Dysfunction of caveolae in muscle disease: A cell biological perspective

Caveolae are highly abundant membrane microdomains on the surface of mammalian skeletal muscle fibers (Fig. 1). Early EM revealed the dense packing of caveolae on the surface of these fibers as well as their association with the neck of transverse (T-) tubules, the extensive surface-connected network of tubules that allows an action potential to reach the interior of the muscle (Dulhunty and Franzini-Armstrong, 1975; Franzini-Armstrong et al., 1975). In isolated adult skeletal muscle fibers, caveolae can occupy up to 50% of the sarcolemmal area (Lo et al., 2015), and even an immature binucleate myotube in a developing zebrafish embryo was estimated to have over 25,000 caveolae (Hall et al., 2023).

The density of caveolae on the surface of muscle fibers suggested a structural role for caveolae in the sarcolemma, a specialized plasma membrane that must cope with constant stress during the repeated cycles of contraction and relaxation throughout the lifetime of the muscle. Early EM studies showed flattening of caveolae in response to membrane stretch, supporting a physical role in preventing membrane damage (Dulhunty and Franzini-Armstrong, 1975). The subsequent discovery of caveolin-3 (CAV3), the major structural protein of muscle caveolae (McNally et al., 1998; Minetti et al., 1998; Way and Parton, 1995) and then other structural components of muscle caveolae ushered in a new molecular era of research into caveolae and the discovery of human diseases linked to muscle caveolar dysfunction (McNally et al., 1998; Minetti et al., 1998).

Caveolae in muscle and non-muscle cells are generated by the combined action of the integral membrane proteins, caveolins, working together with peripheral membrane proteins such as cavins and syndapin/pacsins (Fig. 1 C). In mammalian striated muscle (skeletal and cardiac), caveolae are generated by CAV3, a muscle-specific caveolin, and cavin1, which is ubiquitously expressed. We will focus on aspects specific to skeletal muscle caveolae here. More detailed descriptions of caveolae formation (Parton and Ariotti, 2026) and the involvement of CAV3 in vascular biology and cardiac function have been reviewed elsewhere (An et al., 2024; Zhang et al., 2026). Downregulation or deletion of either CAV3 or cavin1 leads to a loss of morphological sarcolemmal caveolae in cells and in vivo (Minetti et al., 2002; Liu et al., 2008; Liu and Pilch, 2008; Lo et al., 2015). While this has led to a model in which these two components are not only required but sufficient for caveolae formation, elegant studies in mice have demonstrated that syndapin III/PACSIN3 knockout led to a loss of morphological caveolae, but CAV3 and cavin1 were shown to remain at the sarcolemma (Seemann et al., 2017). This intriguing result suggests a requirement for CAV3, cavin1, and syndapin III/PACSIN3 to generate muscle caveolae. CAV3 and cavin1 can both induce membrane curvature in model systems (Kovtun et al., 2014; Walser et al., 2012), but these studies show that in muscle this is insufficient to generate caveolae. In addition to these structural components, a muscle-specific caveolar coat protein, cavin4/MURC, associates with the caveolar domain (Bastiani et al., 2009; Tagawa et al., 2008). Other accessory proteins are less well characterized in muscle but include EHD2 and dynamin2, which can both stabilize caveolae at the plasma membrane and prevent their endocytosis (Moren et al., 2012; Parton et al., 2024; Stoeber et al., 2012).

CAV3, like caveolin-1 (CAV1), is synthesized co-translationally in the endoplasmic reticulum, where it is inserted into the cytoplasmic leaflet of the membrane (Monier et al., 1995; Morales-Paytuvi et al., 2023). It then oligomerizes within the Golgi complex, predicted to form an oligomeric disc ∼14 nm in diameter analogous to the structure of CAV1 determined by cryo-EM (Morales-Paytuvi et al., 2023; Porta et al., 2022) (Fig. 2). Upon arriving at the plasma membrane, the cavin complex is recruited (containing cavin1 and cavin4 in muscle) and is presumably further stabilized by syndapin III/PACSIN3 (Bastiani et al., 2009; Seemann et al., 2017; Tillu et al., 2021). Integral to this process is association with membrane lipids, particularly cholesterol, which is crucial for caveolae formation (Kenworthy et al., 2023). Both CAV1 and cavin1, when expressed in a model cell system, can recruit specific membrane lipids (Zhou et al., 2021) to help facilitate the formation of a metastable domain. How syndapin III fits into this scheme has not been fully revealed. It is unclear if syndapin III functions in a direct structural role through protein–protein interactions, through indirect means by helping to generate the lipid domain required for stable CAV3/cavin1 association, or aids in stabilizing the protein-lipid assembly required for sarcolemmal caveolae formation.

