Interactions between lipid droplets and mitochondria in metabolic diseases

Abstract

Metabolic diseases, a major challenge in global public health, are commonly characterized by insulin resistance, lipid metabolism disorders, and mitochondrial dysfunction, and their pathological processes are often accompanied by the abnormal accumulation of lipids in metabolically active tissues such as the liver, heart, and skeletal muscle. Recently, lipid droplets and mitochondria have been shown to interact with each other through membrane contact sites and play a central role in maintaining cellular metabolic homeostasis. The unique monolayer phospholipid membrane structure and formation process of lipid droplets, along with the double-membrane structure and diverse functions of mitochondria, together form the basis for their interaction. There are two modes of interaction, namely dynamic contact and stable anchoring, which are mediated by a variety of proteins to achieve efficient exchange and metabolic regulation of metabolites such as fatty acids. However, dysregulation of lipid droplet–mitochondria interactions initiates a pathogenic cascade involving fatty acid overload, increased reactive oxygen species generation, and mitochondrial dysfunction. These perturbations drive the pathogenesis of metabolic disorders. This review systematically summarizes the key pathological roles of dysregulated lipid droplet–mitochondrial interactions in globally prevalent metabolic diseases such as diabetes mellitus, metabolic dysfunction-associated fatty liver disease, and obesity. This in-depth analysis of molecular mechanisms clarifies the physiological basis of the regulation of lipid homeostasis in the body and provides a theoretical basis for developing novel therapeutic strategies for these diseases to alleviate the related growing global health burden.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12944-025-02759-4.

Introduction

Lipid droplets (LDs) are surrounded by a monolayer of phospholipid membranes that encapsulate a neutral lipid core, and play crucial roles in lipid metabolism and energy regulation. Their number and size are regulated by the intracellular environment [1]. Mitochondria, which serve as the cell’s energy powerhouse, are critical for integrating energy production, metabolic pathways, and redox signalling [2]. Membrane contact sites (MCSs) refer to the regions where organelles form close contact through specific protein complexes (typically within the range of 10–30 nm), without membrane fusion [3]. LDs and mitochondria form MCS-mediated interfaces to exchange fatty acids (FAs) and coordinate signals, supporting cellular energy metabolism [4]. Multiple types of proteins, including functional proteins that facilitate the exchange of ions, proteins, lipids, or metabolites, and regulatory proteins and recruitment proteins for efficient metabolite transfer, mediate the MCSs between LDs and mitochondria. Moreover, the metabolic state of cells dynamically regulates the formation and dissociation of MCSs. As a result, when different types of cells are under different metabolic conditions, the protein composition within the MSC changes and undergoes adaptive remodelling in response to alterations in metabolic conditions [5, 6]. Therefore, lipid droplet–mitochondria crosstalk (LDMC) is key for coordinating lipid storage, mobilization, and oxidation, directly affecting cellular metabolic homeostasis.

Although first observed in the last century, lipid droplet-mitochondria contact (MLC) has only been studied for the past decade [7]. Thus, this review aimed to clarify the different modes of action and to analyse in depth the tissue-specific mechanisms of key proteins in disease development, to provide a clear theoretical framework and research direction for the development of innovative therapeutic strategies.

The mechanisms of lipid droplet–mitochondrial cross talk

Mitochondria are the core of cellular metabolism and convert FAs stored in LDs into energy in lipid forms such as triglycerides (TAGs) through β-oxidation. Moreover, ATP and other metabolites produced by mitochondria also provide necessary support for the generation and expansion of LDs [8]. MLC can be divided into dynamic contact and stable anchoring. These two types of contact do not exist independently but cooperate with each other to jointly regulate lipid metabolism to meet the energy requirements of cells [9].

Dynamic contact

Dynamic contact is a short-lived and reversible interaction. Protein complexes formed by transient proteins connect LDs with mitochondria, conforming to the “kiss-and-run” model. This contact rapid lipid exchange between LDs and mitochondria and allows the rapid adjustment of lipid distribution and utilization following metabolic demands, supporting the rapid metabolic needs of cells [9].

Synaptosome-associated protein 23 (SNAP23), a soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) protein widely expressed in human skeletal muscle, is a key molecule in the lipid droplet–mitochondrial interaction network. SNAP23 is localized to LDs and may be present near the lipid droplet–mitochondrial contact site [10, 11]. As a key component of the SNARE complex, SNAP23 does not directly mediate the interaction between LDs and mitochondria. Instead, it indirectly regulates the contact efficiency between LDs and mitochondria by promoting the fusion of LDs. This process drives the merging of small LDs into larger LDs of optimal size, thereby ensuring efficient transport and oxidation of FAs from LDs to mitochondria [10]. Recent studies have further revealed that SNAP23, together with synaptic fusion protein 18 (STX18) and SEC22B, is located on the surface of lipid droplets and assembles to form a stable SNARE complex, which plays a crucial role in driving lipid-content mixing during lipid droplet fusion. CIDE C (FSP27) directly binds to the STX18-SNAP23-SEC22B complex, enhancing the lipid mixing ability of the SNARE complex, which in turn drives its mediated lipid droplet fusion process [12]. In addition, SNAP23 function is regulated by the metabolic state. Compared with normal feeding, SNAP23 is specifically recruited to the surface of LDs in the fasting state, and its expression level significantly increases; moreover, the presence of the mitochondrial membrane protein VDAC1 has been detected on LDs, suggesting that SNAP23 facilitates physical contact between the LDs and mitochondria during fasting [13]. STX18-SNAP23-SEC22B complex-mediated fusion of LDs also maintains energy homeostasis in organisms under fasting conditions. On the other hand, SNAP23 mediates the insulin-stimulated binding and fusion of glucose transporter protein 4 (GLUT4) with the plasma membrane and participates in the regulation of LD dynamics through interactions with proteins such as perilipin 2 (PLIN2) and vesicle-associated membrane protein 4 (VAMP4) in a fasting environment [14, 15]. Notably, long-chain acyl-CoA synthetase 1 (ACSL1), a key enzyme in FA metabolism, can interact with SNAP23/VAMP4 under glucose-deprived conditions (simulated fasting) in primary hepatocytes. These findings indicate that the complex of ACSL1 with SNAP23/VAMP4 may serve as a molecular platform that facilitates the anchoring of LDs under fasting conditions to the mitochondrial outer membrane for fatty acid oxidation (FAO). However, silencing SNAP23 does not significantly affect the FAO process, suggesting that other related proteins in the SNARE family may compensate for its function, indirectly reflecting the fact that SNAP23 is not the only protein required to mediate lipid droplet–mitochondrial interactions for FAO [16]. Furthermore, tank-binding kinase 1 (TBK1) binds to ACSL1 to influence lipid metabolism and remains inactive during fasting. By inhibiting inflammatory signalling and synergizing with metabolic kinases such as AMP-activated protein kinase (AMPK), TBK1 recruits ACSL1 to mitochondria, optimizing the efficiency of FAO [17]. However, critical knowledge gaps remain regarding the interactome of SNAP23 and its ability to regulate energy homeostasis under various metabolic conditions.

