Restoring Atrial T-Tubules Augments Systolic Ca Upon Recovery From Heart Failure

Abstract

Background: Transverse (t)-tubules drive the rapid and synchronous Ca²⁺ rise in cardiac myocytes. The virtual complete atrial t-tubule loss in heart failure (HF) decreases Ca²⁺ release. It is unknown if or how atrial t-tubules can be restored and how this affects systolic Ca²⁺.

Methods: HF was induced in sheep by rapid ventricular pacing and recovered following termination of rapid pacing. Serial block-face scanning electron microscopy and confocal imaging were used to study t-tubule ultrastructure. Function was assessed using patch clamp, Ca²⁺, and confocal imaging. Candidate proteins involved in atrial t-tubule recovery were identified by western blot and expressed in rat neonatal ventricular myocytes to determine if they altered t-tubule structure.

Results: Atrial t-tubules were lost in HF but reappeared following recovery from HF. Recovered t-tubules were disordered, adopting distinct morphologies with increased t-tubule length and branching. T-tubule disorder was associated with mitochondrial disorder. Recovered t-tubules were functional, triggering Ca²⁺ release in the cell interior. Systolic Ca²⁺, I_Ca-L, sarcoplasmic reticulum Ca²⁺ content, and sarcoendoplasmic reticulum Ca²⁺ ATPase function were restored following recovery from HF. Confocal microscopy showed fragmentation of ryanodine receptor staining and movement away from the z-line in HF, which was reversed following recovery from HF. Acute detubulation, to remove recovered t-tubules, confirmed their key role in restoration of the systolic Ca²⁺ transient, the rate of Ca²⁺ removal, and the peak L-type Ca²⁺ current. The abundance of telethonin and myotubularin decreased during HF and increased during recovery. Transfection with these proteins altered the density and structure of tubules in neonatal myocytes. Myotubularin had a greater effect, increasing tubule length and branching, replicating that seen in the recovery atria.

Conclusions: We show that recovery from HF restores atrial t-tubules, and this promotes recovery of I_Ca-L, sarcoplasmic reticulum Ca²⁺ content, and systolic Ca²⁺. We demonstrate an important role for myotubularin in t-tubule restoration. Our findings reveal a new and viable therapeutic strategy.

Keywords: calcium; heart diseases; heart failure; myocytes, cardiac; sarcoplasmic reticulum; volume electron microscopy.

Figure 1.

Sheep atrial t-tubules are restored following recovery (Rec) from heart failure (HF). A, Di-4-ANEPPS staining of sheep atrial myocytes showing t-tubule loss in HF and restoration in recovery. B, Half distance to the nearest t-tubule and surface membrane. C, Categorization of t-tubule disorder. D, Typical control (black) and recovery (green) t-tubule orientation; transverse tubules (90) and longitudinal (0) and transverse:longitudinal tubule ratio. Mean data were calculated from skeletonized images for (E) t-tubule length and (F) branching. G, Relationship between branching and t-tubule angle (R²=0.614). Symbols denote cells n; control: n=24 (6 animals) for B to D, and G, n=30 (7 animals) for E and F; HF: n=36 (3 animals); recovery: n=54 (6 animals) for E to G, n=61 (6 animals) for B to D; compared using linear mixed modeling for B, and D to F or simple linear regression for G. Data are presented as mean±SEM. Scale bars: 10 µm. Ctrl indicates control.

Figure 2.

Recovered atrial t-tubules are disordered. A, Di-4-ANEPPS staining of control and recovered atrial myocytes. B, Application of fluo-5N reveals t-tubules open to the cell exterior, occurring in pairs (blue box). C, Tubule pairs were confirmed through intensity plots corresponding to blue lines/boxes in B. Red asterisks denote peaks corresponding to t-tubules. D, Spacing of t-tubules and individual t-tubules forming pairs (red arrows). Symbols denote cells: n=12 (1 animal) for control, n=13 (2 animals) for recovery (3 tubules averaged per cell); compared using Wilcoxon signed-rank test with multiple comparisons. Data are presented as median±IQR. Scale bars: 10 µm. Ctrl indicates control; IQR, interquartile range; Rec, recovery; and tt, t-tubule.

Figure 3.

The diversity of atrial t-tubule remodeling is associated with mitochondrial positioning. A, sbfSEM 3-dimensional reconstructions of control (Ctrl) and recovery (Rec) atrial myocytes (scale bars: 5 µm). B, Common t-tubule morphologies and longitudinal cell views for Ctrl (white) and Rec (gold) with red arrows indicating oak tree and lattice morphologies (scale bars: 2 µm). C, Percentage occurrence of the most common t-tubule morphologies. Mean data from serial block-face scanning electron microscopy (sbfSEM) for (D) t-tubule volume, (E) number of t-tubules per volume, (F) average t-tubule volume, and (G) average branching per t-tubule. Symbols denote cells; n=9 for Ctrl and n=7 for Rec (3 animals for each group) for D to G; compared using Mann-Whitney U test with multiple comparisons. Data are presented as median±IQR. Example 2-dimensional sections and models of (H) control atrial t-tubule (pink) extending along the z-line between a single band of mitochondria (gray arrows/mitochondria) but not penetrating the thicker mitochondrial bed (blue arrows/mitochondria) and (I) a recovered atrial t-tubule (lime), which penetrates thick mitochondrial beds forming irregular branching (scale bars: 2 µm); n=9 cells from 3 animals for Ctrl and Rec.

Figure 4.

