OUP user menu

★ Editor's choice ★

Decreasing mitochondrial fission diminishes vascular smooth muscle cell migration and ameliorates intimal hyperplasia

Li Wang, Tianzheng Yu, Hakjoo Lee, Dawn K. O'Brien, Hiromi Sesaki, Yisang Yoon
DOI: http://dx.doi.org/10.1093/cvr/cvv005 272-283 First published online: 12 January 2015


Aims Vascular smooth muscle cell (VSMC) migration in response to arterial wall injury is a critical process in the development of intimal hyperplasia. Cell migration is an energy-demanding process that is predicted to require mitochondrial function. Mitochondria are morphologically dynamic, undergoing continuous shape change through fission and fusion. However, the role of mitochondrial morphology in VSMC migration is not well understood. The aim of the study is to understand how mitochondrial fission contributes to VSMC migration and provides its in vivo relevance in the mouse model of intimal hyperplasia.

Methods and results In primary mouse VSMCs, the chemoattractant PDGF induced mitochondrial shortening through the mitochondrial fission protein dynamin-like protein 1 (DLP1)/Drp1. Perturbation of mitochondrial fission by expressing the dominant-negative mutant DLP1-K38A or by DLP1 silencing greatly decreased PDGF-induced lamellipodia formation and VSMC migration, indicating that mitochondrial fission is an important process in VSMC migration. PDGF induced an augmentation of mitochondrial energetics as well as ROS production, both of which were found to be necessary for VSMC migration. Mechanistically, the inhibition of mitochondrial fission induced an increase of mitochondrial inner membrane proton leak in VSMCs, abrogating the PDGF-induced energetic enhancement and an ROS increase. In an in vivo model of intimal hyperplasia, transgenic mice expressing DLP1-K38A displayed markedly reduced ROS levels and neointima formation in response to femoral artery wire injury.

Conclusions Mitochondrial fission is an integral process in cell migration, and controlling mitochondrial fission can limit VSMC migration and the pathological intimal hyperplasia by altering mitochondrial energetics and ROS levels.

  • Mitochondria
  • Mitochondrial fission
  • Cell migration
  • Vascular smooth muscle cells
  • DLP1
  • Drp1

1. Introduction

Intimal hyperplasia occurs in response to arterial injury and is an important feature of restenosis and atherosclerotic plaques.1,2 Although (trans)differentiation of endothelial cells, fibroblasts, or other circulating precursors may contribute to the neointima, a majority of neointimal vascular smooth muscle cells (VSMCs) migrate from the underlying medial layer.39 It is well established that PDGF plays a prominent role in recruiting VSMCs to the neointima following arterial injury and in the pathogenesis of atherosclerosis.10,11

Cell migration relies on reorganization of the actin cytoskeleton and myosin motor function, both of which require ATP; therefore, it is predictable that mitochondria are required to generate the energy necessary for cell migration. Mitochondrial morphology is now recognized as an important factor closely associated with the energetic state of mitochondria.1214 While mitochondrial morphologies vary among different cell types, the most prevalent morphology is observed as filamentous tubules that form reticular networks. These mitochondrial tubules undergo fission and fusion mediated by membrane remodelling dynamin family proteins. In mammals, dynamin-like protein 1 (DLP1; also known as Drp1) mediates mitochondrial fission, whereas mitofusin isoforms (Mfn1 and Mfn2) and optic atrophy 1 (OPA1) mediate fusion of the outer and inner membranes, respectively.

Mitochondrial shape has been shown to change in response to internal and external stimuli or stress.1518 Previous studies showed that mitochondrial fission and fusion play a role in cell migration. Increased migration and invasion of metastatic breast cancer cells were associated with up-regulation of DLP1 and fragmented mitochondrial morphology.19 Another study reported mitochondrial redistribution to the uropod during chemokine-induced lymphocyte migration. It was suggested that mitochondrial fission facilitates this observed redistribution.20 However, a direct correlation among mitochondrial morphology, function, and cell migration remains ill-defined.

