Cisd2 delays atrial aging via a modulation of calcium homeostasis that mitigates atrial myopathy

Abstract

Age-associated atrial myopathy results in structural remodeling and a disturbance of atrial conductance. Atrial myopathy often precedes atrial fibrillation (AF) and can facilitate AF progression. However, the molecular mechanism linking aging to atrial deterioration remains elusive. CDGSH iron-sulfur domain-containing protein 2 (CISD2) is a mammalian pro-longevity gene. We used Cisd2 knockout (Cisd2KO) and Cisd2 transgenic (Cisd2TG) mice to investigate pathophysiological mechanisms underlying age-related atrial myopathy. Four findings are pinpointed. Firstly, in both humans and mice, the level of atrial CISD2 declines during natural aging; this correlates with age-associated damage, namely degeneration of intercalated discs, mitochondria, sarcoplasmic reticulum (SR) and myofibrils. Secondly, in Cisd2KO and naturally aged wild-type mice, Cisd2 deficiency causes atrial electrical dysfunction and structural deterioration; conversely, sustained Cisd2 levels protect Cisd2TG mice against age-related atrial myopathy. Thirdly, Cisd2 plays a vital role in maintaining Ca²⁺ homeostasis in atrial cardiomyocytes. Cisd2 deficiency disrupts Ca²⁺ regulation, leading to elevated cytosolic Ca²⁺, reduced SR Ca²⁺, impaired store-operated calcium entry, and mitochondrial Ca²⁺ overload; these compromise mitochondrial function and attenuate antioxidant capability. Finally, transcriptomic analysis reveals that Cisd2 protects the atrium from metabolic reprogramming and preserves into old age a transcriptomic profile resembling a youthful pattern, thereby safeguarding the atrium from age-related injury. This study highlights Cisd2’s crucial role in preventing atrial aging and underscores the therapeutic potential of targeting Cisd2 when combating age-associated atrial dysfunction, which may lead to the development of strategies for improving cardiac health in aging populations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-025-02377-8.

Introduction

During cardiac aging, atrial myopathy occurs and can lead to blood flow stasis, endothelial and endocardial dysfunction, and a hypercoagulable state [45]; these form the Virchow’s triad [53]. Atrial myopathy, which is defined by structural and functional changes present in the atrial myocardium, is strongly associated with calcium (Ca²⁺) homeostasis, especially in the context of age-associated atrial fibrillation (AF). Atrial myopathy is increasingly recognized as a significant contributor to cardiovascular disease. The most common consequence of atrial myopathy during aging is AF, which is predicted to affect 6–12 million people in the USA by 2050 [7] and 17.9 million in Europe by 2060 [26]. However, asymptomatic atrial myopathy usually remains undiagnosed until clinical arrhythmia occurs in the form of AF, atrial tachycardia, or sinus node dysfunction [25]. Previous studies have indicated that aging is associated with atrial remodeling, which can be attributable to atrial dilatation [51], inflammation [23], oxidative stress [24, 51], and aging [30]. Atrial remodeling is associated with intracellular Ca²⁺ dyshomeostasis and a loss of cardiomyocytic sarcomeres, an accumulation of glycogen granules, and an increase in interstitial fibrosis [4]. Currently, clinical diagnostic tools for the identification of early stage of atrial myopathy remain an unmet need [14]. Early recognition and evaluation of active and ongoing detrimental atrial processes, such as fibrosis, inflammation, angiogenesis, apoptosis and thrombosis, are urgently needed to improve atrial myopathy diagnosis, management and intervention. In addition, AF is associated with a five-fold increase in risk of stroke, which is one of the leading causes of mortality and morbidity globally [14, 45].

It is noteworthy that aging triggers significant alterations in Ca²⁺ regulation within atrial cardiomyocytes, leading to a decreased Ca²⁺ transient amplitude and a lower Ca²⁺ current density, both of which are essential for effective myocardial contraction [20]. These changes impair myocardial relaxation and increase susceptibility to AF. Dysregulated Ca²⁺ homeostasis thus can result in incomplete diastolic filling and reduced atrial contractility. Additionally, age-associated atrial interstitial fibrosis and cardiomyocyte hypertrophy can further worsen these Ca²⁺ handling abnormalities, thus creating a favorable environment for the onset of AF [3]. The cumulative impact of these age-associated changes not only predisposes older adults to AF, but also complicates treatment, as the aging heart has a reduced ability to respond to pharmacological interventions that are aimed at restoring the Ca²⁺ balance [20].

Beyond aging, AF itself can complicates the relationship between atrial myopathy and Ca²⁺ homeostasis. AF is associated with abnormal Ca²⁺ signaling, including increased spontaneous Ca²⁺ release from the sarcoplasmic reticulum (SR), which can lead to arrhythmogenic foci and sustain the arrhythmia [28]. This dysregulation is often accompanied by the upregulation of Ca²⁺-calmodulin-dependent protein kinase II (CaMKII), a key player in the pathological remodeling of atrial tissue [9]. The interaction between AF and CaMKII dysregulation creates a self-reinforcing cycle, wherein AF exacerbates atrial myopathy, which leads to further disturbances in Ca²⁺ handling, and this increases the risk of AF recurrence in turn [34]. However, the molecular mechanisms underlying the pathogenesis of AF remain incompletely understood. Thus, to obtaining greater insight into the molecular pathogenesis, is crucial for the development of therapeutic strategies that target both atrial myopathy and Ca²⁺ dysregulation in an attempt to effectively mitigate AF in the aging population, which is a major challenge globally.

The CDGSH iron-sulfur domain-containing protein 2 (CISD2) is a pro-longevity gene in mammals. Our previous studies revealed that Cisd2 mediates lifespan in mice. In Cisd2 knockout (Cisd2KO) mice, Cisd2 deficiency shortens lifespan and causes premature aging. Conversely, in CIsd2 transgenic (Cisd2TG) mice, Cisd2 overexpression delays aging and promotes a healthy longevity [48]. Cisd2 protein is an integral membrane protein localized to the mitochondrial outer membrane, to the ER, and to the mitochondria-associated ER membranes (MAMs) [48]. The ER is the major intracellular Ca²⁺ store and is able to respond to various signals that result in the mobilization of Ca²⁺ transportation between the cytosol and the ER. Furthermore, it is directly involved in extracellular Ca²⁺ influx via a number of highly regulated Ca²⁺ channels. Furthermore, the MAMs are a dynamic structure where ER and mitochondria are reversibly tethered together and this provides a platform that allows rapid exchange of biological molecules, including Ca²⁺, in order to maintain cellular health. Indeed, recent studies have highlighted the critical role of Cisd2 in regulating Ca²⁺ homeostasis. Importantly, Cisd2 directly interacts with sarcoendoplasmic reticulum Ca²⁺ ATPase 2 (Serca2) and regulates its enzymatic activity via the modulation of the redox status of the Serca2 protein in hepatocytes [47] and cardiomyocytes [58]. Serca2 is a Ca²⁺ pump with its ATPase domain being responsible for pumping Ca²⁺ from the cytosol into the ER lumen in order to maintain the ER Ca²⁺ reservoir and regulate cytosolic Ca²⁺-dependent processes. This regulation of Ca²⁺ homeostasis by Cisd2 is vital for preserving mitochondrial integrity and functioning, as dysregulated Ca²⁺ levels contribute to increased oxidative stress and various pathological conditions, including fatty liver diseases and cardiovascular dysfunction [44, 47, 58].

Furthermore, the importance of Cisd2 in maintaining Ca²⁺ regulation is further underscored by its role in cardiac aging [44, 58]; this is particularly relevant to atrial myopathy and AF. Using genetically engineered mouse models, we have demonstrated that in Cisd2KO mice, Cisd2 deficiency accelerates cardiac aging and results in premature electro-mechanical dysfunction of the ventricle. By way of contrast, maintaining a higher level of Cisd2 in Cisd2TG mice protects against age-related structural defects and the functional decline of the heart. In addition, in old Cisd2TG mice, it results in a younger ventricular transcriptome pattern [58]. These suggest that Cisd2 may also play a crucial role in protecting against age-associated atrial myopathy and AF during natural aging.

