Abstract

   Rationale: Improving heart regeneration through reactivating
   cardiomyocyte proliferation holds a great potential for repairing
   diseased hearts. We recently reported that LSD1-dependent epigenetic
   repression of Cend1 transcription is prerequisite for cardiomyocyte
   proliferation and mouse heart development. This study interrogates the
   potential role of this LSD1-CEND1 axis in heart regeneration and
   repair.

   Methods: The cardiomyocyte-specific Lsd1 knockout or overexpression
   mice, Cend1 null mice and cardiomyocyte-specific Cend1 overexpression
   mice were used to determine the role of LSD1-CEND1 axis in heart
   regeneration after experimental injuries. Neonatal and adult mice were
   subjected to apical resection or left anterior descending coronary
   artery ligation, respectively, to establish cardiac injury models.
   Echocardiography and Masson staining were employed to assess cardiac
   function and histopathology, respectively. The molecular changes were
   determined using RNA sequencing, quantitative RT-PCR, Western blotting
   and immunostaining.

   Results: Cardiomyocyte-specific deletion impeded neonatal heart
   regeneration, while overexpression of Lsd1 had the opposite effect. RNA
   sequencing revealed that Cend1, a crucial suppressor of cardiomyocyte
   cycling, was the most significantly elevated gene induced by Lsd1 loss
   during heart regeneration. Cardiomyocyte-specific Cend1 overexpression
   hindered neonatal heart regeneration, while Cend1 loss in nullizygous
   mice had the opposite effect. Cend1 deletion resulted in gene
   expression alterations associated with enhanced cardiomyocyte
   proliferation, neovascularization, and macrophage activation.
   Furthermore, the cardiac regeneration defect caused by Lsd1 loss was
   not observed when experiments were performed with mice that were
   nullizyogus for Cend1. Moreover, we found that either Lsd1
   overexpression or Cend1 deletion could promote heart regeneration and
   repair, and improve cardiac function following experimental myocardial
   infraction in adult mice.

   Conclusion: Our results demonstrate that LSD1-dependent suppression of
   CEND1 is crucial for heart regeneration in neonatal and adult mice
   after experimental injury. These findings suggest LSD1 activation and
   CEND1 inhibition as promising therapeutic strategies to enhance
   endogenous cardiac repair in humans.

   Keywords: heart regeneration, Cend1, LSD1

Introduction

   In mammals, cardiomyocyte proliferation underpins myocardial tissue
   growth during embryonic development, and this proliferative activity in
   cardiomyocytes persists postnatally but is transient and quickly lost
   [65]^1. In preadolescent humans, it is observed that the majority of
   cardiomyocytes exit the cell cycle, with only very few cells
   maintaining the capacity to proliferate. As such, evidence for
   cardiomyocyte regeneration through cell proliferation and replacement
   in healthy and diseased hearts remains elusive [66]^2. Notably,
   clinical studies have shown that the hearts of newborn children with
   severe myocardial infarction appear to undergo spontaneous repair
   through a mechanism that is ostensibly associated with myocardial
   regeneration [67]^3^, [68]^4, yet this capacity is lost in adult heart
   tissues. Moreover, studies of neonatal mouse hearts subjected to apical
   resection (AR) show extensive tissue repair within three weeks, in
   part, through proliferation of preexisting cardiomyocytes to replace
   lost cells [69]^5. Consistent with findings in humans, this capacity
   for heart regeneration in neonatal mice is lost as soon as 7 days after
   birth, with what is documented as near-complete loss of cardiac
   regenerative capacity in adult mice [70]^5. In contrast, studies of
   lower vertebrates such as zebrafish have shown that adult
   cardiomyocytes can re-enter the cell cycle, proliferate and repair
   damaged hearts [71]^6. Such observations have inspired efforts to boost
   cardiomyocyte proliferation as a mean to promote tissue regeneration in
   adult mammalian hearts. Indeed, a growing number of studies over the
   past two decades have provided evidence that the proliferative capacity
   of adult cardiomyocytes may be reactivated through manipulating key
   factors that control cardiomyocyte cycling [72]^7^-[73]^10. Given that
   heart diseases are a leading cause of death worldwide [74]^11,
   identifying the key cellular and molecular regulators that control
   cardiomyocyte cycling and neonatal heart regeneration has clinical
   implications for developing treatments that leverage such self-repair
   mechanisms within the adult heart.

   One molecular regulator of interest is lysine-specific demethylase 1
   (LSD1, also known as KDM1A), the first identified member of a family of
   demethylases that primarily target mono- and di-methylated lysines on
   histone 3 (H3K4me1/2 and H3K9me1/2), so as to regulate gene
   transcription in a positive and negative fashion, respectively
   [75]^12^, [76]^13. Indeed, LSD1 upregulation is well documented in
   multiple types of cancers and to promote tumor cell proliferation
   [77]^14^-[78]^18. By using genetic mouse models, we recently showed
   that LSD1 is essential for cardiomyocyte proliferation during fetal and
   neonatal heart development in mice, and that its downstream target gene
   CEND1 (cell cycle exit and neuronal differentiation 1) controls
   cardiomyocyte proliferation and myocardial growth. Mechanistically,
   LSD1 represses Cend1 transcription by erasing H3K4me2 marks at its
   promoter [79]^19. Moreover, consistent with its function as an inducer
   of neuroprogenitor cell cycle exit [80]^20, CEND1 negatively regulates
   cardiomyocyte cycling through sustaining p53 signaling [81]^19. As a
   corollary, Fei et al. reported that knockdown of Lsd1 inhibited heart
   regeneration in neonatal mice [82]^21. Despite these findings, the
   relevance for a 'LSD1-CEND1 signaling axis' in neonatal and adult
   cardiac regeneration and the underlying mechanism remains elusive.

   In this study, we employed several mouse models to alter expression
   levels of Lsd1 or Cend1 in cardiomyocytes and, through these
   experiments, we have found that the suppression of CEND1 by LSD1 is
   crucial for neonatal heart regeneration. We have also conducted
   experiments to show that either Lsd1 overexpression or Cend1 deletion
   promoted adult heart regeneration and repair, as well as improved
   cardiac function following myocardial infraction. These results suggest
   that targeting the LSD1-CEND1 axis could be beneficial as a strategy
   for triggering heart regeneration in adult mammals.