In this study, we found that PDGF stimulation induced mitochondrial shortening in VSMCs through the function of the mitochondrial fission protein DLP1. Inhibition of mitochondrial fission greatly decreased PDGF-induced VSMC migration, indicating a requisite role of mitochondrial fission in cell migration. We found that inhibition of mitochondrial fission in VSMCs decreased respiration coupling, providing the underlying mechanism by which cell migration is limited in fission inhibition. The in vivo application of our findings in transgenic mice expressing the fission mutant DLP1-K38A demonstrated that decreasing mitochondrial fission greatly reduced neointima formation in the mouse model of intimal hyperplasia.

2. Methods

A detailed Methods section is available in Supplementary material online.

2.1 Primary VSMC culture and inhibition of mitochondrial fission

All the procedures involving animals conform to the US National Institutes of Health regulations and were approved by the Institutional Animal Care and Use Committee of Georgia Regents University. VSMCs were isolated from mouse thoracic aortas as previously described following sacrificing them by CO2 inhalation.21 The dominant-negative mutant DLP1-K38A or DLP1 siRNA was used to inhibit mitochondrial fission.

2.2 Fluorescence imaging

VSMCs were infected with adenovirus carrying mitochondria-targeted GFP (Ad-mitoGFP) to visualize mitochondria. For mitochondrial morphology quantification, morphologies were divided into three classifications: ‘tubular’—greater than half of mitochondria in a cell displaying the long tubular shape; ‘intermediate’—less than half of mitochondria in a cell displaying the tubular shape; and ‘fragmented’—the majority of mitochondria in a cell displaying a short, fragmented shape. Morphometric analyses of mitochondria were performed as described previously using ImageJ.15,22 F-actin and nuclei were stained with rhodamine-conjugated phalloidin and DAPI, respectively. Dihydroethidium (DHE) was used to determine ROS levels as previously described.23 Mitochondrial membrane potential was evaluated with tetramethylrhodamine ethyl ester (TMRE).

2.3 Cell migration assays

The Boyden chamber was used to assess cell migration. The number of migrated cells across the filter was counted after 5-h incubation. For wound healing assays, monolayer cells were scratched and incubated for 6 h in the presence of PDGF.

2.4 Respiration measurements

Whole-cell oxygen consumption rate (OCR) was measured in a sealed chamber as described previously.18,24

2.5 Femoral artery wire injury

Transgenic mice expressing DLP1-K38A in a doxycycline (Dox)-dependent manner (double-transgenic dTg[rtTA/DLP1-K38A]) were described previously.24 Ten- to 12-week-old dTg[rtTA/DLP1-K38A] mice and the age- and sex-matched Tg[rtTA] littermates were used in these experiments. Wire-induced bilateral femoral artery injury was performed as previously described.25 Mice were anaesthetized by ketamine (100 mg/kg) and xylazine (10 mg/kg) i.p. for surgery. An arteriotomy was performed in the left femoral artery for the wire injury with a hydrophilic guide wire, followed by ligature. Sham-operated right femoral arteries experienced arteriotomy and ligature without passage of the wire. Mice received buprenorphine 0.1 mg/kg subcutaneously at the end of the surgery and every 6–12 h until they recovered. Mice were sacrificed by CO2 inhalation at 2–4 weeks post-surgery for femoral artery collection.

2.6 Statistical analyses

Error bars in all graphs represent SEM. Student's t-test (two-tailed, unpaired) was used to compare the two groups. One-way ANOVA was used for multiple groups. A value of P < 0.05 was considered statistically significant.