Perilipin 5 (PLIN5) is a “star protein” that mediates the interaction between LDs and mitochondria. As a conserved member of the perilipin family, PLIN5 is predominantly expressed in highly oxidative tissues, including heart, skeletal muscle, and brown adipose tissue (BAT), and dynamically regulates lipid metabolism and energy supply through different molecular interactions and post-translational modifications [18]. PLIN5 is specifically located in the MLC, and its N-terminal region is located through LD anchoring, whereas the C-terminal region recruits mitochondria to the LD, promoting the transport of FAs between them [19]. Under energy stress conditions such as hunger or exercise, AMPK is activated in skeletal muscle cells. Rab8a, a key effector downstream of AMPK signalling, interacts with PLIN5 when it is in an active state, thereby promoting the association between lipid droplets and mitochondria. The Rab8a-PLIN5 complex recruits adipose triglyceride lipase (ATGL), mobilizes TAGs in LDs, hydrolyses them into FAs, and then transports the FAs to mitochondria for β-oxidation to provide energy [20]. The interaction between PLIN5, ATGL and its activator ABDH5 (α-β-hydrolase domain-containing 5) under energy stress to regulate FA uptake in mitochondria also requires the induction of Rab8a in its active state [21]. In addition, PLIN5, after phosphorylation at the Ser155 site mediated by protein kinase A (PKA), synergistically promotes lipolysis with ATGL; simultaneously, PLIN5 regulates the transcription of peroxisome proliferator-activated receptor alpha (PPARα) target genes, enabling cells to adapt to the metabolic reprogramming required for FAO [22]. Like PLIN5, PLIN2 enhances binding to the mitochondrial membrane protein carnitine palmitoyltransferase 1 A (CPT1A) through phosphorylation to adapt to the dynamic regulatory process of energy stress and further promote lipid mobilization [23]. JMJD8 (Jumonji C domain-containing protein 8) is a novel protein located in the lumen of the endoplasmic reticulum (ER) and can inhibit fasting-induced lipophagy. The observation that suppressing PLIN2 phosphorylation reduces energy production suggests the critical role of JMJD8 and PLIN2 in lipid droplet homeostasis, revealing a possible mechanism for managing fat mobilization under conditions of energy deficit [24].

Stable anchoring

Stable anchors are tightly anchored by transmembrane protein complexes, forming long contact durations and are a strong, irreversible contact mechanism that is typically observed in oxidized tissues (e.g., BAT, skeletal muscle, and cardiac muscle). This contact pattern ensures a continuous lipid supply between LDs and mitochondria, supporting efficient energy production [9].

As previously discussed, PLIN5 meets metabolic demands via dynamic contact; however, its ability to mediate MLC involves a dual regulatory mechanism, in which PLIN5 participates in dynamic regulation and the formation of stable anchoring. As an LD anchoring protein, PLIN5 directly binds to FATP4 (ACSVL4) on mitochondria through its C-terminal domain, where it forms a physical tether. This binding is highly stable, especially under starvation conditions, where the phosphorylation of PLIN5 further enhances its interaction with FATP4, promoting the directional transport of FAs from LDs to mitochondria. Notably, knocking out FATP4 completely abolishes the FAs flux from LDs to mitochondria and partially reduces the colocalization of LDs with mitochondria. These findings suggest that in addition to FATP4, other mitochondrial proteins may cooperate with PLIN5 at the MCS to regulate FA transport [25]. The contact pattern is particularly important in high-energy tissues such as the myocardium because of the continuous FAs transport between LDs and mitochondria to ensure energy supply during myocardial contraction [26]. In different tissues, PLIN5 has significant functional specificity and is closely related to cellular energy requirements and the metabolic environment. In cardiomyocytes, PLIN5 protects cardiomyocytes by maintaining lipid droplet stability; in the liver, it is involved in the regulation of lipid metabolism coupled with mitochondrial function and is involved in antifibrotic processes in hepatic stellate cells (HSCs); and in skeletal muscle, PLIN5 regulates the FAs supply to meet the energy demand under different metabolic conditions [25–27], but the molecular basis of its tissue specificity and regulatory mechanisms has not yet been fully elucidated.

Mitofusin2 (Mfn2) is a transmembrane protein found in the outer mitochondrial membrane. Its function is to promote mitochondrial fusion; it regulates lipid metabolism by interacting with tissue-specific proteins in different tissues. The LD-mitochondrial coupling mediated by mitochondrial-localized Mfn2 is thought to require coordination with another LD-localized protein [28]. In cardiomyocytes, Mfn2 forms a transmembrane complex with heat shock cognate protein 70 (Hsc70) localized on LDs, which supports this hypothesis. On the one hand, Mfn2 can promote the breakdown of lipids in LDs, converting them into small molecules such as FAs and providing substrates for mitochondrial β-oxidation, thereby increasing energy production. On the other hand, this complex can regulate the distribution and storage of lipids in cardiomyocytes, reduce the accumulation of lipids in the cytoplasm, and maintain the balance of lipid metabolism within cardiomyocytes [29]. Mfn2 also interacts with members of the perilipin family. For example, in BAT metabolism, Mfn2 expression couples with LDs through mitochondria and maintains mitochondrial oxidative capacity. The coexpression of perilipin 1 (PLIN1) and Mfn2 promotes lipolysis and stimulates fat oxidation, and their interaction is enhanced after stimulation by adrenaline [30]. In contrast, the combination of Mfn2 and PLIN2 is regulated by microRNA epigenetic factors and LD–mitochondrial contact and affects the reprogramming of hepatic lipid metabolism [31]. However, whether the relationships between Mfn2 and other members of the perilipin family are universally present across different tissues remain unclear.

Synergistic effect of two contact modes

Although existing studies clearly indicate the presence of two modes of contact, these modes do not operate independently. Instead, they share regulatory proteins and metabolic signals to establish a synergistic network, which can even switch between the two modes under specific conditions (e.g., altered metabolic states or changes in the protein regulatory network), serving as a crucial mechanism for cellular adaptation to fluctuations in energy levels.

Endoplasmic reticulum-lipid droplet-mitochondria-related proteins

LDMC involves other organelles in addition to LD and mitochondria. The ER, as the starting point of LD biosynthesis, is directly involved in lipid synthesis and transport, and plays an intermediary role in the interaction between mitochondria and LDs to coordinate energy metabolism [32, 33]. Despite the limitations of the existing research conditions, the mechanism of action of many related proteins on MLC is still ambiguous, and understanding the complexity of three-way contact between LDs and ER and mitochondria is crucial for understanding the complexity of intercellular communication and interactions.

The human VPS13 protein family includes four members, including vacuolar protein sorting-associated protein 13D (VPS13D), which localizes to the LD and mitochondrial MCS under starvation conditions. The VPS13 adaptor-binding domain of this protein can directly bind to the endosomal sorting complex required for transport (ESCRT) protein tumour susceptibility gene 101 (TSG101), forming a complex that promotes the dynamic remodelling of lipid droplet membranes (such as contraction and budding). The lipid-binding capacity of the N-terminal domain of VPS13D to free fatty acids (FFAs) indicates its dual regulatory mechanism in fatty acid metabolism: ESCRT-dependent membrane remodelling coupled with lipid transport [34]. To enable the lipid transport necessary for processes involving mitochondria and LDs, study has indicated that the formation of MCSs among various organelles depends on VPS13A [35]. VPS13 is also present in yeast and relies on the recruitment of the mitochondrial outer membrane protein Mcp1 to localize and bind to form stable intermembrane bridging structures [36]. However, how VPS13D specifically anchors the mitochondrial outer membrane, whether VPS13A can act directly on the MLC, and whether similar mechanisms exist in different species are not clear.