Results

Complementary expression patterns for LSD1 and CEND1 in postnatal murine
heart

   Initially, we examined the mRNA and protein levels of Lsd1 and Cend1
   genes in postnatal murine hearts using qPCR and Western blotting (WB),
   respectively. As shown, the steady-state levels of Lsd1 mRNA decreased
   between postnatal days 1 and 7 (P7) and further declined at P14 and
   then remained unchanged until adulthood (P56) ([83]Figure S1A). Western
   blotting revealed that steady-state levels of LSD1 were reminiscent of
   its mRNA expression pattern, with a significant decrease detected
   between P1 and P7, however, the signal diminished through to adulthood
   ([84]Figure S1B). In contrast, mRNA levels for Cend1 in these tissues
   were significantly elevated at P7 when compared with P1, while
   expression levels at P28 and P56 were similar to P1 levels ([85]Figure
   S1C). Western blotting revealed that CEND1 protein levels significantly
   increased between P1 and P7 and this increase progressed to adulthood
   ([86]Figure S1D). These findings suggest that the expression patterns
   of Lsd1 and Cend1 are complementary in the postnatal murine heart as it
   ages.

Lsd1 deletion impedes neonatal heart regeneration

   To explore the role of LSD1 in neonatal heart regeneration in vivo, we
   transduced cardiomyocyte-specific loss of Lsd1 in mice (Lsd1^MKO)
   through targeted injection of AAV9-cTnT^cre virus into the left
   ventricles of conditional Lsd1^f/f loxp mice at P1, concurrent with AR
   surgery (Figure [87]1A). As shown, when analyzed the mice at P14, we
   found that the LSD1 protein levels in the hearts of Lsd1^MKO mice were
   significantly reduced, compared to mice that received control virus
   (Figure [88]1B-C). The ratios of heart-to-body weight of P21 Lsd1^MKO
   mice were significantly higher than that in age-matched control mice
   (Figure [89]1D). Echocardiography analysis revealed that Lsd1^MKO mice
   had worse cardiac function, evidenced by reduced ejection fraction (EF)
   and fraction shortening (FS), compared to control mice (Figure [90]1E).
   Masson staining showed that more than 80% of control mice did not show
   evidence of fibrosis at the apical region of their hearts, while a far
   lower percentage of Lsd1^MKO mice had this feature (Figure [91]1F),
   suggesting that conditional deletion of Lsd1 impeded cardiac
   regeneration. In addition, WGA staining revealed that Lsd1^MKO mice
   featured cardiomyocytes that were larger in size than control mice
   (Figure [92]1G). Moreover, Lsd1 deletion significantly inhibited
   cardiomyocyte proliferation, shown by reduced numbers of Ki67 (a marker
   for DNA synthesis) and pH3 (a marker for mitosis) positive
   cardiomyocytes in Lsd1^MKO mice, compared to control (Figure [93]1H-I).
   Therefore, these findings demonstrate that LSD1 is important for
   neonatal cardiac regeneration following AR surgery.

Figure 1.

   [94]Figure 1
   [95]Open in a new tab

   Lsd1 deletion influences neonatal heart regeneration following AR
   surgery. (A) Schematic diagram showing experimental procedures. Mice
   subjected to cardiomyocyte-specific Lsd1 knockout (Lsd1^MKO) were
   established by injecting AAV9-cTNT^cre virus into left ventricles of
   their hearts at postnatal day 1 (P1). Control mice received AAV-GFP
   virus. Both Lsd1^MKO and control mice were subjected to apical
   resection (AR). (B) Western blotting analysis of the LSD1 protein
   levels in the hearts of P14 mice (n = 6/group). (C) Co-immunostaining
   of LSD1 and cTNT antibodies in P14 hearts (n = 5/group). The mean
   fluorescence intensity (MFI) was quantified and shown. (D)
   Heart-to-body weight ratios for control (n = 13) and Lsd1^MKO (n = 15)
   mice at P21. (E) Representative images of echo analysis showing cardiac
   function of mice at P21. Bar charts showing the values of ejection
   fraction (EF), fractional shortening (FS). n = 9/group. (F)
   Representative images of Masson staining showing the fibrotic scar in
   the hearts of control and Lsd1^MKO mice at P21. The hearts with or
   without fibrotic scar within apex region were considered
   non-regeneration or regeneration respectively. Bar chart showing the
   quantification for the proportion of regeneration and non-regeneration
   hearts among Control (n = 13) or Lsd1^MKO (n = 9) group. (G)
   Immunostaining of WGA indicating cardiomyocyte size of control (194
   cells from 8 mice) and Lsd1^MKO (285 cells from 8 mice) mice. (H, I)
   Immunostaining with Ki67 (H) and pH3 (I) antibodies indicating
   proliferating cells (arrowhead). Bar charts showing quantification for
   the numbers of Ki67 or pH3 positive cardiomyocytes (CMs) marked by cTNT
   staining in control (n = 5) and Lsd1^MKO (n = 6) mice. HPF, high power
   field. *p < 0.05, **p < 0.01, ***p < 0.001 by unpaired student's
   t-test.

Lsd1 overexpression ameliorates neonatal heart regeneration

   It is recognized that the capacity for cardiac regeneration in mice is
   negligible after P7. To investigate whether Lsd1 gain-of-function might
   influence the regenerative capacity of neonatal hearts beyond P7, we
   transduced cardiomyocyte-specific Lsd1 overexpression (Lsd1^OE) in
   mouse hearts through injection of AAV9-Lsd1 virus into the left
   ventricles of wild type (WT) mice at P7, followed by AR surgery (Figure
   [96]2A). The AAV9-Lsd1 virus is designed to confer over-expression of
   Lsd1 under the control of a cardiac-specific cTnT promoter. qPCR,
   Western blotting and immunostaining confirmed successful Lsd1
   overexpression in cardiac tissues, compared to control treatment
   (Figure [97]2B-D). When analyzing the treated mice at P21, we found
   that control mice showed poor cardiac tissue repair at P21, as
   characterized by insufficient cardiac function and apparent fibrotic
   scar at the apex region (Figure [98]2E-F). In contrast, more than 80%
   of Lsd1^OE mice featured extensive repair of heart tissue at P21 after
   AR surgery at P7, and with normal cardiac function and heart histology
   (Figure [99]2E-F). Moreover, the size of cardiomyocytes in Lsd1^OE mice
   were smaller than that in controls (Figure [100]2G). Also, Ki67, pH3
   and Aurora B (a marker for cytokinesis) immunostaining showed that
   cardiomyocyte proliferation was significantly increased in Lsd1^OE mice
   compared to control (Figure [101]2H-J). In addition, Western blotting
   analysis revealed a significant increase in PCNA protein expression (a
   proliferation marker) in Lsd1^OE mouse hearts relative to controls
   (Figure [102]2K). In contrast, the cell cycle regulator Cyclin D1
   protein levels remained unchanged following Lsd1 overexpression
   ([103]Figure S2A). Notably, histological and echocardiographic analyses
   demonstrated that sole overexpression of Lsd1 had no significant impact
   on either cardiac growth or function in juvenile mice ([104]Figure
   S3A-C). Thus, adenovirus-mediated Lsd1 gain-of-function in P7 hearts
   prior to AR surgery can induce cardiomyocyte proliferation and promote
   heart regeneration.

