Research ArticleCardiovascular Biology

Hypoxia-induced interaction of filamin with Drp1 causes mitochondrial hyperfission–associated myocardial senescence

See allHide authors and affiliations

Sci. Signal.  13 Nov 2018:
Vol. 11, Issue 556, eaat5185
DOI: 10.1126/scisignal.aat5185

Breaking mitochondria and hearts

Blocking the excessive mitochondrial fission mediated by Drp1 that occurs after myocardial infarction prevents the development of cardiac dysfunction. Nishimura et al. (see also the Focus by Boyer and Eguchi) sought to understand why Drp1 activity is increased after myocardial infarction, an unfortunately common cause of mortality and morbidity in developed nations. They found that the cytoskeletal regulator filamin increased the mitochondrial fission activity of Drp1, an effect that led to cardiomyocyte senescence. The association of filamin with Drp1, which was actin dependent, was increased in infarcted heart tissue from mice. Screening identified the antihypertensive drug cilnidipine, initially characterized as a Ca2+ channel blocker, as a small-molecule inhibitor of the interaction between filamin and Drp1. Administering cilnidipine to mice after the induction of myocardial infarction reduced mitochondrial fission, cardiomyocyte senescence, and myocardial dysfunction independently of the drug’s effects on Ca2+ channels. These results identify a mechanism that increases Drp1 activity after myocardial infarction and a drug that could be repurposed to treat this all-too-common event.


Defective mitochondrial dynamics through aberrant interactions between mitochondria and actin cytoskeleton is increasingly recognized as a key determinant of cardiac fragility after myocardial infarction (MI). Dynamin-related protein 1 (Drp1), a mitochondrial fission–accelerating factor, is activated locally at the fission site through interactions with actin. Here, we report that the actin-binding protein filamin A acted as a guanine nucleotide exchange factor for Drp1 and mediated mitochondrial fission–associated myocardial senescence in mice after MI. In peri-infarct regions characterized by mitochondrial hyperfission and associated with myocardial senescence, filamin A colocalized with Drp1 around mitochondria. Hypoxic stress induced the interaction of filamin A with the GTPase domain of Drp1 and increased Drp1 activity in an actin-binding–dependent manner in rat cardiomyocytes. Expression of the A1545T filamin mutant, which potentiates actin aggregation, promoted mitochondrial hyperfission under normoxia. Furthermore, pharmacological perturbation of the Drp1–filamin A interaction by cilnidipine suppressed mitochondrial hyperfission–associated myocardial senescence and heart failure after MI. Together, these data demonstrate that Drp1 association with filamin and the actin cytoskeleton contributes to cardiac fragility after MI and suggests a potential repurposing of cilnidipine, as well as provides a starting point for innovative Drp1 inhibitor development.


Mitochondria are dynamic organelles that interchange their morphology between two distinct arrangements by undergoing fission to generate an elongated interconnected mitochondrial network or fusion to give rise to a fragmented discrete phenotype (1). Although mitochondrial fission and fusion networks are absent and mobility is not normally observed in healthy adult cardiomyocytes, mitochondrial fusion and fission proteins are abundant and indispensable for physiological functions of the heart (2). Defects in mitochondrial dynamics are implicated in the development of cardiovascular diseases (3), as well as neurodegenerative diseases. Although many studies using genetically manipulated mice and/or chemical reagents have highlighted the potential of mitochondrial fusion and fission proteins as therapeutic targets for combating cardiovascular diseases (49), drugs targeting these proteins have not been developed, under the assumption that direct inhibition of fission or fusion proteins will eventually impair the balance between mitochondrial fission and fusion (10).

Originally defined as terminal cell growth arrest, cellular senescence is now appreciated to participate in various biological processes such as development, aging, and age-related diseases (11). Cells undergoing senescence show distinctive functional alterations including telomere shortening, secretome changes, increases in senescence-associated β-galactosidase (SA-β-gal) activity, and changes in the expression of various genes. Although cardiomyocytes are terminally differentiated, SA-β-gal–positive cardiomyocytes are increased in aged heart (12). Senescent cardiomyocytes show decreased contractile performance due to altered abundances of contractile proteins and Ca2+-handing proteins, leading to maladaptive changes in myocardial structure and function (13). Moreover, p53, a key mediator of senescence, accumulates in aged heart and promotes mitochondrial dysfunction by inhibiting Parkin-mediated mitophagy, leading to cardiac dysfunction (14). Premature myocardial senescence, which is observed in several cardiac disease models including myocardial infarction (MI), has been suggested as a major cause of cardiac dysfunction (12, 15, 16). The senescence-associated microRNA (miRNA) miR-34a, which is transcriptionally activated by p53 (17), is increased after MI, and miRNA-34a inhibition reduces DNA damage responses and improves myocardial function through its target phosphatase-1 nuclear targeting subunit (PNUTS) after MI (16). However, the mechanisms underlying defect in mitochondrial dynamics for the induction of premature myocardial senescence are obscure.

Mitochondrial fission is specifically mediated by the dynamin-related guanosine triphosphatase (GTPase) protein 1 (Drp1) (18), which translocates from the cytosol to the outer mitochondrial membrane where it interacts with receptor proteins (19), including fission protein-1 (Fis1), mitochondrial fission factor (Mff), and mitochondrial dynamics proteins of 49 and 51 kDa (MiD49 and MiD51, respectively). Translocated Drp1 oligomerizes and encircles a mitochondrion to constrict the mitochondrial membrane, after which dynamin-2 triggers membrane scission as the final step of mitochondrial division (20). Translocation of Drp1 is regulated by several different posttranslational modifications, including phosphorylation (7), SUMOylation (21), ubiquitination (22), S-nitrosylation (23), and O-GlcNAcylation (24). In addition, we have reported that sulfur deprivation-polysulfidation cycling of cysteine residues in Drp1 regulates its GTPase activity (25). Mitochondrial fusion is also regulated by other mitochondrial membrane–localized GTPases, namely, mitofusins (Mfn1 and Mfn2) in the outer mitochondrial membrane and optic atrophy 1 (Opa1), an inner membrane protein that governs inner mitochondrial membrane fusion (3, 18). These GTPases play important roles in maintaining mitochondrial quality control and the functional integrity of the heart (4, 5, 26, 27).

Drp1 maturation has been suggested to require the interactions of mitochondria with other organelles such as the endoplasmic reticulum (ER) (28), lysosome (29), and cytoskeleton (3033). For example, mitochondria-ER contact precedes Drp1 oligomerization on mitochondrial membranes (28). Cytoskeletal proteins, including myosin II and actin filaments, promote Drp1 assembly at fission sites, leading to efficient mitochondrial fragmentation (3133). However, it remains unclear whether interactions of Drp1 with actin or myosin are sufficient to promote Drp1 activity and assembly at fission sites. It also remains unclear why basal Drp1 activity and mitochondrial fission are not enhanced in actin- and myosin-enriched muscle cells and actin-reorganized cells compared with nonmuscle cell lines. Cells may express a protein that anchors Drp1 to the actomyosin cytoskeleton under pathological conditions. In this study, we identified the actin-binding protein filamin A (FLNa) as a binding partner of Drp1 and demonstrated that FLNa acted as a guanine nucleotide exchange factor (GEF) for Drp1 in coordination with actin cytoskeleton. FLNa overexpression promoted protein complex formation between Drp1 and the actin cytoskeleton at mitochondrial fission sites, leading to progression of hypoxia-induced mitochondrial hyperfission in cardiomyocytes and fibroblasts. Mitochondrial fusion and fission proteins have been suggested to be potential therapeutic targets for combating cardiovascular disease, but clinically applicable drugs have not been identified to date. In this study, we identified cilnidipine as a specific inhibitor of hypoxia-induced mitochondrial hyperfission, a drug initially characterized as a dihydropyridine (DHP)–derivative voltage-dependent L/N-type Ca2+ channel blocker. Using this drug, we demonstrated that cilnidipine disrupted hypoxia-induced Drp1-filamin interaction and improved cardiac function after MI and suggested the potential repurposing of cilnidipine for the treatment of ischemic diseases.


