Orai1 deficiency leads to heart failure and skeletal myopathy in zebrafish

Calcium signaling plays a fundamental role in muscle cells, including regulation of growth and differentiation, metabolism and gene expression. In addition, altered calcium signaling plays an important role in the pathophysiology of cardiac diseases such as myocardial infarction and pathological hypertrophy. Ca²⁺-release-activated Ca²⁺ entry (CRAC) is a established mechanism for nonexcitable cells to replenish intracellular calcium stores and for subsequent Ca²⁺-dependent signaling cascades (Feske, 2007). Calcium-release-activated calcium channel protein 1 (ORAI1), the calcium-selective pore that mediates CRAC current (Prakriya et al., 2006; Vig et al., 2006), was shown to be expressed in isolated cardiac myocytes and is an important regulator involved in hypertrophic growth in neonatal cardiomyocytes (Voelkers et al., 2010). ORAI1 mutations or deficiencies cause myopathy (McCarl et al., 2009). ORAI1-deficient mice die of unknown cause shortly after birth, and the role of ORAI1 in intact hearts is unknown (Gwack et al., 2008; Vig et al., 2008). To examine the role of ORAI1 in the intact heart in vivo, we used reverse genetics in zebrafish, an intriguing model organism for evaluating effects of gene loss on cardiovascular function (Hassel et al., 2009; Nasevicius and Ekker, 2000; Rottbauer et al., 2005; Sehnert et al., 2002), because the absence of blood flow for several days does not affect development of other organs (Shin and Fishman, 2002). We provide evidence that Ca²⁺ entry through ORAI1 is necessary to maintain sarcomere integrity and establish proper z-disc function, in part through regulation of the localization of the calcineurin-associated protein calsarcin. Our results have identified a mechanism by which excitable cells spatially control Ca²⁺-dependent signaling events, independently from excitation contraction coupling.

Immunohistochemistry of mice hearts at different stages showed ubiquitous ORAI1 expression in embryonic tissues, including the heart (Fig. 1A) (Gwack et al., 2008; Vig et al., 2008). Orai1 mRNA expression was largely decreased at 2 months compared with 2 days after birth (supplementary material Fig. S1A). In adult hearts, ORAI1 localized to the plasma membrane of cardiomyocytes and was highly expressed in the vasculature (Fig. 1B). In addition, ORAI1 was detected using a proximity ligation assay in adult hearts (Fig. 1C) as well as in isolated adult cardiomyocytes (supplementary material Fig. S2C), suggesting that ORAI1 is expressed throughout development and in adult cardiomyocytes.

To evaluate the role of ORAI1 in cardiac function in zebrafish, we identified the zebrafish orthologue (zOrai1 for zebrafish Orai1), which shows high evolutionary conservation to the mouse (77%) and human (76%) proteins (supplementary material Fig. S1B). Using whole-mount in situ hybridization, we detected zorai1 transcripts ubiquitously in various tissues, including heart, throughout development (Gwack et al., 2008; Vig et al., 2008). At 24 hours postfertilization (hpf), zorai1 was expressed ubiquitously (not shown). By 48 hpf, zorai1 expression was more restricted, with pronounced expression in brain, liver, somites and heart (Fig. 1D). This pattern was maintained at 72 hpf, with diminishing expression in somites (Fig. 1E).

Next, we inactivated Orai1 by injecting morpholino-modified antisense oligonucleotides targeting the splice donor site of exon 1 (MO-ORAI1) or the translational start site (MO-ORAI1^ATG) into one-cell zebrafish embryos (supplementary material Fig. S1C). Identical phenotypes were obtained with both morpholino oligonucleotides, whereas embryos injected with a control oligonucleotide developed normally. MO-ORAI1-injected embryos were indistinguishable from controls on day 1 of development. By 48 hpf, Orai1-deficient embryos developed severe heart failure, decreased blood circulation and blood congestion at the inflow tract (Fig. 1F–I; supplementary material Movie 1), and most failed to hatch, indicating reduced skeletal muscle force generation. Ventricular contractility, determined by ventricular fractional shortening, was reduced at 48 hpf and further reduced at 72 hpf (Fig. 1J; supplementary material Movie 2). Orai1-deficient embryos developed severe bradycardia by 48 hpf overlapping with onset of reduced contractility (Fig. 1K). Maturation of chamber myocardium proceeded normally with proper growth of ventricular myocardium, and atrial and ventricular cardiomyocytes expressing myosin heavy chains in the regular chamber-specific pattern (Fig. 1L–O; supplementary material Fig. S2B). Furthermore, Orai1-depleted hearts properly expressed cardiac marker genes (supplementary material Fig. S2A). Hence, ORAI1 deficiency caused severe heart failure, bradycardia and skeletal muscle weakness without affecting early cardiogenesis and cardiomyocyte differentiation.