The discovery of CAV3 was soon followed by the identification of numerous pathogenic variants in CAV3 in human patients with a broad spectrum of skeletal muscle diseases (summarized in Fig. 2). Originally classified as limb-girdle muscular dystrophy (LGMD) type 1C, CAV3-related muscle disease has now been excluded from the current LGMD list, given the main clinical features of disease associated with CAV3 variants are rippling muscle disease (RMD), characterized by rippling wave-like electrically silent muscle contractions, and myalgia (Engelman et al., 1998; McNally et al., 1998; Minetti et al., 1998; Galbiati et al., 1999; Galbiati et al., 2000b; Vorgerd et al., 2001; Woodman et al., 2004; Dotti et al., 2006; Parker et al., 2007; Gazzerro et al., 2011; Scalco et al., 2016; Straub et al., 2018; Dewulf et al., 2019; Berling et al., 2023; Shah et al., 2023). Other clinical presentations include proximal and distal myopathies, persistently high serum creatine kinase (CK), muscle hypertrophy, myalgia, and exercise intolerance. Patients often present with overlapping symptoms, and the same genetic variants can be associated with different phenotypes. Pathogenic variants in Cavin1 have also been reported (Dwianingsih et al., 2010; Hayashi et al., 2009; Rajab et al., 2010). Consistent with its more ubiquitous expression pattern, Cavin1 patients are characterized by congenital generalized lipodystrophy type 4, cardiac involvement, and myopathic symptoms including RMD, akin to those observed with CAV3 variants. Intriguingly, there is also a form of immune-mediated RMD (iRMD) characterized by a mosaic pattern of CAV3 expression in muscle (Lo et al., 2011; Schulte-Mattler et al., 2005). More recent studies have identified serum cavin4 autoantibodies as a potential biomarker for iRMD patients (Dubey et al., 2022; Svahn et al., 2023).

The first described missense variant in CAV3, a leucine substitution for a conserved proline at position 104 (CAV3-P104L), was shown to disrupt caveola formation in cells and in the skeletal muscle of patients. CAV3-P104L accumulates in the Golgi complex, is rapidly degraded, and can act as a dominant-negative mutant by perturbing WT CAV3 (Galbiati et al., 1999; Minetti et al., 1998). Cultured cells lacking caveolins can be used to study the difference between WT and pathogenic CAV3 variants (see Table 1). Expression of WT CAV3 (or non-muscle CAV1) induces caveola formation (assayed by EM) and recruitment of endogenous cavin1 to the plasma membrane caveolae by fluorescence microscopy (Hill et al., 2008; Kirkham et al., 2008). In contrast, pathogenic variants such as CAV3-P104L and T78K do not induce caveolar formation or cavin1 recruitment (Benzoni et al., 2024; Hill et al., 2008; Kirkham et al., 2008). However, other characterized variants in CAV3 associated with disease do not cause an absence of sarcolemmal caveolae in human muscle; the effects of these changes are not fully understood. These variants (such as C71W and T78M) generally behave like WT proteins in these assays, resulting in caveolae formation and normal cavin1 recruitment (Benzoni et al., 2024; Hill et al., 2008; Kirkham et al., 2008). We propose a simple classification: Type I variants, which are associated with loss of caveolae in tissues and in cellular assays, and type II variants in which normal caveolae are observed. These features, along with a schematic of representative CAV3 type I and II variants, have been summarized in Fig. 2. Note that this is a simple cell biological classification and is not related to pathological classification of variants for which standard criteria are available (Richards et al., 2015).

