Sarcoplasmic Reticulum-Mitochondria Microdomains: Hugging and Kissing in the Heart

Abstract

Endoplasmic reticulum (ER)-mitochondrial (ER-Mito) interface, termed mitochondrial-ER contacts (MERCs), plays significant roles in the maintenance of bioenergetics and basal cell functions via the exchange of lipids, Ca²⁺, and reactive oxygen species (ROS) in various cell-types/tissues. Genetic deletion of mitofusin 2 (Mfn2), one of the key components of ER-Mito tethering, in cardiomyocytes (CMs) in vivo revealed the importance of the microdomains between mitochondria and sarcoplasmic reticulum (SR), a differentiated form of ER in muscle cells, for maintaining normal mitochondrial Ca²⁺ (mtCa²⁺) handling and bioenergetics in the adult heart. However, key questions remain to be answered: 1) What tethering proteins sustain SR-Mito contact site structure in SR-Mito contact sites in the adult ventricular CMs (AVCMs), the predominant cell type in adult heart; 2) Which MERC proteins operate in AVCMs to mediate specific microdomain functions under physiological conditions; 3) How is the MERC protein expression profile and function altered in cardiac pathophysiology. In this review, we summarize current knowledge regarding the structure, function, and regulation of SR-Mito microdomains in the heart, with particular focus on AVCMs, which display unique membrane organization and Ca²⁺ handling compared to other cell types. We further explore molecular mechanisms underpinning microdomain dysfunction in cardiac diseases and highlight the emerging roles of MERC proteins in the development and progression of cardiac pathology.

Graphical Abstract

1. Introduction

Mitochondrial function is modulated by physical interactions with other organelles, such as the endoplasmic/sarcoplasmic reticulum (ER/SR), peroxisomes, and nucleus (95, 114). Among these, the structural and functional significance of ER/SR-mitochondrial (ER/SR-Mito) interface, termed mitochondrial-ER contacts (MERCs), is well established (2, 89, 96). By virtue of their proximity to Ca²⁺-release sites on the ER/SR (e.g., inositol 1,4,5 trisphosphate receptors [IP₃Rs]) and ryanodine receptors [RyRs]), mitochondria at ER/SR-Mito contact points are exposed to elevated Ca²⁺ concentrations that promotes efficient Ca²⁺ uptake, thereby regulating ATP synthesis and reactive oxygen species (ROS) production (2, 89) in various cell types/tissues, including cardiomyocytes (CMs)/hearts. In addition to Ca²⁺ handling, these contact sites also mediate lipid and ROS exchange, which are critical for the maintenance of mitochondrial bioenergetics (2, 89). Genetic ablation of mitofusin 2 (Mfn2), a key tethering component of ER/SR-Mito contacts, in CMs revealed the importance of the microdomains between mitochondria and SR, a differentiated form of ER in muscle cells, for normal mitochondrial Ca²⁺ (mtCa²⁺) handling and bioenergetics in the heart (13, 41, 86, 93). However, several questions remain: 1) What additional MERC proteins have a structural role at contact sites in the heart beyond Mfn2? 2) Which signaling pathways regulate Ca²⁺, lipid, and ROS at the ER/SR-Mito microdomains during cardiac stress? and 3) How are processes altered in cardiac disease, and what are the downstream effects on mitochondrial and cardiomyocyte (CM) function?

Here, we provide an overview of recent reports on the structural and molecular basis of MERC functions, particularly in adult ventricular cardiomyocytes (AVCMs). As the primary drivers of cardiac contractile force AVCMs exhibit a unique structure characterized by a highly developed network of transverse tubules (T-tubules) that deeply invaginate the plasma membrane, to facilitate efficient electrical signal transmission and coordinated contraction throughout the ventricular muscle in the adult heart (77). Because MERC structure and functions in AVCMs differ significantly from other cell types (see section 2), we attempt to specifically emphasize studies conducted in AVCMs and in vivo animal models rather than the observations and perspectives from non-CMs and CM model cells such as neonatal CMs and induced pluripotent stem cell-derived CMs (see recent reviews (28, 59, 64, 65)).

2. Molecular identity of ER/SR-Mito tethering in adult ventricular cardiomyocytes

In mammalian AVCMs, the predominant cell type in the heart, mitochondria are the most abundant intracellular organelles, numbering in the range of ~ 7000 per cell and occupying up to 35% of the cell volume(106). At least three distinct subpopulations of mitochondria are recognized in AVCMs: interfibrillar, subsarcolemmal, and perinuclear, each distinguished by characteristic morphology and localization (45, 52). Interfibrillar mitochondria, aligned in longitudinal rows between myofibrils, lie in close proximity to the SR. Unlike the ER found in non-muscle cells, the SR in cardiomyocytes is highly specialized for Ca²⁺ cycling and excitation-contraction coupling (78). Consequently, most MERCs in AVCMs are formed by the interaction between interfibrillar mitochondria and junctional SRs located at the Z-bands. This unique spatial arrangement of interfibrillar mitochondria enables privileged signaling crosstalk between the SR and mitochondria, including Ca²⁺, ROS, redox, and pH cascades. Other mitochondrial subpopulations, subsarcolemmal and perinuclear mitochondria are less organized and show variability in morphology. The perinuclear rough ER (76) may also form close contacts with perinuclear mitochondria, contributing to multiple alternative MERC subtypes as shown in dog AVCMs (121).

