Cardiac FGF23 Increases Intracellular Calcium in Atrial Myocytes and the Susceptibility to Atrial Fibrillation Decreased in FGF23f / fMyHCCre /+ Mice

FIGURE 1. Characterisation of the cardiac myocyte‐specific FGF23 knockout mice. (A) Schematic illustration of Cre‐LoxP strategy to generate mice with cardiac myocyte specific FGF23 knockout showing murine FGF23 allele with exons 1–3 and location of LoxP sites (triangles) surrounding exon 2. Top: Floxed FGF23 allele, bottom: FGF23 allele after Cre‐mediated recombination. (B) Representative immunoblot and respective quantification show significantly reduced FGF23 protein levels in total heart lysates of FGF23‐CKO compared to control mice. GAPDH serves as loading control. (C) Representative parasternal long‐axis M‐mode echocardiography images of FGF23‐CKO and Cre mice. (D) Quantification of heart function using echocardiography, respectively, in FGF23‐CKO and Cre mice. (E) FGF23 mRNA and protein expression were increased markedly in overexpressed neonatal mouse atrial myocytes models.

FIGURE 2. Representative simultaneous recordings of surface ECG in FGF23‐CKO mice and control mice following programmed stimulation. (A) Representative recordings of surface ECG in FGF23‐CKO mice and control mice induced atrial fibrillation following burst stimulation. Black arrows showed sinus rhythm. (B) Bar graph summarising the incidence of inducible AF in FGF23‐CKO mice after intraperitoneal injection of isoproterenol. Data were analysed using Fisher's exact test (n = 23 mice for FGF23‐CKO and n = 15 mice for Cre, *p < 0.05 vs. Cre + Iso). (D) No difference was found in duration of AF between two groups (n = 23 mice for FGF23‐CKO and n = 15 mice for Cre).

FIGURE 3. Characteristics of cellular electrophysiology and Ca2+ transients in FGF23 overexpression of NMAMs. (A) Raw diagram of Action potentials recorded in two groups. FGF23 overexpression of NMAMs exhibited a higher incidence of triggering activity (Red arrow) than the control. (n = 20/6 ~ 10 cells/neonatal mice, *p < 0.05 vs. Vector，by Student's t‐test). (B) Raw diagram of L‐type calcium current in two groups (left). Overexpression of FGF23 increased the peak density of ICa,L in neonatal mouse atrial myocytes (right). (n = 10/6 ~ 10 cells/neonatal mice, **p < 0.01 vs. Vector，by Student's t‐test) (C) SSA and SSI for ICa,L in neonatal mouse atrial myocytes of overexpressing FGF23. The SSI curve of. FGF23‐overexpression group shifts significantly to the right. (n = 15/6 ~ 10 cells/neonatal mice, **p < 0.01 vs. Vector，by Student's t‐test) (D) The recovery kinetics process after calci‐electric loss accelerated when FGF23 was overexpressed. (n = 10/6 ~ 10 cells/neonatal mice, p > 0.05) (E) The cardiomyocytes of FGF23‐Overexpression and Vector groups were stimulated at 1 Hz and recorded for 10 s after the stimulation. (F) The calcium transient amplitude in the FGF23‐Overexpression group was…

FIGURE 4. Cellular electrophysiological characteristics of adult atrial myocytes in Cre and FGF23‐CKO mice. (A) FGF23‐CKO had a lower occurrence of delayed afterdepolarizations (blue arrow) or Trigger activities (Red arrow) than control after Iso inducing. (n = 25/3 cells/mice, ***p < 0.001 vs. Cre，by Student's t‐test). (B) ICa,L current traces in 10 mV increments from a holding potential of −40 mV to +50 mV in the Cre and FGF23‐CKO atrial myocytes(left). I‐V curve for ICa,L. At 0 mV, the peak density of ICa,L of FGF23‐CKO decreased from −9.42 ± 0.78pA/pF to −5.75 ± 0.39 pA/pF (right). (n = 10/3 cells/mice, ***p < 0.001 vs. Cre，by Student's t‐test) (C) SSA and SSI for ICa,L in the FGF23‐CKO and Cre group. The slope factor Kact of the FGF23‐CKO group were significantly increased. (n = 10/3 cells/mice, ***p < 0.001 vs. Cre，by Student's t‐test) (D) The recovery curve after inactivation between the two groups, the t1 of the FGF23‐CKO were significantly higher than that of the Cre group. (n = 10/3 cells/mice, ***p < 0.001 vs. Cre，by Student's t‐test)

FIGURE 5. Effects of cardiac FGF23 on intracellular calcium in adult atrial cells. (A) The cardiomyocytes of the adult two groups were stimulated at 1 Hz and recorded for 10 s after the stimulation. (B) The calcium transient amplitude in the FGF23‐CKO group was significantly decreased. (n = 22/3 cells/mice, ***p < 0.001 vs. Cre n = 15/3 cells/mice，by Student's t‐test) (C‐E) There was no significant difference in t to peak 50% (bl%) and calcium release decay time (peak time) between the two groups. (F) Representative line scan images of Ca2+ sparks recorded in a cardiomyocyte adult intact mouse atrial myocyte and 3D imaging of Ca2+ sparks. (G) Bar graphs comparing the Ca2+ spark frequency (CaSpF), FWHM and FDHM between the Cre and FGF23‐CKO groups. The FGF23‐CKO group showed significant decreased values of Ca2+ spark frequency and FWHM. FWHM: Full width half maximum, FDHM: Full duration half maximum. n = 54/3 cells/mice for FGF23‐CKOand n = 42/3 cells/mice for Cre, Value = Mean ± SEM, t‐test were performed, ** p < 0.01 compared with the Cre group. (H) Raw diagram of SR calcium release induced by Caffeine of two groups. (I) Bar graphs comparing the Caffeine ratio, Tau‐caffeine and Tau‐1…

FIGURE 6. Calcium channel and transporter protein expressing in adult and neonatal mice atrial myocytes. (A) Representative western blots and quantification of Ryanodine Receptor 2 (RyR2), SRECA2a, Cav1.2 and NCX1.1 expression in the atrial tissues of mice in the FGF23‐CKO group and Control group with GAPDH as a loading control (n = 6 mice per group). Values are presented as mean ± SE. **p < 0.01, ***p < 0.01 vs. Cre group. (B) Representative western blots and quantification of Ryanodine Receptor 2 (RyR2), Cav1.2 expression in neonatal mice atrial myocytes in the FGF23‐OE group and Vector group with GAPDH as a loading control (n = 3 per group). Values are presented as mean ± SE. *p < 0.05 vs. Vector group.