In this study, we use Cisd2KO and Cisd2TG mice to perform loss-of-function and gain-of-function studies, respectively, in order to further investigate the pathophysiological mechanisms underlying atrial myopathy and AF during aging.

Results

Reduced CISD2 expression is associated with age-related atrial fibrillation (AF) and atrial myopathy in humans

We retrospectively analyzed electrocardiogram (ECG) data from 2,677 participants recruited as part of the Northeastern Taiwan Community Medicine Research Cohort (NTCMRC, ClinicalTrials.gov Identifier: NCT04839796) at Chang Gung Memorial Hospital, Keelung. The mean age of male (n = 1,914) and female (n = 763) patients was 62.7 ± 12.4 (21–96) and 67.4 ± 12.6 (21–100) years old, respectively. The cohort’s median PR interval in male and female patients was 166 ms (interquartile range, 150–182 ms) and 162 ms (interquartile range, 146–178 ms), respectively. The prolongation of PR interval, which is a marker of atrial remodeling, was significantly correlated with age for both sexes (Supplementary Figure S1). The prolongation of the PR interval, which reflects changes in atrio-ventricular node conduction and remodeling [32, 36] and is also recognized as an indicator of cardiac aging and atrial remodeling [57], was significantly correlated with age for both sexes (Supplementary Figure S1). Among the 2,677 retrospectively enrolled participants in the first group, both the mean (Fig. 1A) and coefficient of variation (CV) (Fig. 1B) of the PR interval significantly increased for both sexes from the age of 40, with a progressive rise across each subsequent decade. Similarly, the percentage of AF and atrial flutter also increased with each decade of age (Fig. 1C). The basic preoperative characteristics of the 23 patients from the second group, namely those who underwent cardiac surgery and provided atrial tissue samples for analysis are summarized in Supplementary Table S1. Interestingly, the expression level of CISD2 protein in the atrium is negatively correlated with age for both sexes (R² = 0.5669; Fig. 1D-E).

Fig. 1.

Age-dependent decrease of atrial CISD2 protein level is associated with PR interval prolongation and atrial arrhythmias in humans. In each decade of age, from 20’s until the 90’s, the average of PR intervals (A), the coefficient of variance of the PR intervals (B), and the prevalence of atrial fibrillation or flutter (C) were found to significantly increase during aging in both male and female patients. (D) and (E), Western blot analyses and quantification of CISD2 in human atrial specimens revealed that there was a decrease of atrial CISD2 level and this was significantly associated with aging. (F) Representative IF images of human atrial sections stained with antibodies against Cx43 (green) to localize gap junctions, against pan-cadherin (red) to localize the intercalated discs, and against α-actinin (purple) to stain muscle fibers in male and female patients with high and low CISD2 expression. The sections were also strained with Hoechst (blue) to identify nuclei. (G) Colocalization coefficients of gap junctions (Cx43) and intercalated discs (pan-cadherin) were analyzed using Pearson’s correlation. (H) Representative IF images of human atrial sections stained with antibodies against desmoplakin (green) to localize desmosome, against pan-cadherin (red) to localize the intercalated discs, and against α-actinin (purple) to stain muscle fibers. The sections were also strained with Hoechst (blue) to identify nuclei. (I) Colocalization coefficients of desmosome (desmoplakin) and intercalated discs (pan-cadherin) were analyzed using Pearson’s correlation. The corresponding lower power photomicrographs of (F) and (H) are presented in Supplementary Figure S2A and Supplementary Figure S2B

To study if there is atrial remodeling associated with structural alterations, we performed immunofluorescence (IF) staining of human atrial sections to examine the integrity of the intercalated discs (ICDs) that connects neighboring cardiomyocytes and facilitates the propagation of electrical impulses [40]. Both gap junctions and desmosomes form the major cell junctions that make up the ICDs. Notably, in patients (e.g. male ID12 and female ID05) with higher levels of atrial CISD2, the gap junctions (visualized by connexin 43 staining) were highly colocalized with the ICDs (visualized by pan-cadherin staining); however, the colocalization coefficient decreased as the level of CISD2 decreased (Fig. 1F-G; Supplementary Figure S2A). Previously a genome-wide association study suggested that the integrity of desmosomes is associated with PR prolongation [36]. Accordingly, we analyzed the relationship between the level of CISD2 expression and the colocalization coefficient of desmosomes (visualized by desmoplakin staining) with the ICDs (visualized by pan-cadherin staining). Interestingly, there is a positive correlation between CISD2 protein level and the colocalization coefficient (Fig. 1H-I; Supplementary Figure S2B). In addition, IF staining of α-actinin revealed an overt degeneration of muscle fibers in patients with a lower level of CISD2 expression (Fig. 1F and H). Together, these results suggest that down-regulation of atrial CISD2 expression may contribute to age-associated structural disorganization of ICDs and the degeneration of muscle fibers in the atrium of aging humans (Fig. 1F and H).

Mouse Cisd2 deficiency causes atrial electrical dysfunction and structural deterioration; conversely, a high level of Cisd2 protects against age-associated atrial myopathy

The expression of Cisd2 is down-regulated in an age-dependent manner in various different tissues of WT mice during aging [48]. Strikingly, in the atria, the level of Cisd2 protein is significantly decreased in aging WT mice at 24-month old (24-mo) compared to young WT mice at 3-mo; only approximately 36% of Cisd2 protein remains at 24-mo (Fig. 2A-B). With respect to atrial functioning, ECG analysis revealed irregular and prolonged PR intervals and these were accompanied by atrioventricular block at 24-mo, a pattern that was also observed in the ECG analysis of Cisd2KO mice at 3-mo (Fig. 2C-E). Notably, both the average and the CV of PR intervals in 3-mo Cisd2KO and 24-mo WT mice were significantly increased compared to 3-mo WT mice (Fig. 2G-H). The left atrial area, normalized against body weight, was significantly increased in both Cisd2KO and aged WT mice (22-mo) compared to young WT controls (3-mo to 4-mo), which is indicative of left atrial enlargement (Supplementary Figure S3). These findings suggest that a reduced level of Cisd2 may contribute to age-associated atrial dysrhythmia. In contrast, maintaining a consistently high level of Cisd2 during aging, as observed in Cisd2TG mice (Fig. 2A), significantly mitigated the prolongation of both PR intervals and CV of PR intervals (Fig. 2F-H). Furthermore, a significant decrease in the co-localization of Cx43 and desmoplakin with ICDs was detected in the naturally aged WT mice (24-mo) and the prematurely aged Cisd2KO mice (3-mo); on the other hand, a high level of Cisd2 effectively preserved the normal pattern of Cx43 and desmoplakin co-localization with ICDs in the Cisd2TG mice when old (24-mo) (Fig. 2I-L; Supplementary Figure S4). These findings suggested that Cisd2 is essential to maintaining the structural integrity and the electrical functioning of the mouse atrium. Thus, the age-associated phenotypic abnormalities associated with atrial myopathy appear to be similar in humans (Fig. 1) and in mice (Fig. 2).

Fig. 2.

Cisd2 deficiency in the atrium results in functional and structural deterioration, while a high level of Cisd2 ameliorates age-associated atrial myopathy. (A) and (B), Western blot analyses and quantification of atrial Cisd2 protein levels. (C)-(F), Representative waterfall plots and ECG tracings obtained from 3-mo WT (C), 3-mo Cisd2KO (D), 24-mo WT (E), and 24-mo Cisd2TG (F) mice. Representative atrial dysrhythmic ECGs, namely irregular and prolonged PR intervals and AV block, were found in the 3-mo Cisd2KO and 24-mo WT mice. (G) and (H), Quantification of the average and the CV (coefficient of variance) of ECG PR intervals in various mouse groups. (I) and (K), Representative IF images of atrial sections stained with antibodies against Cx43 (green) to localize gap junctions in (I) and against desmoplakin (green) to localize desmosome in (K). The sections were also stained with antibodies against pan-cadherin (red) to localize the ICDs, and with antibodies against α-actinin (purple) to stain muscle fibers. In addition, the sections were strained with Hoechst (blue) to identify nuclei. (J) and (L), Colocalization coefficient of gap junctions (Cx43) (J) and desmosomes (desmoplakin) (L) with ICDs (pan-cadherin) was analyzed using Pearson’s correlation. The data are presented as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001 by one-way ANOVA with Bonferroni multiple comparison test. The corresponding lower power photomicrographs of (I) and (K) are presented in Supplementary Figure S4A and Supplementary Figure S4B

CISD2 protects the atrium from age-associated ultrastructural deterioration in both humans and mice

To investigate the morphological and compositional properties of aging atrial tissue, transmission electron microscopy was employed to examine the ultrastructural organization of the atrium in order to identify potential defects associated with aging. In human atrial tissues (Fig. 3A-D), significant degeneration of cardiac myofibrils was observed, particularly in patients with low CISD2 levels (Fig. 3C-D). The regular organization of myofibrils was disrupted and there were certain regions that exhibited the presence of disintegrated fibers. Moreover, the ICDs in patients with low CISD2 expression displayed a fragmented and disorganized structure. These ultrastructural injuries thus will undermine structural integrity, resulting in an impairment of the atrium, which will compromise overall heart functioning.