In large animals, a single mitochondrion or two interfibrillar mitochondria with relatively uniform size and shape typically span a single sarcomere from Z-band to Z-band (30, 90). However, the number of mitochondria residing between the Z-bands is greater in rodents, often two or more mitochondria (16, 81, 90), which provides various-sized and shaped interfibrillar mitochondria. Consequently, the MERC functions in AVCMs, especially in the interfibrillar mitochondria, may differ between species. Given potential differences, in this review, we explicitly state the species used in referenced works in this review.

Several ER-mitochondrial tethering complexes have been characterized in non-CMs (see reviews (34, 116, 119)), typically consisting of an ER-resident protein and an outer mitochondrial membrane (OMM) resident protein. Key examples include: 1) Mfn2-Mfn1/Mfn2 homo/heterotypic complexes, 2) IP₃R-voltage-dependent anion-selective channel (VDAC) complex mediated by the chaperone glucose regulated protein 75 (GRP75), 3) vesicle-associated membrane protein-associated protein B (VAPB)-protein tyrosine phosphatase-interacting protein 51 (PTPIP51) complex, and 4) protein B-cell receptor-associated protein 31 (Bap31)-mitochondrial fission protein Fission 1 homologue (Fis1) complex. Additionally, the ER-membrane protein PDZ domain-containing protein 8 (PDZD8) has been proposed as part of the ER-Mito tethers in neurons. However, the mitochondrial counterpart of PDZD8 remains unidentified in any cell type, and the function of PDZD8 in MERCs has not been investigated in the heart (Table 1).

Table 1.

Protein expressions of potential SR/ER-Mito tethering proteins in the adult hearts

Protein/gene name	Subcellular location	Protein expression in human hearts	Protein detection in MAM mass spec in rat hearts	Reports on AVCMs/hearts related to ER/SR-Mito tethering function
Mfn2	SR/ER, OMM	Medium	Yes	(13, 41, 86, 93)
Mfn1	OMM	Not Detected	Yes	(13, 88)
VAPB	SR/ER	Medium	Yes
PTPIP51	OMM	Medium	No	(92)
IP3R1	SR/ER	Low	No	(86)
IP3R2	SR/ER	Medium	No
IP3R3	SR/ER	Medium	No
GRP75	Cytoplasm	High	Yes	(86)
VDAC1	OMM	Medium	Yes	(86)
VDAC2	OMM	High	Yes	(113)
VDAC3	OMM	High	Yes
DJ-1	Cytoplasm	Low	Yes
FUNDC1	OMM	Medium	Yes	(113)
Bap31	SR/ER	Not Detected	Yes
Fis1	OMM	High	Yes
PDZD8	ER	medium	No

The data of the protein expression levels in human hearts are from the Human Protein Atlas (109). Protein detection in MAM protein mass spectrometry was from a recent publication using adult rat hearts (68).

According to the Human Protein Atlas (109), most of the proteins listed above are detectable in adult human heart tissues and CMs (Table 1). Furthermore, quantitative proteomic analysis using mitochondria-associated membrane (MAM)-enriched fractions from adult rat hearts detected most of the candidates of the tethering proteins (68) (Table 1). In the following sub-sections, we summarize recent reports on the tethering structural proteins in CMs/ heart tissue, with particular attention to the Mfn2-Mfn1/Mfn2 complex, the IP₃R-GRP75-VDAC complex, and the VAPB-PTPIP51 complex in AVCMs. To date, there are no reports specifically investigating the roles of Bap31, Fis1, and PDZD8 in SR-Mito tethering in AVCMs/ heart tissue (Table 1).

2.1. Mfn2-Mfn1/Mfn2 complex:

While the IP₃R-GRP75-VDAC complex was the first proposed ER-Mito tethering structure (107), predating the discovery of Mfn2 dimers (21), Mfn2 is now recognized as the most extensively studied tethering complex in rodent and human AVCMs and cardiac tissues (Table 1). Studies using CM-specific Mfn2 knock-out (KO) mice have demonstrated that Mfn2 is a critical determinant of the physiological morphology of SR-Mito microdomains as well as in the maintenance of normal mtCa²⁺ handling and bioenergetics in the adult heart (13). In non-CMs, ER-located Mfn2 can form homodimers (Mfn2-Mfn2) or heterodimers with OMM-localized Mfn1 (21, 83). However, in adult mouse hearts, CM-Specific Mfn1 KO does not significantly affect the size of SR-Mito microdomains (13, 88). These observations suggest that the Mfn2 homodimers play a more dominant role than Mfn2-Mfn1 heterodimers in maintaining SR-Mito contacts under physiological conditions (13, 88). Supporting this, Mfn1 expression in human heart tissue is markedly lower than Mfn2 (or reported as “not detectable” in Human Protein Atlas, see Table 1). Interestingly, despite this, Mfn1 expression level has been proposed as a biomarker for heart failure (HF) in non-responding patients with idiopathic dilated cardiomyopathy(47). This suggests a potential pathological role for Mfn1, though further studies are needed to clarify its function within MERCs in human cardiac disease. Although Mfn2 is widely considered a tethering protein, several controversial studies in non-CMs have shown that Mfn2 deletion increases ER-Mito coupling (15, 32, 56) (see also a review (33)). In the AVCMs, Walsh’s group similarly reported no significant change in SR-Mito contact site size following KO of Mfn2 in mice (88). These discrepancies may result from differences in KO/KD strategies, experimental models (i.e., in vitro vs. in vivo), or imaging and analysis techniques. Nevertheless, even when differences are observed, the reported changes in the ER/SR-Mito distance Mfn2 KO or KD are relatively modest, falling within 15-18 nm range (13, 21), suggesting that Mfn2 likely fine-tunes rather than determines the SR/ER-Mito distance. Given that Mfn2 also plays a canonical role (i.e., OMM fusion), it remains unclear whether and how Mfn2 regulates the MERCs via tethering-independent mechanism. Further mechanistic studies beyond gene deletion or KD approaches are needed to dissect Mfn2’s diverse functions in AVCMs.

