Traditional Northern blot analyses conducted on various human and rodent tissue samples have demonstrated a high level of Shank3 mRNA expression in the heart, which is comparable to, or even higher than the level in the brain (Lim et al., 1999; Wilson et al., 2003). Consistently, bulk tissue RNA sequencing (RNA-seq) datasets available on The Human Protein Atlas (https://www.proteinatlas.org/ENSG00000251322-SHANK3/tissue) indicate “heart muscle” is one of the top-ranked tissues expressing Shank3 mRNA (ninth out of the 54 tissues listed in the consensus dataset). The heart is a complex organ comprising four morphologically and functionally distinct chambers, and is composed of various cell types, such as cardiomyocytes, endothelial cells, fibroblasts, smooth muscle cells, pericytes, and immune cells (Litvinukova et al., 2020). According to the single-cell RNA-seq dataset for heart muscle on The Human Protein Atlas (https://www.proteinatlas.org/ENSG00000251322-SHANK3/single+cell+type/heart+muscle), Shank3 mRNA is predominantly detected in endothelial cells, followed by cardiomyocytes, smooth muscle cells, and fibroblasts, of heart muscle.

At the protein level, however, Shank3 in the heart muscle is indicated as “not detected” on The Human Protein Atlas (https://www.proteinatlas.org/ENSG00000251322-SHANK3/tissue/primary+data). Although it is not immediately clear why there is a discrepancy between mRNA and protein levels, we caution that protein expression analysis should be interpreted with care because the currently available Shank3 antibodies demonstrate varying performance depending on the analyses conducted (Lutz et al., 2022). Furthermore, several direct validations and functional analyses support the expression of Shank3 protein in the heart, especially in cardiomyocytes (as detailed below). However, since most of these studies have used rodent models, it is possible that Shank3 protein expression levels and distributions differ in human hearts. Therefore, further analysis is needed to confirm the expression of Shank3 proteins in human hearts.

Another issue regarding Shank3 expression in the heart is the lack of detailed information about specific isoforms. Both human SHANK3 and rodent Shank3 genes express various isoforms due to multiple intragenic promoters and alternative splicing (Jiang and Ehlers, 2013; Wang et al., 2014) (Figure 1). In the brain, each Shank3 isoform has a distinct regional expression pattern, contributing to the clinical and phenotypic heterogeneity observed in patients and animal models (Lee et al., 2017a; Jin et al., 2018; Yoo et al., 2022). Therefore, it is important to determine whether the heart tissue also exhibits differential expression of Shank3 isoforms depending on the chambers and cell types.

The first molecular function of Shank3 identified in cardiomyocytes was as an interaction partner of phospholipase Cβ1b (PLCβ1b) (Grubb et al., 2011) (Figure 2A). The C-terminal proline-rich sequences of PLCβ1b directly bind to the Src homology 3 (SH3) domain of Shank3 (Grubb et al., 2015), and as a result, PLCβ1b and Shank3 co-localize at the sarcolemma of cardiomyocytes.

In cardiomyocytes, PLCβ1b is activated by the Gq family of heterotrimeric G proteins, thus by G protein coupled receptors (GPCRs) such as α ₁-adrenergic receptors, and resulting in the hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP₂) to generate inositol 1,4,5-triphosphate (IP₃) and sn-1,2-diacylglycerol (DAG) (Filtz et al., 2009). This activation of Gq signals leads to cardiomyocyte hypertrophy, where PLCβ1b, but not its splice variant PLCβ1a, plays a critical downstream role in the process (Filtz et al., 2009). The functional difference between PLCβ1a and PLCβ1b can be attributed to their differential subcellular localizations; unlike PLCβ1b, PLCβ1a does not localize to the sarcolemma where PIP₂ is concentrated in cardiomyocytes (Grubb et al., 2008).

The amino acid sequence difference between PLCβ1a and PLCβ1b is only in the C-terminal region. Specifically, PLCβ1a has a PSD-95, Dlg, and Zo-1 (PDZ)-binding motif in its C-terminal tail instead of proline-rich sequences, and thus cannot bind to the SH3 domain of Shank3. Therefore, the PLCβ1b-Shank3 interaction possibly explains the PLCβ1b-specific localization to the sarcolemma and its role in Gq-induced signaling to mediate cardiomyocyte hypertrophy. Supporting this notion, knock-down of Shank3, expression of full-length PLCβ1b with proline-to-alanine mutations in the C-terminal proline-rich sequences, and expression of only the C-terminal 32-aa sequences of PLCβ1b that possibly disrupt endogenous PLCβ1b-Shank3 interaction impair either PLCβ1b localization to the sarcolemma in cardiomyocytes or Gq-induced cardiomyocyte hypertrophy (Filtz et al., 2009; Grubb et al., 2011; Grubb et al., 2015).