Fig. 3.

CISD2 protects the atrial ultrastructure against age-associated injuries in humans and mice. (A) Representative TEM of the right atrium sampled from a patient with high CISD2. Relatively normal ultrastructures are found, namely intact mitochondria, gap junctions, SR and cardiac myofibrils. (B) Schematic presentation of the ultrastructure of the right atrium shown in (A). (C) Representative TEM of the right atrium sampled from a patient with low CISD2. Mitochondria degeneration, lipofuscin accumulation, and necrotic debris from degenerating myofibrils and organelles can be seen and easily identified. (D) Schematic presentation of the ultrastructure of right atrium shown in (C). (E) Representative TEM for the right atrium of a 24-mo WT mouse. Degeneration of mitochondria and myofibrils can be easily observed. The blurring of the Z line and disarrangement of the ICDs resulting in expanded intercellular spaces can also be detected. (F) Schematic presentation of the ultrastructure of the right atrium shown in (E). (G) Representative TEM of the right atrium of a 24-mo Cisd2TG mouse. The high level of Cisd2 preserves the ultrastructure of organelles and ICDs in Cisd2TG mice. (H) Schematic presentation of the ultrastructure of the right atrium shown in (G). Abbreviations: Cap, capillary; Cm, cardiac myofibril; Cmd, cardiac myofibril degeneration; D, desmosome; EIS, expanded intercellular space; F, fibrosis; FA, fascia adherens; GJ, gap junction; d. ICD; degenerated intercalated disk (ICD f., ICD fragmentation); Lf, lipofuscin; IS, intercellular space; M, mitochondria; MD, mitochondrial degeneration; ZL, Z line; ZLb: Z line breakdown

In addition, a significant expansion of the intercellular space within the atrial tissue was observed in humans with low CISD2 levels (Fig. 3C-D), in 24-mo WT mice with low Cisd2 levels (Fig. 3E-F), and in 3-mo Cisd2KO mice with low Cisd2 levels (Supplementary Figure S5). This widening of the intercellular spaces may indicate the onset of fibrosis, which is characterized by the excessive deposition of extracellular matrix proteins, such as collagen; the collagen replacing functional cardiac tissue. Consequently, this fibrotic remodeling stiffens the myocardium, impairing its contractile function. The extent of fibrosis was notably more pronounced in human atrial samples exhibiting low CISD2 expression (Fig. 3C-D). In these tissues, connective tissue components begin to dominate the intercellular spaces, further diminishing the contractile efficiency of the cardiac muscle; this gives rise to the pathological conditions frequently associated with an aging heart.

Furthermore, several structural elements of ICDs exhibit degeneration during the aging process. The desmosomes, which provide mechanical strength by anchoring intermediate filaments between cells (Supplementary Figure S5A and S5C), appear both weakened and less distinct in both low-CISD2 human samples (Fig. 3C-D), 24-mo WT mice (Fig. 3E-F), and 3-mo Cisd2KO mice (Supplementary Figure S5C and S5D). Similarly, the fascia adherens, a critical component of ICDs that links actin filaments, shows signs of degeneration, which further weakens cell-to-cell adhesion. Additionally, gap junctions, which facilitate electrical coupling between cardiac cells, are reduced in number and disorganized, which is likely to impair the synchronous contraction of the myocardium. These structural abnormalities are very prominent when Cisd2KO mice are investigated (Supplementary Figures S5C and S5D). Collectively, these findings highlight that a loss of CISD2 and a reduction in CISD2 level accelerates cardiac aging by promoting the breakdown of essential cellular structures, and this ultimately contributes to a decline in cardiac function.

Cisd2 is essential to maintaining intracellular Ca²⁺ homeostasis and mitochondrial function in HL-1 atrial cardiomyocytes

In order to gain insights into the molecular mechanism by which Cisd2 deficiency results in atrial dysfunction, we used HL-1, which is an atrial cardiomyocyte cell line [8], carrying a Cisd2KO background (Fig. 4A) to study the role of Cisd2 in Ca²⁺ homeostasis (Fig. 4B; Supplementary Figure S6A). Our results revealed that the basal cytosolic level of Ca²⁺ was significantly elevated in Cisd2KO HL-1 cells (Fig. 4C-D). The activity of Serca2, which is involved in the uptake of Ca²⁺ from the cytosol to the SR, was inhibited using thapsigargin (Tag) treatment. This resulted in the peak Ca²⁺ that is released from the SR being significantly decreased in Cisd2KO HL-1 cells (Fig. 4E), indicating a reduction in the storage of SR Ca²⁺. In addition, the store-operated calcium entry (SOCE) was also decreased in Cisd2KO HL-1 cells (Fig. 4F).

Fig. 4.

Cisd2 deficiency disrupts intracellular Ca²⁺ homeostasis, and causes mitochondrial Ca²⁺ overload and dysfunction in HL-1 atrial cardiomyocytes. (A) Western blot analysis of Cisd2 in HL-1 cells carrying a WT, Cisd2KO, and Cisd2RE genetic background. (B) Schematic illustrating intracellular Ca²⁺ regulation via the SR Ca²⁺pump (SERCA) and a Ca²⁺channel (RyR2), as well as the store-operated calcium entry (SOCE)-related channel (Orai1) and various regulators (TRPC1, STIM1 and STIM2) in cardiomyocytes. Cisd2 maintains intracellular Ca²⁺ homeostasis via the modulation of SERCA2 activity. Nifedipine is a Ca²⁺ channel blocker that inhibits voltage-gated Ca²⁺ channels (VGCCs) leading to a decrease in Ca²⁺ influx from extracellular area to the cytosol. SKF-96,365 is a TRPC1 inhibitor that is able to reduce Ca²⁺ influx via an inhibition of SOCE. (C) The levels of cytosol Ca²⁺ in single HL-1 cells were measured by fluorescence microscopy using Fura-2/AM staining. After measuring the basal level of cytosol Ca²⁺ (first 50 s), thapsigargin (Tag) was added to release Ca²⁺ from the SR. (D) Quantification of the basal cytosol Ca²⁺ levels in the HL-1 cells. (E) Quantification of peak Ca²⁺ release from the SR (maximum minus basal) after Tag treatment as in (C). (F) Quantification of the SOCE levels in the HL-1 cells. (G) and (H), Representative confocal imaging and quantification of STIM1 puncta formation in response to SR Ca²⁺ depletion. (I) Carbonyl cyanide m-chlorophenyl hydrazone (CCCP) was added to release Ca²⁺ from the mitochondria. (J) Quantification of peak Ca²⁺ release from mitochondria (maximum minus basal) after CCCP treatment. (K) and (L), Oxygen consumption rates (OCR) of the HL-1 cells. The indicated chemicals (OA, oligomycin A; FCCP, carbonilcyanide p-triflouromethoxyphenylhydrazone; Rot/AA, rotenone/antimycin A), were added sequentially to determine the ATP-coupled respiration rate, the maximal respiration rate (Max), and the non-mitochondrial respiration rate, respectively, using a Seahorse XF^e 24 analyzer. (M) The ROS levels of WT, Cisd2KO and Cisd2RE HL-1 cells after H₂O₂ (0, 10, and 100 µM) treatment were measured by DCF-DA. In (D)-(F) and (J), the data are presented as mean ± SEM of 60 to 76 cells in two independent experiments. In (H), (L) and (M), the data are presented as mean ± SD from at least three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001 by one-way ANOVA with Bonferroni multiple comparison test