In mice, ORAI1-deficiency causes muscle weakness (Gwack et al., 2008; Vig et al., 2008) and perinatal death, probably from myopathy. Mice lacking STIM1, the calcium sensor of the endoplasmic and sarcoplasmic reticulum required to activate ORAI1, also develop skeletal myopathy and severe immunodeficiency, suggesting that CRAC signaling is essential for normal muscle function (Stiber et al., 2008). All humans with ORAI1 mutations have skeletal myopathy and usually die from severe immunodeficiency and lethal infections by 1 year of age (Feske et al., 2006; McCarl et al., 2009). No assessment of cardiac performance in such patients has been reported so far. We demonstrate for the first time that loss of ORAI1 causes heart failure, without affecting initial myocyte differentiation.

Many skeletal myopathies cause structural and ultrastructural abnormalities in myocytes. We determined whether skeletal muscle weakness in Orai1-deficient embryos reflected altered muscle organization. Immunofluorescence analysis of MO-ORAI1-injected embryos revealed frequent disruption of myofilaments and detachment of myofibrils from myosepta at 48 hpf, with increased severity at 72 hpf, leading to myofiber-free spaces in somites, similar to that seen in mutants deficient for dystrophin–glycoprotein-complex-associated proteins or integrin-linked kinase (Cheng et al., 2006; Guyon et al., 2005; Postel et al., 2008) (Fig. 2A–D). Additionally, at 72 hpf, myofiber density within the somites appeared to be reduced.

Detachment of myofibers is often accompanied by loss of myoseptal integrity (Etard et al., 2010; Parsons et al., 2002; Postel et al., 2008). To evaluate whether loss of Orai1 function affects myoseptal integrity, we performed immunostaining against vinculin, an adaptor protein linking muscle fibres, through integrins, to the extracellular matrix (Critchley, 2000; Postel et al., 2008). At 48 hpf, vinculin level was reduced in myosepta of Orai1-deficient embryos, and vinculin localization at the somite boundaries was disturbed and present ectopically in somites (Fig. 2E,F), indicative for loss of myoseptal integrity (Etard et al., 2010; Postel et al., 2008). At 72 hpf, a stage with advanced myofiber disruption, vinculin levels further declined and mislocalization increased in Orai1-deficient skeletal muscle. Myosepta frequently appeared as tattered, blurry stripes rather than distinct, thin lines separating adjacent somites, as in controls. Thus, loss of Orai1 in zebrafish caused progressive disruption of myofibers and loss of myoseptal integrity, suggesting that Orai1 function is not required for initial muscle assembly but essential for maintaining and straightening of muscle structure. The decreasing myofiber density in somites at 72 hpf suggests that reduced Orai1 disturbs muscle fiber growth and thereby adaptation of muscle force to higher demands during development.

To elucidate the mechanism that effects muscle integrity, we analyzed the ultrastructure of skeletal muscle cells in MO-ORAI1-injected embryos. At 72 hpf, myofibrils in controls were densely packed with precisely aligned and highly organized thick and thin filaments delimited by well-defined z-discs (Fig. 2I). By contrast, even though organized sarcomeric units were generally present, the myofibrils appeared thin and instead of being densely packed, individual myofibrils were separated by cytoplasm (Fig. 2J). Frequently, myofibrils changed from being well organized to disorganized within one to two sarcomeric units, with progressively dispersing thin and thick filaments (Fig. 2J). Additionally, instead of the well-contrasted and distinct z-discs seen in control embryos, the z-lines in MO-ORAI1-injected skeletal muscle cells appeared faint.

