamycin protein complex 1 (mTORC1) signaling axis and activating the adenylate-activated protein kinase (AMPK) pathway [15] . In peripheral blood mononuclear cells from obese patients, CR intervention significantly enhanced AMPK activity and reduced mitochondrial reactive oxygen species (ROS) levels, accompanied by up-regulation of the expression of mitochondrial kinetic-related proteins (MFN2, DRP1, FIS1), suggesting that CR synergizes with mitochondrial autophagy to activate autophagic flow through energy induction [15] . In the Akt2 knockout mouse model, CR improved cardiac metabolic function by regulating the PINK1/Parkin pathway, and its elevated LC3-II/I ratio was positively correlated with the reduction of mitochondrial ROS, revealing a bidirectional regulatory mechanism of mitochondrial function and autophagy activity [17] . Notably, the metabolic protective effect of CR also involves the remodeling of neuroendocrine and intestinal flora. Animal studies have shown that hypothalamic neuropeptide Y (NPY) mimics the CR effect by enhancing CNS autophagic activity [18] , whereas changes in the abundance of gut flora improve insulin sensitivity through the immune-metabolic axis [16] . In addition, CR exhibits specific effects in different metabolic tissues through autophagy modulation. In adipose tissue, CR combined with exercise training significantly reduced visceral fat mass percentage [19] , up-regulated the expression of autophagy-related proteins such as ATG5 and ATG7 [20] , and inhibited the abnormal accumulation of IgG to alleviate macrophage-mediated chronic inflammation [21] . In the liver, CR attenuates the lipotoxicity of NAFLD by up-regulating hepatic autophagy genes such as BECN1 and NBR1 and down-regulating pro-inflammatory gene expression [22] [23] . The cardiovascular system, on the other hand, reduces fibrosis after ischemia-reperfusion injury through CR-induced autophagy in cardiomyocytes and inhibits mTORC1 signaling to delay cardiac aging [12] [24] [25] . It is also important to note that there are significant gender and age differences in the metabolic effects of CR. Studies have shown that young male mice showed a more pronounced decrease in fat mass and improvement in insulin sensitivity after CR intervention, whereas female mice showed a weaker response [26] , and altered miRNA expression profiles in the subcutaneous fat of middle-aged animals suggest that CR may delay metabolic aging through epigenetic regulation [27] .

Although studies have shown that CR mimics (e.g., spermidine and 3,4-dimethoxychalcone) can activate autophagy and ameliorate the metabolic syndrome through pathways such as inhibition of IGF1R signaling [28] [29] , prolonged or excessive CR may trigger problems such as excessive activation of autophagy leading to cell death [30] . Yadin et al. demonstrated that aberrant activation of autophagy aggravated cardiomyopathy [31] , whereas cyclic CR patterns significantly improved cognitive function and intestinal health in aging mice [32] , suggesting that autophagy intensity needs to be cautiously regulated in the context of specific diseases. Therefore, future studies need to combine the genetic background and metabolic status to formulate personalized protocols and deeply resolve the organ-specific mechanism of CR-regulated autophagy. The integration of artificial intelligence and multi-omics technologies may help to reveal the dynamic association of the CR-autophagy-metabolism network and provide theoretical support for the development of targeted intervention strategies, thus promoting caloric restriction as an important pillar of precision therapy for metabolic diseases.

Currently, research on drug-regulated autophagy for the treatment of metabolic diseases primarily centers on metformin, SGLT2 inhibitors, and rapamycin. Table 1 provides a summary of recent advances in the application of these three agents—metformin, SGLT2 inhibitors, and rapamycin—in the treatment of metabolic diseases.