In order to study how CISD2 affects SOCE, we tested the effect of nifedipine (a voltage-gated calcium channel blocker) and SKF-96,365 (a store-operated calcium channel inhibitor) (Supplementary Figure S6B-D). Our results revealed that nifedipine had no significant effect on the basal Ca²⁺ levels or SOCE in Cisd2KO HL-1 cells (Supplementary Figure S6E-G), which suggests that there is no involvement of CISD2 in the voltage-dependent Ca²⁺ channels. However, SKF-96,365 was able to remarkably reduce SOCE in WT HL-1 cells (Supplementary Figure S6H), confirming that SOCE contributes significantly to Ca²⁺ homeostasis. Furthermore, stromal-interaction molecule 1 (STIM1), the primary SR Ca²⁺ sensor, forms puncta that initiate SOCE during SR Ca²⁺ depletion. Confocal imaging revealed robust STIM1 puncta formation in WT HL-1 cells upon Tag treatment; however, the number of STIM1 puncta was significantly decreased in Cisd2KO HL-1 cells, indicating that SOCE activation was disrupted (Fig. 4G-H). Additionally, Western blot analyses of SOCE-associated proteins showed a significant decrease in the mature form of TRPC1 and a reduction in STIM2 levels of Cisd2KO HL-1 cells (Supplementary Figure S6I-M). These findings suggest that decreases in these two proteins may also contribute to the SOCE impairment.

Mitochondria and SR are the two key organelles responsible for buffering Ca²⁺ in cardiomyocytes [54]. Mitochondrial Ca²⁺ storage was measured using carbonyl cyanide m-chloro-phenylhydrazone (CCCP) treatment to release mitochondrial Ca²⁺ into cytoplasm (Fig. 4I). Remarkably, the peak Ca²⁺ released from the mitochondria was significantly increased in Cisd2KO HL-1 cells (Fig. 4J). Accumulation of Ca²⁺ in mitochondria occurs through a low-affinity electrogenic mechanism mediated by the mitochondrial Ca²⁺ uniporter [31]. The elevated cytosolic Ca²⁺ caused by Cisd2 deficiency seems likely to be able to activate this uniporter, which will then lead to mitochondrial Ca²⁺ overload, which will damage mitochondrial function. This is confirmed by a significant decrease in oxygen consumption rate (Fig. 4K-L), as well as a significant decrease in antioxidant ability in Cisd2KO HL-1 cells (Fig. 4M).

Importantly, lentivirus-mediated Cisd2 re-expression (RE) in Cisd2KO HL-1 cells, named Cisd2RE cells, showed a recovery from the disruption that affected intracellular Ca²⁺ homeostasis in Cisd2KO HL-1 cells (Fig. 4C-H; Supplementary Figure S6). This resulted in a restoration of mitochondrial function (Fig. 4I-L) and antioxidant ability (Fig. 4M). Collectively, our results reveal that disrupted Ca²⁺ homeostasis and defects in mitochondria are indeed caused by Cisd2 deficiency and that these defects can be rescued by Cisd2 re-expression in Cisd2KO HL-1 cells.

Additionally, RyR2 phosphorylation, specifically at sites S2808 and S2814, plays a crucial role in regulating cardiac Ca²⁺ release. Excessive elevation of RyR2 phosphorylation contributes to RyR2-mediated SR Ca²⁺ leak, which has been implicated in heart failure and arrhythmias [10, 12]. However, our results reveal that Cisd2 deficiency does not enhance RyR2 phosphorylation in HL-1 cells (Supplementary Figure S7A). Accordingly, the increased cytosolic Ca²⁺ level in HL-1 Cisd2KO cells would seem to be mainly caused by down-regulation of SERCA2 activity, rather than RyR2-mediated Ca²⁺ leak. Furthermore, treatment of tetracaine, which inhibits the opening probability of the RyR Ca²⁺ release channel, reduced cytosolic Ca²⁺ concentration in all of the three groups of HL-1 cells, namely WT, Cisd2KO and Cisd2RE. In addition, the Ca²⁺ peak of SR Ca²⁺ release induced by thapsigargin and the subsequent Ca²⁺ influx of SOCE increased significantly in all three groups of HL-1 cells (Supplementary Figure S7B-F). These results are consistent with the fact that there was no difference in the ratio of RyR2 phosphorylation at S2814 to RyR2 total protein level.

Cisd2 slows down atrial aging in Cisd2TG mice by preserving into old age a transcriptomic profile that is similar to a youthful pattern

To understand the molecular mechanism underlying the role of Cisd2 in atrial aging, we performed RNA sequencing (RNA-seq) and pathway analysis using atrial tissue samples obtained from young (3-mo) WT mice, old (25-mo) WT and old (25-mo) Cisd2TG mice. The mRNA levels of 7,176 expressed genes were quantified. Two sets of pairwise analyses to identify differentially expression genes (DEGs) were performed. These were Set-1: 25-mo WT mice vs. 3-mo WT mice, and Set-2: 25-mo Cisd2TG mice vs. 25-mo WT mice. The normalized counts and DEGs were identified using DESeq2 with the Wald test and a false discovery rate (FDR) cutoff of < 0.05 was applied. Set-1 identified DEGs associated with atrial aging; in total 1,478 genes were significantly up-regulated (861 DEGs) and down-regulated (617 DEGs) in old (25-mo) WT mice. Subsequently, using Set-2, the results revealed a significant change in DEG profile compared to Set-1. A total of 219 aging-associated DEGs (219/1,478 = 14.8%; 150 up-regulated and 69 down-regulated DEGs identified in Set-1) were reverted by Cisd2 overexpression in Cisd2TG mice (Supplementary Figure S8A). Principal component analysis (PCA) of these aging-associated 1,478 DEGs revealed a distinctive pattern to the expression profile differences between 25-mo and 3-mo WT mice. Intriguingly, the expression profiles of 25-mo Cisd2TG mice was shifted and seemed to be approaching those of the 3-mo WT mice, which suggests that Cisd2 overexpression appears to be reversing a proportion of the aging-associated DEGs and moving these mice towards a youthful atrial expression pattern in the atrium in old age (Fig. 5A).

Fig. 5.

A persistently high level of Cisd2 preserves a younger transcriptomic profile in the atrium of Cisd2TG mice during old age. (A) PCA analysis of atrial aging-related DEGs (25-mo WT vs. 3-mo WT; FDR < 0.05) in the atrial tissues of 3-mo WT, 25-mo WT and 26-mo Cisd2TG mice. The PCA was performed using MetaboAnalyst v6.0 (https://www.metaboanalyst.ca/). (B) The biological processes of GO annotation of the Cisd2TG-reverted transcriptome changes (219 DEGs). The grouping of the GO annotation was carried out by STRING v11.5 (https://string-db.org/). FDR < 0.05 was used for selection. (C) The grouping of the KEGG pathways of the Cisd2TG–reverted 219 DEGs was carried out using KOBAS-i (http://bioinfo.org/kobas). A p < 0.05 was used for pathway selection. (D) Canonical pathway analysis by IPA of the Cisd2TG-reverted 219 DEGs. A p < 0.05 was used for pathway selection. (E) The mRNA levels of the integrin-FAK pathway-related DEGs. (F) The mRNA levels of the JAK-STAT pathway-related DEGs. (G) The mRNA levels of the GPCR-RAS pathway-related DEGs. (H) The mRNA levels of the PI3K-AKT-NF-κB pathway-related DEGs. (I) The mRNA levels of the AGEs-RAGE pathway-related DEGs. (J) A graphic summary of the DEGs and pathways associated with natural aging in WT mice and long-lived Cisd2TG mice. In the aged WT mice, various cardiac aging-related pathways have an impact on the atrium during natural aging; these, include the GPCR-RAS, Integrin-FAK, JAK, PI3K-AKT-NF-κB and AGEs-RAGE signaling pathways. In the GPCR-RAS pathway, the expression of the upstream regulators of RAS, including Cxcl16, Grk2, Grk5, and Gng11, were increased in the atrium of the aged WT mice. Moreover, the downstream targets of Ras signaling, including Stk4 and Ets1, exhibit increased expression in the atrium of aged WT mice. Stk4 and Ets1 are involved in cell cycle arrest and cell death,, respectively, as well as cellular senescence. In the JAK-PI3K-AKT-NF-κB pathway, key factors, such as Jak1, Jak3, Akt3 and Nfkb2, exhibit enhanced expression in the atrium of aged WT mice. The upregulation of Akt3 and Nfkb2 is able to promote cardiac hypertrophy, inflammation and fibrosis. In the integrin-FAK pathway, the upstream regulators of FAK, including Sparcl1, Itga1 and Itga6, exhibit increased expression in the atrium of aged WT mice. Activated FAK can modulate downstream signaling regulators, including Ras and PI3K, thereby promoting cardiac hypertrophy, remodeling and fibrosis. In the AGEs-RAGE pathway, dysregulation of the downstream targets of RAGE, including Plcd3 and PKCδ, is observed in the atrium of aged WT mice. Elevated PKCδ can increase ROS levels and promote PI3K signaling. These dysregulated signaling pathways collectively contribute to the aging-related pathological alterations in the atrium of aged WT mice. Cisd2 overexpression reverses these dysregulated gene expression profiles, thereby ameliorating the pathological changes in the atrium of old Cisd2TG mice

