Sequential Changes in Calcium Transients During M Phase Regulate Cardiomyocyte Proliferation

Abstract

Heart muscle growth and regeneration require the proliferation of cardiomyocytes. Rapid pulsatile increases in cytosolic Ca²⁺ concentration, called calcium transients (CaTs), trigger cardiomyocyte contractions, but how cardiomyocytes adapt Ca²⁺ signaling during proliferation is largely unknown. Here, we show that cardiomyocyte proliferation requires changes in Ca²⁺ signaling. Cardiomyocytes undergo a sequence of CaT changes during M-phase: CaT amplitudes begin to decline in prometaphase, reach a minimum in metaphase, rise during anaphase, and return to the original state in daughter cardiomyocytes. Spindle poles show decreased Ca²⁺ levels during prometaphase and metaphase. Localized reduction of Ca²⁺ levels at spindle poles is mediated by dynein 1–dependent SERCA2a accumulation. Active CDK1 induces both the decrease in CaT amplitudes and the accumulation of SERCA2a at the spindle poles, whereas CDK1 inhibition reverses these effects. Forcing an increase in cytosolic Ca²⁺ levels by blocking SERCA2a during prometaphase and metaphase disrupts mitosis and produces binucleated cardiomyocytes, underscoring the essential role of Ca²⁺ signaling changes for cardiomyocyte proliferation.

eTOC summary

Cardiomyocyte contractions require calcium transients. Liu et al. show that mitotic cardiomyocytes suppress calcium transients and remodel the sarco-/endoplasmic reticulum. These changes are required for cardiomyocyte division and identify a link between calcium signaling and cytokinesis in cardiomyocytes.

Introduction

Organ growth and repair require cellular proliferation. While some tissues use the proliferation of undifferentiated stem and progenitor cells for growth and repair, in other tissues, differentiated cells proliferate (Wells and Watt, 2018). Notable examples of the latter are contractile heart muscle cells (cardiomyocytes) (Bersell et al., 2009), insulin-producing β-cells in the pancreas (Dor et al., 2004), Drosophila larval tracheal cells during metamorphosis (Guha et al., 2008; Weaver and Krasnow, 2008), and Arabidopsis root cells after amputation of the stem cell niche (Sena et al., 2009). Thus, the proliferation of differentiated cells seems to be a recurrent principle in biology. However, little is known about the intracellular processes that allow differentiated cells to accomplish division.

Cardiomyocytes are differentiated muscle cells, which proliferate during development before birth and in young mammals. The question arises as to what degree cardiomyocytes alter their sub-cellular structures and functions during the cell cycle (Zhu et al., 2021). Cardiomyocytes are electrically excitable muscle cells with a specialized structure called sarcomere, whose contractions are activated by pulsatile rapid oscillations in intracellular Ca²⁺ levels (calcium transients, CaTs). Although the disassembly of sarcomeres during M-phase has been reported (Ahuja et al., 2004; Bersell et al., 2009; Fan et al., 2015), the mechanisms underlying changes in contractility and the associated Ca²⁺ signaling remain poorly understood.

Ca²⁺ is a ubiquitous and pleiotropic second messenger. The concentration of extracellular Ca²⁺ exceeds cytosolic levels [Ca²⁺]_i (~100 nM) by up to 20,000 fold (Clapham, 2007). Rapid increases in [Ca²⁺]_i levels trigger sarcomere contractions. Cardiomyocytes achieve this by releasing Ca²⁺ very rapidly through ryanodine receptor 2 (RyR2) from the sarcoplasmic reticulum (SR) (Kuo and Ehrlich, 2015; Kushnir and Marks, 2010), a form of endoplasmic reticulum (ER) that is interleaved with the sarcomeres and specialized for the uptake, storage, and rapid release of Ca²⁺. Ca²⁺ release via RyR2, the principal RyR isoform in cardiac muscle, is activated by an initial small influx of Ca²⁺ through plasma membrane Ca²⁺ channels (L-type Ca²⁺ channels), whose opening is triggered by membrane depolarization. Ca²⁺ reuptake into the SR is mediated by the sarco/endoplasmic reticulum calcium ATPase 2a (SERCA2a). This sequential action of membrane depolarization, Ca²⁺ release, sarcomere contraction, and Ca²⁺ reuptake is called excitation-contraction coupling (Bers, 2002). Both physiologic and pathologic mechanisms modulate the efficiency of excitation-contraction coupling in cardiomyocytes to regulate the contraction and relaxation of the heart. While Ca²⁺ signaling has been implicated in cardiomyocyte proliferation (Devilee et al., 2025; Lam et al., 2022; Nguyen et al., 2020), it is unclear whether and how cardiomyocytes modify Ca²⁺ signaling during the cell cycle. The aims of this study were to address this knowledge gap by delineating cell cycle-related changes in the SR, Ca²⁺ signaling, and contractility, characterizing the underlying mechanisms, and examining their significance for cardiomyocyte proliferation.

Results

Human cardiomyocytes exhibit decreased calcium transients in the cell cycle

We developed an approach to examine the physiology of live cardiomyocytes, followed by post hoc determination of their cell cycle status by immunofluorescence microscopy using cell cycle markers. We first examined human induced pluripotent stem cell (iPSC)-derived cardiomyocytes (Fig. 1A), a commonly used cellular system in cardiac research. To detect CaTs, iPSC-derived cardiomyocytes were loaded with the Ca²⁺ indicator, Rhod2. We used an antibody against Ki67 to detect cardiomyocytes in cell cycle (G1-, S-, G2, and M-phase) by post hoc immunofluorescence microscopy (Fig. 1A). We noticed that cycling (Ki67⁺) cardiomyocytes exhibited decreased or absent CaTs (Fig. 1B–C, Video 1).

Figure 1. Cardiomyocytes exhibit reduced calcium transients (CaTs) during the cell cycle.

(A-F) Calcium signaling was assessed using the calcium indicator Rhod2, followed by post hoc immunofluorescence microscopy to identify cycling cardiomyocytes. Red and white arrows indicate cycling and non-cycling cardiomyocytes, respectively. (A-C) Human iPSC-derived cardiomyocytes and (D-F) human fetal cardiomyocytes were stimulated with electrical pulses at 1 Hz. (A, D) Post hoc immunofluorescence microscopy identified cycling human iPSC-derived cardiomyocytes and human fetal cardiomyocytes imaged during live Ca²⁺ recordings. (B, E) CaTs were detected in non-cycling (Ki67⁻), but not in cycling (Ki67⁺) cardiomyocytes (Video 1). (C, F) Quantification showed that the percentage of cycling cardiomyocytes exhibiting CaTs (CaTs⁺, white column) was significantly lower than that of non-cycling cardiomyocytes (C, n = 34 human iPSC-derived cardiomyocytes, Ki67⁺: 17 cells, Ki67⁻:17 cells; F, n = 146 human fetal cardiomyocytes, Ki67⁺: 41 cells, Ki67⁻: 105 cells). (G-L) Cycling mouse cardiomyocytes display reduced CaTs in the intact heart. A transgenic mouse model expressing a genetic Ca²⁺ reporter, GCaMP8 (αMHC-GCaMP8, green cytoplasm, excited by 488 nm laser beam) and a cell cycle reporter, αMHC-Cre;Rosa26-mCherry-Geminin (mCherry-Geminin, red nucleus, excited by 561 nm laser beam) was used to identify proliferating cardiomyocytes. (G) Experimental design of confocal imaging for recording CaTs in neonatal mouse hearts (P1-P3), with intact atria not shown here. (H-L) CaTs were recorded in live mouse hearts (Video 2). Representative areas from two hearts are shown in (H-I) and (J-K), respectively. (H, J) Images show GCaMP8 and Geminin signals. Red and white arrows indicate cycling (Gem⁺, red nuclei) and non-cycling (Gem⁻, the nucleus is not red) cardiomyocytes, respectively. Cell boundaries were defined based on GCaMP8 fluorescence (dashed lines). For visualization, the lookup table (LUT) was adjusted to enhance GCaMP8 signal visibility in cycling cardiomyocytes (Gem⁺, red nuclei) with weaker fluorescence. (I, K) Corresponding Ca²⁺ traces from cardiomyocytes outlined with dashed lines and indicated by arrows in panels (H, J). Electrical field stimulation was applied at 1 Hz. (L) Quantification showed significantly reduced peak CaTs in cycling (Gem⁺) cardiomyocytes (each symbol represents one cardiomyocyte from 3 hearts. Gem⁻: n = 39 cells, Gem⁺: n = 21 cells). F₀ and F_max denote the baseline and peak fluorescence of Ca²⁺ indicators, respectively. Scale bars: 60 μm. Frame rate: (B, E) 10 frames/second, (I, K) 15 frames/second. Statistical analyses: Chi-Square Test (C, F), Student’s t-test (unpaired, Two-tailed) (L). Mean ± SEM is indicated by horizontal lines (L).

Video 1. (Corresponding to Fig. 1A–B) Cycling human iPSC-derived cardiomyocytes exhibit decreased calcium transients (CaTs).

Download video file ^{(347.1KB, mp4)}

Live cell microscopy of human iPSC-derived cardiomyocytes (15 days post differentiation) was used to visualize Ca²⁺ signals detected by Rhod2. Top left panel: Post hoc immunofluorescence microscopy with antibodies against Ki67 and α-actinin identified cell cycle activity in cardiomyocytes. The red arrow indicates a cycling cardiomyocyte, and the white arrow indicates a non-cycling cardiomyocyte (control). Scale bar 60 μm. Top right panel: Movie of Ca²⁺ signals indicated by Rhod2 fluorescence. Electrical field stimulation was applied at 1 Hz. Bottom panel: CaTs (F/F₀) traces of the cardiomyocytes shown in the top panels. Red curve: cycling cardiomyocyte (red arrows in top panels); black curve: non-cycling cardiomyocyte (white arrows in top panels). The CaTs trace represents a phenotype that was rare in all the studied Ki67⁺ cardiomyocytes. Frame rate: 10 frames/second.

The maturity of human iPSC-derived cardiomyocytes is thought to correspond to human fetal cardiomyocytes (Burridge et al., 2012). We therefore examined CaTs in cycling primary human cardiomyocytes isolated from human fetal hearts (gestational age 17 weeks) using the same approach (Fig. 1D, E). The phenotype of no or decreased CaTs in human fetal cardiomyocytes that were in the cell cycle was present in 75%, a 3-fold increase over non-cycling (Ki67⁻) cardiomyocytes (Fig. 1F). In conclusion, human cardiomyocytes can exhibit reduced CaTs during the cell cycle.

The phenotype of decreased CaTs is present in cycling cardiomyocytes in intact hearts

To assess the phenotype of decreased CaTs in unperturbed myocardium, we developed a live-cell microscopy approach to measure CaTs in single cardiomyocytes in intact neonatal mouse hearts (Fig. 1G). To detect CaTs, we used GCaMP8, a fusion protein of calmodulin and green fluorescent protein, in which Ca²⁺ binding to the calmodulin domain induces a conformational change in the green fluorescent protein, resulting in light emission. We used a mouse model expressing GCaMP8 as a transgene under the control of the cardiomyocyte-specific Myh6 promoter (αMHC-GCaMP8) (Lee et al., 2021), which did not alter cardiomyocyte contractility (Fig. S1A–C). To label cycling cardiomyocytes, we cross-bred this strain with mice expressing a genetically encoded fluorescent cell cycle reporter, consisting of a floxed mCherry-Geminin allele in the Rosa26 locus, and the cardiomyocyte-specific αMHC-Cre transgene (Han et al., 2020). αMHC-Cre activates expression of fluorescent mCherry-Geminin in cardiomyocytes, which is deposited in nuclei during S-phase, and degraded at M-phase exit (Han et al., 2020).

For experiments, we resected hearts from neonatal mice (P1-P3) expressing αMHC-GCaMP8; αMHC-Cre; mCherry-Geminin. The hearts were electrically stimulated (1 Hz), and intracellular Ca²⁺ indicated by GCaMP8 was visualized using high-speed confocal microscopy. mCherry-Geminin-positive (Gem⁺, i.e., cycling) cardiomyocytes exhibited significantly reduced CaTs compared with non-cycling (Gem⁻) cardiomyocytes (Fig. 1H–L, Video 2). In conclusion, both human and mouse cardiomyocytes can exhibit decreased CaTs during the cell cycle.

Video 2. (Corresponding to Fig. 1H–K) Cycling cardiomyocytes in intact newborn mouse hearts exhibit decreased CaTs.

Download video file ^{(10.3MB, mp4)}

CaTs in cardiomyocytes of intact newborn mouse hearts (P1-P3) were detected with the live Ca²⁺ reporter GCaMP8 and visualized using intact heart microscopy. Electrical field stimulation was applied at 1 Hz. Results from two acquisitions are shown in three panels in A and B, respectively. In A and B, the top left panels: Cardiomyocytes express αMHC-GCaMP8 (Green cytoplasm) as a Ca²⁺ indicator, and mCherry-Geminin (Red nuclei) specifically expressed in cardiomyocytes (αMHC-Cre;Rosa-mCherry-Geminin) as a cell cycle indicator. Cardiomyocyte boundaries are outlined with colored lines. Red nuclei identify cardiomyocytes in S-, G2-, and M-phase. Red arrows indicate cycling cardiomyocytes, and white arrows indicate non-cycling cardiomyocytes whose CaTs are shown in the lower panels. Top right panels: Background noise was removed using the Nikon Elements Denoise.ai module. Lower panels: CaTs (F/F₀) traces of the cardiomyocytes shown in the top panels. Red curves: Cycling cardiomyocytes (red arrows in top panels); black curves: non-cycling cardiomyocytes (white arrows in top panels). Frame rate: 15 frames/second.

Cardiomyocytes decrease CaTs during prometaphase and re-establish them in telophase

To determine when CaTs begin to decrease during the cell cycle, we first examined CaTs in interphase (G1, S, and G2) cardiomyocytes. To obtain large quantities of cells, we used neonatal rat ventricular myocytes (NRVMs). Interphase cardiomyocytes were identified by staining for the cell-cycle marker Ki67 and confirming the absence of condensed chromosomes (Hoechst-labeled) using post hoc immunofluorescence microscopy. CaTs indicated by Rhod2 in cultured NRVMs were recorded, followed by post hoc immunofluorescence analysis. The results showed that CaTs were not reduced in interphase cardiomyocytes (G1/S/G2, Ki67⁺, without condensed chromosomes) compared with non-cycling cardiomyocytes (Ki67⁻) (Fig. 2A–B). Therefore, we hypothesized that CaTs begin to decrease in cardiomyocytes during the M phase. For simplicity and to emphasize M-phase results, we use “interphase” to represent non-cycling, G1, S, and G2 phases in the following sections unless otherwise specified.

Figure 2. Cardiomyocytes decrease calcium transients (CaTs) during prometaphase and restore them during cytokinesis.

Calcium signaling was examined in neonatal rat ventricular myocytes (NRVMs). Interphase cardiomyocytes (G1/S/G2) were identified by Ki67 staining in the absence of condensed chromosomes, using post hoc immunofluorescence microscopy. Sub-phases of M-phase were identified as described in Fig. S2A and the related Source Data. Rhod2 was used as a Ca²⁺ indicator. Red and white arrows indicate M-phase and interphase NRVMs, respectively, with corresponding CaTs traces presented in red and black curves. (A) Interphase NRVMs (Ki67⁺) exhibit CaTs similar to those observed in non-cycling (Ki67⁻) NRVMs. (B) Analysis of peak CaTs showed that interphase does not affect CaTs in cardiomyocytes (Ki67⁻: n = 88 cells from 3 cultures; Ki67⁺: n = 21 cells from 3 cultures). (C) NRVMs maintained CaTs during prophase (red arrow, dashed outline, Video 3). The corresponding CaTs measurements, indicated by changes in Rhod2 fluorescence, are shown below the cell images. (D) CaTs in NRVM (encircled by a white dashed line) progressing from prometaphase through division into two daughters were recorded using Rhod2 fluorescence. Visualization of CaTs across different cell cycle phases is presented in Video 4. (C-D) F₀ and F_max denote the baseline and peak fluorescence of Ca²⁺ indicators, respectively. (E) Analysis of peak CaTs showed a decrease during prometaphase and full recovery after cell division (Interphase: n = 12 cells; Prophase: n = 5 cells; Prometaphase to Daughter: n = 3 cells). Symbols connected by solid lines indicate the same NRVMs tracked through M-phase to daughter cell formation. (F-G) CaTs were analyzed during distinct sub-phases of M-phase, identified by H3P and Aurora B kinase staining. NRVMs shown in (F) have reduced CaTs in prometaphase and restored CaTs during cytokinesis. (G) Quantification confirmed decreased peak CaTs in prometaphase/metaphase and recovery during cytokinesis (each symbol represents one cardiomyocyte, Interphase: n = 73 cells, Prometa/Metaphase: n = 11 cells, Cytokinesis: n = 8 cells). (H-I) NRVMs in M-phase did not contract during 1 Hz electrical stimulation. (H) Post hoc immunofluorescence labeled NRVMs in M-phase (H3P⁺). Small arrows generated by Particle Image Velocimetry (PIV) indicate the direction and velocity of contraction (arrowhead direction, size, and color scale). (I) Each symbol represents one cardiomyocyte (Interphase: n = 18 cells; M-phase: n = 5 cells). (J-N) NRVMs in M-phase exhibit decreased cytosolic calcium concentration [Ca²⁺]_i. (J) The M-phase cardiomyocyte (red arrow) displayed lower CaTs, indicated by (K) lower peak and baseline [Ca²⁺]_i. Quantification showed that M-phase NRVMs have (L) lower peak and (M) baseline [Ca²⁺]_i, and (N) fold change in [Ca²⁺]_i during systole. (L-N) Interphase: n = 52 cells, M-phase: n = 13 cells. Electrical stimulation (1 Hz) was applied (C, D, F, H). Frame rates: (A, D) 30 frames/second, (C, F) 16 frames/second, (K) 19 frames/second. Scale bars: 40 μm (A, C, D, F, J), 60 μm (H). Statistical analyses: One-way ANOVA followed by Bonferroni’s correction for multiple comparisons (E, G), Student’s t-test (unpaired, Two-tailed) (B, I, L, M, N). Mean ± SEM is indicated by horizontal lines (B, E, G, I, L, M, N). arb. u.: arbitrary unit (E).

