Mitsugumin 29 regulates t-tubule architecture in the failing heart

Transverse tubules (t-tubules) are uniquely-adapted membrane invaginations in cardiac myocytes that facilitate the synchronous release of Ca2+ from internal stores and subsequent myofilament contraction, although these structures become disorganized and rarefied in heart failure. We previously observed that mitsugumin 29 (Mg29), an important t-tubule organizing protein in skeletal muscle, was induced in the mouse heart for the first time during dilated cardiomyopathy with heart failure. Here we generated cardiac-specific transgenic mice expressing Mg29 to model this observed induction in the failing heart. Interestingly, expression of Mg29 in the hearts of Csrp3 null mice (encoding muscle LIM protein, MLP) partially restored t-tubule structure and preserved cardiac function as measured by invasive hemodynamics, without altering Ca2+ spark frequency. Conversely, gene-deleted mice lacking both Mg29 and MLP protein showed a further reduction in t-tubule organization and accelerated heart failure. Thus, induction of Mg29 in the failing heart is a compensatory response that directly counteracts the well-characterized loss of t-tubule complexity and reduced expression of anchoring proteins such as junctophilin-2 (Jph2) that normally occur in this disease. Moreover, preservation of t-tubule structure by Mg29 induction significantly increases the function of the failing heart.

t-tubules adjacent to RyR2 in the apposed SR. When crossed into the well-characterized Csrp3 −/− mouse model of dilated cardiomyopathy, the Mg29 transgene was protective, increasing t-tubule organization and preserving both cardiac contraction and relaxation. Double Sypl2 −/− Csrp3 −/− mice showed a further disorganization of t-tubules and even worse cardiac function compared with Csrp3 −/− mice alone. These data suggest that induction of Mg29 in the diseased heart serves a compensatory role in maintaining t-tubule organization and preserving cardiac function.

Results
Mg29 expression characteristics in the mouse heart. We previously identified that high-expressing myocyte enhancer factor-2A (MEF2A) cardiac-specific transgenic mice develop heart failure by 3 weeks of age and showed induction of Mg29 mRNA expression 21 , a known skeletal muscle-specific gene 10 . We subsequently identified Mg29 mRNA as induced in the hearts of mice following transverse aortic constriction (TAC), myocardial infarction (MI) injury, in transgenic mice expressing activated calcineurin (CnA), and strongly in Csrp3 −/− mice lacking muscle LIM protein (MLP), which is a model of dilated cardiomyopathy (Fig. 1a) 22 . In keeping with the well-characterized loss of t-tubule structure during heart failure, we identified that mRNA for junctophilin-2 (Jph2), which is a major cardiac t-tubule anchoring and structural protein 6,7 , was significantly downregulated in mouse hearts lacking Csrp3 −/− , as well as hearts from mice overexpressing CnA, and in both scar and remote regions of mouse hearts 2 weeks after MI surgery. Notably, however, pressure overload induced by either 2 or 12 weeks of pressure overload induced by transverse aortic constriction (TAC) surgery had no effect on Jph2 mRNA expression in the mouse heart (Fig. 1a).
To model this induction of Mg29 expression in the failing mouse heart we utilized a cardiac-specific binary transgenic system in which the presence of both a tetracycline-transactivator (tTA) protein under the control of the α-myosin heavy chain (α-MHC) promoter and a Mg29 responder transgene together produces overexpression of the Mg29 protein in the absence of doxycycline (DOX), as described previously (Fig. 1b) 23 . Hearts from double transgenic (DTG) mice showed high levels of Mg29 expression similar to native levels in the quadriceps muscle (Fig. 1c). However, this high line developed heart failure on its own so it was not used. DTG mice on DOX chow that had "leak" expression were used for most subsequent experiments because this lower level of expression was still robust for the heart, showing approximately 14-fold greater expression of Mg29 message in the heart compared with WT controls (Fig. 1a) although it was completely innocuous at 5 months of age (as will be discussed later).
Low-level expression of Mg29 resulted in a small but significant increase in Jph2 protein levels, but had no effect on the expression of other Ca 2+ handling proteins localized to the SR or t-tubules ( Fig. 1d and Fig. S1). Since Mg29 has been described as both a t-tubule and SR-resident protein in skeletal muscle 11 , and it was previously identified as a potential regulator of RyR1 activity 24 , we measured Ca 2+ spark frequency and found no effect with Mg29 expression (Fig. 1e). To extend these findings, super-resolution confocal imaging was employed using the recently-described super-resolution radial fluctuations (SRRF) technique 25 to examine the localization of Mg29 and RyR2 in isolated cardiomyocytes from DTG mice. We observed that Mg29 was localized adjacent to RyR2 but not coincident with it (Fig. 1f). These data indicate that in the heart, Mg29 is localized to t-tubules, not the SR, and that low-level expression of Mg29 does not seem to influence RyR2 activity directly.
