Cardiac‐specific overexpression of caveolin‐3 preserves t‐tubular I Ca during heart failure in mice

New Findings What is the central question of this study? What is the cellular basis of the protection conferred on the heart by overexpression of caveolin‐3 (Cav‐3 OE) against many of the features of heart failure normally observed in vivo? What is the main finding and its importance? Cav‐3 overexpression has little effect in normal ventricular myocytes but reduces cellular hypertrophy and preserves t‐tubular I Ca, but not local t‐tubular Ca2+ release, in heart failure induced by pressure overload in mice. Thus Cav‐3 overexpression provides specific but limited protection following induction of heart failure, although other factors disrupt Ca2+ release. Abstract Caveolin‐3 (Cav‐3) is an 18 kDa protein that has been implicated in t‐tubule formation and function in cardiac ventricular myocytes. During cardiac hypertrophy and failure, Cav‐3 expression decreases, t‐tubule structure is disrupted and excitation–contraction coupling (ECC) is impaired. Previous work has suggested that Cav‐3 overexpression (OE) is cardio‐protective, but the effect of Cav‐3 OE on these cellular changes is unknown. We therefore investigated whether Cav‐3 OE in mice is protective against the cellular effects of pressure overload induced by 8 weeks’ transverse aortic constriction (TAC). Cav‐3 OE mice developed cardiac dilatation, decreased stroke volume and ejection fraction, and hypertrophy and pulmonary congestion in response to TAC. These changes were accompanied by cellular hypertrophy, a decrease in t‐tubule regularity and density, and impaired local Ca2+ release at the t‐tubules. However, the extent of cardiac and cellular hypertrophy was reduced in Cav‐3 OE compared to WT mice, and t‐tubular Ca2+ current (I Ca) density was maintained. These data suggest that Cav‐3 OE helps prevent hypertrophy and loss of t‐tubular I Ca following TAC, but that other factors disrupt local Ca2+ release.

2018b). Disruption of Cav-3 signalling with C3SD peptide (Couet, Li, Okamoto, Ikezu, & Lisanti, 1997;Feron et al., 1998) also decreases t-tubular I Ca (Bryant et al., 2014), which impairs local SR Ca 2+ release (Bryant et al., 2014;Bryant et al., 2018a). Interestingly, cardiac hypertrophy and heart failure (HF) are associated with decreased Cav-3 expression (Bryant et al., 2018a), and myocytes from such hearts also show hypertrophy, t-tubular disruption, decreased t-tubular I Ca density and impaired SR Ca 2+ release (Bryant et al., 2015;Bryant et al., 2018a), suggesting that reduced Cav-3 expression may play a role in the phenotypic changes observed in these conditions. In support of this idea, Cav-3 KO results in a progressive cardiomyopathy characterized by ventricular hypertrophy and dilatation and reduced fractional shortening (Woodman et al., 2002), while a loss-of-function mutation in Cav-3, T63S, has been associated with inherited hypertrophic cardiomyopathy (Hayashi et al., 2004). In addition, overexpression (OE) of Cav-3 reduces the functional and phenotypic changes caused by pressure overload induced by transverse aortic constriction (TAC; Horikawa et al., 2011;Markandeya et al., 2015), which normally results in cardiac hypertrophy and failure. However, the cellular changes underlying this cardioprotection remain unclear.
The present study was undertaken, therefore, to investigate how Cav-3 OE alters the response of ventricular myocyte structure and ECC to TAC in mice.

Ethical approval
All animal procedures were approved by the Animal Welfare and Ethics
Data from these mice, bred at the University of Bristol, are compared with data obtained from wild-type (WT) littermates that had undergone either sham operation or TAC that resulted in HF. The WT data have been published previously (Bryant et al., 2018a) so that only mean data, rather than original records, are shown for the WT group, for comparison with the OE data. However, the surgical and experimental procedures were the same, and performed contemporaneously, for the WT and Cav-3 OE groups, and data were obtained using the same techniques and protocols in each group, as described below.
Surgery was performed at 12 weeks of age and myocyte isolations at 20 weeks of age. Mice were kept in a temperature-controlled, enriched environment with ad libitum access to food and water.
Eight weeks of TAC was used to produce pressure overload, since this has previously been shown to result in cardiac hypertrophy and