To determine when CaTs decrease in relation to M-phase, we used live-cell microscopy to record CaTs while simultaneously tracking cell cycle progression in cycling cardiomyocytes over time. We developed a brightfield microscopy approach to identify the sub-phases of M-phase in live cardiomyocytes (Fig. S2A and the related Source Data). CaTs in prophase NRVMs were similar to those in interphase NRVMs (Fig. 2C, Video 3). CaTs began to decrease in prometaphase, and reached to a minimum level in metaphase, then began to return to higher amplitudes in anaphase and telophase, and were restored to interphase levels in the daughter cardiomyocytes (Fig. 2D, Video 4). We confirmed these results by using a different Ca²⁺-indicator, Fluo4, with higher affinity for Ca²⁺ (Fig. S1D–F, Video 5). Taken together, these results show that CaTs decrease in prometaphase and return in telophase (Fig. 2E). To define the decrease and return of CaTs using the post hoc approach, we used an antibody against phosphorylated histone H3 (H3P) along with Hoechst DNA staining to visualize condensed chromosomes for identification of prometaphase and metaphase, and we labeled the cleavage furrow with an antibody against Aurora B kinase to identify cytokinesis (Fig. 2F). We observed a significant decrease in CaTs during prometaphase and metaphase, compared to interphase CaTs (Fig. 2G). During telophase/cytokinesis, CaTs peaks increased significantly compared to metaphase levels (Fig. 2G). In summary, two lines of evidence (life-cell and post hoc microscopy) demonstrate that CaTs begin to decrease during prometaphase and return during cytokinesis. We next examined the functional consequence of CaTs reduction, i.e., sarcomere contractions. Particle imaging velocimetry (PIV) analysis revealed a 13-fold reduction in contractility in M-phase NRVMs compared to interphase cardiomyocytes (Fig. 2H, I).

Video 3. (Corresponding to Fig. 2C) NRVMs do not decrease CaTs during prophase.

Download video file ^{(844.7KB, mp4)}

Live-cell microscopy of NRVMs (P1) was performed to visualize Ca²⁺ signal changes detected by Rhod2 in prophase NRVMs, identified by H3P labeling through post hoc immunofluorescence microscopy. Red arrows indicate a prophase NRVM, and white arrows mark an interphase NRVM (control). The boundary of the prophase NRVM is delineated by white dashed lines.

• Top left panel: Brightfield snapshot showing the morphology of live NRVMs (pre-fixation) during CaTs recording.
• Top right panel: Brightfield snapshot showing the morphology and H3P labeling of NRVMs (Post-fixation) using post hoc immunofluorescence microscopy. Scale bar: 40 μm.
• Bottom left panel: Movie displaying Ca²⁺ signals detected by Rhod2 fluorescence during 1 Hz electrical field stimulation.
• Bottom right panel: CaTs (F/F₀) traces of the NRVMs shown in the top panels. Red curve: prophase NRVM (red arrows); black curve: interphase NRVM (white arrows). Frame rate: 16 frames/second.

Video 4. (Corresponding to Fig. 2D) Cycling NRVMs decrease CaTs in prometaphase and restore them during telophase/cytokinesis.

Download video file ^{(1.1MB, mp4)}

Prolonged live cell microscopy of NRVMs (P1) was performed to visualize Ca²⁺ signal changes detected by Rhod2 during cell cycle progression. Top panels: Brightfield video microscopy visualizes morphological changes of an NRVM progressing from prometaphase (0 min) to cell division (50 min), generating two daughter cells. Red arrows indicate the NRVM in M-phase, and white arrows indicate an interphase NRVM (control). Scale bar 40 μm. Middle panels: Movies of Ca²⁺ signal indicated by Rhod2 fluorescence at 1 Hz field stimulation. Bottom panels: CaTs (F/F₀) traces of the NRVMs shown in the top panels. Red curves: M-phase NRVM (red arrows in top panels); black curves: Interphase NRVM (white arrows in top panels). Frame rate: 30 frames/second.

Video 5. (Corresponding to Fig. S1D–F).

Download video file ^{(2.6MB, mp4)}

A reduction in CaTs was observed in M-phase NRVMs using dual-channel imaging. A mixture of the fluorescent Ca²⁺ dyes Fluo4 AM (green) and Rhod2 AM (red) was loaded into NRVMs (P1). Left: M-phase NRVMs were identified based on morphology (Fig. S2A and related Source Data). M-phase NRVM: green/red double arrows; interphase NRVM (control): white arrow. Scale bar: 40 μm. Middle: Movie of Ca²⁺ signals indicated by Fluo4 fluorescence (top), and CaTs (F/F₀) traces of the NRVMs in color curves (bottom). Right: Movie of Ca²⁺ signals indicated by Rhod2 fluorescence (top), and CaTs (F/F₀) traces of the NRVMs in color curves (bottom). Frame rate: 41 frames/second.

We measured changes in the cytoplasmic calcium concentration [Ca²⁺]_i in nanomolar (Del Nido et al., 1998) (nM, Fig. 2J–N). The peak systolic [Ca²⁺]_i in M-phase NRVMs was 125.7 ± 18.2 nM, a 2.6-fold reduction from interphase cardiomyocytes (Fig. 2L). The baseline [Ca²⁺]_i during the end of diastole in M-phase NRVMs was 68.5 ± 4.5 nM, a 1.8-fold reduction from interphase cardiomyocytes (Fig. 2M). Because cardiomyocyte contractions are triggered by rapid rises of the [Ca²⁺]_i, we compared the ratios of systolic and diastolic [Ca²⁺]_i. M-phase cardiomyocytes had a 1.8-fold systolic increase in [Ca²⁺]_i during systole, which was significantly lower than the 2.9-fold increase observed in interphase cardiomyocytes (Fig. 2N). In conclusion, M-phase cardiomyocytes have significantly lower peak and baseline [Ca²⁺]_i as well as smaller -fold changes in intracellular [Ca²⁺]_i during CaTs.

Ca²⁺-storage and release functions of the SR are unchanged in M-phase

A working cardiomyocyte stores stores large quantities of Ca²⁺ in SR for triggered release through RyR2. We examined the possibility that the reduced CaTs in M-phase could result from reduced SR Ca²⁺ storage or impaired Ca²⁺ release. To investigate this, we used caffeine (1 mM) to maximally activate RyR2 and calculated changes in [Ca²⁺]_i from Rhod2 fluorescence (Fig. 3A, Video 6). Caffeine treatment induced similar peak [Ca²⁺]_i in interphase and M-phase cardiomyocytes (Fig. 3B), indicating that SR Ca²⁺ storage is sufficient and the Ca²⁺ release machinery remains functional in M-phase.

Figure 3. M-phase cardiomyocytes exhibit normal calcium storage and release functions.

Ca²⁺ storage and membrane potentials were examined in NRVMs. M-phase cells were identified as in Fig. S2A. Rhod2 was used as the Ca²⁺ indicator. Red and white arrows indicate M-phase and interphase NRVMs, respectively, with corresponding CaTs traces presented in red and black curves. (A-B) NRVMs in M-phase exhibited normal Ca²⁺ storage in the SR. RyR2-mediated SR Ca²⁺ release was stimulated by 1 mM caffeine and 1 Hz electrical stimulation, followed by ionomycin for intracellular calcium ([Ca²⁺]ᵢ) calibration (A, Video 6). (B) Caffeine-induced peak [Ca²⁺]ᵢ was the same in Interphase (Ctrl) and M-phase NRVMs. Interphase: n = 43 cells, M-phase: n = 11 cells. (C-E) M-phase NRVMs exhibited membrane potential changes in response to electrical stimulation. (C) Membrane potential was detected using the fluorescent membrane potential sensor CytoVolt1. White and red arrows indicate interphase and M-phase NRVMs, respectively. (D) Normalized changes in CytoVolt1 fluorescence from the NRVMs indicated by the arrows in (C), showing membrane potential responses during 1 Hz electrical stimulation. (E) Each symbol represents the percentage of NRVMs with action potentials per culture (n = 3 cultures per group). (F-G) Stimulation of cAMP production with Forskolin (10 μM, 15 minutes) increased peak CaTs in M-phase NRVMs, without 1 Hz electrical stimulation (Video 7). (H) Quantification showed increased peak CaTs in both interphase and M-phase NRVMs following Forskolin treatment (each symbol represents one cardiomyocyte. [Interphase] Ctrl: n = 13 cells, Forskolin: n = 22 cells; [M-phase] Ctrl: n = 4 cells, Forskolin: n = 6 cells). Frame rates: (A) 30 frames/second, (D) 18 frames/second, (G) 50 frames/second. Scale bars: 40 μm (C, F). Statistical analyses: Student’s t-test (unpaired, Two-tailed) (B, E), One-way ANOVA followed by Bonferroni’s correction for multiple comparisons (H). Mean ± SEM is indicated by horizontal lines (B, E, H). arb. u.: arbitrary unit (A).

Video 6. (Corresponding to Fig. 3A) M-phase NRVMs have sufficient Ca²⁺ stored in the sarcoplasmic reticulum.

Download video file ^{(1.6MB, mp4)}

Live cell microscopy of NRVMs (P1) was used to visualize the Ca²⁺ signal detected by Rhod2. Rhod2 fluorescence (F) was used to present CaTs and reflect the total Ca²⁺ release induced by caffeine (1 mM) treatment. M-phase NRVM: red arrows; interphase NRVM (control): white arrows. Top left panel: An M-phase NRVM was identified based on morphology (Fig. S2A and related Source Data). Scale bar 40 μm. Top right panel: Movie of Ca²⁺ signal measured by Rhod2 fluorescence. While Rhod2 fluorescence was recorded, caffeine was added to a final concentration of 1 mM in the medium to maximally open RyR2 channels. This caused a burst release of Ca²⁺ ions, which increased Rhod2 fluorescence. Bottom panel: CaTs (F) traces of the NRVMs shown in the top panels. Red curve: M-phase NRVM (red arrows in top panels); black curve: interphase NRVM (white arrows in top panels). The black arrow indicates the time (~20 s) when caffeine was added. Frame rate: 30 frames/second.

M-phase cardiomyocytes have action potentials

Ca²⁺ release from the SR via RyR2 is mediated by Ca²⁺-induced Ca²⁺ release, which is triggered by membrane depolarizations, called action potentials. We hypothesized that the decrease in CaTs could be due to the absence of action potentials. To examine action potentials, cardiomyocytes were labeled with Cytovolt1 (Di-4-ANBDQBS, CytoCybernetics), a fluorescent voltage-sensitive membrane dye, and membrane potential changes were recorded by live-cell microscopy (Fig. 3C–D). A similar proportion of interphase and M-phase cardiomyocytes exhibited action potentials during 1 Hz stimulation (Fig. 3E). In conclusion, M-phase cardiomyocytes have action potentials.

Calcium transients can be restored during M-phase

Brightfield microscopy revealed thin intercellular connections resembling filopodia between M-phase cardiomyocytes and neighboring cells (Figs. 2D, 3F). CaTs in M-phase cardiomyocytes were synchronous with those of neighboring cardiomyocytes and occurred in the absence of electrical field stimulation (Fig. 3G). The combination of physical intercellular connections and CaTs synchrony under unstimulated conditions suggests the presence of functional electrical coupling between M-phase cardiomyocytes and their neighbors. These observations indicate that CaTs in M-phase cardiomyocytes may be modulated or restored through external intervention (Marks, 2013). To test the hypothesis that increasing cAMP levels could restore CaTs in M-phase cardiomyocytes, we used forskolin, which mimics the active conformation of the stimulatory G protein, Gs, and thus stimulates adenylyl cyclase to produce cAMP. As shown in Fig. 3F–H and Video 7, forskolin increased peak CaTs in M-phase cardiomyocytes to interphase levels, indicating functional regulation of Ca²⁺ signaling during M-phase.

Video 7. (Corresponding to Fig. 3F–G) Forskolin treatment increases peak CaTs in M-phase NRVMs.

Download video file ^{(765.8KB, mp4)}

NRVMs (P1) were treated with forskolin (10 μM, treatment group) for 15 minutes. Ca²⁺ signals were detected using Rhod2 without electrical field stimulation. M-phase NRVMs were identified based on morphology (Fig. S2A). M-phase NRVMs: red arrows; interphase NRVMs: white arrows. Left panels: NRVMs treated with DMSO (Control group) for 15 minutes. Left top panel: Brightfield imaging used to identify M-phase NRVMs based on morphology. Left middle panel: Movie of Ca²⁺ signal indicated by Rhod2 fluorescence. Left bottom panel: CaTs (F/F₀) traces of the NRVMs shown in the top panels. Red curve: M-phase NRVM (indicated by red arrows in the top panels); black curve: interphase NRVM (indicated by white arrows in the top panels). Right panels: NRVMs treated with forskolin (10 μM, treatment group) for 15 minutes. Right top panel: Brightfield imaging used to identify M-phase NRVMs based on morphology. Scale bar: 40 μm. Right middle panel: Movie of Ca²⁺ signal indicated by Rhod2 fluorescence. Right bottom panel: CaTs (F/F₀) traces of the NRVMs shown in the top panels. Red curve: M-phase NRVM (indicated by red arrows in the top panels); black curve: interphase NRVM (indicated by white arrows in the top panels). Frame rate: 50 frames/second.

Cardiomyocytes remodel their SR during M-phase

To examine the structural changes that could correspond to the reduction of CaTs, we visualized the structure of the SR in NRVMs, focusing on two key SR-associated Ca²⁺ regulators: RyR2, the muscle-specific Ca²⁺ release channel, and SERCA2a, the Ca²⁺ reuptake pump. We visualized RyR2 and SERCA2a, together with Troponin I, a component of the interleaved sarcomeres, using immunofluorescence microscopy. In interphase cardiomyocytes, RyR2 and Troponin I displayed characteristic striated patterns (Fig. 4A). In contrast, M-phase cardiomyocytes, identified by mitotic spindle chromosome condensation, lost both the striated Troponin I and the interleaved RyR2 patterns (Fig. 4A). Analysis of sub-phases of M-phase revealed that the striated RyR2 pattern disappeared during prometaphase (Fig. 4A, B, and Fig. S2B). No specific pattern of RyR2 localization was observed during metaphase and anaphase. In telophase/cytokinesis, the striated RyR2 pattern began to reappear at the cell periphery (Fig. 4A, insets).

Figure 4. Cardiomyocytes remodel their sarcoplasmic reticulum during M-phase.

Two SR Ca²⁺ handling proteins, RyR2 and SERCA2a, were visualized to examine SR organization in NRVMs (P1). Mitotic spindles were identified using antibodies against tubulin. Microtubule organizing centers and spindle poles are indicated by yellow stars and confirmed with spindle pole markers pericentrin and γ-tubulin. Different sub-phases of M-phase were determined by the organization of condensed chromosomes (blue) and mitotic spindles. Regions enclosed by brackets or squares are magnified in insets to show the subcellular structures. (A) Interphase and prophase cardiomyocytes exhibited striated RyR2 patterns, which were absent in prometaphase, metaphase, and anaphase but reappeared in telophase. (B) Quantification showed that striated RyR2 patterns were detected in most interphase cardiomyocytes, but were rarely observed in prometaphase, metaphase, and anaphase, and reappeared in telophase (n = 3 cultures per group). (C) Striated SERCA2a patterns were detected in interphase and prophase but diminished during prometaphase. SERCA2a accumulated around spindle poles in prometaphase, peaked in metaphase, weakened in anaphase, and became undetectable in telophase as striations reappeared. (D) Quantification showed that striated SERCA2a patterns were detected in nearly all interphase and prophase cardiomyocytes, but were rarely observed in prometaphase, metaphase, or anaphase. Striated SERCA2a patterns reappeared during telophase (n = 3 cultures per group). (E) Cardiomyocytes with SERCA2a accumulation at spindle poles were quantified. All metaphase cardiomyocytes exhibited SERCA2a accumulation at spindle poles, which declined in anaphase and became undetectable in telophase (n = 3 cultures per group). (F) Co-staining of the spindle pole marker pericentrin and microtubules demonstrates colocalization between spindle poles and microtubule organizing centers. (G) Co-staining of pericentrin and SERCA2a demonstrates SERCA2a accumulation at the spindle poles. (H) Co-staining of the spindle pole marker γ-tubulin and microtubules demonstrates colocalization between spindle poles and microtubule organizing centers. (I) Co-staining of γ-tubulin and SERCA2a demonstrates SERCA2a accumulation at the spindle poles. Scale bars: 10 μm. Inset scale bars: 2 μm. Statistical significance of differences between sub-phases of M-phase and interphase was analyzed with One-way ANOVA followed by Bonferroni’s correction for multiple comparisons. Mean ± SEM is indicated horizontal lines.

