Adducin Regulates Sarcomere Disassembly During Cardiomyocyte Mitosis

Abstract

BACKGROUND:

Recent interest in understanding cardiomyocyte cell-cycle has been driven by potential therapeutic applications in cardiomyopathy. However, despite recent advances, cardiomyocyte mitosis remains a poorly understood process. For example, it is unclear how sarcomeres are disassembled during mitosis to allow the abscission of daughter cardiomyocytes.

METHODS:

Here we use a proteomics screen to identify Adducin, an actin capping protein previously not studied in cardiomyocytes, as a regulator of sarcomere disassembly. We generated a number of adeno-associated viruses (AAVs) and cardiomyocyte-specific genetic gain-of-function models to examine the role of adducin in neonatal and adult cardiomyocytes in vitro and in vivo.

RESULTS:

We identify adducin as a regulator of sarcomere disassembly during mammalian cardiomyocyte mitosis. α/γ-Adducins are selectively expressed in neonatal mitotic cardiomyocytes, and their levels decline precipitously thereafter. Cardiomyocyte-specific overexpression of various splice isoforms and phosphor-isoforms of α-Adducin in identified Thr445/Thr480 phosphorylation of a short isoform of α-Adducin as a potent inducer of neonatal cardiomyocyte sarcomere disassembly. Concomitant overexpression of this α-Adducin variant along with γ-Adducin resulted in stabilization of the adducin complex and persistent sarcomere disassembly in adult mice, which is mediated by interaction with α-Actinin.

CONCLUSION:

These results highlight an important mechanism for coordinating cytoskeletal morphological changes during cardiomyocyte mitosis.

Graphical Abstract

Introduction

Cardiomyocyte loss is a leading cause of heart failure, which is a devastating progressive disease affecting over 30 million patients worldwide¹. Although limited cardiomyocyte renewal occurs in the adult mammalian heart, it is insufficient for restoration of contractile function following cardiomyocyte loss, which results in cardiomyopathy^2-7. Therefore, stimulating cardiomyocyte regeneration is a major goal of heart repair in cardiomyopathy. Unlike some lower vertebrates which possess a life-long cardiac regeneration capacity^8-14, adult mammalian cardiomyocytes permanently withdraw from the cell cycle shortly after birth and have a very limited regenerative potential. Studies in neonatal mouse hearts showed that apical resection or myocardial infarction (MI) of postnatal day 1 (P1) mice leads to a robust regenerative response, however, this capacity is lost by P7 which coinciding with post-natal cardiomyocyte cell cycle arrest^15,16. Although several regulators of the cardiomyocyte cell cycle have been identified^17-23, it is not clear whether there are cytoskeletal-specific factors that regulate mitosis in cardiomyocytes.

The dense myofibrillar composition of cardiomyocytes poses a special challenge during mitosis, which is a consideration that does not exist in other non-striated cells. In an adult cardiomyocyte, sarcomeres occupy over 60% of cardiomyocyte cytoplasm²⁴, and anchor to the plasma membrane to induce deformation of the cell during contraction. An important and interesting feature of regenerating hearts is the disassembly of cardiomyocyte sarcomeres^15,25,26, and their peripheral marginalization during cardiomyocyte mitosis (Figure 1A)²⁷. This phenomenon has long been observed in dividing cardiomyocytes and has proven to be a reliable indicator of cardiomyocyte mitosis in the early postnatal heart^15,17,22,28. However, to date, the mechanisms that regulate sarcomere disassembly are not understood. Moreover, it is unclear whether the regulation of sarcomere disassembly plays a role in the loss of the regenerative capacity of the neonatal heart with advancing postnatal age. In the current study, we set out to identify regulators of cardiomyocyte sarcomere disassembly during mitosis. We found that the cytoskeletal regulatory protein Adducin is a critical regulator of sarcomere disassembly.

Figure 1. Identification of Adducin as a key protein associated with sarcomere disassembly and cardiomyocyte proliferation.

A. Schematic representation of sarcomere disassembly morphology. B. Co-immunoprecipitation coupled with mass spectrometry workflow, utilizing Tnnt2 pulldown to identify associated proteins from total heart extracts 3 days post-MI at P1 (P1MI) and P7 (P7MI). C. Table summarizing key proteins associated with Tnnt2, as identified from Co-IP/MS. Notations: *Peptide Spectrum Matches (PSM) indicates the number of spectra assigned to peptides that contributed to the inference of the protein; **MIC Sin represents the normalized spectral index statistic for each protein, calculated from the intensity of fragment ions in each spectrum assigned to a protein; ***Ratio between groups derived from MIC Sin values. D. Co-IP of Add1 from P1MI and P7MI, probed for Tnnt2 and α-Actinin. E. (Left) Representative WB; (Right) Quantitative comparison of Add1 and Add3 protein expression at three days following sham surgery or MI at P1 and P7. The pH3-S10 is included as a biomarker to indicate the occurrence of MI. F. (Left) Representative WB; (Right) Quantification of endogenous expression of Add1 and Add3 at different postnatal ages. Gapdh serves as a loading control in E and F. G. Percentage of cardiomyocytes expressing Adducins in neonatal hearts. H. Immunostaining for selected cytoskeletal proteins (Add1, Add3, Spectrin, Filamin) in (Left) proliferative and (Right) non-proliferative neonatal cardiomyocytes. Arrows indicate proliferative cardiomyocytes. Arrowheads in Filamin staining refers to the nucleus location. Scale bar: 10 μm (H). Data are presented as mean±sd. Statistical analyses: Two-way ANOVA with Tukey's post-hoc test (E), Kruskal-Wallis with uncorrected Dunn's test (F); * p < 0.05, ** p < 0.01.