Several CAV1 type I variants have been mapped to the 3D structure of caveolin generated by cryo-EM and modeled using AlphaFold. These studies have provided insights into the destabilizing effect of the type I variants that inhibit oligomer formation and onward trafficking from the Golgi complex (Han et al., 2023). Mutation of the Cav1-P132L site (P104L in CAV3) is predicted to disrupt the lateral packing of adjacent protomers within the highly conserved α3-helix of the intramembrane domain of the complex (Han et al., 2023). These variants not only lead to loss of the mutant protein but can also cause degradation of co-expressed WT CAV3. Perturbation of caveolin trafficking also triggers ER stress (Morales-Paytuvi et al., 2023), raising the possibility that the harmful effects of CAV3 type I mutants might not only result from loss of surface caveolae but also from the accumulation of potentially toxic caveolin proteins within the early secretory pathway.

While pathogenic variants in CAV3 and perturbation of caveolae are linked to a number of clinical conditions, caveolae are also linked to muscle disease conditions with different underlying causes. For example, caveolae and CAV3 are increased in Duchenne muscular dystrophy (DMD) (Merrick et al., 2009; Repetto et al., 1999). The overexpression of CAV3 similarly leads to a DMD-like phenotype in mice, raising the possibility that the increased CAV3/caveolae levels contribute to the disease etiology in DMD patients (Galbiati et al., 2000a). However, later studies suggest this appears to be a compensatory mechanism, as loss of dystrophin and partial loss of CAV3 exacerbate the phenotypes in mouse models (Merrick et al., 2009). This study showed that effects on early muscle development are a key feature of loss of both CAV3 and dystrophin, with perturbation of fiber type specification and emergent stem cell function (Merrick et al., 2009). This emphasizes the need to consider not only the clinical features of caveolinopathies as symptoms appear, often late in life, but also the underlying consequences of the loss of the protein during development.

The density of caveolae in muscle suggests that they represent a significant membrane reservoir that could protect the muscle upon stretch (Fig. 3) (Lamaze et al., 2026). Pioneering studies showing that stretch of isolated muscle fibers flattens caveolae (Dulhunty and Franzini-Armstrong, 1975) have been supported by more recent studies using muscle and non-muscle cells with genetic manipulations to perturb caveolae (Dewulf et al., 2019; Lo et al., 2015; Sinha et al., 2011). Genetic loss of caveolae or expression of type I CAV3 variants results in damage to the sarcolemma, as shown in cell models, muscle explants, and zebrafish models in response to increased muscle membrane tension or high muscle activity (Dewulf et al., 2019; Hall et al., 2023; Lo et al., 2015) (Table 1). The elevated CK levels observed in mice lacking cavin1, as well as upregulation of proteins involved in sarcolemmal stability, including dystrophin and dystroglycans, and muscle repair proteins, such as dysferlin, are all consistent with impaired sarcolemmal integrity caused by loss of caveolae (Ding et al., 2017). CAV3 has been linked to the dystrophin–glycoprotein complex (DGC), which is crucial for maintaining sarcolemmal integrity, although it is not considered a fundamental component of the DGC itself (Crosbie et al., 1998; Song et al., 1996). Rather, the association between CAV3 and these components may reflect the close spatial and functional interplay between caveolae, the cortical actin cytoskeleton, and the DGC at the sarcolemma, highlighting their importance in mechanotransduction and in maintaining membrane integrity at the sarcolemma (Wilson et al., 2022).

Direct evidence for caveolar flattening in vivo was provided by whole-embryo EM imaging of zebrafish embryos trapped in a curved state. The stretched fibers on the convex surface of the curved fish showed a loss of caveolae consistent with the amount of membrane required to expand the sarcolemma (Hall et al., 2023). This is in agreement with a model in which caveolae provide a reservoir of membrane to prevent membrane damage in response to extreme muscle shape changes. As caveolae are also localized to the neck of T-tubules (Franzini-Armstrong et al., 1975; Lo et al., 2015), an interesting possibility is that these caveolae can also reduce damage to the T-tubules during muscle shape changes in a similar fashion.

Sarcolemmal caveolae of isolated adult mouse muscle fibers show a striking arrangement as rosettes with multiple caveolae surrounding a single narrow neck that connects to the sarcolemma (Lo et al., 2015). These structures are preferentially disassembled upon increasing membrane tension. The formation of these complex superstructures and their dependence on membrane tension for formation and regulated disassembly has been modeled (Golani et al., 2019).