To our knowledge, ORAI1-deficient myocytes have not been thoroughly analyzed. Consistent with loss of ORAI1, ultrastructural evaluation of STIM1-deficient skeletal muscle similarly showed muscle atrophy (Stiber et al., 2008). Additionally, we found that Orai1 inactivation caused abnormalities of myofibril organization and of the z-line in skeletal and cardiac muscle in zebrafish, characteristics not described for STIM1-deficient mice.

We speculated that defects similar to those in skeletal muscle cells might explain the decreased contractility in Orai1-deficient hearts. Cardiomyocytes from MO-ORAI1-injected embryos similarly had fewer, thinner myofibrils than control cardiomyocyte (Fig. 2K,L). The z-discs between sarcomeres in Orai1-deficient cardiomyocytes were stretched, blurry and often seemingly absent. Thus, Orai1 deficiency induces similar defects in skeletal and cardiac myocytes.

The fact that correctly assembled, although thinner and less densely packed, sarcomeres were present in both skeletal and cardiac tissues suggests that Orai1 loss does not impair initial sarcomerogenesis but disrupts myofibril growth, possibly limiting adaptation of muscle force to increasing demands during development. Our data on vinculin, as well as the aggravation of myofiber disruption, further suggests that ORAI1-deficiency affects sarcomere stability in part by altering the integrity of adhesion complexes linking muscle fibers through adaptor proteins to the extracellular matrix. Because z-discs are crucial for sarcomere stability, the abnormalities in Orai1-deficient skeletal and cardiac z-discs might contribute to the instability and disaggregation of myofibrils (Hassel et al., 2009).

Phenylephrine-induced hypertrophic growth of neonatal rat ventricular cardiomyocytes (NRCM) represents an excellent model to study myofibril addition and sarcomere growth in vitro. To test our hypothesis that ORAI1 is necessary for sarcomeric growth, we decreased ORAI1 expression in neonatal rat ventricular cardiomyocytes with ORAI1-specific short interfering RNA (siRNA). At 48 hours after transfection, ORAI1 expression was reduced by 61% (supplementary material Fig. S3A,B). Treatment with phenylephrine increased cell surface area 2.3-fold at 24 hours together with a remarkable organization and growth of the sarcomeric structure (supplementary material Fig. S3C,D). However, siOrai1-treated cells failed to organize sarcomeres and hypertrophic growth was considerably attenuated (1.4-fold increase of the cell surface area over baseline).

Next, we analyzed the z-disc structure of neonatal cardiomyocytes by immunostaining for actinin and calsarcin. ORAI1 is implicated in regulating calcium–calcineurin–NFAT signaling. Calcineurin binds to calsarcins, a family of striated muscle-specific proteins located at the sarcomeric z-disc, and calsarcin-deficiency results in z-disc remodeling (Frank et al., 2007). In siOrai1-transfected cardiomyocytes, sarcomeres were thinner and less organized than in controls (Fig. 3A). Interestingly, calsarcin failed to localize to z-discs, and global calsarcin expression was decreased (Fig. 3A,B). These findings support our in vivo data in that the Ca²⁺ entry through the ORAI1 channel is essential for proper sarcomeric growth and function in cardiomyocytes, and furthermore indicate that ORAI1 is required for localization of the calcineurin–NFAT signaling component calsarcin to the cardiac z-disc. This finding is supported by the notion that analysis of the nuclear localization of NFAT using an EGFP–NFAT adenovirus is prevented by ORAI1 knockdown (KD; Fig. 3C). In addition the induction of the ‘fetal gene program’ is blunted in ORAI1 KD cardiomyocytes (Fig. 3D,E).

Calsarcin depletion was previously shown to result in the activation of calcineurin–NFAT-mediated hypertrophic gene transcription (Frey et al., 2004). Because calsarcin inhibits calcineurin signaling, global downregulation of calsarcin might represent a compensatory attempt to restore defective calcineurin signaling after ORAI1 knockdown in vivo. Importantly, calsarcin interacts with important components of the z-disc, including telethonin (also known as T-Cap), Cypher (ZASP) and myotilin, all of which can cause cardiomyopathy or muscular dystrophy when dysfunctional, highlighting the importance of calsarcin–calcineurin–NFAT regulation by ORAI1 (Frank et al., 2006; Hayashi et al., 2004; Moreira et al., 2000; Olive et al., 2005; Selcen and Engel, 2005; Vatta et al., 2003). Further analysis in adult cardiomyocytes lacking ORAI1 are needed to determine the molecular mechanism in more detail, but are hindered by the perinatal death of ORAI1-deficient mice (Gwack et al., 2008).