We next examined SERCA2a. Similar to RyR2, the striated pattern of SERCA2a began to disappear during prometaphase (Fig. 4C, D). However, unlike RyR2, SERCA2a began to accumulate at the spindle poles indicated by the microtubule-organizing centers in prometaphase and became concentrated there during metaphase (Fig. 4C, E). SERCA2a accumulation at the spindle poles started to decrease during anaphase (Fig. 4C, E), and the striated SERCA2a pattern reappeared during telophase/cytokinesis (Fig. 4C, D). To verify the spatial correlation between microtubule-organizing centers, spindle poles, and SERCA2a-accumulating regions, we stained for pericentrin and γ-tubulin, two widely accepted and specific markers of spindle poles (Dictenberg et al., 1998; Zhu et al., 2023; Zimmerman et al., 2004) together with tubulin (microtubules) and SERCA2a, respectively. As shown in Fig. 4F–I, the results confirm the established overlap between spindle poles and microtubule-organizing centers, as well as the overlap between spindle poles and SERCA2a accumulation. Because microtubule staining can indicate both spindle structure for identifying M-phase sub-phases and spindle poles, we primarily used microtubule organizing centers indicated by tubulin staining to define spindle poles in this study. In summary, the striated patterns of SR proteins–RyR2 and SERCA2a–became disorganized during prometaphase and reorganized in telophase/cytokinesis. This disorganization and reorganization process was separated by metaphase, during which SERCA2a accumulation around the spindle poles peaked. Although both RyR2 and SERCA2a exhibited similar patterns of disorganization and reorganization at the onset and completion of M-phase, only SERCA2a showed accumulation at the spindle poles, suggesting a potential function, for example, increased local Ca²⁺ reuptake.

SERCA2a accumulation leads to lower Ca²⁺ levels at the spindle poles during M-phase

The localization of SERCA2a at the spindle poles during metaphase (Fig. 4) directed us to investigate the consequence for local Ca²⁺ levels. Using a live-cell tubulin dye (SiR-tubulin) to visualize the mitotic spindle, we examined the spatial distribution of Ca²⁺ levels in prometaphase and metaphase (Fig. 5A). The results indicated that Ca²⁺ levels were lower at the spindle poles compared to adjacent regions. To highlight this spatial pattern, we generated heatmaps of Ca²⁺ levels to reveal regions of low Ca²⁺ levels at the spindle poles, which we refer to as Ca²⁺ valleys Fig. 5B). We calculated the ratio of the Ca²⁺ signal intensity at the spindle poles to that of the corresponding adjacent regions. Prometaphase and metaphase cardiomyocytes exhibited significantly lower Ca²⁺ signal intensity ratios (Fig. 5C). After live-cell Ca²⁺ imaging, we fixed the cells and stained them for SERCA2a. The results showed co-localization of regions with low Ca²⁺ levels and SERCA2a accumulation at the spindle poles (Fig. 5D). In conclusion, the data in Fig. 5A–D suggest that the Ca²⁺ valleys may be caused by SERCA2a accumulation. To test this hypothesis, we inhibited SERCA2a with thapsigargin, which reduced the area of Ca²⁺ valleys by increasing cytosolic Ca²⁺ levels (Fig. 5E–F), supporting a causal relationship between SERCA2a accumulation and the formation of Ca²⁺ valleys.

Figure 5. M-phase cardiomyocytes show decreased calcium levels at the spindle poles.

(A) NRVMs were labeled with Ca²⁺ indicator Rhod2 (red) and microtubule marker SiR-tubulin (green) and examined by live-cell confocal microscopy. During prometaphase, one spindle pole present in the optical plane (white arrow), showed lower Ca²⁺ levels. During metaphase, two spindle poles appeared in the optical plane and exhibited lower Ca²⁺ levels (white arrows). (B) Heatmaps corresponding to (A) illustrate Rhod2 fluorescence intensity and highlight regions of lower Ca²⁺ signal (Ca²⁺ valley) at the spindle poles. The color scale shows Ca²⁺ signal strength from weak to strong. White arrows indicate the same spindle pole region as in (A). (C) Nearly all prometaphase/metaphase NRVMs exhibit reduced Ca²⁺ signal intensity at the spindle poles. Quantification was performed by calculating the ratio of Ca²⁺ signal intensity at the spindle poles to adjacent cytoplasmic regions. Ratio below 1.0 (red dashed line) signifies a Ca²⁺ valley at the spindle pole. Each symbol represents one cardiomyocyte (Prometaphase: n = 17 cells, Metaphase: n = 4 cells). (D) Ca²⁺ valleys overlap with SERCA2a accumulation at spindle poles in live M-phase NRVM. Spindle poles (indicated by arrows) were identified using SiR-tubulin-labeled spindle microtubules. Ca²⁺ distribution was indicated by Rhod2 fluorescence using confocal microscopy. Post hoc immunofluorescence staining identified SERCA2a localization in the same cell. Heatmap highlights Ca²⁺ valleys, with color indicating Ca²⁺ signal strength from weak to strong. (E) Inhibition of SERCA2a with thapsigargin (6 μM) increased Ca²⁺ signal at the spindle poles. Dimethyl Sulfoxide (DMSO) served as the control (Ctrl). Spindle poles were identified using SiR-tubulin-labeled spindle microtubules (white arrows). Heatmaps show Rhod2 indicated Ca²⁺ signal strength before treatment and at 5 minutes, and 8 minutes post-treatment. White arrows indicate Ca²⁺ valleys, a smaller area of Ca²⁺ valley indicates a stronger Ca²⁺ signal around the spindle poles. (F) The area of Ca²⁺ valley was normalized to its baseline measured before treatment. Each symbol represents one M-phase cardiomyocyte (Ctrl: n = 4 cells, Thapsigargin: n = 6 cells). Scale bars: 10 μm (A, B, D, E). Statistical analyses: (C) One-sample t-test comparing values to the control value set at 1.0, (F) One-way ANOVA followed by Bonferroni’s correction for multiple comparisons. Mean ± SEM is indicated by horizontal lines (C, F). arb. u.: arbitrary unit (F).

Cardiomyocyte division requires active Ca²⁺ reuptake by SERCA2a during prometaphase and metaphase

The reduction in peak CaTs and the presence of Ca²⁺ valleys during M-phase suggest that low intracellular Ca²⁺ levels are essential for proper M-phase progression. To test this hypothesis, we inhibited SERCA2a with thapsigargin to increase cytosolic Ca²⁺ levels and assessed the effects using two complementary microscopy techniques: live-cell imaging and immunofluorescence microscopy. Thapsigargin did not disrupt spindle microtubule structure (Fig. S1G–H). In live-cell imaging experiments, we identified cardiomyocytes in the sub-phases of M-phase by examining cell morphology (Fig. 6A). We then introduced thapsigargin and monitored M-phase progression until exit. Lastly, we examined the daughter cardiomyocytes by post hoc immunofluorescence microscopy. Thapsigargin treatment before metaphase (Prometa/Meta; Fig. 6B) resulted in cell division failure. In contrast, treatment after metaphase (Ana/Telo; Fig. 6B) allowed successful completion of cell division. Analysis by post hoc immunofluorescence microscopy revealed that SERCA2a inhibition during prometaphase/metaphase caused cytokinesis failure, indicated by an increase in the percentage of binucleated cardiomyocytes compared with controls (Fig. 6C). However, karyokinesis appeared unaffected, as division failure did not result in an increased number of mononucleated daughter cells (Fig. 6D). In separate immunofluorescence microscopy experiments, thapsigargin treatment (6 hours) increased the proportion of binucleated cardiomyocytes among all cells in culture (Fig. 6E–F).

Figure 6. Altered calcium signaling during prometaphase and metaphase is required for cytokinesis.

NRVMs (P1) were cultured and treated with thapsigargin (6 μM) to inhibit SERCA2a. The control group (Ctrl) received DMSO. (A-D) Exposure to thapsigargin (6 μM) during prometaphase/metaphase caused division failure and binucleation. Red arrows indicate M-phase NRVMs, and white arrows indicate daughter cells observed 4–6 hours after thapsigargin treatment (0 hour refers to pre-identified M-phase NRVMs before treatment). (B) Quantification revealed that initiating thapsigargin treatment during prometaphase/metaphase caused division failure, whereas initiation during anaphase/telophase did not. Each symbol represents one cell culture. Ctrl: n = 110 M-phase cells from 5 cultures; Thapsigargin: n = 143 M-phase cells from 7 cultures. (C-D) Division failure induced by initiating thapsigargin treatment in prometaphase/metaphase NRVMs resulted in the formation of binucleated cells (C; cytokinesis failure), rather than formation of single daughter cells with one nucleus (D; karyokinesis failure). Each symbol represents one cell culture. Ctrl: n = 110 M-phase cells from 6 cultures; Thapsigargin: n = 138 M-phase cells from 8 cultures. (E-M) NRVMs were fixed 6 hours after treatment and analyzed by immunofluorescence microscopy. (E-K) Thapsigargin treatment increased binucleated cardiomyocytes without altering nuclear ploidy. (E) White arrows indicate binucleated cardiomyocytes. (F) Quantification confirmed an increased percentage of binucleated cardiomyocytes among all the NRVMs following thapsigargin treatment. Each symbol represents one cell culture. Ctrl: n = 3 cell cultures, Thapsigargin: n = 3 cell cultures. (G-J) DNA content was measured using Hoechst fluorescence, (G) with telophase daughter nuclei serving as the diploid reference (2N, white arrows) and prometaphase/metaphase chromosomes serving as the tetraploid reference (4N, red arrows). (H-J) Relative DNA content in nuclei normalized by mean DNA content in telophase nuclei (2N), or half mean DNA content in prometaphase/metaphase chromosomes (4N). Each symbol represents one cell. (H) Ctrl: prometa=29 cells, meta=11 cells, telo=108 cells, inter=2070 cells, in 3 cell cultures; (I) Thapsigargin: prometa=22 cells, meta=12 cells, telo=108 cells, inter=1997 cells, in 3 cell cultures; (J) Ctrl: inter=2070 cells in 3 cell cultures; Thapsigargin: inter=1997 cells in 3 cell cultures. (K) Quantification showed a slight but significant decrease in the percentage of polyploid (≥4N) mononuclear NRVMs among all the NRVMs following thapsigargin treatment. Each symbol represents one cell culture. Ctrl: n = 6 cell cultures, Thapsigargin: n = 6 cell cultures. (L-M) Thapsigargin treatment reduced entry of S phase. (L) White arrows indicate BrdU⁺ cardiomyocytes. (M) Quantification demonstrates a decreased percentage of BrdU⁺ cardiomyocytes among all the NRVMs following thapsigargin treatment. Each symbol represents one cell culture. Ctrl: n = 3 cell cultures, Thapsigargin: n = 3 cell cultures. Scale bars: 20 μm (A, E, G), 40 μm (L). Statistical analyses: One-way ANOVA followed by Bonferroni’s correction for multiple comparisons (B), Student’s t-test (unpaired, Two-tailed) (C, D, F, J, K, M). Mean ± SEM is indicated by horizontal lines (B-D, F, H-K, M). arb. u.: arbitrary unit (H-J).

To further examine the potential effect of SERCA2a inhibition on karyokinesis, we measured nuclear ploidy in cardiomyocytes after blocking SERCA2a for 6 hours. Nuclear ploidy was determined using an immunofluorescence microscopy–based method. Based on established principles of cell proliferation, DNA content in prometaphase and metaphase cardiomyocytes was used as the reference for tetraploid nuclei (4N), whereas DNA content in telophase nuclei served as the reference for diploid nuclei (2N) (Fig. 6G). Cardiomyocyte nuclear ploidy was calculated by normalizing DNA content to that of the reference nuclei. The distribution of nuclear ploidy is shown in Fig. 6H–J. Only mononuclear cells were analyzed in the interphase cardiomyocytes. We observed a slight but significant decrease in DNA content in mononuclear interphase cardiomyocyte nuclei after SERCA2a inhibition (Fig. 6J), accompanied by a reduction in the number of cardiomyocytes with polyploid (≥4N) nuclei in interphase (Fig. 6K). Further study showed a reduction in BrdU⁺ nuclei following SERCA2a inhibition (Fig. 6L–M). Because a 6-hour thapsigargin (together with BrdU) treatment is insufficient to span both the M phase (1–2 hours) and the G1 phase (6–12 hours), BrdU⁺ cardiomyocytes were not exposed to thapsigargin and BrdU during the M phase of the preceding cell cycle. Therefore, the results in Fig. 6L–M indicate that SERCA2a inhibition can also reduce S-phase entry, which explains the result in Fig. 6K. Together, these findings support the conclusion that active Ca²⁺ reuptake by SERCA2a is required in prometaphase/metaphase for successful cytokinesis.

CDK1 activity induces SR remodeling and decrease in CaTs

The observations that reduced CaTs and spindle pole-accumulation of SERCA2a occur during M-phase support the hypothesis that these events are mechanistically regulated by the cyclin-dependent kinase 1 (CDK1)/Cyclin B1 complex, which drives M-phase entry and progression. To test the hypothesis, we increased the activity of the complex by using adenoviral transduction with constitutively active CDK1 (T14A/Y15F, CDK1AF, human) (Bicknell et al., 2004) and Cyclin B1 (gene name: CCNB1, human, co-transduction termed CDK1AF+Cyclin B1 or CDK1AF for short). This intervention induced increased M-phase activity (Fig. S3), which was associated with an increased number of cardiomyocytes showing SERCA2a accumulation at the spindle poles (Fig. 7A, B). Cardiomyocytes with induced M-phase activity exhibited lower CaTs than interphase cardiomyocytes that were transduced with CDK1AF + Cyclin B1 (Fig. 7C–E, Video 8). No significant differences in CaTs were observed between interphase cardiomyocytes that were or were not transduced with CDK1AF + Cyclin B1 (Fig. 7E), indicating that the reduction in CaTs is specific to M-phase. In conclusion, experimentally increased CDK1 activity is sufficient to induce M-phase with SERCA2a accumulation and reduced CaTs.

Figure 7. CDK1 activity is sufficient and required for SERCA2a accumulation and decrease in CaTs in M phase.

(A-E) Induction of M-phase entry with active CDK1 triggered SERCA2a redistribution and reduced CaTs in cardiomyocytes. NRVMs (P1) induced into M-phase by forced expression of active CDK1 (CDK1AF+Cyclin B1, Day 1–2) exhibited SERCA2a accumulation at the spindle poles and reduced CaTs. M-phase NRVMs on Day 1 without CDK1 activation served as controls (Ctrl; normal condition). (A) SERCA2a accumulation was observed exclusively in NRVMs undergoing M-phase — either under normal conditions (Ctrl) or upon CDK1AF+Cyclin B1 induction — and was absent in interphase NRVMs in both groups. (B) Quantification of SERCA2a accumulation in NRVMs (n = 3 cell cultures per group). (C-E) NRVMs induced into M-phase (CDK1AF+Cyclin B1) demonstrated decreased CaTs, replicating the reduction observed in the M-phase cardiomyocytes from the Ctrl group, but interphase NRVMs in the CDK1AF+Cyclin B1 group did not demonstrate decrease in CaTs. (C) Representative M-phase NRVMs expressing transduced CDK1AF+Cyclin B1 exhibited characteristic M-phase morphology. CaTs measured in interphase (white arrows) and M-phase NRVMs (red arrows) are shown in (D, Video 8). (E) Peak CaTs were reduced in M-phase NRVMs, but not in interphase NRVMs, in both groups. Each symbol indicates one cardiomyocyte. Ctrl (Interphase: n = 33 cells, M-phase: n = 9 cells); CDK1AF+Cyclin B1 (Interphase: n = 98 cells, M-phase: n = 15 cells). (F–L) CDK1 inhibition eliminated H3P signal, blocked SERCA2a accumulation at the spindle poles, and significantly increased CaTs in M-phase NRVMs. NRVMs were treated with DMSO (20 μL/mL; Ctrl), Ro-3306 (24 μM), MG132 (40 μM), or a combination of Ro-3306 (24 μM) and MG132 (40 μM) for one hour, followed by (F) immunofluorescence microscopy to identify M-phase (H3P⁺, condensed chromosomes, and cell morphology) cardiomyocytes. (G) Ro-3306 treatment decreased the proportion of H3P⁺ NRVMs, including cells arrested in M-phase by MG132. Each symbol represents one cell culture (n = 3 cell cultures per group). (H-L) M-phase NRVMs were identified by examining cell morphology (Fig. S2A, related Source Data) before treatment with MG132 (40 μM, Ctrl), or with a combination of Ro-3306 (24 μM) and MG132 (40 μM) for one hour. (H-I) CDK1 inhibition with Ro-3306 decreased SERCA2a accumulation at spindle poles in NRVMs (H, red arrows) arrested in M-phase by MG132. (I) Each symbol represents one cell culture. N = 3 cultures in each group. (J-L) CDK1 inhibition with Ro-3306 increased CaTs in NRVMs arrested in M-phase by MG132. After treatment, CaTs measured in pre-identified M-phase NRVMs (red arrows) are shown in (J-K, Video 9). (L) CDK1 inhibition increased peak CaTs in NRVMs arrested in M-phase. Each symbol represents one cell. Ctrl[MG132] (Interphase: n = 25 cells; M-phase: n = 5 cells), Ro-3306[MG132] (Interphase: n = 45 cells; M-phase: n = 9 cells). Frame rate: (D, K) 30 frames/second. Scale bars: (A) 10 μm, (C, F, H, J) 20 μm. Statistical analyses: (B, E, G, L) One-way ANOVA followed by Bonferroni’s correction for multiple comparisons, (I) Student’s t-test (unpaired, Two-tailed). Mean ± SEM is indicated by horizontal lines (B, E, G, I, L).

