Inhibition of Hmbox1 Promotes Cardiomyocyte Survival and Glucose Metabolism Through Gck Activation in Ischemia/Reperfusion Injury

All data were analyzed using SPSS20.0 or GraphPad Prism 8 and presented as mean±SD using GraphPad Prism 8. Normality distribution test was first performed for all data. For data with normality distribution, unpaired Student t test, 1-way ANOVA test, or 2-way ANOVA test followed by Tukey post hoc test was used for comparisons of difference between 2 groups or multigroups. For data that did not pass the normality distribution test, nonparametric unpaired Mann-Whitney U test, Kruskal-Wallis test with the original false discovery rate method of Benjamini and Hochberg, or robust 2-way ANOVA test followed by post hoc pairwise Median Test in the rcompanion package was applied for comparisons of difference between 2 groups or multigroups. A P value <0.05 was considered statistically significant.

We previously reported Hmbox1 as a target gene of miR-222, a vital microRNA contributing to exercise-induced physiological cardiac growth.²⁹ Nevertheless, the role of Hmbox1 in physiological cardiac growth remained elusive. In the present study, we identified a significant downregulation of Hmbox1 in heart tissues from a murine swimming exercise model that induces physiological cardiac growth (Figure 1A and 1B). In contrast, Hmbox1 was upregulated in transaortic constriction–induced pathological cardiac hypertrophy (Figure S1A), highlighting differential regulation of Hmbox1 in physiological versus pathological hypertrophy. Next, we isolated NRCMs and neonatal rat cardiac fibroblasts and observed that Hmbox1 was preferentially expressed in cardiomyocytes compared with cardiac fibroblasts at both mRNA and protein levels (Figure S1B and S1C). Although Hmbox1 was downregulated in both cardiomyocytes and cardiac fibroblasts isolated from heart tissue of adult mice subjected to swimming (Figure S1D), functional experiments through upregulating or knocking down Hmbox1 in neonatal rat cardiac fibroblasts showed that Hmbox1 did not influence neonatal rat cardiac fibroblast proliferation or differentiation to myofibroblasts in vitro as determined by immunofluorescent staining for EdU (5-ethynyl-2’-deoxyuridine) and α-SMA (α-smooth muscle actin; Figure S2). We subsequently delved deeper into the function of Hmbox1 in cardiomyocytes.

Exercise-induced physiological cardiac growth is characterized by physiological hypertrophy and increased proliferation markers in cardiomyocytes.^9,29 Our experiments in NRCMs revealed that Hmbox1 knockdown increased cardiomyocyte size, whereas its overexpression decreased it; however, Hmbox1 did not affect cardiomyocyte proliferation (Figure 1C). Moreover, pathological hypertrophy markers atrial natriuretic peptide (Anp) and brain natriuretic peptide (Bnp) remained unaffected by Hmbox1 in cardiomyocytes (Figure 1D), indicating the physiological function of Hmbox1 in regulating cardiomyocyte size.

Meanwhile, we determined Hmbox1 expression in human embryonic stem cell–derived cardiomyocytes treated either by a physiological mechanical stretch using the Flexcell system or a pharmacological stimulus with IGF-1 (insulin-like growth factor 1), which simulated the effect of exercise in vitro.^30,31 Our results showed that both the mechanical stretch to a physiological extent and IGF-1, a canonical mediator for exercise-induced physiological cardiac hypertrophy,³² could downregulate Hmbox1 expression at both the mRNA and protein levels in human embryonic stem cell–derived cardiomyocytes, accompanied by an increased level of Akt phosphorylations, whereas Anp and Bnp expressions were not changed, thus excluding pathological hypertrophy in these conditions (Figure S3). Thus, Hmbox1 was also downregulated in exercise-simulated treatment of human cardiomyocytes. We further determined the function of Hmbox1 in human embryonic stem cell–derived cardiomyocytes and found that reducing Hmbox1 also enlarged human cardiomyocyte size while overexpressing Hmbox1 decreased cardiomyocyte size, without influencing the Anp and Bnp expressions (Figure S4), indicating that Hmbox1 was also able to regulate human cardiomyocyte size physiologically.