• What is the central question of this study?
What is the cellular basis of the protection conferred on the heart by overexpression of caveolin-3 (Cav-3 OE) against many of the features of heart failure normally observed in vivo?
• What is the main finding and its importance?
Cav-3 overexpression has little effect in normal ventricular myocytes but reduces cellular hypertrophy and preserves t-tubular I Ca , but not local t-tubular Ca 2+ release, in heart failure induced by pressure overload in mice. Thus Cav-3 overexpression provides specific but limited protection following induction of heart failure, although other factors disrupt Ca 2+ release.
The aortic arch was exposed via a medial sternal thoracotomy and a silk ligature (6-0) placed between the innominate and left carotid arteries and tied round a 27G needle (0.4 mm OD). Sham animals underwent the same operation but without placement of the banding suture.
Animals were maintained post-operatively for 8 weeks before use.

Echocardiography
In vivo cardiac structure and function were monitored using echocardiography. Animals were anaesthetized (isoflurane 1-3%, Merial Animal Health Ltd, Harlow, UK), heart rate was monitored, and measurements of contractile performance made from M-mode images acquired from the parasternal short axis view using a Vevo 3100 (Fujifilm VisualSonics Inc., Toronto, Ontario, Canada) and MX550D transducer.

Myocyte isolation and detubulation
Animals were killed by cervical dislocation and ventricular myocytes isolated using standard enzymatic digestion via Langendorff perfusion as described previously (Bryant et al., 2014) and used on the day of isolation. Detubulation (DT), the physical and functional uncoupling of the t-tubules from the surface membrane, was achieved using formamide-induced osmotic shock as described previously (Brette & Orchard, 2003;Brette, Komukai, & Orchard, 2002;Kawai et al., 1999); comparison of membrane capacitance and currents in intact and detubulated myocytes enables the distribution of membrane currents and current density between the t-tubule and surface membranes to be determined.

Imaging and analysis of t-tubule structure
Cell width and length were measured from brightfield images of isolated myocytes used for electrophysiology. Cell volume was calculated from these measurements as described previously (Boyett, Frampton, & Kirby, 1991).
Surface and t-tubular cell membranes were labelled by incubating cells with 5 mol l −1 di-8-ANEPPS for 10 min. Image volumes were obtained using an LSM 880 confocal microscope (Zeiss, Carl Zeiss AG, Oberkochen, Germany) in Airyscan 'super-resolution' mode, with a 1.2 NA, ×40 water immersion objective, sampled at 40 nm in-plane and 180 nm along the optical axis. Airyscan uses a 32-channel photomultiplier tube detector that collects a pinhole-plane image at every scan position, thus improving spatial resolution. In super-resolution mode, linear deconvolution provides further improvement to achieve spatial resolution that is 1.7× that of a conventional confocal microscope. The regularity of t-tubule staining was quantified by applying a two-dimensional (2D) fast Fourier transform (FFT) to an offsetsubtracted square region of the cell interior, and the power of the first harmonic normalized to that of the average image intensity (P 1 /P 0 ). Ttubule density was calculated from an intracellular volume marked by hand. The 3D skeleton of the t-tubules was obtained by processing the volumetric data with a tubule-enhancing 3D filter, segmenting using an Otsu threshold in MATLAB R2015a (The Mathworks Inc., Natick, MA, USA), and converting to a skeleton using Skeletonize (2D/3D) in ImageJ (v1.50, NIH, Bethesda, MD, USA). The skeleton was used to calculate t-tubule density (skeleton length divided by the marked intracellular volume, m m −3 ) and local Eigenvectors for t-tubule angles. Tubule orientation is expressed relative to the transverse plane, so that 0 • corresponds to a transverse tubule, while 90 • corresponds to a tubule that extends along the cell (i.e. an 'axial' tubule).