Video 8. (Corresponding to Fig. 7C–D) NRVMs in M-phase induced by forced expression of active CDK1 exhibit decreased CaTs.

Download video file ^{(1.1MB, mp4)}

Live cell microscopy of NRVMs (P1) was used to visualize Ca²⁺ signals detected by Rhod2. The M-phase promoting factor CDK1+Cyclin B1 complex, was expressed using adenovirus-mediated gene transfer (Adv-Cdk1AF+Adv-Cyclin B1). Post hoc immunofluorescence with antibodies against H3P and troponin I identified M-phase NRVMs. M-phase NRVMs: red arrows; interphase NRVMs: white arrows. All cells were stimulated at 1 Hz electrical field. Left: NRVMs were not transfected with adenovirus (Ctrl). Left top panels: H3P (green) and troponin I (purple) identify M-phase and NRVMs, respectively. Movie of Ca²⁺ signal indicated by Rhod2 fluorescence. Left bottom panel: CaTs (F/F₀) traces of the NRVMs shown in the top panels. Red curve: M-phase NRVM (indicated by red arrows in the top panels); black curve: interphase NRVM (indicated by white arrows in the top panels). Right: NRVMs were transfected to express active CDK1AF+Cyclin B1. Right top panels: H3P (green) and troponin I (purple) identify M-phase and NRVMs, respectively. Movie of Ca²⁺ signal indicated by Rhod2 fluorescence. Right bottom panel: CaTs (F/F₀) traces of the NRVMs shown in the top panels. Red curve: M-phase NRVM (indicated by red arrows in the top panels); black curve: interphase NRVM (indicated by white arrows in the top panels). Frame rate: 30 frames/second. Scale bars: 20 μm.

SERCA2a accumulation at the spindle poles requires CDK1 activity

We determined whether CDK1 activity is required for SERCA2a accumulation at the spindle poles during M-phase. To reduce CDK1 activity, we inhibited CDK1 with the chemical Ro-3306, and co-applied the proteasome inhibitor, MG132 to prevent mitotic exit, for 1 hour. We determined the effectiveness of this approach by quantifying the presence of H3P and assessed the M-phase characteristics, e.g., cell morphology and chromosome condensation (Fig. 7F). Treatment with Ro-3306 alone demonstrated effective CDK1 inhibition, indicated by 9.3-fold reduction of the percentage of H3P⁺ cardiomyocytes (Fig. 7G). Treatment with MG132 alone caused a statistically insignificant increase in the percentage of H3P⁺ cardiomyocytes (Fig. 7G), and the combination of Ro-3306 and MG132 resulted in a significant reduction of the percentage of H3P⁺ cardiomyocytes (Fig. 7F, G), indicating that Ro-3306 efficiently inhibited CDK1.

We quantified SERCA2a accumulation at the spindle poles in M-phase NRVMs following CDK1 inhibition with Ro-3306. Prometaphase/metaphase NRVMs were first identified based on cellular morphology (Fig. 7H, Before treatment) and then arrested in M-phase by treatment with MG132 together with the CDK1 inhibitor Ro-3306. After 1 hour of treatment, cardiomyocytes in M-phase were re-identified by post hoc immunofluorescence microscopy (Fig. 7H, After treatment). Based on the results in Fig. 4C, E, only NRVMs that remained in prometaphase or metaphase were selected for quantification of SERCA2a accumulation. As shown in Fig. 7H–I, CDK1 inhibition caused an approximately 3.9-fold decrease in the number of cardiomyocytes exhibiting SERCA2a accumulation, indicating that CDK1 activity is required for SERCA2a accumulation at the spindle poles during M-phase.

Decrease in CaTs during M-phase requires CDK1 activity

We then determined whether CDK1 activity is required for the decrease of CaTs in M-phase. Similarly, we first identified M-phase cardiomyocytes by brightfield live-cell microscopy based on their characteristic morphology (Fig. 7J, Before treatment). Ro-3306 and MG132 were then applied for 1 hour, followed by CaT recordings (Video 9). Cardiomyocytes in M-phase were then re-identified by post hoc immunofluorescence microscopy (Fig. 7J, After treatment). MG132 alone did not alter reduced peak CaTs in M-phase cardiomyocytes (Fig. 7K, L). Addition of the CDK1-inhibitor Ro-3306 resulted in significantly higher peak CaTs in M-phase cardiomyocytes (Fig. 7L), reaching levels similar to interphase cardiomyocytes under the same conditions (Fig. 7L). In conclusion, CDK1 activity is required to sustain the reduction of CaTs in M-phase cardiomyocytes.

Video 9. (Corresponding to Fig. 7J–K) CDK1 inhibition with Ro-3306 increased CaTs in NRVMs arrested in M-phase with MG132.

Download video file ^{(1.2MB, mp4)}

From left to right: Left panels — NRVMs treated with DMSO+MG132 (40 μM) (Ctrl [MG132]) for 1 hour. Right panels — NRVMs treated with Ro-3306 (24 μM)+MG132 (40 μM) (Ro-3306 [MG132]) for 1 hour. From top to bottom: Top panels: Brightfield imaging of M-phase NRVM before and after treatment. Middle panels: Snapshots of post hoc immunofluorescence and movies of Ca²⁺ signal detected by Rhod2 fluorescence. Bottom panels: Ca²⁺ transient (CaTs) traces (F/F₀) of the NRVMs shown in the top panels. Red curves: M-phase NRVMs (indicated by red arrows in the top panels); black curves: interphase NRVMs (indicated by white arrows in the top panels). Frame rate: 30 frames/second. Scale bar: 20 μm.

SERCA2a accumulation at the spindle poles requires microtubules and dynein 1

The poleward localization of SERCA2a suggested the involvement of spindle microtubules and their associated motor proteins. We hypothesized that spindle microtubules guide SERCA2a accumulation and that the pole-directed molecular motor dynein 1 complex mediated this process (Fig. 8A). To examine the requirement of spindle microtubules, we depolymerized microtubules using nocodazole and examined SERCA2a distribution by immunofluorescence microscopy. Nocodazole treatment disrupted spindle structure and increased the proportion of cardiomyocytes in prophase, prometaphase, and metaphase (Fig. 8B–C). These cells exhibited disorganized SERCA2a localization and lacked its accumulation at specific areas (Fig. 8D–E). These results indicate that spindle microtubules are required for SERCA2a accumulation.

Figure 8. SERCA2a accumulation at spindle poles requires spindle microtubules and dynein 1.

(A) Schematic representation of spindle microtubules (red) and the dynein 1 complex (purple), which transports SERCA2a towards the spindle poles. Black arrows indicate the direction of dynein 1 movement towards the microtubule negative ends (red “-”). SERCA2a is shown in green. (B-C) Depolymerization of spindle microtubules with nocodazole arrested proliferating cardiomyocytes in early M-phase. NRVMs (P1) were treated with nocodazole (20 μM) for 60–90 minutes, with DMSO as the control (Ctrl). (B) Cells were fixed and stained with antibodies against troponin I and tubulin, and chromosomes were labeled with Hoechst. Sub-phases of M-phase were identified based on chromosome condensation and microtubule organization. NRVMs in M-phase were categorized into two groups: early M-phase (Pro/Prometa/Meta) or late M-phase (Ana/Telo/Cytokinesis). No late M-phase NRVMs were detected in the nocodazole group. (C) All M-phase NRVMs in the nocodazole group were in early M-phase, indicating that cells were trapped in M-phase. (Ctrl: n = 47 NRVMs from 2 cultures; nocodazole group: n = 46 NRVMs from 3 cultures). (D-E) NRVMs were analyzed using immunofluorescence microscopy for SERCA2a accumulation at spindle poles. (D) Chromosome condensation identified M-phase cardiomyocytes, and spindle microtubules were used to identify spindle poles. (E) Nocodazole treatment prevented SERCA2a accumulation at spindle poles. Each symbol represents one cell culture (N = 4 cell cultures per group). (F-G) SERCA2a did not accumulate at spindle poles in neonatal mouse ventricular myocytes (NMVMs, P1) with Dync1h1 knockout. Control group (Ctrl): αMHC-Cre⁺;Dync1h1^Flox/Wild, Dync1h1 knockout group (Dync1h1 CKO): αMHC-Cre⁺;Dync1h1^Flox/Flox. M-phase was identified by examining condensed chromosomes and spindle microtubules. All 9 prometaphase/metaphase NMVMs from 2 Ctrl hearts exhibited SERCA2a accumulation at spindle poles, whereas none of the 8 NMVMs from 4 Dync1h1 CKO hearts did. (H-I) Chemical inhibition of dynein 1 with ciliobrevin D (120 μM, 60 minutes) decreased percentage of NRVMs with SERCA2a accumulation at spindle poles. Cells in the control group (Ctrl) were treated with DMSO. Each symbol represents one cell culture (N = 3 cell cultures per group). (J-K) Preventing SERCA2a accumulation at spindle poles through induced microtubule depolymerization did not alter peak CaTs in M-phase cardiomyocytes. NRVMs were treated with nocodazole (20 μM, 60–90 minutes) to induce microtubule depolymerization. (J) NRVMs were stimulated with 1 Hz electrical pulses during CaTs recording. F₀ and F_max indicate baseline and peak Rhod2 (Ca²⁺) fluorescence, respectively. (K) Quantification of peak CaTs in cardiomyocytes. Each symbol represents one NRVM (Ctrl: Interphase = 48 cells, M-phase = 12 cells; Nocodazole: Interphase = 49 cells, M-phase = 11 cells). (L-M) Inhibition of dynein 1 with ciliobrevin D (120 μM, 60 minutes) to prevent SERCA2a accumulation at spindle poles did not alter peak CaTs in M-phase NRVMs. (L) F₀ and F_max indicate baseline and peak Rhod2 (Ca²⁺) fluorescence, respectively. (M) Quantification of peak CaTs in M-phase NRVMs. Each symbol represents one cardiomyocyte (Ctrl: Interphase = 69 cells, M-phase = 15 cells; Ciliobrevin D: Interphase = 105 cells, M-phase = 23 cells). Scale bars: (B) 20 μm, and (D, F, H) 10 μm. (J, L) 40 μm. (J) 30 frames/second, (L) 33 frames/second. Statistical analyses: (C) two-way ANOVA, (E, I) Student’s t-test (unpaired, Two-tailed), (K, M) One-way ANOVA followed by Bonferroni’s correction for multiple comparisons. Mean ± SEM is indicated by horizontal lines (E, I, K, M).

Dynein 1 complex is the only known molecular motor for pole-directed movement along spindle microtubules. We inactivated this complex using a Cre-loxP approach to disrupt the gene encoding cytoplasmic dynein 1 heavy chain 1 (Dync1h1) in cardiomyocytes. To this end, we crossed αMHC-Cre⁺;Dync1h1^flox/wild mice with Dync1h1^flox/flox mice (Dahl et al., 2021) and confirmed the reduction of dynein 1 heavy chain 1 protein by immunofluorescence microscopy and Western blot (Fig. S4A–C). M-phase cardiomyocytes from neonatal αMHC-Cre⁺;Dync1h1^flox/flox (Dync1h1 CKO) hearts exhibited disordered localization of SERCA2a without showing SERCA2a accumulation at the spindle poles (Fig. 8F–G), indicating that dynein 1 is not required for the disorganization of striated SERCA2a patterns but is necessary for SERCA2a accumulation. We cross-checked this result in NRVMs by inhibiting dynein 1 with the chemical ciliobrevin D and observed a similar SERCA2a distribution (Fig. 8H–I). Collectively, these results indicate that cytoplasmic dynein 1 is not required for the disorganization of striated SERCA2a patterns in M-phase but is required for SERCA2a accumulation at the spindle poles.

Nocodazole and ciliobrevin D provided the tools to examine the consequence of preventing SERCA2a accumulation for the decrease of CaTs during M-phase. Induction of microtubule depolymerization with nocodazole (Video 10, Fig. 8J–K), and inhibition of dynein 1 with ciliobrevin D (Fig. 8L–M), showed that CaTs remained decreased in M-phase NRVMs (Fig. 8K, M). In conclusion, SERCA2a accumulation at the spindle poles is not responsible for the reduction of CaTs during M-phase.

Video 10. (Corresponding to Fig. 8J) Preventing SERCA2a accumulation through nocodazole-induced microtubule depolymerization does not inhibit the reduction of CaTs during M phase.

Download video file ^{(872.8KB, mp4)}

Ca²⁺ signal in NRVMs (P1) were detected using Rhod2. Nocodazole (20 μM, 60–90 minutes) treatment was applied to induce microtubule depolymerization. DMSO served as the control treatment (Ctrl). Left: NRVMs treated with DMSO (Ctrl). Left top panels: Brightfield imaging was used to identify M-phase NRVMs based on morphology, with a movie showing Ca²⁺ signals detected by Rhod2 fluorescence. Scale bar: 40μm. Left bottom panel: CaTs (F/F₀) of the NRVMs shown in the top panels. Red curve: M-phase NRVM (indicated by red arrows in the top panels); black curve: interphase NRVM (indicated by white arrows in the top panels). Right: NRVMs treated with nocodazole. Right top panels: Brightfield imaging was used to identify M-phase NRVMs based on morphology, with a movie showing Ca²⁺ signals detected by Rhod2 fluorescence. Right bottom panel: CaTs (F/F₀) of the NRVMs shown in the top panels. Red curve: M-phase NRVM (indicated by red arrows in the top panels); black curve: interphase NRVM (indicated by white arrows in the top panels). Frame rate: 30 frames/second.

Dync1h1 gene inactivation reduces cardiomyocyte proliferation

To determine the effect of Dync1h1 inactivation on cardiomyocyte proliferation in mice, we first examined apoptosis in neonatal Dync1h1 CKO hearts. Dync1h1 CKO hearts did not show cardiomyocyte apoptosis (Fig. S4D). We then examined the S- and M-phase activity in vivo using BrdU- and H3P-immunofluorescence. After administering BrdU using intraperitoneal injections on P5 and P6, we examined cardiomyocyte cell cycle activity on P7 using immunofluorescence microscopy. Dync1h1 CKO did not alter the percentage of BrdU⁺ cardiomyocytes, indicating that Dync1h1 is not required for S-phase entry (Fig. 9A–B). However, Dync1h1 CKO decreased M-phase activity in P7 mice (Fig. 9C–D), resulting in a 25% decrease in cardiomyocyte numbers (Fig. 9E). The reduction in the number of cardiomyocytes did not alter heart function (Fig. S4E–F) or cardiomyocyte cell cycle activity in adult mice (Fig. S4G). The results show that Dync1h1 KO reduces cardiomyocyte proliferation, and the mechanisms may involve insufficient SERCA2a accumulation, as well as the reported requirement of Dync1h1 for chromosome segregation (Gonczy et al., 1999).

Figure 9. Cytoplasmic dynein heavy chain 1 (Dync1h1) knockout decreases cardiomyocyte proliferation.

Dync1h1 gene in cardiomyocytes was knocked out by breeding αMHC-Cre⁺;Dync1h1^Flox/Wild mice with Dync1h1^Flox/Flox mice. BrdU (61 μg/g body weight, one injection/day) was administered at P5 and P6 via intraperitoneal injection to label S-phase entry. Hearts were sectioned at P7. Immunofluorescence microscopy showed that Dync1h1 knockout in NMVMs did not alter S-phase entry (A-B; N = 3 hearts per group) but reduced the number of cardiomyocytes in M-phase as indicated by H3P staining (C-D; N = 5 hearts per group), (E) Consequently, total number of cardiomyocytes was reduced in P7 Dync1h1 CKO hearts (Ctrl: n = 4 hearts, Dync1h1 CKO: n = 8 hearts). Control group (Ctrl): αMHC-Cre⁻;Dync1h1^Flox/Flox, Dync1h1 CKO group: αMHC-Cre⁺;Dync1h1^Flox/Flox. Scale bars: 10 μm (A, C). Statistical analyses: (B, D, E) Student’s t-test (unpaired, Two-tailed). Mean ± SEM is indicated by horizontal lines (B, D-E). arb. u.: arbitrary unit (B, D).

Metaphase cardiomyocytes remodel the endoplasmic reticulum

To compare M-phase remodeling of the SR and ER, we examined the distribution of KDEL (Lys-Asp-Glu-Leu) (Karabasheva and Smyth, 2019), an ER retention signal, in M-phase NRVMs (Fig. 10A–B). The KDEL remodeling pattern paralleled that of SERCA2a, progressing from initial disorganization to pronounced accumulation at the spindle poles. To examine the mechanisms underlying KDEL redistribution, NRVMs were treated with nocodazole and ciliobrevin D. Disruption of microtubules with nocodazole abolished KDEL accumulation at the spindle poles, indicating a requirement for an intact microtubule network (Fig. 10C–D). In contrast, inhibition of dynein 1 with ciliobrevin D did not prevent KDEL accumulation at the spindle poles in M-phase cardiomyocytes (Fig. 10C–D). Together, these findings indicate that SR and ER remodeling during M-phase share common features while also exhibit differences, consistent with the involvement of overlapping and distinct mechanisms.

Figure 10. Accumulation of the endoplasmic reticulum marker KDEL at the spindle poles during M-phase requires the mitotic spindle.

