Plakophilins, encoded by four genes (PKP1, 2, 3, and 4), are located on chromosomes 1q32, 12p13, 11p15, and 2q23-q31, respectively [71]. These proteins show complex tissue-restricted expression patterns and are members of the armadillo gene family. The term armadillo was first described in Drosophila and is the historical name for the β-catenin gene, which is thought to regulate homodimerization of a-catenin to mediate actin branching and bundling. Armadillo repeat domains interact with binding partners to regulate cell-contact- and cytoskeleton-associated protein interactions and a wide range of other functions [72,73]. PKP proteins (1, 2, and 4) contain 9 armadillo repeats (repetitive amino acid sequences of approximately 42 residues) with a flexible insert between repeats 5 and 6 [71]. PKP1 and PKP2 each have two isoforms, one with a shorter “a” variant and another with a longer “b” variant generated by alternative splicing. PKP4 isoforms are similar to those of PKP1 and PKP2, but the two PKP3 isoforms are produced by alternative promoter usage. PKP1, 2, and 3 and their isoforms can be found in desmosomes and in the nucleus, the location of which is influenced by 14-3-3 proteins [47]. PKP2 phosphorylation at serine residue 82 by Cdc25C-associated kinase 1, for example, creates a 14-3-3 binding site that is located mainly at the plasma membrane. PKP2 mutants that cannot bind 14-3-3 are found mostly in the nucleus [74]. PKP4 is involved in Rho regulation during cytokinesis.

The PKP2 gene, the main isoform in the human heart, encodes a protein of 881 amino acids. It is especially prominent in the right ventricle [75]. Mutations in PKP2 are typically inherited and closely associated with familial forms of the disease that typically impact the right ventricle. From patient studies, Gerull et al. described nearly 25 different heterozygous mutations from individuals of Western European descent. Of these, 12 were insertion–deletion, 6 were nonsense, 4 were missense, and 3 were splice site mutations. Most of the mutations were present in the C-terminal half of the molecule [76]. In a Polish population of patients that met ITFC criteria for ACM, Biernacka et al. identified five frameshift, two nonsense, two splicing, and one missense mutations in PKP2. Through comparison studies, they found that patients with a PKP2 mutation were younger at diagnosis, more often had negative T waves, had higher left ventricular ejection fracts, and were less likely to have symptoms of heart failure than ACM patients without PKP2 mutations [77]. Individuals in the non-PKP2 group tended to have a better prognosis with less SCD or need for a heart transplant [77]. Syrris et al. investigated 100 Caucasian patients with PKP2-associated ACM. Nine different mutations were identified, including five that had not previously been reported in PKP2 (R413X, A733fsX740, L586fsX658, V570fsX576, and P533fsX561); these five mutations included one which caused a premature truncation of PKP2 and four which were frameshift mutations [78]. Dalal et al. reported that males were more likely to have structural and conduction abnormalities than females [79], and the phenotypes within the tested cohort ranged from asymptomatic to severe (including SCD) [79]. Van Tintelen et al. studied ACM cases in a Dutch population, and of the 34 patients that satisfied the index criteria for ACM, 23 were familial of which 16 (70%) were caused by PKP2 mutations [80]. PKP2 mutations also have been associated with SCD, particularly in young athletes, and with left ventricular involvement during late stages of the disease [81].