Mitoguardin2 (MIGA2) is a mitochondrial outer membrane protein that forms ER-mitochondrial contact sites by interacting with VAMP-associated protein B and C (VAPB), a large hydrophobic lumen in its interior that extracts and internalizes lipids for transmembrane transport [37]. In adipocytes, MIGA2 is thought to participate in the formation of lipid droplet–mitochondria interfaces, and the amphiphilic fragment of the C-terminal end of MIGA2 may bind directly to the surface of LDs to promote LD-mitochondrial junctions, driving de novo lipogenesis in adipocytes [38].

CLSTN3β, an ER membrane protein at the ER–LD contact site, was initially considered to be present only in adipose tissue, limiting LDs amplification by blocking cell death-inducing DFFA-like effector (CIDE)-mediated LDs fusion [39]. Lauren F et al. first reported the essential role of CLSTN3β in liver metabolism. During fasting and on a ketogenic diet, the activation of PPARα induces an increase in CLSTN3β expression. By inhibiting CIDE protein activity, it expands the surface area of LDs and enhances lipolysis. Moreover, it may interact with mitochondrial proteins to form lipid transfer channels, increasing the number of LDs around the mitochondria and shortening the transport distance. Importantly, the excess acetyl-CoA produced in the mitochondria can activate the mitochondrial integrated stress response (mISR), increasing the production of FAO and ketone bodies [40]. However, conclusive evidence indicating a direct interplay between CLSTN3β and specific proteins on LDs or mitochondria is lacking, and whether other proteins are involved in the targeting of CLSTN3β under conditions of CIDE protein inhibition remains unclear.

On the basis of proximity labelling proteomics, a polymeric protein complex composed of extended synaptotagmin (ESYT)1, ESYT2, and VAPB was discovered at the LD–ER–mitochondria interface. Knockout of any protein in this complex restricts FAO derived from LDs, leading to reduced ATP production, remodelling of the cellular lipidome, and triggering stress induced by lipid toxicity. This finding indicates that its function is consistent with that of all proteins found at the MCS. Moreover, β-adrenergic stimulation enhances the lipolytic activity of LDs [32]. ESYT deficiency results in mitochondrial phospholipid remodelling, characterized by a reduction in cardiolipin and phosphatidylethanolamine (PE), impairing oxidative phosphorylation efficiency. These findings strongly support the role of ESYTs in lipid transport [41].

Dual localization of the lipid transfer protein ORP5/8 is observed at ER-plasma membrane and ER-mitochondria contact sites. Research has indicated an interaction between ORP5/8 and PTPIP51, which resides on the outer membrane of mitochondria. This interaction, mediated by its lipid transfer ORD domain, promotes the transfer of phosphatidylserine (PS) from the ER to mitochondria, thereby promoting PE synthesis. This process is crucial for retaining mitochondrial crista architecture and respiratory function, providing the foundation for intracellular lipid transport through synergistic interactions with the MIB/MICOS complex [42, 43]. Another study revealed a novel function of ORP5/8 in lipid droplet biogenesis in which ORP5/8 specifically localizes to PS-rich mitochondria-associated membrane (MAM) subdomains, and recruits and interacts with seipin, a key regulator of LD formation, to establish a three-way contact point between the ER, mitochondria, and LDs. This efficiently initiates and positively regulates LDs nucleation and growth. Their functional loss specifically impairs LD formation originating at the MAM, leading to overall disruption of LD biogenesis. This discovery reveals a novel role for ORP5/8 in lipid metabolism and, more importantly, establishes the MAM as a critical platform for LD biogenesis, revealing novel insights of how organelle interactions are regulated to maintain lipid homeostasis [43]. Additionally, under specific conditions (such as oleic acid treatment, which induces LDs formation), ORP5 is localized to the ER‒LD contact site and interacts with LDs via its ORD ligand-binding domain. When ORP5 is absent, PI(4)P accumulates extensively on the surface of the LD and PS cannot be effectively transported to the LD. This alteration in phospholipid composition promotes the biosynthesis of neutral lipids (e.g., TAG) and alters the physical properties of the LD surface (e.g., membrane tension and curvature), ultimately leading to abnormal enlargement of the LD. Notably, its homologue ORP8 does not exhibit this phenomenon, primarily because ORP8 lacks the ability to target LDs and cannot be recruited to ER‒LD contact sites to perform its function [44]. In general, ORP5/8, as a bridge connecting ER‒mitochondria‒LD, can coordinate the regulation of cellular lipid homeostasis.

LD–mitochondria contact mainly manifests as dynamic contact and stable anchoring, and is regulated by related proteins to control FA transfer. Additionally, since LDs originate from the ER, ER‒LD–mitochondria interactions collectively participate in the regulation of metabolism (Fig. 1). Although some studies have identified proteins involved in FA transfer in MLC, most studies focus on single proteins or small protein complexes for unfolding; therefore, systematic and comprehensive studies to clarify whether these known components work together synergistically at the contact site are lacking, and the specific contact modes through which efficient transfer of FA is achieved are still unclear.

Fig. 1.

Two lipid droplet–mitochondria contact modes and proteins related to the endoplasmic reticulum–lipid droplet–mitochondria network. This figure depicts two different modes of contact between lipid droplets (LDs) and mitochondria: dynamic contact and stable anchoring. Dynamic contact connects LDs and mitochondria through transient protein complexes that mediate rapid lipid exchange and maintain normal cellular physiological processes under conditions such as energy stress. Stabilizing anchoring supports energy production through the formation of spike-like complexes by transmembrane proteins that ensure a continuous lipid supply between LDs and mitochondria, and the two work in concert to regulate intracellular lipid metabolism and energy homeostasis. In addition, the endoplasmic reticulum (ER), as the starting point of LD biosynthesis, plays an important role in the coordination of cellular metabolism through the construction of an ER–LD–mitochondria ternary interaction network

Metabolic diseases caused by dysregulation of lipid droplet–mitochondria interactions

Metabolic diseases, including diabetes mellitus (DM), metabolic dysfunction-associated fatty liver disease (MAFLD), cardiovascular diseases (CVDs), obesity, metabolism-associated neoplasms, gout and thyroid disorders, are a group of disorders marked by abnormal regulation of the body’s processing of nutrients and energy homeostasis [45]. Their aetiology is complex and involves interactions among genetics, epigenetics, adverse living environments and lifestyles, metabolic abnormalities, and microbial regulation [46]. Among these, metabolic syndrome (MS) is a syndrome of multiple metabolic abnormalities featured by abdominal obesity, insulin resistance, hypertension, and hyperlipidaemia and is a common pathological basis for diseases such as DM, MAFLD, and CVDs [47]. LD-mitochondrial dysfunction further triggers lipotoxicity, oxidative stress, insulin resistance, and mitochondrial dysfunction [48, 49]. During this process, the accumulation of FFAs plays a particularly critical role. Specifically, in obese states, insulin resistance induces β-cells to secrete more insulin, which further exacerbates insulin resistance and inhibits lipolysis, resulting in a malignant cycle. These mechanisms collectively increase the risk of CVDs, diabetes, and stroke through pathways involving lipid metabolism disorders, chronic inflammation, endothelial dysfunction, and oxidative stress [50]. Therefore, MS reflects the pathologic characteristics of multiple metabolic diseases that interact through overlapping pathways.

