Stress-Induced Proteasome Sub-Cellular Translocation in Cardiomyocytes Causes Altered Intracellular Calcium Handling and Arrhythmias

Abstract

The ubiquitin-proteasome system (UPS) is an essential mechanism responsible for the selective degradation of substrate proteins via their conjugation with ubiquitin. Since cardiomyocytes have very limited self-renewal capacity, as they are prone to protein damage due to constant mechanical and metabolic stress, the UPS has a key role in cardiac physiology and pathophysiology. While altered proteasomal activity contributes to a variety of cardiac pathologies, such as heart failure and ischemia/reperfusion injury (IRI), the environmental cues affecting its activity are still unknown, and they are the focus of this work. Following a recent study by Ciechanover's group showing that amino acid (AA) starvation in cultured cancer cell lines modulates proteasome intracellular localization and activity, we tested two hypotheses in human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs, CMs): (i) AA starvation causes proteasome translocation in CMs, similarly to the observation in cultured cancer cell lines; (ii) manipulation of subcellular proteasomal compartmentalization is associated with electrophysiological abnormalities in the form of arrhythmias, mediated via altered intracellular Ca²⁺ handling. The major findings are: (i) starving CMs to AAs results in proteasome translocation from the nucleus to the cytoplasm, while supplementation with the aromatic amino acids tyrosine (Y), tryptophan (W) and phenylalanine (F) (YWF) inhibits the proteasome recruitment; (ii) AA-deficient treatments cause arrhythmias; (iii) the arrhythmias observed upon nuclear proteasome sequestration(-AA+YWF) are blocked by KB-R7943, an inhibitor of the reverse mode of the sodium-calcium exchanger NCX; (iv) the retrograde perfusion of isolated rat hearts with AA starvation media is associated with arrhythmias. Collectively, our novel findings describe a newly identified mechanism linking the UPS to arrhythmia generation in CMs and whole hearts.

Keywords: amino acids starvation; arrhythmias; induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs); intracellular Ca2+ handling; proteasome sub-cellular compartmentalization; ubiquitin–proteasome-system (UPS).

Figure 1

The effect of 2-day amino acid-deficient media on proteasome intracellular distribution in iPSC-CMs (CMs, clone FSE). CMs were cultured in complete medium, control; medium lacking all amino acids, -AA; and -AA medium supplemented with the 3 aromatic amino acids tyrosine (Y), tryptophan (W), and phenylalanine (F), -AA+YWF. (A) Representative immunofluorescence images from the three experimental groups; green: staining of the proteasome α6 subunit; red: staining of the cardiac marker α-actinin; blue: DAPI staining of the nucleus. In control the proteasome is mostly concentrated in the nucleus, in -AA the proteasome translocates from the nucleus to the cytoplasm, and in -AA+YWF the proteasome translocation is prevented, and its nuclear signal intensified. Clone: 24.5; scale bar = 50 μm. (B) A summary showing the ratio between the proteasome α6 subunit density in the nucleus and cytoplasm, N/C. In each cell, the density was defined as signal intensity divided by the respective areas of the nucleus and cytoplasm. The confocal fluorescent images were analyzed using ImageJ quantification tool (see details in Section 4: Materials and Methods). Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test followed the Kruskal–Wallis test; **** p < 0.0001. Control n = 154 CMs; -AA n = 145 CMs; -AA+YWF n = 113 CMs. (C) The effect of 2-day amino acid-deficient media on proteasome activity in iPSC-CMs (CMs). Live imaging of proteasome activity was performed using the Me4BodipyFL-Ahx3Leu3VS probe in CMs cultured as in panel A. Green: Me4BodipyFL-Ahx3Leu3VS, a probe for proteasome activity; blue: Hoechst staining the nucleus. The intracellular distribution of proteasome activity is compatible with the α6 subunit distribution. Clone: FSE; scale bar = 20 μm. In all treatments the white arrows point to the nucleus.