Figure 2.

   [105]Figure 2
   [106]Open in a new tab

   Lsd1 overexpression ameliorates neonatal heart regeneration following
   AR surgery. (A) Schematic diagram showing experimental procedures. Mice
   subjected to cardiomyocyte-specific Lsd1 overexpression (Lsd1^OE) were
   generated by injecting AAV9-cTNTP-Lsd1 virus into left ventricle of the
   hearts of wide type (WT) mice at postnatal day 7 (P7). Control mice
   received control virus. Both Lsd1^OE and control mice were subjected to
   apical resection (AR) at P7. (B and C) The mRNA and protein levels of
   LSD1 in the hearts of P21 mice were measured by qPCR (B, n = 5) and
   Western blotting (C, n = 3) respectively. (D) Co-immunostaining of LSD1
   and cTNT antibodies in P21 hearts (n = 5/group). The mean fluorescence
   intensity (MFI) was quantified and shown. (E) Representative images of
   echo analysis showing cardiac function of mice at P21. Bar charts
   showing the values of ejection fraction (EF), fractional shortening
   (FS). n = 6/group. (F) Representative images of Masson staining showing
   the fibrotic scar in the hearts of P21 mice. The hearts with or without
   fibrotic scar within apex region were considered non-regeneration or
   regeneration respectively. Bar charts showing quantification for the
   proportion of regeneration and non-regeneration hearts among control (n
   = 14) or Lsd1^OE (n = 15) group. (G) Immunostaining of WGA to measure
   cardiomyocyte sizes in heart tissues from control (826 cells from 6
   mice) and Lsd1^OE (663 cells from 6 mice) mice. (H-J) Immunostaining
   with Ki67 (H), pH3 (I) and Aurora B (J) antibodies indicating
   proliferating cells (arrowheads). Bar charts showing the quantification
   for the numbers of Ki67, pH3 or Aurora B positive cardiomyocytes (CMs)
   marked by cTNT staining in control (n = 5-6) and Lsd1^OE (n = 5-6)
   mice. (K) Western blotting analysis of the PCNA protein levels in the
   hearts of P21 mice (n = 4/group). HPF, high power field. *p < 0.05, **p
   < 0.01, ***p < 0.001 by unpaired student's t-test.

Cend1 is repressed by Lsd1 and inhibits neonatal heart regeneration

   To explore the downstream targets of LSD1 that increase heart
   regeneration, we analyzed the transcriptomic profiles of heart tissues
   collected from P14 control and Lsd1^MKO mice that received AR surgery
   at P1. As shown, the targeted Lsd1 loss in the heart tissues of
   Lsd1^MKO mice led to 1342 differentially expression genes. Among these,
   we identified 699 and 643 genes that were upregulated and
   downregulated, respectively (Figure [107]3A). We recently identified
   that Cend1 is a key negative regulator of cardiomyocyte cycling
   suppressed by Lsd1 during heart development [108]^19 and,
   interestingly, the RNA-seq results showed that Cend1 was the most
   significantly elevated gene resulting from Lsd1 loss concurrent with AR
   surgery in our experiment (Figure [109]3B). This upregulation was
   confirmed at the translational level by immunostaining and Western
   blotting (Figure [110]3C-D). In a parallel experiment in which Lsd1 was
   conditionally overexpressed in the setting of AR surgery (that is, in
   Lsd1^OE mice), Cend1 was downregulated in the hearts of these mice,
   compared to controls (Figure [111]3E, [112]Figure S4A). Given that
   Cend1 inhibits cardiomyocyte proliferation during heart development
   [113]^19, our current results suggest that Cend1 may function as a
   crucial negative regulator of cardiomyocyte proliferation and heart
   regeneration. To test this hypothesis, we induced
   cardiomyocyte-specific Cend1 overexpression (Cend1^OE) mice through
   injection of AAV9-Cend1 virus into the left ventricles of WT mice at
   P1, followed by AR surgery (Figure [114]3F). The AAV9-Cend1 virus
   confers over-expression of Lsd1 under the control of a cardiac-specific
   cTnT promoter. qPCR and immunostaining confirmed significant Cend1
   overexpression in cardiomyocytes (Figure [115]3G, [116]Figure S4B). In
   this experiment, mouse heart repair at P21 was poor for Cend1^OE mice
   compared to control, as characterized by poor cardiac function and
   fibrotic scar at the apex region, in comparison to comprehensive heart
   repair in over 90% of control mice (Figure [117]3H-I). Moreover, the
   cardiomyocyte size in Cend1^OE mice was significantly larger, compared
   to that in controls (Figure [118]3J). Furthermore, immunostaining
   revealed that Cend1 overexpression resulted in significant reductions
   in Ki67 and pH3, indicative decrease in cardiomyocyte proliferation
   (Figure [119]3K-L). Notably, histological and echocardiographic
   analyses revealed that Cend1 overexpression had no significant impact
   on either cardiac growth or function in juvenile mice under
   physiological conditions ([120]Figure S4C-E).

Figure 3.

   [121]Figure 3
   [122]Open in a new tab

   Cend1 is repressed by Lsd1 and its overexpression impedes neonatal
   heart regeneration. (A-B) Heart ventricle tissues were isolated from
   P14 control and Lsd1^MKO hearts which received apical resection (AR)
   surgeries at P1. Volcano plots of RNA-sequencing datasets showing
   differentially expressed genes between two groups (A). A Heatmap
   showing the top 10 upregulated genes induced by Lsd1 loss (B). (C)
   Co-immunostaining of LSD1 and cTNT antibodies in P21 hearts (n =
   4/group). The mean fluorescence intensity (MFI) was quantified and
   shown. (D) Western blotting analysis of the CEND1 protein level in P21
   hearts (n = 6/group). (E) qPCR analysis of the Cend1 mRNA level in P21
   hearts (n = 5/group). (F) Schematic diagram showing experimental
   procedures. Mice subjected to cardiomyocyte-specific Cend1
   overexpression (Cend1^OE) were generated by injecting AAV9-cTNTP-Cend1
   virus into the left ventricles of the hearts of wide type (WT) mice at
   P1, while the mice received AAV9-GFP virus were used as controls. Both
   Cend1^OE and control mice were subjected to AR surgeries at P1. (G)
   qPCR analysis of the Cend1 mRNA level in the hearts of P14 mice (n =
   4/group). (H) Representative images of echo analysis showing cardiac
   function of mice at P21. Bar charts showing the values for ejection
   fraction (EF) and fractional shortening (FS). n = 9/group. (I)
   Representative images of Masson staining showing the fibrotic scar in
   the hearts of P21 mice. The hearts with or without fibrotic scar within
   apex region were considered non-regeneration or regeneration
   respectively. Bar chart showing the quantification for the proportion
   of regeneration and non-regeneration hearts among control (n = 9) or
   Cend1^OE (n = 11) group. (J) WGA staining indicating cardiomyocyte size
   of control (878 cells from 5 mice) and Cend1^OE (722 cells from 5 mice)
   mice. (K, L) Immunostaining with Ki67 (K) and pH3 (L) antibodies
   indicating proliferating cells (arrowheads). Bar charts showing the
   quantification for the numbers of Ki67 or pH3 positive cardiomyocytes
   (CMs) marked by cTNT staining in control (n = 5) and Cend1^OE (n = 5)
   mice. HPF, high power field. *p < 0.05, **p < 0.01, ***p < 0.001 by
   unpaired student's t-test.