The role of PKP2 mutations in dACM has been studied in animal models and in human cells. Cerrone et al. found that loss of PKP2 in a stable, lentivirus-modified HL-1 cell line, a cell line originally derived from the AT-1 mouse atrial CM tumor lineage, resulted in reduced sodium current (I_Na) relative to controls [68,94]. This was the first to demonstrate that loss of Na_V1.5 at the ICD could be caused by a genetic variation in PKP2. These PKP2 mutations could also contribute to Brugada syndrome, which is characterized by arrhythmias associated with very fast heart rates. Interestingly, the authors speculated that the effects of PKP2 mutations may not have been due to its function as a component of the desmosome, but rather that cytoskeletal alterations affecting ion channel trafficking may be important for this phenotype [68]. Chelko et al. showed that buccal mucosa cells cultivated in vitro from patients with PKP2 mutations generally had a reduction in PKP1, JUP, and connexin 43 (Cx43) immunostaining; however, DSP was not affected. Interestingly, the loss of PKP1 staining in these cells was more consistent than the loss of PKP2 staining. Treatment of the cultured cells with SB216763, a glycogen synthase kinase 3-β (GSK3β) inhibitor previously reported to reverse ACM disease phenotypes in zebrafish [95], could normalize desmosome protein distributions in the cells through a process thought to involve Wnt/β-catenin signaling [96]. In zebrafish subjected to PKP2 knockdown using morpholinos, Moriarty et al. found that heart development was stifled by the presence of PKP2 mutations [97]. It was accompanied by cardiac edema, heart looping defects/blood pooling, reduced heart rate, and changes to the desmosome structures and numbers. Importantly, this phenotype could be rescued in the zebrafish by injection of PKP2 mRNA.

Animal models have provided additional direct evidence of the role of PKP2 in ACM. Using transgenic animal models, Grossman et al. developed a null mutation by homologous recombination in PKP2 [76]. During development, mouse homozygote embryos developed right ventricular (RV) wall thinning, had hemopericardium at embryonic day 10 with blood aggregates in the intraperitoneal cavity that led to small holes in the lining (endothelium) of the heart and developing vasculature, and did not survive beyond ED11 [76]. Cerrone et al. reported that missense PKP2 mutations with PKP2 deficits in CMs were correlated with sodium current (I_Na) deficiency, with reduced numbers of sodium channels at the ICD and with an increased separation of microtubules from the cell end, which was defined by the midline of N-cadherin clusters. PKP2 deficiency thus affected the ability of microtubules to reach the ICD, which likely impacted the delivery to the ICD of proteins related to sodium channel function [68]. Using a stable cardiac expression mutation of PKP2 (c.2203C>T), Cruz et al. demonstrated that the R735X mutation functioned as a dominant-negative variant that led to an exercise-dependent dACM phenotype with impaired global right ventricular systolic function and regional wall motion abnormalities [98]. In summary, these and other studies from human and mouse models show that mutations present in PKP2 account for ~70% of all diagnoses attributable to desmosome mutations, typically affect males more than females, and are often associated with SCD in young athletes. Furthermore, PKP2 functionally plays an essential role in the development of desmosomal and ICD structures.

In humans, desmoglein proteins are encoded by four genes (DSG1-4) located as a gene cluster on chromosome 18q12. These proteins are all members of the cadherin cell adhesion molecule superfamily that mediates calcium-dependent cell–cell adhesion. This family of proteins is involved in numerous biological processes that include cell recognition, communication, and signaling as well as angiogenesis and morphogenesis. DSG genes are comprised of either 15 or 16 exons, and their expression patterns display a tissue and differentiation-specific distribution [99]. DSG1 and 3 are mostly found in stratified squamous epithelia among other tissues, while DSG4 is present mostly in human hair follicle. DSG2 is the most widely expressed isoform, and it is the only DSG protein present in CMs. It is composed of 1118 amino acids, and it is a calcium-binding transmembrane glycoprotein.