Figure 2

Amino acid-deficient media cause arrhythmias in iPSC-CMs (CMs). (A) Representative action potentials recorded from spontaneously firing CMs cultured for 2 days in control (complete medium), -AA and -AA+YWF. The treated CMs contain both regular (left-hand traces) and arrhythmogenic action potentials. Arrhythmias are presented as skipped beats (red symbols), oscillatory pre-potentials (black symbols) and delayed afterdepolarizations (DADs, blue symbols). (B) The percentage of arrhythmogenic CMs in control, -AA and -AA+YWF. An arrhythmia is defined as a minimum of three arrhythmogenic events. (C) The occurrence of arrhythmias, defined as the mean number of arrhythmogenic events/minute. In the -AA bar, the mean ± SD are 0.0514 ± 0.146. Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test followed by the Mann–Whitney test; ** p < 0.01. Control, n = 11 CMs; -AA, n = 9 CMs; -AA+YWF, n = 37 CMs.

Figure 3

Action potential characteristics recorded from iPSC-CMs (CMs) under different experimental conditions. Control (complete medium): n = 22 CMs; -AA: 2-day treatment, n = 9 CMs; -AA+YWF: 2-day treatment, n = 59 CMs; -AA+Leptomycin B (LMB, 5 ng/mL): 2-day treatment, n = 13 CMs. Beat rate: beats per minute; APA: action potential amplitude; Peak: action potential peak; dV/dt_max: maximal upstroke velocity of phase zero depolarization; MDP: maximal diastolic potential; APD₉₀: action potential duration at 90% repolarization. Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test preformed on all characteristics. Except for dv/dt_max, all characteristics were normally distributed. For the normally distributed characteristics, we used one-way analysis of variance (ANOVA) followed by Tukey post hoc test. For the non-normally distributed characteristics, we used the Kruskal–Wallis test. There were no statistically significant differences among the four groups.

Figure 4

Analysis of Beat Rate Variability (BRV) of spontaneous action potentials recorded from control, and iPSC-CMs (CMs) treated for 2 days with -AA or -AA+YWF. (A) Representative action potentials recorded from experiments used for the BRV analysis. (B) Inter-beat intervals (IBI) versus time plots of the experiments shown in (A). The black arrow in the insert illustrates an expanded IBI scale for control and -AA CMs, showing minimal IBIs dispersion. In contrast to these two groups, -AA+YWF CMs exhibit bimodal firing pattern, as clearly illustrated by the insert, red arrow. (C) Superimposed IBI histograms of the three groups. Note the double-peak histogram of the -AA+YWF group. (D) Superimposed Poincaré plots of the three groups; whereas the plots of control and -AA are condensed (‘cigar-like’, see insert), the -AA+YWF plot is highly dispersed, composed of more than one cloud. (E) BRV measures in control, -AA and -AA+YWF. Mean IBI: mean inter-beat-interval; coefficient of variation; SD1 and SD2: Standard deviation 1 and Standard deviation 2, respectively. Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test followed by the Kruskal–Wallis test; * p < 0.05, ** p < 0.01, *** p < 0.001, ns not significant. Control n = 10 CMs; -AA n = 9 CMs; -AA+YWF n = 17 CMs.

Figure 5

Analysis of Beat Rate Variability (BRV) of extracellular electrograms recorded from control (complete medium) and iPSC-CMs monolayers treated for 2 days with -AA or -AA+YWF. (A) Representative extracellular electrograms recorded from spontaneously firing iPSC-CMs clusters used for the BRV analysis. (B) Inter-beat intervals (IBI) versus time plots of the experiments shown in (A). (C–E) IBI histograms of control, -AA and -AA+YWF, respectively. Note the double-peak histogram of the -AA+YWF group. (F) Superimposed Poincaré plots of the three groups; whereas the plots of control and -AA are condensed (‘cigar-like’), the -AA+YWF plot is highly dispersed, and composed of four clouds as expected from bimodal firing pattern. (G) BRV measures in control, -AA, and -AA+YWF. Mean IBI: mean inter-beat-interval; coefficient of variation; SD1 and SD2: Standard deviation 1 and Standard deviation 2, respectively. Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test followed by the Kruskal–Wallis test. * p < 0.05, ** p < 0.01, *** p < 0.001. Control, n = 8 monolayers; -AA, n = 6 monolayers; -AA+YWF, n = 6 monolayers.