   Next, we used nullizygous Cend1^-/- mice to explore if Cend1 repression
   is necessary for the regeneration of neonatal heart. First, we observed
   that under physiological conditions, Cend1^-/- mice showed no
   significant differences in cardiac histology and function compared to
   control mice ([123]Figure S4F-H). Subsequently, we subjected both
   Cend1^-/- and wild type control mice to AR surgery at P7 and assessed
   cardiac repair at P14 and P21 (Figure [124]4A). qPCR analysis confirmed
   that Cend1 transcripts were undetectable in Cend1^-/- hearts (Figure
   [125]4B). The heart-to-body weight ratios of Cend1^-/- mice were
   significantly reduced, compared to controls at P21 (Figure [126]4C).
   Furthermore, more than 90% of control mice exhibited poor heart repair
   at P21, evidenced by impaired cardiac function and obvious fibrotic
   scar at the apex region (Figure [127]4D-E). In contrast, over 80% of
   Cend1^-/- mice featured significant heart repair, characterized by
   normal cardiac function and histological features (Figure [128]4D-E).
   Moreover, cardiomyocytes in Cend1^-/- mice were significantly smaller
   than that those in controls (Figure [129]4F). In agreement, levels of
   Ki67, pH3 and Aurora B were significantly increased in Cend1^-/- mice
   (Figure [130]4G-I). Furthermore, the PCNA protein levels were markedly
   increased in the hearts of Cend1^-/- mice compared to controls (Figure
   [131]4J). In contrast, Cyclin D1 protein levels remained unchanged
   following Cend1 deletion ([132]Figure S2B). Together, these findings
   suggest that Cend1 repression is critical for heart regeneration in
   neonatal mice.

Figure 4.

   [133]Figure 4
   [134]Open in a new tab

   Loss of Cend1 is permissive for juvenile heart regeneration. (A)
   Schematic diagram showing experimental procedures. Control (WT) and
   Cend1^-/- mice were subjected to apical resection (AR) surgeries at P7.
   (B) qPCR analysis of the Cend1 mRNA level in P14 hearts (n = 6/group).
   (C) Heart-to-body weight ratios of mice at P21 (n = 10/group). (D)
   Representative images of echo analysis showing cardiac function of mice
   at P21. Bar charts showing the values of ejection fraction (EF) and
   fractional shortening (FS) for control (n = 10) and Cend1^-/- (n = 8)
   mice. (E) Representative images of Masson staining showing the fibrotic
   scar in the hearts of P21 mice. The hearts with or without fibrotic
   scar within apex region were considered non-regeneration or
   regeneration respectively. Bar charts showing the quantification for
   the proportion of regeneration and non-regeneration hearts among each
   group (n = 23/group). (F) Immunostaining of WGA indicating
   cardiomyocyte size of control (345 cells from 8 mice) and Cend1^-/-
   (360 cells from 8 mice) mice. (G-I) Immunostaining with Ki67 (G), pH3
   (H) and Aurora B (I) antibodies indicating proliferating cells
   (arrowheads). Bar charts showing the quantification for the numbers of
   Ki67, pH3 or Aurora B positive cardiomyocytes (CMs) marked by cTNT
   staining in control (n = 5-6) and Cend1^-/- (n = 5-6) mice. (J) Western
   blotting analysis of the PCNA protein levels in P21 hearts (n =
   4/group). HPF, high power field. *p < 0.05, **p < 0.01, ***p < 0.001 by
   unpaired student's t-test.

   To determine the mechanism by which Cend1 loss promoted neonatal heart
   regeneration, we performed RNA-sequencing to compare transcriptomic
   variations between control and Cend1^-/- heart tissue at P14. Among the
   3136 differentially expressed genes, 1932 were upregulated and 1204
   were downregulated ([135]Figure S5A). GO analysis revealed that these
   upregulated genes were involved in cell cycle regulation, angiogenesis
   and macrophage activation, while the downregulated genes were
   associated with sarcomere, cardiomyocyte differentiation and cardiac
   muscle contraction ([136]Figure S5B). Notably, heart tissues ablated
   with Cend1 showed elevated expression levels for M2 macrophages markers
   such as ll1ra, CD163, Mrc1, Pecam1, Ccl24, Igf1, Arg1, CD86, Cxcl1
   ([137]Figure S5C). In addition, GSEA assays revealed that the DEGs
   associated with Cend1 knockout were enriched in vascular endothelial
   growth factor production, angiogenesis and extracellular matrix (ECM)
   regulation ([138]Figure S5D). In line with these transcriptomic
   changes, Cend1^-/- hearts had more capillary vessels marked by
   VE-cadherin and reparative macrophages immunostained by CD206
   ([139]Figure S5E-F). Notably, the mRNA and protein levels for Agrn, OPN
   (encoded by Spp1) and Versican, three extracellular matrix proteins
   known to promote heart regeneration [140]^22^-[141]^24, were
   significantly upregulated in the heart tissues of Cend1^-/- mice
   compared to controls (Figure [142]5G-I). These results demonstrate that
   Cend1 loss leads to transcriptomic changes associated with
   cardiomyocyte proliferation, angiogenesis and macrophage polarization.

Figure 5.