3. Results

3.1 Migratory stimulation induces mitochondrial shortening in VSMCs

VSMC migration from the arterial media to the neointima is important in the development of intimal hyperplasia after arterial wall injury. Mitochondrial morphology changes in response to environmental stimuli and stresses. Therefore, we first tested the effect of chemotactic stimulation on mitochondrial morphology in VSMCs. PDGF is a strong chemoattractant that plays a prominent role in recruiting VSMCs to the neointima. We found that PDGF stimulation induces rapid mitochondrial shortening in VSMCs. In untreated control VSMCs, mitochondria were observed as networks of long filaments. Upon adding PDGF (10 ng/mL), mitochondria rapidly became shortened within 10 min in a majority of the cells (Figure 1A and B). In cell counting based on different mitochondrial morphologies, the number of cells containing intermediate and fragmented mitochondria was markedly increased after PDGF treatments, which was sustained up to 6-h incubation in PDGF (Figure 1C). In addition, we used morphometric analyses to calculate average form factor and aspect ratio for mitochondria in VSMCs treated with PDGF. A form factor is defined as the reciprocal value of the circularity, which represents mitochondrial elongation and complex shapes such as branching.26 Aspect ratio is the ratio of maximum-to-minimum axis, reflecting mitochondrial length.26 Mitochondria in cells treated with PDGF had lower form factor and aspect ratio values (Figure 1D), indicating quantifiably shorter mitochondria. A less pronounced difference in the aspect ratio is presumably due to underestimated values for the long but curved mitochondria in control cells. Furthermore, mitochondrial size distribution shows the prevalence of smaller mitochondria in PDGF-treated cells (Figure 1E). In these analyses, the average number of mitochondria was ∼300 in control and ∼800 in PDGF-treated VSMCs, suggesting that short mitochondria were likely formed by the fission process. These data demonstrate that PDGF induces mitochondrial shortening in VSMCs.

Figure 1

PDGF stimulation induces rapid mitochondrial shortening in VSMCs. (AC) Mitochondria in VSMCs were visualized with mitochondria-targeted GFP. In untreated quiescent VSMCs, mitochondria were observed as networks of long filaments (A). PDGF (10 ng/mL) stimulation induced rapid mitochondrial shortening within 10 min (B). Scale bar: 20 µm. (C) Cell counting shows that the number of cells containing intermediate and fragmented mitochondria was markedly increased after PDGF stimulation, which was sustained up to 6-h incubation. P < 0.0001 to 0 min. Data from three independent experiments. (D and E) Morphometric analyses of mitochondria in VSMCs treated with PDGF for 30 min show decreased values for both form factor and aspect ratio in VSMCs treated with PDGF, indicating mitochondrial shortening in PDGF-treated cells (D). ****P < 0.0001. A frequency plot for mitochondrial size distribution also shows the prevalence of smaller mitochondria in PDGF-treated cells (E). n = 9, #P < 0.0001; $P < 0.01.