Methods

Detailed descriptions of materials and methods and information on statistical tests are provided in the article or the Supplemental Material. All data and methods used in this analysis will be made available to researchers upon request. The use of animals in this study was approved by the Institutional Animal Care and Use Committee of the University of Texas Southwestern Medical Center. Experimental analysis was performed in a double-blind fashion. Statistical analysis and plotting were performed using GraphPad Prism (v.10.1.0) (GraphPad Software Inc, San Diego, CA). Data were tested for normality before parametric statistics were applied using the Kolmogorov-Smirnov test. Data were analyzed by the two-tailed Mann-Whitney U test or the two-tailed Student unpaired t-test to compare means when there were 2 experimental groups, by Kruskal-Wallis with uncorrected Dunn's test or one-way ANOVA followed by Tukey post-hoc test to compare means among ≥3 groups, or by two-way ANOVA to compare means when there were ≥2 independent variables. Differences between the groups were considered significant at P<0.05.

Results

Adducin is highly expressed in neonatal regeneration, closely associated with disassembled sarcomeres.

Immunostaining of cardiac troponin T (Tnnt2), a crucial sarcomere protein, in our previous studies, revealed a pivotal process related to the neonatal cardiomyocyte proliferation: sarcomere disassembly and relocation to the sub-membranous area in mitotic cardiomyocytes¹⁵. This presence of Tnnt2 in disassembled sarcomeres can thus be used to further explore novel disassembly-regulated proteins. Utilizing mass spectrometry (MS) following Co-Immunoprecipitation (Co-IP) with a Tnnt2 antibody, we examined Tnnt2-associated proteomes at two stages: three days after MI in day-one (P1MI) and day-seven (P7MI) postnatal hearts, correlating to high and low cardiomyocyte mitosis and disassembly activities, respectively (Figure 1B). Our findings showed a significant cytoskeletal protein presence (Table-S1), notably highlighting the differential association of α-Adducin (Add1) and γ-Adducin (Add3) with Tnnt2 during the regenerative window (Figure 1C). Although detected at lower levels, both Adducins were identified exclusively in the P1MI samples, not unlike α-actinin. Spectrin, a well-known Adducin interacting protein, also predominantly interacted with Tnnt2 in the P1MI but less expressed in P7MI (Figure 1C). Reverse Co-IP using Add1 antibody as a bait in P1MI and P7MI hearts confirmed the interaction of Add1 with Tnnt2 and also its strong association with α-Actinin in P1MI hearts (Figure 1D). These results suggest the relevance of Adducin presence and cardiomyocyte proliferation.

We further examined the relationship between Adducin expression and cardiomyocyte proliferation using Western blot (WB). Following P1MI, a period of intense proliferation marked by high levels of the mitotic marker phosphor-histone H3-Ser10 (pH3), Add1 expression significantly increased, compared to P1Sham. In contrast, P7MI hearts displayed a marginal, non-significant increase in Adducin expression, compared to P7Sham. Add3 levels remained unchanged between MI and Sham groups (Figure 1E). Additionally, both Add1 and Add3 showed decreased expression with increasing postnatal age (Figure 1F). Therefore, we assessed whether the Add1-Tnnt2 interaction aligns more with robust cardiomyocyte proliferation than maturation. We conducted Co-IP using Adducin or Tnnt2 antibodies as baits on P1MI or sham surgery neonatal hearts (Figure S1A,B). The results confirmed a stronger Add1-Tnnt2 interaction in higher proliferation status P1MI (Figure S1C,D). We noted a mild but noticeable Add1-Tnnt2 interaction in P1Sham, attributable to the ongoing cardiomyocyte proliferation at this stage ((Figure S1C,D). Conversely, to confirm Adducin capability in stimulating cardiomyocyte mitosis, we knocked down Add1 and Add3 in neonatal rat ventricle myocytes (NRVM) via siRNA and monitored proliferation through pH3 staining (Figure S2A). WB confirmed the successful knockdown, particularly with tandem siRNA targeting both Add1 and Add3 (Figure S2B). siRNA-mediated Adducin knockdown led to reduced mitotic rates (Figure S2C), underlining their importance in neonatal cardiomyocyte proliferation.

Immunohistochemistry revealed Adducin expression in about 30% of neonatal cardiomyocytes (Figure 1G), with Add1, Add3, and Spectrin specifically associated with disassembled sarcomeres in proliferating cells, localizing in the membranous and submembranous areas (Figure 1H). Conversely, α-Filamin displayed a nuclear staining pattern in non-proliferating cells and a diffuse cytoplasmic pattern in proliferating cardiomyocytes. siRNA-mediated Add1 and Add3 loss led to a decline in sarcomere disassembly rates (Figure S2D). These observations reinforce the role of Adducin in sarcomere disassembly during cardiomyocyte proliferation in the neonatal heart.

Overexpression of Add1 isoform 2 results in transient disassembly of sarcomeres.

Alternative splicing plays an important role in heart development and cardiomyopathies, significantly impacting post-transcriptional regulation^29-33. While previous studies focused on the full-length Add1 isoform 1 (Add1_i1), there is another major isoform, Add1 isoform 2 (Add1_i2), differentiated by the exclusion of exon 15 (Figure 2A). Add1_i2, the shorter isoform lacking about 100 amino acids at the C-terminus, also declines in expression with postnatal age in mouse hearts, as observed in Add1_i1 (Figure 2B).

Figure 2. Overexpression of Add1 isoform 2 disassembles cardiomyocyte sarcomeres.