A general ER peptide sequence, KDEL (Lys-Asp-Glu-Leu), was stained to examine ER in NRVMs (P1). Mitotic spindles were identified using antibodies against tubulin (red). Microtubule-organizing centers and spindle poles are indicated by yellow stars. Different sub-phases of M-phase were determined by the organization of condensed chromosomes (blue) and mitotic spindles. (A) In interphase and prophase cardiomyocytes, KDEL exhibited striated patterns. During prometaphase, KDEL began to accumulate around the spindle poles (yellow star). The accumulation was maximal in metaphase, weakened in anaphase, and became undetectable in telophase as ER striations reappeared. (B) Quantification showed that KDEL accumulation at spindle poles was present in all metaphase cardiomyocytes, declined during anaphase, and became undetectable in telophase. Each symbol represents one cell culture (n = 3 cell cultures). (C-D) ER remodeling requires intact microtubules, but not dynein 1. NRVMs were treated with nocodazole (20 μM, 90 minutes) to disrupt microtubules or with ciliobrevin D (120 μM, 60 minutes) to inhibit dynein 1. Control cells (Ctrl) received DMSO. (C) NRVMs in prometaphase or metaphase were imaged and analyzed by immunofluorescence microscopy for KDEL accumulation at the spindle poles. (D) Disruption of microtubules with nocodazole prevented KDEL localization at the spindle poles, whereas inhibition of dynein 1 with ciliobrevin D did not. Each symbol represents one cell culture (n = 3 cell cultures per group). (E) Summary of Ca²⁺ valley formation, SERCA2a redistribution, and CaTs changes in proliferating cardiomyocytes. Baseline [Ca²⁺]ᵢ levels in interphase and M-phase cardiomyocytes are denoted by dashed lines. (F) Molecular and cellular model of CDK1-induced changes of Ca²⁺-signaling and SR structure in M phase. Scale bars: (A, C) 10 μm. Statistical significance of differences between sub-phases of M-phase and interphase was analyzed with One-way ANOVA followed with Bonferroni’s correction for multiple comparisons (B, D). Mean ± SEM is indicated by horizontal lines (B, D).

Cardiomyocytes derived from human iPSCs exhibit structural and functional changes during M-phase

We next extended our analysis to commercially available human induced pluripotent stem cell (iPSC)–derived cardiomyocytes. Entry into M-phase was induced by treatment with neuregulin 1 (NRG1; 100 ng/mL), as previously described (Bersell et al., 2009). Consistent with our observations in NRVMs (Fig. 4C–D), human iPSC-derived cardiomyocytes exhibited a disorganized SERCA2a distribution during M-phase (Fig. S5A–B). In these cells, SERCA2a accumulated in proximity to the mitotic spindle, whereas spindle poles showed relatively low SERCA2a signal (Fig. S5A), in contrast to the enrichment at spindle poles observed in NRVMs (Fig. 4C). We next examined the functional consequences of these structural changes. Inhibition of dynein 1 with ciliobrevin D restored CaT amplitudes in M-phase human iPSC-derived cardiomyocytes to levels comparable to those observed in interphase cells (Fig. S5C–D). Together, these results indicate that cardiomyocytes derived from human iPSCs undergo SR disorganization and reduced CaTs during M-phase. Comparison with NRVMs reveals differences in the spatial distribution of SERCA2a and in functional responses, consistent with the involvement of both shared and distinct mechanisms.

DISCUSSION

Our study identifies dynamic and reversible remodeling of Ca²⁺-handling structures and signaling during cardiomyocyte mitosis and reveals an unexpected integration of Ca²⁺ homeostasis with cell division. Building on our experimental observations, we propose a three-phase model for Ca²⁺ remodeling during M-phase (Fig. 10E). During prometaphase, the normally striated organization of the SR becomes disordered, coinciding with a reduction in CaT amplitudes and cytoplasmic Ca²⁺ levels. During metaphase, SERCA2a associated with the SR redistributes toward the spindle poles, accompanied by localized regions of reduced Ca²⁺ concentration at these sites, which we refer to as Ca²⁺ valleys. Perturbation of these Ca²⁺ valleys uncovered their functional importance for cytokinesis and, by extension, cardiomyocyte proliferation. Finally, during cytokinesis, both SR organization and Ca²⁺ signaling are restored to their pre-mitotic state. While this framework captures shared principles of Ca²⁺ remodeling during cardiomyocyte division, our results also indicate that the spatial organization and functional consequences of these changes can vary across cardiomyocyte types and maturation states.

The consequences of these structural and functional changes are profound, as M-phase cardiomyocytes stop contracting. However, the duration of these changes is short, as M-phase lasts only 1–2 hours (Baniol et al., 2021; Hahn et al., 2009), and only few cardiomyocytes are in M-phase at any given time. For example, in human infants, only 0.01 – 0.04% of cardiomyocytes are in M-phase (Mollova et al., 2013). Consequently, the profound structural and functional changes at the level of single M-phase cardiomyocyte are balanced by their transient nature and infrequent occurrence, ensuring that the overall pump function of the heart is not compromised by proliferation of individual cardiomyocytes. Neonatal rat and mouse cardiomyocytes, as well as fetal human cardiomyocytes, show similar changes in the SR and CaTs, although human iPSC-derived cardiomyocytes exhibit subtle differences.

The heart functions as a functional syncytium of electrically coupled cardiomyocytes, allowing for the rapid propagation of action potentials and ensuring synchronous contraction. Our results show that action potentials are present in M-phase cardiomyocytes, suggesting that cardiomyocytes separate reduced Ca²⁺-release from action potentials in M-phase. Electrical coupling is present between M-phase cardiomyocytes and their neighboring cells. In this way, the integrity of the electrical syncytium could be maintained, while individual M-phase cardiomyocytes can undergo significant structural and functional changes. Although peak CaTs in M-phase cardiomyocytes are reduced, several lines of evidence suggest that the individual functional elements of Ca²⁺-induced Ca²⁺ release can be activated: First, Ca²⁺-release from the SR can be stimulated by caffeine treatment. Second, although electrical stimulation did not increase CaTs in M-phase cardiomyocytes under baseline conditions, forskolin treatment without external electrical stimulation enhanced excitation coupling between neighboring cells and the resulting Ca²⁺ release, thus restoring CaTs amplitudes to interphase levels. In conclusion, the complex subcellular structures required for Ca²⁺-induced Ca²⁺-release do not seem to disassemble into non-functional individual components during M-phase. Instead, these subcellular structures appear to separate in a coordinated manner into functional subunits. This could ensure that the contractile function is rapidly restored during cytokinesis.

Thapsigargin-mediated inhibition of SERCA2a underscores the importance of low Ca²⁺ levels for M-phase progression. This study revealed that elevating Ca²⁺ levels during prometaphase and metaphase, but not anaphase or telophase, causes cytokinesis failure, leading to one binucleated daughter cardiomyocyte. It is interesting to note that thapsigargin-induced increase of Ca²⁺ levels during prometaphase or metaphase did not alter karyokinesis, suggesting that Ca²⁺-regulated mechanisms may differ between karyokinesis and cytokinesis. The finding that SERCA2a inhibition reduces S-phase entry indicates that elevated Ca²⁺ levels can affect cardiomyocyte proliferation at multiple stages of the cell cycle.

Rapid, phasic oscillations in cytosolic Ca²⁺ levels are a defining feature of cardiomyocytes, distinguishing them from the slow and tonic Ca²⁺ release observed in many other cell types (Berridge et al., 2003; Fearnley et al., 2011). In cardiomyocytes, [Ca²⁺]_i typically ranges from ~0.1 μM during relaxation to ~1 μM during contraction (Bers and Guo, 2005; Dewenter et al., 2017), whereas in most non-excitable cells, [Ca²⁺]_i remains relatively stable at around 0.1 μM (Bootman and Bultynck, 2020). While a noticeable Ca²⁺ release—i.e., an increase in cytosolic Ca²⁺ levels—during the metaphase-to-anaphase transition has been reported in non-cardiomyocytes (Poenie et al., 1986), we observed a reduction in CaTs during M-phase in cardiomyocytes. It is widely accepted that the mechanisms governing M-phase progression are conserved across diverse cell types. However, the contrasting pattern of Ca²⁺ modulation suggests that precise Ca²⁺ modulation is required to activate Ca²⁺-sensitive pathways specifically during M-phase, and more precisely, during the metaphase-to-anaphase transition. These findings underscore the need for further investigation of how Ca²⁺ dynamics are specifically modulated to support cell division in specialized cells such as cardiomyocytes.

Significant structural changes of the ER during mitosis have been previously reported in other cell types (Ellenberg et al., 1997; Lu et al., 2009; Nourbakhsh et al., 2021; Smyth et al., 2012; Yang et al., 1997). Our results demonstrate in cardiomyocytes that the SR markers RyR2 and SERCA2a, as well as the ER marker KDEL, undergo pronounced reorganization during M-phase; however, only SERCA2a and KDEL accumulate at the spindle poles, whereas RyR2 does not. Consistent with our findings, KDEL has been reported to localize to spindle poles in other cell types (Karabasheva and Smyth, 2019), whereas the ER marker calnexin was not reported to exhibit spindle pole accumulation (He and Davie, 2006). Mechanistically, spindle pole–associated accumulation of SERCA2a requires dynein 1 activity, whereas KDEL redistribution seems to occur independently of dynein 1. Our new findings, together with the literature, underscore the heterogeneity of SR and ER remodeling during mitosis and add to the view that distinct organelle subdomains could be differentially regulated as cells progress through M-phase.

In summary, our results show that cardiomyocytes enter M-phase before they change their SR and ER structures and before they reduce CaTs (Fig. 10E). Stimulating M-phase entry by increasing the activity of CDK1 is sufficient to induce these structural and Ca²⁺ signaling changes (Fig. 10F). Furthermore, our results show that CDK1 activity is required to maintain these changes during M-phase. However, activity of the CDK1/cyclin B complex without achieving M-phase entry does not seem to be sufficient to induce these changes. Thus, CDK1 activity provides a mechanistic connection between M-phase, the structural changes of the SR, and functional changes in Ca²⁺ signaling, although the specific CDK1 phosphorylation targets and mechanisms have to be further defined. Additionally, our results indicate that the reduction in CaT amplitude is not caused by disorganization of striated SERCA2a patterns or by subsequent SERCA2a accumulation at the spindle poles, suggesting that other mechanisms downstream of CDK1 may contribute to this reduction in CaT amplitude. Our results define a general blueprint for changes of Ca²⁺-signaling in M-phase cardiomyocytes. The expression of this blueprint may differ between different types of cardiomyocytes and between other cell types, whose function requires Ca²⁺ signaling.

Materials and Methods

Approval of research involving human samples and animals

The Institutional Review Board (IRB) at the University of Pittsburgh (protocol PRO14100621) approved research involving human hearts and samples. All animal experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Pittsburgh and Weill Cornell Medical College.

Human heart samples

Human fetal hearts (17-week gestational age) were collected from the Health Tissue Bank at Magee Women’s Hospital, University of Pittsburgh, approved under IRB protocol (PRO14100621).

Cell lines

The CiPS001–13 human iPS cells used for Fig. 1A–C were obtained as a gift from Dr. Dan Roden and Kevin R. Bersell (Vanderbilt University). The CSC-C2847SC human iPSC-derived cardiomyocytes kit used for Fig. S5 was purchased from Creative Bioarray (Cat# CSC-C2847SC).

Animal models

For this study, we bred a mouse line that expresses a calcium reporter (αMHC-GCaMP8, Michael Kotlikoff, Cornell University) (da Silva Lopes et al., 2011) simultaneously with the mCherry-Geminin cell cycle reporter system (René Maehr, UMass Chan Medical School) (Sakaue-Sawano et al., 2008). This new mouse line enabled us to identify cardiomyocytes with the cardiomyocyte-specific GCaMP8 expression in the cytoplasm and specifically characterize them as either in the cell cycle (Geminin⁺) or not (Geminin⁻). We also bred a mouse line expressing αMHC-Cre (The Jackson Laboratory, Strain #:011038, RRID:IMSR_JAX:011038) and Dync1h1^flox (Wolfgang Baehr, University of Utah)(Dahl et al., 2021) to generate cardiomyocyte-specific Dynein 1 heavy chain knockout mice (αMHC-Cre⁺; Dync1h1^flox/flox). Mice and rats (Charles River, Sprague Dawley, Strain: Crl:CD(SD), RRID:RGD_734476) were used at P1-P3, or P7 for cardiomyocyte and intact heart isolations, utilizing the natural distribution of male and female mouse pups.

Animal experiments

For this study, we bred a mouse line that expresses a calcium reporter (αMHC-GCaMP8) (da Silva Lopes et al., 2011) simultaneously with the mCherry-Geminin cell cycle reporter system (Sakaue-Sawano et al., 2008). This new mouse line enabled us to identify cardiomyocytes with cardiomyocyte-specific GCaMP8 expression in the cytoplasm and to specifically characterize them as being in the cell cycle (Geminin⁺) or not (Geminin⁻). We also bred a mouse line expressing αMHC-Cre and Dync1h1^flox to generate cardiomyocyte-specific Dynein 1 heavy chain knockout mice (αMHC-Cre⁺;Dync1h1^flox/flox). We used mice and rats at P1, P3, or P7 for cardiomyocyte and intact heart isolations, utilizing the natural distribution of male and female mouse pups.

Verification of Genetic Mouse Models

αMHC-GCaMP8 Mouse

To determine whether GCaMP8 expression in cardiomyocytes affects cellular contraction, cardiomyocytes were isolated from newborn mouse pups expressing GCaMP8. The isolated cells were cultured for one day. Cells were then incubated in medium containing SiR-actin (1 nM; Cytoskeleton, Cat. #CY-SC001) for 1 hour to label actin filaments. After three washes with fresh medium, cardiomyocyte contraction was recorded using a Nikon AXR confocal microscope equipped with a resonant scanner. Changes in sarcomere length during contraction were measured from the recorded videos. Sarcomere length was determined by measuring the distance between dark regions (M-line positions) located between adjacent actin stripes labeled with SiR-actin. To improve accuracy, the lengths of two consecutive sarcomeres were measured and averaged. Measurements were taken at the end of relaxation (diastole) and contraction (systole). The sarcomere shortening fraction was calculated using the following equation:

$$\text{Sarcomere}\text{shortening}\text{fraction}=\frac{L_{\text{Relaxation}}-L_{\text{Contraction}}}{L_{\text{Relaxation}}}\times 100\%$$

• $L_{\mathit{\text{Relaxation}}}:$ Sarcomere length at the end of

• $L_{\mathit{\text{Contraction}}}$ Sarcomere length at the end of

Three consecutive contraction cycles were analyzed to obtain three sarcomere shortening fraction values. The mean of these values was used as the representative sarcomere shortening fraction for each cell in subsequent analyses.

αMHC-Cre⁺;Dync1h1^flox/flox mouse

We used Western Blot assay to confirm Dync1h1 knockout.

Protein Lysate Preparation and Measurement.

Mouse hearts were minced with a razor blade and digested with HALT Protease & Phosphatase Inhibitor (Invitrogen, Cat# PI78442) diluted in RIPA buffer at a ratio of 500μL per 10mg of tissue. Tissue was homogenized via sonication, vortexed, and rocked in a cold room for 10 minutes, followed by agitation using a 30G needle before being centrifuged at 10,000 rcf for 15 minutes at 4°C. The supernatant was aliquoted and stored at −80°C. Protein concentration was measured using the Pierce BCA Protein Assay Kit (Fischer Scientific, Cat# 23227) and a nanodrop spectrophotometer. Western Blot. Protein samples were prepared for 50 μg/lane. Samples were combined with RIPA lysis buffer and 5x Laemmli Buffer (Fischer Scientific, Cat# 39000), then boiled at 97 °C for 10 minutes. After cooling and centrifugation, samples were loaded onto Novex Tris-Glycine Protein gels (4–12%) (Invitrogen, Cat# XP04120BOX) and run at a constant 150 volts for 2 hours 15 minutes. After electrophoresis, gels were transferred to a nitrocellulose membrane using a transfer apparatus with a western sponge, blotting pads, and nitrocellulose transfer paper. The transfer was performed overnight at 34 volts in a 4 °C refrigerator. The next day, the nitrocellulose membrane was washed in TBS-T and blocked with 5% dry milk diluted in TBS-T for 1 hour at room temperature. The membrane was then incubated overnight with primary antibodies (Dync1h1 1:500 (Proteintech, Cat# 12345–1-AP), Vinculin 1:3000 (Sigma Aldrich, Cat# V9131)) in 5% milk. Membrane were then washed with TBS-T and treated with horseradish peroxidase secondary antibody at dilutions of 1:500 and 1:3000, respectively, for 1 hour at room temperature. After secondary antibody incubation, the membrane was treated with ECL prime (Sigma Aldrich, Cat#GERPN2236), made according to the manufacturer’s instructions, and applied to the membrane for 3 minutes in the dark. Imaging was performed using a BioRad ChemiDoc Imaging System.

Inclusion and exclusion criteria

If no M-phase cardiomyocytes were identified in a given cell isolation or individual mouse heart, all cardiomyocytes from that isolation or heart were excluded from analysis. Calcium transients in M-phase cardiomyocytes were analyzed from recorded videos of Rhod2, Fluo4, or GCaMP8 fluorescence, using interphase cardiomyocytes captured in the same videos as controls. Videos were excluded from analysis if interphase cardiomyocytes did not exhibit visible CaTs.

Mouse and rat cardiomyocyte isolation and culture

Neonatal mouse and rat cardiomyocytes were isolated from P1-P2 pups using the Cellutron Neonatal Myocyte Isolation System (Cellutron, Cat# nc-6031) according to the manufacturer’s instructions. All procedures were performed under sterile conditions in a tissue culture hood unless otherwise specified.