2.2. IP₃R-GRP75-VDAC complex:

Glucose regulated protein 75 (GRP75), a product of the HSPA9 gene, also known as Mortallin, PBP74, and mitochondrial HSP70 (mtHSP70) was initially identified as a molecular chaperone that functionally links the metabolic flow, Ca²⁺, and cell death signaling between the ER and mitochondria by physically connecting VDAC isoform 1 (VDAC1) at the OMM with ER Ca²⁺-release channels, IP₃Rs (107). Co-immunoprecipitation experiments by De Stefani and Rizzuto’s group demonstrated that GRP75 interaction is isoform-specific: GRP75 forms a complex with VDAC1, but not VDAC2 and VDAC3(25). These data support the role of VDAC1 as the primary conduit for Ca²⁺ transit among the VDAC isoforms involved in intrinsic apoptosis (25). Additionally, they showed that at least two IP₃R isoforms, IP₃R1 and IP₃R3, can bind to GRP75 (25). Given the high sequence homology among the three IP₃R isoforms (35) and evidence that all three are required for maintaining ER-Mito contact sites independent of their role in Ca²⁺ flux (3), it is plausible that IP₃R2 can also associate with GRP75 to form a tethering complex. In adult heart tissues, IP₃Rs, GRP75, and VDACs are all expressed, although expression levels of VDACs vary among the isoforms (see Table 1). In mouse AVCMs, the presence of an IP₃R1-GRP75-VDAC1 complex likely exists since the association of IP₃R1, GRP75, and VDAC1 was detectable by immunoblotting with subcellular fractionated proteins from mouse hearts (86) (Table 2). While IP₃R-, GRP75-, and VDAC-KO mice have been generated, there are no published studies examining whether any structural and functional alterations in SR-Mito contact sites occur in the hearts of these animals. Interestingly, DJ-1, a protein originally linked with autosomal recessive early-onset Parkinson’s disease, has been proposed as one of the regulators of the IP₃R3-GRP75-VDAC1 complex (63). DJ-1 has also been reported to exert cardioprotective effects under various pathological stress conditions(29, 102, 103). However, whether these effects are mediated via the modulation of the IP₃R-GRP75-VDAC1 complex in AVCMs remains unclear.

Table 2.

Proteins associated with SR-Mito Ca²⁺ transport in adult ventricular myocytes (AVCMs)

Protein	Location	Notes
RyR2	SR	RyR2 is specifically located on the SR subdomain facing the T-Tubules, which is the opposite side of mitochondria (36, 71). Functional RyR2 is also expressed at the network SRs that wrap the interfibrillar mitochondria (71).
SERCA2a	SR	SERCA2a is abundant in the junctional SR, which faces mitochondria (60) and is detectable in MAMa from adult mouse and rat hearts (68, 74).
phospholamban (PLB)	SR	PLB, a regulator of SERCA, is detected in MAMs from mouse hearts (74).
sarcolipin (SLN)	SR	No (or very low) expression of SLN, a potent inhibitor of the SERCA in AVCMs under physiological condition. In the Duchenne muscular dystrophy mouse model, SLN is overexpressed and detectable in MAMs from AVCMs (74).
IP3Rs	SR	In the adult mouse hearts, IP3R-GRP75-VDAC1 interaction was detectable, suggesting that IP3Rs expression at the MAMs of AVCMs (86). IP3Rs were not detectable in MAM protein mass spectrometry using adult rat hearts (68).
MCU	IMM	MCU is concentrated at the junctional SR-Mito contact sites in mouse AVCMs (22). MCU is expressed in all areas of IMM equally in the rabbit AVCMs (66).
EMRE	IMM	EMRE is concentrated on the junctional SR-Mito contact sites in mouse AVCMs (22).
MICUs	IMM	MICU1, but not MICU2, was detected in MAM protein mass spectrometry using adult rat hearts (68).
VDACs	OMM	All three isoforms were detected in MAM protein mass spectrometry using adult rat hearts (68).
NCLX	IMM	NCLX is excluded from the junctional SR-Mito contact sites (23)

FUN14 domain containing 1 (FUNDC1) is a highly conserved OMM protein that can bind to IP₃R2 in neonatal CMs to modulate ER Ca²⁺ release (113). Importantly, FUNDC1 is also required to maintain the integrity of MERC structures (113) (Table 1). Since FUNDC1 KO mice showed significant alterations in mitochondrial morphology and oxygen consumption rate, further studies are needed to precisely understand whether the effects of FUNDC1 on MERC structures are mediated via the Ca²⁺-flux of IP₃R2, alterations of mitochondrial morphology, and/or structural changes in the IP₃R-GRP75-VDAC tether complex. In summary, all the components of the IP₃R-GRP75-VDAC complex and its regulatory proteins are expressed in the adult heart. However, the relative contribution of the IP₃R-GRP75-VDAC complex and Mfn2 dimers (see section 2.1) to SR-Mito tethering in AVCMs is still unknown.