The 219 aging-associated DEGs that were reverted in Cisd2TG mice were further annotated by Gene Ontology (GO) enrichment and KEGG pathway analyses to elucidate their potential implications regarding cardiac function and aging [1, 29]. Interestingly, functional enrichment analysis of the GO classification (biological processes) revealed that these DEGs are associated with metabolism, cell death & inflammation, stress response, proteostasis and cell adhesion (Fig. 5B; Supplementary Figure S8B-F). Moreover, KEGG pathway analysis indicated that these DEGs are involved in a number of pathways associated with cardiac aging and disease, including adhesion molecules, aging, cardiac disorders (arrhythmogenic and dilated cardiomyopathy), cell death, and inflammation (Fig. 5C). Notably, the AGE (advanced glycation end product)-RAGE (receptor for AGE) signaling pathway, is significantly enriched in the DEGs that are reverted in the Cisd2TG mouse atrium (Fig. 5C); the AGE-RAGE pathway has been previously linked to various cardiac disorders such as AF, cardiac inflammation and oxidative stress [52, 60]. Furthermore, the KEGG pathway analysis indicated that these DEGs are also involved various other important signaling pathways related to cardiac homeostasis, including PI3K-AKT, JAK-STAT, Ras and NF-κB signaling (Fig. 5C) [5, 16, 19, 21, 38, 41, 42]. Additionally, the Ingenuity Pathway Analysis (IPA) of the 219 aging-associated DEGs identified several significant pathways (p < 0.05 and absolute Z score > 1). These formed three major functional groupings: (1) aging & cardiac disorders, (2) signaling pathways-related to AF and hear failure, and (3) immune response & inflammation (Fig. 5D). Remarkably, several signaling pathways, namely focal adhesion kinase (FAK), JAK, PI3K/AKT/NF-κB, and GPCR, are significantly activated in the atria of old WT mice (Fig. 5D); activation of these pathways has been shown to be associated with cardiac hypertrophy, remodeling and fibrosis, cell death and inflammation [5, 18, 19, 35, 42]. The expression levels of the DEGs involved in the above-mentioned pathways, including the Integrin-FAK, JAK, GPCR-RAS, PI3K-AKT-NF-κB and AGEs-RAGE pathways, are presented herein (Fig. 5E-I). Importantly, overexpression of Cisd2 attenuates the age-associated upregulation of these DEGs in the atria of Cisd2TG mice at 25-mo.

In summary, these transcriptomic results reveal that a persistently high level of Cisd2 is able to modulate several signaling pathways related to atrial aging, thereby mitigating age-associated atrial myopathy, which includes cardiac hypertrophy and remodeling, oxidative stress, cell death, inflammation, fibrosis, and senescence (Fig. 5J). Moreover, the transcriptomic findings are consistent with the phenotypic observation of these old mice and the molecular mechanism characterized from Cisd2KO HL-1 atrial cardiomyocytes.

A variety of metabolic pathways, including amino acid and lipid metabolism, oxidative stress and mitochondrial UPR, are disturbed in the atria of Cisd2KO mice

To get insights into the adverse effects of Cisd2 deficiency on the atrium, we performed transcriptomic pathway analysis using atrial tissues obtained from Cisd2KO and WT mice at 3-mo. The pairwise DEG analysis was performed; the normalized counts and DEGs were obtained using DESeq2 (FDR < 0.05; Cisd2KO vs. WT, either up-regulated or down-regulated). This analysis identified a total of 89 DEGs (40 up-regulated and 49 down-regulated) in the atria of Cisd2KO mice (Supplementary Figure S9). PCA analysis of the 89 Cisd2KO-affected DEGs revealed a dramatic difference in the expression profiles between Cisd2KO and WT mice (Fig. 6A). The Cisd2KO-affected DEGs were further annotated by enrichment analysis of KEGG pathways to evaluate the implications of these expression changes in terms of cardiac functioning. Many pathways were identified by KEGG, including proteostasis, metabolism related to NAD⁺, lipid and amino acid, cell death, inflammation and various signaling pathways (Fig. 6B). The IPA analysis of the Cisd2KO-affected DEGs identified several significant pathways (p < 0.05). These formed six major functional groupings: (a) the unfolded protein response (UPR) and redox homeostasis, (b) NAD⁺ and Sirtuin, (c) lipid metabolism, (d) amino acid metabolism, (e) cell death and inflammation, and (f) signaling pathways (Fig. 6C).

Fig. 6.

A variety of metabolic pathways, namely amino acid metabolism, lipid metabolism, NRF2-mediated oxidative stress and mitochondrial UPR, are activated in the atrium of Cisd2KO mice. (A) PCA analysis of Cisd2KO-related DEGs (3-mo Cisd2KO vs. 3-mo WT; FDR < 0.05). The PCA was performed by MetaboAnalyst v6.0 (https://www.metaboanalyst.ca/). (B) The grouping of the KEGG pathways of Cisd2KO-related DEGs was carried out using KOBAS-i (http://bioinfo.org/kobas). A p < 0.05 was used for pathway selection. (C) Canonical pathway analysis by IPA of the Cisd2KO-related transcriptome changes. A p < 0.05 was used for pathway selection. (D) A heatmap illustrating that Cisd2 deficiency affects the expression of a panel of DEGs (Cisd2KO vs. WT, FDR < 0.05) that are involved in the following: (i) amino acid metabolism, (ii) NAD⁺ and Sirtuin-PPARα, (iii) NRF2-mediated oxidative stress response, and (iv) mitochondrial unfolded protein response (mtUPR). (E) A graphic summary of the DEGs and pathways associated with Cisd2 deficiency in the atrium of Cisd2KO mice. In the Cisd2KO mice, various cardiac aging-related biological processes and pathways have an impact on the atrium during premature aging. (i) Amino acid metabolism: the expression of several enzymes involved in valine and isoleucine degradation, including Dbt, Acadsb, Hsd17b10, Aldh6a1 and Aox1, is significantly enhanced in Cisd2KO mice. Additionally, Aldh6a1, an enzyme involved in alanine metabolism, is significantly increased in the Cisd2KO atrium. (ii) The NAD⁺ and Sirtuin-PPARα pathway: Nampt, the enzyme converting nicotinamide (NAM) to nicotinamide mononucleotide (NMN), is significantly increased in Cisd2KO atrium, leading to increases in NAD⁺ production and Sirtuin 1 (Sirt1) activity. Subsequently, Sirt1 modulates its downstream targets by inhibiting Hif1α; this in turn regulates the TCA cycle. In addition, Sirt1 activates PPARα and its downstream targets (Abca1, Acox1 and Gpd1) expression, thereby influencing lipid metabolism in the atrium. (iii) The NRF2-mediated oxidative stress response: Cisd2 deficiency leads to increase of ROS levels and activates NRF2 and its downstream target gene expression, namely Aox1, Dnajc7 and Fom1. The enhanced Aox1 and Fom1 are able to further increase ROS levels, thus establishing a vicious cycle of oxidative stress in the atrium of Cisd2KO mice. (iv) Mitochondrial dysfunction and mtUPR: Cisd2 deficiency causes mitochondrial dysfunction and enhances the expression levels of mtUPR-related chaperones, namely Dnajc7 (Hsp40) and Hspa9 (mtHsp70), in the atrium. Taken together, these changes would seem to contribute to the premature aging phenotype that affects the atrium of Cisd2KO mice