Preparation of Solutions

On the day of isolation, all solutions were freshly prepared using sterile techniques in 50 mL conical tubes. Solution D1 was prepared by diluting 5 mL of D1 stock with 45 mL of autoclaved distilled water, followed by sterile filtration and storage on ice. Solution D2(EC) was prepared by combining 20 mL of D2 stock with 28 mL of autoclaved distilled water and 2 mL of EC supplement. The solution was sterile-filtered and maintained at 37°C in a water bath.

Solution D3 was prepared by mixing 15 mL of D3 with 25 mL of NS medium, followed by sterile filtration and storage in the tissue culture hood.

Heart Isolation

Hearts were isolated and immediately transferred into 12 mL tubes containing 5 mL of ice-cold D1 solution. Samples were maintained on ice until further processing.

Tissue Processing and Cell Isolation

A sterile 10 cm Petri dish was prepared inside the tissue culture hood. Dissection tools, including blades, were sterilized with 70% ethanol prior to use. Isolated hearts were washed three times with 5 mL of D1 solution to remove residual blood cells. The tissues were then transferred to the Petri dish, and atrial regions were removed. Ventricular tissue was supplemented with a small volume of D2(EC) buffer and minced into approximately 0.5 mm fragments. The resulting tissue fragments were transferred into 12 mL tubes containing 5 mL of D2(EC) buffer.

Pre-digestion

Tissue samples were subjected to an initial digestion step to remove dead cells. Tubes were placed in an incubator shaker and incubated for 10 minutes. Following incubation, the supernatant was carefully removed using vacuum aspiration. During this step, the D2(EC) solution was maintained at 37°C in a water bath.

Enzymatic Digestion and Cell Collection

Fresh 5 mL aliquots of D2(EC) buffer were added to each tube containing tissue fragments, followed by incubation for 15 minutes in the shaker incubator. After digestion, the supernatant containing dissociated cells was transferred to tubes containing D3 solution and gently mixed. This digestion and collection process was repeated until complete tissue dissociation was achieved.

Cell Purification

Cell suspensions collected in D3 buffer were centrifuged at 700 rpm for 10 minutes. The supernatant was carefully aspirated, and the cell pellet was resuspended in 10 mL of DMEM supplemented with 10% fetal bovine serum (FBS). The resulting cell suspension was transferred to 10 cm Petri dishes and gently distributed to ensure even coverage of the culture surface. Cells were incubated at 37°C for 60 minutes. Following incubation, the cell suspension was collected and passed through a 70 μm cell strainer into a clean 50 mL tube to remove cell aggregates and undigested tissue fragments. Cells were subsequently counted using a hemocytometer.

Surface Coating

Ibidi culture dishes (ibidi dish grid-500, ibidi# 81166, Fisher Scientific, Cat#50-305-809) or 4-well chamber slides (Fisher Scientific, Cat#12-565-7) were coated with fibronectin to promote cell adhesion. Briefly, 300 μL of fibronectin solution ((20 μg/mL, Fisher Scientific, Cat# PHE0023) was added to each dish and incubated at room temperature to allow uniform surface coverage. Following incubation, excess fibronectin solution was carefully removed immediately prior to cell seeding.

Cell Seeding

Cell concentration was determined and adjusted to 8 × 10⁵ cells/mL. Fibronectin solution was removed from the ibidi dishes, and 500 μL of the prepared cell suspension (4 × 10⁵ cells per dish) was added to each dish. After plating, dishes were left undisturbed for 15 minutes to allow cells to settle and adhere. Dishes were then carefully transferred to a 37°C incubator.

Cell culture

Cells were cultured for 3 hours under standard conditions. Subsequently, dishes were gently agitated to dislodge non-adherent cells, and the medium was immediately removed. Fresh DMEM/High Glucose (Fisher Scientific, Cat# 11960044), supplemented with 10% FBS (2 mL per dish) and 1% (100 U/mL) penicillin-streptomycin (Thermo Fisher, Cat#15140122) was added, and cells were returned to the incubator for continued culture.

Immunofluorescence staining

Cultured cardiomyocytes were fixed with 3.7% formaldehyde (Fisher Scientific, Cat#AC410731000) or 4% paraformaldehyde (Fisher Scientific, Cat#J19943.K2) for 12 minutes at room temperature. Cells were immersed in a permeabilizing and blocking solution (0.5% Triton X-100, Fisher Scientific, Cat#BP151–500; 5% donkey serum, Abcam, Cat#ab7475; or goat serum, Sigma Aldrich, Cat#G9023) in PBS for 30 minutes at room temperature. Subsequently, cells were incubated with primary antibodies at a 1:200 dilution or according to the vendor’s recommendations at 4°C overnight. Secondary antibodies were applied at 1:200 to 1:400 dilution for 1 hour at room temperature. The nuclei were stained with Hoechst 33342 (1:1,000 dilution, Fisher Scientific, Cat# H3570) for 5 minutes at room temperature. The cells were then mounted with a mounting medium containing 1% N-propyl-gallate (Sigma Aldrich, Cat#2370) dissolved in glycerol (Sigma Aldrich, Cat#G5516–1L) and sealed with nail polish. List of primary antibodies: Goat polyclonal to Cardiac Troponin I (Abcam, Cat#ab56357; RRID: AB_880622), Rat monoclonal [YL1/2] to Tubulin (Abcam, Cat#ab6160; RRID: AB_305328), Rabbit polyclonal to RyR2 (Thermo Fisher, Cat#PA5–87416; RRID: AB_2804131), Mouse monoclonal [IID8] to SERCA2 ATPase (Abcam, Cat#ab2817; RRID: AB_2061427), Mouse monoclonal to α-Actinin (Sarcomeric) (Sigma Aldrich, Cat#A7811; RRID: AB_476766), Rabbit monoclonal [SP6] to Ki67 (Abcam, Cat#Ab16667; RRID: AB_302459), Rabbit polyclonal to phospho-Histone H3 (Ser10) (Millipore, Cat#06–570; RRID: AB_310177), Mouse monoclonal [mAbcam 3609] to Aurora B (Abcam, Cat#ab3609; RRID: AB_449204), Rabbit KDEL Polyclonal Antibody (Thermo Fisher, Cat#PA1–013; RRID: AB_325593), Recombinant rabbit monoclonal to γ-tubulin (Thermo Fisher, Cat# MA5–38194; RRID:AB_2898111), Rabbit polyclonal to Pericentrin (Thermo Fisher, Cat# PA5–116926; RRID: AB_2901556). List of secondary antibodies: Donkey anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 488 (Thermo Fisher, Cat# A-21202; RRID: AB_141607), Donkey anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 568 (Thermo Fisher, Cat#A10037; RRID:AB_2534013), Donkey anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 594 (Thermo Fisher, Cat#A-21203; RRID:AB_141633), Donkey anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 647 (Thermo Fisher, Cat#A-31571; RRID:AB_162542), Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 488 (Thermo Fisher, Cat#A-21206; RRID:AB_2535792), Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 594 (Thermo Fisher, Cat#A-21207; RRID:AB_141637), Anti-Rabbit IgG (H+L), highly cross-adsorbed, CF^™ 647 antibody produced in donkey (Sigma-Aldrich, Cat# SAB4600177; RRID:AB_3741793), Donkey anti-Rat IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 488 (Thermo Fisher, Cat#A-21208; RRID:AB_2535794), Anti-Rat IgG (H+L), highly cross-adsorbed, CF^™ 568 antibody produced in donkey (Sigma-Aldrich, Cat#SAB4600077; RRID:AB_2827516), Donkey Anti-Rat IgG H&L (Alexa Fluor^® 647) preadsorbed (Abcam, Cat#ab150155; RRID:AB_2813835), Goat anti-Guinea Pig IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 594 (Thermo Fisher, Cat#A-11076; RRID:AB_2534120).

Microscope image acquisition and postprocessing

Images and videos were acquired with a Nikon confocal microscope (A1R or AXR) or a Nikon (TiE or Ti2) epifluorescence microscope using the Nikon NIS-Elements software (Version 6.02.01, https://www.microscope.healthcare.nikon.com/products/software/nis-elements, RRID:SCR_014329). The numerical aperture of the Nikon Plan Fluor 40x Oil objective lenses is 1.3. The numerical aperture of the Nikon Plan Apo λ 100x Oil is 1.45. Cameras used to record calcium transients are Andor Zyla (Model# VSC-00311) and HAMAMATSU (Model# C15440–20UP). The Elements Denoise.ai module was utilized to reduce noise in the acquired images. Calcium transients (CaTs) data and membrane potential (Cytovolt1) data were extracted from raw videos, i.e., not processed with the Elements Denoise.ai module, recorded using the NIS-Elements software. The images and videos were processed using Fiji software (https://imagej.net/software/fiji/, RRID: SCR_002285) for publication. Both the calcium signal (Rhod2 fluorescence) and the membrane potential signal (Cytovolt1 fluorescence) were saved as grayscale values for each pixel in the raw Nikon microscope data files (ND2 format). In ND2 files capturing calcium signals, higher grayscale values indicate stronger Rhod2 fluorescence and, consequently, a higher calcium signal. Similarly, in ND2 files recording membrane potential signals, higher grayscale values reflect stronger Cytovolt1 fluorescence. Changes in grayscale values across time frames correspond to changes in membrane potential in response to field stimulation. The exact time points for each video frame are stored in the ND2 files and was extracted using Fiji. Only videos in which all interphase cardiomyocytes neighboring the M-phase cardiomyocytes exhibited CaTs were included in the final data analysis. To extract CaTs data from cardiomyocytes in the calcium fluorescence recordings, the target cells were outlined in Fiji using the Polygon Selections tool, and the selected areas were added to the ROI Manager. A blank region, located in an area without cells, was selected as a background reference and used for subtraction of background noise. The Multi Measure function in ROI Manager was used to measure the Mean gray value of each selected cardiomyocyte and the background region across all frames of the recorded CaTs video. Background values were subtracted from the corresponding cellular signals to eliminate background noise. Time values corresponding to each video frame were extracted from the same ND2 files. The minimum value of CaTs was defined as F₀, and calcium levels at each time point were calculated using the formula CaTs = F/F₀. The maximum CaTs (F/F₀) value in the entire CaT trace was reported as the Peak CaTs or amplitude in this study. In the representative images, the CaT tracing curves were not filtered or smoothed using any mathematical methods. Cytovolt1 fluorescence was measured to detect the membrane potential changes (action potentials). The derivation of membrane potential changes followed the definition of the lowest Cytovolt1 fluorescence as baseline and the definition of noise background. To acquire images for quantifying cycling cardiomyocytes (e.g., M-phase cells), five to ten areas per sample were randomly selected under a confocal microscope. A sufficient number of cells (500–2000 cells per image) were captured by adjusting the stitching settings in the large-area acquisition function provided by the Nikon NIS-Elements software. To avoid bias during quantification, investigators were blinded to the treatment groups by concealing the labels until quantification was completed.

Identification of interphase and sub-phases of M-phase

We used phospho-Histone 3 (H3P) immunofluorescence microscopy to identify M-phase and Hoechst immunofluorescence microscopy to distinguish sub-phases of M-phase. For the latter, we examined the morphology of condensed chromosomes and, when possible, in the experimental design, the presence and morphology of spindle microtubules (highlighted with a Tubulin antibody in fixed cells or the SiR-tubulin dye in live cells). This is necessary because, although H3P identifies M-phase, it does not distinguish sub-phases of M-phase.

To identify live M-phase neonatal rat ventricular myocytes (NRVMs) for live cell calcium measurement experiments, we first identified M-phase by cell morphology using bright field microscopy (Fig. S2A, and Source Data). M-phase was then confirmed with post hoc immunofluorescence microscopy. To this end, immediately after live cell calcium measurement, cells were fixed, stained with antibodies against α-actinin or troponin I (cardiomyocyte markers) and Ki67 or H3P (cell cycle markers), and imaged with a Nikon TiE epifluorescence microscope. The morphology of M-phase NRVMs identified under the bright field was associated with the Ki67 or H3P fluorescent markers, identifying M-phase using the following criteria (Fig. S2A):

Interphase –

Nucleus with a smooth boundary and no chromosome condensation.

Prophase –

Nucleus with a smooth boundary, H3P or Ki67 staining, and condensed chromosomes.

Prometaphase –

Microtubule organization in a spindle shape. Chromosomes are condensed and not yet aligned at the midzone. Filopodia-like structures start protruding from the cell periphery, making the M-phase cardiomyocyte outline different from interphase.

Metaphase –

Chromosomes aligned at the midzone. The mitotic spindle is properly shaped and symmetrically aligned with the chromosomes at the equator.

Anaphase –

Aligned sister chromosomes are separated along the spindle axis. Filopodia remain. The midzone exhibits a rod shape.

Telophase and cytokinesis –

Round daughter nuclei are present. Cleavage furrow ingression creates a dumbbell-shaped cell morphology.

A detailed description of identifying cell cycle phases by examining cell morphology under a bright field microscope can be found in Fig. S2A.

Quantification of sarcoplasmic/endoplasmic reticulum remodeling

We used antibodies against RyR2 and SERCA2a to study SR remodeling and an antibody against KDEL to study endoplasmic reticulum remodeling.

Cardiomyocytes with striated RyR2/SERCA2a pattern – When RyR2 or SERCA2a striations are present on two side of the associated sarcomere, we labeled the cardiomyocyte as having a ‘striated RyR2 pattern’ or ‘striated SERCA2a pattern,’ respectively. The percentages of NRVMs marked as ‘Cardiomyocytes with striated RyR2 pattern’ and ‘Cardiomyocytes with striated SERCA2a pattern’ were calculated for each cell cycle phase (Fig. S2B). If RyR2, SERCA2a, or KDEL accumulated at the spindle poles as indicated by spindle microtubules, the cells were labeled as “Cardiomyocytes with RyR2/SERCA2a/KDEL accumulation at the spindle pole(s)”.

SR remodeling in cardiomyocytes with dynein 1 knockout - To inactivate the cytoplasmic dynein 1 complex with gene knockout in cardiomyocytes, αMHC-Cre⁺;Dync1h1^flox/wild mice were crossbred with Dync1h1^flox/flox mice. The aMHC-Cre⁺;Dync1h1^flox/flox mice were the test group, and the αMHC-Cre⁺;Dync1h1^flox/wild mice were the control group. Cardiomyocytes from newborn mouse pups (P1) were isolated and cultured overnight, then fixed with 3.7% formaldehyde. Cells were stained with antibodies against SERCA2a, Tubulin, and Troponin I, followed by DNA staining with Hoechst. The distribution of the SR as indicated by SERCA2a was studied under a Nikon A1R confocal microscope.

Measuring CaTs of single cardiomyocytes in intact mouse hearts

Mice expressing the genetically encoded calcium indicator αMHC-GCaMP8 were cross-bred with mice expressing fluorescent cell cycle reporter αMHC-Cre;Rosa26-mCherry-Geminin to generate pups carrying both the fluorescent cell cycle reporter mCherry-Geminin and the calcium indicator GCaMP8 in cardiomyocytes. Mouse pups (P1-P3) were euthanized via decapitation after hypothermia. Hearts were isolated and placed in an ibidi dish (150 μm thick bottom) containing warm medium (DMEM, high glucose, 10% FBS, Blebbistatin 25 μM). The hearts in the dish were stabilized using a parafilm mold with a piece of cover glass on top and stimulated with a 1 Hz electrical field. The calcium signal, indicated by the calcium reporter GCaMP8, was recorded using a Nikon A1R confocal microscope with a high-speed resonant scanner (15–30 fps). Under the control of cardiomyocyte-specific promoter, Myh6 (αMHC), cardiomyocytes express the GCaMP8 calcium sensor, regardless of their cell cycle status. The GCaMP8 indicator emits green fluorescence only when Ca²⁺ ions are bound. Cardiomyocytes in diastole, when Ca²⁺ concentration is low, emit low levels of fluorescence. Although individual cardiomyocytes may exhibit different expression levels of GCaMP8, the experimental approach normalizes the GCaMP8 fluorescence signal change over the cardiac cycle to its own fluorescence in diastole (called F₀). To outline the boundaries of cardiomyocytes exhibiting low GCaMP8 fluorescence, we increased the brightness of the F₀ image using ImageJ/Fiji software to ensure cell boundaries were visible and then outlined the cells. The ratio of the instantaneous calcium fluorescence intensity (F) over the baseline calcium fluorescence (F₀) was calculated as the calcium transients (CaTs). Changes in CaTs over time were traced and presented as CaTs vs. time curves. The peak values of the CaTs vs time curves of each cardiomyocyte were collected (Peak CaTs) and statistically analyzed. For presentation, the images and videos were processed using the Elements Denoise.ai module to improve the image quality.

Live-cell calcium measurements to acquire CaTs in cultured cardiomyocytes

Cultured neonatal rat cardiomyocytes (NRVMs), human primary fetal cardiomyocytes, and human induced pluripotent stem cell-derived cardiomyocytes were studied using this live cell calcium measurement technique. Rhod2-AM (Biotium, Cat#50023) was used as a calcium indicator. Cardiomyocytes cultured in ibidi dishes were loaded with Rhod2-AM (10 μM, in culture medium) or Fluo4-AM (10 μM, in culture medium) for 6 minutes at 37 °C. The cells were then washed with culture medium. After washing, fresh media was added, and the cells were placed in a stage-top incubator (37 °C, 5% CO₂, Tokai Hit) installed on the stage of a Nikon TiE or Ti2 epifluorescence microscope.