Diabetes

Current research on LDMC mainly focuses on type 2 diabetes mellitus (T2DM). T2DM is marked by inadequate insulin secretion or impaired insulin sensitivity in peripheral tissues, ultimately leading to insulin resistance and disturbances in glucose and lipid metabolism [51]. Among other things, abnormal LD‒mitochondria interactions are involved in various pathological processes, such as insulin resistance, β-cell apoptosis, oxidative stress, and inflammatory responses, which are highly important for elucidating the pathogenesis of T2DM (Fig. 2).

Fig. 2.

Comparison of lipid droplet–mitochondria interactions between healthy individuals and type 2 diabetes mellitus patients. This figure demonstrates that lipid droplet–mitochondrial contact under normal physiological conditions supports insulin sensitivity and energy homeostasis by dynamically regulating the balance between lipolysis and glucose metabolism. In the T2DM state, the abnormal enlargement of lipid droplets leads to triglyceride accumulation, the accumulation of lipotoxic metabolites caused by the blockage of free fatty acid transport to the mitochondria, the activation of inflammatory signalling pathways, mitochondrial dysfunction, and the increase in reactive oxygen species, which collectively prevent insulin signalling; moreover, the normal function of β-cells is disrupted, and apoptosis gradually occurs

In T2DM, dysregulation of LDMC has implications for both insulin resistance and pancreatic β-cell function. On the one hand, in insulin-sensitive tissues, FFAs produced by the lipolysis of excess lipid droplets cannot be oxidized efficiently via mitochondria and accumulate intracellularly to form toxic lipid metabolites, like diacylglycerol and ceramides, which activate signalling pathways such as protein kinase C (PKC) and inflammatory pathways, blocking insulin signalling and insulin resistance [52]. Lipid storage in skeletal muscle and mitochondrial function are key players in obesity-induced insulin resistance. Compared with skeletal muscle in healthy individuals, skeletal muscle in patients with T2DM is characterized by a more fragmented, dysfunctional mitochondrial network and larger LDs, resulting in an effective inability of mitochondria to access FFAs released from LDs, inducing lipotoxicity and inhibiting insulin signalling pathways [53]. Furthermore, mitochondrial dysfunction drives excessive production of reactive oxygen species (ROS), thereby compromising insulin sensitivity and aggravating inflammatory states [54]. Another study using quantitative transmission electron microscopy for single-fibre morphological analysis revealed that, in addition to the extremely large LDs in muscle fibres, these fibres have position-specific defects in subsarcolemmal mitochondria in patients with T2DM, indicating that T2DM exhibits marked cellular heterogeneity in intramuscular lipid storage [55]. Aberrant expression or modification of lipid droplet-associated proteins may affect the structural stability of LDs and FAs release [51, 56]. Exercise serves as a crucial therapeutic intervention for T2DM. It can enhance the contact interface between LDs and mitochondria, thereby promoting fat oxidation. Notably, this effect is not influenced by metabolic disease status and is observed in healthy individuals, obese individuals, and T2DM patients [57]. Despite exercise transiently increasing contact length, the LD-mitochondrial interaction mechanism in the T2DM basal state remains defective, undermining the long-term benefits of exercise as a long-term intervention for T2DM patients. However, some studies have shown that high-intensity interval training can ameliorate diabetic intramuscular metabolism through a coordinated remodelling of LDs storage and an increase in mitochondrial content. This shift toward a non-diabetic phenotype leads to a reduction in lipid accumulation, improving related complications [58, 59]. These observations may serve as a guide for health education on the right type of exercise for patients in clinical practice.

On the other hand, dysregulation of LDMC affects the normal function of pancreatic β-cells. In a normal physiological state, LDs and mitochondria closely collaborate through specific molecular mechanisms, and FAs in LDs can be transported to mitochondria in an orderly manner for β-oxidation to provide energy for cells. However, in T2DM patients, chronic overnutrition breaks this collaboration. When the storage capacity of LDs is overloaded, FFAs and lipid intermediates that fail to enter the mitochondria accumulate around the mitochondria, damaging mitochondrial membrane integrity and oxidative phosphorylation function [60]. A decrease in oxidative phosphorylation capacity leads to a reduction in ATP synthesis, and ATP deficiency directly affects proinsulin transcription and translation efficiency, thus inhibiting insulin synthesis [61, 62]. In this case, when the dynamic balance of LDs is disrupted, β-cells consequently exhibit a reduction in LD quantity and impaired morphological function [63]. In addition, abnormalities in the ER-LD-mitochondria signalling complex result in aggravated ER stress, oxidative stress, calcium homeostasis, and mitochondrial autophagy dysregulation in β-cells, ultimately triggering β-cell apoptosis and insulin secretion dysfunction [64].

Metabolic dysfunction-associated fatty liver disease (MAFLD)

MAFLD (formerly known as non-alcoholic fatty liver disease) is a chronic heterogeneous liver disease featured by the aberrant accumulation of lipids in hepatocytes and mitochondrial network imbalance. LD-mitochondria dysregulation promotes the progression from simple fatty liver disease to steatohepatitis through metabolic imbalance, oxidative stress, and fibrosis signal activation, eventually leading to liver fibrosis, cirrhosis, and hepatocellular carcinoma (HCC) [65, 66]. Notably, mitochondrial heterogeneity and dysfunction of key regulatory proteins in hepatocytes accelerate the stage-specific deterioration of the disease process directly by affecting lipid turnover and energy homeostasis (Fig. 3).

Fig. 3.

Comparison of lipid droplet–mitochondria interactions between healthy individuals and MAFLD patients. This figure shows that the normal liver facilitates fatty acid transport through lipid droplet–mitochondrial contact under physiological conditions, maintaining energy metabolism and homeostasis of LDs morphology and number. During the development of MAFLD, peridroplet mitochondria (PDM) and cytoplasmic mitochondria (CM) become dysfunctional and are unable to fully cope with changes in the organism, leading to a decrease in fatty acid β-oxidation, excessive accumulation of triglycerides (TAGs) and pathologic enlargement of the LD, which leads to structural disintegration of the LD, activation of the inflammatory pathway, activation of hepatic stellate cells (HSCs), mitochondrial dysfunction, and the ROS-mediated DNA damage cascade

Hepatocytes have two distinct mitochondrial subpopulations: peridroplet mitochondria (PDM) and cytoplasmic mitochondria (CM). In the early stages of MAFLD, in response to increased FFAs uptake, mitochondria are recruited around LDs to form PDM, and PDM supply ATP to promote TAG synthesis, driving LD expansion to counteract lipotoxic stress in hepatocytes. This process is an adaptive response of the body [67, 68]. As the disease progresses to an advanced stage, the amount of PDM decreases, and mitochondria are unable to overcome the additional FFAs. With continued external FAs “overload”, increased lipotoxicity causes mitochondrial dysfunction, hepatic damage, subsequent inflammation, and fibrosis [68]. PDM show higher bioenergetic activity and pyruvate oxidizing capacity in both healthy and diseased livers. Nonetheless, PDM are not capable of preventing advanced disease progression in MASLD. In contrast, CM demonstrate increased FAO capacity during MASLD progression, which increases with disease progression and exerts compensatory catabolic effects [69]. In addition to this, the more lipid droplet–mitochondrial contacts in the CM, the higher the ROS levels and the lower the mitochondrial membrane potential, whereas the PDM showed an opposite trend, which suggests that the PDM exhibit relatively delayed ATP production [70].