It is estimated that 5–10% of cases with dACM carry a pathogenic DSG2 mutation variant, with the majority being rare missense mutations [99]. Several nonsense, insertion, deletion, and splice site variants that lead to frameshifts and premature termination codons have been observed [100,101,102]. DSG2 pathogenic variants often show prominent left ventricular involvement during early stages of the disease. From human studies, Pilichou et al. found nine heterozygous DSG2 mutations in a series of unrelated dACM patients, including five missense, one nonsense, two insertion/deletion, and one splice-site mutation [101]. These DSG2 mutations in dACM patients appeared to predominantly affect the left ventricle with desmosome structure remodeling, a loss of myocytes, and fibrofatty tissue replacement [101]. In 2021, Lao et al. published a case report of a 28-year-old woman who presented with left ventricular dominant dACM who had a heterozygous pathogenic variant in exon 15 (c.3059–3062del, p.Glu1020Alafs*18) [103]. Clinically, this patient had epicardial and midmyocardial fatty infiltration in the left ventricles, regional dyskinesis, and reduced left ventricular ejection fraction [103]. Using next generation sequencing (NGS) and single nucleotide polymorphism (SNP) arrays on samples from two individuals with ACM, Brodehl et al. reported a homozygous splice mutation (c.378+1G>T) and a nonsense mutation (p.L772X) coupled with a large deletion in DSG2 that appeared to exhibit an autosomal recessive inheritance pattern in patients [104]. Whole exome sequencing uncovered a homozygous mutation (c.1592T>G) regulated by DNA modification that resulted in amino acid sequence changes, protein structure effects, and splice site changes in samples. The patient with this biallelic mutation had increased levels of cardiac troponin I, myocardial edema in the lateral ventricular wall and apex, and an electrocardiogram with multiple premature ventricular beats. This unusual mutation in DSG2 resulted in an initial diagnosis of pediatric myocarditis, which was subsequently classified as ACM. Due to the early onset and distinct clinical features, an expanded clinical feature spectrum of DSG2-associated dACM was developed to account for this phenotype [105]. These studies, among others, established a basis for dACM that involves DSG2 and left ventricular disease phenotypes [106,107,108].

As described above, Chelko et al. utilized buccal mucosa cells to assess dACM in patient cells. Using these human DSG2 mutant cells, they found that PKP1 was not affected by DSG2 mutations, in contrast to what was observed for PKP2; however, these mutations led to a reduction in Cx43 and JUP, which could be improved by treatment with a GSK3β inhibitor [96]. In mouse models, DSG2^−/− embryos die during early embryogenesis, while DSG2^+/− embryos often die at or shortly after implantation. In blastocysts, the distribution of DSP in the desmosomal plaque was disturbed; however, neither E-cadherin nor β-catenin appeared to be affected [109]. Pilichou et al. demonstrated the impact of DSG2 mutations by creating a transgenic mouse line with cardiac-restricted expression [110]. Mice with a N271S mutation developed clinical features of dACM, which included aneurysms, spontaneous ventricular arrhythmias, cardiac dysfunction, biventricular dilation, and sudden cardiac death (SCD) at a young age. CM necrosis was also observed. DSG2 mutations thus can promote myocardial damage that may lead to myocardial atrophy, inflammation, fibrofatty replacement of myocytes, and calcification [110]. Krushce et al. developed a model of dACM in mice lacking exons 4–6 (targeted deletion) of DSG2 [111], which led to an increase in sudden cardiac death. These mice showed an increase in CM cell death and necrosis as well as fibrotic lesions. Proliferation of cells within fibrotic lesions (increased scar/fibrous tissues) were more pronounced in young versus old mice. With aging, however, these mice developed fibrous tissues and ventricular dilation, all hallmarks of dACM. Rizzo et al. studied transgenic mice that overexpressed mutant DSG2 (N271S). This mutation was a homologue of the DSG2 N266S mutation previously identified in an dACM patient by Pilichou et al. [101]. Rizzo et al. found that DSG2 mutations induced widening of the ICD at the level of the area composita, a finding that coincided with slowed conduction, a reduction in action potential upstroke velocity, and a reduction in I_Na. They provided further evidence of an in vivo interaction between DSG2 and Na_v1.5 as a molecular mechanism responsible for the slowed conduction and presence of arrhythmias in dACM prior to any overt structural changes [112] Kant et al. described a transgenic mouse line with CM-restricted ablation. DSG2 protein levels in hearts were reduced to <3% of normal, and all the animals developed dACM with severe morphological alterations during the postnatal period. The phenotypes included chamber dilation, calcifying CM necrosis, inflammation, interstitial and focal replacement fibrosis, and conduction defects with an abnormal Cx43 distribution [113]. Chelko et al. reported on DSG2 homozygous mutant mice lacking exons 4 and 5 that recapitulated dACM by 8 weeks of age compared with controls. Phenotypically, these mice had abnormal JUP and Cx43 at the ICD and extensive biventricular fibrosis but exhibited a normal distribution of CDH2 relative to controls. Inhibition of GSK3β could improve the lifespan of these mice through improvements in ventricular ectopy, function, and myocardial fibrosis/inflammation [114]. Another cardiac-restricted knockout DSG2 model was developed [115]. Qiu et al. later showed that activation of PPARɑ provided a cardioprotective effect through its phosphorylation of STAT3 and SMAD3. This inhibited cardiac fibrosis in ACM, suggesting a possible treatment for DSG2-mediated dACM [116]. Altogether, these findings are consistent with DSG2 mutations leading to ACM; however, it is noteworthy that a reduction in DSG2 by > 50% was required for overexpression of DSG2 N271S to lead to ACM, at least in mouse models. In summary, these studies showed that DSG2 mutations often affect the left or both ventricles in humans, compromise sodium channel and connexin distributions, which promote arrhythmias, and adversely affect ICD formation.