Premature myocardial senescence associated with Drp1-dependent mitochondrial fission is induced in the peri-infarct region of mouse hearts after MI

We have previously reported that formation of reactive oxygen species (ROS)–dependent electrophiles, such as 8-nitroguanosine 3′5′-cyclic monophosphate (8-NO2-cGMP), mediates H-Ras–dependent myocardial early senescence at the late stage of heart failure after MI (34). Here, we identified the region of the left ventricular (LV) myocardium that was the predominant cause of myocardial senescence after MI. The SA-β-gal–positive cell area was significantly increased in size the infarct zone and peri-infarct zone of the LV myocardium at 4 weeks after MI (Fig. 1A and fig. S1A). To identify the molecular event upstream of cardiac senescence, we analyzed MI heart samples at an earlier time point. Sarcomere structures in the infarct zone were almost absent, and myocardium was eventually replaced by myofibroblasts by 1 week after MI. In contrast, sarcomere structures in the peri-infarct zone were not disrupted, and almost all striated cardiomyocytes were still healthy (Fig. 1B). Electron microscopic analysis also revealed that mitochondria in peri-infarct zone myocardium were smaller and more rounded than those in sham-operated myocardium (Fig. 1, B and C, and fig. S1B). Although blood flow was apparently supplied properly to the peri-infarct zone, the protein abundances of hypoxia-inducible factor 1α (HIF-1α) and heme oxygenase-1 (HO-1) were significantly increased in the peri-infarct zone myocardium at 1 but not 4 weeks after MI (fig. S2A), indicating that hypoxic stress was transiently induced at 1 week after MI. Because the abundances of mitochondrial proteins, such as Mn-superoxide dismutase and tafazzin, did not change (fig. S2A), moderate hypoxic stress may cause mitochondrial fragmentation but not decreased mitochondrial protein quantity. Hypoxic stress increases Drp1 activity (8, 35), and accordingly, the GTP-binding activity of Drp1 was significantly increased, although Drp1 protein abundance was slightly reduced in the peri-infarct zone myocardium at 1 week after MI (Fig. 1D). Exposing neonatal rat cardiomyocytes (NRCMs) to 1% hypoxia significantly increased the GTP-binding activity of Drp1 in vitro (Fig. 1E). The GTP-binding activities of the fusion-accelerating proteins Mfn1, Mfn2, and Opa1 were not changed in the peri-infarct zone myocardium compared with sham-operated myocardium (fig. S2B) or NRCMs with hypoxia (fig. S2C). Consistent with previous report about reduced Opa1 protein levels in human heart with ischemic cardiomyopathy (36), Opa1 protein abundance was slightly reduced after MI (fig. S2B). These results suggest that Drp1 activity is the main determinant of mitochondrial morphology after MI. Depending on the mitochondrial fragment length, we classified the mitochondrial morphology of NRCMs into three forms: tubule, intermediate, and vesicle (fig. S3, A and B). Consistent with Drp1 activation (Fig. 1E), hypoxia treatment significantly increased the number of NRCMs with vesicle-type mitochondrial fragmentation compared to those with tubule-type fragmentation, which was the predominant mitochondrial morphology in normoxic NRCMs (Fig. 1F). Hypoxic stress was reduced at 4 weeks after MI (fig. S2A), which is also when myocardial senescence occurred (Fig. 1A). ROS are critical to inducing cellular senescence (37), and hypoxia/reoxygenation but not hypoxia alone increased ROS production in NRCMs (fig. S4A). Hypoxia itself did not induce NRCM senescence (fig. S4B), whereas hypoxia/reoxygenation treatment significantly increased the number of SA-β-gal–positive cardiomyocytes (Fig. 1G). We also used p53 as an alternative senescence marker because cytosolic p53 is increased in abundance in aged hearts (14). The population of cells containing p53 puncta in α-actinin–positive NRCMs was significantly increased in cells subjected to hypoxia/reoxygenation (fig. S4C). Hypoxia/reoxygenation-induced myocardial senescence was attenuated by application of the mitochondria-targeted superoxide scavenger (2-(2,2,6,6-tetramethylpiperidin-1-oxyl-4-ylamino)-2-oxoethyl)triphenylphosphonium chloride (mito-TEMPO) (fig. S4D). These results suggest that hypoxia-induced mitochondrial fission after mitochondrial ROS production is involved in the induction of senescence, although it remains unclear whether this cascade is sufficient for senescence. Knockdown of Drp1 by one of two independent siRNAs (fig. S5A) prevented hypoxia-induced mitochondrial fragmentation (fig. S5B) and hypoxia/reoxygenation-induced senescence of NRCMs (Fig. 1G). In addition, the hypoxia/reoxygenation-induced cardiomyocyte senescence and mitochondrial ROS production were significantly suppressed by the Drp1 inhibitor Mdivi-1 (fig. S5, C and D) (38). These results suggest that hypoxic stress–induced Drp1 activation participates in mitochondrial fragmentation–associated premature senescence of cardiomyocytes in the peri-infarct zone myocardium after MI.

Fig. 1 Mitochondrial fission–associated premature myocardial senescence in peri-infarct zone myocardium after MI.

(A) Myocardial senescence determined by SA-β-gal staining in mouse heart 4 weeks after MI (n = 5 mice per group). IZ, infarct zone; RZ, remote zone. Scale bars, 500 μm (left panels) and 50 μm (middle and right panels). (B) Representative electron micrographs of peri-infarct zone myocardium 1 week after MI depicting two subpopulations of mitochondria: interfibrillar (IF) and perinuclear (PN) mitochondria. N, nucleus (n = 5 mice per group). Scale bars, 1 μm. (C) Quantification of mitochondrial area and circularity (n > 350 or 125 perinuclear or interfibrillar mitochondria from 5 mice per group). (D) Guanosine 5′-triphosphate (GTP)–binding activity and protein abundance of Drp1 in the peri-infarct zone myocardium at 1 week after MI (n = 5 mice per group). GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (E) GTP-binding activity of Drp1 in NRCMs under normoxia (N) or hypoxia (H) for 16 hours (n = 3 independent experiments). (F) Hypoxia-induced change in mitochondrial morphology. Left: Representative images of mitochondrial morphologies of NRCMs loaded with MitoTracker Green. Right: Mitochondrial morphologies of NRCMs classified into three groups: vesicle, intermediate, and tubule (n = 3 independent experiments). Scale bars, 10 μm (white) and 5 μm (yellow). (G) Effects of Drp1 knockdown on hypoxia/reoxygenation (H/R)–induced cardiomyocyte senescence. NRCMs transfected with negative control small interfering RNA (siRNA) (siNC) or two different siRNAs for Drp1 (siDrp1 #1 and #2) were incubated for 16 hours of hypoxia, followed by 12 hours of reoxygenation (n = 3 independent experiments). Left: Representative images of SA-β-gal staining of NRCMs. The arrowheads show SA-β-gal–positive cells. Right: Quantification of SA-β-gal–positive cells. Scale bars, 50 μm. Data are means ± SEM. *P < 0.05, **P < 0.01 by unpaired t test (D to F) or one-way analysis of variance (ANOVA) (G).