Mg29 expression partially preserves t-tubule structure and cardiac function in Csrp3 −/− mice. Given that Mg29 was previously described as an important regulator of t-tubule structure in skeletal muscle 12,13 , and that Csrp3 −/− mice develop heart failure 22 that is associated with reduced t-tubule architecture, we crossed Mg29 DTG mice into the Csrp3 −/− background. At 12-15 weeks of age, Csrp3 −/− mice expressing Mg29 had significantly improved t-tubule integrity compared to control Csrp3 −/− mice expressing only the single tTA transgene (Fig. 2a,d), including a reduction in longitudinal t-tubule elements that is a hallmark of heart failure (Fig. 2a,c). Surprisingly, wildtype Mg29 DTG mice alone showed increased t-tubule integrity at baseline (Fig. 2a,d) reflecting an increase in total t-tubule density (Fig. 2a-c) but not regularity. Although the increase in t-tubule density and organization observed here was subtle, it resulted in a more substantial and significant restoration in systolic function in the Csrp3 −/− background at 15 weeks of age (Fig. 3b), as well as significantly improved relaxation (Fig. 3c) with no change in resting heart rate (Fig. 3e). Csrp3 −/− mice expressing Mg29 also demonstrated trends toward reduced heart weight ( Fig. 3a) and improved fractional shortening (Fig. 3d, full data in Tables S1, S2) at 5 months of age, perhaps reflecting reduced cardiac remodeling due to the improvement in cardiac function.
Mg29 expression does not alter cardiac disease progression during pressure overload. We next sought to determine whether expression of Mg29 might affect the organization of t-tubules after 12 weeks of pressure overload induced by TAC surgery given that long-term pressure overload induces some aspects of heart failure. However, this stimulus was ultimately different from the type of dilated heart failure observed in Csrp3 −/− mice, which have a dramatic downregulation in Jph2 expression, while 12 weeks of TAC in WT mice showed no reduction in Jph2 mRNA expression (Fig. 1a). Indeed, while mice subjected to TAC surgery developed some aspects of heart failure as evidenced by significantly reduced cardiac fractional shortening (Fig. 4e, full data in Tables S3, S4), increased heart weight (Fig. 4f) and pulmonary edema (Fig. 4g), these mice showed relatively little remodeling of t-tubule structure (Fig. 4a) perhaps due to the continued presence of Jph2. Both the density of transversely-oriented tubule elements and the combined index of t-tubule integrity were significantly decreased but not different with or without Mg29 transgene expression ( Fig. 4b-d). Consistent with these results, Mg29 expression did not benefit cardiac remodeling/function after TAC as it did in Csrp3 −/− mice ( Fig. 4e-g). These data suggest a paradigm in which Mg29 expression is only beneficial in a disease context where the native t-tubule organizing protein Jph2 is significantly reduced and t-tubular organization is more dramatically altered. . Hearts subjected to either 2 or 12 weeks of transverse aortic constriction (TAC) surgery showed induction of Mg29 but no change in Jph2 levels. Mg29 Low transgenic hearts (described below) showed approximately 14-fold induction of Mg29 message compared to WT control hearts. Hearts from 6 control mice, 6 two-week TAC mice, 10 twelve-week TAC mice, 5 Csrp3 −/− mice, 6 CnA transgenic mice, 5 two-week MI mice, and 4 Mg29 Low transgenic mice were utilized in this experiment. *P < 0.05 compared to expression in control mice. (b) Schematic representation of double transgenic (DTG) system used to express Mg29 in the mouse heart. (c) Immunoblot demonstrating expression of Mg29 protein in the DTG mouse heart (Mg29 High) at comparable levels to that seen in wild-type quadriceps muscle (WT Quad). Administration of DOX in the chow reduced Mg29 levels (Mg29 Low). Cropped western blots images are shown although the full-length gel files can be viewed in Supplemental Fig. 3. Photoshop was used to process the images, although they were not manipulated to alter the data. (d) Immunoblots of Ca 2+ handling proteins in DTG hearts expressing Mg29 or tTA-expressing control hearts. Phospholamban (PLN) and its phosphorylated forms at S16 and T17 were also unchanged. Cropped western blots images are shown although the full-length gel files can be viewed in Supplemental Fig. 3. Photoshop was used to process the images, although they were not manipulated to alter the data. (e) Ca 2+ spark frequency measured in myocytes isolated from DTG hearts expressing Mg29 or tTA-expressing control hearts. Number of myocytes and animals used are shown in the graph. (f) Super-resolution confocal microscopy of an isolated adult cardiomyocyte from a DTG heart showing that Mg29 (red) and RyR2 (green) are adjacent but not overlapping. Scale bar is 10 µm.

