Cardiac T-Tubule Microanatomy and Function

Abstract

Unique to striated muscle cells, transverse tubules (t-tubules) are membrane organelles that consist of sarcolemma penetrating into the myocyte interior, forming a highly branched and interconnected network. Mature t-tubule networks are found in mammalian ventricular cardiomyocytes, with the transverse components of t-tubules occurring near sarcomeric z-discs. Cardiac t-tubules contain membrane microdomains enriched with ion channels and signaling molecules. The microdomains serve as key signaling hubs in regulation of cardiomyocyte function. Dyad microdomains formed at the junctional contact between t-tubule membrane and neighboring sarcoplasmic reticulum are critical in calcium signaling and excitation-contraction coupling necessary for beat-to-beat heart contraction. In this review, we provide an overview of the current knowledge in gross morphology and structure, membrane and protein composition, and function of the cardiac t-tubule network. We also review in detail current knowledge on the formation of functional membrane subdomains within t-tubules, with a particular focus on the cardiac dyad microdomain. Lastly, we discuss the dynamic nature of t-tubules including membrane turnover, trafficking of transmembrane proteins, and the life cycles of membrane subdomains such as the cardiac BIN1-microdomain, as well as t-tubule remodeling and alteration in diseased hearts. Understanding cardiac t-tubule biology in normal and failing hearts is providing novel diagnostic and therapeutic opportunities to better treat patients with failing hearts.

I. INTRODUCTION

Cardiac transverse tubules (t-tubules) are highly branched invaginations of cardiomyocyte sarcolemma that are rich in ion channels important for excitation-contraction (EC) coupling, maintenance of resting membrane potential, action potential initiation and regulation, and signaling transduction. Specific to striated muscle cells, t-tubules are invaginations of the sarcolemma, penetrating into the intracellular space of myocytes. These tubular invaginations form a complex network with transverse tubules that are interconnected within the cytoplasm by longitudinal tubules. The transverse tubules occur around myofilaments, anchoring to sarcomeric z-discs through costameres. The phospholipid-rich t-tubule lipid bilayers are shaped and maintained by membrane scaffolding proteins and intracellular cytoskeleton, as well as surrounding extracellular matrix. T-tubule membrane contains microdomains that compartmentalize transmembrane ion handling proteins and signaling molecules. Together, t-tubules are effective organelles that serve as a centralized signaling hub controlling cardiac contractile function and electrophysiology.

The most recognized function of t-tubules is regulation of cardiac EC coupling by concentrating voltage-gated L-type calcium channels (LTCCs) and positioning them in close proximity to calcium sense and release channels, ryanodine receptors (RyRs), at the junctional membrane of sarcoplasmic reticulum (jSR). The closely approximated LTCCs and RyRs form calcium releasing units, dyads, where calcium transients are initiated following each beat-to-beat action potential. With recent advances in high-resolution imaging technologies, detailed t-tubule microdomain structural organization and newly recognized functions are starting to emerge. For instance, the t-tubular invagination is not smooth, but consists of extensive membrane microfolds sculpted by a cardiac isoform of a BAR domain containing protein bridging integrator 1 (BIN1 or amphiphysin 2) (91). Cardiac BIN1 (cBIN1) organized membrane microfolds generate diffusion barriers which slow t-tubule-associated extracellular ionic diffusion, preserving electrical stability of myocytes during tachycardia-related fluctuations of ionic concentrations in the local extracellular and possibly intracellular environment (91). In addition, cBIN1-microfolds also organize local microdomains for efficient and dynamic regulation of cardiac dyad function, regulating EC coupling (66, 93, 94).

Cardiac t-tubules are also substantially remodeled during heart failure (38, 52, 89, 97, 98, 108, 130-134, 180, 202, 207, 208, 220, 224, 230). In addition to the observed loss and gross diminishment of the t-tubule network, protein components at t-tubules including ion channels and signaling molecules are reported to be either reduced at the t-tubule surface or redistributed elsewhere to non-t-tubule sarcolemma (20, 93, 134). Loss of complex t-tubules in heart failure not only causes impaired contractile function due to disrupted dyads and the resultant EC uncoupling, but also alters local concentration gradients and increases susceptibility to ventricular arrhythmia (66, 85, 91, 157, 163, 179, 181, 213). The molecular and cellular mechanisms of t-tubule remodeling in failing cardiomyocytes, in particular the alterations in t-tubule membrane ultrastructure and subdomains, is an active area of interest among cardiac biologists.

This review summarizes and cites the current knowledge of cardiac t-tubule morphology, structure, components, and physiological functions, with emphasis on the organization and function of microdomains within t-tubules. The dynamic life cycle of t-tubule membrane and protein components, as well as the current understanding of microdomain remodeling in heart failure, will also be discussed.

II. CARDIAC T-TUBULE STRUCTURE

A. History of T-Tubule Structure-Related Research

Striated muscle cells have a unique membrane system, the transverse-tubules (t-tubules), or more precisely, the transverse-axial-tubular system. T-tubules are sarcolemma invaginations formed by membrane tubules that are primarily perpendicular to myocyte longitudinal edges. These membrane tubular structures are interconnected in a complex network. Retzius first proposed the existence of t-tubules over 130 years ago, in 1881, when exploring the quick inward spread of action potentials penetrating into muscle cells (see the 1967 Croonian lecture by Huxley) (95). The first visual evidence of tubular structures around myofilament was in 1897 by light microscopy in Nystrom's study (95), when he injected mammalian heart muscle with India ink to track extracellular space. Later, in 1924, Tiegs visualized Z-lines as a spiral structures where inward spread of action potential occurs (95). However, it was not until the 1950s that Huxley and his peers reported, in various muscle cells, the structure for inward spread as tubular membrane structures at the I band (which spans the Z-line) (96). The primary element of cardiac t-tubules is transverse to the long axis of myocytes and occurs periodically at sarcomeric z-discs of myofilament (126). By using transmission electron microscopy (TEM) imaging methods in 1957, Lindner (126) identified that these transverse cardiac t-tubules are open to extracellular space. The longitudinal axial component of t-tubules was identified later in 1970s (201), adding the complexity of the t-tubule system. The 1960s and 1970s was a period of improved TEM and sample preparation technology which led to prolific t-tubule morphology research (60, 62, 153, 190, 191, 201), revealing the organization of t-tubule networks (see details in sect. IIB). The discovery of dyads and calcium releasing units at t-tubules in late 1980s and early 1990s (31, 59) led to function-driven research of cardiac t-tubules. In the past 5 years, with advance in super-resolution fluorescent microscopy and three-dimensional TEM tomography reconstruction, direct visualization into membrane ultrastructure and protein localization has provided unprecedented details of membrane microdomains within cardiac t-tubules, which will be discussed in more detail in section IIC.

Present studies of cardiac t-tubules include myocytes from the atria (109), ventricles, and conducting system (3) and are performed across different species including small murine, large mammals, and human hearts (20). Cardiac t-tubules have classic features that distinguish them from skeletal t-tubules. It is well documented that t-tubules of some form exist in all cardiomyocytes and that the t-tubule system in mammalian ventricular cardiomyocytes is the most extensive (see Figure 1 for a schematic illustration of ventricular cardiomyocyte organziation and t-tubule membrane structure occurring near z-discs). Recently identified t-tubule membrane subdomains also provide a structural foundation and clearer understanding of t-tubule regulation of cardiomyocyte function. This section reviews the gross morphology, membrane ultrastructure and microdomains, and unique features of cardiac t-tubules.

FIGURE 1.

Schematic illustration of the internal structures of an adult ventricular cardiomyocyte. T-tubules, which are enriched with voltage-gated L-type calcium channels, are positioned closely near the sarcoplasmic reticulum, the primary internal calcium store. Sarcomeres form myofibrils, which are responsible for cardiomyocyte contraction upon calcium release. The Golgi apparatus and microtubules serve as the “loading dock” and “highways,” respectively, to deliver ion channels to specific subdomains on the plasma membrane. Mitochondria provide the energy needed for the contraction of cardiomyocytes. Intercalated discs located at the longitudinal sides of each ventricular cardiomyocyte mediate the cell-to-cell propagation of action potentials.

B. Gross Morphology and Geometry

Early TEM images in glutaraldehyde-fixed cardiac tissues infiltrated with tracers such as ferritin, lanthanum salts, or horseradish peroxidase identified some of the key features of cardiac t-tubule system (60, 62, 153, 190, 191, 201) including 1) t-tubules are continuous extension of sarcolemma with openings clearly connected to extracellular space; 2) orientation of t-tubules indicates these membrane tubules are primarily transverse and perpendicular to myocyte longitudinal edges, with an axial portion running in parallel between myofilaments connecting the transverse tubules; and 3) t-tubules wrap around the z-disc and form tubular structures with the flat ends of sarcoplasmic reticulum known as terminal cisternae.

Since the 1990s, fast development of modern fluorescent light microscopy techniques has made it feasible to image cardiac t-tubules in intact cardiac tissue and live cardiomyocytes. Consistent with the earlier TEM findings, two-photon and confocal images of membrane fluorescent dye-labeled t-tubules reveal that these membrane invaginations occur orderly at every ∼1.8- to 2-μm intervals perpendicular to the long axis of cardiomyocytes, which forms a radial “spokelike” organization in transverse section of myofilament near Z-disc regions (198). Three-dimensional (3D) volume reconstruction of stepper motor acquired z-stacks of two-dimensional XY frame images acquired at incremental depth provide spatial visualization of the t-tubule gross morphology and network, allowing global analysis of t-tubule localization, size, and branching points. On the basis of studies using these imaging techniques together with electrophysiological tools, current knowledge of t-tubule morphology and geometry includes the ratio between transverse and axial elements (60 vs. 40%) (198), luminal diameter (ranges from 20 to 450 nm with an average of 200–300 nm for rat cardiomyocytes), average length between branching points (∼6.7 μm), total volume percent of cardiomyocyte (varies between 0.8% in mouse and 3.6% in rat cardiomyocytes), and percentage of membrane capacitance (varies from 21 to 64% with an average around 30%) (159, 161–163, 198).