PLIN5 plays different roles in hepatic lipid metabolism depending on the cell type. In hepatocytes, PLIN5 expression enhances β-oxidation capacity by promoting dynamic contact between LDs and mitochondria, reducing lipid accumulation in liver cells [71]. PLIN5 also upregulates mitochondrial respiratory chain-related genes, stabilizes the efficiency of the electron transport chain to reduce ROS levels, and sequesters harmful mitochondrial proteins to LDs to alleviate cellular stress, decreasing the risk of liver damage and HCC [72]. Recent data suggest that sodium diethyldithiocarbamate (DDC) may amplify mitochondrial dysfunction in metabolic dysfunction-associated steatohepatitis (MASH) by contributing to the formation of PDM through the upregulation of PLIN5 [68]. In their quiescent state, HSCs contain abundant cytoplasmic LDs. Upon activation, these cells become the primary drivers of hepatic fibrosis. In high-fat diet-induced MAFLD, during the activation of HSCs driven by hepatic lipid overload and inflammation, the loss of cytoplasmic LDs is associated with a concomitant reduction in PLIN5 levels, demonstrating a negative correlation with HSC activation, further emphasizing the role of PLIN5 in MAFLD fibrosis [73]. The role of Rab GTPases in hepatic steatosis has been demonstrated. Rab18 regulates the expression of PLIN2 and peroxisome proliferator-activated receptor γ (PPARγ), exacerbating liver injury and lipid accumulation in patients with MAFLD [74]; Rab2A regulates the stability of hepatic PPARγ protein downstream of the AMPK-TBC1D1 axis, thereby promoting the occurrence and development of hepatic steatosis through the expression of PPARγ in response to energy/nutrient status [75].

The downregulation of Mfn2 is among the key factors driving the progression of MAFLD, with its mechanism of action originating from mitochondrial–ER contact/MAM dysfunction. As a major protein in the MAM complex, Mfn2 regulates the transfer of PS. Its expression deficiency directly impairs this process, leading to reduced synthesis of PE and phosphatidylcholine (PC). This process induces ER stress and perturbs hepatic lipid homeostasis, becoming an early driver of hepatic steatosis [76, 77]. Concurrently, phospholipid deficiency and loss of Mfn2 impair mitochondrial function, inhibiting FAO and further exacerbating lipid accumulation. As a compensatory mechanism, hepatocytes form numerous LDs to segregate excess lipids—an adaptive protective response. However, persistent mitochondrial dysfunction and LD overload ultimately disrupt normal LD‒mitochondrial interactions, causing malignant intracellular lipid accumulation, triggering lipotoxicity, oxidative stress, and cell death, thereby accelerating and maintaining the pathological state of MAFLD [67]. Furthermore, Mfn2 deficiency results in lipid accumulation within adipose tissue and inhibits mitochondrial oxidative phosphorylation. Surprisingly, however, studies involving tissue-specific knockout of Mfn2 in adipose tissue revealed that improved insulin sensitivity and liver protection in mice on a high-fat diet [30]. This paradox highlights the tissue-specific nature of Mfn2 function and the intricate compensatory mechanisms involved in systemic metabolic networks. Recent studies have found that different splice variants of Mfn2, ERMIT2 and ERMIN2, are specifically distributed in the ER but have different biological functions. ERMIN2 regulates endoplasmic reticulum morphology, whereas ERMIT2 positions itself at ER–mitochondria junctions to mediate inter-organelle connectivity through interactions with mitochondrial proteins. The splice variants of Mfn2 significantly increase the efficiency of calcium transport from the ER to the mitochondria and facilitate phospholipid translocation, expanding the biological functions of Mfn2, which is essential for interorganelle communication and energy metabolism [78].

Additionally, aflatoxin B1 (AFB1), a pervasive carcinogenic mycotoxin in food, drives hepatocarcinogenesis by stimulating pathological interactions between mitochondrial p53 and lipid droplet-associated PLIN2, and this aberrant crosstalk promotes dysfunctional LD–mitochondria contacts, triggering lipid accumulation in hepatocytes. This mechanism suggests that environmental toxins may promote the transformation of MAFLD into a malignant phenotype by interfering with lipid droplet–mitochondria interactions and metabolic processes [79]. As a protective measure against a spectrum of metabolic disorders, exercise plays a crucial role in MAFLD. Aerobic exercise reduces the number of LDs in contact with mitochondria in MAFLD liver and slows MAFLD progression by increasing the rate of ATP-synthesis-coupled respiration and FAO in the PDM, enhancing PLIN2-LIPA axis-mediated adipogenesis, and decreasing the size of hepatic LDs [80–82]. Therefore, future studies should focus on the multiple mechanisms through which LD-mitochondrial dysregulation induces the pathological process of MAFLD, develop corresponding molecularly targeted drugs, and incorporate lifestyle interventions to provide new directions for the treatment of MAFLD.

Cardiovascular diseases (CVDs)

The function of the heart as a high-energy-demanding organ is highly dependent on the oxidative phosphorylation capacity of mitochondria and efficient lipid metabolism collaboration between LDs and mitochondria. Ectopic lipid deposition and circulating lipid metabolism disorders are not only major pathological features of metabolic diseases, but also independent risk factors for cardiovascular diseases [83, 84]. Simultaneously, mitochondrial dysfunction can trigger corresponding pathological processes, including the induction of oxidative stress, the dysregulation of intracellular calcium cycling, activation of apoptotic pathways, and alterations in lipid metabolism, all of which drive the development and progression of CVDs [85].

PLIN5 is a regulator of myocardial lipid metabolism. It decreases FFA peroxidation by inhibiting the lipolysis of intracellular lipid droplets, consequently protecting the myocardium from lipotoxicity. PLIN5 has different functions in different cardiac disorders. In ischaemic cardiomyopathy, PLIN5 stabilizes LDs, promotes lipolysis, and maintains mitochondrial function, indirectly alleviating oxidative stress caused by mitochondrial oxidation [86]. Research has revealed that PLIN5 protects the heart by inhibiting PPARα, activating the PI3K/AKT signalling pathway, and improving mitochondrial energy metabolism [87]. Some microRNAs, such as miR-370 and miR-292-5p, protect the myocardium during myocardial ischaemia/reperfusion (I/R) injury by activating PLIN5 [88, 89]. When PLIN5 is absent, damage to mitochondrial structure occurs in the presence of I/R injury, which manifests as mitochondrial deformation, increased oedema, decreased matrix density, and blurring of mitochondrial cristae, as well as exacerbation of the damaging effects of ROS [90]. In addition, PLIN5 deficiency disrupts the energy metabolism homeostasis of cardiomyocytes, which exacerbates pressure overload-induced cardiac hypertrophy and failure through a mechanism driven by enhanced myocardial FAO and oxidative stress [84, 91]. In cardiomyocytes in the context of persistent FAO hyperactivity, PLIN5 deficiency reduces glucose uptake, causing insulin resistance. Moreover, insulin signalling is impaired by decreased GLUT4 expression and lower levels of AKT phosphorylation. This expression is particularly evident in the presence of free fatty acids. Notably, elevated NADH levels promote lactate production, further aggravating glucose metabolism disorders and inducing cardiac hypertrophy [84]. In another model of ethanol-induced cardiotoxicity (EIC), the downregulation of PLIN5 contributed to a reduction in LD-mitochondrial contact and the blockade of FAs flux in ethanol-treated cardiomyocytes; the opposite was true for PLIN5 overexpression, which may be an adaptive response of the body to protect against the effects of EIC [92]. Moreover, PLIN5 overexpression leads to myocardial steatosis, causing mild impairment of cardiac function and structural changes in the heart (although it does not significantly affect the body), accompanied by reduced expression of PPARα target genes and decreased mitochondrial function [93]. Although the function of PLIN5 differs among cardiac diseases, these findings collectively reveal the important role of PLIN5 in cardiac energy metabolism and the complexity of its regulation of cardiac metabolism.