Some natural compounds (e.g. resveratrol, inulin, catechins, etc.) may provide low-toxicity and synergistic therapeutic strategies for metabolic diseases by activating the AMPK/SIRT1 pathway or inhibiting mTOR signaling, enhancing autophagy activity, promoting degradation of lipid droplets and restoration of mitochondrial function, and ameliorating hepatic lipid accumulation and oxidative stress. Inulin is a group of natural polysaccharides, that has been shown to indirectly enhance hepatic autophagy to inhibit the progression of non-alcoholic steatohepatitis (NASH) by enriching Parabacteroides distasonis to produce pentadecanoic acid, restoring intestinal barrier function and reducing endotoxemia [78] . Resveratrol is a polyphenol that is believed to have anti-inflammatory and immunomodulatory activities. Animal experiments have shown that the administration of resveratrol induces autophagy in the organism, and the increase in the levels of the markers beclin-1 and Atg5 protein and the decrease in the expression of p62 are effective in preventing hepatic lipid accumulation [79] . Its potential to attenuate hepatocellular lipid accumulation and inflammation was also supported in a methionine-choline-deficient NASH model [80] . Further studies have shown that the combination of resveratrol and metformin also attenuates hepatic steatosis by activating autophagy and reducing lipid accumulation via the cAMP/ AMPK/SIRT1 signaling pathway [81] . In addition, randomized controlled trials have shown that a green Mediterranean diet rich in polyphenols, such as catechins, also significantly reduces intrahepatic fat content, and its effect is closely related to autophagy-mediated lipid clearance [82] . Notably, some natural active ingredients may affect host metabolism by regulating gut flora metabolites (e.g., short-chain fatty acids) through targeted modulation of related proteins such as RPN11 and ACC1 [83] - [85] , suggesting a multidimensional effect of natural compounds in ameliorating lipotoxicity and metabolic disorders. The above evidence suggests that natural compounds often achieve metabolic protection by synergistically modulating oxidative stress, mitochondrial metabolism and autophagy pathways, e.g. targeting ACC1 degradation or inhibition of RPN11 improves both insulin sensitivity and fibrotic progression [83] - [85] .

However, the clinical application of natural compounds still faces many challenges. Low bioavailability is a major bottleneck. Studies have shown that the effects of resveratrol on hepatic steatosis are significantly hampered in the presence of autophagy inhibitors (3-methyladenine, Bafilomycin A1, and chloroquine) or interventions targeting autophagy-related genes (Atg5 or beclin-1), suggesting that the beneficial effects of resveratrol are closely linked to its regulation of autophagy [86] [87] . Individual heterogeneity (e.g. PNPLA3 I148M gene variant) and differences in gut flora also significantly affect efficacy [88] [89] , and individualized regimens need to be developed based on genetic background and microbial composition analysis. In addition, excessive activation of autophagy may exacerbate tissue damage. For example, the multi-targeted effects of natural compounds may interfere with the dynamic balance between autophagy and other cell death pathways (e.g., apoptosis) under specific pathological conditions [87] . Future research needs to combine precision medicine with interdisciplinary technologies to break through existing limitations. The development of natural derivatives targeting key autophagy nodes (e.g., ULK1, LC3-II), combined with single-cell sequencing to resolve the dynamics of tissue-specific autophagy, is expected to achieve precise regulation. Secondly, synergistic gut-liver axis intervention strategies (e.g., combining prebiotics with polyphenolic compounds) may ameliorate metabolic disorders through the dual action of flora metabolites (pentadecanoic acid) and autophagy pathways [78] . Furthermore, the use of artificial intelligence-assisted drug design accelerates the high-throughput screening of natural compound libraries, e.g. using hepatocyte organoid models to predict multi-target synergistic effects, optimizing candidate molecules, and investigate their mechanisms and drug targets [88] . For clinical translation, stratified clinical trials based on biomarkers (e.g. plasma OxPLs) are needed to clarify the causal association between autophagy modulation and metabolic endpoints (e.g. liver fibrosis reversal) [90] . By integrating multi-omics technologies and delivery system innovations, natural compounds are expected to be upgraded from “dietary interventions” to “targeted autophagy drugs”, providing a new direction for personalized treatment of metabolic diseases.