These observations raise the possibility that a loss of mechanical stability underlies the diseases associated with muscle dysfunction. This appears to be a reasonable hypothesis in view of the results in these model systems, the analysis of mouse models of caveola loss, and some of the clinical features of caveolinopathies such as sustained high CK levels and rhabdomyolysis (Berling et al., 2023; Scalco et al., 2016), in association with the presence of atrophic muscle fibers (Berling et al., 2023). However, like many muscle disease conditions, the clinical symptoms associated with caveolinopathies are usually late-onset, relatively mild, and disease severity is also highly variable, even between individuals of the same family carrying the same mutation. This suggests that even with type I variants, associated with a loss of muscle caveolae, there are considerable compensatory mechanisms, or other features of the environment, to allow muscle fibers to survive despite loss of such an abundant surface feature. Consistent with this, cavin1-null mouse skeletal muscle was shown to have increased deposition of ECM components that could potentially protect a more fragile sarcolemma (Ding et al., 2017). In a non-muscle tissue from CAV1-null mice, loss of caveolae was shown to be accompanied by increased ECM deposition in the adipose tissue (Ding et al., 2017; Martin et al., 2012). This was shown to help compensate for the more fragile plasma membrane that was revealed upon cell isolation, accompanied by ECM removal (Martin et al., 2012). Taken together, these findings, from whole animals and in model cellular or explant systems, are consistent with the importance of caveolae in maintaining the integrity of the sarcolemma by buffering against physical stresses.

Examination of mouse muscle fibers lacking sarcolemmal caveolae through genetic loss of cavin1 reveals dramatic large-scale perturbation of the sarcolemma (Lo et al., 2015). In addition to loss of caveolae, the entire sarcolemma of isolated fibers shows an irregular, ruffled appearance and highly increased endocytosis (Chaudhary et al., 2014; Lo et al., 2015). This suggests an inhibitory effect of caveolae on endocytosis and is consistent with the observation that loss of caveolae (CAV1 or cavin1) is associated with increased clathrin-independent endocytosis in non-muscle cells (Chaudhary et al., 2014). It is again worth noting that the morphological changes observed in the isolated adult mouse muscle fibers are less obvious in intact tissue, possibly indicative of compensation in the complex in vivo environment as outlined above. Loss or mutation of CAV3 has also been linked to loss of specific proteins from the sarcolemma, such as dysferlin (Cai et al., 2009; Matsuda et al., 2001), a membrane repair protein localized to the sarcolemma and T-tubules where it is implicated in T-tubule biogenesis and function (Ampong et al., 2005; Bansal and Campbell, 2004; Kerr et al., 2013; Klinge et al., 2010). In the absence of caveolin, dysferlin was shown to be rapidly internalized by a clathrin-independent endocytic pathway in a model fibroblast system (Hernandez-Deviez et al., 2006; Hernandez-Deviez et al., 2008). This could be rescued by WT caveolin (CAV1 or CAV3) but not variants in the conserved scaffolding domain of CAV3 (G55S and C71W). Despite the differences from WT CAV3, in these cell biological assays, both these variants that were initially linked to muscular dystrophies were subsequently detected in the general population and have been suggested to be polymorphisms rather than disease-inducing type II variants (de Paula et al., 2001).

Loss of skeletal muscle CAV3 has also been linked to dysregulation of other membrane processes. Genetic loss of CAV3 in a mouse model caused the development of insulin resistance, with decreased insulin-stimulated glucose uptake and serum lipid dysregulation (increases in resting triglyceride levels and smaller increases in free fatty acids and cholesterol (Oshikawa et al., 2004). The upstream mechanisms underlying this dysregulation are not yet clear, but downstream signaling pathways were shown to be attenuated.

The effects of caveolar perturbations on the membrane lipid composition of the sarcolemma are also likely to be very significant, although most of the work in this area comes from studies of non-muscle cells. Loss of caveolae (through loss of CAV1 or cavin1) or expression of a CAV3 variant, CAV3-C71W, affects specific Ras isoforms that are dependent on specific lipids for their dynamic compartmentalization on the inner leaflet of the plasma membrane (Ariotti et al., 2014; Carozzi et al., 2002). Membrane lipid nanoscale regulation was shown to be coupled to dynamic disassembly of caveolae in response to increased membrane tension (Ariotti et al., 2014), suggesting that the mechanical response of the caveolar system is not restricted to physical membrane buffering.