In vivo examination of the role of Hmbox1 in exercise-induced physiological cardiac growth involved injecting mice with AAV9 expressing Hmbox1 (overexpressing-Hmbox1-AAV9) and then subjecting them to swimming exercise (Figure 1E). We observed that neither cardiac systolic function nor diastolic function was markedly influenced in the swimming-exercised mice with Hmbox1 overexpression or not (Tables S1 and S2). After evaluating cardiac function, mice heart tissues were harvested, and we verified the efficiency of OE-Hmbox1-AAV9 in enhancing Hmbox1 expression in mice hearts (Figure 1F). It is interesting that the swim-induced increase in heart weight and the heart weight/tibia length ratio were reduced in mice overexpressing Hmbox1 (Figure 1G). Wheat germ agglutinin staining further indicated that upregulating Hmbox1 abolished the exercise-induced enlargement of the cross-sectional area of the myocardium (Figure 1H). However, the exercise-induced increase in Ki67-labeled cardiomyocytes was not influenced by Hmbox1 overexpression (Figure 1I), which was corroborating our in vitro findings of the noninvolvement of Hmbox1 in cardiomyocyte proliferation. Similarly, pathological hypertrophy markers Anp, Bnp, and β-Mhc (β-myosin heavy chain) were unaffected by Hmbox1 overexpression in mice hearts (Figure 1J). These data indicate that although Hmbox1 regulates physiological hypertrophy, it does not affect cardiomyocyte proliferation. This supports the idea that downregulating Hmbox1 is pivotal for exercise-induced physiological cardiac hypertrophy.

Considering exercise-regulated molecules might confer myocardial protection in pathological conditions, we assessed Hmbox1 expression in clinical and experimental postinfarct cardiac tissues, and then examined its function in cardiomyocytes upon pathological stress. Heart samples from 2 deceased patients with MI who had received vascular revascularization treatment were collected from the infarct area (Table S3). The myocardial slices were stained with double immunohistochemistry for C4d/Hmbox1, displaying a markedly high expression of Hmbox1 in MI zones adjacent to the necrotic cardiomyocytes (Figure 2A). These observations suggest that increased Hmbox1 may be involved in human MI disease progression.

In parallel, we assessed Hmbox1 expression in experimental mouse models of cardiac I/R injury. Hmbox1 was significantly upregulated in the border zone of I/R heart in the acute (24 hours) phase (Figure 2B), and in the remodeling I/R heart in the chronic (3 weeks) phase as well (Figure 2C). Likewise, Hmbox1 was upregulated in isolated NRCMs upon OGDR stress, which mimics cardiac I/R injury in vitro (Figure 2D). OGDR-induced cardiomyocyte apoptosis was then determined in Hmbox1 knocked down or overexpressed NRCMs. We first confirmed that Hmbox1 was efficiently knocked down by sh-Hmbox1 lentivirus (Figure S5A), and overexpressed by OE-Hmbox1 lentivirus in NRCMs (Figure S5C), either with OGDR stress or not. Hmbox1 knockdown inhibited OGDR-induced cardiomyocyte apoptosis, as evidenced by reduced terminal deoxynucleotidyl transferase mediated dUTP nick-end labeling (Tunel)–positive cardiomyocytes, diminished Bax/Bcl2 ratio, and decreased cleaved caspase3 expression (Figure 2E; Figure S5A). On the contrary, Hmbox1 overexpression aggravated cardiomyocyte apoptosis (Figure S5B and S5C). Meanwhile, we cultured human embryonic stem cell–derived cardiomyocytes and found an increase of Hmbox1 expression in OGDR-stressed human cardiomyocytes (Figure 2F), whereas reducing Hmbox1 prevented cardiomyocyte apoptosis (Figure 2G; Figure S6). Our data from both human and murine samples indicate that Hmbox1 is upregulated in postinfarct myocardium, whereas downregulation of Hmbox1 reduces cardiomyocyte apoptosis in both human and murine cardiomyocytes.