Desmocollins (DSCs) are type I transmembrane glycoproteins and members of the cadherin cell adhesion molecule superfamily. These proteins are encoded by three genes (DSC1, DSC2, and DSC3) located as part of two clusters on chromosomes 18q12 [117]. Each gene encodes a pair of proteins with a larger “a” form and shorter “b” form with differences in their C-terminus. Full length DSC2 encodes a protein that is 901 amino acids in length; however, gene transcripts vary in length due to two initiation sites of transcription and two splice variants. Only two protein isoforms are formed from these transcripts due to alternative splicing at exon 16. DSC3 has a very short variant of 214 amino acids that has an incomplete coding sequence. Functionally, DSC proteins provide tensile strength to the desmosomes, and they act as molecular sensors and mediators of signal transduction [118]. DSC1 and DSC3 proteins are prevalent in epidermis and some other tissues. DSC2 is the only desmocollin isoform present in CMs, and the translated protein is localized mainly to the desmosomes within the ICD.

DSC2 variants are generally rare in patients with ACM. Mutations in this gene predominantly affect left ventricular electrical activity during early stages of the disease; however, a pathogenic role of these mutations in the right ventricular form of dACM has also been consistently documented (1–2% of cases) [119]. Gehmlich et al. reported two missense (R203C and T275M) mutations, one of which led to a premature termination codon (predicted to not produce a functional protein), and another one causing a frameshift mutation (A8a97fsX900). The prematurely terminated protein likely led to haploinsufficiency, while the other led to defects in proteolytic cleavage at the N-terminal cadherin domains. As a consequence, the DSC2 proteins fail or only partially localize to the desmosomes of the ICD [120]. In a 58-year-old male who had ventricular arrhythmias with a left bundle branch block, Heuser et al. identified a heterozygous splice-acceptor-site mutation in intron 5 (c.631–2A>G) of DSC2, which leads to a cryptic splice-acceptor site and the creation of a downstream premature termination codon [121]. Gerull et al. reported a homozygous founder mutation in DSC2 in two large families of the Alberta Hutterite population that led to a truncation mutation (c.1660C>T). Immunostaining of endomyocardial biopsies confirmed the presence of truncated DSC2 protein at the ICD [6]. The subgroup of affected individuals maintained a carrier frequency of 9.4% (1 in 10.6) among 1535 Schmiedeleut Hutterites [6]. Brodehl et al. reported a four-base pair (bp) DSC2 deletion (c.1913_1916delAGAA, p. Q638LfsX647^hom) that caused a frameshift mutation in dACM patients and resulted in a loss of heterozygosity with segmental interstitial uniparental isodisomy [122]. Through transmission electron microscopy, the ultrastructure of the myocytes from this patient displayed a widened ICD (i.e., gap) located in the left ventricular myocardium [122].