Loss of Mg29 further reduces t-tubule organization and cardiac function in
investigate if the induction of Mg29 that normally occurs in the setting of heart failure due to Csrp3 deletion was truly compensatory, we crossed Sypl2 −/− mice into the Csrp3 −/− background and assessed t-tubule organization and cardiac performance. While Sypl2 −/− mice lacking Mg29 protein alone showed no significant alterations in t-tubule organization apart from a small decrease in longitudinal element density ( Fig. 5c), mice lacking both Mg29 and MLP protein showed significantly reduced t-tubule element density compared with loss of MLP protein alone (Fig. 5a,b). They also showed increased longitudinal t-tubule element density (Fig. 5a,c) and reduced t-tubule integrity (Fig. 5a,d), all of which indicate a worsening of the heart failure phenotype. Consistent with these observations, fractional shortening was reduced even further in the double null mice (Fig. 5e, full data Tables S5, S6) and heart weight normalized to body weight (HW/BW) was significantly increased compared with loss of MLP protein alone (Fig. 5f), although pulmonary edema was not different (Fig. 5g).
To extend these studies, we again sought to determine whether Mg29 induced during long-term pressure overload (Fig. 1a) plays a role in t-tubule organization and cardiac functional performance. To that end, we subjected wild-type or Sypl2 −/− mice to 12 weeks of TAC surgery, but found that loss of Mg29 protein had no effect on cardiac remodeling or functional performance apart from a slight reduction in intraventricular septum thickness ( Figure S2). Collectively these results suggest that the induction of Mg29 that occurs in mouse models of heart failure where Jph2 expression is lost, such as in Csrp3 −/− mice, produces a compensatory response that antagonizes the progression of heart failure through a mechanism involving stabilization of t-tubules. However, in mouse models of heart failure where Jph2 expression is not reduced, augmenting Mg29 expression is of no protective value.

Discussion
Reduced t-tubule density and organization commonly accompany heart failure in human patients 2, 5 and in multiple animal models 2 . This rarefaction of t-tubule structure is likely caused by reduced expression or loss of anchoring proteins such as Jph2 that position t-tubules at the z-line 6,7 , resulting in fragmentation of the t-tubule network. T-tubule fragmentation interrupts coupling between LTCCs and RyR2 in the interior of the myocyte, leading to the dyssynchronous Ca 2+ release and reduced force generation that typifies heart failure.
Mg29 plays an important role in anchoring t-tubules in skeletal muscle 12,13 , as muscle fibers from Sypl2 −/− mice show swollen, fragmented or even absent t-tubules. While Mg29 is highly expressed in mammalian skeletal muscle we were unable to detect this protein in the mouse heart, suggesting it serves no baseline functional role in this tissue. Indeed, we found no significant difference in gravimetric measures or functional performance of the hearts of Sypl2 −/− mice up to 22 weeks of age ( Figure S2), nor were t-tubules in cardiomyocytes from these hearts overtly changed compared with wild-type controls (Fig. 5a-d). In the failing mouse heart we observed induction of Mg29 mRNA expression (Fig. 1a) 21 and both Mg29 mRNA and protein were recently shown to be upregulated in failing guinea pig hearts 20 .