To assess the in vivo effects of targeting Hmbox1, we used a murine model of cardiac I/R injury. Mice were injected with AAV9, which mediated Hmbox1 knockdown through shRNA (sh-Hmbox1-AAV9), and then subjected to I/R injury (Figure S7A). Mouse hearts were harvested 24 hours after I/R, and the border zone of I/R hearts was used to determine the expression of Hmbox1 and to examine the apoptosis-related molecular changes. Our results showed that sh-Hmbox1-AAV9 significantly downregulated Hmbox1 in the heart tissues (Figure S7B and S7C). Functionally, Hmbox1 knockdown reduced the infarct size (Figure 3A) and inhibited myocardial apoptosis (Figure 3B; Figure S7C). At 3 weeks after I/R (Figure S8A through S8C), Hmbox1 knockdown also preserved cardiac function and inhibited cardiac apoptosis and remodeling (Figure 3C through 3F; Table S4; Figure S8C). Moreover, downregulation of Hmbox1 was associated with increased Akt phosphorylation, a key signaling pathway for cell survival, in I/R heart tissues (Figures S7D and S8D). Hmbox1 also influenced the Akt/mTOR/P70S6K pathway in cardiomyocytes in vitro (Figure S9). Our data further showed that Akt activation was necessary to mediate the effect of reduced Hmbox1 in promoting cardiomyocyte size and inhibiting cardiomyocyte apoptosis (Figure S10). This suggests that reducing Hmbox1 can activate the Akt/mTOR/P70S6K signaling, which might contribute to its protective effect against cardiac I/R injury and myocardial apoptosis.

To investigate the mechanisms underlying the protective effects of reducing Hmbox1 in myocardial tissue, we performed RNA-seq of heart tissues from both sham and 3-week-I/R mice injected with either sh-Hmbox1-AAV9 or CTL-AAV9. The RNA-seq profiling data has been uploaded to the Gene Expression Omnibus (GEO: GSE241079). We identified 119 upregulated and 38 downregulated genes in the I/R group compared with the sham group (fold change >2.0; P<0.05; Figure 4A). Concurrently, 83 genes were upregulated and 124 genes were downregulated in the I/R mice treated with sh-Hmbox1-AAV9 relative to those treated with CTL-AAV9 (fold change >2.0; P<0.05) (Figure 4B). Hmbox1 is recognized as a novel transcriptional repressor that negatively regulates target gene expression.¹² From the RNA-seq data, 12 genes were found to be downregulated in I/R hearts but upregulated after sh-Hmbox1-AAV9 intervention, as shown in the Venn diagram. We further used the Cistrome Data Browser and applied the Tools “Check a putative target” to predict the potential targets that Hmbox1 might regulate at the promoter regions and found that Myl4, Pllp, and Gck had the top 3 scores among the 12 genes (Figure 4C). We continued to assess the expression of these genes in the animal I/R model and the OGDR-stressed NRCMs by quantitative reverse-transcription polymerase chain reaction. Only Gck was consistently downregulated in the myocardium after both I/R injury and OGDR stress, but was upregulated after Hmbox1 inhibition at the mRNA level (Figure 4D; Figure S11A) and the protein level as well (Figure S11B and S11C), which was consistent with the RNA-seq data. According to the quantitative reverse-transcription polymerase chain reaction validation experiments, although Myl4 was upregulated in the I/R heart tissues compared with the sham group, and was downregulated in the I/R mice injected with sh-Hmbox1 AAV9 compared with those injected with CTL-AAV9 (Figure 4D), the results from quantitative reverse-transcription polymerase chain reaction for Myl4 were not consistent with the RNA-seq analysis, which was at least in part attributed to the within-group variations of the RNA-seq data for Myl4. In addition, Hmbox1, known as a transcriptional repressor, was not able to negatively regulate Myl4 expression in NRCMs (Figure 4D). Considering the above reasons, Gck was selected to be further studied. Kyoto Encyclopedia of Genes and Genomes enrichment analysis of metabolism pathways further showed that Gck could be involved in glycolysis (Figure 4E), highlighting the possible regulatory role of Hmbox1 in myocardial metabolism upon I/R injury. Collectively, we identified Gck as the most likely downstream target of Hmbox1 during cardiac I/R injury.

To further elucidate the regulation of Gck by Hmbox1 in cardiomyocytes, we infected primary NRCMs with either OE-Hmbox1 or sh-Hmbox1 lentivirus. We observed that Hmbox1 negatively regulated Gck at both mRNA and protein levels (Figure 4F and 4G; Figure S11D and S11E). Both the luciferase reporter experiment and the chromatin immunoprecipitation polymerase chain reaction assay confirmed the ability of Hmbox1 to directly bind to and enrich in the promoter region of Gck (Figure 4H and 4I). These findings suggest that Hmbox1 operates as a transcriptional repressor, negatively regulating Gck in cardiomyocytes.