Figure 6

Two-day treatment of iPSC-CMs (CMs) with -AA+Leptomycin B (LMB, 5 ng/mL, an inhibitor of exportin1) recapitulates the effects of -AA+YWF on proteasome intracellular distribution. CMs were cultured for 2 days in control (complete medium), -AA or -AA medium supplemented with LMB (-AA+LMB). (A) Representative immunofluorescence images from the three experimental groups; green: staining of the proteasome α6 subunit; red: staining of the cardiac marker α-actinin; blue: DAPI staining of the nucleus. In control, the proteasome is mostly concentrated in the nucleus, whereas in -AA+YWF and -AA+LMB the proteasome nuclear signal is markedly augmented. Clone: FSE; scale bar = 50 μm. The white arrows point to the nucleus. (B) A summary showing the ratio between the proteasome α6 subunit density in the nucleus and cytoplasm, N/C. In each cell, the density was defined as signal intensity divided by the respective areas of the nucleus and cytoplasm. The confocal fluorescent images were analyzed using ImageJ quantification tool (see details in Section 4: Materials and Methods). Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test followed by the Kruskal–Wallis test; **** p < 0.0001. Control, n = 103 CMs; -AA+YWF, n = 113 CMs; -AA+LMB, n = 101 CMs.

Figure 7

Leptomycin B (LMB, 5 ng/mL) prevents the translocation by -AA of the proteasome from the nucleus to the cytoplasm (-AA+LMB) and causes arrhythmias. (A) Representative action potentials recorded from spontaneously firing iPSC-CMs cultured in control (complete medium, 3 experiments) and in -AA+LMB (6 experiments); arrhythmias are shown in all 6 experiments. (B) Analysis of the arrhythmias caused by 2-day treatment with -AA+LMB medium versus control. The percentage of arrhythmogenic cardiomyocytes in control and -AA+LMB. (C) The occurrence of arrhythmias: the mean number of arrhythmic events/minute in each group. Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test followed by the Mann–Whitney test between the control and -AA+LMB; p = 0.0568. Control, n = 8 CMs; -AA+LMB, n = 13 CMs.

Figure 8

Ca²⁺ transients in control (complete medium) and in iPSC-CMs (CMs) treated for 2 days with -AA or -AA +YWF. (A) Representative Ca²⁺ transients from the three experimental groups. (B) A scheme illustrating the Ca²⁺ transient characteristics analyzed. (C) Ca²⁺ transient amplitude (R-Amplitude). (D) Maximal rate of [Ca²⁺]_i rise (+d[Ca²⁺]_i /dt). (E) Maximal rate [Ca²⁺]_i decay (−d[Ca²⁺]_i /dt. Statistical analysis was performed by means of the Kolmogorov–Smirnov normality test, followed by one-way analysis of variance (ANOVA) and Tukey post hoc test; * p < 0.05, ** p < 0.01. Control, n = 8 clusters; -AA, n = 7 clusters; -AA+YWF, n = 9 clusters.

Figure 9

The response of the Ca²⁺ transient to caffeine in control (complete medium) and in iPSC-CMs treated for 2 days with -AA or -AA+YWF. (A) Representative Ca²⁺ transients illustrating the response to caffeine in the three experimental groups. The red arrows indicate the addition of caffeine. (B) Recovery time calculated as the time from the peak of caffeine-induced Ca²⁺ rise to the first measurable Ca²⁺ transient. (C,D) The percent change in caffeine-induced Ca²⁺ transient amplitude, and fold change in the area of caffeine-induced Ca²⁺ transient compared to the pre-caffeine Ca²⁺ transient, respectively. The statistical analysis was performed by means of the Kolmogorov–Smirnov normality test, followed by one-way analysis of variance (ANOVA) and Tukey post hoc test. There were no statistically significant differences among the three groups. Control, n = 8 clusters; -AA, n = 7 clusters; -AA+YWF, n = 9 clusters.