Other proteins, like PLIN2, not only participate in lipolysis but also regulate lipophagy. When PLIN2 is absent, the lipophagy process is impaired, and TAG levels increase, causing lipid accumulation in cardiomyocytes. Under normal physiological conditions, heart function is not significantly affected, but under pathological conditions, heart function is severely impaired. These results are similar to the effects observed when PLIN5 is absent. PLIN2 deficiency does not affect mitochondrial function in cardiomyocytes, demonstrating that the effect of PLIN2 on mitochondrial function may be slight [94]. Atherosclerosis is closely linked to the progressive buildup of foam cells. PLIN2, a vital lipid protein of macrophages and vascular smooth muscle cell-derived foam cell lipid droplets, increases under the stimulation of oxidized low-density lipoprotein, resulting in the aberrant amplification of LDs and accelerating foam cell formation and plaque progression [90]. In addition to cardiomyocytes, the accumulation of LDs in endothelial cells increases the susceptibility to AS. Environmental pollution is another cause of cardiovascular damage [95]. Exposure to oil mist particulates induces alterations in cardiac enzymes and causes substantial accumulation of LDs within myocardial tissue, disrupting FAs metabolism in the myocardium. Furthermore, it targets and inhibits the PPARα signalling pathway, resulting in mitochondrial damage, ultimately precipitating myocardial cell apoptosis and myocardial tissue injury [96].

Obesity

With changes in lifestyles, the global obesity rate has risen significantly, becoming a global public health challenge and is linked to an increased risk of the incidence of various diseases [97]. Obesity is a chronic progressive disease, and its main pathological mechanism involves the abnormal accumulation of fat, which indirectly leads to secondary damage to multiple organ systems through systemic metabolic disorders [98].

When adipose overload occurs, LDs increase, TAG overaccumulates in LDs, and the entry of FFAs into mitochondria is blocked, which in turn triggers lipotoxicity and mitochondrial dysfunction; this metabolic abnormality ultimately causes multiorgan damage through oxidative stress, the activation of inflammation-related pathways, and an imbalance in energy metabolism [99]. The CIDE protein is a key protein that facilitates the fusion of LDs to form larger LDs and binds to lipid droplet-associated proteins (such as PLIN1and RAB8A) to regulate the stability of the fusion complex [100]. When CIDE proteins are deficient, LD fusion and growth are inhibited, and many small LDs are formed. These small LDs have a relatively larger total surface area, which not only increases the protein level and activity of ATGL but also promotes the release of FFAs. The released fatty acids combine with activated PPARα, upregulating the expression of mitochondria-related enzymes (e.g., CPT1A) and ultimately increasing energy expenditure [101, 102]. Notably, the regulation of the CIDE protein is tissue-specific. The downregulation or absence of its expression results in uncontrolled lipolysis and ectopic lipid deposition in white fat, whereas the overexpression of the CIDE protein in brown fat inhibits thermogenesis and exacerbates obesity [103]. Thus, lipid homeostasis may be restored by modulating antagonistic effects in white and brown adipose tissue. Furthermore, research has indicated that a long-term high-fat diet causes the ectopic deposition of lipids in non-adipose tissues (such as the liver, heart, and skeletal muscle), which is related to insulin resistance, the release of inflammatory mediators, energy metabolism disorders, and target organ damage. These metabolic pathways interact with each other, triggering a cascade of pathological metabolic responses in the body [104–106]. Under lipotoxic conditions, accumulated lipid derivatives inhibit the phosphorylation of insulin receptor substrates (IRSs), obstruct the activity of GLUT4, and induce the activation of inflammatory pathways such as the JNK signalling pathway, accelerating insulin resistance in skeletal muscle and liver [107, 108]. Modification of mitochondrial function using GSK3 inhibitors reduces LD size and TAG levels in hepatocytes from obese patients, demonstrating that LDs may be larger in hepatocytes from obese patients, reflecting the importance of the dynamic regulation of LDs [109]. Therefore, targeting tissue-specific and mitochondrial function repair may become a new method for treating obesity and related metabolic diseases.

Cancer

Both LDs and mitochondria are thought to induce the occurrence and development of tumors. In tumour cells, the distinguishing feature is that LDs are more numerous, larger, and more motile [110]. To support their rapid proliferation and the formation of endosomal organelles, cancer cells utilize LDs as a source of metabolic energy and essential membrane components [111]. Further findings have revealed that cancer cells deficient in LDs are more prone to apoptosis, and LDs also help to remove ROS from cells to alleviate cellular stress, suggesting that lipid droplets are critical for tumor cell growth [112]. Critically, the accumulation of LDs may affect both chemical-induced and immunological death pathways in tumor cells, thereby leading to chemotherapy resistance [113].

Many studies have shown the mechanism of LD occurrence in cancer and it is a potential biomarker of cancer. Fluorescent probe technology has been developed to dynamically monitor changes in LDs (especially viscosity and pH) to distinguish normal cells from cancer cells, providing strong evidence for early cancer diagnosis [114–116]. Moreover, the study of nanomaterials targeting LDs as drug carriers can be used for cancer treatment with a high safety profile, further emphasizing the importance of lipid droplets for clinical translation [117, 118]. Mitochondrial dysfunction is closely related to increases in the number and size of LDs, possibly due to excess liposomes [119]. Fluorescent probes based on the ability to target bilobal organelles can be used to successfully visualize the dynamics of lipid droplet–mitochondrial interactions [120].

Through combined analysis of multiple techniques, PDM were identified within lung adenocarcinoma (LUAD) cells; these mitochondria exhibit densely arranged orthodox type I cristae and are vertically aligned at the MLC to support increased oxidative phosphorylation (OXPHOS) activity [121]. On the basis of these findings, Ruijuan Cai et al. constructed a lipid droplet–mitochondrial risk scoring model using key identified lipid droplet-mitochondrial-related genes (LMRGs), this model showed good predictive performance for the prognosis of LUAD patients. In particular, drug sensitivity analyses have indicated that this risk prediction model can aid in predicting the efficacy of clinical immunotherapies, chemotherapies, and targeted therapies in patients with LUAD [122]. Phosphofructokinase-liver type (PFKL) is among the key rate-limiting enzymes in glycolysis. Under energetic stress, PFKL phosphorylates PLIN2 to enhance it binding with CPT1A, which facilitates LD and mitochondrial bolus tethering and enhances β-oxidation processes, thus supporting tumour cell survival and proliferation. Notably, the phosphorylation levels of PFKL and PLIN2 correlate with poor prognosis in HCC patients [23]. Although some studies on the biomedical technology through which LDMC regulates tumour metabolism have been conducted, a lack of direct evidence currently exists to elucidate the mechanism through which related proteins mediate tumour proliferation in vivo, invasion and apoptosis, which needs to be further explored in vivo experiments, to provide a clinical translational basis for the mechanistic explanation of LDMC and targeted interventional therapies.