Next, we performed function-rescue experiments to evaluate the involvement of Gck in the function of Hmbox1. In NRCMs cotreated with sh-Hmbox1 and Gck siRNA (si-Gck), we observed that downregulation of Gck significantly reversed the increased cell size and antiapoptotic effect observed with Hmbox1 knockdown (Figure S12A and S12B; Figure 4J and 4K). Using a glucokinase activator, GKA50, we observed that GKA50 could inhibit cardiomyocyte apoptosis under OGDR stress (Figure S12C) and efficiently increased Gck activity in cardiomyocytes (Figure S12D). Meanwhile, we demonstrated that Gck activation in NRCMs counteracted the negative regulation of cell size and the proapoptotic effect seen with Hmbox1 overexpression (Figure S12E and S12F). Taken together, these findings underscore the role of Hmbox1 as a transcriptional repressor targeting Gck. By reducing Hmbox1, physiological hypertrophy is promoted, and cardiomyocyte apoptosis is inhibited through Gck activation.

Gck, also known as glucokinase, encodes hexokinase-IV, a member of the hexokinase protein family. This enzyme catalyzes the initial step of glycolysis: the phosphorylation of glucose to form glucose-6-phosphate.³³ Impaired glycolysis and mitochondrial respiration are critical contributors to cardiomyocyte death during cardiac hypoxic and ischemic stress.^34,35 However, the role of Hmbox1 inhibition in regulating myocardium metabolism remains largely unknown. Using the XFe96 extracellular flux analyzer, we observed that OGDR stress significantly reduced mitochondrial respiration in isolated cardiomyocytes, as indicated by reduced levels of oxygen consumption rate, basal respiration, maximal respiration, and spare respiratory capacity (Figure 5A). Concurrently, ATP production also diminished in cardiomyocytes subjected to OGDR stress (Figure 5A). However, the Hmbox1 inhibition mitigated the OGDR-induced reduction of mitochondrial respiration (Figure 5A). Conversely, OGDR stress increased the glycolysis proton efflux rate and enhanced the basal and compensatory glycolysis levels in isolated cardiomyocytes (Figure 5B). We hypothesize that this metabolic shift is a self-responsive adaptation of cardiomyocytes to OGDR stress. It is interesting that Hmbox1 inhibition further amplified glycolysis in OGDR-stressed cardiomyocytes (Figure 5B).

To further investigate the role of Gck activation in Hmbox1-regulated metabolism, cardiomyocytes were cotreated with si-Gck and sh-Hmbox1 lentivirus. Results demonstrated that Gck knockdown significantly diminished the beneficial effects of Hmbox1 inhibition on mitochondrial respiration and glycolysis (Figure 5C and 5D). Collectively, this evidence directly suggests that reducing Hmbox1 improves mitochondrial respiration and glycolysis in cardiomyocytes under OGDR stress by activating Gck.

To specifically assess the functional role of Hmbox1 in cardiomyocytes in vivo, we used cardiac myocyte–specific Hmbox1 cKO (Hmbox1^fl/fl Cre⁺) mice and control (Hmbox1^fl/fl Cre^-) mice and established a cardiac I/R injury model (Figure 6A and 6B). We first confirmed efficient Hmbox1 suppression in mouse heart tissues (Figure 6C and 6I). Three weeks after I/R injury, Hmbox1 cKO mice demonstrated preserved cardiac function and diminished cardiac remodeling (Figure 6D through 6G; Table S5). Cardiac-specific Hmbox1 cKO mice exhibited a reduced Bax/Bcl2 ratio and caspase3 cleavage in heart tissues, alongside increased Akt phosphorylation (Figure 6H; Figure S13), suggesting attenuated myocardial apoptosis. It is important to note that these mice also displayed elevated Gck expression in heart tissues (Figure 6I and 6J), indicating the efficacy of cardiac-specific Hmbox1 ablation in activating Gck in vivo.