Besides cross-linking thin filaments and stabilizing sarcomeres by transmission of force generated during contraction, z-discs are nodal points for mechanotransduction (i.e. stretch sensing) (Frank et al., 2006). Given the ultrastructural abnormalities and mislocalization of calsarcin (an important Ca²⁺-signaling component) in ORAI1-deficient z-discs, we hypothesized that ORAI1 is required to establish a functioning, properly signal-transducing z-disc. Therefore, we assessed expression of the stretch-responsive gene encoding atrial natriuretic factor (anf) by whole-mount RNA in situ hybridization (Bendig et al., 2006). At 48 hpf anf was markedly upregulated, and still slightly increased at 72hpf (Fig. 3F–I), indicative of cardiac dysfunction and impaired stretch sensing in zebrafish (Bendig et al., 2006). Calcineurin interacts directly with the mechanosignaling transduction protein muscle-LIM-protein (MLP), and calsarcin and calcineurin are displaced from the z-disc upon MLP depletion (Heineke et al., 2005). Animals lacking calcineurin or calsarcin express higher levels of Anf, endorsing our data in zebrafish (Frey et al., 2004; Schaeffer et al., 2009).

Our results show that ORAI1 is necessary for muscle growth and physiological adaptation to increasing performance requirements. To determine whether ORAI1 signaling contributes to pathological aspects of hypertrophy, we analyzed ORAI1 expression in mice after myocardial infarction or transaortic constriction (TAC) to induce pressure overload-mediated hypertrophy. Sham-operated mice served as controls. After TAC, ORAI1 protein and mRNA increased significantly (P<0.001; Fig. 4A–C) and the fetal gene program was reactivated (data not shown). ORAI1 expression also increased after myocardial infarction (Fig. 4D–F). Thus, ORAI1 might contribute to the pathology of hypertrophy, even in humans, although the molecular mechanism is unknown.

In summary, we confirmed that loss of Orai1 causes skeletal muscle weakness in zebrafish (Gwack et al., 2008; McCarl et al., 2009). Structural and ultrastructural analysis revealed that the weakness is due to myofibrillar disruption and progressive mechanical instability in skeletal myocytes, both distinctive features of muscular dystrophy in humans and zebrafish (Emery, 2002; Postel et al., 2008). Furthermore, we show, for the first time, that Orai1 deficiency in zebrafish causes heart failure. Sarcomeres in Orai1-deficient hearts exhibit lateral growth and ultrastructural abnormalities at the cardiac z-disc with decreased expression and displacement of calsarcin from the z-disc. Loss of Orai1 also caused defective signal transduction at the cardiac z-disc. Our findings link ORAI1-mediated calcium signaling to sarcomere physiology, in part by affecting z-disc composition and function mediated by the calcineurin-calsarcin-NFAT signaling pathway. Thus, the molecular machinery mediating SOCE might represent a valuable pharmacological target to modulate Ca²⁺-dependent signaling to improve the outcome of pathological cardiac hypertrophy.

Care and breeding of zebrafish (Danio rerio) was carried out as previously described (Westerfield, 1995). Morpholino-modified antisense oligonucleotides were designed against the translational start site of zebrafish ORAI1 [NM_205600] (MO-ORAI1ATG 5′-AGTGCTCGCTCCGACTCATCTTCAT-3′) and the splice donor site of zebrafish ORAI1 exon 1 (MO-ORAI1 5′-AAACAGCGCGGAGACTCACCATTGC-3′). Functional assessment of cardiac contractility was carried out essentially as described previously (Rottbauer et al., 2005).

Surgical procedures on mice have been described previously (Muraski et al., 2007). Briefly, myocardial infarction was produced by ligating the left anterior descending coronary artery with an 8-0 suture (Ethicon). For trans-aortic constriction (TAC), the aorta was ligated between the innominate artery and the left common carotid artery by using a 7-0 polypropylene suture (Ethicon) with an overlying 27-gauge needle to produce a discrete stenosis. All animal experiments were performed according to the relevant regulatory standards.