Obstructive sleep apnoea syndrome (OSAS)

OSAS is a respiratory disease characterized by chronic intermittent hypoxia, which is closely related to metabolic disorders, and its pathological features may be potentially related to lipid droplet–mitochondrial imbalance [123]. A prospective study reported a positive correlation between the quantity of LDs in pharyngeal constrictor muscles and the apnoea hypopnea index (AHI) in patients with OSAS and found myogenic fibre lesions, mitochondrial oedema, and intracellular LD accumulation in myocytes, indicating that abnormalities in lipid metabolism may cause airway collapse by altering muscle structure [124]. Chronic hypoxia causes LD accumulation through activation of the JNK/SREBP/ACC signalling pathway and induces mitochondrial oxidative stress, resulting in a vicious cycle of LD accumulation and mitochondrial damage [125, 126]. OSAS can induce a systemic inflammatory response that interferes with the expression of lipid metabolism genes and inhibits lipolytic enzyme activity, leading to aggravated LD accumulation [127]. As research has been limited to single organelles, there is no direct evidence that LDMC causes OSAS, but the hypoxia–lipid droplet accumulation–mitochondrial damage cascade induced by metabolic disorders in OSAS indirectly suggests that the lipid droplet–mitochondrial axis may be a potential regulatory node.

Discussion

LDMC is a key link in lipid metabolism, but the specific protein complexes and regulatory pathways involved remain unknown. Previous studies have suggested that metabolic channeling is a possible mechanism for lipid transport from LDs to mitochondria. This model posits that following the hydrolysis of TAGs at the LDs surface, the produced FAs can be directly and efficiently transferred to mitochondria via MCSs. This facilitates their activation, transport, and β-oxidation, resulting in the formation of a spatially highly organized metabolic pathway [4, 128]. Notably, although membrane fusion refers to FA exchange between LDs-LDs or mitochondria-mitochondria, the MCS examined in this study provides a pathway that does not depend on membrane fusion [129]. This MCSs-based mechanism prevents the accumulation of FFAs in the cytoplasm and the resulting lipotoxicity, while also helping maintain energy and redox homeostasis. The LD–mitochondria contact-associated molecules proposed in this study, such as PLIN5, FATP4, and Mfn2, may mediate membrane contact structures that provide the physical basis for metabolic channels. Moreover, SNARE complexes or lipid transporters (e.g., the VPS13 family) may participate in regulating the formation and function of these channels. Therefore, LD-mitochondrial MCSs represent the key structural foundation and physical platform for channeling lipid utilization metabolism. However, this metabolic pathway is not confined to LDs and mitochondria but involves other organelles such as the ER, lysosomes, and peroxisomes [130]. The development of novel molecular tools for real-time monitoring of FAs transport at MCSs will play a critical role in further understanding the mechanism of lipid transport in metabolic pathway formation.

Although multiple proteins involved in this process have been identified, the specific role of certain proteins remains unclear. For example, it is known that VPS13A is located at the MCS between the ER and mitochondria/LD and VPS13C is located at the MCS between the ER and late endosomes/lysosomes/LD, but whether they are involved in the direct lipid droplet–mitochondrial MCSs and the underlying molecular mechanisms are unknown [131]. Similarly, whereas it has been shown that SNAP23 may not be an essential factor in FAO and its knockdown in hepatocytes does not disrupt lipid droplet–mitochondrial contact, its exact function and the possible existence of compensatory roles for other SNAP isoforms (e.g., SNAP25) need to be further verified by a complete knockout model [16].

Organizational variability in protein function is also one of the focuses of current research. The most typical example is PLIN5, which primarily stems from differing metabolic demands in various tissues and distinct molecular interaction networks involving PLIN5. Similar observations have been made for Mfn2. PLIN5 has been identified as a double-edged sword; it mitigates myocardial injury and oxidative stress following ischaemia reperfusion by coordinating LD and mitochondrial function, yet its overexpression causes myocardial steatosis and cardiac dysfunction, suggesting that its protective effect may occur within an optimal range [90, 93]. Moreover, the function of PLIN5 is dynamically regulated by AMPK/PKA and other signalling pathways and phosphorylation modifications, providing a theoretical basis for the precise regulation of PLIN5 as a therapeutic target. Therefore, therapeutic strategies targeting PLIN5 must focus on precisely regulating its activity to restore it to an optimal balance point under specific pathological conditions. Utilizing the tissue specificity of key proteins to develop novel targeted drugs will be beneficial for reshaping lipid metabolism homeostasis and providing more precise treatments for metabolic diseases.

In addition to the above protein-mediated mechanisms, genetic factors play important roles in regulating lipid biology and certain metabolic diseases. Genome-wide association studies (GWASs) have indicated that mutations at multiple genetic loci (such as PNPLA3, TM6SF2, GCKR, and MBOAT7) significantly increase the risk of MAFLD. In the liver, the gene products of these loci perform distinct functions: PNPLA3 participates in LD remodelling and hydrolysis; TM6SF2 contributes to the assembly and secretion of very low-density lipoproteins (VLDLs); and GCKR and MBOAT7 influence TAG synthesis by regulating de novo lipogenesis (DNL) and phospholipid remodelling, respectively [132–134]. Similarly, these gene products also drive the progression of MAFLD through distinct molecular mechanisms, but their common core outcome is the disruption of hepatic lipid homeostasis, leading to excessive lipid accumulation within the liver [135]. However, these mutations have varying effects on systemic lipid metabolism. Among them, PNPLA3 and TM6SF2 mutations reduce circulating lipid levels, thereby protecting the heart, whereas GCKR mutations increase TAG levels and cause atherogenic dyslipidaemia [136]. PNPLA3 I148M mutation confers the greatest risk among known genetic factors for MAFLD. Its mutant protein abnormally accumulates on LD surfaces. It loses its own lipase activity and predominantly inhibits the activity of other lipases (such as ATGL), thereby suppressing TAG breakdown and LDs remodelling. This process triggers a cascade of downstream events, including lipotoxicity, ER stress, and inflammation [137]. Additionally, the pathological mechanisms by which MBOAT7 rs641738 and TM6SF2 E167K deletions induce MAFLD deserve attention. Studies have indicated that MBOAT7 deficiency primarily induces large-bubble steatosis with relatively mild mitochondrial damage but significant endoplasmic reticulum stress. Conversely, TM6SF2 deficiency specifically leads to microvesicular steatosis accompanied by more severe mitochondrial morphological abnormalities, increased ROS production, and suppressed oxidative phosphorylation. These mechanisms collectively drive the progression of MAFLD towards MASH and liver fibrosis [138]. Other genetic targets also hold promise for developing new therapeutic and preventive strategies against metabolic disorders. Although variants in the CREBRF gene significantly influence obesity and diabetes risk, they exhibit marked racial differences, suggesting that population genetic background must be considered in clinical applications [139]. GPR151 also represents a potential therapeutic target for obesity and T2DM, yet its precise mechanisms influencing disease risk remain to be elucidated. Other targets, such as PCSK9 and ANGPTL3, may offer effective strategies for cardiovascular disease prevention and treatment, highlighting the feasibility of translating genetic mechanisms into clinical therapies. However, the development of corresponding inhibitors remains in the exploratory phase [140]. Therefore, constructing polygenic risk scores (PRSs) by integrating these genetic mutations with diverse effects will facilitate a more accurate assessment of metabolic disease progression risk and clinical prognosis. Concurrently, it is imperative to delve deeper into the mechanisms by which genetic factors molecularly drive metabolic diseases, thereby establishing a theoretical foundation for advancing clinical translation.