An additional effect of the loss of caveolae might be related to their concentration of cholesterol. A number of studies have emphasized the importance of cholesterol in forming caveolae and suggested that high levels of cholesterol exist within the caveolar domain (Doktorova et al., 2025; Hailstones et al., 1998; Kenworthy et al., 2023; Ortegren et al., 2004; Zhou et al., 2021). In a cell that has lost caveolae, this “sink” would be lost, and an imbalance in cholesterol could arise. One interesting consequence of this imbalance is inhibition of mTORC1 signaling in cells and tissues in which caveolae have been perturbed due to cholesterol accumulation in the lysosomes; myotubes lacking CAV3 showed increased lysosomal cholesterol and attenuation of signaling (Shah et al., 2023). As mTORC1 is the master regulator of the cellular response to feeding/starvation, lack of caveolae leading to lysosomal accumulation of cholesterol would have severe consequences for muscle function, exercise-induced hypertrophy, and muscle growth. Somewhat enigmatically, genetic loss of cavin1 (which also caused loss of CAV3 protein in the skeletal muscle fibers) was shown to be associated with increased muscle mass and hyperactivation of mTORC1 downstream of Akt (Ding et al., 2017).

While caveolae can play a physical mechanoprotective role by providing a membrane reservoir, caveolae disassembly upon mechanical perturbation can also lead to the release of cavins and other caveolar proteins into the cytosol to mediate other effects (Fig. 3). For example, in non-muscle cells, EHD2 redistributes to the nucleus upon mechanical stimulation to regulate transcriptional pathways (Torrino et al., 2018) and cavin1 and cavin3 can interact with cytoplasmic targets (McMahon et al., 2019; McMahon et al., 2021; Wu et al., 2023). Cavin4 is localized within the nucleus of myofibers lacking cavin1 and caveolae (Lo et al., 2015), suggesting that release from caveolae in response to physiological stimuli could also trigger a similar pathway and interactions with downstream components. Downregulation of cavin4/MURC was shown to impair differentiation of C2C12 myoblasts, impairing expression of myogenin and reducing activation of ERK (Tagawa et al., 2008).

CAV3 has also been linked to the regulation of other pathways. Perturbed membrane buffering in muscle fibers with decreased levels of CAV3 led to dysregulation of the IL6/STAT3 signaling pathway (Dewulf et al., 2019). In this system, myotubes from patients with type I CAV3 variants (P28L and R26Q) showed constitutive hyperactivation of the IL6/STAT3 pathway and increased expression of downstream targets. Loss of CAV3 through expression of CAV3 dominant-acting mutants has also been shown to cause muscle atrophy in mouse models with CAV3 mutants, causing increased activation of myostatin target genes (Ohsawa et al., 2006). Myostatin is a negative regulator of skeletal muscle growth and differentiation, suggesting a model in which CAV3 acts to reduce muscular atrophy and increase cell volume. However, in a different model of caveola deficiency, mice with genetic loss of cavin1 showed loss of myostatin, increased muscle mass, and hypertrophy (Ding et al., 2017).

Pioneering EM showed that the precursors of T-tubules in skeletal muscle appear to be formed from chains of interconnected budded structures, with the morphology of caveolae (Ishikawa, 1968). The later demonstration that the newly discovered muscle-specific caveolin isoform, CAV3, was associated with these early surface-connected structures (Parton et al., 1997), and that perturbation of cholesterol affected both caveolae and T-tubules (Carozzi et al., 2000) strengthened the idea that T-tubules could be formed through the repeated formation of chains of caveolae without budding. More recently, an elegant combination of high-resolution EM methods demonstrated that T-tubules form in areas rich in caveolae, particularly in striking ring-shaped structures dependent on the key BAR domain protein, Bin1 (Lemerle et al., 2023). Enrichment of caveolae in specific regions of these rings was associated with the emergence of Bin1-positive membrane tubules, consistent with a role for these structures in the early stages of T-tubule biogenesis (Fig. 3).