DSC2 mutations have also been studied in human cells and animal models. The use of patient buccal mucosa cells with DSC2 mutations gave results that were remarkably similar to those observed for DSG2, but distinct from those reported for PKP2 [96]. To better understand the consequence of the splice-acceptor-site mutation in intron 5 of the DSC gene in humans described above, Heuser et al. cloned the zebrafish ortholog of DSC2 and used morpholinos to target the translation start site and the splice-acceptor sites for exons 6 and 11. The morpholino phenotypes showed a significant dose-dependent bradycardia with chamber dilation, pericardial edema, and abnormal cardiac contractility. Reduced desmosome plaque areas and loss of desmosomal midlines in developing embryos were also observed. Importantly, these phenotypes could be rescued through co-injection of the morpholinos with wild-type human DSC2 mRNA, but not by injection of mutant human DSC2 mRNA [121]. In transgenic mice with cardiac-restricted overexpression, tagged DSC2 proteins localized primarily to the ICD; however, some remained cytoplasmic. These mice developed severe myocardial necrosis, as well as fibrosis and calcification of both ventricles and the septum. An acute inflammatory response involving chemokine, cytokine, or toll-like receptor signaling, as well as macrophage infiltration into the myocardium were also reported [123]. Neither DSC2 heterozygous (+/G790del) nor homozygous (G790del/G790del) mice showed structural or functional defects in the right ventricle or developed lethal arrhythmias. Only at six-months of age did the homozygous mutant mice show modest left ventricular dysfunction with decreased cell shortening and prolonged intracellular Ca²⁺ transients. Spontaneous Ca²⁺ transients were also observed in response to isoproterenol [124]. These results in human and animal models provide direct evidence that DSC2 mutations can lead to disrupted ICD and DSC2 proteins that are linked with dACM phenotypes. More recently, Pohl et al. investigated the in vitro impact of prodomain variants of the DSC2 protein. Prodomains and conserved positions within the prodomain are thought to be important for the subcellular transport of DSC2 to the plasma membrane. hiPSC-CMs were transfected with prodomain variants and immunostained with wheat germ agglutinin, α-actinin, DSC2, calnexin (endoplasmic reticulum), and N-acetylgalactosaminyltransferase 2 (Golgi apparatus). Variants (p.D30N, p.V52A/I, P.G77V/D/S, p.V79G, and p.I96V/T) were analyzed via confocal microscopy for their plasma membrane localization. The imaging revealed that variants P.G77V/D/S and p.V79G were expressed but did not localize to the plasma membrane. Instead, these two variants remained in the endoplasmic reticulum or Golgi apparatus, showing that prodomain variants in DSC2 could prevent subcellular transport to the plasma membrane [125]. In summary, DSC2 mutations in humans affect left and right ventricles, and from animal model studies, desmosomal plaques are reduced, mice develop necrosis, fibrosis, and calcification of both ventricles and the septum, and calcium handling is affected.

Desmoplakin (DSP) is encoded by a single gene located on chromosome 6p24 that gives rise to two isoforms: DPI and DPII, which are composed of 2871 and 2272 amino acids, respectively. DPI is the predominant form present in the heart. The protein functions as a homodimer with a dumbbell-shaped conformation [126]. Unique to this desmosomal protein is the presence of a “plakin domain”, comprising six spectrin repeat domains separated by SH3 domains. The plakin family consists of proteins characterized by a multimodular structure that enables them to function as cross-linkers of the cytoskeleton (microfilaments, microtubules, and intermediate filaments). This fosters cytoskeletal component interactions with each other and with junctional complexes (adhesion molecules) on the plasma membrane to help control cell shape and polarity through modulation of cytoskeletal dynamics. These domains also help regulate processes like cell adhesion, migration, polarization, or signaling pathways. The N-terminal globular head domain is composed of a-helical bundles required for localization of the protein to the desmosome, where it interacts with the N-terminal region of PKP2 and with the C-termini of DSC and DSG2. The middle region of DSP contains a coiled-coil rod domain responsible for homodimerization. The C-terminal is composed of three plakin repeat domains (A, B, and C) required for cytoskeletal component binding.