Drp1 associates with FLNa under hypoxia

To investigate how Drp1 is activated under hypoxic conditions, we screened for proteins that bound to Drp1 in a hypoxia-dependent manner. Proteomic analyses revealed that endogenous cytoskeletal proteins, such as the actin-binding protein FLNa (39) and β-actin, were cross-linked and coprecipitated with FLAG-tagged Drp1 in HeLa cells and that these interactions were increased after exposure to hypoxic stress (Fig. 2A and fig. S6). To examine whether hypoxic stress induced the direct interaction of Drp1 with these cytoskeletal proteins, we performed immunoprecipitation assays without a cross-linker. Hypoxic stress did not promote the interaction of FLAG-Drp1 with endogenous FLNa but not with endogenous β-actin (Fig. 2B), suggesting that only FLNa directly associates with Drp1. Dephosphorylation of Drp1 at Ser637 participates in mitochondrial fission, and hypoxic stress induces dephosphorylation of Drp1 Ser637 in endothelial progenitor cells (40). We investigated whether the phosphorylation state at Ser637 on Drp1 alters the interaction with FLNa. However, Drp1 phospho-deficient S637A and phospho-mimic S637D mutants associated with FLNa to the same extent (Fig. 2C). Green fluorescent protein (GFP)–Drp1 was mainly localized in the cytoplasm, whereas mCherry-FLNa showed punctate structures in normoxic H9c2 cardiac myoblasts. After hypoxia, GFP-Drp1 formed punctate structures and colocalized with mCherry-FLNa (Fig. 2D). To investigate whether Drp1 and FLNa interaction occurred in mitochondria, we performed subcellular fractionation. The Drp1-FLNa interaction was mainly observed in the mitochondrial fraction (Fig. 2E). Moreover, FLNa was diffusely detected in cardiomyocytes of the hearts of sham-operated mice, whereas Drp1 was detected as punctate dots. In the peri-infarct zone myocardium 4 weeks after MI, the signals for FLNa and Drp1 were increased and were colocalized (Fig. 2F). These results suggest that FLNa binds to Drp1 under hypoxic conditions.

Fig. 2 FLNa is a hypoxia-dependent binding partner of Drp1 in peri-infarct zone myocardium after MI.

(A) Coomassie brilliant blue–stained images of SDS–polyacrylamide gel electrophoresis (PAGE) gels to identify proteins interacting with FLAG-Drp1. HeLa cells expressing FLAG-Drp1 were exposed to normoxia or hypoxia. (B) Hypoxia-dependent interaction of FLAG-Drp1 with endogenous FLNa in HeLa cells. (n = 3 independent experiments). (C) Interaction of Myc-FLNa with FLAG-Drp1 S637D (SD) and S637A (SA) mutants in HeLa cells under hypoxia (n = 3 independent experiments). (D) Colocalization of GFP-Drp1 with mCherry-FLNa in H9c2 cells under normoxia or hypoxia (n = 3 independent experiments). Scale bars, 10 μm. (E) Interaction of Myc-FLNa with FLAG-Drp1 in mitochondrial (Mito) or cytosolic (Cyto) fractions of HeLa cells (n = 3 independent experiments). (F) Representative images and distribution of Drp1 (red) and FLNa (green) in mouse LV myocardium 4 weeks after MI. Lower panels display high magnification images. The boxed regions in peri-infarct zone myocardium of MI-treated heart or sham-treated hearts were magnified and color separated (n = 3 mice per group). Scale bars, 60 μm (white) and 10 μm (yellow).

The GTPase domain of Drp1 interacts with the rod-2 segment of FLNa

Next, we identified the protein domains required for the interaction of Drp1 and FLNa. Drp1 contains a GTPase, middle domain (MID), variable (Var), and GTPase effector domain (GED) domain (41). The GTPase domain of Drp1 was sufficient for the interaction with FLNa (Fig. 3A). FLNa consists of an N-terminal actin-binding domain and 24 immunoglobulin (Ig) repeats. The rod-1 segment (Ig1-15) is the secondary actin-binding region. Mutants lacking the actin-binding regions (Ig1-24 and Ig16-24) are still associated with Drp1 (Fig. 3B). The Ig24 dimerization domain did not associate with Drp1 (fig. S7), indicating that the rod-2 segment (Ig16-23) of FLNa is sufficient for the interaction with Drp1.

Fig. 3 The interacting regions of Drp1 and FLNa.

(A) Interaction of truncated FLAG-Drp1 (GTPase) with endogenous FLNa under hypoxia (n = 3 independent experiments). (B) Interaction of FLAG-Drp1 with actin-binding–deficient Myc-FLNa truncated mutants (Ig1-24 or Ig16-24) under hypoxia (n = 3 independent experiments).

FLNa is indispensable for hypoxia-induced mitochondrial fission

Actin and myosin filaments recruit Drp1 to mitochondrial fission sites and induce ring-like oligomeric assembly of Drp1 (3033), but the mechanism for how Drp1 accumulates in these cytoskeletal filaments remains obscure. Consistent with previous reports, disruption of the actin cytoskeleton by latrunculin B (LatB; Fig. 4A) prevented hypoxia-induced mitochondrial fission in H9c2 myoblasts (Fig. 4B). To examine whether FLNa was required for mitochondrial fission, we tested the effect of FLNa knockdown on hypoxia-induced mitochondrial fission, using two different siRNAs for FLNa (siFLNa #1 and #2; fig. S8A). Each siRNA for FLNa inhibited hypoxia-induced mitochondrial fragmentation (Fig. 4C) in H9c2 cells. In addition, FLNa knockdown prevented hypoxia-induced mitochondrial fission in nonmuscle cardiac fibroblasts (fig. S8B). We evaluated the mitochondrial function of NRCMs by analyzing mitochondrial membrane potential and lactate production. Knockdown of FLNa suppressed not only the hypoxia-induced reduction in mitochondrial membrane potential and lactate production (fig. S8, C and D) but also the hypoxia/reoxygenation-induced senescence of NRCMs (Fig. 4D). Drp1 is mainly localized in the cytosol but forms punctate foci at fission sites during mitochondrial fission (31). FLNa knockdown prevented the targeting of Drp1 punctate structures to mitochondria under hypoxia (Fig. 4E). These results suggest a critical role of FLNa in Drp1-mediated mitochondrial fission under hypoxia in both muscular and nonmuscular cells.

Fig. 4 FLNa is required for hypoxia-mediated Drp1 activation and mitochondrial fission–associated myocardial senescence.