Whole-mount in situ hybridization and immunofluorescence staining of zebrafish embryos was carried out essentially as described previously (Hassel et al., 2009). The following antibodies were used: MF20 (1:20; Developmental Studies Hybridoma Bank, developed by D. A. Fishman) and anti-vinculin (1:50; Sigma). 72 hpf control and MO-ORAI1-injected embryos were used for histological analysis of the heart. Histological sections were prepared as described previously (Rottbauer et al., 2001). Transmission electron microscopy analysis was carried out as described by Kurrasch et al. (Kurrasch et al., 2009).

Immunohistochemistry of mouse sections have been described before (Muraski et al., 2007). In brief, hearts were flushed with formalin for 15 minutes, excised, and fixed in formalin for 24 hours at room temperature. Sections were cut and deparaffinized using standard procedures. Primary antibodies were applied overnight at 4°C in nitro blue tetrazolium. The Orai1 signal was detected with a Tyramide Signal Amplification kit (Perkin Elmer), used according to the manufacturer's recommendations. To reveal nuclei, slides were treated with To-pro3 iodide (Molecular Probes; 1:10,000) for 20 minutes in 1× TN (Tris-based neutral buffer), washed, and a coverslip place on top. Micrographs were acquired with a Leica TCS-SP2 confocal laser scanning microscope. Neonatal rat ventricular cardiomyocytes (NRCMs) grown on glass coverslips were treated with Orai1 or scrambled siRNA as described before (Voelkers et al., 2010). After 48 hours cells were fixed, permeabilized and stained using standard procedures. A proximity ligation assay to detect ORAI1 was used. Blocking, staining, hybridization, ligation, amplification and detection steps were performed according to manufacturer's instructions (OLINK).

Whole-cell lysates of hearts and NRCM lysates were prepared in 1× SDS sample buffer and immunoblotted. Primary antibodies used were against GAPDH (Invitrogen) and Orai1 (Santa Cruz Biotechnology). Calsarcin antibody was a gift from Norbert Frey (Kiel, Germany). The day after application of the antibodies, blots were washed three times with Tris-buffered saline plus 0.1% Tween-20 (TBS-T), probed with fluorescent or alkaline phosphatase–conjugated secondary antibodies (1:2000 in blocking solution; Jackson Immunoresearch Laboratories) for 2 hours at room temperature, washed three times with TBS-T, and scanned.

Ventricular cardiomyocytes from 1- to 2-day-old rat neonatal hearts (NRCMs) were prepared by trypsin digestion as described previously (Voelkers et al., 2010). Adult ventricular cardiomyocytes were isolated using standard procedures as previously described (Most et al., 2006). In a subset of experiments NRCMs transfected with the siRNAs were transfected on the same day with an EGFP–NFAT adenovirus and the cellular localization of EGFP–NFAT were analyzed 48 hours after transfection and 24 hours after treatment with phenylephrine.

Custom-designed synthetic orai1 siRNA and scrambled siRNA as a negative control were from Applied Biosystems. NRCMs were transfected with orai1 and control siRNA oligonucleotides (25 nM) by using HiPerfect transfection reagent according to the manufacturer's instructions (Qiagen).

Total RNA was isolated from frozen heart or cultured cells with Quick-RNA MiniPrep (Zymo Research) and reverse-transcribed into cDNA with an iScriptcDNA Synthesis kit (Bio-Rad). Real-time PCR was performed on all samples in triplicate with a QuantiTect SYBR Green PCR kit (Qiagen) according to the manufacturer's instructions. All primer sequences are shown supplementary material Table S1.

For RT-PCR analysis, total RNA from hearts derived from control and MO-orai1-injected embryos at 72 hpf was extracted using Trizol reagent (Invitrogen). RNA was subsequently treated with DNase and reverse transcribed using a random hexamer primer and SuperScript II reverse transcriptase (Invitrogen). Gene-specific primer sequences are available upon request. Amplification involved 35–40 cycles. Elongation factor I alpha (elfa) was used as endogenous control (McCurley and Callard, 2008).

We thank Jinny Wong for excellent technical support, Gary Howard for outstanding editorial support with the manuscript and Norbert Frey for the anti-calsarcin antibody.