Recovered t-tubules are associated with restored systolic Ca²⁺. A, Representative Ca²⁺ transients and (B) systolic Ca²⁺ transient amplitude. C, Example normalized systolic Ca²⁺ transients. D, Typical experimental time course for Ca²⁺ including application of 10 mmol/L caffeine. E, Rate of decay of the systolic Ca²⁺ transient (_SYS). F, Sarcolemmal (k_CAFF) and (G) SERCA (k_SR)-dependent rates of Ca²⁺ removal. Symbols denote cells n; control (Ctrl): n=20 (9 animals) for B and E, n=16 (8 animals) for F and G; heart failure (HF): n=21 (9 animals) for B and E, n=15 (8 animals) for F and G; recovery (Rec): n=19 (7 animals) for B and E, n=16 (6 animals) for F and G; compared using linear mixed modeling. Data are presented as mean±SEM.

Figure 5.

Recovered t-tubules trigger Ca²⁺ release. A, Representative time series from a recovery atrial cell showing early calcium release (fluorescence). B, Atrial myocytes showing; triggered calcium release, membrane staining, merge of calcium release, and membrane staining and Ca²⁺ rise time. C, Summary data for 50% Ca²⁺ rise time. D, Correlation between rise time and distance to t-tubules for all data points (left) and for t-tubule half distance (right, R²=0.7599). E, Dys-synchrony of Ca²⁺ release. Symbols denote cells n; control (Ctrl): n=6 (4 animals) for C and D, n=8 (2 animals) for E; heart failure (HF): n=4 (2 animals) for C to E (control and HF to confirm our previous work); recovery (Rec): n=7 (3 animals) for C and D, n=27 (5 animals) for E; compared using Kruskal-Wallis test with multiple comparisons. Data are presented as median±interquartile range (IQR). Scale bars: 10 µm.

Figure 6.

Peak I_Ca-L and sarcoplasmic reticulum (SR) Ca content are restored following recovery of the atrial t-tubule network. A, Example voltage step (upper) used to elicit I_Ca-L (lower) in control (Ctrl; black), heart failure (HF; red), and recovery (Rec; green) atrial myocytes, and data (right) for peak I_Ca-L. B, Normalized I_Ca-L (left) and data (right) for the I_Ca-L inactivation rate constant. C, I_Ca-L integral. D, Quantitative assessment of free and total Ca in atrial myocytes. Caffeine-evoked free Ca²⁺ transients (top) are associated with inward I_NCX (middle), which was integrated in a cumulative manner and corrected for cell volume and Ca²⁺ removal via pathways other than Na⁺-Ca²⁺ exchanger (NCX) to give a measure of total Ca (lower). E, SR Ca content as calculated from D. F, Typical cellular buffer curves showing the relationship between free and total Ca. G, Cellular Ca²⁺ buffering power. Symbols denote cells n; control: n=22 (12 animals) for A and C; n=16 (11 animals) for B; n=25 (16 animals) for E and G; HF: n=35 (14 animals) for A to C, n=21 (12 animals) for E and G; recovery: n=21 (6 animals) for A and C, n=15 (6 animals) for B; n=17 (6 animals) for E and G; compared using linear mixed modeling or simple linear regression. Data are presented as mean±SEM.

Figure 7.

Detubulation (Detub) of recovered atrial myocytes converts the recovery Ca²⁺ transient back to a heart failure (HF)-like phenotype. A, Di-4-ANEPPS staining of recovery sheep atrial myocytes showing t-tubule loss following detubulation. Scale bars: 10 µm. B, Half distance to nearest t-tubule and surface membrane. C, Cellular capacitance of recovery atrial myocytes. D, Representative and mean data for peak I_Ca-L in recovery (Rec; gray) and detubulated (Detub) Rec (gray) atrial myocytes. E, Representative systolic Ca²⁺ transients and mean systolic Ca²⁺ transient amplitude. F, Example normalized systolic Ca²⁺ transients, and mean rate of decay of the systolic Ca²⁺ transient (k_SYS). Symbols denote cells n; Rec n=61 (6 animals) for B, n=26 (7 animals) for C, n=20 (6 animals) for D to F; Detub n=19 (3 animals) for B, n=10 (4 animals) for C to F; compared using linear mixed modeling. Data are presented as mean±SEM.

Figure 8.

MTM1 (myotubularin 1) and Tcap (telethonin) influence t-tubule density and structure. A, Western blots and mean data for BIN1 (bridging integrator 1/amphiphysin II), Tcap, MTM1, JPH2 (junctophilin 2) in control (Ctrl), heart failure (HF), and recovery (Rec) sheep atrial tissue. Symbols denote animals N for A, Control: N=5 for BIN1 and JPH2, N=6 for Tcap and MTM1; HF: N=7 for BIN1 and JPH2, N=8 for Tcap and MTM1; recovery: N=7 for BIN1, JPH2, Tcap, and MTM1, compared using ANOVA with multiple comparisons. B, NRVMs transfected with BIN1, Tcap, or MTM1 or triple transfection of BIN, MTM1, and Tcap. C, Fraction of colocalization with BIN1. Mean data summarizing: (D) fractional area of cells occupied by tubules; (E) average skeleton length; and (F) average branching of each tubule structure. Symbols denote cells n for C to F; Ctrl vector n=45 cells (5 L); BIN1: n=79 cells (7 L); BIN1+Tcap: n=98 cells (8 L); BIN1+MTM1: n=84 cells (6 litters); and triple transfection: n=77 cells (5 L); compared using linear mixed modeling. Data are presented as mean±SEM. Scale bars: 10 µm.