(A) Morphological changes in actin cytoskeleton of H9c2 cells in response to the actin-depolymerizing drug LatB (20 nM, 18 hours). Actin cytoskeleton was labeled with Alexa 594–phalloidin (n = 3 independent experiments). (B) Effects of LatB on hypoxia-induced mitochondrial fission. H9c2 cells were treated with 20 nM LatB for 1 hour before being exposed to hypoxia (n = 3 independent experiments). Scale bars, 50 μm. DMSO, dimethyl sulfoxide. (C) Effect of FLNa knockdown on hypoxia-induced mitochondrial fission. H9c2 cells were treated with negative control siRNA (siNC) or two different siRNAs for FLNa (siFLNa #1 and #2) (n = 3 independent experiments). Scale bars, 50 μm. (D) Effect of FLNa knockdown on hypoxia/reoxygenation-induced accumulation of p53 in NRCMs. The population of senescent cells near p53 punctae in α-actinin–positive NRCMs was quantified (n = 3 independent experiments). Scale bars, 20 μm. (E) Effects of FLNa siRNA on the localization of GFP-Drp1 and mitochondria in cardiac fibroblasts under normoxia or hypoxia. Mitochondria were stained with MitoTracker Red (n = 3 independent experiments). Scale bars, 10 μm. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01 by one-way ANOVA.

FLNa acts as a GEF for Drp1 in coordination with an actin cytoskeleton

Hypoxic stress increases FLNa abundance in neurons (42). The protein abundance of β-actin and FLNa was increased in the peri-infarct zone myocardium after MI (Fig. 5A). To examine whether the increased abundance of these proteins affected Drp1 function, we analyzed the GTP-binding activity of Drp1 in human embryonic kidney–293 (HEK293) cells overexpressing FLNa or β-actin. Overexpression of FLNa promoted the GTP-binding activity of Drp1 (Fig. 5B), whereas β-actin overexpression did not affect Drp1 activation, although actin filaments reportedly promote Drp1 GTPase activity (31). GEFs increase the steady-state GTPase activity of G protein by accelerating guanosine diphosphate (GDP)/GTP exchange. We evaluated the GEF activity of FLNa by measuring the steady-state GTP hydrolysis mediated by Drp1. The GTP hydrolysis of recombinant Drp1 was increased by the addition of EDTA, which enhances spontaneous GDP/GTP exchange, as expected, and by the addition of FLNa purified from mammalian cells (Fig. 5C). In general, GEFs preferentially bind to the GDP-bound form or the nucleotide-free form of G protein. FLNa is associated with the Drp1 K38A mutant, which is deficient in GTP binding (Fig. 5D) (43). These results suggest that FLNa acts as a GEF for Drp1. Drp1 was not activated by the overexpression of the actin-binding–deficient FLNa mutants (Ig1-24 and Ig16-24; Fig. 5E) that could still interact with Drp1 (Fig. 3B), indicating that FLNa mediates actin-dependent Drp1 activation in cells. Consistent with its ability to increase the GTP-binding activity of Drp1, FLNa overexpression induced mitochondrial fission in H9c2 cells (Fig. 5F). Some FLNc protein variants, including A1539T, cause familial hypertrophic cardiomyopathy, and overexpression of FLNc A1539T induces the aggregation of actin filaments (44). Overexpression of the FLNa A1545T mutant, which corresponds to the FLNc A1539T mutant, caused more severe mitochondrial fragmentation than overexpression of wild-type FLNa (Fig. 5F). These results suggest that hypoxia-induced ternary complex formation between Drp1, FLNa, and actin filaments underlies Drp1 activation, followed by mitochondrial fission.

Fig. 5 FLNa drives Drp1 activation and mitochondrial fission in coordination with actin filaments.

(A) Protein abundances of actin and FLNa in peri-infarct zone myocardium at 4 weeks after MI (n = 3 mice per treatment). (B) Effect of FLNa or β-actin overexpression on Drp1 activation in HEK293 cells (n = 3 biological replicates). (C) The time course of GTP hydrolysis by Drp1 (600 nM) in the absence or presence of FLNa (200 nM) or EDTA. Right: Calculated GTP hydrolysis rate of Drp1 (n = 3 independent experiments). (D) Interaction of Myc-FLNa with FLAG-Drp1 WT or K38A (KA) under hypoxia (n = 3 independent experiments). (E) Effect of actin-unbound FLNa mutant on Drp1 activation in HEK293 cells (n = 3 biological replicates). (F) Effect of FLNa overexpression on mitochondrial fission. H9c2 cells expressing mCherry-FLNa wild-type (WT) or A1545T were incubated under normoxia and stained with MitoTracker Green (n = 3 independent experiments). Scale bars, 50 μm. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01 by unpaired t test (A) or one-way ANOVA (B, C, E, and F).

Cilnidipine prevents hypoxia-induced mitochondrial fission by disrupting the Drp1-FLNa interaction

Drp1-mediated mitochondrial fission is an attractive target for combating cardiovascular diseases. Mdivi-1 was originally identified as a pharmacological inhibitor of Drp1 and blocks Drp1-mediated mitochondrial fission (38). Although Mdivi-1 treatment improves cardiac function in several cardiac disease models (79, 4547), the selectivity of Mdivi-1 for Drp1 has been challenged (48, 49). Alternatively, drug repositioning is an attractive strategy for the application of known drugs to new treatment indications while reducing the risk of failure because of adverse toxicology. For these reasons, we screened compounds that inhibit hypoxia-induced mitochondrial fission among antihypertensive drugs approved by the Japanese government. We used cardiac fibroblasts for drug screening because their mitochondrial morphology enables more sensitive evaluation of mitochondrial fission in response to hypoxic stress (compare fig. S9, A and B, to fig. S3, A and B). Moreover, FLNa is involved in hypoxia-induced mitochondrial fission in cardiac fibroblasts (fig. S8B) as it is in H9c2 cells (Fig. 4C). Among the antihypertensive drugs we tested, we found that cilnidipine, an antihypertensive drug with a DHP-based structure that targets voltage-dependent L/N-type Ca2+ channels, specifically suppressed hypoxia-induced mitochondrial fragmentation (Fig. 6A). Cilnidipine also suppressed hypoxia-mediated mitochondrial fission in NRCMs (Fig. 6B). Other drugs, including DHP-derived L-type Ca2+ channel blockers and other antihypertensive drugs, did not affect hypoxia-induced mitochondrial hyperfission, suggesting that pharmacological inhibition of L/N-type Ca2+ channels did not contribute to mitochondrial fission. Cilnidipine also suppressed hypoxia/reoxygenation-induced cardiomyocyte senescence and ROS production (Fig. 6C and figs. S4C and S5C), whereas amlodipine, an L-type Ca2+ channel blocker, did not affect hypoxia/reoxygenation-induced myocardial senescence (Fig. 6C).

Fig. 6 Identification of cilnidipine as an inhibitor of hypoxia-induced mitochondrial fission by disrupting Drp1-FLNa complex.

(A) Effects of antihypertensive drugs on mitochondrial fission. Cardiac fibroblasts were treated with each drug (1 μM) for 1 hour before being exposed to hypoxia. The percentages of cells with vesicle-type mitochondrial fragmentation were quantified (n = 3 to 6 independent experiments). Cilnidipine, CIL. (B) Effects of cilnidipine on hypoxia-mediated mitochondrial fragmentation. Left: Representative images of mitochondrial morphologies of NRCMs loaded with MitoTracker Green. Right: Mitochondrial morphologies of NRCMs treated with cilnidipine (1 μM) for 1 hour before 16 hours of normoxia (N) or hypoxia (H) (n = 3 independent experiments). Scale bars, 10 μm. (C) Effects of cilnidipine on hypoxia/reoxygenation-induced cardiomyocyte senescence. NRCMs were treated with 1 μM cilnidipine or amlodipine (AML) for 1 hour before 16 hours of hypoxia, followed by 12 hours of reoxygenation (n = 5 independent experiments). Scale bars, 20 μm. (D and E) Interaction of FLAG-Drp1 with endogenous FLNa under hypoxia and effects of cilnidipine (D) or Mdivi-1 (E) treatment. HeLa cells expressing FLAG-Drp1 were treated with hypoxia with or without cilnidipine (3 μM) or Mdivi-1 (10 μM) [n = 5 (D) and n = 3 (E) independent experiments]. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01 by one-way ANOVA.