A. (Top) Schematic detailing the exon-intron structure of the Add1 gene. The red rectangle highlights exon15, the alternative splicing target. (Bottom) Depicts alternative splicing of exon15 in mRNA: inclusion results in Add1_i2 with 636 amino acids, while exclusion produces Add1_i1 with 732 amino acids. The corresponding protein schematics are displayed alongside. B. (Left) Representative WB; (Right) Quantitative comparison of Add1_i2 endogenous expression across different ages, with Gapdh as an internal control. C. Representative images of NRVMs transduced with AAV6 viral particles for Add1_i1, Add1_i2, and GFP control, at day 3. Cells were immunostained with α-Actinin (Actn2, red) to visualize sarcomere structure. Green fluorescence indicates Add1_i1, Add1_i2, or GFP expression. High-magnification views of selected areas are shown to the right, depicting varying states of sarcomere assembly: (i) and (iii) partially disassembled sarcomeres; (ii) and (iv) fully disassembled sarcomeres; (v) assembled sarcomeres. Dotted lines mark cardiomyocyte borders. D. Percentage of sarcomere disassembly in response to Adducin overexpression in C. E. Generation strategy for Add1_i2 transgenic mice. F. H&E staining of WT and Add1_i2 transgenic hearts. G. HW/BW comparison in control and Add1_i2 transgenic mice. H. (Top) Images showing Add1_i2 expression patterns in transgenic mice at P14 and P28. (Bottom) Insets show enlarged views of individual cardiomyocytes. I. Proportion of cardiomyocytes expressing Add1_i2 in transgenic mice. J. (Top) Images representing sarcomere patterns in Add1_i2 transgenic mice at P14 and P28. (Bottom) Insets show enlarged views of individual cardiomyocytes. Dotted lines indicate the outer and inner borders of sarcomeres. K. Comparisons of sarcomere disassembly scored with age-matched WT littermates indicate significant sarcomere thinning and central clearance in Add1_i2 transgenic mice at P14. Scale bar 10 μm (C, H, J); 1 mm (F). Data are presented as mean±sd. Statistical analyses: Kruskal-Wallis with uncorrected Dunn's test (B), Mann-Whitney U test (two-tailed) (I),two-way ANOVA with Tukey's post-hoc test (D, K; *p < 0.05, **p < 0.01, ***p < 0.001 and ****p<0.0001.

To explore the impact of Add1 on sarcomere disassembly, we infected NRVMs with either Add1_i1 or Add1_i2 using Adeno-associated virus (AAV) 6, with AAV6- green fluorescent protein (GFP) serving as control. We categorized sarcomere disassembly into three patterns: assembled, partially disassembled, and fully disassembled. Partially disassembled sarcomeres, identifiable by α-actinin staining in red, contain some visible cytoplasmic sarcomeres (Figure 2Ci,iii), while fully disassembled sarcomeres lack them completely (Figure 2Cii,iv). Overexpression of Add1_i1 resulted in 10.8% of cells with fully disassembled sarcomeres, while Add1_i2 led to 49.6% (Figure 2D). In contrast, GFP-infected NRVMs showed no sarcomere disassembly (Figure 2Cv). These results suggest that Add1_i2 overexpression led to a higher rate of complete sarcomere disassembly compared to Add1_i1.

Furthermore, we generated a cardiac-specific transgenic (TG) mouse model expressing Add1_i2 under an α-Myosin heavy chain (α-MHC) promoter (Figure 2E). The Add1_i2 TG mice exhibited no significant differences in cardiac size, morphology as indicated by Hematoxylin and Eosin (H&E) staining (Figure 2F), or heart weight to body weight ratio (HW/BW) (Figure 2G). Postnatal analysis of these mice revealed a notable decrease in Add1-positive cardiomyocytes at P28 compared to P14 hearts (Figure 2H,I). Importantly, at P14, Add1_i2 TG displayed a significant increase in sarcomere disassembly, evidenced by reduced assembly scores (Figure 2J,K) and increased pH3-positive (pH3⁺) cardiomyocytes (Figure S3A), without any change in cardiomyocyte size (Figure S3B). However, by P28, alongside a decrease in Add1_i2 expression, these hearts showed no significant change in assembly scores (Figure 2J,K) or left ventricular (LV) systolic function, assessed by echocardiography (Figure S3C), compared to wild-type (WT) mice. These results suggest that cardiomyocyte-specific overexpression of Add1_i2 results in a transient increase in sarcomere disassembly, which does not continue into adulthood.

Add1 T445/T480 phosphorylation extends neonatal cardiomyocyte sarcomere disassembly.

We explored the role of Adducin phosphorylation in cardiomyocyte sarcomere disassembly by examining its phospho-isoforms expression in neonatal mouse hearts (Figure 3A). Studies indicate that phosphorylation at Add1_Threonine (T) 445 and 480 (T445/T480) sites enhances Spectrin recruitment to F-actin³⁴. We found Add1_pT445 at the cellular cortex in myocytes with disassembled sarcomeres and in the nuclei of non-proliferating cardiomyocytes (Figure 3A). Additionally, phosphorylation by cyclin-dependent kinase 1 (CDK1) at Serine (S) 12 and Ser355 associates Add1 with mitotic spindles³⁵, and we noted Add1_pS355 localization in the cytoplasm of mitotic cardiomyocytes (Figure 3A). Add1 also has protein kinases A (PKA) and C phosphorylation sites at Ser716 and Ser726 (S714 and S724 in mice), Ser408, Ser436, and Ser481, which inhibit Spectrin-F-actin binding^36,37. However, our study found no phosphorylation at S724 or S481 in neonatal cardiomyocytes (Figure 3A).

Figure 3. Role of Add1_i1 phosphorylation on sarcomere disassembly.