Recent studies have revealed that abnormal metabolism of lipid and branched chain amino acid (BACC; Valine, Leucine and Isoleucine) are associated with cardiac aging and cardiovascular disease [6, 15, 33, 55]. Interestingly, both KEGG and IPA analyses of the Cisd2KO-affected DEGs identified pathways related to fatty acid oxidation and BCAA degradation (Fig. 6B-C). Furthermore, several enzymes involved in the degradation of valine and isoleucine are upregulated in the atrium of Cisd2KO mice (Fig. 6D). Moreover, PPARα and its downstream target genes in lipid metabolism, including Abca1, Acox1 and Gpd1, are significantly enhanced in the Cisd2KO atrium (Fig. 6D). Together, these results suggest that Cisd2 deficiency leads to dysregulation of lipid and amino acid metabolism in the atrium. In addition, several DEGs are involved in aging-related and stress response pathways; these include NAD⁺ metabolism and Sirtuin-PPARα signaling (Nampt, PPARα and Pdk1), mitochondrial UPR (mtUPR) (Hspa9 and Dnajc7) and the NRF2-mediated oxidative stress response (Aox1 and Fmo1). All of these DEGs are upregulated in the atrium of Cisd2KO mice (Fig. 6D), indicating that Cisd2 deficiency leads to oxidative stress, mitochondrial dysfunction and the activation of mtUPR in the atrium.

In summary, these transcriptomic analyses reveal that Cisd2 deficiency results in the dysregulation of a range of aging-related atrial disorders in premature aging Cisd2KO mice. These include dysregulation of amino acid and lipid metabolism, the oxidative stress response, mtUPR and a compensatory enhancement in NAD⁺ metabolism (Fig. 6E). Furthermore, the transcriptomic findings correlate well with the phenotypic and mechanistic analyses carried out on Cisd2KO mice and Cisd2KO HL-1 cardiomyocytes.

Discussion

This study demonstrates that a persistently high level of Cisd2 slows down atrial aging. There are four key findings. Firstly, in both humans and mice, an age-dependent decrease in CISD2 levels can be found in the atrium; this is correlated with various types of age-associated damage, namely degeneration of intercalated discs, of mitochondria, of the sarcoplasmic reticulum (SR) and of myofibrils. Secondly, in Cisd2KO and naturally aged WT mice, Cisd2 deficiency causes atrial electrical dysfunction and structural deterioration; conversely, a sustained level of Cisd2 protects Cisd2TG mice from atrial myopathy during aging. Thirdly, Cisd2 plays a vital role in maintaining Ca²⁺ homeostasis in atrial cardiomyocytes by regulating SR Ca²⁺ uptake from the cytosol, thus maintaining intracellular Ca²⁺ balance. Cisd2 deficiency disrupts Ca²⁺ regulation, leading to elevated basal Ca²⁺ level in the cytosol, a reduction of Ca²⁺ in the SR, impaired SOCE, and an overloading of mitochondria with Ca²⁺. These in turn compromise mitochondrial functioning and reduce antioxidant capability. Finally, transcriptomic analysis reveals that Cisd2 protects the atrium from metabolic reprogramming and preserves in Cisd2TG mice a transcriptomic profile that leans towards a youthful pattern in old age, thereby safeguarding the Cisd2TG atrium from age-related injury. On the other hand, a variety of metabolic pathways, including amino acid and lipid metabolism, oxidative stress and mitochondrial UPR, are significantly disturbed in the atrium of Cisd2KO mice.

ROS plays a critical role in the degradation of ICDs

Cisd2 deficiency results in elevated ROS levels (Fig. 4M), which can then cause oxidative damage to cellular components and lead to cellular senescence and apoptosis [49]. In ICDs, ROS have been shown to disrupt the trafficking and localization of essential proteins like Cx43, which is crucial for maintaining gap junction communication [2]. This oxidative stress-induced mis-localization of Cx43 impairs electrical coupling between cardiomyocytes, thus increasing the likelihood of arrhythmias. Our results show that decreased CISD2 expression disrupts the localization of Cx43 and compromises the integrity of ICDs in the atrium in both the human and mouse model systems. However, maintaining a higher Cisd2 level during aging preserves the integrity of ICDs, thereby supporting proper atrial conductance.

Enhancing Cisd2 to mitigate metabolic reprogramming in the atrium

Metabolic reprogramming is a hallmark of cardiac dysfunction and plays a critical role in its progression. The healthy heart predominantly relies on fatty acid oxidation to meet its energy demands [13, 55]. However, during aging, the heart undergoes significant metabolic shifts that are characterized by a reduced reliance on fatty acid oxidation and a switch toward glycolysis as the primary energy source [13, 55]. This metabolic adaptation, although it can be observed in conditions such as heart failure induced by cardiac ischemia, is inefficient and contributes to an energy deficit, which impairs contractile function in the heart [55]. Intriguingly, a comparative analysis of DEGs and pathways highlights the distinct roles of Cisd2 in aging metabolism and in cardiac metabolism. In the long-lived Cisd2TG mice, the atrium exhibits a younger transcriptome pattern, indicating that Cisd2 has a cardioprotective effect against aging. Overexpression of Cisd2 appears to safeguard the atrium against aging-associated dysregulation of genes and pathways that are associated with metabolism, proteostasis, stress response, cell death and immune response; this promotes longevity and cardiac health. In contrast, in the Cisd2KO and old WT mice, Cisd2 deficiency accelerates aging and affects numerous biological processes within the atrium, including enhanced amino acid metabolism, upregulation of lipid metabolic genes such as PPARα as well as its downstream targets (Abca1, Acox1, and Gpd1), and dysregulation of NAD⁺ metabolism including the Sirtuin-PPARα signaling pathway. Cisd2 deficiency also exacerbates oxidative stress via the activation of NRF2 and its downstream targets (Aox1, Dnajc7, and Fom14). This establishes a vicious cycle of ROS production. Furthermore, the mitochondrial dysfunction present in Cisd2KO mice enhances the expression of mtUPR-related chaperones, namely Dnajc7 and Hspa9, which then contributes to the dysregulation of proteostasis. These findings underscore the critical role of Cisd2 in maintaining cardiac health by regulating various metabolic pathways, controlling oxidative stress responses, and modifying proteostasis; thus overexpression of Cisd2 confers protection against atrial aging, while, on the other hand, Cisd2 deficiency leads to premature atrial abnormalities.

The potential role of ETS1 and STK4 in atrial senescence

Cellular senescence, which is defined as a state of irreversible cell cycle arrest, is recognized as a key factor in the development of AF. Senescent cells accumulate in the aging heart and these contribute to atrial fibrosis and dysfunction [13]. ETS1 and STK4 are two downstream proteins involved in GPCR-RAS signaling (Fig. 5J). They have been implicated in cellular senescence. Although direct evidence linking ETS1 and STK4 to cardiac aging is limited, their general functions may provide insights into the possible mechanisms behind this phenomenon. As shown in Fig. 5J, our transcriptomic study revealed that up-regulation of ETS1 and STK4 may contributes to pathways implicated in cellular senescence, as evidenced by elevated expression of ETS1 and STK4 in the atrium of old WT mice, where Cisd2 has been shown to be significantly decreased. However, Cisd2 overexpression down-regulates the expression of ETS1 and STK4 in Cisd2TG mice in their old age, indicating that novel interventions targeting ETS1 and STK4, perhaps inhibiting their activity, could be a potential strategy to combat cellular senescence.