DSP pathogenic variants have been associated mostly with left ventricular dominant dACM; however, recessive forms have also been described. The first reported cases of autosomal recessive DSP mutation were by Kaplan et al. They identified a homozygous point mutation (7901delG), which led to a premature stop codon and truncation of the protein in the C-terminal portion of the molecule [127]. This truncation caused generalized striated keratoderma and dilated left ventricular CMP that ultimately evolved into biventricular cardiomyopathy. Histological assessments of a patient’s heart revealed a unique CMP characterized by ventricular hypertrophy with dilatation, ultrastructural abnormalities of ICD with decreased amounts of DSP, JUP, and Cx43, and a failure of desmin to localize to the ICD. There was, however, no evidence of fibrofatty infiltration. Autosomal dominant forms of DSP mutations in dACM have also been described. Rampazzo et al. identified a mutation (S299R) in exon 7 of DSP that modifies a putative phosphorylation site in the N-terminal domain binding JUP [128]. This mutation resulted in biventricular dilation associated with woolly hairs and palmoplantar keratoderma [128]. Bauce et al. subsequently reported two novel missense mutations (R1775I, R1255K) and one intron–exon splicing region mutation (c.423-1G>A -intron 3) in four families that demonstrated left ventricular involvement and a high occurrence of sudden cardiac death (SCD) [129]. Patients experienced myocardial enzyme release and chest pain with ST segment elevation in their electrocardiograms and had a mutation in DSP that was identified as c.423-1G>A. Sen-Chowdhry et al. later published a cohort study of 200 patients with dACM probands. These individuals underwent cardiac magnetic resonance imaging, which indicated high prevalence of left ventricular involvement with dACM and a higher rate of ventricular arrhythmias [130].

DSP mutations in modified cell lines and transgenic mouse models have established the role of this protein in ACM. Garcia-Gras et al. found that DSP null mice experienced early embryonic lethality [131] as did transgenic lines with cardiac deletion of DSP [51]. Although most cardiac null DSP mice showed growth arrest at E10-E12, some DSP^−/− mice survived the embryonic period but died, usually within the first 2 weeks post-partum. Heterozygote adults experienced premature death, displayed enlarged cardiac chambers, and had poorly organized myocytes with large areas of patchy fibrosis. Fat droplet accumulation was also observed predominantly at the site of fibrosis. Target genes of canonical Wnt/β-catenin signaling were decreased in these animals, while transcripts for adipogenic genes (C/EBPa and adiponectin were increased). Garcia-Gras et al. also utilized HL-1 cells expressing siRNAs against DSP. Suppression of DSP transcripts led to nuclear localization of JUP and a two-fold reduction in canonical Wnt/β-catenin signaling. The cells had increased transcripts for adipogenic and fibrogenic genes, as well as accumulation of fat droplets. Huang et al. utilized Myh6-Cre mice to selectively delete DSP in CMs. These mice showed growth arrest at ~E11.5, similar to prior observations by Garcia-Gras et al. [51,132]. Stevens et al. found that homozygote DSP (R452G) mice died during embryonic development prior to E10, whereas heterozygote mice with this mutation survived. Adult mice did not display any changes in cardiac structure or function, nor did they have baseline arrhythmias or electrical abnormalities in the absence of stress. However, with pressure overload induced by transverse aortic constriction, the mutant mice progressed to heart failure more quickly than normal. These mutant mice exhibited chamber dilation with reduced fractional shortening. They also developed more severe T-wave inversions and displayed a fragmented pattern in the QRS complex. Catecholaminergic (epinephrine) challenge resulted in an increased prevalence and severity of arrhythmias, which included ventricular tachycardia, bigeminy, a higher risk of atrioventricular block, and premature ventricular contractions. Morphologically, these animals showed altered Cx43 localization, which were disrupted further following stress. These findings highlight the role of cardiac stress in the development of dACM disease phenotypes [133]. Finally, another study demonstrated the role of DSP in Carvajal-Huerta syndrome. A frameshift mutation at 38,288,978 bp of chromosome 13 in the DSP gene led to a C-terminus truncation and a phenotype consistent with this syndrome and caused abnormal ruffled hair, epidermal blistering, abnormal ECGs, and ventricular fibrosis [134]. In summary, patients with DSP mutations have clinical features that include arrhythmias of left ventricular origin, T wave inversions, and prominent regional myocardial fibrosis in the left ventricle as well as autosomal-dominant and recessive inheritance traits. The animal models also show defects in desmosome protein localization, intracellular signaling cascades, electrical abnormalities, and enhanced fat deposition.