   [143]Figure 5
   [144]Open in a new tab

   Cend1 deletion rectifies the regeneration defects caused by Lsd1 loss.
   The Lsd1^f/f;Cend1^-/- and Lsd1^f/f mice at postnatal day 1 (P1) were
   injected with AAV9-cTNT^cre virus to generate Lsd1 and Cend1 double
   knockout (DKO) and Lsd1 single knockout mice (Lsd1^MKO) respectively,
   while Lsd1^f/f mice received AAV9-GFP virus and these were used as
   controls. The mice were subjected to apical resection (AR) surgeries at
   P1 and analyzed at P21. (A) Representative images of Masson staining
   showing the fibrotic scar in the hearts of P21 mice. The hearts with or
   without fibrotic scar within apex region were considered
   non-regeneration or regeneration respectively. Bar charts showing the
   quantification for the proportion of regeneration and non-regeneration
   hearts among control (n = 13), Lsd1^MKO (n = 9) or DKO (n = 15) group.
   (B) Immunostaining of WGA indicating cardiomyocyte size of control (330
   cells from 8 mice), Lsd1^MKO (360 cells from 8 mice) and DKO (330 cells
   from 7 mice) mice. (C, D) Immunostaining with pH3 (C) and Ki67 (D)
   antibodies indicating proliferating cells (arrowhead). Bar charts
   showing the quantification for the numbers of Ki67 or pH3 positive
   cardiomyocytes (CMs) marked by cTNT staining (n = 8/group). *p < 0.05,
   **p < 0.01, ***p < 0.001 by One-way ANOVA with Tukey tests.

LSD1 supports neonatal heart regeneration through suppressing CEND1

   To determine the functional importance for LSD1-dependent repression of
   Cend1 in neonatal heart regeneration, we generated compound mutant,
   Lsd1^f/f; Cend1^-/- mice homozygous for Cend1 knockout and the
   conditional (loxp) Lsd1 allele (hereafter refers to DKO). With these
   DKO mice, we injected AAV9-cTnT^cre virus into their left ventricles at
   P1 when AR surgery was performed simultaneously. Masson staining of
   tissues showed that the majority of control and DKO mice had extensive
   heart repair at P21, in contrast to all Lsd1^MKO mice that failed to do
   so, as evidenced by the presence of fibrotic scars at the apex regions
   of hearts (Figure [145]5A). When measuring the cardiomyocyte size, we
   found that it was comparable between DKO and control mice, while
   cardiomyocytes were significantly larger in Lsd1^MKO mice (Figure
   [146]5B). Moreover, the Ki67 and pH3 positive cells were significantly
   reduced in Lsd1^MKO mice compared to control. However, these marker
   expression in DKO mice was significantly elevated compared to Lsd1^MKO
   mice, manifesting enhanced proliferation in DKO mice (Figure
   [147]5C-D). Thus, conditional loss of LSD1 in Cend1 knockout mice is
   beneficial for AR surgery-induced neonatal heart regeneration.

LSD1 suppression of CEND1 is beneficial for adult heart repair

   Given that it mediated neonatal heart regeneration, we asked whether
   the LSD1-CEND1 axis might be relevant to adult heart regeneration and
   repair. To achieve this, we administered AAV-Lsd1 virus to adult (8
   weeks) WT mice that confers Lsd1 transgene expression under the control
   of a cTnT promoter, to generate Lsd1^OE mice. In parallel, control mice
   received AAV-GFP virus. Next, these virus-injected mice were subjected
   to acute myocardial infraction (AMI) through left anterior descending
   coronary artery ligation (LAD) (Figure [148]6A). The successful
   overexpression of Lsd1 in the hearts of Lsd1^OE mice was confirmed
   through Western blotting analysis (Figure [149]6B). The heart-to-body
   weight ratios of Lsd1^OE mice were significantly smaller compared to
   controls (Figure [150]6C). Also, the cardiac function of Lsd1^OE mice
   was significantly improved over control mice 28 days after AMI
   surgeries, evidenced by increased values of EF and FS (Figure [151]6D
   and [152]Figure S6A). Consistently, Lsd1^OE mice showed reduction in
   fibrotic scars in their heart tissues when compared to controls (Figure
   [153]6E). WGA staining further showed that the cardiomyocyte size in
   Lsd1^OE mice was significantly smaller than that in control hearts
   (Figure [154]6F). Consistently, Lsd1 overexpression following AMI
   induced significant cardiomyocyte proliferation, compared to AMI alone
   in adult mice (Figure [155]6G-I). In addition, the PCNA protein levels
   were significantly increased in the hearts Lsd1^OE mice compared to
   controls (Figure [156]6J). In contrast, Cyclin D1 protein levels were
   unaffected by Lsd1 overexpression ([157]Figure S2C). It is important to
   note that Lsd1 overexpression had no detectable impact on cardiac
   structure or function under baseline conditions ([158]Figure S3D-F).

Figure 6.

   [159]Figure 6
   [160]Open in a new tab

   Lsd1 overexpression promotes adult heart repair. (A) Schematic diagram
   showing experimental procedures. Adult wild type (WT) mice were
   injected with AAV9- cTnTP-Lsd1 via tail vein to generate mice with
   cardiomyocyte-specific Lsd1 overexpression (Lsd1^OE), while mice that
   received AAV9-GFP were used as controls. Two weeks after virus
   injection, the mice were subjected to generation of acute myocardial
   infraction (AMI) model by left anterior descending artery ligation
   surgeries. (B) The protein levels of LSD1 in the hearts of D7 mice were
   measured by Western blotting (n = 4/group). (C) Heart-to-body weight
   ratios for control (n = 12) and Lsd1^OE (n = 12) mice at 4 weeks post
   AMI surgeries. (D) Bar charts showing the values of ejection fraction
   (EF) and fractional shortening (FS) of control (n = 10-11) and Lsd1^OE
   (n = 10-11) mice at indicated times. (E) Representative images of
   Masson staining showing the fibrotic scar in the hearts. The bar chart
   showing the quantification for the proportion of fibrotic scar among
   total myocardium in control (n = 11) and Lsd1^OE (n = 9) mice. (F)
   Immunostaining for WGA indicating cardiomyocyte size of control (481
   cells from 5 mice) and Lsd1^OE (681 cells from 5 mice) mice. (G-I)
   Immunostaining with Ki67, pH3 and Aurora B antibodies indicating
   proliferating cells (arrowheads). Bar charts showing the quantification
   for the numbers of Ki67 (G), pH3 (H) or Aurora B (I) positive
   cardiomyocytes (CMs) marked by cTNT staining in control (n = 5) and
   Lsd1^OE (n = 5) mice. (J) Western blotting analysis of the PCNA protein
   levels in D28 hearts (n = 4/group). HPF, high power field. *p < 0.05,
   ***p < 0.001 by unpaired student's t-test.