The advent of super-resolution fluorescent microscopy imaging techniques can, in theory, improve XY spatial resolution from 250 nm with confocal microscopy to 10–20 nm, while retaining robust protein localization (227). In a study using stimulated emission depletion (STED) imaging, t-tubule structure and lumen are visualized in mouse cardiomyocytes with new structural details (220). In addition to the fluorescent microscopy, scanning ion conductance microscopy (SICM) has also been used to image live cardiomyocyte surface topology, providing visualization of openings of transverse tubular invaginations (77, 154). Meanwhile, the new two-axis tilted electron microscopy imaging in combination with three-dimensional tomography reconstruction also provides new TEM features of t-tubule ultrastructure, including the identification of variable t-tubule luminal diameter (87) with dilation near junction (228), ryanodine receptor feet at the dyad junctional membrane (43, 87), and membrane microfolds (91) (see more details in sect. IIC). The use of super-resolution microscopy helps identify LTCC and RyR localization within these microfolds (66).

C. Membrane Ultrastructure and Microdomains

The meshlike cardiac t-tubule network with interconnected transverse and longitudinal tubules is now increasingly appreciated as a nonuniform tubular network consisting of variable lumen diameters, differential membrane ultrastructure, and multiple types of frequently occurring subdomains. These membrane ultrastructure and subdomains not only add spatial complexity to t-tubule lumen, but also provide structural foundation necessary to fulfill the functional needs of t-tubules in regulation of EC coupling, extracellular ion diffusion, membrane excitability, and cellular signaling. In this section, we discuss main membrane microdomains within t-tubule membrane invaginations.

1. cBIN1 microfolds: microdomains supporting LTCC-RyR dyads

Recently, a t-tubule subdomain was found to consist of membrane microfolds sculpted by cBIN1. With the use of 3D confocal and TEM imaging, it was identified that t-tubule invaginations are actually not formed by simply straight and smooth membrane on the sides, but rather with complex and layered membrane microfolds (91). As shown in Figure 2A, low-magnification 2D TEM image (left) indicates that t-tubules are contoured tubules with calcium-dependent electron-dense precipitations along tubular membrane. High-magnification 2D TEM images and 3D tomography (Figure 2A, right panels) of transverse and axial cross sections of cardiac t-tubules reveal that redundant membrane layers and microfolds exist along tubular membrane (91). While it is intriguing to some readers that such dramatic membrane structural feature received sparse attention, reports of contoured t-tubules in mammalian cardiomyocytes have been reported for decades. As early as in 1970, Forssmann and Girardier (62) found in rat cardiomyocytes that t-tubules are extensively tortuous with hairpin-like bends (Figure 2B). Following this initial study, contoured cardiac t-tubules with extra membrane layers and microfolds can be identified in multiple cardiac t-tubule studies including a recent study from Franzini-Armstrong's group in 2015 (Figure 2C) (117), as well as the apparent variable luminal areas within t-tubules and extended membrane subdomains by super-resolution fluorescent imaging with STED (Figure 2D) (220) and stochastic reconstructive optical reconstruction microscopy (STORM) (Figure 2E) (101). By inspection, cBIN1 generated microfolds are more concentrated at the neck of t-tubules close to tubular openings at the sarcolemma (Figure 1E in Ref. 91). Removal of these microfolds in cardiac conditional Bin1-deleted adult mouse ventricular cardiomyocytes not only increases t-tubule luminal diameter, but also increases the visibility of t-tubule openings via SICM imaging of cardiomyocyte surface topology (91). While increasing evidence in support of the existence of microfolds within t-tubules starts to emerge, in our experience their fragility places an emphasis on cardiomyocyte preservation protocols for electron microscopy. We have found that, for TEM imaging (Figure 2A), freshly extracted whole mouse hearts need to be maintained with immediate perfusion on a Langendorff system and then, while the heart is still beating, the myocytes should be fixed in calcium-containing buffer introduced via the perfusion needle securing the aorta directing flow into the coronaries and retrograde through the heart. The variation in sample preparation and imaging techniques can possibly explain the time it has taken to fully appreciate the complexity of microfolds within t-tubule invaginations.

FIGURE 2.

Images of t-tubule microfolds. A: TEM images reproduced from Hong et al. (91). B: TEM images reproduced from Forssmann and Girardier (62). C: TEM image reproduced from Lavorato et al. (117), with permission from Springer. D: STED images reproduced from Wagner et al. (220). E: STORM images reproduced from Jayasinghe et al. (101).

Sculpting of microfolds within cardiac t-tubules is dependent on BIN1, which is a phospholipid-adherent membrane scaffolding protein. BIN1, alternatively known as amphiphysin 2, is a member of the BIN1-Amphiphysin-Rvs (BAR) domain containing protein superfamily, which consists of membrane deformation proteins involved in multiple cellular processes across different tissues and organs. BIN1, and in general the BAR domain protein superfamily, contributes to membrane trafficking and remodeling, cytoskeleton dynamics, DNA repair, cell cycle progression, and apoptosis (173). With 49% sequence homology (64), BIN1 and its closest cousin amphiphysin 1 both belong to the N-BAR subfamily containing an NH₂-terminal BAR domain. One of the most well-studied functions of amphiphysins is their involvement in membrane trafficking and endocytosis (9). Interestingly, BIN1 was first identified as a MYC oncoprotein interacting protein (36, 47), which serves as a tumor suppressor by inhibiting MYC function (221). Follow-up studies identified that more than 10 BIN1 protein isoforms are encoded by a single gene with 20 exons (Figure 3). All the BIN1 protein isoforms contain a conserved NH₂-terminal lipid binding BAR domain encoded by exons 1–9 (exon 7 is skipped outside of the neuronal system), an alternatively spliced linkage coiled-coil region, the MYC binding domain encoded by a ubiquitously alternatively spliced exon 17, a constitutive exon 18, and a conserved COOH-terminal exons 19–20 encoded Src homology 3 (SH3) domain critical for interaction with cytoskeleton and other intracellular protein partners. The middle linkage region is heavily alternatively spliced with tissue and organ specificity including a skeletal specific exon 11 encoding the phosphoinositide (PI) binding domain, a cardiac alternatively spliced exon 13 encoding proline-rich domains, and neuronal co-spliced exons 13–16 encoding a clathrin and AP2 binding domain (CLAP) (22, 72).

FIGURE 3.

Bin1 splice variants in adult mouse cardiomyocytes. Cartoon of Bin1 exons and the splice variants found in adult mouse cardiomyocytes, as well as brain and skeletal muscle. BAR, Bin-amphiphysin-Rvs domain; PI, phosphoinositide binding domain; CLAP, clathrin/AP2 binding region; MBD, myc binding domain; SH3, SH3 domain. For details, see Reference 89.

Cardiac t-tubule membrane microfold formation requires the cardiac isoform BIN1+13+17 with inclusion of both the cardiac alternatively spliced exon 13 and the ubiquitously alternatively spliced exon 17 (91, 222) (Figure 3). The ability of BIN1+13+17 to induce microfolds is likely through its NH₂-terminal BAR domain together with the proline-rich domains encoded by cardiac exon 13 because the NH₂-terminal BAR domains oligomerize into concave-shaped surfaces with positively charged patches crucial for interactions with membrane lipids. Through these interactions membrane is sculpted around BIN1 organized curvature. The crystal structure of the amphiphysin N-BAR domain reveals a dimeric interaction between helixes from each monomer to form a 6-helix bundle (169). With the elongated banana shape of the dimers (169) and the general function in membrane curvature formation, BIN1 has thus been referred to as the “banana” molecule (28, 79, 174). The interface between the monomers is largely hydrophobic, and the curvature of the dimer results from how the monomers intersect at highly conserved kinks (28, 225) at the end of the 6-helix bundle. These kinks are dependent on proline residues, isomerization of which represents a rate-limiting step in dimer formation (79). Domains immediately following the N-BAR in BIN1 can further strengthen its binding affinity to phospholipids for curvature formation. For instance, cardiac isoform BIN1+13+17 contains proline-rich domains encoded by the cardiac spliced exon 13. These extra proline residues likely create additional kinks in BIN1+13+17 dimers, increasing concavity of the surface of the banana molecules needed for microfold formation within t-tubules. On the other hand, inclusion of exon 17 can further separate exon 13-encoded proline-rich domains from the NH₂-terminal SH3 domain, preventing intramolecular binding between the two domains as observed in BIN1+13. As a result, the BIN1+13+17 isoform is particularly well suited to promote N-WASP (neuronal Wiskott-Aldrich syndrome protein)-dependent actin polymerization to maintain and stabilize t-tubule membrane microfolds to sarcomeric z-discs (91). Interestingly, the skeletal muscle specific splicing of exon 11 encodes highly positively charged amino acids, increasing skeletal BIN1-BAR's binding affinity to phospholipids. A recent study of exon 11 skipping in zebrafish skeletal muscle reported a dilated t-tubule pattern similar to the cardiac phenotype, indicating a similar function of exon 11 containing skeletal BIN1 in organizing microdomains in skeletal t-tubules (194).