A. Endogenous phospho-Adducin patterns in neonatal mouse hearts: phospho-sites (T445/T480, S12/S355, S714/S724, S408/S436/S481) co-stained with sarcomeric proteins Tnnt2 or α-Actinin. Arrows and arrowheads indicate colocalization with disassembled and assembled sarcomeres, respectively. B. NRVM showing Add1_pT445 expression, stained with α-Actinin (red) and pT445 (green), noting its translocation from the nucleus to the cytoplasm during mitosis. C. (Left) Representative WB; (Right) Quantitative comparison of Add1_pT445 endogenous expression in relative to Gapdh. Tnnt2 is used as an age indicator. D. Representative images of NRVM transduced with AAV6 viral particles for phospho- or non-phospho mimic T445/T480. (Top left) Adducin mutants are expressed with 3 tandem FLAG epitopes at the N-terminus. Cells were immunostained with α-Actinin (red) to visualize sarcomere structure. Adducin mutants are detected by FLAG antibody (green). High magnification images of representative region (insets) were shown on the right. Dotted lines were drawn around the cell border. E. Quantitative analysis of disassembled sarcomeres induced by Adducin overexpression in D. F. Generation strategy for cardiac-specific Add1_i1 phospho transgenic mice. G. H&E staining of control and cardiac-specific Add1_i1 phospho transgenic. H. HW/BW in control and phospho-transgenic mice. I. Images showing phospho Add1_i2 expression patterns in transgenic mice at P14 and P28, with (Bottom) high magnification views of selective phospho-Adducin mimic overexpression cardiomyocytes. J. Proportion of cardiomyocytes expressing pAdd1 from I. K. Sarcomere patterns in phospho Add1_i1 transgenic mice at P14 and P28, with insets showing high magnification images. Dotted lines are drawn around the outer and inner edges of cardiomyocytes. L. Quantitative analysis of sarcomere disassembly. Scale bar 10 μm (A, B, D, I, K); 1 mm (G). Data are presented as mean±sd. Statistical analyses: Kruskal-Wallis with uncorrected Dunn's test (C) or Mann-Whitney U test (two-tailed) (J) or two-way ANOVA with Tukey's post-hoc test (E, L); *p < 0.05, **p < 0.01, ***p < 0.001 and ****p<0.0001.

Furthermore, we correlated phospho-Adducin expression with cardiomyocyte mitosis in vitro (Figure 3B) and in vivo (Figure S4A) using immunofluorescent staining. Add1_pT445 transitions from nuclear during prophase to cytoplasmic and mitotic chromosome-associated in metaphase with sarcomere disassembly initiation, and remains cytoplasmic through telophase. WB analysis showed that Add1_pT445 expression decreases with age, paralleling Add1 (Figure 3C). These results suggest that phosphorylation of Add1 at T445 and T480 might play a significant role in sarcomere disassembly during cardiomyocyte mitosis.

To determine if Add1 phosphorylation at T445/T480 regulates cardiomyocyte sarcomere disassembly, we generated Add1 mutants mimicking phosphorylated (phospho-mimic) and non-phosphorylated (phospho-silent) states. We produced these variants using glutamic acid (Glu) for the phospho-mimic and alanine (Ala) for the phospho-silent forms with recombinant AAV6. We analyzed sarcomere morphology in NRVMs by staining with α-Actinin antibody and identifying Add1-expressed cells as FLAG-positive. Our results indicated that nearly 60% of cells overexpressing the phospho-mimic Add1(T445E/T480E) showed complete sarcomere disassembly, while those with the phospho-silent form Add1(T445A/T480A) mostly exhibited partial disassembly (Figure 3D,E). Mutants for other phosphorylation sites (S12/S355 and S714/S724) were also tested but did not show a significant effect on complete disassembly (Figure S4B,C). These results led us to focus on the T445/T480 sites for their prominent role in sarcomere disassembly.

To determine if forced expression of phospho-mimic Add1(T445E/T480E) induces sarcomere disassembly in cardiomyocytes in vivo, we developed an inducible cardiac-specific TG by breeding pTRE-Add1(T445E/T480E) with αMHC-tTA mice. This line expresses phospho-mimic Add1(T445E/T480E) in cardiomyocytes in the absence of doxycycline (Tet-off system) (Figure 3F). These mice showed no significant changes in heart morphology (Figure 3G) or HW/BW (Figure 3H) at P28. However, at P14, cardiomyocytes expressing the phospho-mimic Add1 displayed Add1 cytoplasmic localization and sarcomere disassembly, as evidenced by gaps in Tnnt2 and nucleus staining (Figure 3I). This disassembly was not evident at P28, with decreased Add1(T445E/T480E) (Figure 3I,J) and cardiomyocyte morphology returning to normal (Figure 3K,L). In addition, lack of significant differences in proliferation rate (Figure S5A) or cardiomyocyte size (Figure S5B) support the notion that phospho-mimic Add1 overexpression temporarily enhances sarcomere disassembly during early development, but this effect is not sustained to adulthood, not unlike Add1_i2 TG line.

Concomitant expression of Add1 and Add3 is required for protein stabilization.

Previous research focused on Add1, due to its role in dimer formation with other Adducin isoforms. Notably, Add1 knockout models indicated a reduction in Add3 expression. We observed that sarcomere disassembly in Add1 TG lines was impaired after P14, coinciding with diminished endogenous Add3 expression by this age. This result suggests that the reduced disassembly capacity of Add1 TG lines beyond P14 might be due to the loss of Add3. To explore the interplay between Add1 and Add3, we conducted in vitro transfections using Add1_i1 and i2, both with and without Add3. To avoid cross-reactivity between Adducin isoforms, Add1 variants were tagged with FLAG while Add3 were tagged with TY1. Cells were treated with 80 μM cycloheximide (CHX) 24 hours post-transfection to inhibit protein synthesis and collected at various intervals thereafter. WB analysis showed that Add1 levels decreased over time in the absence of Add3. Conversely, in the presence of Add3, the levels of Add_i1 and i2 were maintained and ended up higher than those of Add1 alone at later time points (Figure 4A). These results suggest that the limited sarcomere disassembly observed in single Add1 TG lines beyond P14 might be related to the absence of endogenous Add3 at later developmental stages.

Figure 4. Co-expression of non-phosphorylated Add1_i2 with Add3 induces partially sarcomere disassembly.