2.3. VAPB-PTPIP51 complex:

Vesicle-associated Membrane Protein-Associated Protein B (VAPB) is part of the VAMP-associated protein family and is a key player for facilitating tethering at membrane contact sites between the ER and other intracellular membranes such as mitochondria or plasma membrane (47). So far, investigation on the roles of VAPB in CMs is still in its early stages (73, 104). To date, no reports show the specific role of VAPB in the formation of ER/SR-Mito contact sites and its contributions to the MERC functions in the CMs (Table 1). A single report showed that at least its binding partner PTPIP51, also known as RMDN, may regulate SR-Mito physical contact sites in the mouse AVCMs (92) (Table 1); PTPIP51 KD has been shown to decrease SR-Mito contacts in mouse hearts in vivo, suggesting the potential existence of the VAPB-PTPIP51 complex in the SR-Mito contact sites in AVCMs.

3. SR-Mito contact-site proteins and their biological functions in adult ventricular cardiomyocytes

ER-Mito microdomains play crucial roles in the exchanges of Ca²⁺ and lipid as well as the hubs for cellular signaling pathways that impact processes such as autophagy, apoptosis, inflammasome formation, and metabolic modulation (2, 7, 18, 38, 89, 94, 96). These functions are orchestrated by diverse proteins enriched within the MERC regions. A recent quantitative proteomics study identified 1871 proteins in the MAM fractions from adult rat hearts (68), most of which are presumably from SR-Mito contact sites of AVCMs. Among them, 216 proteins were previously reported as “consensus MAM proteins”. Notably, in cardiac MAMs, mitochondrial proteins were the predominant subcellular contributors. In contrast, in other tissues, such as the liver, brain, and testis, ER-resident proteins were the largest group (68). These findings suggest that SR-Mito contact sites in AVCMs differ significantly from ER-Mito contact sites in other cell types in both structure and function. Proteins identified in cardiac MAMs spanned a range of functional categories (from most to least abundant): Metabolism (e.g., ATP bioenergetics and lipid metabolism), organelle organization (e.g., mitochondrial fission/fusion), signaling (e.g., apoptosis and immune response), post-translational modification (e.g., phosphorylation and dephosphorylation), transport (e.g., vesicle and cation transport), and autophagy.

To date, the most thoroughly investigated MERC function in AVCMs is Ca²⁺ handling, which is central to excitation-contraction-metabolism coupling in the heart (see reviews (24, 111, 120)). This has been dissected using in vivo knockout (KO) and knockdown (KD) mouse models targeting specific Ca²⁺-handling proteins (see Table 2). As outlined in Section 2, mitochondrial dynamics proteins, notably Mfn1 and Mfn2, have also been shown to modulate SR-Mito contact sites in AVCMs. In addition, several studies have begun to explore the consequences of KO/KD of lipid metabolism-associated proteins at SR-Mito interfaces. In this section, we overview the recent reports that investigated two MERC functions in AVCMs, Ca²⁺ and lipid transport. Although other potential MERC functions, such as bioenergetics, apoptosis, immune response, and autophagy, are of great interest, they are hampered by an extremely limited number of publications assessing these functions in the context of SR-Mito contact sites of AVCMs. Therefore, these MERC functions and their potential relationship with cardiac physiology and pathology will be discussed in the following sessions as a perspective.

3.1. Ca²⁺ transport:

We previously showed the first indication of the existence of Ca²⁺ communication between SR and mitochondria through the SR-mitochondrial microdomain in CMs; the application of Ca²⁺ chelator 1,2-Bis(2-aminophenoxy) ethane-N,N,N′,N′-tetraacetic acid (BAPTA) to the cytosol in saponin-permeabilized CMs was able to suppress the global cytosolic Ca²⁺ transient but not the mitochondrial Ca²⁺ transient following rapid caffeine-induced SR Ca²⁺ release, demonstrating the privileged communication between SR and mitochondria that relies upon local communication at the micrometer scale (100). Similarly, Maack’s group demonstrated SR-Mito Ca²⁺ transport in the intact AVCMs after the physiological SR Ca²⁺ release by L-type Ca²⁺ channel activation (51).

In non-CM cells, it is widely established that IP₃Rs, the major Ca²⁺ release units at the ER membrane, face toward ER-Mito contact sites and are involved in Ca²⁺-mediated crosstalk between ER and mitochondria (26) (Fig. 1A, left); The IP₃Rs release Ca²⁺ from ER lumen to the ER-Mito interface where microdomains of Ca²⁺ reach concentrations of 10–30 μM (17). This Ca²⁺ flows into the mitochondrial intermembrane space (IMS) via VDACs at the OMM, then crosses the inner mitochondrial membrane (IMM) through pores created by the mtCa²⁺ uniporter (MCU)-protein complex (mtCUC) at the IMM into the mitochondrial matrix (Fig. 1A, left).