   For functional relevance, we performed LAD ligation on adult Cend1^-/-
   (n = 20) and WT (Control, n = 16) mice (Figure [161]7A). At 28 days
   after AMI, the survival rates of WT mice were lower (at 22.5%) compared
   to Cend1^-/- mice, where only one Cend1^-/- mouse died after AMI
   (Figure [162]7B). Moreover, the heart-to-body weight ratios of
   Cend1^-/- mice were significantly different from those of WT mice
   (Figure [163]7C). Moreover, Cend1^-/- mice had increased EF and FS
   values (Figure [164]7D, [165]Figure S6B), indicating that Cend1 loss
   led to enhanced recovery of cardiac function following AMI.
   Consistently, Cend1^-/- mice had significant less fibrotic scars in
   their hearts compared to controls (Figure [166]7E). In addition, the
   cardiomyocyte size in Cend1^-/- mice was significantly smaller than it
   in controls (Figure [167]7F). Concomitant with the improved heart
   repair phenotypes in Cend1^-/- mice after AMI, immunostaining revealed
   elevated expression of Ki67, pH3 and Aurora B in cardiomyocytes (Figure
   [168]7G-I). Like Lsd1^OE mice, Cend1^-/- mice exhibited significant
   upregulation of PCNA protein levels in myocardial tissues compared to
   controls (Figure [169]7J). In contrast, Cyclin D1 protein levels were
   comparable between two groups ([170]Figure S2D). Taken altogether, our
   findings suggest that LSD1 induces cardiomyocyte proliferation and
   promotes adult heart repair likely through downregulation of the cell
   cycle inhibitor Cend1.

Figure 7.

   [171]Figure 7
   [172]Open in a new tab

   Adult heart repair following experimental AMI is enhanced in Cend1
   knockout mice. (A) Schematic diagram showing experimental procedures.
   Adult wild type (WT, control) and Cend1^-/- mice were subjected to
   generation of acute myocardial infarction (AMI) model by left anterior
   descending artery ligation surgeries. (B) A chart showing the survival
   curve of control (n = 20) and Cend1^-/- (n = 16) mice. (C)
   Heart-to-body weight ratios for mice at 28 days after AMI. n =
   12/group. (D) Bar charts showing the values of ejection fraction (EF)
   and fractional shortening (FS) of control (n = 12) and Cend1^-/- (n =
   12) mice at indicated times. (E) Representative images of Masson
   staining showing the fibrotic scar in the hearts. Bar chart showing the
   quantification for the proportion of fibrotic scar among total
   myocardium in control (n = 12) and Cend1^-/- (n = 12) mice. (F)
   Immunostaining of WGA indicating cardiomyocyte size of control (479
   cells from 8 mice) and Cend^-/- (435 cells from 8 mice) mice. (G-I)
   Immunostaining with Ki67, pH3 and Aurora B antibodies indicating
   proliferating cells (arrowheads). Bar charts showing the quantification
   for the numbers of Ki67 (G), pH3 (H) or Aurora B (I) positive
   cardiomyocytes (CMs) marked by cTNT staining in control (n = 5-6) and
   Cend^-/- (n = 5-6) mice. (J) Western blotting analysis of the PCNA
   protein levels in D28 hearts (n = 4/group). HPF, high power field. *p <
   0.05, **p < 0.01, ***p < 0.001 by unpaired student's t-test.

Discussion

   We recently demonstrated that LSD1-dependent epigenetic suppression of
   Cend1 is essential for cardiomyocyte proliferation and heart
   development in mice [173]^19. Here we extended that study to find that
   repression of Cend1 by LSD1 is crucial for cardiomyocyte proliferation
   during neonatal and adult heart regeneration after experimental injury
   ([174]Figure S7).

   The neonatal mouse heart subjected to either AR or LAD surgeries can
   extensively repair itself through a mechanism that involves
   proliferation of existing cardiomyocytes [175]^5^, [176]^25, but this
   repair capacity is largely lost after P7 because the majority of
   cardiomyocytes exit cell cycle beyond this postnatal stage [177]^5
   through mechanism that so far has not been well characterized. In our
   current study, we found that the expression level of LSD1 was high in
   the hearts of P1 mice but decreased after P7, while the level of CEND1
   featured a complementary pattern of expression. Indeed, their dynamic
   expression pattern is observed during the critical period for neonatal
   heart tissue regeneration, and this motivated our investigation that
   LSD1-CEND1 signaling could be relevant to cardiomyocyte cell cycle exit
   and proliferation in the context of injury and repair in early and,
   potentially, adult hearts. Our hypothesis is supported by the evidence
   in our current study that collectively indicates that, loss of Lsd1 or
   overexpression of Cend1 in heart tissues can impede heart regeneration
   in neonatal P1 mice following AR surgery, while Lsd1 overexpression or
   loss of Cend1 significantly promote heart regeneration in juvenile P7
   mice. This is consistent with previous findings that Lsd1 knockdown in
   mice curtails neonatal heart regeneration [178]^21.

   In our recent report, we identified that Cend1 was upregulated
   following Lsd1 loss and, furthermore, genetic deletion of Cend1 could
   correct the cardiomyocyte proliferation defect and embryonic lethality
   in cardiomyocyte-specific Lsd1 knockout mouse embryos [179]^19. In line
   with our previous findings, RNA sequencing results in the current study
   identified Cend1 as the most significantly upregulated gene in the
   hearts of Lsd1 knockout mice that were subjected to AR surgeries.
   Moreover, genetic deletion of Cend1 was able to rescue the impaired
   cardiac regeneration in Lsd1 knockout mice. Mechanistically, our recent
   work demonstrated that LSD1 represses Cend1 transcription through
   H3K4me2 demethylation at its promoter [180]^19. Emerging evidence
   highlights the context-dependent nature of LSD1 function, wherein its
   biological outcomes are dictated by stage-specific binding partners. In
   embryonic stem cells, LSD1-CoREST complexes maintain pluripotency by
   silencing differentiation genes [181]^26^, [182]^27. During myeloid
   lineage commitment, LSD1-NuRD interactions promote terminal
   differentiation [183]^28. Our findings extend this paradigm to cardiac
   regeneration, where LSD1-mediated Cend1 suppression appears to
   facilitate proliferative responses. However, the specific co-factors
   recruiting LSD1 to the Cend1 locus in neonatal versus adult
   cardiomyocytes remain undefined.