A. HEK293T were transfected with 3xFLAG-Add1 variants, either with or without 3xTY1-Add3, and treated with cycloheximide (CHX) to halt protein translation; cell collection at 0, 24, 48 hours post-treatment. (Left) Representative WB; (Right) Assessment of relative protein levels of Add1_i1 and Add1_i2 (normalized FLAG to Gapdh), in the absence or presence of Add3. B. Generation strategy for cardiac-specific Add1_i2/Add3 dTG mice. Both expression constructs were driven by an α-MHC promoter. C. qPCR analysis of Add1_i2 mRNA in 2-month-old low expression dTG(L) hearts, normalized to WT. D. (Left) H&E staining and (Right) HW/BW comparison in 2-month-old dTG(L) mice. E. (Left) Echocardiography and (Right) left ventricular systolic function in 2-month-old dTG(L) mice. F. Sarcomere patterns in 2-month-old WT and dTG(L) hearts and high-magnification images of specific regions are provided. The dotted lines indicate cardiomyocyte borders, with green insets highlighting partially disassembled sarcomeres. Arrows indicate corresponding Adducin expression. G. qPCR analysis of Add1_i2 mRNA in 2-month-old low expression dTG(H) hearts, normalized to WT. H. (Left) H&E staining and (Right) HW/BW comparison in 2-month-old dTG(H) mice. I. (Left) Echocardiography and (Right) left ventricular systolic function in 1-month-old dTG(H) mice. J. (Top) Sarcomere pattern in 1-month-old dTG(H) hearts with robust disassembly. (Bottom) High-magnification images of specific regions show the regions with a lack of disassembly. K. (Left) Evidence of cardiomyocyte fragmentation and (Right) High magnification of the inset indicate colocalization of Add1 and Tnnt2 in fragmented sarcomere (arrows). Scale bar 10 μm (F, J, K); 1 mm (D, H). Data are presented as mean±sd. Statistical analyses: repeated measures two-way ANOVA with Tukey's post-hoc test (A) or Mann-Whitney U test (two-tailed) (C, E, G, H, I) were used to determine statistical significance; *p < 0.05, **p < 0.01, ***p < 0.001 and ****p<0.0001.

The Add1/Add3 double transgenic mouse model exhibits partial sarcomere disassembly in adult hearts.

We generated cardiac-specific double transgenic (dTG) mouse line those co-expresses both Add1_i2 and Add3 (Figure 4B) to investigate their involvement in disassembly. Immunostaining confirmed the co-expression and colocalization of these isoforms in cardiomyocytes (Figure S6A). To further understand the expression dynamics, we divided dTG mice into low and high Add1_i2 mRNA expression groups (Figure 4C,G, respectively). Notably, Add1_i2 endogenous expression is low post-P14, making transgenic mRNA expression levels seem too high in comparison to the WT.

The low expression group (dTG-L) showed normal cardiac morphology and HW/BW at two-month-old (Figure 4D) but slightly reduced LV systolic function (Figure 4E), compared to WT. Intriguingly, in cardiomyocytes where Adducin was overexpressed, we observed partial sarcomere disassembly characterized by disorganized sarcomeres lacking clear striations (Figure 4F). However, the typical pattern of complete sarcomere disassembly and clearance was not evident.

The high expression group (dTG-H) had about double the Add1 mRNA level of dTG-L (Figure 4G) and exhibited ventricular dilation (Figure 4H, left), increased HW/BW (Figure 4H, right), and markedly reduced LV systolic function (Figure 4I) at two-month-old. Partial sarcomere disassembly with focal localization of Tnnt2 to the intercalated disk was observed (Figure 4J), although some myocytes appeared with no significant sarcomere disorganization (Figure 4J, lower row). Additionally, these mice also showed cardiomyocyte disruption and signs of increased cell death (Figure 4K), confirmed by TdT-mediated dUTP nick end labeling (TUNEL) assay (Figure S6B). Cardiomyocyte size analysis indicated a higher proportion of smaller cells in dTG-H (Figure S6C). These findings suggest that Add3 and Add1 co-expression stabilizes the Adducin complex, leading to partial sarcomere disassembly, with the high-expression group also experiencing cardiomyocyte death and diminished heart function.

Phospho-mimic Add1/Add3 double transgenic mouse sustains sarcomere disassembly.

The transgenic mouse expressing a phospho-mimic Add1(T445E/T480E) showed neonatal sarcomere disassembly, but not in adults (Figure 3G), likely due to reduced Add3 expression. To investigate further, we developed a phospho-double transgenic (p-dTG) mouse line co-expressing phospho-mimic Add1_i2(T445E/T480E) and WT_Add3 (Figure 5A, top). These mice had larger hearts as indicated by H&E staining (Figure 5A, bottom) and a trend of increasing HW/BW (Figure 5B), with slightly lower left ventricular ejection fraction (LVEF) at two months (Figure 5C), compared to WT. qPCR analysis confirmed increased expression levels of Add1_i2 (Figure 5D) and immunostaining showed co-localization of Add1 and Add3 in p-dTG cardiomyocyte cytoplasm (Figure S7A). Interestingly, p-dTG cardiomyocytes displayed disassembled sarcomeres across developmental stages (P7, P14, P21, and 2 months), with more pronounced disassembly in early neonatal stages (Figure 5Ei,ii,iii,iv). In adult cardiomyocytes, partial disassembly was primarily observed around the nuclei and cell periphery, with complete clearance of sarcomeres in these zones. Furthermore, analysis of proliferation markers, including pH3 and aurora B kinase, indicated an increase in proliferative cardiomyocytes at P21 and 2 months (Figure S7B,C), without changes in cell size (Figure S7D) or nucleation (Figure S7E). These findings suggest that the concomitant expression of these proteins promotes persistent sarcomere disassembly and potentially cardiomyocyte proliferation.

Figure 5. Co-expression of phosphor mimic Add1_i2 with Add3 promotes sarcomere disassembly.

