Dual effect of cardiac FKBP12.6 overexpression on excitation-contraction coupling and the incidence of ventricular arrhythmia depending on its expression level

Abstract

FKBP12.6, a binding protein to the immunosuppressant FK506, which also binds the ryanodine receptor (RyR2) in the heart, has been proposed to regulate RyR2 function and to have antiarrhythmic properties. However, the level of FKBP12.6 expression in normal hearts remains elusive and some controversies still persist regarding its effects, both in basal conditions and during β-adrenergic stimulation. We quantified FKBP12.6 in the left ventricles (LV) of WT (wild-type) mice and in two novel transgenic models expressing distinct levels of FKBP12.6, using a custom-made specific anti-FKBP12.6 antibody and a recombinant protein. FKBP12.6 level in WT LV was very low (0.16 ± 0.02 nmol/g of LV), indicating that <15% RyR2 monomers are bound to the protein. Mice with 14.1 ± 0.2 nmol of FKBP12.6 per g of LV (TG1) had mild cardiac hypertrophy and normal function and were protected against epinephrine/caffeine-evoked arrhythmias. The ventricular myocytes showed higher [Ca²⁺]_i transient amplitudes than WT myocytes and normal SR-Ca²⁺ load, while fewer myocytes showed Ca²⁺ sparks. TG1 cardiomyocytes responded to 50 nM Isoproterenol increasing these [Ca²⁺]_i parameters and producing RyR2-Ser2808 phosphorylation. Mice with more than twice the TG1 FKBP12.6 value (TG2) showed marked cardiac hypertrophy with calcineurin activation and more arrhythmias than WT mice during β-adrenergic stimulation, challenging the protective potential of high FKBP12.6. RyR2^R420Q CPVT mice overexpressing FKBP12.6 showed fewer proarrhythmic events and decreased incidence and duration of stress-induced bidirectional ventricular tachycardia. Our study, therefore, quantifies for the first time endogenous FKBP12.6 in the mouse heart, questioning its physiological relevance, at least at rest due its low level. By contrast, our work demonstrates that with caution FKBP12.6 remains an interesting target for the development of new antiarrhythmic therapies.

Keywords: Calcium; Catecholaminergic polymorphic ventricular tachycardia (CPVT); Excitation-contraction coupling; FK506 binding proteins; Heart; Ventricular arrhythmias.

Fig. 1.

Quantification of endogenous and overexpressed FKBP12.6 in mouse hearts. A, Immunoblots of FKBP12.6 from left ventricular (LV) homogenates of WT mice. A GST-tagged recombinant protein (RP-FKBP12.6-GST) was used to quantify precisely FKBP12.6 expression. Fifty μg of total proteins were loaded for each WT sample. B, Linear regression of the standard curve of the RP-FKBP12.6-GST absorbance measured on the immunoblot showed in panel A. C, Immunoblots of FKBP12.6 in WT and TG mice with 10 μg of loaded proteins from each LV homogenates and four concentrations of RP-FKBP12.6-GST. D, Quantification of FKBP12.6 expression in LV homogenates from WT and both TG mouse lines (N = 4 mice for each). Statistical analysis was conducted with One-way ANOVA (P_ANOVA = 3.772e-6) followed by post-hoc Sidak test (P_S). E, Immunoblots of FKBP12.6 in LV homogenates after loading 1.5 μg and 75 μg of proteins from TG and WT LV homogenates, respectively.

Fig. 2.

Prevention of ventricular arrhythmias induced by a pro-arrhythmogenic cocktail in TG1 mice, but not in TG2 mice. A, ECG recordings of conscious control (WT, top), TG1 (middle), and TG2 (bottom) mice in the daytime recorded before (basal condition; left) and after epinephrine/caffeine (Epi/Caff) injection (i.p., 2 and 120 mg/kg, respectively; right). The scale bar represents 200 ms for each example. B, Heart rate before and after Epi/Caff injection, in WT (blue circles; N = 10), TG1 (orange down-triangles; N = 5), and TG2 mice (red up-triangles; N = 4). For Epi/Caff, the heart rate was determined during the maximal effect of the drug injection. Horizontal lines represent the means for each condition. Two-way ANOVA repeated measures (P_genotype < 0.0001; P_treatment < 0.0001; P_interaction = 0.0796) with the factor main effect shown. C, Example of sustained VT observed in a TG2 mouse after stress (Epi/Caff injection). D, Percentage of mice with sustained VT episodes in basal conditions and after Epi/Caff injection, shown in colour, and, in white, of those without VT, in WT (blue; N = 10), TG1 (orange; N = 5) and TG2 (red; N = 4) mice.