Plakoglobin (JUP) was one of the first desmosomal genes reported to be associated with dACM in human patients. It is encoded on human chromosome 17q21 and translates into a cytoplasmic protein of 745 amino acids. JUP is a member of the armadillo superfamily, and it contains 12 armadillo repeats, flanked by N- and C-terminal domains. Functionally, it links cadherins to the actin cytoskeleton, is essential for the normal development of ICD, and influences the arrangement and function of the cytoskeleton and the arrangement of cells within a tissue. It is the only protein known to be a common constituent of sub-membranous plaques. JUP forms links in these plaques with IFs and may contribute to desmoplakin and desmosomal cadherin protein interactions. JUP also may be important for cross-talk between AJs and desmosomes [135,136].

The number of patients that suffer from JUP mutations is relatively small. A homozygous deletion was identified via genetic linkage analysis to be the cause of the autosomal recessive Naxos disease. Specifically, a two-base pair (bp) deletion in JUP on chromosome 17q21 (PK2157del2) led to a premature stop codon [137] with loss of myocyte integrity and junction disruption, leading to cell death and fibrofatty replacement. Like Carvajal-Huerta syndrome, these mutations led to cardiocutaneous phenotypes characterized by woolly hair and palmoplantar keratoderma. Histologically, the hearts had a typical pattern of dACM with fibrofatty replacement of the right (mostly subepicardial and midmyocardial layers) and left ventricles. Surviving myocytes were surrounded by fibrotic tissue and embedded within fatty tissue and Cx43 was reduced at the ICD. Autosomal dominant mutations for JUP have also been described. Asimaki et al. reported a JUP mutation at S39_K40insS in a patient of German descent. From a right ventricular biopsy, they observed reductions in plakoglobin localization, JUP, and Cx43 [138]. Using human embryonic kidney (HEK) cells and transfection of an expression vector containing the mutation, they reported that, relative to controls, the transfected HEK cells showed increased proliferation, displayed lower rates of apoptosis, and had reduced numbers and sizes of desmosomes. These data suggest that this mutation disrupted the mechanical integrity of cells. Lui et al. reported a 24-year-old male admitted to the hospital for syncope during sports activity who had an electrocardiogram with inverted T-waves [139]. Sanger sequencing revealed an autosomal dominant mutation of c.1729C>T/p.R775C of JUP. Using AC16 human cardiomyocytes, derived from the fusion of primary cells from human ventricle with SV40 transformed uridine auxotroph human fibroblasts, a mutant AC16 CM cell line (R577C) was developed and compared with control cells. Their results indicated decreased expressions of DSG2 and Cx43, which was speculated to contribute to the disruption of desmosomes and intermediate junction stability.