Short-time live cell calcium measurement for CaTs in specific cell cycle phases

Cardiomyocytes of interest in specific cell cycle phases, i.e., interphase, prophase, prometaphase/metaphase, anaphase, telophase, and cytokinesis, were identified by morphology and confirmed by post hoc immunofluorescence microscopy. CaTs, indicated by Rhod2 fluorescence, were recorded in videos for 5 – 10 seconds, with or without 1 Hz electrical field stimulation. Immediately after live cell calcium measurement, cells were fixed with 3.7% formaldehyde, followed by immunofluorescence staining of H3P or Aurora B kinase to indicate the cell cycle activity. Grey values of Rhod2 fluorescence intensity were extracted from the recorded videos using Nikon Elements software or Fiji software. The grey values representing CaTs (F/F₀) were analyzed using Microsoft Excel (RRID: SCR_016137) and GraphPad PRISM (Ver. 10, RRID: SCR_002798). CaTs (F) versus time curves were normalized by the baseline (F₀) of each curve. Peak values (peak CaTs) were visually determined.

Long-time live cell calcium measurement to track CaTs throughout the M-phase

Cardiomyocytes in prometaphase were identified by morphology under brightfield microscopy. CaTs were measured for 5 – 10 seconds every 10 minutes until cells divided and generated two daughter cells. The acquired calcium signals were processed to generate CaTs (F/F₀) versus time graphs. Adjacent interphase cardiomyocytes (control) within the same videos were analyzed similarly. The Peak CaTs (F/F₀) from each video were renormalized using the mean Peak CaTs (F/F₀) value of the interphase cells within the same video. Finally, the renormalized peak CaTs (arbitrary unit, a.u) of mitotic cardiomyocytes in different subphases were compared with the peak CaTs (arbitrary unit, a.u) of interphase cardiomyocytes.

Determination of Ca²⁺ signal strength at the spindle poles

Live cardiomyocytes were incubated with an engineered tubulin marker SiR-tubulin (1 nM, Cytoskeleton, Cat# CY-SC002) at 37°C for 60 minutes, then loaded with Rhod2-AM for 6 minutes and washed three times with fresh medium. Labeled cardiomyocytes were examined under a Nikon A1R confocal microscope to identify the M-phase based on spindle microtubule morphology. Acquired images of identified M-phase cardiomyocytes were processed using the Denoise.ai module in NIS Element software to remove background noise. Ca²⁺ signal strength at spindle poles, indicated by the mean grey value, was quantified using ImageJ/FIJI and compared with adjacent regions.

Ratio of Ca²⁺ signal strength at the spindle poles over Ca²⁺ signal strength at the adjacent regions was calculated.

• Ratio = 1 indicates no difference.

• Ratio > 1 indicates stronger Ca²⁺ signal strength at the spindle poles.

• Ratio < 1 indicates weaker Ca²⁺ signal strength at the spindle poles.

A multicolor Ca²⁺ heatmap was generated to visualize the distribution of localized Ca²⁺ signal strength in M-phase cardiomyocytes using the 3D surface plot function of the Fiji software.

Quantification of the cytosolic Ca²⁺ concentration [Ca²⁺]_i in cultured cardiomyocytes

A modified method from a previously published protocol (Del Nido et al., 1998) was used.

The cytosolic Ca²⁺ concentration [Ca²⁺]_i was calculated using the following equation:

$${\left[{\text{Ca}}^{2+}\right]}_{\text{i}}={\text{K}}_{\text{d}}\left({\text{F}}_{\text{t}}-{\text{F}}_{\text{b}}\right)/\left({\text{F}}_{\text{max}}-{\text{F}}_{\text{t}}\right)$$

where K_d=720 nM (constant for cultured cells loaded with Rhod2) represents the calcium dissociation constant of Rhod2 (Du et al., 2001). F_b = 0 (background fluorescence before Rhod2 was loaded). F_t represents time-dependent fluorescence during the CaTs period.

Prior to Rhod2-AM loading, M-phase cardiomyocytes were identified based on morphological criteria, and their positions were recorded using NIS-Elements software (Nikon). Cellular autofluorescence was measured in the Rhod2 channel and used to determine background fluorescence (F_b), which was negligible under the experimental conditions (Fig. 3A).

Cells were then loaded with Rhod2-AM. Following dye loading, cultures were washed with fresh medium to remove residual extracellular dye. Pre-identified cardiomyocytes were relocated under the microscope, and Rhod2 fluorescence was recorded for 5–10 seconds. For calcium transient (CaT) analysis, live-cell imaging was performed to acquire fluorescence time-lapse recordings. To determine maximal fluorescence (F_max), Rhod-2 signals were recorded during the addition of ionomycin (60 μM; Biotium, Cat#59007), which induces Ca²⁺ influx and saturates the dye signal. Saturated fluorescence values (F_max) were subsequently extracted from cardiomyocytes of interest for analysis.

Measurement of caffeine-induced Ca²⁺ release from SR

After loading Rhod2-AM (10 μM, 6 minutes) in cultured NRVMs (P1, day 1), cells were washed with fresh medium. A total of 1,960 μL of fresh medium was added to ibidi dishes (ibidi dish grid-500, ibidi# 81166; Fisher Scientific, Cat# 50-305-809). The dishes were placed on a Nikon TiE epifluorescence microscope. M-phase cardiomyocytes were identified by examining cell morphology using bright field microscopy and stimulated with an electrical field (50 V/cm) at 1 Hz. Under live cell recording of Rhod2 fluorescence, 40 μL of stock caffeine solution (50 mM, Sigma Aldrich, Cat# C0750) was added, resulting in a final caffeine concentration of 1 mM. Cardiomyocytes were then treated with ionomycin (60 μM) under Rhod2 fluorescence recording to calculate cytosolic Ca²⁺ concentrations [Ca²⁺]_i, as described in the previous section. The response of the cells, reflected by Rhod2 fluorescence changes, was analyzed from recorded videos. CaTs curves ([Ca²⁺]_i in nM) calibrated by ionomycin treatment were generated, and the peak [Ca²⁺]_i in nM were extracted.

Examination of membrane potential changes through live cell measurement of Cytovolt1 fluorescence

The engineered fluorescent membrane potential indicator, Cytovolt1 (CytoCybernetics), was used to study the membrane potential of cardiomyocytes in the M-phase. Neonatal rat cardiomyocytes cultured in ibidi dishes were loaded with Cytovolt1 (50 μM) for 10 minutes at 37 °C. Blebbistatin (25 μM, Sigma Aldrich, Cat# B0560–5MG) was added to inhibit cardiomyocyte contractions and reduce cell motion artifacts. The ibidi dishes were then placed in a stage-top incubator (37 °C, 5% CO2, Tokai Hit) installed on a Nikon TiE fluorescence microscope. Under electrical field stimulation (50 V/cm, 1 Hz), the fluorescence emitted by Cytovolt1 was recorded. The membrane potential in cultured NRVMs was measured using the Cytovolt1 fluorescent membrane potential dye. Changes in membrane potential in response to 1 Hz field stimulation were calculated by normalizing Cytovolt1 fluorescence (1-F/F₀), where F₀ represents the maximum Cytovolt1 fluorescence value, and F is the real-time Cytovolt1 fluorescence value. If the peak change in membrane potential was at least twofold greater than the noise level, the cardiomyocyte was classified as having an action potential; otherwise, it was classified as not having an action potential (Fig. 3C–D). The function of cardiomyocytes is not uniformly restored in culture; therefore, it is expected that some cells do not exhibit changes in membrane potential in response to electrical stimulation. We manually quantified cardiomyocytes with or without changes in membrane potential following electrical stimulation. In Fig. 3C–D, representative examples of cardiomyocytes that do or do not respond to electrical stimulation with a change in membrane potential are shown. No additional selection criteria were applied specifically to interphase or M-phase cardiomyocytes during quantification. Thus, the results shown in Fig. 3E reflect the influence of M-phase activity on membrane potential changes in response to electrical stimulation. Fig. 3E indicates that M-phase activity does not affect membrane potential responses to electrical stimulation.

Human induced pluripotent stem cell (iPSC) culture (CiPS001-13 human iPSCs)

CiPS001-13 human iPSCs, generously provided by Drs. Kevin Bersell and Dan Roden (Vanderbilt University), were maintained in mTeSR 1 media (StemTechnologies, Cat# 85850) supplemented with 1% Antibiotic-Antimycotic (Life Technologies, Cat# 15240062) on BD hESC-qualified Matrigel (Fisher, Cat# 08-774-552 Corning, Cat# 354277), diluted according to the manufacturer’s COA. Cells were passaged every five days using ReLeSR dissociation reagent (Stem Cell Technologies, Cat# 5872). Rho-associated kinase inhibitor Y27632 (ROCK, 10 μM, EMD, Cat# 680001) was added to the media for the first 24 hours after passaging. Cells were cultured in mTeSR 1 media with changes every two days until the next passage.

Generation of human iPSC-derived cardiomyocytes. CiPS001-13 human iPSC cells (p48) were split using ReLeSR dissociation reagent, as described above, and seeded onto 12-well growth factor reduced (GFR) Matrigel coated plates (Fisher, Cat# CB-40230A Corning, Cat# 356230, concentration of 0.08 mg/mL in DMEM/F12, Life Tech, Cat# 11330-032) at a ratio of 1:40. Cells were grown in mTeSR 1 media for four days until reaching 80–90% confluency. On day 0, the start of differentiation, the media was switched to differentiation media (RPMI 1640, Life Technologies Cat#11875-093; 2 % B27 minus insulin, Life Technologies Cat#A1895601; 1 % Pen/Strep, Life Technologies Cat#15140122) supplemented with 12 μM GSK3 inhibitor CHIR99021 (Selleck, Cat#S2924). After 24 hours, the medium was replaced with 2 mL of fresh differentiation media and incubated for 48 hours. On day 3, cells were incubated with 2 mL of fresh differentiation media containing 5 μM WNT inhibitor IWR-1 (Sigma, Cat# I0161). On day 5, the media containing IWR-1 was removed and replaced with 2 mL of fresh differentiation media, with media changes occurring every other day until dissociation. On day 7 of differentiation, B27 minus insulin was replaced with B27 Supplement (Life Technologies Cat# 17504044). Contracting cells were observed on day 9. On day 11, post dissociation, differentiation media was replaced with RPMI 1640 w/o glucose (Life Technology Cat# 11879-020) containing 5 μM sodium D-Lactate (Sigma Aldrich, Cat#L4263), 2 % B27 supplement, and 1 % Antibiotic-Antimycotic. Replating of human iPSC-derived cardiomyocytes. On day 15 post differentiation, cells were replated onto GFR Matrigel coated 35 mm ibidi dishes (ibidi dish grid-500 ibidi# 81166, Fisher Scientific, Cat# 50-305-809). Cells were briefly washed with PBS, followed by incubation with 500 μL TrypLE (Life Technology Cat# 12605010) for approximately 10–12 minutes at 37 °C. TrypLE was inactivated by adding an equal amount of stop media consisting of 50% DMEM, 50% heat inactivated FBS (Life Technology, Cat# 10437028), and 10 μL/mL DNAse (Roche, Cat# 10104159001). Cells were centrifuged at 800 rpm for five minutes at room temperature. Cell pellets were resuspended in a differentiation medium containing glucose and plated in ibidi dish (200-300×10⁵ cells/dish) for further culture.

Human induced pluripotent stem cell culture (CSC-C2847SC human iPSC-CMs)

Human iPSC-derived cardiomyocytes (hiPSC-CMs) were recovered and cultured according to the manufacturer’s instructions. SuperCult Human Cardiomyocytes Plating Medium was thawed overnight at 4°C and used within 2 weeks. Ibidi dishes or chamber slides were coated with 500 μL of gelatin-based solution per well and incubated at 37°C for ≥1 hour prior to seeding. For thawing, plating medium was equilibrated to room temperature. Cryopreserved hiPSC-CMs were transferred from liquid nitrogen to dry ice for 10 minutes and then incubated in a 37°C water bath for 4 minutes without submerging the vial cap. The vial was disinfected with 70% ethanol and transferred to a biosafety cabinet. Cells were gently transferred to a 50 mL conical tube. The vial was rinsed with 1 mL plating medium, added dropwise (~90 seconds) with gentle swirling. Subsequently, 8 mL plating medium (or 3.5 mL for smaller aliquots) was added gradually, with the first 1 mL added dropwise (30–60 seconds), followed by the remainder. The suspension was mixed by gentle inversion. Cell viability and concentration were determined using Trypan Blue and a hemocytometer. Cells were adjusted to 4 × 10⁵ cells/mL, and 500 μL was seeded per dish. Dishes were left undisturbed for 15 minutes before incubation at 37°C. After 3 hours, 1.5 mL plating medium was added, and cells were cultured for 48 hours. Non-adherent cells were removed by gentle washing. Cells were then maintained in medium containing 100 ng/mL recombinant human neuregulin (rNRG1; R&D Systems, Cat#396-HB-050/CF) for 4 days, with medium changes every 2 days, prior to experimental use.

Human fetal primary cardiomyocytes isolation and culture

Human fetal hearts were collected from the Health Tissue Bank at Magee Women’s Hospital, approved under IRB protocol (PRO14100621). Cardiomyocytes were isolated using the Cellutron Neonatal Myocyte Isolation kit according to the manufacturer’s instructions. Immediately after isolation, cells were frozen in liquid nitrogen for later use. Frozen cells were thawed and cultured in IMDM (Fisher Scientific, Cat#SH3022801) containing 5% FBS (Life Technology Cat#10437028) and 100 ng/ml rNRG1. Approximately 200,000–300,000 cells were seeded in culture medium onto 35 mm optical plastic dishes (ibidi dish grid-500 ibidi# 81166, Fisher Scientific, Cat#50-305-809) pre-coated with 400 μL fibronectin (50 μg/ml, Fisher Scientific, Cat#PHE0023).

Particle image velocimetry (PIV)

Cultured cardiomyocytes were recorded using a Nikon TiE epifluorescence microscope to examine cell morphology in bright field. Local cell movements were analyzed using ImageJ/Fiji with the PIV plugin (https://imagej.nih.gov/ij/index.html) (Tseng et al., 2012). To calculate peak contraction velocity, velocity vectors in the regions of interest for both non-cycling and cycling cardiomyocytes were extracted from videos. The largest velocity vectors in each contraction cycle (including systole and diastole) were collected as the peak contraction velocity.

Adenoviral transduction of cultured cardiomyocytes

Cyclin-dependent kinase 1 (CDK1) is a master regulator of M-phase entry and is regulated by phosphorylation. CDK1AF is a human CDK1 whose sites of phosphorylation (T14A/Y15F) by Wee1/Myt1 were mutated. The effect of these mutations on CDK1 protein function is loss of inhibition, i.e., constitutive activity (Bicknell et al., 2004). Overexpression of CDK1AF and its obligatory partner, Cyclin B, is a validated experimental approach to increase M-phase entry in cells. CDK1AF has been used to increase cardiomyocyte proliferation (Bicknell et al., 2004; Mohamed et al., 2018).

We validated the effects of adenovirus-mediated transduction of CDK1AF and CCNB1 (Cyclin B1) in NRVMs. Immunofluorescence microscopy revealed that 14.5% of cardiomyocytes co-expressed CDK1AF and cyclin B1 at levels higher than endogenous controls (Fig. S3A–C). While nearly 100% of cardiomyocytes expressed either or both proteins after transduction, only part of those with both CDK1AF and Cyclin B1 above a threshold entered M-phase (Fig. S3B–D). Using separate adenoviruses for each gene, we found that their combination induced M-phase activity in 15% of cardiomyocytes, representing a 21-fold increase over controls (Fig. S3E–F).

To promote M-phase, adenoviral transduction was used to overexpress active CDK1 and Cyclin B1. Freshly isolated NRVMs were cultured overnight at 37 °C in 5% CO₂. The following day, cells were transfected with Adv-CDK1AF (Vendor product name: Ad-Cdc2AF, Vector Biolabs, Cat#1765, MOI: 200 for rat cells; MOI: 100 for mouse cells) and Adv-CCNB1 (Vendor product name: Ad-Cyclin-B1, Vector Biolabs, Cat#1762, MOI: 200 for rat cells; MOI: 100 for mouse cells) for 24 hours (rat cells) or 48 hours (mouse cells), followed by live cell calcium measurements or immunofluorescence studies.

Chemical treatment of cultured cardiomyocytes

Freshly isolated NRVMs were cultured overnight at 37 °C in 5% CO₂. The following day, cells were treated with one of the following compounds: ciliobrevin D (120 μM, Sigma Aldrich, Cat# 250401-10MG) for 60 minutes to inhibit cytoplasmic dynein 1; nocodazole (20 μM, Sigma Aldrich, Cat#SML1665) for 60–90 minutes to induce depolymerization of spindle microtubules: Thapsigargin (6 μM, Abcam, ab120286) for various durations to inhibit SERCA2a; Ro-3306 (24 μM, Fisher Scientific, Cat#NC1193372) for either 15 or 60 minutes to inhibit CDK1; MG132 (40 μM, Fisher Scientific, Cat# 50-311-2916) for 60 minutes to inhibit both the proteasome and calpain. Dimethyl sulfoxide (DMSO) was used as the vehicle control (Ctrl) for all chemical treatments. Following treatment, live-cell calcium imaging or immunofluorescence microscopy was performed.