3.2 PDGF-induced mitochondrial shortening in VSMCs is dependent on DLP1-mediated fission

To test whether mitochondrial fission is increased upon PDGF treatment, we first examined cellular distribution of the fission protein DLP1. In IF, DLP1 in untreated control VSMCs distributed throughout the cytoplasm with lower-level colocalization with mitochondria (Figure 2A). In contrast, in cells treated with PDGF, many DLP1 puncta became associated with short mitochondria, whereas cytosolic DLP1 was decreased (Figure 2A). To substantiate the increased mitochondrial association of DLP1 in PDGF stimulation, we isolated the mitochondrial fraction from VSMCs treated with PDGF and assessed the level of DLP1. We found that PDGF stimulation of VSMCs substantially increased DLP1 levels in the mitochondrial fraction in both 10- and 30-min treatments with no changes in total DLP1 levels (Figure 2B). These observations suggest that the PDGF-induced mitochondrial shortening is accompanied by increased DLP1 in mitochondria. Phosphorylation of DLP1 at two different serine residues has been shown to regulate the association of DLP1 with mitochondria. Phosphorylation at the upstream serine residue (corresponding to serine 616 in human DLP1) increases mitochondrial fission,2730 whereas phosphorylation at the downstream serine (serine 637) increases or decreases fission depending on kinases.3134 We found that DLP1 phosphorylation at serine 616 was increased in PDGF-treated VSMCs (Figure 2C). Because fibroblasts are responsive to PDGF,35 we also examined DLP1 phosphorylation in this cell type. Both NIH 3T3 and mouse embryonic fibroblasts (MEFs) showed significantly increased phosphorylation at serine 616 upon PDGF treatment (Figure 2C). No clear changes in phosphorylation at serine 637 were observed with PDGF treatment (not shown). To verify PDGF specificity for DLP1-mediated fission, we examined the dose-dependent effect of PDGF on DLP1 phosphorylation at serine 616. More than five-fold increase in DLP1 phosphorylation was observed with lower concentrations of PDGF (0.5 and 1 ng/mL), and more drastic increases in higher PDGF concentrations up to 20 ng/mL (Figure 2D). Furthermore, a selective inhibitor of PDGF receptor kinase, AG1295, prevented DLP1 phosphorylation, indicating a specific effect of PDGF on DLP1 phosphorylation and mitochondrial fission (Figure 2E). These results indicate that PDGF stimulation increases mitochondrial fission by inducing DLP1 binding to mitochondria and phosphorylation at the serine 616 of DLP1.

Figure 2

DLP1 mediates PDGF-induced mitochondrial shortening. (A) IF shows increased DLP1 colocalization with mitochondria in VSMCs after PDGF treatment. Quiescent VSMCs expressing mitochondria-targeted GFP (green) were incubated with and without PDGF for 30 min. Scale bar: 10 µm. (B) DLP1 levels were increased in the mitochondrial fractions from VSMCs incubated in PDGF for 10 and 30 min. Representative data of three independent experiments. (C) PDGF treatment increased DLP1 phosphorylation at serine 616 in VSMCs, NIH 3T3, and MEFs. (D) A PDGF dose-dependent increase in DLP1 phosphorylation at serine 616. Numbers indicate a fold increase in band density normalized to total DLP1. (E) The PDGF receptor inhibitor AG1295 (10 µM) blocks the DLP1 phosphorylation. (F) VSMCs expressing the fission mutant DLP1-K38A maintained an elongated mitochondrial morphology after PDGF stimulation.

To further examine whether mitochondrial shortening induced by PDGF requires the DLP1-mediated fission process, we inhibited mitochondrial fission by expression of the dominant-negative mutant DLP1-K38A and examined mitochondrial morphology. Mitochondria in cells expressing DLP1-K38A were more elongated than those in normal VSMCs. Upon PDGF stimulation, these cells maintained the elongated mitochondria (Figure 2F), indicating that fission inhibition prevented PDGF-induced mitochondrial shortening. Time-lapse imaging revealed that mitochondria in DLP1-K38A-expressing VSMCs remained elongated upon PDGF treatment (see Supplementary material online, Movie S2), whereas control cells began forming short and condensed mitochondria at ∼5 min after addition of PDGF (see Supplementary material online, Movie S1). These observations demonstrate that the mitochondrial fission protein DLP1 is responsible for PDGF-induced mitochondrial shortening.

3.3 Inhibition of mitochondrial fission in VSMCs prevents PDGF-induced lamellipodia formation and cell migration

Because we found that chemotactic stimulation induces a rapid increase in mitochondrial fission in VSMCs, we next examined whether mitochondrial fission is important for PDGF-induced VSMC migration. Lamellipodia are actin-mediated membrane protrusions at the leading edge of polarized cells, a characteristic feature of migrating cells. Therefore, we examined lamellipodia in control and DLP1-K38A-expressing cells after PDGF stimulation. In control VSMCs, phalloidin staining indicated that PDGF stimulation induced lamellipodia formation (Figure 3A). In contrast, we found that the inhibition of mitochondrial fission by DLP1-K38A significantly decreased PDGF-induced lamellipodia formation while maintaining elongated mitochondrial morphology (Figure 3B and B′). Quantification shows a four-fold decrease in the number of cells with lamellipodia among DLP1-K38A-expressing cells when compared with control cells (Figure 3C).