Next, we examined the effect of cilnidipine on hypoxia-induced interaction between Drp1 and FLNa. The hypoxia-induced formation of Drp1-FLNa complexes was suppressed by cilnidipine, but not by Mdivi-1 in HeLa cells (Fig. 6, D and E). HeLa cells do not express voltage-dependent Ca2+ channels, suggesting that this cilnidipine effect is independent of blockage of voltage-dependent L/N-type Ca2+ channels. These results indicate that cilnidipine prevents hypoxia-induced mitochondrial fragmentation by disrupting Drp1-FLNa complex formation.

Cilnidipine improves cardiac function after MI by inhibiting Drp1-mediated myocardial senescence

Ca2+ channel blockers, especially cardiac-specific non-DHP Ca2+ channel blockers, generally exacerbate cardiac dysfunction in patients with chronic heart failure (50). We investigated whether inhibition of Drp1 by cilnidipine had beneficial effects on heart failure after MI in mice. Because pharmacological pretreatment or early posttreatment (3 hours to 3 days) with amlodipine reportedly improves cardiac function after MI in rodents (5154), we compared the cardioprotective effects of cilnidipine and amlodipine starting 7 days after infarction. Continuous treatment of mice with cilnidipine (20 mg/kg per day) from 7 days after MI improved cardiac function compared with vehicle-treated mice at 4 weeks after MI (Fig. 7A). Because the same dose of amlodipine treatment increased mortality (fig. S10A), we analyzed the effect of a low daily dose of amlodipine (2.5 mg/kg per day). Posttreatment of mice with amlodipine did not have a cardioprotective effect (Fig. 7A), suggesting that cilnidipine exerts cardioprotection through a different molecular mechanism than amlodipine. Moreover, echocardiographic and Millar catheter analyses showed that cilnidipine treatment from 1 day after MI significantly protected LV function compared with vehicle-treated mice at 4 weeks after MI (tables S1 and S2). We used heart samples from mice treated with cilnidipine 1 day after infarction for further analyses. Cilnidipine treatment significantly reduced LV hypertrophy compared with vehicle-treated mice at 4 weeks after MI (Fig. 7, B and C). The mortality and LV hypertrophy after MI in N-type Ca2+ channel Cav2.2-deficient (Cav2.2−/−) mice were similar to those in wild-type (Cav2.2+/+) mice (Fig. 7D and fig. S10B). Cilnidipine treatment significantly improved LV hypertrophy and function after MI in both Cav2.2+/+ and Cav2.2−/− mice (Fig. 7D and table S3). These results suggest that inhibition of L/N-type Ca2+ channel activity does not participate in the ability of cilnidipine to prevent heart failure. Furthermore, cilnidipine significantly improved mitochondrial morphology (Fig. 7E) and decreased Drp1 activity (Fig. 7F) and SA-β-gal–positive areas (Fig. 7G) in the peri-infarct zone in the LV myocardium, as well as LV remodeling, including the increase in interstitial fibrosis (Fig. 7H) and expression of hypertrophy- and fibrosis-related genes (Fig. 7I). Cilnidipine failed to suppress HO-1 protein induction in the peri-infarct zone myocardium after MI (Fig. 7J), indicating that cilnidipine may not reduce hypoxic stress. Collectively, these results suggest that cilnidipine improves cardiac function after MI by suppressing Drp1-mediated myocardial cell senescence, although Ca2+ channel blockers are believed to be contraindicated for treatment of heart failure.

Fig. 7 Improvement of cardiac function after MI through suppression of Drp1-mediated myocardial senescence by cilnidipine.

(A) Effect of cilnidipine on ejection fraction (EF) in mice 4 weeks after MI. An osmotic pump filled with cilnidipine (20 mg/kg per day) or amlodipine (2.5 mg/kg per day) was implanted intraperitoneally 7 days after MI (n = 5 to 8 mice per treatment). (B) Effects of cilnidipine on the heart weight (HW)/body weight (BW) ratio in mice 4 weeks after MI. Cilnidipine (30 or 100 mg/kg per day) was infused at 1 day after MI (n = 4 to 10 mice per treatment). (C) Hematoxylin and eosin–stained images of hearts at 4 weeks after MI. Scale bars, 1.5 mm. Right: Cell surface area (CSA) of cardiomyocytes (n = 4 to 5 mice per treatment). (D) Effects of cilnidipine (30 mg/kg per day) on increased heart weight/body weight ratio in Cav2.2+/+ and Cav2.2−/− mice 4 weeks after MI (n = 3 to 6 mice per treatment). (E) Effects of cilnidipine (30 mg/kg per day) on mitochondrial morphology in peri-infarct zone myocardium 1 week after MI. Left: Representative images of interfibrillar mitochondria. Right: Quantification of mitochondrial circularity. n > 200 mitochondria from three mice per group. Scale bars, 1 μm. (F) Effects of cilnidipine (30 mg/kg per day) on Drp1 activity in peri-infarct zone myocardium 1 week after MI (n = 3 mice per treatment). (G) Effect of cilnidipine (30 mg/kg per day) on the proportion of SA-β-gal–positive area in peri-infarct zone myocardium 4 weeks after MI (n = 3 to 4 mice per treatment). Scale bars, 100 μm. (H) Picrosirius red–stained images of peri-infarct zone myocardium 4 week after MI. Scale bars, 100 μm. Right: Collagen volume fractions (CVFs) (n = 4 mice per treatment). (I) mRNA amounts of hypertrophy-related (ANP and α-SKA) and fibrosis-related (ACE, Periostin, Collagen 1a1, Collagen 3a1, TGF1, TGF2, TGF3, and CTGF) genes in peri-infarct zone myocardium at 1 week after MI with or without cilnidipine (30 mg/kg per day) (n = 3 to 4 mice per treatment). (J) Effects of cilnidipine on protein abundances in peri-infarct zone myocardium at 1 week after MI (n = 3 mice per treatment). *P < 0.05, **P < 0.01 by one-way ANOVA (A to H and J); ††P < 0.01 compared to vehicle sham, #P < 0.05, ##P < 0.01 compared to vehicle MI by one-way ANOVA (I).