A. (Top) Schematic strategy for generating cardiac-specific phospho-mimic Add1_i2/Add3 double transgenic mice (p-dTG), with T445 and T480 Glu substitutions in Add1_i2 to mimic phosphorylation, driven by an α-MHC promoter. (Bottom) H&E staining of 2-month-old control and p-dTG hearts. B. HW/BW in WT and p-dTG mcie in P21 and 2-month-old mice. C. (Left) Echocardiography and (Right) left ventricular systolic function in P21 and 2-month-old mice p-dTG mice. D. qPCR results for Add1_i2 in 2-month-old p-dTG. E. Sarcomere structure in p-dTG hearts at P7 (i), P14 (ii), P21 (iii) and 2-month-old (iv). High-magnification images of selected regions show sarcomere absence in areas with high Adducin expression (arrows). Dotted lines outline cardiomyocyte borders. F. Identification of 7 kinase candidates potentially phosphorylating (Left) T445 and (Right) T480 using 245 Ser/Thr peptide kinase assays, with corrected kinase activity values depicted. G. Verification of Nek1, Dcam1, and Irak4 kinase activity in phosphorylating T445 and T480 using Serine/Threonine Kinase dose-response assays. Each kinase was tested at three concentrations in triplicates. H. Endogenous expression pattern of (Left) Irak4 and (Right) Nek1 in the regenerating neonatal heart, with positive signals indicated (arrows). I. (Top) Schematic of Irak4 expression in AAV6 constructs. Irak4 (green) overexpression in NRVMs causes sarcomere disassembly (stained with α-actinin, red) disassembly (i) and translocation of Add1_pT445 (red) from nucleus to cytoplasm (ii). Arrowheads indicate Adducin is expressed in the nuclei in Irak4 negative cells. J. (Top left) Schematic of Irak4 expression in AAV9 constructs. (Top right) Experimental design for neonatal AAV injection. (Bottom) Results post AAV9-Irak4-GFP injection in CD1 pups at P1, showing intense sarcomere disassembly in cardiomyocytes compared to littermate controls at P15. Scale bar 10 μm (E, H, I, J); 1 mm (A). Data are presented as mean±sd. Statistical analyses: Two-way ANOVA with Bonferroni's multiple comparisons test (B, C), Mann-Whitney U test (two-tailed) (D); *p < 0.05, **p < 0.01, ***p < 0.001 and ****p<0.0001.

Given the variations in transgene copy numbers or integration sites could affect protein overexpression, we verified and assessed the Add1_i2 and Add3 expression in our TG lines using WB. All TG lines showed significantly higher transgene protein levels compared to WT littermates (Figure S8). The dTG line exhibited double the Add1_i2 levels, likely due to reduced Add1 degradation from Add3 overexpression. However, Add1_i2 levels were low in the p-dTG line, possibly due to our antibody's limited detection of phospho-mimic Adducin protein.

Identification of Irak4 as the kinase responsible for Adducin phosphorylation in cardiomyocytes.

To understand the phosphorylation mechanism of Add1 at T445/T480, crucial for sarcomere disassembly, we conducted a kinase screening assay with 245 kinases. This initial screen revealed seven kinases active on T445 (Figure 5F, left, Table_S2). Peptide of T480 only showed a significant response to kinase Vrk1 (Figure 5F, right). We then performed a secondary screen on the top hits outlined above using a HitConfirmation^™ dose-response assay. The result confirmed NIMA-related kinase 1 (Nek1), Interleukin 1 Receptor Associated Kinase 4 (Irak4), and Doublecortin-like and CAM kinase-like 2 (Dcamkl2) as key kinases for T445 phosphorylation, with a less pronounced effect on T480 (Figure 5G, Table_S3). In P4 hearts, both Irak4 and Nek1 colocalized with disassembled sarcomeres (Figure 5H).

We then explored the role of Irak4 in inducing sarcomere disassembly. In vitro overexpression of Irak4 by AAV6 infection led to NRVMs sarcomere disassembly (Figure 5I, i), correlated with cytoplasmic phosphorylated Add1-pT445 (Figure 5I,ii), while in Irak4 negative cells, Add1 remained nuclear. However, Nek1 overexpression in NRVMs showed no apparent effect on sarcomeres or Adducin phosphorylation (data not shown). To assess sarcomere changes in vivo, we generated an AAV9-Irak4 virus. We injected this virus into neonatal mouse hearts at P1 and collected hearts at P15, a time point well beyond cell cycle exit of neonatal cardiomyocytes. Immunostaining on heart sections demonstrated that cardiomyocytes with Irak4 overexpression displayed prominent sarcomere disassembly (Figure 5J). These results indicate that Irak4 phosphorylates Adducin in vitro and in vivo and induces sarcomere disassembly. Collectively, our results identify the Adducin heterodimer as a key regulator in cardiomyocyte sarcomere disassembly during mitosis.

Identification of binding partners for the Adducin complex during regenerative and non-regenerative phases.

Our findings suggest phosphorylated Adducin heterodimers play a key role in cardiomyocyte sarcomere disassembly. To identify Adducin's binding partners in cardiomyocytes, we conducted Add1 Co-IP/MS analysis on protein extracts from P1 and P7MI hearts. This approach, similar to our previous Tnnt2 pulldown but employed Add1 antibody as bait, revealing that Add1 shares many binding targets with Tnnt2 (Figure 6A, Table_S4). This result suggests a significant interaction of Add1 with sarcomeric proteins in neonatal hearts.

Figure 6. Analysis of Co-IP/MS with purified Adducin complex.