Western blotting
Following myocyte isolation, aliquots were pelleted by centrifugation, the supernatant removed and the cell pellet snap frozen in liquid nitrogen and stored at −80 • C. Once all samples had been acquired, the pellets were processed simultaneously by thawing directly into lysis buffer containing 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, complete protease inhibitors, 8 g ml −1 calpain inhibitor I, 8 g ml −1 calpain inhibitor II, 50 mM sodium fluoride, 1 mM sodium orthovanadate, and 16 mM sodium pyrophosphate, homogenized by pipetting and incubated on ice for 15 min. After centrifugation at 13,000 g 4 • C for 15 min the supernatants were collected, protein concentrations estimated using the Pierce BCA protein assay (Thermo Fisher Scientific, Waltham, MA, USA) and adjustments made to allow for equal protein loading on SDS-PAGE.
Ten-microgram samples of the myocyte lysates were run on 4-15% gradient SDS-PAGE gels and transferred onto Immobilon-P membrane.
Gels were first probed with the antibody to Cav-3 or JPH-2, then stripped using a commercial stripping solution (Restore TM western blot stripping buffer, Thermo Fisher Scientific) and re-probed with the loading control antibody to GAPDH, before being stripped and reprobed for JPH-2 or Cav-3. The density of the bands was measured using ImageJ and normalized to GAPDH.

I Ca recording
Myocytes were placed in a chamber mounted on a Nikon Diaphot inverted microscope. Membrane currents and cell capacitance were recorded using the whole-cell patch-clamp technique, using an Axopatch 200B, Digidata 1322A A/D converter and pClamp 10 (Molecular Devices, LLC, San Jose, CA, USA). Pipette resistance was typically 1.5-3 MΩ when filled with pipette solution (see below), and pipette capacitance and series resistance were compensated by ∼70% to optimize the measurement of membrane current. Currents were activated from a holding potential of −80 mV by step depolarization to −40 mV for 200 ms (to inactivate the sodium current) followed by steps to potentials between −50 and +80 mV for 500 ms, before repolarization to the holding potential, at a frequency of 0.2 Hz.
Absolute I Ca amplitude (pA) in intact myocytes was measured as the difference between peak inward current and current at the end of the depolarizing pulse; absolute I Ca in the t-tubular and surface membranes was calculated from measurements of I Ca and membrane capacitance in intact and detubulated myocytes with correction for incomplete detubulation as described previously (Bryant et al., 2015).
I Ca was normalized to cell capacitance (pF; an index of membrane area) to calculate I Ca density (pA pF −1 ). I Ca density in the t-tubule membrane was calculated from the loss of membrane current and capacitance following DT; I Ca density in the surface membrane was calculated from currents measured in DT myocytes with correction for incomplete detubulation as described previously (Bryant et al., 2015).
DT efficiency was not significantly different between cell types.

Measurement of SR Ca 2+ release
Intracellular Ca 2+ and membrane potential were recorded simultaneously along single t-tubules as described previously (Bryant et al., 2015). Briefly, myocytes were loaded with the Ca 2+ indicator Fluo-4/AM (5 mol l −1 for 25 mins; Thermo Fisher Scientific) and the voltage sensitive dye di-4-AN(F)EPPTEA (0.5-1 g ml −1 for 15 min; kindly supplied by Dr Leslie Loew; Yan et al., 2012). Cells were imaged using a Zeiss LSM 880 (see above) with the confocal pinhole set to 1 Airy unit. Line-scan images along a selected t-tubule were recorded at wavelengths between 518 and 560 nm for Ca 2+ , and 590 and 700 nm for voltage, at a rate of 0.51 ms/line, with an excitation wavelength of 514 nm. A negative deflection in di-4-AN(F)EPPTEA fluorescence was used to determine the time of the AP upstroke, and the latency from the AP upstroke to the initial (>5 SD above average pre-stimulus value) and maximum rate of rise of Ca 2+ was determined at each point along the Fluo-4 line-scan image. The SD of latencies for each cell was used as a measure of the heterogeneity of release. Whole-cell Ca 2+ transients were obtained using line-scans along the long axis of cells loaded with Fluo-4/AM only. Cells were field-stimulated at 0.2 Hz at 1.5 × threshold using parallel Pt electrodes.

Data presentation
Data are expressed as mean ± SD (of N animals for in vivo data and of n cells from N animals (n/N) for cellular measurements). Data normality was assessed using the Shapiro-Wilk test and subsequent testing was performed using Student's t test or the Mann-Whitney U test, one-way ANOVA, or the Kruskal-Wallis test, as appropriate.
I Ca density-voltage relationship curves were analysed using repeated measures (RM) ANOVA with voltage and intervention as factors. Single myocyte properties including those elicited by a step depolarization to a single voltage were analysed with two-way ANOVA; post hoc tests used the Bonferroni correction. The errors in derived variables (specifically I Ca density at the t-tubule and surface membranes), and the subsequent statistical analysis (unpaired Student's t test), were calculated using propagation of errors from the source measurements (Bryant et al., 2015). The limit of statistical confidence was taken as P < 0.05, and is denoted by * between treatments (e.g. sham vs.
TAC) for a given phenotype and by † between phenotypes (WT vs.