Fig. 1. Structural difference in Ca²⁺ transport system at MERCs in adult ventricular cardiomyocytes (AVCMs) and non-AVCMs.

A. Schematic diagrams of MERC Ca²⁺ transport system in non-CM (top) and AVCMs (bottom). Major Ca²⁺ releasing units of ER (i.e., IP₃Rs) face towards mitochondria, whereas major Ca²⁺ releasing unit of SR (i.e., RyR2) face towards T-Tubules. The Ca²⁺ concentration after the ER/SR Ca²⁺ release was shown as heat maps. Distributions of MCU and NCLX within the mitochondria in AVCMs are simplified, and the SR-Ca²⁺ uptake system is abbreviated (see details in Section 3.1). B. Summary of the key differences in the structural MERC Ca²⁺ transport system in AVCMs and non-CMs.

In AVCMs, beat-to-beat Ca²⁺ release from junctional SR occurs almost exclusively via ryanodine receptor type 2 (RyR2) and not from the IP₃Rs (Fig. 1B). Moreover, the majority of RyR2s were localized along the T-tubule side of the SR rather than the SR-Mito microdomains (Fig. 1A, right and 1B) (36, 71). Indeed, this unique localization of RyR2s at junctional SR tightly controls Ca²⁺-induced Ca²⁺ release (CICR) in CMs (4), but high local concentrations of Ca²⁺ after CICR in the microdomains between the T-tubule membrane and SR should diffuse near OMM to trigger mitochondrial Ca²⁺ uptake via VDACs and mtCUC. However, the actual exposure time to high local Ca²⁺ concentrations (10-20 μM) at MERCs, that can trigger mtCa²⁺ uptake via mtCUC (50), might be quite brief (i.e., approximately 10 ms) (112). Moreover, the mtCUC channel conductance from adult mouse hearts is extremely small compared to that from other organs (31). Therefore, Lederer’s group proposed that physiological mtCa²⁺ uptake via mtCUC during the heartbeats is small and inherently inefficient due to the greater distance from RyR2 at junctional SR to the OMM (50–100 nm) compared to that from IP₃R at ER to the OMM in non-CMs (~20 nm) (91), thus does not significantly impact any global Ca²⁺ handling profile in AVCMs (112) (Fig. 1B). However, we previously demonstrated biased localization and subunit composition of mtCUC complexes at the SR-Mito contact sites: the EMRE-containing mtCUC is preferentially expressed at the MERC area rather than other IMM areas that do not face junctional SR (22). This specialized mtCUC distribution may compensate for the spatial (i.e., RyR2-OMM distance) and functional (i.e., low mtCUC channel conductance) disadvantages and facilitate effective Ca²⁺ transfer from SR to mitochodnria. Furthermore, cardiac mitochondria are relatively deficient in the mtCUC component MICU1, conveying enhanced sensitivity to low cytosolic Ca²⁺ concentrations(43). In contrast to specialized mtCUC distribution at MERCs in the AVCMs (22), we also reported that mitochondrial Ca²⁺ extrusion via the Na⁺/Ca²⁺ exchanger (NCLX) is excluded from IMM regions facing the junctional SR-Mito contact sites (23). These asymmetries in the localization of Ca²⁺ uptake and extrusion machinery may enhance mtCa²⁺ handling at minimal energetic cost. While our idea is based on functional, biochemical, and imaging assays, Bers’s group reported relatively uniform MCU distribution over the mitochondrion in the rabbit AVCMs (66). Takeuchi and Matsuoka reported that NCLX is preferentially localized in MERC area in mouse AVCMs (108). This discrepancy may reflect species differences and/or the different methodologies selected, but further computational modelling is needed to establish whether efficient mtCa²⁺ influx/efflux is driven by geometric positioning or specialized MERC organization as we propose.

Although mass spectrometry and MAM Western blotting have detected RyR2 in cardiac MAM fractions (68, 74), its expression is likely much lower than that in T-tubule side, as shown in the studies using electron microscopy (36, 71). One possibility is that a small RyR2 subset residing in the MERCs formed by junctional SRs and interfibrillar mitochondria contributes to localized Ca²⁺ release at the MERCs. Indeed, there are several reports suggesting the functional coupling of RyRs and VDAC2 in the mouse and zebrafish hearts (98, 101) and physical interactions between RyR2 and VDAC2 at MERCs were shown in the rat neonatal CMs (79). However, due to the limited number of RyR2 within the MERCs formed by junctional SR and interfibrillar mitochondria in the AVCMs, further studies are required to precisely determine whether this small number of RyR2 facing OMM has a significant impact on the junctional SR-Mito Ca²⁺ transport. Moreover, as suggested by Kim’s group, these RyR2-VDAC2 interactions exist only in the specific population of mitochondria, such as subsarcolemmal mitochondria, but not in interfibrillar mitochondria (79).

Another possibility is that RyR2 expressed at the networked SRs, which reside between the Z-lines and interacts significantly with interfibrillar mitochondria (71), may provide an additional and faster Ca²⁺ transport pathway to the OMM of interfibrillar mitochondria. This specific population of RyR2 could contribute to propagating CICR (70), but the relative contribution to mtCa²⁺ transport in interfibrillar mitochondria remains unclear.