Immunofluorescence microscopy-based nuclear ploidy quantification

Cultured cardiomyocytes were fixed in 4% paraformaldehyde (PFA) and subsequently immunostained for troponin I (cardiomyocyte marker) and tubulin (microtubule marker), with DNA counterstained using Hoechst 33342. Imaging was performed on a Nikon Ti2 epifluorescence microscope equipped with an ORCA-Fusion BT Digital CMOS camera. Excitation intensity, exposure time, and camera gain were optimized to prevent signal saturation and maintained at constant levels throughout each sample’s acquisition session. For data analysis, fields of view containing at least one M-phase cardiomyocyte (preferably in prometaphase, metaphase, or telophase) were selected. ImageJ/Fiji software was utilized to quantify fluorescence intensity. The integrated fluorescence of all chromosomes in prometaphase and metaphase cardiomyocytes established the tetraploid (4N) reference, whereas the mean integrated fluorescence of the two daughter nuclei in telophase cardiomyocytes provided the diploid (2N) reference. When both stages were present in a single image, telophase nuclei were prioritized as the ploidy reference. The integrated nuclear fluorescence of all mononuclear interphase cardiomyocytes was then measured and normalized against the respective reference nucleus from the same field of view.

Quantification of total number of cardiomyocytes in mouse hearts

To quantify the total number of cardiomyocytes in P2 mouse hearts, live cardiomyocytes were isolated from single P2 mouse hearts using the Cellutron isolation kit, as previously described. Freshly isolated cardiomyocytes were suspended in 2 mL medium. The number of cells per milliliter was counted using a standard hemocytometer method to obtain the total number of cardiomyocytes in the hearts. To quantify the total number of cardiomyocytes in P7 mouse hearts, cardiomyocytes were isolated using the fixation-digestion method (Liu et al., 2021). Hearts were resected, minced into ~1 mm tissue blocks, and fixed in 3.7% formaldehyde at room temperature for 90 minutes. Following fixation, the tissue was washed and digested in enzyme solution containing Collagenase B (3.6 mg/mL, Sigma, Cat# 11088831001) and Collagenase D (4.8 mg/mL, Sigma, Cat# 11088882001) in PBS at 37 °C while shaking at 10 rpm for 24 hours. After 24 hours, the supernatant containing cardiomyocytes was collected, and an equal volume of fresh enzyme solution was added. Digestion steps were repeated until no visible tissue remained. Collected supernatants were briefly centrifuged and resuspended in 2 mL PBS. The number of cardiomyocytes was quantified using a hemocytometer.

Quantification of cardiomyocyte cell cycle activity in mouse hearts with Dync1h1 knockout

The cell cycle activity of cardiomyocytes in mouse hearts with Dync1h1 knockout was studied by quantifying cycling cardiomyocytes labeled with Ki67, BrdU, and H3P using an immunofluorescence microscope. In P7 mouse hearts, cardiomyocytes were labeled with BrdU and H3P antibodies. BrdU was administered via intraperitoneal injection at a dose of 61.42 μg/g body weight on P5 and P6, with one injection per day.

Heart sections for Immunofluorescence microscopy

Fresh resected mouse hearts were washed in PBS containing 25 μM KCl and fixed immediately in 3.7% formaldehyde at room temperature for 8 hours. The fixed hearts were washed in PBS and immersed in 30% sucrose (ThermoFisher, Cat#A15583.36) at 4 °C for 24 hours. After being embedded in optimum cutting temperature (OCT) compound, the frozen block was trimmed, and the exposed heart tissue was cut into 15 μm thick sections with a Leica CM1950 cryostat and adhered to glass slides (Color Frost, Fisher). The sectioned tissue slides were stored in −20°C. Before antibody staining, tissue slides were thawed at 57 °C for 5 minutes. Sections were then fixed with 4% paraformaldehyde (PFA) for 5 minutes and rinsed with 1X PBS. The tissue was blocked with 5% normal donkey serum (NDS) for 15 minutes and then treated with 0.3M glycine for 20 minutes. The tissue was then treated with a permeabilization solution containing 5% NDS, 0.1% Triton X-100, and 0.05% Tween20 for 20 minutes. Primary antibody was applied at room temperature for 2 hours and then washed with PBS. Secondary antibody was applied at room temperature for 30 minutes. Hoechst was applied at 1:1000, and slides were mounted using Fluromount.

Cardiomyocyte apoptosis in heart sections

Apoptosis in cardiomyocytes from mouse heart sections with Dync1h1 knockout was assessed using the ApopTag^® Red In Situ Apoptosis Detection Kit (Millipore, Cat# S7165), following the manufacturer’s instructions. Tissue slides were thawed at 57 °C for 5 minutes prior to apoptosis detection and antibody staining. Sections were fixed with 4% paraformaldehyde in PBS and permeabilized using 0.25% Triton X-100. Samples were then incubated with 100 μL of TdT Reaction Buffer (Component A) for 10 minutes at 37 °C. After removal of the buffer, 50 μL of the prepared TdT reaction mixture was added and incubated for 60 minutes at 37 °C in a humidified chamber. Sections were washed twice with 3% BSA in PBS for 5 minutes each. For the Click-iT Plus reaction, a 10X Click-iT Plus TUNEL Reaction Buffer Additive was freshly prepared by diluting the 100X stock 1:10 in deionized water and used on the same day. The Click-iT Plus TUNEL reaction cocktail was prepared according to the manufacturer’s protocol, mixed thoroughly by vortexing, and applied (50 μL per section) within 15 minutes of preparation. Sections were incubated with the cocktail for 30 minutes at 37°C, protected from light, followed by a single 5-minute wash with 3% BSA in PBS. Heart sections were blocked with donkey serum buffer and incubated overnight at 4 °C with primary antibodies against Troponin I (1:200). Following two washes with 3% BSA in PBS, samples were incubated with secondary antibodies at room temperature for 60 minutes. DNA was stained using Hoechst 33342 solution (1:1000), followed by PBS washes, and sections were mounted with cover glass using mounting medium.

Quantification and statistical analysis

Numeric results are presented as mean ± SEM. Fluorescence data obtained from live cell calcium measurement experiments was analyzed using GraphPad Prism 10 software (GraphPad Software). The differences between groups were tested for statistical significance with the unpaired two-tailed Student’s t test, One-sample t-test, or ANOVA with the Bonferroni post-test. Data distribution was assumed to be normal, but this was not formally tested. Differences at P = 0.05 and lower were considered statistically significant.

Supplementary Material

Figure S1

Figure S1. Validation experiments confirm no unpredicted impact of indicators or drugs on contractility, observed M-phase CaTs, or spindle architecture. (A-C) Expression of GCaMP8 in cardiomyocytes does not alter their contractility. Cardiomyocytes were isolated from newborn (P1) mice expressing GCaMP8 (GCaMP8⁺) and those not expressing GCaMP8 (GCaMP8⁻). (A) Formula used to calculate sarcomere shortening fraction. (B) Live cardiomyocytes were labeled with the fluorescent actin dye SiR-Actin to indicate thin filaments (purple arrows). The dark areas, indicated by black arrows, represent M-lines. Sarcomere contractions were recorded under a fluorescence microscope (frame rate: 50 frames/second). Sarcomere length was measured at the end of relaxation (diastole) and contraction (systole). To ensure accuracy, the lengths of two serially connected sarcomeres were measured and averaged. (C) The sarcomere shortening fraction in cardiomyocytes expressing GCaMP8 (GCaMP8⁺: n=8 cells from three hearts) was comparable to that in the control group (GCaMP8⁻: n=3 cells from one heart). (D-F) Dual-channel imaging shows reduced CaTs in M-phase cardiomyocytes. Freshly isolated NRVMs (P1) were cultured overnight. (D) The same cells were loaded with Fluo4 AM and Rhod2 AM, which have distinct excitation/emission spectra (Fluo4: Ex/Em: 494/506 nm and Rhod2: Ex/Em: 552/581 nm). Fluorescence from both dyes was simultaneously recorded through separate channels using a confocal microscope equipped with resonant scanner. A decrease in peak CaTs within the same Ca²⁺ release cycle was detected in M-phase NRVMs in both channels. Red arrows indicate cardiomyocytes in M-phase. (E) Fluo4 and (F) Rhod2 confirm a consistent reduction in CaTs during M-phase. Calcium transient dynamics are visualized in Video 5. Frame rate: 41 frames/second. (G-H) Thapsigargin treatment does not disrupt spindle microtubules in M-phase cardiomyocytes. Freshly isolated NRVMs (P1) were cultured overnight and treated with thapsigargin (6 μM, 120 minutes; n = 3 cultures). Spindle microtubules in M-phase were assessed using immunofluorescence microscopy. DMSO-treated cells (120 minutes; n = 10 cultures) served as the negative control, while nocodazole-treated cells (20 μM, 60–90 minutes; n = 7 cultures) served as the positive control. No M-phase cardiomyocytes in the thapsigargin or DMSO groups showed disrupted spindle microtubule architecture. In contrast, all M-phase cardiomyocytes in the nocodazole group displayed disrupted spindle microtubules. Scale bars: (D) 40 μm, (G) 10 μm. Statistical analyses: (C) Student’s t-test (unpaired, Two-tailed), (H) One-way ANOVA followed by Bonferroni’s correction for multiple comparisons. (C, H) Mean ± SEM is indicated by horizontal lines.

Figure S2

Figure S2. Criteria for identifying sub-phases of M-phase and striated patterns of sarcoplasmic reticulum proteins in neonatal rat cardiomyocytes. (A) Interphase and sub-phases of M-phase in live NRVMs were distinguished based on cell morphology revealed by brightfield microscopy. The sub-phases of M-phase in fixed NRVMs were identified by sarcomere structure (Troponin I), organization of condensed chromosomes (Hoechst), and H3P revealed with immunofluorescence microscopy. Additional NRVMs exhibiting typical M-phase morphology in live cells, confirmed by post hoc immunofluorescence, are provided in the Source Data. Text in the table describes features unique to each subphase of M phase and those shared across M-phase subphases. (B) Striated patterns of RyR2 and SERCA2a are defined by the presence of RyR2/SERCA2a stripes on both sides of individual sarcomeres. The red lines indicate RyR2/SERCA2a stripes, while the white arrows point to the corresponding sarcomeres. Scale bars: (A) 20 μm, (B) 2 μm.

Figure S3

Figure S3. Forced expression of CDK1AF and Cyclin B1 induces M-phase entry in neonatal cardiomyocytes. Fresh isolated NRVMs (P1) were cultured overnight, then transfected with adenovirus (Adv-CDK1AF+Adv-Cyclin B1) and incubated at 37°C for 24 hours. (A-C) Adenovirus-mediated gene transduction increased CDK1 and Cyclin B1 levels in the NRVMs. Both endogenous and exogenous CDK1 in the NRVMs were labeled with the same CDK1 antibody, and both endogenous and exogenous Cyclin B1 were labeled with the same Cyclin B1 antibody. During image acquisition, excitation light intensity, detector sensitivity (gain), and acquisition time were kept constant for each channel throughout the experiment. (A) CDK1 and Cyclin B1 were detected in all cardiomyocytes using immunofluorescence microscopy. Representative images of the control group (Ctrl) and treatment group (CDK1AF; Cyclin B1) are presented using the same upper and lower threshold settings for each channel in the look-up table (LUT). (B-C) Co-expression of CDK1 and Cyclin B1, indicated by fluorescence intensity, in single cardiomyocytes. Each symbol represents one cardiomyocyte. (B) The highest expression levels of CDK1 and Cyclin B1 in cardiomyocytes from the Ctrl group (n = 959 cells) are indicated by the vertical (green) and horizontal (red) dashed lines, respectively. (C) The same lines were applied to the treatment group (n = 1262 cells), where the shaded area indicates the cardiomyocytes expressing higher levels of both CDK1 and Cyclin B1 than the Ctrl group. In the treatment group, 14.5% of the cardiomyocytes exceeded the CDK1 levels in the Ctrl group (right side of the green vertical line) and 27.5% exceeded the Cyclin B1 levels in the Ctrl group (above the red horizontal line). Symbols beneath the color shadow represent cardiomyocytes with higher levels of both CDK1 and Cyclin B1 than those in the control group. N=100 M-phase cardiomyocytes were quantified by identifying condensed chromosomes in the 915 cells within the shaded area in (C)—that displayed higher fluorescence levels of CDK1 and cyclin B1, as visually observed, which is 7.92% of all the 1262 cells. (D) M-phase cardiomyocytes identified by the presence of condensed chromosomes do not always correspond precisely with cells co-expressing CDK1 and Cyclin B1. White arrow: A cardiomyocyte displaying condensed chromosomes and co-expressing both CDK1 and Cyclin B1. Green arrow: A cardiomyocyte with condensed chromosomes expressing only CDK1. Red arrow: A cardiomyocyte co-expressing CDK1 and Cyclin B1 without exhibiting condensed chromosomes. (E-F) Increasing CDK1 activity elevated the proportion of M-phase (H3P⁺) cardiomyocytes. (F) Each symbol represents an independent cell culture (n = 3 cell cultures per group). Scale bar: (A) 100 μm, (D) 10 μm, (E) 50 μm. Statistical analysis: (F) Student’s t-test (unpaired, Two-tailed). Mean ± SEM is indicated by horizontal lines.

Figure S4

Figure S4. Inactivation of Dync1h1 gene in cardiomyocytes during cardiac development neither induces apoptosis nor impacts heart function. The Dync1h1 gene in cardiomyocytes of a genetic mouse model (αMHC-Cre;Dync1h1^flox) was constitutively inactivated during cardiac development. (A) Immunofluorescence microscopy revealed decreased fluorescence of dynein cytoplasmic 1 heavy chain 1 (the protein encoded by the Dync1h1 gene) in neonatal mouse hearts (P1-P2) following Dync1h1 CKO. Ctrl: αMHC-Cre⁻;Dync1h1^F/W, Dync1h1 CKO: αMHC-Cre⁺;Dync1h1^F/F. (B) Western blot analysis of heart tissue from neonatal (P7) and adult (4–6 months) mice confirmed reduced Dync1h1 protein levels (Source Data is available). Ctrl: αMHC-Cre⁺;Dync1h1^F/W, Dync1h1 CKO: αMHC-Cre⁺;Dync1h1^F/F. Vinculin was used as the loading control. (C) Dync1h1 protein levels (grey values normalized to Vinculin) were further normalized to the mean value of Dync1h1 protein in the Ctrl group, P7 mice (Ctrl: n = 3 hearts, Dync1h1 CKO: n = 2 hearts), adult mice (Ctrl: n = 3 hearts, Dync1h1 CKO: n = 3 hearts). (D) TUNEL assay demonstrates that inactivation of the Dync1h1 gene does not induce cardiomyocyte apoptosis in newborn (P1-P2) mouse hearts. White arrows indicate apoptotic cells identified by TUNEL. Ctrl: αMHC-Cre⁻;Dync1h1^F/F, Dync1h1 CKO: αMHC-Cre⁺;Dync1h1^F/F. Heart section thickness: 20 μm. (E-F) Inactivation of the Dync1h1 gene in cardiomyocytes during cardiac development did not alter (E) heart-body weight ratio or (F) cardiac function in adult (4–6 months) mice (Ctrl: n = 7 mice, Dync1h1 CKO: n = 5 mice). (G) Adult mouse cardiomyocytes were isolated using the fixation-digestion method. No cardiomyocytes in the cell cycle, as indicated by Ki67 immunofluorescence, were found in either Ctrl or Dync1h1 CKO groups. Ctrl: αMHC-Cre⁺;Dync1h1^F/W, Dync1h1 CKO: αMHC-Cre⁺;Dync1h1^F/F. Scale bar: (A) 40 μm, (D) 500 μm. Statistical analysis was performed using Student’s t-test (unpaired, Two-tailed) (C, E, F). Mean ± SEM is indicated by horizontal lines.

Figure S5

Figure S5. Human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) undergo changes in the sarcoplasmic reticulum and exhibit reduced CaTs during M-phase. The structure of SR in human iPSC-CMs, identified using SERCA2a, was examined via immunofluorescence confocal microscopy. Human iPSC-CMs were recovered according to the vendor’s protocol and cultured in the vendor’s medium containing 100 ng/ml rNRG1 for 48 hours. (A-B) Remodeling of SR in human iPSC-CMs during M-phase. (A) Striated SERCA2a patterns (indicated by white arrows) were observed in interphase cardiomyocytes but were rarely detected in M-phase. M-phase human iPSC-CMs were identified by the presence of mitotic spindles (tubulin, red). The inset highlights SERCA2a striations in the region marked by the white arrow. A region of low SERCA2a signal was clearly observed at the spindle poles (indicated by red arrows) in M-phase cells. (B) The number of cardiomyocytes exhibiting detectable SERCA2a striations decreased during M-phase compared with interphase cardiomyocytes (n = 3 cultures). (C, D) Human iPSC-derived cardiomyocytes exhibited decreased CaTs during M phase, whereas inhibition of cytoplasmic dynein heavy chain 1 with ciliobrevin D increased CaTs in M-phase cells, as shown by CaT tracing curves. F₀ and F_max indicate baseline and peak Rhod2 (Ca²⁺) fluorescence, respectively. (D) Quantification of peak CaTs in M-phase iPSC-derived cardiomyocytes. Each symbol represents one cardiomyocyte. Ctrl: n = 11 cells, Ciliobrevin D: n = 13 cells. Scale bars: (A) 20 μm, (C) 40 μm. Inset scale bar: (A) 2 μm. Statistical significance: (B) Student’s t-test (unpaired, Two-tailed), (D) One-way ANOVA followed by Bonferroni’s correction for multiple comparisons. (B, D) Mean ± SEM is indicated by horizontal lines.

Data availability

This study includes no data deposited in external repositories. All materials are available from the corresponding author upon reasonable request and without undue delay.