While a number of studies have shown perturbation and disorganization of T-tubules upon CAV3 knockdown, expression of CAV3 variants, or loss of cavin1 (Lemerle et al., 2023; Lo et al., 2015), the idea that T-tubule formation is dependent on caveolae, involving their repeated formation without scission to form a vesicle, now appears to be an oversimplification. T-tubule formation is clearly a robust process, and perhaps alternative compensatory mechanisms can lead to the same final morphological feature. However, it is clear that caveolae are not essential to make T-tubules; mice completely lacking caveolae (due to loss of CAV3 or cavin1) still form T-tubules that, although aberrant, are similar to native T-tubules in both general morphology and function (Hall et al., 2020; Lo et al., 2015; Minetti et al., 2002). The same is true in patients expressing mutant forms of CAV3, in which mild perturbation has been described (Lemerle et al., 2023). However, detailed analysis of T-tubule development using combinations of knockout models, such as CAV3/CAV1 double KO mice (Park et al., 2002), might be required to pinpoint the precise role of caveolae in T-tubule development.

A closer look at the T-tubules in mouse fibers lacking caveolae through loss of cavin1 or CAV3 shows a mild upregulation of T-tubule components, some disorganization of T-tubule elements in the fiber periphery, and the presence of an increased number of longitudinal T-tubule elements (Lo et al., 2015; Ding et al., 2017; Minetti et al., 2002). As these elements, unlike completely transverse T-tubules, are sensitive to longitudinal stretch, this has been linked to the reported RMD phenotype associated with specific CAV3 variants (Lamb, 2005; Berling et al., 2023). The skeletal muscle of caveolinopathy patients was shown to exhibit abnormal branching and swelling of the T-tubules (Lemerle et al., 2023). As CAV3 has been detected at low levels in adult T-tubules (Ralston and Ploug, 1999), it may also play an organizing role in the mature T-tubule. It is interesting to speculate that RMD and excitation–contraction coupling disorders may share a common defect in caveolin-dependent T-tubule/triad microdomains that integrate mechanosensation, membrane excitability, and calcium signaling. In this view, CAV3 is not only a structural component of caveolae, but an organizer of specialized membrane domains associated with the T-tubule system and DHPR–RyR coupling. Disruption of these caveola-associated microdomains would compromise both membrane mechanoprotection and the fidelity of voltage-induced calcium release, providing a mechanistic basis for the electrically silent, mechanically triggered contractions characteristic of RMD.

It is currently difficult to arrive at a satisfying unifying hypothesis to explain all the complex links between caveolae and T-tubules, but it is certainly apparent that the underlying mechanisms for generating the two domains are similar in their dependence on cholesterol (Carozzi et al., 2000). Caveolae also link the T-tubules to the cell surface in mature mammalian muscle fibers. Dysmorphic T-tubules in the absence of caveolae also show the close links between the two processes. Further clues come from studies of the muscle-specific cavin isoforms in zebrafish muscle, cavin4a and 4b (the orthologues of the single mammalian cavin4) (Lo et al., 2021). Loss of cavin4a and 4b leads to a disruption of the usual maturation process occurring in the developing T-tubules. CAV3 is initially associated with the developing T-tubules but then is lost or removed from the T-tubules during development to generate the high level of sarcolemmal caveolae and lack of caveolae in the T-tubules characteristic of mature skeletal muscle (Lo et al., 2021). In the absence of cavin4a/b, this maturation process is perturbed, and CAV3 is retained in the T-tubules. This appears to physically disrupt the T-tubules, which are now highly perturbed, covered in spiraling caveola-like domains, and are dysfunctional as determined using live-cell imaging of calcium release from the sarcoplasmic reticulum (Lo et al., 2021). A similar phenotype was observed in mice in which cavin4 was downregulated during embryonic development (Lo et al., 2021). One explanation for this striking observation is that the early stages of T-tubule formation show similarities to caveolae formation, for example, in their cholesterol dependence, and as a result, caveolar components are also recruited to the forming tubules. However, as the muscle matures, the caveolae must be removed from the T-tubules (Lemerle et al., 2023; Lo et al., 2015; Lo et al., 2021), a process that requires cavin4, which binds to Bin1 to coordinate this process. Caveolae are maintained at high density on the sarcolemma to enable them to perform their functions in signaling and mechanoprotection, whereas the mature T-tubule has a low density of CAV3 and no morphological caveolae (Lo et al., 2015; Lo et al., 2021; Ralston and Ploug, 1999). As outlined above, the general perturbation of cell surface properties in muscle lacking the usually highly abundant caveolae is likely to have diverse effects on the biophysics, domain organization, and dynamics of the sarcolemma, with resulting impacts on the formation of endocytic structures, including the related surface-derived precursor T-tubules.