Together with the previously identified ability of BIN1 in localizing and clustering LTCCs to cardiac t-tubules, these cardiac BIN1+13+17 created microfolds are likely the membrane microdomains supporting cardiac dyad (LTCC-RyR) formation. The role of BIN1 in facilitating microtubule-dependent targeted delivery of Ca_V1.2 channels to t-tubule membrane (94) will be discussed in section III. Confocal imaging and biochemical results identified that Cav1.2 is localized to BIN1 molecules at cardiac t-tubules. BIN1 decreases occur in human (93) and animal models of heart failure (23, 134), which affects LTCC surface expression at the t-tubules of failing cardiomyocytes. New super-resolution STORM imaging further identified that BIN1-microfolds are not only enriched with Cav1.2 channel clusters but also attract RyRs from the jSR membrane (see supplemental movie on the Physiological Reviews website), further supporting the role of BIN1 in organizing cardiac dyads (65, 66). Thus BIN1-induced t-tubule microfolds create local microdomains bringing sarcolemmal LTCCs to couple with RyRs from jSR membrane. BIN1-regulated coupling of LTCCs to RyRs also helps explain previous observed allosteric activation of RyR following calcium channel activation independent of Ca²⁺ influx (73). In summary, BIN1 microdomains provide a physical understanding of how cardiac calcium releasing units are organized (Figure 4). More detailed discussion into the multimodal role of BIN1 in formation of LTCC-RyR dyads and extracellular ion diffusion will be further discussed in sections IV and V.

FIGURE 4.

Schematic illustration of BIN1-microdomain organization within cardiac t-tubules. Top: the “banana”-shaped molecule cBIN1 regulates cardiac t-tubule function through: 1) facilitating microtubule-dependent forward trafficking of Cav1.2 channels (LTCC, L-type calcium channels) to cBIN1 organized membrane microfolds at t-tubules (targeted delivery); 2) clustering of LTCCs and RyRs at cBIN1-microfolds based microdomains; 3) creating extracellular ion slow diffusion zone within t-tubule lumen; 4) organizing microdomains for dyad formation and regulation by β-adrenergic signaling. Bottom: cartoon of mammalian adult ventricular cardiomyocytes, which are rod-shaped striated muscle cells with t-tubule invaginations occurring periodically near z-discs of myofilaments.

2. Caveoli (caveolae): β-AR/LTCC signaling hubs

The other well-identified membrane subdomains are caveoli, which are flask-shaped membrane invaginations with diameter typically around 90 nm (123). In muscle cells, caveoli are detergent-resistant membrane structures organized by the plasma membrane binding protein caveolin-3, rich in cholesterol and sphingomyelin. Caveoli are found to be attached to both general sarcolemma (0.04 μm² membrane area/μm³ cell volume) and t-tubule membrane (0.03 μm² membrane area /μm³ cell volume) in cardiomyocytes (159), increasing total plasma membrane area by 14–21% (159). Within t-tubules of cardiomyocytes, caveoli cluster at necks of the t-tubules in the subsarcolemmal space (151, 238). Attachment of caveolae to t-tubules not only increases the spatial complexity of t-tubule lumen, but also differentially concentrates membrane ion channels and signaling molecules for localized regulation at these membrane subdomains. In cardiomyocytes, caveolae can house ion channels HCN4, Cav1.2, Kv1.5, Kir6.2/Sur2a, Nav1.5, and NCX1, and signaling molecules such as β-adrenergic receptors (β-AR). For example, caveolin-3 is critical in concentrating β₂-AR-cAMP signaling to the t-tubules (229). A subpopulation of LTCCs at caveolae, which is under the regulation of β₂-AR stimulation (14), is reported to be important in hypertrophic signaling but not muscle contraction (137). Thus caveolae have a critical role in myocyte function by creating microdomains with compartmentalized specific lipid species, membrane proteins, and signaling molecules, facilitating functional regulation.

The membrane scaffolding proteins responsible for caveolae formation are caveolins, which are a family of proteins with binding affinity to plasma membrane lipids particularly cholesterol and sphingomyelin. There are three caveolin isoforms, caveolin-1, caveolin-2, and caveolin-3, with caveolin-3 specific to the muscle cells (200). Caveolin-3 has been indicated in muscle t-tubule development and is associated with developing skeletal t-tubules, while knockout of cavolin-3 can induce abnormal t-tubules in skeletal muscle (70). Furthermore, as caveolin-3 organized caveloae is critical in compartmentalized regulation of ion channels, mutations in caveolin-3 are associated with channelopathies including long QT syndrome (5).

3. Ankyrin B microdomains: NCX/NKA/InsP₃ macromolecular complex

In addition to BIN1-microfolds and caveoli, a third set of microdomains organized by the membrane scaffolding protein ankyrin B has been identified to be essential in the formation of the macromolecular complex of sodium calcium exchanger (NCX), sodium-potassium-ATPase (NKA), and inositol trisphosphate receptor (InsP₃) (147). Distinct from the dyad macromolecular complex, the NCX/NKA/InsP₃ complex is critical in removing calcium from the luminal clefts between the t-tubule and jSR membrane, facilitating calcium decline during muscle relaxation (147). Therefore, by balancing calcium entry through LTCC-RyR dyads and calcium removal through NCX anchored at ankyrin B-microdomains, intracellular calcium equilibrium can be achieved to maintain coordinated beat-to-beat heart contraction.

The ankyrin B membrane microdomains are formed by ankyrin B protein, which belongs to a family of intracellular proteins with binding affinity to actin and spectrins. Three separate genes ANK1, ANK2, and ANK3 encode ankyrin R, ankyrin B, and ankyrin G, respectively, that are ubiquitously expressed with multiple alternatively spliced forms of each ankyrin. However, cardiomyocytes express three ankyrin B isoforms including the ankyrin B-220 kDa and two recently identified isoforms ankyrin B-188 kDa and ankyrin B-212 kDa (231), and one 190-kDa ankyrin G, with distinct functions in organizing NCX/NKA/InsP₃ complex at t-tubules or trafficking Nav1.5 to intercalated discs, respectively [see more details in reviews (39, 146)]. A loss-of-function mutation in human ankyrin B as well as heterozygous ankyrin B knockout mouse are associated with cardiac features including sinus bradycardia, abnormal heart rate variability, defects in cardiac conduction, and increased ventricular arrhythmia (148). The ankyrin B syndromes are distinct from classical long QT syndrome, but can manifest as sick sinus syndrome with bradycardia and increased risk of sudden cardiac death. Results from cell biological studies indicate that lack of ankyrin B microdomains alters membrane ion channel localization to their physiological subdomains, disrupting intracellular calcium homeostasis and promoting arrhythmias (24, 146, 148, 149).

The composition, organization, function, regulation, and turnover of these t-tubule-related membrane microdomains remain active areas of research. Further understanding of microdomain content and regulation may provide new insight into the biological functions of t-tubules in striated muscles, both in normal physiology and during disease development.

D. Peri-Membrane Cytoskeleton Scaffolds at Cardiac T-Tubules

T-tubule membrane is supported by a unique set of peri-membrane scaffolds formed by both extracellular cell adhesion molecules and intracellular membrane scaffolding proteins tethered through cortical cytoskeleton to costamere and sarcomere structural proteins. Different from skeletal t-tubules, cardiac t-tubule membrane is anchored to the Z-discs of sarcomeres. In this section, peri-t-tubule membrane scaffolds and cytoskeleton as well as the t-tubule membrane anchors, including F-actin/N-WASP/tropomysin, costamere dystrophin, and t-tubule membrane-associated dysferlin, will be discussed.

1. Actin filaments, tropomyosin, and N-WASP

Actin filaments have been indicated in the organization and maintenance of t-tubule structure and function. Stabilization of cardiac actin filament by cytochalasin D preserves t-tubules in rodent ventricular cardiomyocytes during extended in vitro culture, which normally loses t-tubule invaginations (91, 118, 214). Actin depolymerization by latrunculin A, on the other hand, disrupts membrane microfolds at t-tubules (91). The integrity of actin filaments is heavily regulated by many actin-regulating proteins including G-actin monomer binding proteins, nucleating proteins, depolymerizing proteins, capping proteins, and polymerizing proteins. These regulatory proteins are important in the organization and function of the myocyte contractile apparatus, cardiac sarcomeres that are primarily composed of cardiac actin and myosin. For example, muscle specific tropomyosin isoforms are associated with sarcomeric actins to form thin filaments, responsible for myocyte contraction following intracellular calcium elevation. Interestingly, in addition to sarcomere organization, recent studies indicate that nonsarcomeric actin is involved in vesicle trafficking (197), and a non-muscle tropomyosin (Tm5NM1) localized adjacent to Z-line can define actin filaments that are associated with t-tubules, contributing to t-tubule organization and function (219).

Furthermore, BIN1+13+17 facilitates N-WASP-dependent actin polymerization, responsible for the maintenance of BIN1-microfolds at t-tubules. N-WASP is a ubiquitously expressed protein involved in transduction of signals from receptors to actin cytoskeleton. N-WASP is bound to the actin nucleator Arp2/3 protein complex, and remains inactive through autoinhibition. Upon activation by small GTPase Cdc42, PIP₂, amphiphysin 1, or BIN1, the functional VCA domain of N-WASP is released from autoinhibition, which subsequently activates Arp2/3-dependent actin polymerization (203). In cardiomyocytes, N-WASP is localized at the z-discs (211). When interacting with t-tubule membrane associated BIN1+13+17, N-WASP activation promotes F-actin polymer formation, stabilizing BIN1+13+17-membrane microfolds to z-discs (91).