Figure 3

Inhibition of mitochondrial fission diminishes PDGF-induced cell migration. (AC) Lamellipodia assessed by phalloidin staining in PDGF-treated VSMCs. PDGF induced lamellipodia formation in control cells, whereas fission inhibition by DLP1-K38A significantly decreased lamellipodia formation. Scale bar: 20 µm. n = 50–100, **P < 0.01. (D) Immunoblotting of DLP1 showing DLP1 silencing and overexpression of the dominant-negative DLP1-K38A. The arrow indicates overexpressed DLP1-K38A. DLP1-K38A was generated using a larger spliced variant of DLP1. (E) PDGF induces VSMC migration. Cell migration was assessed in the Boyden chamber after 5-h incubation in PDGF. n = 12, ****P < 0.0001. (F) Inhibition of mitochondrial fission prevents PDGF-induced VSMC migration. Mitochondrial fission was inhibited by DLP1 silencing or DLP1-K38A expression. Non-targeted RNA was used as a control. n = 12, ****P < 0.0001 vs. control. (G) Boyden chamber assays with MEFs show greatly decreased cell migration of DLP1-KO cells. n = 8 for wild-type (WT) and 12 for DLP1-KO, ***P < 0.001. (H) MEF migration assessed by wound healing assay. Data show a significant decrease in migration of DLP1-KO MEFs after 6-h PDGF incubation. Representative images from three independent experiments.

Next, we directly tested whether mitochondrial fission is important for cell migration. Boyden chamber assays indicated that PDGF treatment greatly increased VSMC migration (Figure 3E). We found that inhibition of mitochondrial fission prevents this PDGF-induced VSMC migration. Mitochondrial fission was inhibited by DLP1 silencing or DLP1-K38A expression (Figure 3D). Upon fission inhibition in VSMCs, the number of cells having migrated after 5-h PDGF incubation was markedly decreased to the unstimulated level (Figure 3F). We also tested DLP1-knockout MEFs (DLP1-KO MEFs)36 for PDGF-induced cell migration. Similar to VSMCs with fission inhibition, DLP1-KO MEFs showed greatly reduced cell migration after 5-h PDGF incubation, compared with wild-type MEFs (Figure 3G). Wound healing assay also showed a significant decrease in migration of DLP1-KO MEFs after 6-h PDGF incubation (Figure 3H). Our data demonstrated that PDGF stimulation activates mitochondrial fission, and that inhibiting fission greatly diminishes chemotactic cell migration.

3.4 PDGF stimulation of VSMCs activates mitochondrial energetics and ROS production for cell migration

Cell migration is an energy-demanding process, which may require an activation of mitochondrial energetic function. Therefore, we examined mitochondrial functional parameters in PDGF-stimulated VSMCs. We found that OCR was immediately and significantly increased upon addition of PDGF (1.25 ± 0.026 fold, n = 5; Figure 4A). We also found that the VSMCs incubated in PDGF for 5 h had a substantially higher OCR than untreated cells (Figure 4B). Leak and maximum OCRs measured with oligomycin and carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP), respectively, were not significantly different in untreated vs. PDGF-treated cells. We also examined mitochondrial inner membrane potential using the potentiometric probe TMRE. Quantification of TMRE fluorescence upon addition of PDGF indicated a gradual increase in the membrane potential, which became significant after 8–10 min of PDGF incubation (Figure 4C). The PDGF receptor inhibitor AG1295 abolished the PDGF-induced increase of TMRE fluorescence (Figure 4C). In prolonged PDGF incubation, quantification showed significant hyperpolarization at 30 min, apparently reaching a plateau afterwards (Figure 4D). These results are consistent with the increase of OCR upon PDGF addition, suggesting that mitochondrial energetic function is enhanced in response to PDGF stimulation in VSMCs.