In this perspective, we have attempted to highlight some of the key studies that have sought to unravel the complex effects of caveolar dysfunction in skeletal muscle. Clearly removal or reduction of such an abundant surface component, as occurs in type I caveolinopathies or upon genetic ablation of CAV3 or cavin1, can have a multitude of effects, many of which could be indirect as a result of the profound reorganization of the cell surface that must occur; we predict changes in biophysical properties, under resting conditions and particularly in response to stress, as well as membrane lipid changes based on studies of non-muscle cells. We have also highlighted work showing that the effects of caveolar loss can be detected early in development, raising the possibility that the late onset of clinical symptoms is a result of complex mechanisms to compensate for this loss of a surface organelle. Cell biological studies have revealed effects of type II variants in specific cellular assays; their contribution to caveolinopathies is even less well understood and remains an intriguing area of research.

The role of caveolae in the T-tubule system and the dysfunction that leads to RMD are fascinating aspects of caveolar biology that are still largely unresolved in terms of mechanistic explanation, with caveolae linked to T-tubule formation and development as well as to the function of the mature T-tubule. More than one process for T-tubule formation has been described (Franzini-Armstrong, 1991), and unraveling the role of caveolae in these different processes may be required. How other aspects of skeletal muscle function relate to caveolae are also unresolved; for example, how neuromuscular junction (NMJ) formation and action potentials affect the behavior of the caveolae at the sarcolemma. CAV3 is also concentrated at the NMJ, playing a role in the clustering of the nicotinic acetylcholine receptor (Hezel et al., 2010), and recent evidence suggests an additional role in the development of junctional folds in postnatal mouse NMJs (Kwan et al., 2023).

In terms of the major functions of caveolae in skeletal muscle, mechanoprotection and sensing of mechanical stress, linking to multiple downstream pathways, clearly appear to be important. If the membrane-organizing role of caveolae, proteins, and lipids is perturbed, this could also have multiple effects on many different signaling pathways and processes such as endocytosis. Release of cholesterol that would be “trapped” in caveolae of WT muscle could also have diverse effects, including effects on mTORC1 signaling, upon perturbation of caveolae. We have not discussed the extensive literature on cardiac muscle caveolae and links to cardiac disease in this review, but it currently remains unclear how the insights from each system relate to the other or how the same organelle could operate in fundamentally different ways in the two tissues. It appears that the effects of CAV3 loss or mutation in cardiac muscle may be partially abrogated by the expression of CAV1 in the heart (Benzoni et al., 2024), but the reason for the presence of both these proteins in cardiac muscle is still unknown. These questions highlight the many mysteries still surrounding these enigmatic pits, but also the importance of understanding how caveolar dysfunction leads to diverse disease conditions.

N. Ariotti is supported by a Human Frontier Science Program Grant (RGP/011/2023) and Australian Research Council (ARC) Discovery Project (DP260101196). R.G. Parton is supported by an ARC Laureate Fellowship (FL210100107) and ARC Discovery Project (DP260101196). T.E. Hall and H.P. Lo are supported by the National Health and Medical Research Council Ideas Grant (2027559). We acknowledge support from the Australian Functional Genomics Network funded by the Medical Research Future Fund (MRF2007498) and administered by the Murdoch Children’s Research Institute.

Author contributions: Harriet P. Lo: conceptualization, investigation, visualization, and writing—original draft, review, and editing. Stéphane Vassilopoulos: investigation, resources, visualization, and writing—original draft, review, and editing. Nicholas Ariotti: visualization and writing—original draft, review, and editing. Thomas E. Hall: conceptualization, funding acquisition, project administration, and writing—original draft, review, and editing. Robert G. Parton: conceptualization, funding acquisition, project administration, visualization, and writing—original draft, review, and editing.