2. Costamere complex and t-tubule membrane-associated scaffolds

Costameres are cytoskeleton-rich structures surrounding the sarcomere Z-discs right below t-tubule membrane, serving as anchor points between sarcolemma and myofilament. Costameres consist of two major protein complexes, the dystrophin-glycoprotein (DAG) complex and the vinculin-talin-integrin complex. The vinculin-talin-integrin complex is involved in the bidirectional transmission of contractile force between cardiomyocytes and extracellular matrix, with vinculin as the protein linking membrane and actin cytoskeleton. By clustering with vinculin complex (106), dystrophin also has a primary mechanical role in maintenance of cell membrane integrity. Dystrophin is a large 427-kDa protein product encoded by the Duchenne muscular dystrophy gene. In addition to the full-length dystrophin Dp427, a short isoform Dp71 is also expressed in mouse cardiomyocytes (141). Dystrophin is present at cardiac t-tubules from the time tubules develop. Dp427 is found in both general sarcolemma and t-tubules, whereas Dp71 is specifically expressed in t-tubules (58). The dystrophin-DAG complex provides a structural link between cytoskeleton and extracellular matrix. Sarcolemma fragility secondary to dystrophin degradation has been observed in dilated cardiomyopathy (107).

As discussed in other sections, t-tubule membrane-associated scaffolds such as BIN1 (sect. IIC), caveolin-3 (sect. IIC), and junctophilin (sect. VA) are important to maintain normal cardiac t-tubule structure and function. Additionally, another membrane-inserted protein, dysferlin, is also understood to be an important regulator of t-tubule membrane trafficking and repair during injury (6, 7, 46, 54). Dysferlin is a 230-kDa membrane inserted protein with binding affinity to both calcium and phospholipid (44), which is enriched in subsarcolemma vesicles and can be quickly recruited to the membrane injury site (7) to facilitate membrane repair (54). In cardiomyocytes, dysferlin is found to be present in both sarcolemma t-tubules and intercalated discs (30). For t-tubule membrane repair, it has been proposed that upon insult, dysferlin can be recruited to t-tubules along with its interacting partners including BIN1, caveolin-3, and EHDs to induce repair (46). Without dysferlin, abnormal lipid vesicle trafficking impairs membrane repair and accumulates neutral lipids like glycerol, which further promotes detubulation and myopathy (46). Dysferlin deficiency is associated with dilated cardiomyopathy (84, 114, 226), further indicating its important role in maintenance of normal cardiac t-tubule function.

3. Myofilament structural proteins

The Z-disc organizing protein α-actinin, with the cardiac isoform being α-actinin 2, is the structural protein at both Z-disks and peri-Z-disc costameres. Using F-actin filament, α-actinin is linked to the t-tubule membrane-associated cardiac BIN1+13+17 for the stabilization of t-tubule microdomains around myofilament Z-discs. Whether sarcomere α-actin, costameric γ-actin, and/or cortical β-actin fibers are involved in α-actinin-dependent microfold stabilization to Z-discs remains unidentified. Tcap is another myofilament aligning protein (86, 111) that is critical in the maintenance of the integrity of t-tubule network. Mutations of Tcap have been associated with abnormal t-tubule morphology (111) and development of hypertrophic and dilated cardiomyopathies.

Supported by these numerous membrane-associated proteins, scaffolding proteins, cytoskeleton, costamere linking proteins, sarcomere structural proteins, and extracellular matrix, t-tubules are well-organized rigid membrane organelle with limited fluidity. Yet, t-tubule membrane system can also be extremely dynamic and labile (see more detailed discussion in sect. V). It remains to be understood how these regulatory proteins work with each other to maintain a rigid yet dynamic membrane system.

E. Variation in Cardiac T-Tubule Geometry

A high degree of variability in t-tubule geometry has been reported in myocytes from different cardiac chambers and across species. T-tubules have been found in cardiomyocytes from all mammalian species studied so for (mouse, rat, rabbit, dog, sheep, pig, human), but are inconsistently reported or devoid from other phyla including amphibians, reptiles, and birds (see review in Ref. 19). Across mammalian species, it is well accepted that ventricular cardiomyocytes have the most developed and mature t-tubule system (see reviews in Refs. 19, 20, 80), whereas atrial myocytes have more scarce and irregular t-tubules (109, 121, 175, 215). Occasionally, t-tubules are identified to be present in the Purkinje conducting system (51, 160). In mammalian ventricular cardiomyocytes, although the Z-line localization of transverse tubules remains conserved across species, large variability is found in the geometry and size of t-tubules. It appears that t-tubules in rodents and small mammals are denser, deeper, and narrower with more structural complexity than t-tubules from large mammals (i.e., rabbit, pig, and human). Quantitative measurements of overall t-tubule morphology and spacing, based on traditional fluorescent microscopy which cannot discern membrane microfolds but is excellent and quantifying the larger structures, estimate segment branch length (L) to be ∼9 μm and diameter (D) to be ∼200 nm in mouse (91), ∼7 μm (L) and ∼250 nm (D) in rat (198), and ∼2 μm (L) and ∼400 nm (D) in rabbit (183) and human (26) myocytes. The size and complexity of t-tubule network seem to correlate with the heart rate of each species (20). Conceptually, higher heart rates as in rodents require faster action potential propagation into cell interior, allowing for efficient intracellular Ca²⁺ diffusion to SR to maintain the synchronicity and efficacy of beat-to-beat myofilament contraction (16, 20, 237). Humans and large mammals have resting heart rates less than 100 beats/min, so relatively larger and more widely spaced t-tubules (20, 90) can fulfill the same functional needs to maintain synchronized ventricular contraction during each heartbeat.

Although a t-tubule network is a membrane structure common to the striated muscles, cardiac t-tubules have unique characteristics that are distinct with regard to tubular size, location, and numerical count. Compared with the small skeletal tubules with diameter averaged at 20–40 nm, cardiac t-tubules are much larger with estimated diameter to be 100–400 nm depending on species. Meanwhile, cardiac t-tubules are present within 0.5 μm of Z-discs (198), whereas skeletal t-tubules are localized to the interface between A band and I band of the sarcomere. Thus cardiac t-tubules appear once every sarcomere length spaced out at ∼1.8–2 μm along the longitudinal edge of the myocytes. The skeletal t-tubules, on the other hand, consist of two sets of transverse tubules at each A band/I band interface within individual sarcomeres of skeletal myofilaments.

III. TRANSMEMBRANE PROTEINS AT CARDIAC T-TUBULES

Continuously extended from surface sarcolemma, t-tubules are lipid bilayers embedded with transmembrane or lipid-associated proteins. Studies indicate t-tubules have a different lipid composition relative to general sarcolemma. Previous studies from skeletal muscle identified that t-tubules are enriched in cholesterol (27) and phospholipids, and have a higher phospholipids/cholesterol ratio (167, 178, 204), with phosphotidylserine and sphingomyelin as the two most abundant phospholipid species. The enrichment of phospholipids may be enhanced by the BIN1 organized membrane microfolds which preferentially bind to and concentrate negatively charged phosphoinositides (PIP₂) (120). The full lipid profile of cardiac t-tubules awaits careful lipidomic studies. In contrast, the protein components of t-tubules are well studied. This section will focus on the current knowledge of the transmembrane protein composition at t-tubule membrane, focusing on ion handling proteins and signaling molecules.

A. Ion Handling Proteins (Channels/Transporters/Pumps)

The expression of transmembrane ion channels, ion transporters, and pumps have been well characterized in cardiac t-tubules. The calcium handling proteins that are important in cardiac EC coupling, in particular the LTCCs, are mostly enriched in t-tubules. Other Na⁺ and K⁺ channels and handling proteins are also found to be present in t-tubules, although to a lesser degree of enrichment. Electrophysiology studies in normal and osmotic shock detubulated cardiomyocytes identified the relative amount of membrane currents at t-tubule membrane versus non-t-tubule sarcolemma (see review in Ref. 19).

1. Ca²⁺ handling proteins

The most cardiac t-tubule “signature” protein is the calcium transient initiator LTCC, consisting of one α subunit with multiple auxiliary β, α₂δ₁, and γ subunits. The I_Ca conducting pore is formed by the large 24 transmembrane domain containing an α-subunit, whose trafficking and gating properties are regulated by the auxiliary subunits (42). The predominant splice variant of α-subunit expressed in cardiomyocytes is α_1c, or Ca_V1.2 (212). Ca_V1.2 LTCCs are critical in cardiac development and function (186). Global deletion of Ca_V1.2 induces abnormal cardiac morphogenesis with embryonic lethality (186). On the other hand, in postdevelopment adult hearts, inducible deletion of Ca_V1.2 leads to reduced contractility and increased susceptibility to stress-induced heart failure (76). As the cell surface dyadic channel, LTCCs need to be concentrated to t-tubules for better coupling with RyRs from jSR membrane, allowing efficient CICR and normal EC coupling (163). Immunocytochemistry data indicate that LTCCs are concentrated at t-tubules, and electrophysiology measurement identified that 80% of the surface calcium current mediated by LTCC is at t-tubules. Such a high concentration of LTCCs at t-tubule membrane is achieved by microtubule-dependent targeted delivery of LTCCs to t-tubules, a process facilitated by the t-tubule membrane curvature protein BIN1 (94) (Figure 4). Furthermore, LTCCs at t-tubules are not distributed evenly, but rather clustered to membrane microdomains formed by BIN1-microfolds. These LTCCs clusters will couple with RyRs channels on the nearby jSR membrane to form functional calcium releasing units bridged by BIN1 molecules (Figure 4). These TEM visible electron-dense feetlike structures, dyads, are the initiation sites where beat-to-beat calcium transient starts. During each action potential, membrane depolarization activates voltage-gated LTCCs to induce calcium influx. The initial calcium ions entered through LTCCs at the junctional dyads diffuse a short distance to activate the nearby calcium sensing and releasing RyRs at jSR membrane, inducing a large release of calcium from SR lumen to cytosol. This is a sequence of events known as calcium-induced calcium release (CICR). How LTCC-RyR dyads are organized by BIN1-microfolds to control CICR and EC coupling (66) will be further discussed in section IV focusing on the functional regulation of dyads.