In addition to the main SR Ca²⁺ releasing channel RyR2, Dedkova and Ritter’s group showed that all three isoforms of IP₃Rs are expressed in the mouse AVCMs, and G_q protein-coupled receptor stimulation can trigger the Ca²⁺-release from SR via these IP₃Rs, promoting mitochondrial Ca²⁺ uptake (99). While the localization of IP₃R at the MERCs in AVCMs remains to be directly confirmed, these data suggest that IP₃Rs-dependent Ca²⁺ transfer may contribute to SR-mitochondrial signaling outside of the beat-to-beat context.

In summary, there are three potential Ca²⁺ releasing pathways from SR pathways that underlie Ca²⁺ transfer to interfibrillar mitochondria in AVCMs: 1) limited localization of RyR2 at MERCs formed by junctional SRs, 2) RyR2 expressed at the network SR, and 3) the IP₃R-mediated Ca²⁺ release at MERCs. Additionally, we also need to take into consideration that the activity of SR/ER Ca²⁺-ATPases (i.e., Ca²⁺ uptake into SR) is capable of modulating the amplitude and the speed of the Ca²⁺ elevation and reduction within the MERC region (91). Indeed, SERCA2a and its regulatory protein phospholamban were detected from adult mouse hearts (40, 74) (Table 2). Future studies are needed to define how these spatial and molecular mechanisms integrate into the broader Ca²⁺-signaling landscape of the AVCMs.

3.2. Lipid transport:

Cardiolipin and phosphatidylglycerol are mitochondria-specific phospholipids that are synthesized within mitochondria (46). Among these, cardiolipin, a critical regulator for mitochondrial respiration (96), exists in its highest concentrations in the heart, reflecting the organ’s high oxidative demand. Mitochondria can also synthesize phosphatidylethanolamine (46). However, to synthesize these lipids, phosphatidylcholine, phosphatidylinositol, phosphatidylserine, and sterols must be imported from neighboring organelles, including ER. For optimal Lipid transport, ER-Mito contact distance needs to be around 10 nm, a size shorter than the efficient Ca²⁺ transport from IP₃Rs (39). Recent studies showed that MERCs possess multiple lipid-synthesizing enzymes and transport machinery (see review (96)). Many of these enzymes and transport proteins are also detectable in MAM proteomes, including those from adult rat AVCMs (68).

Among MERC-localized tethering proteins, Mfn2 and PTPIP51 have emerged as critical regulators of lipid metabolism in addition to their structural roles (see Section 2). Specifically, Mfn2 facilitates the transfer of phosphatidylserine and PTPIP51 phosphatidic acid from ER to mitochondria, respectively (44, 117). Although CM-specific KO and KD of Mfn2 and PTPIP51 in mice have been tested (see section 2), it has not been precisely investigated whether any changes in the lipid metabolism in AVCMs occur after the ablation of Mfn2 or PTPIP51. Given a lipid overload-induced dysfunction of cardiac mitochondria (49), lipid metabolism at MERCs in AVCMs may play a critical role in maintaining bioenergetics and membrane integrity in cardiac mitochondria.

4. Role of SR-Mito contact sites in cardiac pathophysiology

As shown above, recent research in non-CM cells highlighted the significance of MERCs in the various critical cellular processes under physiological and pathophysiological conditions. As we described in the above section, MERC functions in non-AVCMs and AVCMs should have critical differences such as mtCa²⁺ homeostasis, because major MERCs in AVCMs are the SR-Mito contact sites (i.e., not ER-Mito). These structural and functional differences in AVCMs may also provide unique contributions to the development and progression of cardiac pathophysiology.

4.1. Ischemic/reperfusion injury:

Although it is still controversial whether basal Ca²⁺ concentration at the mitochondrial matrix is involved in maintaining basal bioenergetics in the AVCMs (37, 61, 111), physiological mtCa²⁺ loading via mtCUC is required for the "fight-or-flight" response that enables the heart to match its workload with ATP production during adrenergic stimulation (53, 72) (Fig. 2A). The size and organization of SR-Mito contact sites is critical for this response. For instance, Mfn2 KO in AVCMs blunts bioenergetic response against the cytosolic Ca²⁺ elevation after β-adrenoceptor stimulation, indicating that proper SR-Mito architecture is necessary for efficient mtCa²⁺-mediated metabolic coupling (13). (Fig. 2A). The mtCa²⁺ overload via mtCUC and subsequent opening of the mitochondrial permeability transition pore (mPTP), ROS generation, and apoptotic signaling activation are well-established phenomena, especially under ischemia/reperfusion (I/R) (37). Ramasam’s group showed that under I/R stress, SR-Mito distance becomes significantly shorter, which exacerbates mitochondrial Ca²⁺ overload (118) (Fig. 2B). They also found that a canonical effector for Rho small GTP-binding proteins, diaphanous-related formin 1 (DIAPH1) is overexpressed during I/R, whereupon it binds to Mfn2, which modulates SR-Mito distance, mitochondrial turnover, mitophagy, and oxidative stress (85, 118); Strikingly, cardiac specific DIAPH1 KO increases SR-Mito distance without impacting the basal cardiac function, and promotes AVCMs to be protected against I/R injury (118). This observation suggests that a significant decrease in SR-Mito distance is detrimental to cardiac stress, such as I/R. Indeed, Ovize’s group also showed that inhibition or deletion of the regulatory proteins for IP3R-GRP75-VDAC complex protects AVCMs from hypoxia-reoxygenation injury, passively via modifying in the amount/size of SR-Mito microdomains (40, 86). In line with this idea, we also tested the impact of enhancing the SR-Mito interaction by the cardiac-specific introduction of synthetic linker construct in vivo (84). However, synthetic linker expression does not impact on the overall cardiac function or promote mitochondrial Ca²⁺ overload-mediated apoptotic signaling activation despite the changes in SR-Mito contact site morphology. This result, combined with the report of the basal cardiac function of DIAPH1 KO mice (118), suggests that chronic alterations of SR-Mito distance in the adult hearts may not affect the basal cardiac functions, possibly via an adaptive, compensatory mechanism. (84). Nevertheless, integrated SR-Mito structures are essential during prenatal cardiac development, and disruptions in MERC structure during this period are known to cause severe defects (14)