On the other hand, the exploration of LDMC has remained confined to functional analyses at the level of single proteins or reduced-scale complexes; however, the molecular modes through which different proteins (such as transmembrane bridging, membrane curvature modulation, or enzymatic channeling) achieve the specific transfer of FAs are still major challenges that need to be addressed. Since LDs originate from the ER, the regulatory role of the ER in lipid droplet–mitochondrial interactions highlights the importance of an in-depth study of ER–LD–mitochondrial ternary interactions. Moreover, the biological techniques for targeting LDs or mitochondria have been intensively investigated, but research on targeting dual organelles is still insufficient because of the immaturity of the LDMC mechanism and the limitations of the technology. This “point-and-line” approach has led to two core gaps: first, the spatial and temporal coordination of multiprotein components in the LD-mitochondrial MCSs has not yet been established; second, the dynamic regulatory network governing transmembrane transport remains largely unclear, with no overarching systematic dissection yet available. Future research should focus on a paradigm shift from point-and-line to surface, systematically exploring the ER–LD–mitochondrial network to form an integrated understanding of MCS mechanisms. On this basis, precise intervention strategies targeting MLC should be developed to provide novel approaches and technologies for diagnosing and treating related metabolic diseases.

Strengths and limitations

This review analysed the lipid transport mechanism from a novel perspective, systematically distinguished two modes of LDMC, dynamic contact and stable anchoring, and emphasized the critical role of ER‒LD–mitochondrial ternary interactions in lipid metabolism. This review also provided insights into the tissue-specific functions of key proteins at the MCS and their association with metabolic diseases, suggesting potential therapeutic strategies for targeting MCS-associated proteins and evaluating the clinical translational potential of related genetic factors. Finally, future research directions were identified, and a new theoretical framework for understanding metabolic disease mechanisms and developing clinical intervention strategies was provided. However, some of the mechanisms discussed in this review are still not supported by in vivo experimental evidence due to the limitations of existing research and technology, which limit the in-depth exploration of related issues.

Conclusions

This review systematically elucidated that the interaction between LDs and mitochondria is achieved through the close apposition of their membrane structures at contact sites (rather than membrane fusion). Their pivotal role in maintaining cellular metabolic homeostasis and their critical significance in the pathogenesis and progression of various metabolic diseases were discussed, emphasizing the importance of LD-mitochondrial membrane contact sites in lipid metabolism. On the basis of the lipid droplet–mitochondrial interaction mechanism, tissue-specific targeted drugs can be developed against key contact site proteins (e.g., PLIN5 and Mfn2) to block lipotoxicity accumulation from the pathophysiological link by regulating lipid transport and metabolism; the relevant genetic molecules have the potential to be used as biomarkers for predicting disease progression or treatment response, especially MAFLD, which can facilitate the early identification of high-risk patients and provide a basis for the development of individualized interventions to improve the long-term prognosis of patients. In addressing the worldwide health challenge of prevalent metabolic disorders, this study offers insights and innovative approaches grounded in the principles of metabolic homeostasis regulation.

Abbreviations

LDs

Lipid droplets

MCSs

Membrane contact sites

FAs

Fatty acids

LDMC

Lipid droplet-mitochondria crosstalk

MLC

Lipid droplet-mitochondria contact

TAGs

Triglycerides

SNAP23

Synaptosome-Associated protein 23

SNARE

Soluble N ethylmaleimide sensitive factor attachment protein receptor

STX18

Syntaxin 18

GLUT4

Glucose transporters type 4

PLIN2

Perilipin 2

VAMP4

Vesicle-associated membrane protein 4

ACSL1

Long-chain acyl-CoA synthetase 1

FAO

Fatty acid oxidation

TBK1

Tank-binding kinase 1

AMPK

AMP-activated protein kinase

PLIN5

Perilipin 5

BAT

Brown adipose tissue

ATGL

Adipose triglyceride lipase

ABDH5

α-β-hydrolase domain-containing 5

PKA

Protein kinase A

PPARα

Peroxisome proliferator-activated receptor alpha

CPT1A

Carnitine palmitoyltransferase 1 A

JMJD8

Jumonji C domain-containing protein 8

ER

Endoplasmic reticulum

HSCs

Hepatic stellate cells

Mfn2

Mitofusin2

Hsc70

Heat shock cognate protein 70

PLIN1

Perilipin 1

VPS13D

Vacuolar protein sorting-associated protein 13D

ESCRT

The endosomal sorting complex required for transport

TSG101

Tumour susceptibility gene 101

FFAs

Free fatty acids

MIGA2

Mitoguardin-2

VAPB

VAMP associated protein B and C

CIDE

The cell death-inducing DFFA-like effector

mISR

Mitochondrial integrated stress response

ESYT

Extended synaptotagmin

PE

Phosphatidylethanolamine

PS

Phosphatidylserine

MAM

Mitochondria-associated membrane

DM

Diabetes mellitus

MAFLD

Metabolic dysfunction-associated fatty liver disease

CVDs

Cardiovascular diseases

MS

Metabolic syndrome

T2DM

Type 2 diabetes mellitus

PKC

Protein kinase C

ROS

Reactive oxygen species

HCC

Hepatocellular carcinoma

PDM

Peridroplet mitochondria

CM

Cytoplasmic mitochondria

DDC

Diethyldithiocarbamate

MASH

Metabolic dysfunction-associated steatohepatitis

PPARγ

Peroxisome proliferator-activated receptor γ

PC

Phosphatidylcholine

AFB1

Aflatoxin B1

I/R

Ischaemia/reperfusion

EIC

Ethanol-induced cardiotoxicity

IRSs

Insulin receptor substrates

LUAD

Lung sdenocarcinoma

OXPHOS

Oxidative phosphorylation

LMRGs

Lipid-droplet-mitochondrial-related genes

PFKL

Phosphofructokinase-liver type

OSAS

Obstructive sleep apnea syndrome

AHI

Apnea hypopnea index

GWASs

Genome-wide association studies

VLDLs

Very low-density lipoproteins

DNL

De novo lipogenesis

PRSs

Polygenic risk scores

Authors’ contributions

G.W. and B.S: conceptualization, original drafting, visualization, and writing - review & editing. H.L.: visualization and editing. M.H. and H.X.: review and editing. H.L. and X.W.: review and visualization, and suggestions. M.T.: supervision, funding acquisition, project administration, review & editing. All authors read and approved the final manuscript.

Funding

This work was supported by the General Program of Natural Science Foundation of Jiangxi province (No. 20224BAB206110 to Mingfu Tong), Science and technology program of traditional Chinese Medicine of Jiangxi Provincial (No. 2023A0338 to Mingfu Tong), and the College Students’ innovation and entrepreneurship training program (No. 202311843027 to Jianmin Ye and Mingfu Tong).

Data availability

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.