Structural connections between mitochondria and the cytoskeleton play key roles in various mitochondrial functions, including transport and fission. Microtubules are important for mitochondrial motility and transport (55). The GTPase mitochondrial Rho (Miro) and the kinesin adaptor Milton complex drives kinesin-dependent mitochondrial transport among microtubules (55). The Miro-Milton complex regulates mitochondrial morphology in a Ca2+-dependent manner (55). In the case of mitochondrial fission, ER tubules wrap around mitochondria to facilitate constriction. This ER-mitochondria communication through contact sites known as mitochondria-associated ER membranes (MAMs) is important for mitochondrial fission. Actin filament formation at this region through ER-localized inverted formin 2 (INF2) and mitochondrial Spire1C is important for the accumulation and oligomeric ring formation of Drp1 (31, 56). Myosin II is also involved in this process (33). In this study, we identified FLNa as a binding partner of Drp1. FLNa is a scaffold for more than 90 binding proteins involved in cell adhesion and migration, including channels, receptors, intracellular signaling molecules, and transcriptional factors (39). Our results showed that FLNa acted as a protein scaffold to functionally couple Drp1 with actin at fission sites and activated Drp1 through GEF activity under hypoxic conditions. However, it is still unclear how Drp1-FLNa interaction is regulated under hypoxic conditions. The mitochondrial outer membrane protein FUN14 domain-containing 1 (FUNDC1) accumulates at MAMs and recruits Drp1 to MAMs and mitochondrial fission in response to hypoxia (57). Other binding proteins such as FUNDC1 or posttranslational modifications of Drp1 may regulate its interaction with FLNa.

Humans have three filamin genes, which encode FLNa, FLNb, and FLNc proteins. These proteins share the same domain structures and have about 70% sequence similarity overall. FLNa is ubiquitously distributed, and mutations in FLNa cause an X-linked form of familial cardiac valvular dystrophy (58, 59). FLNa gene deletion in mice leads to a severe defect in cardiac morphogenesis (60), suggesting a critical role of FLNa in cardiac systems. FLNc is expressed in both cardiac and striated muscle. Various mutations in the FLNc gene have been identified in patients with familial hypertrophic cardiomyopathy, and the FLNc mutants affect actin and FLNc organization (44). Future clinical studies are necessary to determine whether FLNc participates in Drp1-mediated mitochondrial fission and senescence and whether cilnidipine could be used to treat FLNc-related cardiomyopathy.

Chronic heart failure after MI is a major cause of mortality and morbidity worldwide. In this study, we demonstrated that Drp1-dependent mitochondrial fragmentation occurred in the peri-infarct zone in mouse heart after MI. Fragmented mitochondria in the myocardium are also observed in patients with dilated cardiomyopathy (61), and pharmacological inhibition of mitochondrial fission by Mdivi-1 attenuates heart failure in mice induced by pressure overload (7, 45), ischemia/reperfusion (8, 46), and doxorubicin (9). These results suggest that abnormal mitochondrial fission in the myocardium leads to the progression of heart failure and that inhibition of mitochondrial fission is an attractive target for stress-induced cardiomyopathy. However, the mitochondrial fission/fusion cycle is physiologically critical to maintain mitochondrial quality control. Genetic ablation of Drp1 leads to mitochondrial enlargement and evokes cardiomyocyte necrosis and lethal dilated cardiomyopathy (5). Therefore, long-term inhibition of mitochondrial fission, which would block physiological fission for quality control, might not be an appropriate therapeutic strategy.

In the present study, cilnidipine treatment attenuated the progression of heart failure after MI. Cilnidipine suppressed hypoxia-mediated mitochondrial fragmentation and subsequent senescence by inhibiting Drp1-FLNa complex formation. FLNa colocalized with Drp1 in peri-infarct zone myocardium of MI-treated heart, whereas the interaction of FLNa and Drp1 did not appear to occur under normoxia (Fig. 2). Therefore, cilnidipine would effectively prevent hypoxia-mediated mitochondrial fragmentation without disrupting the fission/fusion balance under normoxia.

An abnormal balance between mitochondrial fission and fusion has been linked to various diseases, including not only cardiac diseases but also neurologic diseases, cancer, and diabetes (62). Our results suggest that hypoxia induced the functional coupling of Drp1 and FLNa to enhance mitochondrial fission, and cilnidipine selectively inhibited the hypoxia-induced Drp1-FLNa interaction. Because FLNa is ubiquitously distributed and also correlated with cancer and Alzheimer’s disease (63, 64), elucidation of the universality of the Drp1-FLNa interaction for mitochondrial quality control may contribute to expanded approval of cilnidipine for other mitochondria-related diseases.


Reagents and antibodies

Amlodipine was from Tokyo Chemical Industry Co. Ltd. Mdivi-1, sildenafil, atenolol, propranolol, metoprolol, thiocarbohydrazide solution, durcupan resin, c-Myc peptide, and GTP-agarose were from Sigma-Aldrich. Dulbecco’s modified Eagle’s medium (DMEM), diltiazem, nifedipine, nisoldipine, nicardipine, nitrendipine, verapamil, losartan, uranium acetate, anti-Myc, and anti–β-actin were from Wako. MitoTracker Green FM, MitoTracker Red CM-H2XRos, and MitoSox Red were from Invitrogen. Collagenase II was from Worthington. Osmium tetroxide was from Electron Microscopy Science. Bovine serum albumin (BSA), sodium cacodylate, and phenol red–free DMEM were from Nacalai Tesque. Anti-Drp1 (H-300), anti-GAPDH (FL-335), anti-SOD2 (Mn-SOD: N-20), anti-Mfn1 (H-65), and anti-filamin 1 (FLNa: H-300 and E-3) were from Santa Cruz Biotechnology. Anti-p53 (7F5) and horseradish peroxidase (HRP)–conjugated anti-mouse and rabbit IgG were from Cell Signaling Technology. Anti-tafazzin, anti–HIF-1α, anti-sarcomeric α-actinin (EA-53), and anti-Mfn2 were from Abcam. Anti-Opa1 was from BD Biosciences. Anti–HO-1 was from Enzo Life Sciences. EGTA and Hepes were from Dojindo.


All experiments using mice and rats were approved by an Ethics Committee of National Institutes of Natural Sciences and carried out in accordance with their guidelines. C57BL/6J mice and Sprague-Dawley (SD) rats were purchased from Japan SLC Inc. The Cav2.2−/− mice and their wild-type littermates (Cav2.2+/+) were generated as previously described (65). All mice were maintained in a specific pathogen–free area on a 12-hour light/dark cycle.

Induction of MI

Surgery to induce MI was performed on 6-week-old male C57BL/6J, Cav2.2+/+, and Cav2.2−/− mice as described (34). All surgical procedures were performed in mice anesthetized with a mixture of domitor (0.75 mg/kg; Nihon Zenyaku Kogyo), midazolam (4 mg/kg; Sandoz), and butorphanol (5 mg/kg; Meiji Seika Pharma). A mini-osmotic pump (ALZET) filled with vehicle (saline) or cilnidipine (20, 30, or 100 mg/kg per day) or amlodipine (2.5 or 20 mg/kg per day) was implanted intraperitoneally into 6-week-old male C57BL/6J mice at 1 or 7 days after MI.

Transthoracic echocardiography and cardiac catheterization

Echocardiography and cardiac catheterization were performed in mice anesthetized with a mixture of domitor, midazolam, and butorphanol. Echocardiography was performed by Nemio-XG echocardiography (Toshiba) with a 14-MHz transducer at 4 weeks after MI. Cardiac functions were measured using a micromanometer catheter (Millar 1.4F, SPR 671, Millar Instruments) at 4 weeks after MI.

Morphological analysis

Mouse hearts were removed, washed in phosphate-buffered saline (PBS), and fixed with 10% neutral-buffered formalin (Nacalai Tesque). For quantitative assessment of collagen I and III deposition, hearts were embedded in paraffin, sectioned at a thickness of 3 μm, and stained with picrosirius red using 0.1% Direct red 80 (Polysciences). For assessment of cardiomyocyte cross-sectional areas, specimens were stained with hematoxylin and eosin (Sigma). Three regions were randomly selected for each specimen in the LV area, and the average values were calculated using BZ-II Analyzer software (Keyence).