ETS1, a transcription factor, is down-regulated in long-lived human individuals, suggesting its potential involvement in aging [56]. ETS1 positively regulates ribosomal protein gene (RPG) expression and its knockdown has been shown to reduce RPG expression. This has been shown to alleviate cellular senescence in human dermal fibroblasts and embryonic lung fibroblast cells [56]. On the other hand, STK4, also known as MST1, is a serine/threonine kinase that phosphorylates FOXO1 and facilitates its nuclear translocation. Nuclear FOXO1 promotes SMS1 expression, which is essential for sphingomyelin biogenesis, a process associated with cell proliferation and stemness maintenance. Down-regulation of SMS1 is linked to aging, and STK4 phosphorylation is inhibited by ROR2, a receptor tyrosine kinase. This suggests it might be possible to slow down senescence by modulating STK4 and FOXO1 activity [11].

Potential role of the PI3K/Akt signaling in the atrium

The PI3K/Akt signaling plays a crucial role in a number of cardiovascular diseases, including AF, by significantly impacting atrial function. Activation of the pathway occurs through diverse stimuli including growth factors and G-protein coupled receptors; these initiate the activation of PI3K, which subsequently brings about the activation of Akt [16, 41]. Akt, a serine/threonine protein kinase, is central to the regulation of key cellular functions and various aspects of metabolism [41]. Once activated, Akt phosphorylates downstream targets, including the proteins mTOR, GSK-3β, and FoxO, which are essential for cell growth, proliferation, and survival [16]. The PI3K/Akt pathway plays an integral part in regulating cellular senescence [13, 16]. Previous findings suggest that Akt, a pivotal component of the PI3K/Akt pathway, is able to indirectly influence atrial function by modulating processes such as cellular senescence and fibrosis, which are critical in the pathogenesis of AF. Further investigations are warranted to establish a direct link between PI3K/Akt and specific atrial functions. Research focusing on the role of PI3K/Akt in regulating ion channels, Ca²⁺ handling, and electrical signaling in atrial cardiomyocytes could provide valuable insights into its contribution to atrial function and AF development. Our transcriptomic results have revealed that the PI3K/Akt signaling, which is significantly activated in in the aging atria of old WT mice, may play a role in triggering atrial aging through various biological processes related to cell cycle arrest & cell death, cellular senescence, and inflammation (Fig. 5J). It is noteworthy that PI3K/Akt signaling is significantly down-regulated in the long-lived Cisd2TG mice, which highlights the potential for targeted the suppression of PI3K/Akt signaling in order to mitigate aging-related atrial damage.

Reducing cytosolic Ca²⁺ overload is a potential therapeutic strategy for AF treatment

Recent research has indicated that inhibition of late sodium currents by ranolazine, which is a medication used to treat heart related chest pain, is able to mitigate AF by reducing cytosolic Ca²⁺ overload and prevent mitochondrial dysfunction [2]. The HARMONY trial has also indicated that ranolazine treatment can help to prevent AF initiation as well as maintenance [43]. Studies have demonstrated that ranolazine decreases cytosolic Ca²⁺ concentration in cardiomyocytes during ischemia and enhances Ca²⁺ uptake by the SR, both of which are crucial for effective cardiac function [37]. This modulation of Ca²⁺ dynamics is particularly relevant in the setting of AF, where disrupted Ca²⁺ handling is a key contributor to arrhythmogenesis [17]. Additionally, clinical evidence suggests that ranolazine is able to reduce the frequency of AF episodes, likely due to its effects on Ca²⁺ homeostasis and myocardial electrical stability [17]. The drug’s ability to stabilize cardiac ryanodine receptors, which are critical for Ca²⁺ release from the SR, further supports its role in preventing arrhythmias associated with Ca²⁺ dysregulation [39].

Limitations and perspectives

There are several limitations to this study. Firstly, while we demonstrated an age-related increase in the prevalence of AF, we were unable to quantify the AF burden due to the absence of continuous rhythm monitoring. Both prevalence and burden are clinically relevant in an aging population, as they are independently associated with increased risks of stroke, heart failure, and cognitive decline [22]. Secondly, this study utilized bulk RNA sequencing to investigate transcriptomic changes associated with Cisd2 deficiency in atrial tissue. While bulk RNA sequencing provides valuable insights into overall gene expression alterations, it lacks the resolution to distinguish cell type-specific effects. In particular, the role of Cisd2 in non-myocyte populations, such as atrial fibroblasts, immune cells, and endothelial cells, remains unclear. Indeed, preliminary analysis of publicly available single-cell portals [50] indicates that CISD2 is expressed in various different cardiac cell types present in human atrial tissues (Supplementary Figure S10). However, these current datasets do not allow for the assessment of age-related changes in CISD2 expression across these distinct cell types. Accordingly, single-cell RNA sequencing and other high-resolution transcriptomic approaches would therefore be needed to dissect whether CISD2 has differential effects on different atrial cell types, including those cell types closely associated with pro-fibrotic and inflammatory pathways, both of which are implicated in cardiac aging [27]. Future studies incorporating single-cell RNA sequencing, particularly utilizing human samples, will be critical to elucidate the age-dependent expression levels of CISD2 in different atrial cell types and the precise cellular mechanisms involved in atrial remodeling and dysfunction.

Conclusion

This study highlights the crucial role of Cisd2 in preserving the atrial structure and preventing age-associated functional decline. We demonstrate that the levels of atrial Cisd2 declines during natural aging in both humans and mice. Our mouse model further reveals that Cisd2 deficiency disrupts Ca²⁺ homeostasis, compromises mitochondrial function, and impairs antioxidant capacity to buffer atrial ROS. Accordingly, enhancing Cisd2 by activators, such as hesperetin [46, 59], in order to ameliorate AF via a reduction in cytosolic and mitochondrial Ca²⁺ overload, should provide an alternative and promising therapeutic strategy for treating AF. In conclusion, these insights underscore the therapeutic potential of targeting Cisd2 and its associated pathways as a means of combating age-associated atrial dysfunction thus helping the development of strategies for improving cardiac health in aging populations.

Materials and methods

Human subjects

The study protocol adheres to the ethical guidelines outlined in the 1975 Declaration of Helsinki and received approval from the Institutional Review Board of Chang Gung Medical Foundation (IRB No: 201800686B0, 202000077B0A3 and 201801875A3). Informed consent was obtained from all participants in the study. This study involved two distinct patient groups. The first group consisted of 2,677 participants whose ECG data were retrospectively analyzed as part of the Northeastern Taiwan Community Medicine Research Cohort (NTCMRC; ClinicalTrials.gov Identifier: NCT04839796) at Chang Gung Memorial Hospital, Keelung. The second group consisted of 23 patients who underwent cardiac surgery at Chang Gung Memorial Hospital between January 2019 and December 2021. Atrial tissue specimens were collected from individuals aged 20 years or older who were diagnosed with cardiac disease that required open-heart surgery and who had provided written informed consent. The preoperative clinical characteristics of these patients are summarized in Supplementary Table S1. Patients were excluded if they or their family members refused to participate in the study, if they had been diagnosed with an active infection, if they had chronic renal failure requiring long-term hemodialysis or peritoneal dialysis, or if they had additional cardiac conditions, such as thoracic or abdominal aortic aneurysm, or had received preoperative extracorporeal membrane oxygenation support.

Mice

Cisd2KO and Cisd2TG mice, both having a C57BL/6 background, were housed in a pathogen-free facility under a 12-hour light/dark cycle at 23°C with ad libitum food and water. Protocols were approved by the IACUC of Chang Gung Memorial Hospital (No. 2017103002) and National Yang Ming Chiao Tung University (No. 1040104r).

Electrocardiography (ECG) and echocardiography

Mice underwent ECG recording under 1.5% isoflurane anesthesia using a PowerLab system (ML866, ADInstruments, Colorado Springs, CO). PR intervals from 1,500 beats were analyzed using LabChart software (v7.3.1, ADInstruments), with the statistical comparisons being carried using the Mann-Whitney U test (p < 0.05). To assess left atrial dimensions and area via echocardiography, we employed the VisualSonics Vevo 2100 Imaging System (VisualSonics, Toronto, Ontario, Canada). Left atrial dimensions were measured in the parasternal long-axis view at end-systole, and the left atrial area was calculated from the apical four-chamber view, normalized to body weight.