While autophagy, as a core regulatory mechanism of cellular metabolic homeostasis, shows immense potential in treating metabolic diseases, its clinical application faces multiple challenges and scientific controversies. While protective at basal levels, excessive or dysregulated autophagy can become cytotoxic. This can occur through crosstalk with apoptosis pathways (e.g., sustained Beclin-1 activation can promote apoptosis via Bcl-2 dissociation), or via impaired cellular function resulting from unrestrained mitophagy leading to bioenergetic failure [91] . The current study reveals that the “double-edged sword” effect of autophagy significantly limits its therapeutic window—for example, rapamycin ameliorates hepatic lipid accumulation through activation of autophagy [52] [53] [68] , but its systemic mTOR inhibition may trigger decreased insulin sensitivity and immune suppression and immunosuppression [51] [56] [70] [71] . Whereas in NASH, over-activated autophagy may accelerate abnormal mitochondrial degradation and instead exacerbate oxidative stress and fibrosis [83] - [85] [92] . This paradoxical effect suggests that future drug development needs to focus on tissue-specific regulatory strategies (e.g., targeting hepatic lipophagy or myocardial mitochondrial autophagy) [93] [94] rather than global activation of autophagy pathways. In addition, population heterogeneity and differences in pathological mechanisms further increase therapeutic complexity. Hepatic lipotoxicity (e.g., activation of the diacylglycerol-PKC-ε pathway) may be preferentially ameliorated by autophagy modulation in obese patients [95] , whereas β-cell dysfunction in diabetic patients may be attenuated by excessive autophagy leading to degradation of proteins associated with insulin secretion [7] [8] [51] [96] [97] . Genetic factors [88] [89] and gender differences (e.g., women are more sensitive to lifestyle interventions) [98] reinforce the need for individualized treatment strategies. Therefore, the integration of metabolic phenotypic stratification (e.g., hepatic fat content, insulin secretion capacity) with multiple groups of biomarkers (e.g., exosomal miR-690, plasma acetyl coenzyme A levels) will be a potential direction for precision therapy [89] [95] [99] [100] . At the clinical translational level, innovation in autophagy activity detection technology remains a key issue. Existing methods rely on invasive biopsies or indirect indicators [101] , which are unable to monitor autophagy flux dynamically. Despite the emerging potential of non-invasive detection technologies based on exosomal miRNA or imaging histology [94] , their specificity and mechanistic relevance still need to be validated on a large scale. In the future, a standardized autophagy activity assessment system should be established, combined with artificial intelligence-assisted metabolic imaging analysis for dynamic optimization of therapeutic regimens.

Based on the above challenges, the clinical application of autophagy modulation still needs to seek breakthroughs in innovative strategies. In terms of interventions, priority should be given to promoting safer lifestyle interventions (e.g. caloric restriction/intermittent fasting) and optimizing existing drug combination regimens—e.g. combining metformin with GLP-1 receptor agonists to synergistically modulate autophagy through AMPK activation and enteroglycoprokinetic effects [102] [103] , or developing tissue-specific delivery systems (e.g. liver-targeted ACC1 inhibitors [83] [84] ) to reduce systemic side effects. At the technical level, there is a need to accelerate the development of novel modulators targeting key nodes of autophagy (e.g. ULK1, LC3-II) and to resolve tissue-specific regulatory networks using single-cell sequencing and organ-like models. In addition, stratified clinical trials based on genetic risk scores (e.g., PNPLA3 variants [89] ) and metabolomic profiles will drive the shift from “broad-spectrum intervention” to “precision repair” in therapeutic strategies.

In summary, autophagy regulation opens up a promising new pathway for the treatment of metabolic diseases, but its clinical translation needs to strike a balance between mechanistic depth and therapeutic safety. Through interdisciplinary integration and innovative clinical study design, it is promising to make the leap from basic discovery to clinical benefit and provide safer and more effective personalized therapeutic solutions for metabolic syndrome patients.

We would like to express our special thanks to Dr. Hongmei Wu and Mingxue Zhang for their linguistic assistance on the revision of this manuscript.

*The authors contributed equally to this work and shared first authorship.

^#Corresponding author.