   Notably, CEND1 overexpression not only inhibited proliferation but also
   induced cardiomyocyte hypertrophy, suggesting its dual role in
   coordinating cell cycle exit and growth adaptation. Given our previous
   demonstration that CEND1 activates p53 signaling in cardiomyocytes
   [184]^19, we propose a mechanistic model wherein CEND1-mediated p53
   upregulation simultaneously represses cell cycle progression through
   canonical p21 activation, and promotes hypertrophic growth via
   induction of metabolic reprogramming [185]^29^, [186]^30. This duality
   mirrors the established role of p53 in orchestrating both proliferative
   arrest and compensatory hypertrophy in pressure-overloaded hearts
   [187]^31^, [188]^32, though further studies are needed to dissect
   CEND1-specific contributions to these processes. Together, our findings
   demonstrate that Cend1 is a prominent, negative regulator of
   cardiomyocyte proliferation and that is suppressed by LSD1, although
   other factors may be at play. For example, Mahmoud et al. reported that
   the transcription factor MEIS1 negatively regulates cardiomyocyte
   cycling and that Meis1 deletion could induce cardiomyocyte
   proliferation in mice after P7 [189]^9. When considered together, it is
   plausible that CEND1 and MEIS1 would act as brakes for cardiomyocyte
   cycling in the context of heart tissue regeneration and repair. In
   addition, multiple positive regulators of cardiomyocyte cycling such as
   YAP1 [190]^33, GATA4 [191]^34, Tbx20 [192]^35^, [193]^36, NRG1
   [194]^37, IGF-1, HGF [195]^38, Pkm2 [196]^39, PPAR [197]^40, ERBB2
   [198]^41^, [199]^42, FSTL1 [200]^43, mir302-367 [201]^44, mir-31a-5p
   [202]^45, has-miR-590 [203]^46, has-miR-199a [204]^47 have also been
   identified to be critical for heart repair through enhanced
   cardiomyocyte cycling. In this light, we speculate that the mechanism
   driving sufficient cardiomyocyte proliferation in adult mammalian
   hearts might be achieved through a gene therapy approach that
   simultaneously activates proliferation alongside de-repression of the
   negative regulators for cardiomyocyte cycling.

   Heart repair involves many molecular and cellular processes including
   inflammatory response, immune cell migration and activation,
   cardiomyocyte proliferation as well as angiogenesis [205]^48. In the
   present study, transcriptomic profiling of Cend1^-/- hearts revealed
   that the upregulated genes in Cend1^-/- mice were involved in G1/S
   phase transition, vasculature development and macrophage activation.
   Immunostaining experiments confirmed that there were more vessels and
   M2 macrophages in the hearts of Cend1^-/- mice, compared to controls.
   Enhanced angiogenesis is known to promote heart regeneration and repair
   following myocardial infraction in several animal models [206]^49. In
   addition, M2 macrophages can promote heart regeneration and repair
   through their clearance functions as well as paracrine signaling
   [207]^24^, [208]^50. Together with our current findings, we surmise
   that Cend1 deletion led to promotion of heart repair through elevated
   cardiomyocyte proliferation, angiogenesis and macrophage activation.
   However, the precise molecular mechanisms underlying CEND1-mediated
   gene expression regulation remain to be elucidated and represent an
   important direction for future investigation. To further dissect the
   cell-autonomous functions of CEND1, subsequent studies employing cell
   type-specific conditional knockout models (targeting cardiomyocytes,
   endothelial cells, fibroblasts, and macrophages) will be particularly
   valuable for understanding its compartmentalized roles in cardiac
   regeneration.

   The composition of extracellular matrix (ECM) is highly dynamic and
   associated with cardiomyocyte proliferation during heart development
   and repair following injuries [209]^51^, [210]^52. Several ECM proteins
   can induce cardiomyocyte proliferation and promote heart regeneration
   following myocardial infraction in mammalian models. Agrn, mainly
   produced by endothelial cells, was the first identified ECM protein
   which promotes heart regeneration following myocardial infraction in
   mice [211]^53 and pigs [212]^22. More recently, Versican [213]^23 and
   SPP1 [214]^24, respectively derived from fibroblasts and macrophages,
   have been shown to induce heart regeneration and improve cardiac
   function following myocardial infraction in mice. Interestingly, the
   expression levels for these three ECM proteins were significantly
   elevated in the hearts of Cend1^-/- mice. Future studies will define
   the mechanisms through which CEND1 regulates the expression of Agrn,
   Versican and Spp1 to influence heart repair following injury.

   Our study demonstrates that LSD1 overexpression or CEND1 deletion
   enhances cardiac regeneration in both neonatal and adult mouse models.
   While neonatal mice exhibited more robust regenerative capacity
   consistent with known developmental windows of cardiac plasticity, the
   observed functional improvements in adult mice - though statistically
   significant - warrant cautious interpretation due to fundamental
   differences between murine and human cardiac biology and the inherent
   limitations of adult mammalian regeneration. These findings provide
   important mechanistic insights into the LSD1/CEND1 regulatory axis in
   cardiac repair while highlighting the need for further validation in
   higher mammalian systems before considering clinical translation. These
   results nevertheless identify LSD1 activation/CEND1 inhibition as
   promising new directions for developing therapeutic strategies to
   promote endogenous cardiac repair in humans.

Methods

Animals

   Mice subjected to neonatal cardiomyocyte-specific Lsd1 knockout
   (Lsd1^MKO) were generated by injecting AAV9-cTnT^cre (10 μl per pup)
   into the left ventricle of hearts of Lsd1^f/f mice [215]^54 at
   Postnatal day 1 (P1). Mice subjected to cardiomyocyte-specific Lsd1
   overexpression (Lsd1^OE) were generated by injecting AAV9-cTnTP-Lsd1
   into the left ventricle (10 μl per pup) of hearts of P7 wild-type (WT)
   mice or through tail vein (100 μl per mice) of adult WT mice at 8
   weeks. Mice subjected to cardiomyocyte-specific Cend1 overexpression
   (Cend1^OE) were generated by injecting AAV9-cTnTP-Cend1 into the left
   ventricle (10 μl per pup) of hearts of WT mice at P1. The mice injected
   with equal volume of AAV9-GFP virus were used as controls. Cend1^-/-
   mice were purchased from GemPharmatech Co., Ltd and crossed with
   Lsd1^f/f mice to generate Lsd1^f/f; Cend1^-/- mice which were subjected
   Lsd1 and Cend1 double knockout (DKO) in heart tissues by injecting
   AAV9-cTnT^cre (10 μl per pup) into their left ventricle of hearts at
   P1. Mice care and experimentation procedures were carried out using
   protocols authorized by the Institutional Animal Care and Use Committee
   (IACUC) of Xian Jiaotong University. All mice were housed under typical
   environmental conditions of 45-65% relative humidity, temperatures of
   21-24 °C, a 12 h light/dark cycle and free access to water and food in
   the animal Research Center of Xian Jiaotong University.