Figure 10

Rapid pacing causes arrhythmias in -AA+YWF (2-day treatment) but not in control or -AA (2-day treatment) iPSC-CMs (CMs). The pacing protocol consists of 20-pulse trains at 0.5, 1.0, 1.5, and 2.0 Hz, with a 20 s pause interval between each pacing session. (A) Representative action potentials recordings of the pacing protocols, illustrating that in contrast to control (upper row) and -AA (2^nd row from top), rapid pacing induced arrhythmias in -AA+YWF (3^nd and 4^rd rows from top). The red arrow points to a delayed afterdepolarization. (B–D) Percentage of CMs presenting arrhythmias in response to pacing at 1.0 Hz, 1.5 Hz, and 2.0 Hz. Compared to control and -AA, 37–50% of -AA+YWF CMs generated arrhythmias in response to rapid pacing. Control, n = 6 CMs; -AA, n = 5 CMs; -AA+YWF, n = 8 CMs.

Figure 11

KB-R7943 (a blocker of NCX reverse mode) blocks the arrhythmias in iPSC-CMs (CMs) treated for 2 days with -AA+YWF. (A) Four representative experiments show that KB-R7943 (after 4–5 min superfusion) blocked the arrhythmias either at 3 μM or 10 μM. Collectively, KB-R7943 blocked the -AA+YWF-induced arrhythmias in 11 out of 12 experiments. (B) CMs treated for two days with -AA+YWF were superfused with Tyrode`s solution, followed by Tyrode’s solution containing 15 µL of dimethyl sulfoxide (DMSO, after 4–5 min superfusion)—the KB-R7943 solvent.

Figure 12

KB-R7943 (a blocker of the NCX reverse mode) does not affect action potential characteristics of 2-day treated -AA+YWF iPSC-CMs (CMs). Action potential characteristics were measured pre-KB-R7943 and post-KB-R7943 (after 4–5 min of superfusion), when the arrhythmias were blocked. Beat rate: beats per minute; APA: action potential amplitude; Peak: action potential peak; dV/dt_max: maximal upstroke velocity of phase zero depolarization; MDP: maximal diastolic potential; APD₉₀: action potential duration at 90% repolarization. The statistical analysis was performed on all characteristics by means of the Kolmogorov–Smirnov normality tests. Except for dv/dt_max, all characteristics were normally distributed. For the normally distributed characteristics, we used t-test. For the non-normally distributed characteristics, we used the Mann–Whitney test. There were no statistically significant differences between the two groups. n = 10 CMs in each group.

Figure 13

AA-deficient solutions cause arrhythmias in Langendorff retrogradely perfused isolated rat hearts for 1.5 h. Hearts were perfused with Tyrode’s solution with either all amino acids (control), no amino acids (-AA), or only the 3 aromatic amino acids tyrosine (Y), tryptophan (W), and phenylalanine (F) (-AA+YWF). (A) Representative ECG recordings illustrating regular firing in a control heart, and ventricular arrhythmias (Vent Arr, VA) in -AA and -AA+YWF perfused hearts. The horizontal blue lines in the main figure and inserts show the abrupt transition (marked by the blow arrows) from a regular ECG to VA bursts. Black scale bar = 0.5 s. Rhythm irregularities were associated mostly with ectopic ventricular rhythms, but occasionally from supraventricular arrhythmias and irregularities in AV conduction. (B,C) Analysis of the malignant VA (VT and VF) in isolated rat hearts perfused (1.5 h) with control solution, -AA or -AA+YWF. (B) The VA occurrence in the perfused hearts. Whereas the control hearts were non-arrhythmogenic, the arrhythmias caused by -AA+YWF were more pronounced than by -AA. (C) The fractional (percentage) duration of the arrhythmias during the 1.5 h experiment. The statistical analysis was performed by means of the Kolmogorov–Smirnov normality test, followed by the Kruskal–Wallis test; ** p < 0.01. Control, n = 7 rats; -AA, n = 33 rats; -AA+YWF, n = 25 rats.