The effect of TAC on cardiac structure and function
Cardiac structure and function were assessed in vivo using echocardiography; exemplar records from Cav-3 OE mice are shown in Figure 1a; WT data have been shown previously (see Methods).
These data suggest, therefore, that Cav-3 OE per se has little effect on cardiac structure or function, but reduces the hypertrophy observed following TAC. The magnitude of these changes was similar to that reported previously in Cav-3 OE mice following TAC (Horikawa et al., 2011, and see below).  (f) Heart weight to tibia length ratio (HW:TL, mg/mm). (g) Lung weight to tibia length ratio (LW:TL, mg/mm). N = 6 sham and 6 TAC Cav-3 OE mice, and 7 sham and 7 TAC WT mice. ***P < 0.001 between treatments for a given phenotype (WT or Cav-3 OE); †P < 0.05 between phenotypes for a given treatment (sham or TAC). The WT data have been published previously (Bryant et al., 2018a) n/N = 41/10, TAC 192.0 ± 57.9 m, n/N = 21/5, P < 0.01; Cav-3 OE: Analysis of t-tubule structure using 2D FFT showed a significant decrease in t-tubule regularity in both WT and Cav-3 OE myocytes following TAC (P 1 /P 0 , Figure 2d). Since P 1 /P 0 depends, inter alia, on t-tubule density and orientation, further detailed analysis in 3D was performed, which showed that this reduction in regularity was due, at least in part, to a significant decrease in t-tubule density (Figure 2e) and changes in tubule orientation, with a decrease in the fraction and T-tubule orientation (degrees from transverse plane; f) in WT and Cav-3 OE myocytes. ***P < 0.001 between treatments for a given phenotype (WT or Cav-3 OE); †P < 0.05 between phenotypes for a given treatment (sham or TAC). The WT data have been published previously (Bryant et al., 2018a) of transverse (0-15 • ) tubules (Figure 2f). However, the changes in t-tubule structure were not significantly different between WT and Cav-3 OE myocytes.

The effect of TAC on cell morphology
To ensure that Cav-3 expression was increased in OE mice, we used western blotting. Figure 3a shows exemplar blots (left) and corresponding mean densitometric analysis (right) showing that Cav-3 expression was increased ∼2-fold in OE compared to WT myocytes.
Cav-3 expression was not significantly altered following TAC in Cav-3 OE myocytes, so that its expression level remained higher than in sham and TAC WT myocytes. Since t-tubule structure was altered following TAC in Cav-3 OE myocytes, we also investigated the expression of JPH-2, which has been implicated in t-tubule and dyad formation. In summary, Cav-3 OE appears to have little effect on cardiac and cell morphology, and in vivo cardiac function, and the response to TAC was qualitatively similar in WT and Cav-3 OE mice; however, both heart weight and cell volume were significantly smaller in Cav-3 OE mice than in WT, following TAC.

I Ca distribution and regulation
To determine the distribution of I Ca between the surface and t-tubular membranes, I Ca was measured in intact and DT myocytes. Figure 4 shows exemplar records of I Ca recorded at 0 mV from intact (Figure 4a, top) and DT (Figure 4b, top) myocytes isolated from sham and TAC Cav-3 OE hearts, with the corresponding mean current density-voltage relationships shown below. Absolute I Ca and I Ca density were not significantly different in WT and Cav-3 OE myocytes (Figure 4c,d).
However, the increase in cell capacitance caused by TAC in WT mice