A. The table enumerates key proteins associated with Add1, as identified through Co-IP/MS, with a similar strategy as 1C. B. Schematic of the purification process of short and long Adducin complexes. C. Schematic of Co-IP/MS by purified adducin complex pulldown D. Bar graph for Canonical Pathways enriched in each sample. FDR: false discovery rate. E. Network plot of cytoskeleton proteins from the pull-down assay interact with Adducin long or Adducin short isoforms at different regenerative time points (P1 and P21). Different color labels indicate common and unique proteins among sample groups (Red: all sample groups, Yellow: common to all groups, Green: common to two or three groups, Blue: unique to each group). F. PLA result shows the direct interaction between Add1_pT445 and α-actinin in P4 heart tissue. Red dots (arrows) represent fluorescent signals of close interactions. Interaction between α-actinin and Tnnt2 serves as a positive control. G. Immuno-EM image details the subcellular location of Add1_pT445 in cardiomyocytes with disassembled sarcomeres in neonatal P1MI hearts. Two red boxes are magnified in the right 2 panels showing Add1_pT445 at (i) z-disks, and (ii) at z-disks associated with the plasma membrane in cardiomyocytes with disassembled sarcomeres. Red arrows indicate the location of Adducin. Scale bar 1 μm (G); 10 μm (F).

Considering that Adducin functions as an α/γ duplex, the aforementioned single antibody-based pulldown might not fully capture its binding pattern, especially given the limited Adducin expression in adult hearts. To investigate why sarcomere disassembly is more pronounced in neonatal than in adult cardiomyocytes, including those in dTG hearts, we performed a pulldown study using a purified Add1/Add3 complex as bait with heart extracts from disassembly (P1) and non-disassembly (P21) stages. We cloned both Add1 and Add3 genes into the plasmid pETDuet, which co-expresses two open reading frames (ORFs) together and forms a duplex naturally in cells, generating long (Add1_i1/Add3) and short (Add1_i2/Add3) forms (Figure 6B). After incubating these purified duplexes with heart lysates and binding them to Ni-NTA beads, we analyzed bound proteins using MS (Figure 6C). We also included a negative control where heart lysates were mixed with Ni-NTA beads. Log2(sample/control) fold changes greater than 1.5 were considered for further Ingenuity Pathway Analysis. Figure 6D is a bar graph of statistically significant canonical pathways enriched in the sample. Interestingly, the only pathway upregulated at P1 compared to P21 was the actin cytoskeleton signaling.

Network plots showed how this pathway's proteins interact with long or short Adducins at different stages (Figure 6E). Interestingly, both Adducin long and short forms specifically interacted with certain sarcomeric proteins, such as α-actinin at P1, but not P21. In adult cardiomyocytes, α-Actinin is localized to the Z-disks, where it is critical for organizing sarcomere structure. This finding suggested that Adducin and α-Actinin interaction could be key for sarcomere disassembly during cardiomyocyte proliferation phase. Therefore, we further explore their interaction by using proximity ligation assay (PLA) with an anti- Add1_pT445 antibody in neonatal hearts, using Tnnt2 binding to α-Actinin as a positive control. The results revealed a strong association of Add1_pT445 with α-Actinin at this stage (Figure 6F). Furthermore, using Immunogold electron microscopy (Immuno-EM), we examined Add_pT445 subcellular localization in neonatal cardiomyocytes, finding it at the Z-disks of sarcomeres (Figure 6Gi), where α-Actinin is also located, and also associated with disassembled sarcomeres near the cellular cortex (Figure 6Gii). Considering the role of Adducin in sarcomere disassembly and its association to α-Actinin, we hypothesized that this interaction might extract α-actinin from sarcomeres, leading to disassembly. To investigate whether the loss of α-actinin could promote disassembly, we developed an ACTN2-deficient human induced pluripotent stem cell (iPSC) cardiomyocyte line using Clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9), leading to an expected ACTN2 frameshift. WB analysis confirmed the effective deletion of ACTN2 (Figure S9A). Remarkably, after differentiation to cardiomyocytes, these cells exhibited a sarcomere disassembly phenotype similar to Adducin transgenics, with actin mainly localizing to the cellular cortex (Figure S9B). These insights provide a deeper understanding of the molecular interactions and localization between Adducin and α-Actinin in relation to sarcomere disassembly during neonatal cardiomyocyte mitosis.

Discussion

Numerous published reports have outlined various mechanisms of cardiomyocyte cell cycle regulation in recent years. These mechanisms encompass diverse pathways including direct cell cycle regulators, transcription factors, micro RNAs, Lnc-RNAs as well as non-myocyte factors including extracellular matrix, immunological, and environmental factors among others. However, striated muscle-specific mechanisms that regulate sarcomere organization during cardiomyocyte mitosis remain poorly understood. In the current report we demonstrate that the cytoskeleton protein Adducin regulates sarcomere disassembly during cardiomyocyte mitosis. In non-contractile cells, previous studies have identified mechanisms that regulate cytoskeletal assembly and actin sheets dimerization. For example, the cytoskeletal regulatory protein Adducin is a major regulator of actin bundling, and actin-spectrin interaction^38,39. The Adducin family is comprised of 3 members, namely α-, β-, or γ-Adducins. The fully assembled Adducin is comprised of heterodimers of homologous alpha and beta or gamma subunits. Despite the known role of Adducin in actin assembly in non-contractile cells,its role in contractile cells is entirely unknown.

Our results stem from a proteomics screen of proteins that differentially associate with Tnnt2 during an early regenerative timepoint of the newborn mouse heart. Although several cytoskeleton proteins appeared to be differentially associated with Tnnt2 during this early time point, we focused on Adducin given its known cytoskeleton regulatory role in non-contractile cells. We provide several lines of evidence that support the association of adducin with sarcomeric proteins; first, immunostaining demonstrates that α-Adducin is localized to the cytoplasm and membrane only in cardiomyocytes with disassembled sarcomeres. Second, Co-IP using Add1 antibody or purified α/γ-Adducin complex suggests a possible association of adducin with sarcomeric proteins, in particular α-Actinin. Third, we identified a membrane-associated phospho-isoform of Add1, which binds α-Actinin by PLA, and colocalizes with membrane-associated sarcomeres by Immuno-electron microscopy labeled with nanogold particles. These results indicate that Add1 is a sarcomere associated cytoskeletal protein in cardiomyocytes.