Fig. 2. Structural alterations of SR-Mito microdomains in the adult ventricular cardiomyocytes and their potential impact on the cardiac functions under pathophysiological conditions.

A. Physiological mtCa²⁺ loading via SR-Mito microdomains and mtCUC is required for the "fight-or-flight" response that enables the heart to match its workload with ATP production. B. The mtCa²⁺ overload via cytosolic Ca²⁺ elevation and narrowed SR-Mito contact sites under I/R promotes mitochondrial depolarization, ROS generation, mPTP opening, and apoptotic signaling activation. C. The mtCa²⁺ reduction due to the wide SR-Mito contact sites in heart failure causes lower ATP production and higher mitochondrial ROS generation.

4.2. Heart failure:

Chronic HF is the leading cause of mortality in cardiovascular disease (8). It is characterized by both structural remodeling (e.g., detubulation and myofilament degradation) (57, 97) and transcriptional changes, including the reactivation of fetal gene expression profiles(75). As described in Sections 2 and 3, SR-Mito tethering structures and MERC proteins facilitate communication between these organelles and regulate various cellular processes critical for AVCM functions.

Recently, co-existence of HF and increased SR-Mito distance (and/or decrease in the amount/size of contact sites) was reported in the animal models (19, 113) (Fig. 2C). However, the investigation of the MAM protein expression profiles in the human HF has just begun, and the information is still limited. Using topographic electron microscopy analysis and human AVCMs isolated from the ischemic, ischemic-dilated, and dilated cardiomyopathy patient hearts. Elrod’s group recently reported a significant increase in the SR-Mito distance compared to those from nonfailing control hearts (55) (Fig. 2C). Moreover, like HF animal models (19, 113), Elrod’s group also showed decreased expression of the major tethering proteins in failing hearts compared to controls, suggesting that SR-Mito tethering structure is likely diminishing under human HF (55). Increased SR-Mito distance can lower the Ca²⁺ concentration at the mitochondrial matrix, which impairs heart function by decreased ATP production and higher ROS production (42, 62, 105) (Fig. 2C).

Cardiac hypertrophy, often a precursor to non-ischemic HF, further illustrates dynamic MERC remodeling. During early hypertrophy, AVCMs increase their expression of MAM proteins, followed by a progressive decline as the disease advances (69). This suggests that loss of tethering proteins may contribute to the transition from compensatory hypertrophy to HF. However, cardiac-specific KO of Mfn1/2 or PTPIP51 itself does not lead to overt HF, despite mild dysregulation of mitochondrial bioenergetic impairments. (13, 41, 92). These findings imply that loss of SR-Mito contacts alone is insufficient to cause HF in the adult heart but may facilitate or exacerbate disease progression under stress conditions(19, 113). Nevertheless, further studies are needed to precisely understand whether MERC remodeling initiates HF pathogenesis or occurs as a secondary consequence of broader structural and metabolic changes. A detailed understanding of MERC involvement in adaptive vs. maladaptive remodeling will help clarify its role in HF.

4.3. Diabetic cardiomyopathy:

Diabetic cardiomyopathy, a major complication of type 2 diabetes mellitus, is frequently associated with Ca²⁺ signaling abnormalities in AVCMs (1, 90). HF patients with diabetes experience worse clinical outcomes, including a higher risk of mortality and hospitalization compared to those without diabetes (87). Understanding the cellular mechanisms of unique diabetic cardiomyopathy is therefore essential for the development of targeted therapy. Importantly, weighted gene co-expression network analysis showed that MERC-related genes were clustered into a module correlated with diabetic cardiomyopathy (69), suggesting the critical link between MERC protein expression and SR-Mito communication in disease progression. In a high-fat, high-sucrose diet mouse model, Paillard’s group reported a significant increase in the number of tighter interacted SR-Mito contact sites (less than 10 nm) in AVCMs compared to control animals (27). Such tight organelle associations facilitate lipid transport but may impair efficient Ca²⁺ transport from IP₃Rs to mitochondria in non-CMs (39). Moreover, the normalization of the size of SR-Mito contact sites with the genetic introduction of an artificial tethering construct rescued the abnormality in the mtCa²⁺ homeostasis and bioenergetics, underscoring the consequences of SR-Mito remodeling (27). Challenges in this field are as follows: 1) lack of evidence of SR-Mito contact sites in human diabetic cardiomyopathy (i.e., whether the alteration trend in human is similar direction with the animal model or not), 2) identifying the SR-Mito distance that produces most efficient Ca²⁺ transport to the mitochondria specifically in the AVCMs since the major SR Ca²⁺ releasing sites in AVCMs is RyR2 (see section 3), 3) understanding the relative contribution of MERC changes and global mitochondrial morphology changes (48). Lastly, clarifying the exact SR-Mito distance suitable for lipid transport in AVCMs is particularly important to understand the pathology of lipid overload-mediated mitochondrial dysfunction in the heart. Answering these questions will enhance the understanding of how MERC remodeling contributes to diabetic cardiac pathology and may reveal therapeutic targets to restore mitochondrial function in this context.