In addition to regulating cardiomyocyte hypertrophy, PLCβ1b can also affect calcium homeostasis in cardiomyocytes by producing DAG (Figure 2A). DAG activates protein kinase Cα (PKCα), which in turn regulates the phosphorylation status of phospholamban (PLN) via the inhibitor-1 (I-1) and protein phosphatase-1 (PP-1) pathway (Woodcock and Grubb, 2015). Specifically, PKCα phosphorylates I-1, which activates PP-1 to dephosphorylate PLN. When PLN is dephosphorylated, it inhibits the sarco-endoplasmic reticulum ATPase 2a (SERCA2a), a pump critical for maintaining calcium homeostasis in cardiomyocytes (MacLennan and Kranias, 2003). SERCA2a transports calcium ions from the cytosol back to the sarcoplasmic reticulum (SR) during the cardiac contraction cycle. Dysregulation of SERCA2a can results in defects in intracellular calcium homeostasis, which is implicated in heart failure (Zhihao et al., 2020). Since the PLCβ1b-Shank3 interaction is critical for proper subcellular localization and activity of PLCβ1b in producing DAG, it is conceivable that Shank3 is involved in regulating calcium homeostasis in cardiomyocytes by indirectly regulating the activity of SERCA2a (Woodcock and Grubb, 2015).

To date, several Shank3 mutant mouse models, including conventional and conditional knock-out, knock-in, and overexpression mice, have been generated and characterized (Jung and Park, 2022). Among them, various knock-out mouse models, which have different deletions of Shank3 exons (out of total 22 exons in the mouse Shank3 gene), exhibit both shared and distinct neurobehavioral phenotypes, possibly due to the deletion of specific Shank3 isoforms based on the affected exons (Jin et al., 2020; Delling and Boeckers, 2021) (Figure 1).

Taking into account this phenotypic heterogeneity, a recent study investigated the morphological and functional changes in the heart of adult Shank3 exon four to nine heterozygous (Δexon 4–9 ^+/− ) and homozygous (Δexon 4–9 ^−/− ) knock-out mice (Assimopoulos et al., 2022). Since Shank3 exon four to nine encodes the N-terminal ankyrin repeats (ANK) of Shank3, Shank3 isoforms containing SH3, PDZ, proline-rich, and the C-terminal sterile alpha motif (SAM) domains can still be normally expressed in the mice (Jiang and Ehlers, 2013). The Shank3 Δexon 4–9 ^+/− mice displayed significantly increased left ventricular (LV) anterior and posterior wall thickness, while Shank3 Δexon 4–9 ^−/− mice only displayed an increase in LV anterior wall thickness compared to wild-type mice (Assimopoulos et al., 2022). However, other cardiac parameters, including aorta diameter, heart rate, LV fractional shortening, stroke volume, and chamber diameter, remained normal in both Shank3 Δexon 4–9 ^+/− and Δexon 4–9 ^−/− mice.

Another study demonstrated the effects of Shank3 dosage on the morphological and functional changes in the heart after myocardial infarction (MI) by investigating the phenotypes in adult Shank3 knock-out and overexpressing mice (Man et al., 2020). Unfortunately, detailed information, such as deleted Shank3 exons and the promoter driving Shank3 overexpression, about the Shank3 knock-out and overexpressing mice, respectively, were not provided in the paper. Nevertheless, both morphological (infarct size) and functional (echocardiographic parameters, including LV ejection fraction and fractional shortening) changes after MI were significantly aggravated in Shank3 knock-out mice, while they were alleviated in Shank3 overexpressing mice. Mechanistically, after MI induction, increased apoptosis and decreased autophagy were observed in Shank3 knock-out mice, while the opposite changes were seen in Shank3 overexpressing mice compared to wild-type mice. The effects of Shank3 dosage on apoptosis and autophagy were replicated in primary neonatal cardiomyocytes after hypoxia condition (Man et al., 2020), confirming the protective role of Shank3 in cardiomyocytes after MI. The exact mechanism by which Shank3 regulates apoptosis and autophagy in cardiomyocytes remains unknown, but crosstalk between Shank3 and mitochondria (as detailed below) or the mechanistic target of rapamycin complex 1 (mTORC1) signaling (Bidinosti et al., 2016; Lee et al., 2017b), a negative regulator of autophagy (Kim and Guan, 2015), may be involved.