During relaxation, distal SR membrane localized sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) reuptake elevated cytosol calcium into SR storage. Meanwhile, the elevated intracellular calcium at the cleft between t-tubule and SR membrane will also be extruded through sarcolemmal NCX1, a calcium handling protein anchored at ankyrin-B microdomains within t-tubules that is critical for intracellular calcium homeostasis.

2. Na⁺ handling proteins

In most excitable cells, the voltage-gated sodium channel-mediated Na⁺ current is responsible for the rapid action potential upstroke. Sodium channels consist of four pore-forming α-subunits and one or two auxiliary β-subunits. Ventricular cardiomyocytes express several different pore-forming subunits of the voltage-gated Na⁺ channels including the primary cardiac isoform Nav1.5 and the brain-type isoforms Nav1.1, Nav1.3, and Nav1.6 (136), with different intracellular localization and functional importance. Sodium current measured in control and detubulated rat ventricular cardiomyocytes indicate more uniform distribution of cardiac Nav1.5 current along sarcolemma and t-tubules (in channels/μm², 11 on total sarcolemma vs. 10 on t-tubules), and a significant t-tubule concentration of current mediated by neuronal sodium channel isoforms (in channels/μm², 1 on total sarcolemma vs. 2.5 on t-tubules).

Although Nav1.5 is also localized to cardiac t-tubules (55), this primary cardiac α-subunit Nav1.5 is preferentially localized to intercalated discs with auxiliary β₂- and/or β₄-subunits, mediating rapid action potential propagation from cell to cell across the myocardium. In contrast, the brain isoforms of Nav1.1, Nav1.3, and Nav1.6 together with β₁- and/or β₃-subunits, are localized to t-tubules (135) and near t-tubule openings (56). These t-tubule localized brain Nav isoforms have more negative voltage dependence of gating and more rapid activation and inactivation kinetics, facilitating fast action potential inward penetration from myocyte surface to interior via t-tubule invagination. However, the role of these channels in EC coupling remains controversial. T-tubule localized neuronal isoforms of sodium channels have been indicated as contributors to more effective CICR in some studies (136), and removal of Nav1.6 has been shown to prolong calcium transient duration (155). Other studies indicate that, although these neuronal sodium channels are concentrated at cardiac t-tubules, they are not required for EC coupling (21).

NKA uses ATP hydrolysis-derived energy to pump 3 Na⁺ and 2 K⁺ ions across membrane, hence maintaining membrane potential in cardiomyocytes. Three catalytic α (α₁, α₂, α₃) subunits of NKA are expressed in the heart, with α₁ as the dominant isoform (206). However, the α₁ isoform is more uniformly expressed, whereas isoforms α₂ and α₃ are mainly located in the junctional membrane within t-tubules (105). Studies have shown that isoform α₂ expression is four times higher in t-tubules versus non-t-tubule sarcolemma (10). Within t-tubules, NKA colocalizes with NCX1 at ankyrin B microdomains, contributing to the regulation of CICR and calcium transient. NKA and NCX1 and the jSR localized InsP₃ receptors are clustered to a junctional subdomain organized by a membrane scaffolding protein ankyrin B. Distinct from BIN1-organzied LTCC-RyR dyads, ankryin B similarly tethers cytoskeleton to facilitate targeted delivery of NCX1 and NKA (147), resulting in the concentrated expression of the functional NKA and NCX1 at t-tubules (∼3- to 3.5-fold higher than external sarcolemma) (48).

3. K⁺ handling proteins

Most potassium channels are more uniformly distributed between t-tubules and surface sarcolemma in cardiomyocytes. Detubulation of cardiomyocytes only decreases the steady-state current I_ss [mediated by Kv1.2/2.1 (184) or TASK-1 (104)], without affecting the density of other potassium currents including transient outward current I_to (mediated by Kv4.2/Kv4.3), inward rectifier current I_k1 (mediated by Kir2.x), and delayed rectifier current I_k (I_kr mediated by HERG, and I_ks mediated by KvLQT1) (113). Interestingly, immunofluorescence and TEM imaging data demonstrated that the I_to generating Kv4.2 is concentrated at both t-tubules and intercalated discs (8). The lack of current density change after detubulation may be due to the I_to current at the intercalated discs. Antibody labeling also supports that I_k1 generating Kir2.x, including Kir2.1 (35), Kir2.2 (122), and Kir2.3 (143) channels, are localized to t-tubules. Other ATP-sensitive inward rectifiers (Kir6.x) are also localized to t-tubules. Both I_KATP and I_k1 decrease along with t-tubule membrane loss when cardiomyocytes are cultured in vitro (34, 145). Interestingly, K_ATP channel release from the Golgi can be dependent on stress-induced PKA activation, resulting in rapid insertion of K_ATP channels into t-tubule membrane (1). If the cargo is in the close proximity to the membrane, there may a direct vesicle-mediated Golgi to membrane translocation pathway (172). Immunofluorescence data also indicate that the cardiac I_kr channels, which consist of two α-subunits hERG1a and hERG1b, are localized to t-tubules (103).

B. Signaling Molecules at Cardiac T-Tubules

The most well-studied signaling pathway at t-tubules is the G protein-coupled β-adrenergic receptor (β-AR) pathway, which mediates the sympathetic control of myocardial function. Cardiomyocytes express both β₁-AR and β₂-AR, which are coupled to the Gs-AC-cAMP cascade (see review in Ref. 233). Through the activation of cAMP-dependent protein kinase A (PKA) and the subsequent phosphorylation by PKA of critical proteins involved in calcium handling and contractile function, β-ARs modulate catecholamine-dependent rate, force, and speed of myocyte contraction. The protein targets of β-AR-activated PKA include sarcolemmal t-tubule LTCCs (67, 68), SR membrane RyRs (139) and phospholamban (PLB) (125), and intracellular myofilament proteins (troponin I and C proteins) (12). In addition to the stimulatory G_s pathway which β₁-AR is solely coupled to, β₂-AR is also known to couple to the inhibitory G_i protein (234, 235). Thus, within β₂-AR-enriched caveolae (14, 229) and t-tubule (77), the β₂-AR to G_i pathway can negate β₂-AR-G_s mediated cAMP-PKA signaling.

In the t-tubule system, several factors affect the compartmentalization of the β-AR signaling. First, the differential distribution pattern of β₁-AR and β₂-AR reveals that β₁-AR is homogeneously spread throughout sarcolemma and t-tubules, whereas β₂-AR is more selectively confined to t-tubules (77, 154). Second, LTCCs, the ion channel targets of G_s-activated PKA, are concentrated to t-tubule membrane with particular clustering to BIN1-microfolds (94). Third, a number of signaling molecules in the β-AR-Gs-PKA cascade are also localized to t-tubules, including phosphodiesterase 3 (PDE3), PKA, AKAPs, and AC. The β_1/2-ARs mediated G_s-cAMP signaling at t-tubules is therefore more confined to subsarcolemmal microenvironment right below the t-tubule openings, where most BIN1-microfold clustered dyadic LTCCs and RyRs are also enriched. The structural restriction of AKAP on PKA diffusion (78, 110) further confines β-AR-G_s-PKA signals close to submembrane microdomains to more effectively modulate the phosphorylation and activity of surface LTCCs without affecting intracellular protein targets. Compartmentalization of β-AR signaling to cardiac dyads is also linked by cBIN1-microdomains, which both organize LTCC-RyR dyads and undergo fast reassembly upon acute β-adrenergic activation (66).

Other than the secondary messenger cAMP-dependent function in contractile regulation, β-AR signaling pathway also contributes to regulation in apoptosis and cell survival. The β_1/2-AR-cAMP-PKA pathway stimulates apoptosis by activating MAPK family, particularly ERK1 and ERK2 (32, 41, 241). The antiapoptotic effect of β₂-AR activation, on the other hand, is more likely mediated through the Gβγ subunits, which are released from G_i upon β₂-AR activation by catecholamine (240). Thus t-tubule localized β₂-AR also signal through G_i-Gβγ pathway to activate the pro-survival PI3K and AKT signals (240).

Given the differential role of β₁-AR/β₂-AR in regulating cell contractile function and apoptosis/cell survival, t-tubule-dependent-compartmentalized β₂-ARs play a critical role in regulating these signaling pathways to maintain normal cardiac function. In fact, redistribution of β₂-AR with altered β₂-AR-cAMP signaling occurs during heart failure (77, 134, 154), indicating the importance of t-tubule compartmentalized β₂-AR-cAMP signaling in normal cardiac function and during disease development.

IV. FUNCTION OF CARDIAC T-TUBULES

Exploration of the functional importance of t-tubules in cardiac calcium transients has been a very active field of research since late 1980s. The dyad membrane structure formed by t-tubules and jSR terminal cisternae is critical in organizing EC coupling of heart muscle. In 1989, Fabiato (59) proposed the model of CICR as the molecular mechanism underlying EC coupling, which is further supported by the later identification of calcium sparks as the elementary events (31). The sarcolemma voltage-gated LTCCs, which are activated during membrane depolarization following each action potential, conduct the initial calcium influx to induce the subsequent series of events of intracellular CICR. The intracellular calcium sense and release units are RyRs on the jSR membrane. In addition to dyad formation for efficient CICR and normal EC coupling, t-tubules also create a diffusion barrier for extracellular ions, protecting electrical stability of the cardiomyocytes during fluctuations in extracellular bulk environment. The recently identified BIN1 microdomains contribute to both dyad microdomain formation (65, 94) and generation of a “fuzzy space” (119) restricted diffusion zone for extracellular ions, thus serving as a major regulator of the primary functions of cardiac t-tubules (Figure 2).