The induction of Mg29 expression observed in Csrp3 −/− mice was clearly of functional importance as deletion of the Sypl2 gene further reduced t-tubule structure and organization and it worsened heart failure in Csrp3 −/− mice (Fig. 5). Another interesting aspect of our data is that even a relatively modest improvement in t-tubule structure and organization due to Mg29 induction produced a disproportionately greater rescue of cardiac function in the Csrp3 −/− background. Thus, the rarefaction or reorganization of t-tubules that occurs in the failing heart is likely of significant disease relevance. Because Mg29 also serves as a regulator of store-operated Ca 2+ entry, possibly via interaction with transient receptor potential canonical 3 (TRPC3) [14][15][16] , and as an activator of RyR1 activity 24 , we examined the localization of Mg29 using super-resolution confocal imaging 25 . Although it was previously demonstrated by immunogold electron microscopy that Mg29 is localized to both t-tubules and junctional SR in skeletal muscle 11 , in the hearts of our Mg29 DTG mice we observed strict localization only to regions adjacent to but not overlapping with RyR2, suggesting Mg29 is present in t-tubules. Thus, induction of Mg29 expression the failing heart appears to be tailored to regulating t-tubule structure and function, although it certainly could indirectly impact SR Ca 2+ release by affecting the coordination between these 2 apposed membrane surfaces. Hence, preservation of t-tubule structure should preserve normal excitation-contraction coupling and necessarily improve cardiac function just through this structural mechanism. While we did not measure Ca 2+ sparks in Csrp3 −/− hearts with transgenic overexpression of Mg29, it is possible that Ca 2+ spark frequency may even be decreased compared with Csrp3 −/− alone, as the preserved t-tubule structure found with Mg29 expression should result in fewer "orphaned" RyR2 multimers and the improved coupling could reduce the Ca 2+ sensitivity required for coordinated activation of RyR2, which may also reduce SR Ca 2+ leak.
Structurally, Mg29 is composed of a central MARVEL domain that comprises the majority of the protein (75%), flanked by shorter N-and C-terminal cytosolic domains 10,11 . MARVEL domains are also present in related proteins such as synaptophysin, which is similarly involved in membrane apposition processes and induction of membrane curvature 26 . Since our data demonstrate that Mg29 is localized to t-tubules in the heart it could serve as a structural brace to surround the t-tubule as previously hypothesized 27 . Jph2 is the major t-tubule anchoring protein expressed in the heart 6, 7 , however its expression is frequently reduced in heart failure. Indeed, our results suggest that the effects of Mg29 are only uncovered in heart failure models where Jph2 is downregulated. During long-term pressure overload where Jph2 expression is retained and t-tubules are better preserved, the transgenic expression of Mg29 has no additional effect on t-tubule structure and cardiac disease progression (Fig. 4). Thus, the primary role of Mg29 in the heart appears to be t-tubule stabilization to partially counter the loss of native anchoring proteins such as Jph2 during heart failure. Indeed, we also find that transgenic mice expressing low levels of Mg29 have increased cardiac expression of Jph2 protein ( Fig. 1d and Fig. S1), which may further aid Mg29 in preserving t-tubule structure during failure. Finally, while previous data have shown that Jph2 can stabilize t-tubule structure and cardiac function following pressure overload 28 , it also interacts with and may regulate RyR2 activity 29 , making it difficult to determine whether its role in t-tubule stabilization is solely responsible for the observed improvement in cardiac function. Our results using Mg29, which only affects t-tubules, support the notion that t-tubule stabilization has a critical role in cardiac disease progression through an obvious mechanism of regulating the coupling between the t-tubule and SR membranes to maintain coordinated Ca 2+ release for contraction. Hence, controlled expression of Mg29 in the failing heart, such as with viral gene therapy approaches, may offer a novel way to correct the reduction in coordinated coupling between the sarcolemma and SR that occurs in heart failure, similar to the strategy recently proposed for Jph2 30 .

Methods
Cloning and transgenic mice. A cDNA encoding the Sypl2 gene was cloned into the murine α-MHC promoter expression vector 23 and used to inject newly-fertilized oocytes to generate transgenic mice (FVB/N background). To test the inducibility of the transgenic system, we administered DOX chow and observed loss of greater than 50% of the Mg29 expression in the heart. Because of lethality we observed in double transgenic (DTG) mice in the absence of DOX, all experiments using transgenic mice were conducted in the presence of DOX to produce low expression of Mg29 protein, with the exception of the experiment shown in Fig. 1F, in which DOX was removed at weaning and mice were aged at least an additional 4 weeks in order to allow for a higher level of protein expression. Sypl2 null mice 13 and Csrp3 null mice 22 were previously described. For experiments in Fig. 1, mice were in a pure FVB/N background. For experiments in Figs 2 and 3, mice are in a mixed FVB/N and C57Bl/6 background. Mice in Fig. 2 were 12-15 weeks of age, mice in Fig. 3a,d,f were 20-22 weeks of age, and mice in Fig. 3b,c,e were 15-17 weeks of age. For pressure overload experiments in Fig. 4 and Figure S2, TAC surgeries were performed on mice at 8-10 weeks of age. For experiments in Fig. 5, mice were in a C57Bl/6 background. All experiments in Fig. 5 were performed at 5-6 weeks of age (except for Sypl2 −/− mice, which were examined at 6-10 weeks of age), as Csrp3 −/− mice develop more rapid disease onset in the C57Bl/6 background in our hands. Both sexes of mice were used and no animals were discarded in the statistical analysis except where noted. Experiments involving animals were approved by the Institutional Animal Care and Use Committee of Cincinnati Children's and in accordance with the National Institutes of Health Guidelines for the care and use of laboratory animals.