Apical resection

   Apical resection (AR) was performed on P1 or P7 mice as described
   previously [216]^5. Briefly, neonatal pups were anesthetized by ice
   freezing for 5-10 min and then subjected to intercostal incisions
   between the third and fourth intercostal space to separate the
   pericardium and expose the heart. Injury was induced by surgically
   cutting the left ventricle apex (~15% of the ventricular myocardium).
   Finally, the chest wall and skin incision were both closed with 7-0
   sutures. The neonatal pups of the sham group underwent the same
   procedure without apical resection. After surgery, all neonatal pups
   were transferred onto 37℃ heating pads for resuscitation.

Histology and immunostaining

   Mouse hearts were collected and fixed in 4% paraformaldehyde for 24-48
   h at 4°C and then processed for paraffin embedding. The heart tissues
   were cut at a 5 μm thickness. Masson's trichrome staining and
   Hematoxylin and Eosin (H&E) were performed according to standard
   procedures. For immunostaining, tissue sections were prepared from
   paraffin-embedded tissue samples and dewaxed, then rehydrated through a
   series of graded ethanol washes with increasing proportional volumes of
   water, and followed by antigen retrieval through boiling with sodium
   citrate buffer (Sigma, S1804) (pH 6.0). The tissues were then blocked
   with 5% horse serum (Solaribo, SL042) at room temperature (RT) for 1 h
   and subsequently incubated with primary antibodies ([217]Table S1)
   diluted in blocking buffer at 4 ℃ overnight. The following day, tissues
   sections on the glass slides were incubated with fluorescent-labelled
   secondary antibodies at RT for 1h. Nuclear counter staining was
   performed with DAPI staining (Beyotime, C1006). The stained slides were
   subjected to image acquisition using a confocal microscope.

Western blotting

   Heart tissues were homogenized and lysed in RIPA buffer containing
   protease inhibitors. After spinning at 4°C for 15 minutes, supernatants
   were collected and subjected to protein concentration quantification
   through a BCA assay kit. Defined quantities of lysates were denatured
   with SDS loading buffer, separated by 10% polyacrylamide gel
   electrophoresis and transferred to PVDF membranes. The membranes were
   then blocked in 5% BSA in Tris buffered saline with Tween-20 (TBST)
   solution for 1 h at RT and incubated with primary antibodies diluted in
   TBST containing 5% BSA at 4°C overnight. On the following day, the
   membranes were washed with TBST for three times and incubated with
   second antibodies conjugated with HRP at RT for 1h. After three times
   of washing with TBST, the membranes were subjected to color development
   by using an ECL detection kit and the chemiluminescence immune
   detection system.

RNA extraction and quantitative PCR (qPCR)

   Total RNA was isolated from mice hearts using the Trizol solution
   (Invitrogen 9088901). For cDNA synthesis, 1 μg of total RNA was reverse
   transcribed with a HiScript II reverse transcriptase Kit (Vazyme R222).
   Then qPCR was performed using the SYBR Green PCR Master Mix (Genstar
   A304-05) and gene-specific primers ([218]Table S2). Results were
   calculated by using the 2-^△△CT method and standardized by
   glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

RNA-seq data analysis

   Heart tissues were collected from P14 control, Lsd1^MKO and Cend1^-/-
   mice and subjected to RNA-sequencing. The data were processed and
   analyzed as described previously [219]^19. In brief, the raw sequencing
   reads were first assessed for quality using Fastp (version 0.23.1)
   [220]^55 and selected for further analysis with the following excluding
   criteria: (1) containing adapter contamination, (2) more than 10% of
   bases are uncertain, (3) the proportion of low quality (Phred quality <
   5) bases is over 50%. Next, the selected reads were mapped to the mouse
   reference genome (GRCm39) using the Hisat2 (v2.0.5) [221]^56. The read
   numbers and FPKM for each gene were calculated using FeatureCounts
   (v1.5.0-p3) [222]^57. After this, differential expression genes (DEGs)
   were identified using DEseq2 package47 in R 4.2.1, and p values were
   adjusted using the Benjamini and Hochberg's approach. Following
   exclusion of genes with less than 50 counts, genes of significantly
   differential expression were determined with padj ≤0.05 and |log2
   (foldchange)| ≥ 0.5. Finally, The DEGs were subjected to Gene Ontology
   (GO) and KEGG pathways enrichment analysis using the clusterProfiler
   package (R 4.2.1). GO and KEGG terms with corrected p < 0.05 were
   considered significantly enrichment. Gene Set Enrichment Analysis
   (GSEA) of the complete pre-ranked gene list was carried out using the
   GSEA_4.1.0 software and the Hallmark gene set collection within the
   Molecular Signatures Database (MSigDB).

Echocardiography

   Mice were first anesthetized by inhalation of isoflurane mixed with
   pure oxygen at a concentration of 2% to 3% for induction of anesthesia
   and 1.5% to 2% for maintenance of anesthesia. Then, body hair from the
   left precordial chest region as removed using depilatory cream.
   Transthoracic echocardiography was performed using the VINNO 70
   (VisualSonics) equipped with an X10 transducer operating at 23 MHz
   frequency. All measurements were obtained under light anesthesia (1.5%
   isoflurane) with maintained body temperature at 37°C. Standard
   parasternal long-axis and short-axis views were acquired at the
   papillary muscle level, with all analyses performed by an investigator
   blinded to experimental groups. The ejection fraction (EF) and
   fractional shortening (FS) features were both measured on M-mode
   images.

Experimental Myocardial Infarction

   Adult mice were subjected to surgical ligation of the left anterior
   descending coronary artery (LAD) to establish myocardial infarction
   (MI). In brief, the mice were anesthetized by inhalation of isoflurane
   mixed with pure oxygen. Subsequently, a 1 cm incision was made in the
   precardiac area of the mice, followed by blunt separation of the
   anterior thoracic muscle and intercostal muscles. Then gently rotating
   forceps were inserted into the chest through the left anterior 3-4
   costal space. A 6-0 silk thread was ligated approximately 2 mm below
   the left atrium at a depth of 2-3 mm. The successful ligation was
   confirmed by observing the color change of distal vessel from red to
   pale. Finally, the chest and skin incision were closed by using 4-0
   silk sutures. The mice were warmed up on a heating pad at 37°C for
   several minutes until recovery.

Statistical analysis

   All statistical analyses were conducted using Graph Pad Prism 9.0
   software. All experimental results were from at least three biological
   replicates and presented as mean ± standard deviation. Comparison
   between two groups were conducted using a two-tailed Student's t-test,
   while one-way analysis of variance was used for comparisons among
   multiple groups. A p-value less than 0.05 was considered statistically
   significant.

Supplementary Material

   Supplementary figures and tables.
   [223]thnov15p6313s1.pdf^ (4.9MB, pdf)

Acknowledgments