A. Dyad and EC Coupling

Normal ventricular EC coupling requires a calcium transient initiated at the cardiac t-tubules. As discussed earlier, the current accepted model of intracellular calcium transient development is CICR (59). At each heartbeat, in response to action potential-triggered inward sodium current, sarcolemmal membrane depolarization activates voltage-gated LTCCs at the t-tubule membrane. The initial calcium influx mediated by activated LTCC subsequently induces a massive calcium release from SR stores. Proposed as early as in late 1980s, this CICR model was further supplemented by the identification of the jSR membrane localized RyRs (100, 168). In the early 1990s, calcium sparks from the SR were then identified as the “elementary units” of the calcium transient (31). The complexes consisting of LTCCs at t-tubules and RyRs at jSR membrane (∼1:4 ratio) form cardiac dyads (13, 25). Close physical association of jSR membrane localized RyRs with sarcolemma LTCCs (31, 59, 100, 168) is required for optimal CICR needed for efficient EC coupling. Such a close association between the two dyadic channels is achieved by LTCC localization to t-tubules (11, 185) and clustering to microdomainds induced by BIN1 microfolds, which also recruit RyRs to jSR membrane for LTCC-RyR couplon formation (Figure 3). Upon membrane depolarization, the initial calcium influx through LTCCs and the close association between LTCCs and RyRs (∼15 nm) at dyads permit efficient CICR and subsequent sarcomeric contraction (156).

1. LTCC trafficking and clustering to dyads through BIN1

The calcium transient initiator LTCCs need to be concentrated and clustered to t-tubule microdomains for better coupling with RyRs to form functional dyads (163). Precise t-tubule localization of LTCCs relies on the membrane scaffolding protein BIN1 for several reasons. First of all, LTCC trafficking from the Golgi apparatus occurs on microtubules which deliver LTCCs to BIN1 containing membrane (94). This microtubule-dependent channel delivery highway with specificity, known as targeted delivery, was introduced via previous studies using connexin 43 trafficking to intercalated discs as a model system (93, 94, 188, 195). Briefly, targeted delivery describes that, after exiting the Golgi, membrane vesicles containing transmembrane ion channels are rapidly directed on microtubule highways across the cytoplasm to their specific membrane microdomains (Figure 5). The microtubules with negative ends are anchored at the microtubule organizing center (MTOC) near Golgi, and the fast growing positive ends are approaching outward ready to be captured by membrane anchor protein complexes at specific plasma membrane subdomains. Specificity of delivery requires a combination of the individual channel protein or truncated isoform (Figure 6), the plus-end-tracking proteins at the growing ends of microtubules, and the membrane anchor protein complex responsible for capturing. In the case of Cav1.2 channel delivery, t-tubule-associated BIN1 anchors growing microtubules with Cav1.2 cargoes, allowing the delivery of Cav1.2 to occur at BIN1 molecules (94). BIN1 contains a membrane curvature BAR domain, the middle coiled-coil regions, and an SH3 protein-protein interaction domain (see sect. II). Compelling data are that deletion of the coiled-coil and SH3 domains abrogates the ability of BIN1 in facilitating Cav1.2 trafficking without altering membrane invagination (94). Thus targeted delivery of Cav1.2 is achieved through interaction with the BIN1 molecules at t-tubules, rather than the t-tubule structures themselves. The non-BAR domains in BIN1 molecules likely attract Cav1.2 channels through interaction with the COOH terminus of the channel. The involvement of BIN1 in trafficking Cav1.2 is further confirmed in a series of subsequent studies using BIN1-deficient ventricular cardiomyocytes (91, 93). Alternative translation of connexin 43 provides smaller isoforms (Figure 6) that assist with forward trafficking of the full-length protein (196). Fragments of the LTCC COOH terminus occur and have been attributed to cleavage (69). It is not known if these fragments could be formed by alternative translation instead, if whether they assist with forward trafficking.

FIGURE 5.

Schematic representation of ion channel targeted delivery. Channel proteins are sorted into vesicular carriers and docked onto microtubules at the trans-Golgi network (TGN) and subsequently delivered to their subcellular destinations in cooperation with actin “rest stops” along the route. Microtubule plus-end binding proteins interact with anchor proteins of specific membrane subdomains, allowing targeted delivery of cargo proteins. In the case of connexin 43 (Cx43) trafficking, the interaction between the microtubule plus-end binding protein EB1, the channel itself, and the adherens junction complex ensures the targeted delivery of Cx43 hemichannels to the intercalated discs. For LTCC delivery to t-tubules, key components are the LTCC channels, a +TIP protein, and cardiac bridging integrator 1 (cBIN1). The microtubule +TIP protein associated with LTCC delivery has not yet been identified.

FIGURE 6.

Alternative translation of connexin 43. In addition to full-length 43-kDa connexin 43 (GJA1-43K), there are six NH₂-terminal truncated smaller Cx43 isoforms (GJA1-32K, GJA1-29K, GJA1-26K, GJA1-20K, GJA1-11K, and GJA1-7K) resulting from internal methionine residues that initiate ribosomal translation, a process known as alternative translation. At least one of the isoform, GJA1-20k, has been identified as necessary for the full-length channel to traffick to the surface membrane.

In addition to facilitate forward trafficking of Cav1.2 channels, BIN1 also plays a critical role in clustering Cav1.2 channels that are already localized at the t-tubule membrane. Through formation of BIN1 microfolds, these unique membrane structures form curved membrane microdomains enriched with Cav1.2 channels. As indicated in our recent super-resolution STORM imaging (XY-resolution at 10–20 nm, Z-resolution at 50 nm), Cav1.2 form discrete clusters at cuplike microdomains created by BIN1 microfolds. The role of BIN1 in clustering Cav1.2 channels is further supported by the observed smaller Cav1.2 clusters in adult mouse ventricular cardiomyocytes with Bin1 heterozygous deletion (91). Clustering of Cav1.2 channels at these t-tubule microfolds alters channel kinetic properties including gating and open probability (53, 152). How BIN1 microdomains regulated Cav1.2 clustering alters coupled gating of LTCCs at the t-tubule surface is an interesting and important area of future consideration.

2. RyR recruitment to dyads through BIN1

The other critical dyadic component is the RyR homotetramer at the jSR membrane apposing LTCCs at t-tubules. In the RyR tetramer, the COOH-terminal transmembrane pore domain of each RyR monomer comes together to form a calcium-conducting pore across the SR membrane. The four NH₂-terminal cytoplasmic domains from the tetramer constitute a large mushroomlike platform containing binding sites (182) for various modulators including calstabin (216), PKA and mAKAP (muscle A kinase anchoring protein), phosphatases PP1 and PP2A (139, 140), as well as calmodulin (236) and Ca²⁺/calmodulin kinase II (CaMKII) (223). Biophysical properties of RyR can be regulated by these modulators via alteration in receptor posttranslational modifications including phosphorylation states, luminal calcium (81), or stabilization of receptor closed states. The macromolecular complex centering on RyR channels facilitates efficient spatial and timely regulation of channel function.

It is important to note that the arrangement of RyR at dyad is neither uniform nor static. The receptor cluster contains 10–300 RyR homotetromers (63) with a spatial separation of 0.6–1 μm between clusters (199), the organization of which can be altered on a time scale of minutes by Mg²⁺ concentration, receptor phosphorylation, as well as reorganization of BIN1 microdomains (2). It was recently identified that t-tubule-attached BIN1 recruits RyRs into dyads, an interaction that is enhanced after RyR phosphorylation (66). The enrichment of positively charged residues within the BAR domain of BIN1 may explain its preferential binding to negatively charged hyperphosphorylated receptors, in a way similar to the known binding affinity of BIN1 to negatively charged phospholipids. Interestingly, upon acute β-AR activation, BIN1 is redistributed to t-tubules, a result likely due to β-AR activation-induced phospholipid accumulation at the t-tubule lipid bilayer. Subsequently, phosphorylated RyRs are further recruited to BIN1 microdomains, allowing efficient RyR coupling with LTCC for functional dyad formation, improving EC coupling with preserved electrical stability during acute stress (Figure 7) (66).

FIGURE 7.

Cartoon of BIN1-dependent recruitment of phosphorylated ryanodine receptors (RyR) to dyads during acute stress. Acute isoproterenol (ISO) redistributes BIN1 to cardiac t-tubules and subsequently recruits phosphorylated RyRs to couple with LTCCs at dyads, increasing excitation-contraction coupling gain while maintaining electrical stability. In Bin1 HT cardiomyocytes with less BIN1 available, insufficient RyR recruitment leads to accumulation of hyperactive phosphorylated RyRs outside of dyads, increasing spontaneous calcium release and promoting arrhythmias.

3. Junctophilin-2 in t-tubule-jSR junctional membrane complex formation

Another well-studied protein family involved in t-tubule and jSR organization is the junctophilins (JPs), which have the apparently unique ability to directly bind to both t-tubule and SR membranes. In mammalian tissue, three junctophilin proteins JP-1, JP-2, and JP-3 are encoded by different genes, with JP-2 expressed in heart. JP-2 has a COOH-terminal hydrophobic segment spanning the SR membrane, a linkage α-helical domain, and the remaining cytoplasmic region consisting of eight lipophilic “membrane occupation and recognition nexus (MORN)” domains (115). The MORN domains are associated with t-tubules, approximating t-tubule sarcolemma to the jSR membrane. It remains unclear whether JP-2 facilitates BIN1 bridged LTCC-RyR dyad formation and function. Whether JP-2 is critical in EC coupling and contributes to heart failure progression is also controversial. Some studies indicate that JP-2 is downregulated in animal models of hypertrophic and dilated cardiomyopathy, and conditional knockdown of JP-2 in the heart causes systolic heart failure in murine models (218). Other studies report no significant changes of JP-2 expression in failing hearts (23), or with recovery of heart function (134).