Mouse heart tissues were embedded in optimal cutting temperature (O.C.T.) compound (Sakura Finetek) and snap-frozen with liquid nitrogen. Samples were sectioned at a thickness of 20 μm. Glass-mounted fresh frozen sections were fixed in 4% paraformaldehyde in PBS at room temperature for 10 min. After rinsing with PBS, sections were blocked in PBS with 1% BSA, 0.3% Triton X-100 at room temperature for 30 min and then incubated overnight with primary antibodies directed against FLNa and Drp1 at 4°C, followed by incubation with secondary fluorescent anti-mouse and anti-rabbit antibodies. Sections were incubated with the Vector TrueVIEW Autofluorescence Quenching Kit (Vector Laboratories Inc.) for 1 min, before being mounted with VECTASHIELD HardSet Antifade Mounting Medium (Vector Laboratories Inc.). Images were obtained using an A1 confocal microscope (Nikon).

Transmission electron microscopy

Mouse LV tissues were prefixed with 2% paraformaldehyde solution containing 0.15 M sodium cacodylate and 2 mM CaCl2 (pH 7.4) for 3 hours on ice and cut into 1- to 2-mm cubes. After washing with 0.15 M cacodylate solution, the tissue blocks were immersed in a solution containing 2% osmium tetroxide, 1.5% potassium ferrocyanide, 0.15 M sodium cacodylate, and 2 mM CaCl2 (pH 7.4) for 1.5 hours at room temperature. After washing with distilled water, the tissue blocks were immersed in thiocarbohydrazide (0.01 mg/ml) solution for 40 min and post-fixed with 2% osmium for 1 hour. En bloc staining was performed by sequentially immersing the tissue blocks in a solution of 1% uranium acetate overnight at 4°C and an aqueous solution of lead aspartic acid for 60 min with oven-drying. After dehydration in a graded ethanol series and acetone, the specimens were embedded in durcupan resin. The surface (thickness of 70 nm) of the resin-embedded tissue was exposed using a diamond knife on an Ultracut UC7 (Leica Microsystems) and imaged with a Veleta CCD camera (Olympus) equipped on a JEOL1010 microscope (JEOL). We defined “perinuclear mitochondria” as mitochondria around nuclear and “interfibrillar mitochondria” as a single row of mitochondria between the myofibrils. Mitochondria for each animal were traced, and quantification of area of perinuclear mitochondria and circularity of interfibrillar mitochondria was performed by ImageJ software (National Institutes of Health).

Cell culture and transfection

NRCMs and fibroblasts were prepared from the ventricles of 1- to 2-day-old SD rats as described (34). H9c2 cardiac myoblasts were purchased from the American Type Culture Collection and cultured in DMEM supplemented with 10% fetal bovine serum (FBS). Plasmid DNA was transfected using Viafect (Promega) for HeLa cells or a Neon transfection system (Invitrogen) for cardiomyocytes, fibroblasts, and H9c2 cells. For siRNA knockdown, cells were transfected with each siRNA (20 nM) using Lipofectamine RNAiMAX reagent (Invitrogen) for 72 hours. Stealth siRNAs for rat FLNa (#1, RSS308835; #2, RSS308836) and rat Drp1 (#1, RSS300120; #2, RSS300122) were from Invitrogen.

Western blotting

Mouse hearts were homogenized using a Physcotron (Microtec) in lysis buffer [20 mM Hepes (pH 7.4), 100 mM NaCl, 3 mM MgCl2, 1% NP-40, 50 mM NaF, 1 mM Na3VO4, and 20 mM β-glycerophosphate] containing a protease inhibitor cocktail at 4°C. Cleared lysates were mixed with sample buffer. Proteins separated by SDS-PAGE were electrotransferred onto polyvinylidene difluoride membranes (Millipore) at 2 mA/cm2 for 1 hour. After blocking with 1% BSA, the membrane-bound proteins were incubated with primary antibodies, followed by HRP-conjugated secondary antibodies. Antibody-bound proteins were detected using the Western Lightning Plus-ECL Kit (PerkinElmer).

GTP-agarose pulldown assay

GTP-agarose pulldown assays were performed as described (24). Tissue samples were homogenized using the Physcotron in ice-cold GTP-binding buffer [50 mM Hepes (pH 7.4), 1% Triton X-100, 10% glycerol, 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, and 1% protease inhibitor cocktail]. NRCMs were washed with ice-cold PBS, collected in GTP-binding buffer, and homogenized by sonication (Qsonica, Waken B Tech) for 3 s. The lysate was centrifuged (16,000g for 15 min at 4°C), and an aliquot of the supernatant (100 μg protein) was incubated with 20 μl of GTP-agarose beads (equilibrated in GTP-binding buffer) for 45 min at room temperature. The beads were centrifuged (1000g for 1 min at 4°C) and washed twice with GTP-binding buffer. The GTP-bound proteins were eluted with sample buffer containing dithiothreitol (DTT) and subjected to SDS-PAGE.

Measurement of mitochondrial functions

Mitochondrial membrane potential was analyzed with 5′,6,6′- tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1, Invitrogen). NRCMs were incubated with JC-1 (0.5 μM) for 20 min. After changing the medium to phenol red–free DMEM, digital images were taken at 40× magnification using a fluorescence microscope (BZ-X700, Keyence), with five regions randomly selected for each sample. The mitochondrial membrane potential was represented as the ratio of green-to-red fluorescence. Mitochondrial energy state was evaluated by lactate production. Lactate concentration in culture medium was measured by Determinar LA (Kyowa Medic).

Mitochondrial morphology

Mitochondrial morphology was detected with MitoTracker Green FM and fluorescence microscopy using the BZ-X700 microscope with a sectioning function. Cells were serum-starved for 48 hours and incubated under normoxic (21% O2) or hypoxic (1% O2) conditions for 16 hours in a multigas incubator (Panasonic). For LatB treatment, cells were treated with 20 nM LatB 1 hour before hypoxia. The cells were then loaded with MitoTracker Green (50 nM) for 20 min. For colocalization of Drp1 and mitochondria, fibroblasts expressing GFP-Drp1 were incubated with MitoTracker Red CM-H2XRos (50 nM) for 20 min. After incubation, the cells were washed with PBS, changed to phenol red–free DMEM, and imaged. To analyze actin structure, cells were incubated with 20 nM LatB for 18 hours, and actin cytoskeleton was stained with Alexa 594–phalloidin. Ten images were randomly taken for each sample. The average mitochondrial fragment length per cell was quantified using ImageJ, and the mitochondrial morphology was classified into three groups (vesicle, intermediate, and tubule) according to the average mitochondrial fragment length.

Mitochondrial ROS production

Mitochondrial ROS were examined using MitoSOX Red. Cells were incubated with MitoSOX (5 μM) for 10 min at 37°C, washed with PBS, changed to phenol red–free DMEM, and imaged by confocal microscopy (FV10i, Olympus).