Fig. 3.

Increased Ca²⁺ transient amplitude and accelerated Ca²⁺ transient decay in TG1 but not in TG2 ventricular myocytes. A, Line-scan confocal images of [Ca²⁺]_i transients in cardiomyocytes loaded with Fluo-3 AM and field-stimulated at 2 Hz. WT (top), TG1 (middle), and TG2 (bottom) myocytes, before (left) and during 50 nM isoprenaline (Iso) perfusion (right). Above each image, the fluorescence trace. B, Amplitude of the [Ca²⁺]_i transients before and during Iso application, expressed as F/F₀, where F is the peak of the fluorescence trace and F₀ the basal fluorescence. WT are shown as blue circles (N = 11 mice; n = 42 cells), TG1 as orange down-triangles (N = 4; n = 22), and TG2 as red up-triangle (N = 5; n = 27). ANOVA of aligned rank transformed data was used (P_genotype = 0.0031; P_treatment < 2.22e-16; P_interaction = 0.7856) with the factor main effect shown. C, Cardiomyocyte shortening obtained during field stimulation (2 Hz) in WT (N = 11; n = 34), TG1 (N = 4; n = 14), and TG2 (N = 5; n = 19). ANOVA of aligned rank transformed data was used (P_genotype = 0.0623; P_treatment < 2e-16; P_interaction = 0.1709) with the factor main effect shown. D, Time constant of the [Ca²⁺]_i transient decay in WT (N = 11; n = 42), TG1 (N = 4; n = 22), and TG2 (N = 5; n = 27). ANOVA of aligned rank transformed data was used (P_genotype = 2.750e-5; P_treatment < 2.22e-16; P_interaction = 0.0868) with the factor main effect shown. E, Immunoblots of SERCA2a protein expression at the top. At the bottom, quantification of the SERCA2a protein expressed as median [5–95%] in whisker boxes normalized by β-actin and expressed as fold-change of that determined in WT mice from the TG1 and TG2 lines (N = 6–11). Non-parametric Kruskal-Wallis test was used (P_KW = 0.4068) followed by Dunn’s test.

Fig. 4.

Unaltered SR Ca²⁺ load in TG1 and TG2 ventricular myocytes. A, Line-scan confocal images of caffeine-evoked [Ca²⁺]_i transients before (top) and during 50 nM Iso perfusion (bottom) in WT (top of the panel), TG1 (middle of the panel), and TG2 (bottom of the panel) myocytes. Above each image, the fluorescence trace. B, Amplitude of caffeine-evoked [Ca²⁺]_i transients (peak expressed as F/F₀), before and during 50 nM Iso perfusion. N = 4–13 mice; n = 18–46 cells. ANOVA of aligned rank transformed data was used (P_genotype = 0.6163; P_treatment = 0.0023; P_interaction = 0.6126) with the factor main effect shown. C, Time constant of the caffeine-evoked [Ca²⁺]_i transient decay in WT, TG1, and TG2 cells (N = 4–13 mice; n = 16–39 cells). ANOVA of aligned rank transformed data was used (P_genotype = 0.7708; P_treatment = 0.0383; P_interaction = 0.6672) with the factor main effect shown. D, NCX protein expression level, with immunoblots at the top and quantification at the bottom, expressed as median [5–95%] in whisker boxes. Protein levels were normalized by β-actin and expressed as the fold-change of that determined in WT (N = 4–11). Non-parametric Kruskal-Wallis test was used (P_KW = 0.1337) followed by Dunn’s test.

Fig. 5.