B. Ion Diffusion and Membrane Excitability

One of the functional attributes of t-tubules is the presence of slow diffusion zones close to t-tubule sarcolemma (164, 189, 209). Restricted diffusion occurs both at extracellular and intracellular spaces near the t-tubule membrane with accumulated ions such as calcium, which was originally observed as a theoretical “fuzzy space” (119). Growing evidence indicates that extracellular ions diffuse slowly within the t-tubule network, compared with extracellular bulk environment. In fact, ion diffusion coefficients within the t-tubule network are normally 5–10 times slower than those in the extracellular bulk environment (189, 209). Fluctuations in environmental extracellular ion concentrations therefore occur with delayed lag phase before reaching to t-tubule membrane and its complement of ion channels. For instance, the change of extracellular calcium is delayed by up to 2.3 s to reach the t-tubules in guinea pig myocytes (16). This lag phase due to restricted diffusion within the t-tubules can have a significant effect on muscle cell homeostasis and stability, particularly during environmental fluctuations induced by acute stress. Increased heart rate and quick transmembrane ion flux together with limited diffusion can rapidly accumulate outward current ions like potassium (209) and deplete inward current ions like calcium (164, 165), affecting the driving force of ion channels at t-tubule membrane and shortening the action potential duration (166). Shortened action potential limits t-tubule membrane excitability and electrical instability, serving as a crucial protective mechanism in preventing detrimental ventricular arrhythmias.

Although t-tubule forms a complex interconnected network (198), this network feature does not explain the highly restricted diffusion near the t-tubule sarcolemma. The structural foundation of the restricted diffusion zone was not revealed until the recent discovery of t-tubule membrane microfolds created by cardiac BIN1+13+17. In wild-type adult mouse ventricular cardiomyocytes, along with tortuous t-tubule (91) path meandering into the myocyte interior, BIN1 microfolds are present at t-tubule membrane, bringing spatial complexity to the lumen. These spatial divided subdomains with narrow connecting points to the central lumen help trap extracellular ions, preventing ionic diffusion. For example, the small extracellular gaps between layered membrane microfolds can trap Ca²⁺ and K⁺. Thus, in addition to dyad microdomain organization, BIN1 microfolds also create spatial pockets, serving as reservoirs for extracellular ions and other signaling molecules. In BIN1-deficient cardiomyocytes, these microfolds and so formed spatial pockets are lost and tortuous t-tubules become more straight and dilated, normalizing rapid ion diffusion at t-tubules. Mathematical modeling further confirms that removal of this diffusion barrier increases fast ion diffusion from t-tubule network to outside bulk environment. When BIN1 is reduced as occurs in heart failure, loss of BIN1 microfolds removes the protective ion diffusion barrier, increasing membrane excitability and promoting arrhythmias (91). In short, the microfold-forming cardiac BIN1+13+17 plays a pivotal role in maintaining and organizing t-tubule ultrastructure essential for cardiomyocyte electrical stability and homeostasis. Interestingly, recent studies in zebrafish skeletal muscle indicate that skeletal BIN1 may also have a similar function in maintaining t-tubule ultrastructure (194). Whether a similar membrane microdomain-formed slow diffusion zone exists in narrow skeletal tubular network with functional significance remains to be identified.

C. Other Microdomain-Related Functions

Other t-tubule microdomains, in addition to dyads, serve as signaling centers by concentrating transmembrane receptors, ion channels, and signaling molecules. For instance, caveolae are enriched with a subset of LTCCs (4) as well as β₂-AR (229), permitting efficient calcium signaling regulation by the β-adrenergic system (5). In addition, ankyrin B-organized NCX/NKA/InsP₃ macromolecular complex (147) facilitates calcium removal during muscle relaxation. Coordination of these t-tubule microdomains is necessary for achieving timely intracellular calcium equilibrium required for beat-to-beat heart contraction. Recent evidence indicates that crosstalk between BIN1 microdomains and caveolae likely exists for better control of β-adrenergic-regulated calcium entry through LTCCs, especially during acute stress as well as chronic dysregulation in heart failure. It is likely that caveloae microdomains are indeed related to BIN1 microfolds, since intracellular localization of caveolin-3 is altered in cardiomyocytes missing the Bin1 allele (116). The role of BIN1 in regulation of caveolae localization and function remains to be tested. Nevertheless, in addition to organizing microfolds-based dyads, BIN1 may have a significant role in other t-tubule microdomains as well.

V. LIFE CYCLE OF CARDIAC T-TUBULES

Despite their structural complexity, t-tubules are extremely labile (112). T-tubules are absent in embryonic myocytes (187); develop only after birth (82, 187); dedifferentiate in cultured myocytes (118, 130, 145), a process that can be slowed by actin stabilization (118, 214); remodel during heart hypertrophy and failing (130, 132, 133, 224); and can recover during functional recovery of the hearts (133). In addition to membrane turnover, other t-tubule components including transmembrane ion channels and receptors are also extremely dynamic. By regulating protein trafficking and recycling, cardiomyocytes maintain a steady state of continuously replenished t-tubule pool of ion channels and signaling molecules for functional homeostasis. Meanwhile, intracellular cytoskeleton components including cortical actin, microtubules, and costameres are actively regulated to control membrane integrity and protein function. This flexible function of t-tubules allows them to accommodate both healthy and stress environments.

A. Membrane Biogenesis, Maintenance, and Turnover

Unlike the sarcoplasmic reticulum membrane system, which develops in the fetal stage, cardiac t-tubules are absent in embryonic hearts. In rodent and rabbit hearts, the process of t-tubule biogenesis initiates a few days after birth (82, 83, 187), along with the rise of the left ventricular pressure and working volume. Mature t-tubules are only fully developed by the end of the first month of life. The mechanism of cardiac t-tubule development remains unclear. The Franzini-Armstrong group proposed in 2007 that the inward penetration of t-tubule invagination is driven by accrual of membrane lipids and specific proteins (50). Up to this date, limited progress has been made to further understand the lipids, proteins, and other molecules involved in biogenesis process of t-tubules. The need to resolve the molecular mechanisms of postnatal cardiac t-tubule development becomes even more urgent since the fast development in the stem cell research field over the last few years. In 2007, Yamanaka's group successfully reprogrammed differentiated human somatic cells into an undifferentiated pluripotent state for the generation of induced pluripotent stem (iPS) cells (210), which can then be redifferentiated into most any terminally differentiated cell types including cardiomyocytes. These iPS-induced cardiomyocytes contain essential contractile apparatus including sarcomeres, yet they do not have a mature t-tubule network. It is of great interest to know how to mature these iPS-differentiated cardiomyocytes into adult phenotype with characteristic interconnected t-tubules and efficient EC coupling machineries.

T-tubule networks and especially BIN1 organized microdomains can also be extremely dynamic. The gross t-tubule network in isolated cardiomyocytes loses integrity when cardiac origin cells are cultured in vitro, and t-tubule loss and remodeling are also a signature of failing cardiomyocytes. On one hand, the SH3 domain of BIN1 binds to dynamin 2 and is involved in dynamin 2-dependent endocytosis (22, 171), indicating a possible intracellular recycling of BIN1 microfolds through endocytic vesicle formation (Figure 8, adapted from BAR-regulated endocytosis in non-cardiac cells in Ref. 232). On the other hand, the recently identified BIN1 microfolds are lost in isolated cardiomyocytes during extended in vitro culture (91), indicating external release of BIN1 microfolds. Loss of BIN1 microfolds can be rescued by replenishment of the microfold-forming BIN1 isoform and stabilization of the filamentous actin cytoskeleton. Furthermore, BIN1 is blood available, and plasma BIN1 correlates with cardiac function (92). It is possible that plasma BIN1 originates from cardiac t-tubules as a result of BIN1-microfold turnover and release. In addition to stabilizing membrane microfolds to Z-discs, BIN1 may also regulate cytoskeleton dynamics for microparticle release (see speculative release mechanism in Figure 9), balancing the amount of microfolds and microfold-based calcium regulatory microdomains at t-tubule membrane. The equilibrium between microfold organization and turnover is achieved by the amount of BIN1 protein at the t-tubule membrane. The released BIN1-containing microparticles therefore carry a signature of cardiomyocyte t-tubule membrane, which can be quantified in plasma level, and the resulted plasma cardiac BIN1 level can be used as a marker of myocardial health (92). How this BIN1-mediated membrane turnover and release is linked to other membrane vesicle trafficking processes, such as injury-related membrane repair (45, 46) especially during acute stress and insult, will be interesting to pursue in the future.

FIGURE 8.

Schematic of endocytosis from cBIN1 microdomains. Adapting BAR domain regulated endocytosis from noncardiac cells (229), we expect that for t-tubules, cBIN1 interacts with cortical actin to generate an endocytic neck, and the SH3 domain of BIN1 the binds to dynamin 2 for scission and internalization of the endocytic vesicle.

FIGURE 9.

BIN1-microdomain release from cardiac t-tubules. A schematic of how fission can potentially occur between two neighboring invaginated cBIN1 microfolds, resulting in external release of cBIN1 microfolds that result in blood-borne microparticles.