Myocardial senescence

Myocardial senescence was evaluated by senescence-associated β-gal activity or cytosolic p53 accumulation. SA-β-gal activity was determined using the Senescence β-Galactosidase Staining Kit (Cell Signaling Technology) according to the manufacturer’s instructions. NRCMs were pretreated with drugs for 1 hour, subjected to 16 hours of hypoxia followed by 12 hours of reoxygenation, fixed with fixative solution for 15 min at room temperature, and incubated overnight with β-gal staining solution at 37°C. For mouse hearts, β-gal staining solution was used at a final pH of 5.0. For immunofluorescence to detect p53, cells were fixed with 4% paraformaldehyde for 10 min and incubated with blocking buffer (1% FBS in PBS containing 0.05% Triton X-100) for 1 hour. Cells were incubated with anti-p53 and anti–α-actinin antibodies and then with fluorescence-conjugated secondary antibodies. Samples were imaged with the BZ-X700 microscope. The population of senescent cells accompanying p53 punctae in α-actinin–positive cardiomyocytes was analyzed.


HeLa cells were lysed in GTP-binding buffer, and the cell lysates were clarified by centrifugation (16,000g for 15 min). FLAG-tagged proteins were immunoprecipitated from the supernatants containing 20-μl bed volume of anti-FLAG M2 agarose beads (Sigma). The immune complexes were washed three times and eluted with FLAG peptide (300 μg/ml). To identify Drp1-binding proteins, HeLa cells expressing FLAG-Drp1 were incubated under normoxia or hypoxia for 16 hours. After washing with PBS, the cells were incubated with 100 μM dithiobis (succinimidyl propionate) cross-linker for 30 min at room temperature. The cross-linking reaction was quenched by addition of 50 mM tris (pH 7.5) in PBS. After two washes with PBS, the cells were lysed and immunoprecipitated. Proteins were separated by SDS-PAGE and analyzed by Coomassie brilliant blue staining. Bands were identified after in-gel trypsin digestion and mass spectrometry. Mass spectrometric analysis was performed once.


HeLa cells expressing Myc-FLNa and FLAG-Drp1 were collected from five 10-cm dishes, and mitochondrial and cytosolic fractions were separated using a Mitochondria/Cytosol Fractionation kit (BioVision) according to the manufacturer’s instructions. Mitochondrial pellets were resuspended into the GTP-binding buffer, and mitochondrial lysates were clarified by centrifugation (16,000g for 15 min). FLAG-Drp1 was immunoprecipitated from mitochondrial and cytosolic supernatants.

Steady-state GTP hydrolysis activity of Drp1

Recombinant His-tagged Drp1 was purified from Escherichia coli using Ni–nitrilotriacetic acid agarose and stocked in buffer [20 mM Hepes (pH 7.4), 100 mM NaCl, 3 mM MgCl2, 1 mM DTT, 10% glycerol, and 10 μM GDP]. FLNa was purified from HEK293 cells expressing Myc-FLNa. Myc-FLNa was immunoprecipitated using anti–Myc-tag monoclonal antibody agarose (Medical & Biological Laboratories Co. Ltd.) and collected in elution buffer [20 mM Hepes (pH 7.4), 50 mM KCl, 2 mM MgCl2, 1 mM DTT, and c-Myc peptide (500 μg/ml)]. Drp1 (0.6 μM) was mixed with or without FLNa (0.2 μM) in hydrolysis buffer [20 mM Hepes (pH 7.4), 50 mM KCl, 10 mM MgCl2, and 1 mM DTT]. To promote spontaneous GDP/GTP exchange of Drp1, 1 mM EDTA and 2 mM MgCl2 were used instead of 10 mM MgCl2. GTP hydrolysis was determined by incubating Drp1 with 100 μM GTP for 60 min at 37°C. At each time point, the reaction was stopped by adding 25 mM EDTA, and released free phosphate was measured using a Malachite Green Phosphate Assay kit (BioAssay Systems).

Real-time reverse transcription polymerase chain reaction

Total RNA was isolated from frozen mouse hearts using an RNeasy Fibrous Tissue Mini Kit (Qiagen) according to the manufacturer’s instructions. Complementary DNA was synthesized with Prime Script RT (Takara Bio). Real-time polymerase chain reaction (PCR) was performed using Power SYBR Green PCR Master Mix (Applied Biosystems) or the QuantiTect Probe RT-PCR Kit (Qiagen) and TaqMan probes with an ABI PRISM 7500 Real-Time PCR System (Applied Biosystems) according to the manufacturer’s instructions. The primers are listed in table S4.

Statistical analysis

The results are presented as means ± SEM from at least three independent experiments. Statistical comparisons were carried out by two-tailed Student’s t test or one-way ANOVA, followed by a Tukey or Newman-Keuls comparison procedure. Values of P < 0.05 were considered to be statistically significant.


Fig. S1. Mitochondrial fission and premature myocardial senescence in peri-infarct zone myocardium after MI.

Fig. S2. Changes in protein abundances and activities in peri-infarct zone myocardium after MI.

Fig. S3. Mitochondrial morphology in NRCMs.

Fig. S4. Reoxygenation-induced mitochondrial ROS production plays an indispensable role in myocardial senescence.

Fig. S5. Effects of gene silencing and pharmacological inhibition of Drp1 on mitochondrial morphology and ROS production.

Fig. S6. Identification of FLNa by mass spectrometric analysis.

Fig. S7. No interaction between Drp1 and the Ig24 dimerization domain of FLNa.

Fig. S8. Effect of FLNa knockdown on hypoxia-induced mitochondrial fission and function.

Fig. S9. Mitochondrial morphology in cardiac fibroblasts.

Fig. S10. Mortality in mice after MI.

Table S1. Echocardiographic parameters of mice with cilnidipine administration.

Table S2. Cardiac parameters of mice with cilnidipine administration.

Table S3. Echocardiographic parameters of Cav2.2(+/+) or Cav2.2(−/−) mice with cilnidipine administration.

Table S4. List of Taqman probe and primer sets for real-time PCR.


Acknowledgments: We thank H. Ishihara, K. Murata, and N. Miyazaki (electron microscopy facilities in NIPS, Okazaki, Japan) for technical guidance in the transmission electron microscopy imaging and the Functional Genomics Facility, NIBB Core Research Facilities for technical support in the Mass Spectrometry, Spectrography and Bioimaging Facility. We also thank A. Sherwin from Edanz Group ( for editing a draft of this manuscript. Funding: This work was supported by grants from JST PRESTO Program (13417243 to M.N.), JSPS KAKENHI [15K14959 (to M.N.) and 17K15464 (to A.N.)], and Innovative Areas [Research in a Proposed Research Area ‘Oxygen Biology’ (26111011)] from the Ministry of Education, Culture, Sports, Science and Technology. This work was also supported by the Platform Project for Supporting Drug Discovery and Life Science Research [Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)] from AMED (JP18am0101091) and the Japan Research Foundation for Clinical Pharmacology (to M.N.). Author contributions: M.N., A.N., T.S., and T. Ishikawa designed the research. A.N., T.S., T. Tanaka, K.S., N.K., N.S., and T. Toyama performed experiments. T.N.-T., S.Y., Y.S., K.K., T. Ide, Y.K., T.A., A.O., and Y.M. contributed new reagents/analytic tools. A.N., T.S., T. Tanaka, T. Ishikawa, T. Ide, Y.K., T.A., A.O., Y.M., and M.N. analyzed and interpreted data. A.N., T.S., and M.N. wrote the paper. Competing interests: A patent (JPWO2016080516A1) has been filed in the International Patent System for part of this work. M.N. and T.I. are named as inventors on this patent. The other authors have declared that they have no competing interests. Data and materials availability: The mass spectrometry data have been deposited in jPOSTrepo ( with the identifier JPST000503. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article