Figure 4

PDGF stimulation of VSMCs activates mitochondrial energetics. (A) Measurements of VSMC oxygen consumption indicated a significant increase in respiration upon addition of PDGF. A representative trace of five independent experiments. (B) VSMCs incubated in PDGF for 5 h had a significantly higher basal OCR than untreated cells. Leak and maximum OCRs measured with oligomycin and FCCP, respectively, showed no differences. n = 7 for control and 8 for PDGF, **P < 0.01. (C) TMRE fluorescence gradually increased upon addition of PDGF, indicating an increase in inner membrane potential. The PDGF receptor inhibitor AG1295 (10 µM) blocked the TMRE fluorescence increase in PDGF incubation. Vehicle control had no effect on TMRE fluorescence. (D) TMRE fluorescence measured in cells treated with PDGF for 0, 30, 60, and 120 min. n = 100–163, ****P < 0.0001.

PDGF has been shown to increase reactive oxygen species (ROS), which are a signalling component necessary for cell migration.3740 Although PDGF and other growth factors are known to increase ROS through the NADPH oxidase,38,39 ROS from mitochondria have also been shown to play a role in cell migration.41,42 We observed mitochondrial hyperpolarization in PDGF treatment, which is conducive to an ROS increase from mitochondria. To examine whether mitochondria produce ROS upon PDGF stimulation, we treated VSMCs with the mitochondria-targeted antioxidant mito-TEMPO during PDGF incubation. ROS levels increased significantly with PDGF incubation; however, mito-TEMPO blunted this PDGF-induced ROS increase (Figure 5A and B). Furthermore, mild uncoupling of mitochondria by low concentration FCCP (100 nM) also prevented the PDGF-induced ROS increase (Figure 5C), confirming that ROS are generated from the mitochondrial respiratory chain through increased inner membrane potential. These data demonstrate that PDGF stimulation increases ROS production from mitochondria in VSMCs.

Figure 5

PDGF stimulation of VSMCs increases mitochondrial ROS production for cell migration. (A and B) ROS levels were evaluated by DHE in VSMCs. Ethidium fluorescence shows increased ROS in VSMCs incubated in PDGF for 60 min. Mito-TEMPO (10 µM, 30 min prior to PDGF treatment) abolished the PDGF-induced ROS increase. AU: arbitrary unit. n = 20, ***P < 0.001. (C) Mild uncoupling of mitochondria by 100 nM FCCP prevented the PDGF-induced ROS increase. n = 26, ****P < 0.0001. (D) Mitochondrial energetics and ROS production are necessary for PDGF-induced VSMC migration. Treatment with oligomycin, mito-TEMPO, or FCCP significantly reduced migration of VSMCs stimulated with PDGF. n = 6, ***P < 0.001 and ****P < 0.0001 vs. ‘None’.

Because our experimental results indicated that PDGF stimulation increases mitochondrial energetics and ROS production from mitochondria, we next examined whether these components play a role in PDGF-induced VSMC migration. To this end, PDGF-induced VSMC migration was assessed in the presence of the ATP synthase inhibitor oligomycin, mito-TEMPO, or FCCP. As shown in Figure 5D, oligomycin treatment significantly reduced migration of VSMCs stimulated with PDGF, suggesting that mitochondrial ATP production is necessary for cell migration. Likewise, decreasing mitochondrial ROS production by treatment with mito-TEMPO or 100 nM FCCP greatly limited the migration of VSMCs upon PDGF incubation. Mild uncoupling by 100 nM FCCP did not decrease ATP production;24 therefore, the decreased cell migration under these conditions is likely due to diminished ROS levels (Figure 5C and D). These data indicate that mitochondrial energy production and mitochondria-generated ROS are important in VSMC migration.