B. Ion Channel Trafficking and Half-Life

Transmembrane ion channels and ion handling proteins are dynamically turned over to maintain a functional pool at the t-tubule surface. In cardiomyocytes, the half-life of individual ion channels is remarkably short, on the order of hours. The half-lives of t-tubule localized calcium and potassium channels, and sodium-calcium exchangers, are all on the order of 10 h (33, 37, 49, 57). Even the more extended half-life of sodium channels is ∼35 h (138). For an overall turnover scale of several to tens of hours, the bulk of the ion channels in each cardiomyocyte are regenerated and have to be newly positioned to t-tubule membrane in the course of any day. To maintain such a steady state and continuously replenished t-tubule pool for functional homeostasis, ion channels at the t-tubule membrane need to be tightly regulated at both forward trafficking and recycling pathways. In fact, ion channels often contain multiple signaling sequences that regulate forward trafficking including the rate of their synthesis, organization at endoplasmic reticulum such as multimerization with auxiliary subunits, and traffic through Golgi apparatus before arrival at the plasma membrane (144, 205, 239). On the other hand, t-tubule pool of ion channels is also regulated by retrograde trafficking including internalization, recycling, and degradation. Furthermore, it is possible that ion channel half-life at plasma membrane may occur at a different time scale. For instance, the half-life of active NCX in the plasma membrane is likely to be under 12 h (57, 127), which is shorter than the long half-life of 33 h measured using metabolic labeling in lysates from neonatal cardiomyocytes (193). Similarly, the functional half-life of the Cav1.2 channel is ∼3 h (33), whereas the total protein turnover occurs at a much slower rate of 25 h (29). The same might be true for other ion channels including the potassium and sodium channels. The precise measurement of channel life cycles at differential plasma membrane subdomains is limited in the literature, which will require a combination of pulse-chase experiments with biochemical fractionation and extensive use of life cell imaging.

To provide more insights into ion channel turnover in t-tubules versus sarcolemmal membrane, future studies are required in primary cardiomyocytes from adult ventricles. The results will also help determine how microdomains in t-tubules affect ion channel stability compared with non-t-tubule plasma membrane. Nevertheless, it is fair to conclude from studies to date that dynamic turnover of ion channels at t-tubule membrane is essential for normal function of cardiomyocytes, and allows timely regulation of functional channels under normal physiological conditions. Furthermore, it has already been established that mutations which affect forward trafficking, recycling, and degradation of ion channels, or their interaction with microdomain-organizing proteins such as caveolin-3 and, ankyrin B can result in major cardiac diseases (17; review in Ref. 40).

VI. ALTERATION OF T-TUBULE STRUCTURE AND FUNCTION IN HEART FAILURE

Heart failure is the fastest growing cardiovascular disorder with a high mortality and morbidity affecting more 40 million patients worldwide (129, 176, 177), particularly in the developed countries including the United States. The primary mortality involves cardiac pump failure as a consequence of reduced cardiac contractility and sudden cardiac death due to detrimental ventricular arrhythmias. The definitive therapy for end-stage heart failure is a heart transplant, which unfortunately remains as a limited option due to poor organ availability (102), and a need for subsequent care and monitoring. The alternative therapy for end-stage heart failure is left ventricular assist devices (LVAD) (124, 192), and for arrhythmia management are implantable cardioverter defibrillators (ICDs) (217). However, the cost of device implantation of either LVAD or ICD and the morbidity associated with implanted devices are enormous. Lack of an accurate understanding of cardiomyocyte health and recovery potential limits the ability in making critical decisions with regard to heart transplant priorities and criteria for LVAD and ICD implantation (18, 170). New diagnostic and prognostic tests that accurately evaluate myocardial health and arrhythmogenic potential at the cellular level could significantly improve decision making for LVAD and ICD implantation with improvement in patient care. New therapeutic tools with the ability to break pathophysiological cycling of heart failure progression are acutely needed to help postpone and rescue disease development in millions of patients with heart failure.

For efficient development of new diagnostic, prognostic, and therapeutic tools of human heart failure, a more complete understanding of the biology of failing cardiomyocytes is required. The pathophysiological hallmark of failing cardiomyocytes is weakened calcium transients (74, 75, 128). Normal transients are initiated at cardiac dyads, where t-tubule LTCCs (11, 185) are closely associated with SR RyRs (31, 59, 100, 168) for optimal CICR following each action potential. Failing calcium transients can be attributed to t-tubule remodeling (132, 133, 224) and dyad uncoupling (15, 74, 75, 88, 131, 132). T-tubule remodeling has been considered as a characteristic mark of the transition from hypertrophy to failure (224) and a contributing mechanism of heart failure progression (see reviews in Refs. 19, 20, 61, 80, 99, 132). In addition to the destruction of the gross t-tubule network, protein components at t-tubules including ion channels and signaling molecules are also reported to be either reduced or redistributed out of t-tubules. Loss of complex t-tubules in heart failure not only desynchronizes CICR and impairs contractility, but also impairs local electrophysiological homeostasis and increases susceptibility to ventricular arrhythmia.

Of heart failure-associated molecules, the cardiac t-tubule membrane deformation protein BIN1 is emerging as a crucial master regulator of healthy t-tubule function, which when reduced can promote heart failure progression (93, 116). The role of BIN1 as a cardiac t-tubule adaptor protein was first noted when constitutive Bin1 knockout in mice caused perinatal lethality due to cardiomyopathy (150). Later it was identified that BIN1 localizes to cardiac t-tubules, facilitating microtubule-dependent forward delivery of Cav1.2 channels to t-tubule membrane (94). Follow-up studies from independent research groups identified that BIN1 is transcriptionally decreased in acquired human heart failure (93), as well as multiple animal models of heart failure (23, 134). Decreased BIN1 level in failing cardiomyocytes reduces LTCC surface expression (23, 93, 134), contributing to the pathophysiology of diseased calcium transients and electrical instability in failing hearts (Figure 10). In zebrafish, morpholino-mediated knockdown of the Bin1 gene induces significant contractile dysfunction and cardiomyopathy (93). Mice with cardiac conditional Bin1 knockout after birth have increased tendency to develop cardiomyopathy with either older age or pressure overload (116). Of the four BIN1 isoforms, cardiac specific BIN1+13+17 with inclusion of exons 13 and 17 is the primary BIN1 isoform localized to cardiac t-tubules and responsible for cardiac t-tubule membrane ultrastructure formation (91). Whether aberrant splicing and mutation in cardiac spliced exons plays a role in cardiomyopathy development remains to be observed. Although the mechanism of reduced BIN1 expression in heart failure remains to be determined, BIN1-based microdomain is a potential therapeutic target for the development of new heart failure therapies.

FIGURE 10.

Reduced cBIN1 microdomains in failing cardiomyocytes. In failing cardiomyocytes with reduced cBIN1 and cBIN1-induced microdomains, expression of ion channels on the cell surface and the morphology of t-tubules are altered. As highlighted in the cartoon, changes following reduced BIN1 transcription include 1) impaired microtubule-dependent targeted delivery of LTCC results in reduced t-tubule surface expression of LTCCs; 2) disruption of cBIN1 microdomains impairs sufficient LTCC clustering within t-tubules; 3) loss of dense membrane microfolds and resultant t-tubule dilation cause removal of the protective slow diffusion zone within t-tubule lumen; and 4) impaired recruitment of RyRs causes accumulation of hyperphosphorylated RyRs outside dyads, increasing orphaned leaky RyRs.

In addition to serving as a target for therapeutic development, BIN1 is a potential biomarker of myocardial health and reserve. BIN1 is blood available, and plasma BIN1 correlates with cardiac health (92), through continuous turnover of BIN1-microdomains released from the t-tubule membrane. In a cohort of arrhythmogenic right ventricular cardiomyopathy patients, plasma BIN1 decreases in patients developing symptomatic heart failure (92). Low plasma levels of BIN1 correlates with heart failure status and predicts future arrhythmia incidences (92). The recently cloned microdomain forming cBIN1 (91) provides an opportunity to develop a cardiac specific BIN1 blood test to aid in the diagnosis and management of patients with failing hearts.

In the clinical literature, a clear distinction now exists between heart failure with reduced ejection fraction (HFrEF) and heart failure with preserved ejection fraction (HFpEF), with HFpEF incidence exceeding that of HFrEF (158). The diagnosis of HFpEF is made in patients who present with clinical heart failure despite the absence of left ventricular systolic dysfunction. Unlike HFrEF, HFpEF has few available treatment options (142, 158). Newer therapies such as cardiosphere-derived cell-based treatments are still in the research phase (71). In unpublished studies we have found that plasma-based cBIN1 decreases in patients with HFpEF just as it does for patients with HFrEF. These types of studies suggest that future classification of failing hearts will be based less on measures of overall organ function such as ejection fraction, and more on the molecular details of myocyte, and in particular t-tubule, pathophysiology.

VII. CONCLUSIONS

In this review, we provided an overview of the structural and functional organization of the cardiomyocyte t-tubule system that is required for normal contractility and electrical stability of the heart. The life cycle of t-tubule membrane and proteins, as well as t-tubule remodeling during heart failure, is also discussed. The membrane microdomains within the cardiac t-tubular invaginations are highlighted with particular focus on the recently discovered microdomains sculpted by the cardiac isoform of the membrane curvature formation protein BIN1 (cBIN1). Evidence is emerging that cBIN1-formed membrane microfolds are needed for both normal dyad formation and creation of a “fuzzy space” of slow ionic diffusion zone within the t-tubule network. The reversible reduction of BIN1 transcription in failing myocardium further indicates the crucial role of BIN1 in heart failure pathophysiology. Measurement of cBIN1 levels at the cardiac t-tubule membrane can potentially be monitored through a simple blood test, likely due to the continuous turnover of BIN1 microfolds into circulation. Thus BIN1 microdomain is a promising new target for the development of novel diagnostic and therapeutic tools for human acquired heart failure. More generally, t-tubules are now being approached as structurally complex, multifaceted organelles vital to cardiomyocyte health and function.

GRANTS

This work is funded by National Heart, Lung, and Blood Institute Grants R00-HL10975 (to T. Hong) and R01-HL094414 (to R. M. Shaw) and by American Heart Association Grants IRG-16IRG27780031 and BGIA-16BGIA27770151 (to T. Hong) and EIA-13EIA4480016 (to R. M. Shaw).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.