4.4. Cardiomyopathy associated with Duchenne muscular dystrophy:

Duchenne muscular dystrophy (DMD) frequently exhibits cardiomyopathy. Babu’s group discovered overexpression of sarcolipin (SLN), a potent inhibitor of the SERCA pump, in the ventricles of mouse DMD models and human DMD (110). They further demonstrated that the overexpressed SLN is preferentially localized in the MAMs of DMD AVCMs (Table 2), which potentially increases Ca²⁺ concentration in the MERC area by inhibiting SERCA2a activity (i.e., inhibiting Ca²⁺ uptake into junctional SR), leading to mitochondrial dysfunction and elevated cellular oxidations (74). These reports strongly support the idea that SERCA2a is indeed an important factor for determining the microdomain Ca²⁺ concentration within the MERC region in AVCMs (see 3.1). While the SERCA2a and its regulatory proteins are in the MERC regions, further studies are needed to determine whether these proteins can modulate the microdomain structure in AVCMs, given that interaction between SERCA2 and Mfn2 has been shown in non-CM cells (115).

5. Summary and future perspectives

In this review, we highlighted the recent reports on the structural and molecular basis of SR-Mito contact sites in AVCMs, which are functionally and morphologically distinct from ER-Mito contact sites observed in other cell types. As summarized above, most studies to date focus on Ca²⁺ transfer from the SR to mitochondria, while other MERC functions, such as downstream mitochondrial bioenergetics, apoptosis, ROS production, and mitophagy, are generally examined as downstream consequences of altered Ca²⁺ handling. Although lipid metabolism at MERCs has been well recognized in the context of cardiac pathology (20), the specific role of SR-Mito contact sites in AVCM lipid exchange remains largely unexplored. Similarly, while immune signaling is a known function of MERCs (12, 80), no studies have directly examined this in the AVCM context. Future studies will provide some clues to assess whether SR-Mito contact sites possess these roles beyond Ca²⁺ transport in the AVCMs.

In this review, we emphasized the roles of SR-Mito contact sites in AVCMs, a major form of MERCs in this cell type, since SR and ER are recognized as functionally distinct internal membrane compartments (78) (Fig. 1A). However, a small population of the ER-Mito contact sites should also exist in AVCMs in addition to SR-Mito contact sites. As briefly mentioned in Section 2, ER-Mito contact sites in AVCMs are likely located at the perinuclear area where perinuclear mitochondria (52) and the majority of IP₃Rs (82) are located. . Such perinuclear mitochondria would be ideally positioned to engage in retrograde signaling toward the nuleus by direct and indirect routes, including via signaling crosstalk with redox-sensitive IP3 receptors (5, 11)Precise dissections of each MERC function in the different subcellular locations, such as applying live AVCM imaging to quantitatively measure MERC functions in both the dyads and perinuclear area (67) may uncover the potential roles of ER-Mito contact sites in the perinuclear mitochondria of AVCMs.

As mentioned in Section 2, MERC function from AVCMs might be slightly different across species, particularly in large animals, including humans, because of differences in the morphology of the interfibrillar mitochondria and the cytosolic Ca²⁺ environment (e.g., slower heartbeats), although majority of the cited research in this review is from rodent models. While sex differences in expressed genes encoding mitochondrial proteins, and the machinery for SR Ca²⁺ release, mtCa²⁺ uptake, and mitochondrial ROS have been reported (10, 16, 54), there are currently no reports specifically investigating the potential sexual dimorphism in MERC structures and functions in AVCMs. Collecting this information may be critical to develop human-specific and sex-specific computational models for mtCa²⁺ influx/efflux at MERCs in AVCMs, which would serve as valuable tools for precisely understanding the pathological importance of MERCs in the development of human cardiac diseases.

Lastly, atrial CMs are the second major CM type in the heart, which have distinct features of plasma membrane and SR membrane structures and Ca²⁺ handling mechanism (6, 9). A recent study using the sinoatrial node-specific Mfn2 KO mice showed the critical role of SR-Mito contact sites and Mfn2 in sinoatrial node automaticity (93). The decrease in the number of SR-Mito contact sites and Mfn2 protein levels was shown in right atrial appendages of patients with persistent atrial fibrillation compared with control patients (58). These reports clearly showed the importance of Ca²⁺ transport at SR-Mito contact sites in atrial CM function under physiological and pathophysiological conditions. Future studies are needed to test whether other tethering machineries (see section 2) also exist in atrial CMs and, if so, what their relative contributions are compared to Mfn2.