In a separate animal experiment, we injected Hmbox1^fl/fl mice with AAV9 vectors expressing cardiac-specific cTnT (cardiac troponin T) promoter–driven Cre (induced Cre-AAV9, iCre-AAV9), followed by cardiac I/R injury or sham operation a week later (Figure S14A). Three weeks after I/R injury, Hmbox1 was significantly reduced in mice heart tissues (Figure S14B and S14H). Interestingly, cardiac-specific Hmbox1 knockdown markedly improved cardiac function, alleviated cardiac apoptosis and remodeling (Figure S14C through S14G; Table S6; Figure S15), and was accompanied by an increase in cardiac Gck expression (Figure S14H and S14I). Collectively, this evidence strongly supports the protective role of cardiac myocyte–specific Hmbox1 suppression against cardiac I/R injury and its activation of Gck in vivo.

To identify potential upstream regulators of Hmbox1, we consulted the PROMO (prediction of transcription factor binding sites) database, which predicted GR, STAT4, IK-1, and ETS1 as candidates. Given the downregulation of Hmbox1 in exercised hearts and its upregulation during cardiac I/R injury, we first examined the regulation of these genes in these conditions. Validation experiments revealed that only ETS1 exhibited downregulation in OGDR-stressed NRCMs and heart tissues after I/R injury and upregulation in the hearts of swimming-exercised mice (Figures S16, S17A, and S17B; Figure 7A and 7B). This suggests an inverse relationship between ETS1 and Hmbox1 regulation in these scenarios. Subsequent studies demonstrated that ETS1 repressed Hmbox1 expression in primary NRCMs (Figure 7C; Figure S17C and S17D). Chromatin immunoprecipitation polymerase chain reaction assay analysis revealed an enrichment of ETS1 in the promoter region of Hmbox1 (Figure 7D). Gain- and loss-of-function experiments in NRCMs showed that ETS1 overexpression increased cell size and inhibited cardiomyocyte apoptosis, whereas ETS1 knockdown had the opposite effects (Figure S18). Rescue experiments further revealed that Hmbox1 knockdown abolished the effects of reduced ETS1 on cardiomyocyte size and apoptosis (Figure S19). These findings indicate that ETS1 acts as a predominant upstream inhibitor of Hmbox1 in the myocardium.

ETS1 is known to play pivotal roles in cardiac development and myocardial survival.^36–38 However, its function in cardiac I/R injury in vivo remained unclear. Here, we introduced the myocardial injection of ETS1-expressing (OE-ETS1) lentivirus and subsequently established a cardiac I/R injury model in mice. We observed that ETS1 overexpression reduced infarct size and preserved cardiac function after I/R injury (Figure 7E and 7F; Table S7). Efficacy validation revealed that the OE-ETS1 lentivirus increased cardiac ETS1 expression while suppressing Hmbox1 mRNA expression in heart tissues with I/R injury (Figure 7G; Figure S20A). Mice overexpressing ETS1 exhibited diminished cardiac remodeling and reduced fibrosis- and pathological hypertrophy–associated gene expressions (Figure 7H; Figure S20B). Moreover, ETS1 overexpression resulted in reduced Hmbox1 and elevated Gck protein levels, and enhanced Akt phosphorylation levels in I/R-injured mouse hearts (Figure 7I; Figure S20C). Collectively, these findings reveal that ETS1 negatively regulates Hmbox1 in the heart. Overexpression of ETS1 effectively downregulates Hmbox1 and activates Gck, providing protection against cardiac I/R injury.

Considering that therapeutic approaches are mostly applied after the onset of cardiac I/R injury in clinical practice, we further investigated the therapeutic efficacy of AAV9-mediated Hmbox1 knockdown. Mice were injected with either sh-Hmbox1-AAV9 or CTL-AAV9 within 30 minutes after myocardial reperfusion (Figure 8A). After validating Hmbox1 knockdown in heart tissues (Figure 8B and 8G), we demonstrated that the sh-Hmbox1-AAV9 treatment effectively preserved cardiac function and mitigated cardiac remodeling 3 weeks after I/R injury (Figure 8C through 8F; Table S8), whereas cardiac diastolic function was unaltered in the model (Table S9). Mice treated with sh-Hmbox1-AAV9 exhibited decreased expression of apoptosis-related proteins (Figure S21A) and increased Gck expression and Akt phosphorylation in heart tissues (Figure 8G; Figure S21B and S21C). Collectively, these data suggest that AAV9-mediated Hmbox1 reduction remains a viable therapeutic strategy for protection against cardiac I/R injury.

Supplemental Methods

Figures S1–S21

Tables S1–S10

References 52–54