The role of Shank3 in age-related cardiac dysfunction has also been demonstrated in a study (Wang et al., 2022) (Figure 2B). This study found that aged (18-month-old) wild-type mice exhibited cardiac dysfunction, hypertrophy, and fibrosis along with increased cardiac Shank3 expression compared to young mice. Conversely, in cardiac-specific Shank3 conditional knock-out (cardiac cKO) mice, which were generated by crossing floxed-Shank3 mice and cardiac α-myosin heavy chain-Cre mice, age-related cardiac changes were alleviated. This indicates that increased Shank3 expression contributes to age-related cardiac dysfunction. At the molecular and cellular level, cardiomyocytes in Shank3 cardiac cKO mice exhibited suppressed age-related increase in mitochondrial oxidative stress and apoptosis, and age-related decrease in mitophagy. Mechanistically, Shank3 interacts with calcium-calmodulin-dependent protein kinase II (CaMKII) and inhibits its translocation to mitochondria, which is enhanced in the aged heart as Shank3 expression increases. Reduced CaMKII translocation to mitochondria subsequently inhibited Parkin phosphorylation and translocation to mitochondria, critical steps in mitophagy activation (Vives-Bauza et al., 2010). Therefore, increase of Shank3 expression in the aged heart inhibits mitophagy in a CaMKII-dependent manner (Wang et al., 2022). These results somewhat contrast with the protective role of Shank3 in MI-induced cardiac dysfunction (Man et al., 2020). This difference in findings may be attributed to the different mouse models used (adult conventional vs. aged conditional knock-out) and/or the experimental setting (MI vs. aging).

Numerous Shank3 interactors in the PSD have been identified using candidate-based and unbiased approaches, such as yeast two-hybrid screening and mass spectrometry-based proteomic analysis, followed by functional validation studies (Sakai et al., 2011; Han et al., 2013; Wang et al., 2020). Some of these synaptic interactors are also highly expressed in the heart and have already been shown to play essential roles in proper cardiac function. Specifically, we observed that 297 proteins from a recent Shank3 interactome (Wang et al., 2020) were present in at least one of three cardiac expression datasets (Supplementary Table S1). Therefore, in addition to the functions described above, it is possible that Shank3 can exert further molecular functions by interacting with and regulating these proteins in cardiomyocytes.

The Homer family proteins, comprising Homer1–3, act as scaffolds and well-established interactors of Shank3 in the PSD, where they together form a highly-organized lattice-like structure (Hayashi et al., 2009; Sheng and Kim, 2011). Moreover, Homer interacts with and regulates the activity and subcellular localization of several calcium-regulatory proteins, such as L-type calcium channels, ryanodine receptors (RyRs), IP₃ receptors (IP₃R), PLCβ, and transient receptor potential canonical (TRPC) channels, in various cell types (Worley et al., 2007). These interactions, including the Homer-Shank3 interaction, are mediated by the N-terminal Ena/VASP homology1 (EVH1) domain of Homer, which recognizes the proline-rich motifs in the binding partners (Supplementary Figure S1A). Long Homer proteins (the Homer1b/c isoforms, Homer2, and Homer3), but not short Homer1a, also contain a coiled-coil domain at the C-terminal region, which mediates the formation of tetrameric Homer (Hayashi et al., 2009) (Supplementary Figure S1B). Therefore, Homer can organize cellular microdomains through its interactions with calcium regulators and its multimerization, to efficiently and tightly regulate intracellular calcium signaling (Figure 2C). One example of such regulation in muscle cells, including cardiomyocytes, is the Homer1-dependent regulation of the excitation-contraction coupling by forming the Ca_v1.2-Homer1-RyR2 complex (Worley et al., 2007). Homer1b/c solidifies the complex, while Homer1a relaxes it, thereby controlling the membrane depolarization-induced activation of the calcium-induced calcium release (Huang et al., 2007). It remains to be further investigated whether and how Shank3 is involved in Homer1-dependent regulation of calcium signaling in cardiomyocytes. A previous study has shown that Homer1c interacts with PLCβ1b and Shank3 in cardiomyocytes, mediating Gq-induced cardiomyocyte hypertrophy (Grubb et al., 2012).

Shank3 interacts with numerous actin-binding and -regulatory proteins in the PSD (Han et al., 2013; Wang et al., 2020), and Shank3 can also directly bind to actin through its N-terminal domain (Salomaa et al., 2021). Consistently, abnormalities in the synaptic actin cytoskeleton are considered to be the primary molecular mechanisms underlying the synaptic and behavioral deficits observed in Shank3 mutant mice (Duffney et al., 2015). In cardiomyocytes, the F-actin filament, also known as the thin filament, is the key component that mediates the contractile function, and many actin-binding proteins, such as α-actinin, are concentrated at the Z-disc where one end of thin filament is anchored (Sequeira et al., 2014). In the skeletal muscle sarcomere, Shank3 interacts with α-actinin and co-localizes in Z-discs, and loss of Shank3 resulted in shortened Z-discs (Lutz et al., 2020). Therefore, Shank3 may have similar roles in cardiomyocytes, organizing the Z-disc by interacting with actin, α-actinin, and other actin-binding proteins (Figure 2D).