Importantly, gain of function studies in vitro and in vivo demonstrate that forced Adducin expression induces sarcomere disassembly. We show that expression of several isoforms of Adducin induces sarcomere disassembly in primary neonatal cardiomyocytes in culture. In addition, both single TG Add1_i2 and Add1_i1 phospho-mimic manifest a non-homogeneous expression pattern of adducin which resulted in extension of the sarcomere disassembly window until P14, but not adulthood. Protein stability assays demonstrated that co-expression of Add1 or Add3 is necessary for the stability of both proteins, and this was confirmed in vivo in the dTG lines which displayed stable expression of both Add1 and Add3 through adulthood. We also demonstrate that although the non-phosphomimic dTG lines display persistent adducin expression, this did not result in pronounced sarcomere disassembly even at high expression levels, while the same dTG construct with a constitutively active T445/T480 phospho-mimic resulted in sarcomere disassembly that persisted to adulthood.

One important limitation of the current findings is the incomplete disassembly we observed in adult cardiomyocytes even in the phospho-mimic dTG. Although we noted complete disassembly in the early postnatal period, most adult cardiomyocytes displayed incomplete disassembly that appeared to be localized to the zone surrounding nuclei and in the periphery of cardiomyocytes with complete clearance of the sarcomeres in these zones. This incomplete disassembly likely explains the lack of severe decline in LV systolic function in this dTG line. The mechanism of this differential effect of adducin on early postnatal compared to adult sarcomeres is unclear but might be related to changes in sarcomere phenotype with age. To address this question, we used purified alpha (either short or long isoform)-gamma Adducin protein complex to identify binding partners in early neonatal and adolescent hearts. The results demonstrate that the Adducin complex differentially binds to cytoskeletal proteins depending on the postnatal age. Specifically, both short and long forms of the Adducin complex appear to bind selectively to the cardiac-specific α-Actinin in P1 but not P21 hearts. PLA and immunogold-EM support the notion that Adducin associates with α-Actinin in Z-disks. Although Adducin has not been previously shown to bind α-Actinin, α-Actinin belongs to the Spectrin superfamily, and Spectrin is a known binding partner of Adducin in non-contractile cells. Given the known role of α-actinin in anchoring actin to Z-disks in cardiomyocytes, future studies should explore the potential role of α-actinin in the regulation of sarcomere disassembly downstream of Adducin. Collectively, these results identify an important regulatory mechanism of sarcomere disassembly, which links cytoskeleton organization to cell cycle regulation in mammalian cardiomyocytes and provides insights into the challenges of mitosis in actively contractile cells.

Clinical Perspective.

WHAT IS NEW?

• Mammalian hearts can regenerate spontaneously in the first few days of life but lose this ability as cardiomyocytes stop dividing shortly after birth. Sarcomere disassembly during mitosis is crucial for this early heart regeneration.

• Here we identify Adducin, a protein previously not studied in the heart, as a critical regulator of sarcomere disassembly during cardiomyocyte mitosis. Forced expressions of Adducin in adult hearts result in sarcomere disassembly and increased cardiomyocyte mitosis.

• Adducin induces sarcomere disassembly by interacting with alpha-actinin.

WHAT ARE THE CLINICAL IMPLICATIONS?

• These results identify Adducin as a regulator of sarcomere disassembly during cardiomyocyte mitosis during mammalian neonatal heart regeneration.

• Modulating Adducin expression promotes sarcomere disassembly and enhances the regenerative capacity of adult cardiomyocytes.

• These findings indicate that targeting sarcomere disassembly is plausible therapeutic strategy to promote myocardial regeneration.

Non-standard Abbreviations and Acronyms

MI

myocardial infarction

P

postnatal day

Tnnt2

cardiac troponin T

Actn2

α-Actinin

MS

mass spectrometry

Co-IP

co-Immunoprecipitation

P1MI

three days after MI in day-one

P7MI

three days after MI in day-seven

IB

immunoblotting

Add1

α-Adducin

Add3

γ-Adducin

WB

western blot

pH3

phosphor-histone H3-Ser10

NRVM

neonatal rat ventricle myocytes

Add1_i1

full-length Add1 isoform 1 (long form)

Add1_i2

Add1 isoform 2 (short form)

AAV

Adeno-associated virus

GFP

green fluorescent protein

TG

transgenic

α-MHC

α-Myosin heavy chain

H&E

Hematoxylin and Eosin

HW/BW

heart weight to body weight ratio

pH3⁺

pH3-positive

LV

left ventricular

WT

wild-type

T

Threonine

S

Serine

Add1_pT445

phosphorylation at Add1_Threonine 445 and 480 (T445/T480)

CDK1

cyclin-dependent kinase 1

PKA

protein kinases A

Glu

glutamic acid

Ala

alanine

CHX

cycloheximide

dTG

double transgenic

dTG-L

the Adducin low expression dTG group

dTG-H

the Adducin high expression dTG group

TUNEL

TdT-mediated dUTP nick end labeling

LVEF

left ventricular ejection fraction

p-dTG

phospho-double transgenic

Nek1

NIMA-related kinase 1

Irak4

Interleukin 1 Receptor Associated Kinase 4

Dcamkl2

Doublecortin-like and CAM kinase-like 2

ORFs

open reading frames

PLA

proximity ligation assay

Immuno-EM

Immunogold electron microscopy

iPSC

induced pluripotent stem cell

CRISPR

Clustered regularly interspaced short palindromic repeats

Cas9

CRISPR-associated protein 9

PSM

Peptide Spectrum Matches

LAD

left anterior descending

LVIDd

left ventricular internal diameters during end-diastole

LVIDs

left ventricular internal diameters during end-systole

EF

ejection fraction

FS

fractional shortening

PFA

paraformaldehyde

PBS

phosphate-buffered saline

qPCR

Quantitative reverse transcription–polymerase chain reaction assay

WGA

Wheat Germ Agglutinin