F I G U R E 3 Effect of Cav-3 OE and TAC on Cav-3 and JPH-2 protein expression. (a) Left panel, exemplar western blots of Cav-3 (18 kDa) and GAPDH (37 kDa) in isolated myocyte lysates from three WT sham (lanes 1), three Cav-3 OE sham (lanes 2) and three Cav-3 OE TAC (lanes 3) mice;
Right panel, densitometry analysis of Cav-3 western blots (N = 5 animals in each group in duplicate), compared with previously published data (Bryant et al., 2018a) showing the effect of TAC on Cav-3 expression in WT mice (left two bars). (b) Left panel, exemplar western blots of JPH-2 (95 kDa) and GAPDH (37 kDa); right panel, densitometry analysis of JPH-2 western blots (N = 5 animals in each group in duplicate), compared with previously published data (Bryant et al., 2018a) showing the effect of TAC on JPH-2 expression in WT mice (left two bars). The blots in the left panels are from the same gels which were stripped and re-probed for the different proteins and were therefore obtained sequentially (see Methods). Data in each group (WT or Cav-3 OE) in the right panels are expressed as a percentage of the mean of the WT data in that group. **P < 0.01, between treatments for a given phenotype (WT or Cav-3 OE); † †P < 0.01, † † †P < 0.001 between phenotypes the surface membrane of WT myocytes following TAC, so that I Ca density at the surface membrane is unaltered following TAC in these cells (Figure 4c,d). The lack of change of I Ca density in intact Cav-3 OE myocytes following TAC, despite a decrease at the surface membrane, suggests that I Ca density at the t-tubule membrane is increased.
Calculation of t-tubular I Ca showed that absolute I Ca is significantly increased so that I Ca density is maintained in the t-tubules of Cav-3 OE myocytes following TAC (Figure 4c,d), in contrast to the lack of change of absolute I Ca and thus decrease in t-tubular I Ca density observed in WT myocytes following TAC (Figure 4c,d). Thus, I Ca density at the cell surface decreases and t-tubular I Ca density is preserved in Cav-3 OE mice following TAC, whereas WT mice show no change at the cell surface and a decrease in t-tubular I Ca density in response to TAC.
We have previously shown that incubating cells with C3SD, which mimics the scaffolding domain of Cav-3 and has no effect on cell capacitance (Bryant et al., 2014), decreases I Ca density in intact WT control myocytes (Bryant et al., 2018a;Kong et al., 2017), and that this effect is lost following Cav-3 KO and in TAC-induced HF (Bryant et al., 2018a). To investigate whether changes in this regulatory pathway might underlie the different I Ca distributions observed following TAC in OE mice, we determined the response to C3SD in myocytes from sham and TAC Cav-3 OE mice. Figure 4e,g show that in sham Cav-3 OE myocytes I Ca density was increased by pre-treatment with C3SD, as reported previously in myocytes from unoperated Cav-3 OE mice (Kong et al., 2017) but in contrast to the decrease of I Ca density observed in sham WT myocytes in response to C3SD (Bryant et al., 2018a; Figure 4g). However, C3SD had no effect on I Ca density in Cav-3 OE myocytes following TAC (Figure 4f,h); this loss of response to C3SD following TAC is similar to that reported in WT myocytes following TAC (Bryant et al., 2018a;Figure 4h). This suggests that this regulatory pathway is lost in both cell types and is not, therefore, responsible for the different distribution of I Ca density observed in OE and WT myocytes following TAC. Two-way repeated measures ANOVA: mV P < 0.001, TAC P < 0.001, interaction P < 0.001. (c,d) Absolute I Ca (c) and I Ca density (d) at 0 mV in intact Cav-3 OE myocytes, and at the cell surface and t-tubules (obtained as described in Methods), compared with previously published data (Bryant et al., 2018a) from WT mice. **P < 0.01, ***P < 0.001 between treatments for a given phenotype (WT or Cav-3 OE); †P < 0.05, † †P < 0.01, † † †P < 0.001 between phenotypes for a given treatment (sham or TAC). (e) Mean I Ca density-voltage relationships from sham untreated (n/N = 19/5) and C3SD treated (n/N = 12/3) Cav-3 OE myocytes. Two-way repeated measures ANOVA: mV P < 0.001, C3SD P = 0.014, interaction P < 0.001). (f) Mean I Ca density-voltage relationships from TAC untreated (n/N = 22/5) and C3SD treated (n/N = 17/5) Cav-3 OE myocytes. Two-way repeated measures ANOVA: mV P < 0.001, C3SD P = 0.13, interaction P = 0.4. (g) Mean I Ca density at 0 mV in sham untreated and C3SD treated Cav-3 OE myocytes, compared with previously published data (Bryant et al., 2018a; see Methods) from sham untreated (n/N = 16/5) and C3SD treated (n/N = 17/5) WT myocytes. Two-way ANOVA: OE P < 0.001, C3SD P = 0.4, interaction P < 0.001. (h) Mean I Ca density at 0 mV in TAC untreated and C3SD treated Cav-3 OE myocytes, compared with previously published data (Bryant et al., 2018a; see Methods) from TAC untreated (n/N = 19/5) and C3SD treated (n/N = 15/5) WT myocytes. Two-way ANOVA: OE P < 0.001, C3SD P = 0.14, interaction P = 0.84. *P < 0.05, **P < 0.01 between treatments for a given phenotype (WT or Cav-3 OE); † † †P < 0.001 between phenotypes for a given treatment