Western blotting. Hearts were excised from DTG mice expressing the tet-transactivator (tTA) and the inducible promoter for Mg29 expression, in either the presence or absence of DOX, frozen in liquid nitrogen and stored at −80 °C. Ventricles were homogenized in either a modified radioimmunoprecipitation assay (RIPA) buffer containing 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 7.4, and protease inhibitors or a buffer containing 20 mM Tris-HCl, pH 7.5, 250 mM NaCl, 1% Triton X-100, 10 mM MgCl 2 , 0.5 mM dithiothreitol, and protease inhibitors. Homogenates were centrifuged at 14,000 rpm for 10 min and supernatants were used for blotting. Eight to one hundred and twenty micrograms of protein was used, subjected to SDS-PAGE, and transferred to PVDF membranes. Immunoblots were performed using the appropriate primary antibody and fluorescent conjugated secondary antibodies (LI-COR, Lincoln, NE, USA) in combination with an Odyssey CLx Infrared Imaging System (LI-COR Immunofluorescence and super-resolution confocal imaging. Cardiac myocytes from adult transgenic mice were isolated as previously described 33 and imaged with a Nikon A1 confocal microscope using a GaAsP detector. Immunofluorescence was performed using primary antibodies against Mg29 (Sigma-Aldrich, St. Louis, MO, USA) and RyR2 (Abcam, Cambridge, United Kingdom). Super-resolution processing of images was accomplished using super-resolution radial fluctuations (SRRF) 25  Invasive hemodynamics and echocardiography. Hemodynamic measurements were performed as previously described 35 . Briefly, mice were anesthetized by intraperitoneal injection of pentobarbital (6 mg/100 g body weight). A high fidelity, solid state 1.2F pressure catheter (Transonic Scisense Inc, London, ON, Canada) was inserted into the left ventricle (LV) via right carotid exposure and retrograde introduction of the pressure catheter into the LV, and the animal brought to 37 °C. Data were collected by PowerLab 8/36 (ADInstruments, Sydney, Australia) and recorded using LabChart 7 Pro (ADInstruments, Sydney, Australia). Ten seconds of recorded data was averaged for each time point. Echocardiography was performed on mice after isoflurane inhalation for anesthesia (dosage was to effect) using a Hewlett Packard 5500 instrument with a 15-MHz microprobe and measurements were taken on M-mode in triplicate for each mouse and averaged. Analgesia was not used in these experiments because echocardiography is non-invasive, and hemodynamic assessment was a terminal procedure.
Pressure overload and myocardial infarction surgeries. Eight to ten week-old mice of the relevant genotypes were subjected to transverse aortic constriction (TAC) or sham surgical procedure as previously described 36 . Mice were anesthetized by isoflurane inhalation to effect. Doppler echocardiography was performed on mice after TAC in order to determine pressure gradients across the aortic constriction. Mice with pressure gradients of less than 45 mmHg and fractional shortening greater than 30% at 12 weeks post-TAC were excluded from the results, as this indicated an unsuccessful surgery. Myocardial infarction (MI) was induced via permanent surgical ligation of the left coronary artery, as previously described 37 . Both sexes of mice were used for both TAC and MI surgeries. Mice were given buprenex as pain relief at a final concentration of 0.03 mg/ml by i.p. injection. Mice were then transferred to 30 °C incubators and monitored while in recovery.

Statistics.
Results are presented in all cases as mean ± SEM. Statistical analysis was performed using Prism 5 (Graphpad Software, La Jolla, CA, USA). Experiments were analyzed using Student's t-test or one-way ANOVA with Newman-Keuls multiple comparisons test. P-values less than 0.05 were considered significant.