HL-1 Cisd2KO cell line

The HL-1 cell line, a gift from Prof. William C. Claycomb (Louisiana State University, Baton Rouge, LA, USA), was maintained according to his protocols. The Cisd2 gene was disrupted in HL-1 cells using the CRISPR/Cas9n (D10A) system. Guide RNA (gRNA) sequences for the mouse Cisd2 sequence were designed using Optimized CRISPR Design (http://crispr.mit.edu): gRNA1: GGCAGCGGACGCCGCCG, gRNA2: CGTGGCCCGCATCGTGA, gRNA3: CCGGCGGCAGGATGGTC. The gRNA/Cas9n (D10A) expression vectors, obtained from Dr. Tsai-Yu Tzeng (VYM Genome Research Center, Taiwan), were used to transfect HL-1 cells using PolyJet™ reagent (SignaGen Laboratories). The gRNA1/Cas9n and gRNA3/Cas9n plasmids expressed RFP, while gRNA2/Cas9n plasmid expressed EGFP. Double-positive EGFP/RFP cells were sorted using a BD FACSAria™ IIu flow cytometry system. Sorted cells (5,000-10,000) were seeded into 10-cm dishes and incubated at 37°C with 5% CO₂ until colonies were formed. Colonies were then picked, amplified in 24-well plates, and verified as Cisd2 KO cells by PCR, DNA sequencing, and Western blot analysis.

HL-1 Cisd2RE cell line

A mouse Cisd2 expression vector was packaged into lentivirus by the National RNAi Core Facility, Academia Sinica, Taiwan. Cisd2KO HL-1 cells were seeded the day before infection. After overnight incubation, lentivirus with polybrene was added to the medium the cells were growing in. After 24 hours, the medium was removed, and then the cells were washed twice with PBS. Post-infected cells were selected with puromycin. Recovery times were adjusted based on selection efficiency and subsequent experimental requirements.

Measurement of intracellular Ca²⁺ levels

Cytosolic Ca²⁺ levels were measured at 37°C using the fura-2 fluorescence ratio method. HL-1 cells were loaded with 2 μM fura-2/AM (Invitrogen, San Diego, CA, USA) in Claycomb medium for 30 minutes. ER Ca²⁺ was then depleted using 2 μM thapsigargin (Tag) in Ca²⁺-free buffer for 10 minutes. SOCE was triggered by replacing the buffer with 0–2 mM Ca²⁺ for 5 minutes. Fluorescence excitation (340/380 nm) and emission (510 nm) were monitored using a Polychrome IV monochromator, Olympus IX71 microscope, and an IMAGO CCD camera (Till Photonics, Grafelfing, Germany); this was followed by analysis using TILLvisION 4.0.

Measurement of mitochondrial oxygen consumption rate (OCR)

OCR was measured using an XFe24 Seahorse Analyzer (Seahorse Bioscience, MA, USA). Cardiomyocytes were cultured in gelatin and fibronectin-coated XF24 plates (0.02% gelatin and 5 µg/mL fibronectin), and respiration was recorded following treatment with oligomycin (1 μM), FCCP (1 μM), and rotenone/antimycin A (1 μM each). The results are expressed as pmol/min/μg protein.

Western blotting

Proteins from heart tissue were extracted in RIPA buffer, separated by SDS-PAGE (Bio-Rad), and transferred to PVDF membranes (PerkinElmer). Membranes were probed with various antibodies, namely Cisd2, Gapdh (MAB374, Millipore), α-actinin (A7811, Sigma), connexin 43 (C8093, Sigma), and desmoplakin (25318-1-AP, Proteintech). The proteins were detected using the ECL system (WBKLS0500, Millipore).

Immunofluorescence (IF) staining and colocalization coefficients

Optimal cutting temperature-embedded cryosections were stained with antibodies against pan-cadherin (C3678, Sigma), desmoplakin (CBL173, Millipore), connexin 43 (C8093, Sigma), and α-actinin (A7811, Sigma). Additional staining consisted of Hoechst to detect nuclei. Colocalization of connexin 43, desmoplakin, and pan-cadherin at intercalated discs were analyzed using Image J’s Pearson–Spearman correlation plugin.

Transmission electron microscopy (TEM)

Heart tissue samples were fixed in glutaraldehyde/paraformaldehyde, post-fixed with osmium tetroxide and potassium hexanoferrate, and then embedded in Epon (EMS, USA) for sectioning. TEM was used to assess ultrastructural abnormalities, including mitochondrial degeneration, disruption of the fascia adherens, and fragmentation of the gap junctions.

RNA isolation, sequencing, and transcriptomic analysis

To obtain sufficient RNA for RNA sequencing, we combined four atrial tissue samples (left and right) from two mice into one tube for all of the groups of mice. Three independent groups of mouse atrial tissue samples were collected in this study. Atrial tissue samples were homogenized using needles and total RNA were extracted using a TRI Reagent (Sigma, T9424); this was followed by phenol and chloroform purification. RNA-seq and pathway analyses were conducted as previously described [46]. Briefly, RNA-seq was performed by the Genomics Center for Clinical and Biotechnological Applications, National Yang Ming Chiao Tung University using single-end sequencing with a depth of at least 20 million reads per sample. After mapping, unique gene reads were analyzed as expected counts in order to assess gene expression. A total of 7,176 genes were analyzed after filtering to identify genes expressed in the mouse atrium (minimal expected counts >400; and detected in at least 50% samples). Normalized counts and DEGs were obtained using DESeq2 with the Wald test, and the false discovery rate (FDR) cutoff was controlled to be below 0.05. Enrichment analyses included Gene Ontology (GO) biological process annotation and KEGG pathway analysis; these were conducted using the online tools STRING (https://string-db.org) and KOBAS-i (http://bioinfo.org/kobas), respectively. Canonical pathway analysis was performed using QIAGEN Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems®, www.ingenuity.com). The normalized counts were transformed into z-scores (normalized counts minus mean and divided by the standard deviation [SD]) and these z-scores were used to generate heatmaps using Multi Experiment Viewer 4.9 software (mev.tm4.org).

Statistical analysis

Data are presented as mean ± standard deviation (SD) or mean ± standard error of the mean (SEM) as indicated in the figure legends. Comparisons between two groups were performed using unpaired two-tailed Student’s t tests. Comparisons among multiple groups were conducted using one-way ANOVA with the Bonferroni multiple comparison test. Statistical significance was defined as p < 0.05. All statistical analyses were conducted using GraphPad Prism software (v9.0, GraphPad Software, San Diego, CA, USA).

Abbreviations

AGEs

advanced glycation end products

ARV

Arrhythmogenic right ventricular

FAK

Focal adhesion kinase

GPCR

G-protein-coupled receptors

JAK

Janus kinases

RAGE

receptor for advanced glycation end products

ROS

reactive oxygen species

Authors’ contributions

CHY co-designed the study, conducted the human and mouse phenotypic examinations and drafted the manuscript. ZQS conducted the Western blotting and mitochondrial experiments, performed transcriptomic pathway analysis and drafted a portion of the manuscript. LHC contributed to cell culture and the Ca²⁺ image experiments. CS contributed to mouse breeding and the IF image experiments. TYT performed the CRISPR-mediated Cisd2KO in HL-1 cells. CYT contributed to transcriptomic analysis. WTC co-designed, conducted and interpreted the Ca²⁺ study in HL-1 cells and drafted a portion of the manuscript. CHK co-designed, conducted and interpreted the TEM results and drafted a portion of the manuscript. TFT designed the experiments, analyzed and interpreted the results, and wrote the final manuscript. All authors have read and approved the final version of the manuscript, and this has been ensured to be the case.

Funding

The work was supported by grants from the National Science and Technology Council (MOST110-2314-B-182 A-113-MY3 and NSTC 113-2314-B-182 A-067-MY3 to CHY; MOST 110-2320-B-A49A-529-MY3 and NSTC 112-2320-B-A49-011-MY3 to TFT) and from the National Health Research Institutes (NHRI-EX112-11239SI to CHY; NHRI-11A1-CG-CO-07-2225-1, NHRI-12A1-CG-CO-07-2225-1, NHRI-13A1-CG-CO-07-2225-1 and NHRI-14A1-CG-CO-07-2225-1 to TFT), and from Chang Gung Memorial Hospital (CMRPG3K2241 to CHY).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.