Ca 2+ release following TAC
To determine whether the preservation of t-tubular I Ca in Cav-3 OE myocytes helps to maintain Ca 2+ release, we investigated the latency and heterogeneity of Ca 2+ release along a single t-tubule from the time of membrane depolarization (Figure 5a), which showed that Cav-3 OE had no significant effect on latency, nor did it affect the increase in latency of both the initial and maximum rate of rise of Ca 2+ observed following TAC (Figure 5b, top), suggesting that TAC-induced impairment of local Ca 2+ release is unaffected by Cav-3 OE. However, Cav-3 OE caused a significant decrease in the heterogeneity of Ca 2+ release compared to WT myocytes, and although heterogeneity increased following TAC in both cell types, it remained smaller in Cav-3 OE myocytes (Figure 5b, bottom) suggesting more uniform Ca 2+ release along the t-tubules following Cav-3 OE. However, TAC had little effect on the early (release) phase of the whole cell Ca 2+ transient, which was not significantly different in WT and Cav-3 OE myocytes and showed no change in either time to peak or amplitude (Figure 5c). Figure 6 shows the ratio of the data from Cav-3 OE mice to those from WT mice, in sham (left) and in mice in HF following TAC (right). (c) Whole cell Ca 2+ transient amplitude (F/F 0 ) and time to peak (ms) measured from Ca 2+ transients of sham (n/N = 25/5) and TAC (n/N = 22/4) Cav-3 OE myocytes compared with WT sham (n/N = 53/13) and TAC (n/N = 14/3) myocytes. *P < 0.05, **P < 0.01, ***P < 0.001 between treatments for a given phenotype (WT or Cav-3 OE); †P < 0.05, † †P < 0.01 between phenotypes for a given treatment (sham or TAC). The WT data have been published previously (Bryant et al., 2018a) significantly smaller in OE than in WT, consistent with previous work (Horikawa et al., 2011;Markandeya et al., 2015). In addition, I Ca amplitude and density were greater in intact OE than in WT myocytes, as a result of a small decrease at the cell surface but a large increase at the t-tubules. Thus, the decrease in t-tubular I Ca normally observed following TAC (Bryant et al., 2018a) is prevented by Cav-3 overexpression, and although Cav-3 OE exerts an anti-hypertrophic effect and increases I Ca , these effects only become apparent following TAC.

DISCUSSION
The present data show that TAC caused qualitatively similar changes in Cav-3 OE mice to those reported previously in WT mice (Bryant et al., 2018a): cardiac hypertrophy and failure, with disrupted cell structure and function. However, the cardiac and cellular hypertrophy associated with TAC were smaller in OE animals and t-tubular I Ca density was maintained, although I Ca density at the surface membrane decreased; this contrasts with the decrease in t-tubular I Ca density with no change at the cell surface observed in WT myocytes. Thus Cav-3 OE appears to confer limited but specific protection against the effects of TAC.