Fewer myocytes with Ca²⁺ sparks in TG1 than in WT and TG2 mice. A, Line-scan confocal images of Ca²⁺ sparks before (left) and during 50 nM Iso perfusion (right) in quiescent WT (top), TG1 (middle), and TG2 (bottom) cardiomyocytes. B, Ca²⁺ spark frequency in basal conditions in WT (blue circles), TG1 (orange down-triangles), and TG2 (red up-triangles) myocytes (N = 7–15 mice; n = 72–138 cells per group). ANOVA of aligned rank transformed data was used (P_ANOVA = 0.1073). C, Percentages of cells with Ca²⁺ sparks before and during their stimulation with 50 nM Iso, in WT, TG1, and TG2 mice (N = 4–10 mice; n = 16–21 cells per group). Generalized linear mixed model for repeated measures was used. D, Ca²⁺ wave frequency in basal conditions in WT (blue circles), TG1 (orange down-triangles), and TG2 (red up-triangles) myocytes (N = 7–15 mice; n = 72–138 cells per group). ANOVA of aligned rank transformed data was used (P_ANOVA = 0.0258).

Fig. 6.

FKBP12.6 overexpression did not modify RyR2 phosphorylation before and following β-adrenergic stimulation. A, RyR2 protein expression level in TG1 and TG2 mouse lines, with immunoblot (top) and protein quantification (bottom), expressed as median [5–95%] in whisker boxes. Protein levels were normalized to β-actin and expressed as fold-change of that determined in WT mice (N = 5–6). Non-parametric Kruskal-Wallis test was used (P_KW = 0.0222) followed by Dunn’s test. B, Immunoblot (top) and quantifications (bottom) of RyR2 phosphorylation at the PKA site (Ser2808), and, C, at the CaMKII site (Ser2814) normalized to the total RyR2 in cardiomyocytes, in WT, TG1, and TG2 (N = 5–9), in basal conditions and under Iso stimulation (100 nM). ANOVA of aligned rank transformed data was used ((B) Ser2808: P_genotype = 0.0002; P_treatment = 0.0101; P_interaction = 0.1928; (C) Ser2814: P_genotype = 0.4541; P_treatment = 0.6341; P_interaction = 0.3539).

Fig. 7.

FKBP12.6 overexpression mitigated proarrhythmic SR Ca²⁺ waves and stress-induced bidirectional ventricular tachycardia in RyR2^R420Q CPVT mice. A, ECG of CPVT (top) and DT (bottom) mice before (left) and under Epi/Caff stimulation (i.p. 2 and 120 mg/kg, respectively; right). B, Percentage of mice with bidirectional ventricular tachycardia (BDVT) episodes under Epi/Caff stimulation, shown with colour in CPVT (pink; N = 4) and DT (purple; N = 5) mice. C, Duration of BDVT during Epi/Caff stimulation in CPVT (pink squares; N = 4) and DT (purple diamonds; N = 5) mice. Non-parametric Mann-Whitney test was used. D, Line-scan confocal images of caffeine-evoked [Ca²⁺]_i transients in CPVT RyR2^R420Q (top) and DT (bottom) myocytes, with above each image the fluorescence trace. The thick line above each image indicates 10 mM caffeine perfusion, after the 2 Hz cell stimulation. E, Peak of the [Ca²⁺]_i transients expressed as F/F₀, and F, time constant of the [Ca²⁺]_i transient decay in CPVT (pink squares; N = 4 mice; n = 51 cells) and DT myocytes (purple diamonds; N = 4 mice; n = 48 cells). ANOVA of aligned rank transformed data was used. G, Peak of the caffeine-evoked [Ca²⁺]_i transients (expressed as F/F₀) in CPVT (pink squares; N = 4 mice; n = 22 cells) and DT (purple diamonds; N = 4 mice; n = 18 cells) myocytes. ANOVA of aligned rank transformed data was used. H, Line-scan confocal images of Ca²⁺ waves in quiescent CPVT-RyR2^R420Q (top), and DT (bottom) myocytes. I, Ca²⁺ wave frequencies in basal conditions in CPVT (pink squares; N = 4 mice; n = 41 cells) and DT (purple diamonds; N = 4 mice; n = 43 cells) myocytes. ANOVA of aligned rank transformed data was used. J, Arrhythmic diastolic events (Ca²⁺ waves + triggered activities) in CPVT in basal conditions (pink squares; N = 4 mice, n = 19 cells) and after incubation with 1 μM JTV-519 (red wine squares; N = 3 mice; n = 20 cells). ANOVA of aligned rank transformed data was used. All data are shown as median [5–95%] in whisker boxes.