Plakoglobin animal models have been instrumental to understanding of how mutations in this gene promote dACM. Ruiz et al. showed that JUP null mice die between E10 and E16 due to cardiac defects [135]. Pericardial cavities of mutant embryos were often swollen, filled with blood, and the heart walls were often ruptured, but still contracting. Histologically, the ICD was grossly altered, with typical desmosomes no longer detectable. In place of the desmosomes, AJ plaques were prominent. Interestingly, typical desmosomes were present in epithelial organs, suggesting that JUP is essential for the normal development of myocardial desmosomes and proper formation of the ICD. In a study on ten-month-old heterozygous JUP-deficient mice compared with wild-type siblings, Kirchoff et al. showed that mutant mice exhibited increased right ventricular volume, reduced right ventricular function, and spontaneous ventricular ectopy [140]. There was no reported change to size and function of the left ventricle in the affected mutants. Isolated, perfused mutant hearts demonstrated spontaneous ventricular tachycardia and had prolonged right ventricular conduction times. Fabritz et al. studied the effects of endurance training (7 weeks of daily swimming) on wild-type and littermate heterozygous plakoglobin-deficient mice [141]. Mutant mice demonstrated right ventricle enlargement relative to the wild-type mice. Mouse hearts were subsequently isolated and perfused, and these hearts exhibited ventricular tachycardias and reduced right ventricular longitudinal conduction velocity. Histology on the excised hearts revealed reduced myocardial JUP but no changes to CDH2. In a zebrafish JUP knockdown model, Martin et al. developed and implemented a morpholino against the AUG region common to plakoglobin-1a and 1b. Zebrafish with the mutation demonstrated decreased heart size, reduced heartbeat, cardiac edema, and reflux of blood between the developing heart chambers. Lombardi et al. reported the overexpression of a truncated form of JUP (23654del2) [142]. In this mouse model, truncated JUP mutants demonstrated fibrofatty development, cardiac dysfunction, and premature death. Cardiac progenitor cells isolated from mutant mice were found to have increased adipogenesis, increased adipogenic factors, and reduced levels of adipogenesis inhibitors. Li et al. reported the development of mice with CM-restricted ablation of JUP [143]. Plakoglobin-deficient mutants experienced SCD as early as 1 month of age and typically had an average lifespan of 4.6 months. Mutant hearts also had many of the features observed in the clinic, including ventricular dilation, cardiac fibrosis, cardiac dysfunction, ventricular aneurysm, and spontaneous ventricular arrhythmias. These data are all consistent with JUP, like the other desmosomal proteins, having a critical role in the formation of desmosomes and ICD in myocardial tissues, an influence on cardiac electrical conduction, and a disease phenotype restricted, at least initially, to the right ventricle.

Desminopathies, caused by mutations in DES, represent one of the most common IF human disorders, some of which have been linked directly to ACM [12,144,145,146,147]. It is located on chromosome 2q35 and is prevalent in heart and in skeletal muscle. It is not, however, typically considered to be part of the desmosome [146,148], but people include a short discussion of this type III IF protein in ACM since it binds DSP and links the desmosomes to numerous intracellular CM components, such as the myofibril Z-discs, cell nucleus, mitochondria, and several other organelles [149,150,151,152]. One of the earliest discovered DES mutations linked to ACM is a missense N116S mutation, which resulted in disturbance of the aggresomes of skeletal muscle fibers, possibly due to impaired desmin filament formation [144]. In a meta-analysis of 159 patients with various DES variants, 60% of the cohort experienced cardiac conduction disease or arrhythmias with atrioventricular block [153]. Bermúdez-Jiménez et al. found that the largest known family with a p.E401D mutation in this gene primarily exhibited familial ACM, suggesting that the prevalence of mutant DES ACM may be underestimated. Histological and molecular analyses of these tissues revealed abnormal cell growth, reduced desmin mRNA, and disrupted ICD [146], the last of which is suggestive of dACM. Further insights into the pathogenesis of desminopathies and, by extension, desmin-related ACM, have been made possible through the development of a knockout DES murine model, which many laboratories have used to recapitulate desmin-related CMP phenotypes [154]. Using a desmin null mouse model, researchers have identified a link between a compromised desmin network and dysregulated mitochondrial function that leads to oxidative stress and cell death [155].