Cardiac structure and function
Echocardiography showed that Cav-3 OE had little effect on cardiac structure or function in sham animals ( Figure 1). Following TAC,  (Bryant et al., 2018a). The x-axis shows the change relative to WT: an increase in Cav-3 OE mice compared to WT results in a value >1, while a decrease results in a value <1. The coloured bands delineate different groups of data that correspond to those shown in Figs 1 (green, top) to 5 (grey, bottom). The data in red are significantly different in Cav-3 OE and WT mice (statistical analysis performed using original data). LVVD, left ventricular volume at diastole, l; LVVS, left ventricular volume at systole, l; stroke volume, l; ejection fraction, %; HW:TL, heart weight to tibial length ratio, mg mm −1 ; LW:TL, lung weight to tibial length ratio, mg mm −1 ; cell volume, pl; t-tubule power, P 1 /P 0 ; t-tubule density, m m −3 ; Cav-3, caveolin-3 expression, %; JPH-2, junctophilin expression, %; intact capacitance, whole cell capacitance, pF; intact I Ca , whole-cell I Ca amplitude at 0 mV, pA; intact I Ca density, whole-cell I Ca density at 0 mV, pA pF −1 ; surface capacitance, surface membrane capacitance, pF; surface I Ca , surface membrane I Ca amplitude at 0 mV, pA; surface I Ca density, surface membrane I Ca density at 0 mV, pA pF −1 ; t-tubule capacitance, t tubule membrane capacitance, pF; t-tubule I Ca , t-tubule membrane I Ca amplitude at 0 mV, pA; t-tubule I Ca density, t-tubule membrane I Ca density, pA pF −1 ; Ca 2+ latency, calcium transient latency from action potential depolarization, ms; Ca 2+ heterogeneity, calcium transient heterogeneity during upstroke, ms; Ca 2+ amplitude, calcium transient amplitude, F/F 0 ; Ca 2+ time to peak, ms in HW:TL and LW:TL in Cav-3 OE mice. However, although cardiac function was not significantly different in WT and Cav-3 OE mice following TAC, the hypertrophic response to TAC was smaller in Cav-3 OE than in WT mice. These changes are similar to those reported previously following 4 weeks TAC in Cav-3 OE mice (Horikawa et al., 2011) and suggest that Cav-3 OE has maintained anti-hypertrophic effects following TAC, consistent with the idea that Cav-3 inhibits the hypertrophic p42/44 mitogen-activated protein kinase (Woodman et al., 2002) and calmodulin-dependent calcineurin/nuclear factor of activated T cells (Markandeya et al., 2015) pathways.
Following TAC, cardiac function was not significantly different, and there were similar decreases in ejection fraction, in WT and Cav-3 OE mice. However, stroke volume and cardiac output decreased significantly in Cav-3 OE but not in WT mice following TAC. This implies that an increase in heart size in WT (cf. the larger scatter in Figure 1b) compared to Cav-3 OE mice helped to maintain stroke volume and thus cardiac output despite similar decreases in ejection fraction. Thus, the anti-hypertrophic effect of Cav-3 OE may impair the ability of the heart to maintain cardiac output.
Previous work has shown that 4 weeks' TAC caused only a small (not significant) decrease in cardiac function in Cav-3 OE mice, but a significant decrease in WT mice (Horikawa et al., 2011;Markandeya et al., 2015). The contrast with the current work may be due to the longer (8 weeks) exposure to TAC in the present study. Taken together these data suggest that, while the anti-hypertrophic effect of Cav-3 is maintained, the deleterious effect on cardiac function, while slowed in onset, can still occur in Cav-3 OE mice following TAC. However, the observation that Cav-3 OE has little effect on either the size or the function of the heart in the absence of TAC ( Figure 6; Horikawa et al., 2011;Markandeya et al., 2015) suggests that Cav-3 expression is normally sufficient to inhibit hypertrophy and enable normal ECC.

Cell structure
In the absence of TAC, Cav-3 OE had little effect on cell size or structure, as reported previously (Kong et al., 2017), but cells were larger and t-tubule structure was disrupted following TAC, as in WT myocytes (Bryant et al., 2018a). However, the increase in cell volume following TAC was significantly smaller in OE than in WT myocytes, providing a mechanism for the reduced hypertrophy observed in the whole heart and consistent with the suggestion that caveolin inhibits hypertrophic pathways (Galbiati et al., 1998;Markandeya et al., 2015;Woodman et al., 2002). Knockout and loss-of-function mutations of Cav-3 are associated with hypertrophic cardiomyopathy, further supporting a role for Cav-3 as an inhibitor of cardiac hypertrophic signalling pathways (Hayashi et al., 2004;Woodman et al., 2002).

Cell function
Cav-3 OE had little effect on the function of myocytes from sham hearts. However, the distribution of I Ca was markedly different in OE and WT myocytes following TAC. TAC-induced heart failure in WT mice is associated with a decrease in t-tubular I Ca density due to cellular hypertrophy with no change in absolute current, and no change in I Ca density at the cell surface (Bryant et al., 2018a). In contrast, in OE myocytes, there was no change in t-tubular I Ca density following TAC, because absolute current increased proportionally with membrane area, but I Ca density at the cell surface decreased, since absolute I Ca was unchanged despite cellular hypertrophy. Thus, following TAC, Cav-3 OE maintains I Ca density at the t-tubular membrane but not at the cell surface.
Previous work has shown that chronically decreasing Cav-3 expression, via either KO or TAC, leads to decreased t-tubular I Ca density as a result of an increase in membrane area (Bryant et al., 2018a), whereas acute disruption of Cav-3 activity using C3SD peptide decreases I Ca with no change of capacitance (Bryant et al., 2018a;Kong et al., 2017). However, in agreement with previous work (Kong et al., 2017), the present data show that C3SD increases I Ca in OE myocytes although, as in WT myocytes, this regulation was lost following TAC.
These data suggest that Cav-3 alters I Ca density by (at least) two mechanisms. The first is by altering cell growth. It has previously been suggested that Cav-3 is anti-hypertrophic (Woodman et al., 2002), the t-tubules, and C3SD decreases t-tubular I Ca in WT myocytes, although this effect is lost following TAC (Bryant et al., 2018a). It has been suggested that Cav-3-dependent stimulation of I Ca is due to co-localization of LTCCs with components of the protein kinase A pathway (Bryant et al., 2014). However, recent work has shown LTCC clustering leading to co-operative gating (Ghosh et al., 2018); if Cav-3 plays a role in this process, loss of Cav-3 activity would be expected to decrease I Ca . The present study shows that C3SD increases I Ca in sham Cav-3 OE myocytes, suggesting that the levels of Cav-3 expression achieved in OE myocytes may inhibit these Cav-3 dependent pathways.
Such inhibition could occur as the result of autoinhibition or because there is abnormally located Cav-3 in OE myocytes which competes with normally localized proteins. Reducing Cav-3 expression, activity or regulation, via TAC or C3SD, may relieve this inhibitory effect to produce the increases in I Ca observed in these conditions; the loss of effect of C3SD on I Ca following TAC is also indicative of such loss of Cav-3 dependent regulation.
In contrast to t-tubular I Ca , I Ca density at the cell surface decreased following TAC in Cav-3 OE mice, due to cellular hypertrophy with little change in absolute I Ca . Thus, following TAC, in OE myocytes t-tubular I Ca appears to be maintained at the expense of I Ca at the surface membrane, whereas in WT myocytes t-tubular I Ca density decreases with no change at the cell surface. This suggests that changes in Cav-3 expression occur predominantly at the t-tubules, where it competes with that at the surface membrane to bind the proteins which localize I Ca .
Regardless of the mechanism underlying the maintenance of ttubular I Ca , the latency and heterogeneity of local Ca 2+ release at the t-tubule still increased following TAC which, with the disruption of ttubule morphology also observed following TAC, would be expected to desynchronize Ca 2+ release and thus impair contraction. However, the whole cell Ca 2+ transient showed little change, suggesting that it is dominated by other factors and is not responsible for the impaired cardiac performance observed following TAC. However, the Ca 2+ transient was monitored at a low stimulation frequency (0.2 Hz) compared with the mouse's normal heart rate, to enable comparison with I Ca , which was recorded at this frequency to allow recovery from inactivation between voltage clamp pulses. It remains possible, therefore, that changes in local Ca 2+ release may decrease Ca 2+ transient amplitude, and thus impair cardiac performance, at physiological frequencies. However, the impairment occurred in the presence of increased heart size, so that reduced stroke volume and ejection fraction do not necessarily imply reduced contractility because myocyte contraction will have to overcome the higher wall tension that will result from the increase in heart size (law of Laplace).
The lack of effect of Cav-3 OE on the latency of Ca 2+ release despite the recovery of t-tubular I Ca suggests that the decrease in I Ca is not the primary cause of disruption of Ca 2+ release following TAC. This may be because of redundancy in the Ca 2+ -induced Ca 2+ release process (Cannell, Berlin, & Lederer, 1987) in mice, and/or because the LTCCs, and thus I Ca , which are regulated by Cav-3 are predominantly extradyadic (Glukhov et al., 2015;Sanchez-Alonso et al., 2016), which could also explain why similar decreases in I Ca amplitude produced by Ca 2+ channel blockers, which will affect all Ca 2+ channels, inhibit release (Bryant et al., 2015). The TAC-induced disruption of Ca 2+ release may,