3.5 Inhibition of mitochondrial fission in VSMCs increases inner membrane proton leak

It has been shown that deficiency in mitochondrial fission increases inner membrane proton leak/uncoupling in other cell types.18,24,43,44 Because an increase of inner membrane proton leak decreases not only the efficiency of mitochondrial ATP synthesis but also ROS production from mitochondria, we reasoned that the limited cell migration observed in fission inhibition was due to increased proton leak. Therefore, we compared the respiration coupling states of VSMCs with and without DLP1 silencing. Owing to variations in experiments and cell status, it is not optimal to determine proton leak based on measured leak respiration in the presence of oligomycin. Instead, a respiration ratio within individual measurements can be used to determine coupling status. In intact cell respiration, the ratio of the uncoupler (FCCP)-induced maximum rate (state 3u) to the leak rate in the presence of oligomycin (state 4o) is analogous to the respiratory control ratio (RCR) of isolated mitochondria,45 which defines how well the inner membrane is sealed or respiration coupling. The slope of the linear regression in the plot of state 3u against state 4o represents the cellular quasi-RCR, referred to as j3u/4o (Figure 6A).46 DLP1-silenced VSMCs had a significantly lower j3u/4o value (6.1 ± 0.34 vs. 8.5 ± 0.69 in control, P = 0.0138), indicating decreased respiration coupling. The reciprocal value of the j3u/4o was presented as the leak ratio, representing the extent of proton leak out of the maximum uncoupled respiration. The leak ratio was higher in DLP1-silenced VSMCs, indicating an increase in proton leak (Figure 6B). These results indicate that decreasing mitochondrial fission in VSMCs reduces the respiration coupling efficiency.

Figure 6

Inhibition of mitochondrial fission decreases respiration coupling efficiency and ROS level. (A and B) DLP1-silenced VSMCs showed significantly decreased OXPHOS coupling. The FCCP-induced maximum rate (state 3u) was plotted against the leak rate in the presence of oligomycin (state 4o) (A). Linear regressions show a decreased slope in fission inhibition, indicating decreased respiration coupling. The ratio of state 4o/state 3u is presented as the extent of proton leak (B). DLP1-silenced VSMCs show increased proton leak compared with control cells. n = 7 for control and 6 for DLP1 siRNA, *P < 0.05. (C and D) Inhibition of mitochondrial fission diminishes ROS levels. VSMCs expressing DLP1-K38A (arrows in C) display significantly reduced ROS levels (C), as shown in quantification of ethidium fluorescence (D). n = 24, ***P < 0.001.

Because increased proton leak limits ROS production from the respiratory chain, we examined the effect of fission inhibition on PDGF-induced ROS production. We co-infected Ad-DLP1-K38A and Ad-mitoGFP in VSMCs in a low titer to compare ROS levels in control and DLP1-K38A-expressing cells under the same experimental conditions. We consistently observed that VSMCs expressing DLP1-K38A showed lower levels of ROS compared with adjacent uninfected cells in PDGF incubation (Figure 6C and D). Taken together, these experimental results demonstrate that chemotactic stimulation of VSMCs increases mitochondrial energetics and ROS levels, which is accompanied by increased mitochondrial fission. Our data establish that inhibiting mitochondrial fission limits VSMC migration by negating the energetic and ROS increase through the increased proton leak.

3.6 Expression of DLP1-K38A in mice decreases intimal hyperplasia following arterial wall injury

Our in vitro experimental results obtained using VSMCs demonstrated that the inhibition of mitochondrial fission prevents cell migration, predicting a beneficial effect of targeting mitochondrial fission in decreasing intimal hyperplasia prevalently associated with restenosis and atherosclerosis. To test the in vivo efficacy of decreasing mitochondrial fission, we used our transgenic mice expressing DLP1-K38A in a Dox-inducible manner.24 Immunolabelling showed the transgene expression in femoral arteries of Dox-induced mice (Figure 7A′). In addition, VSMCs isolated from the transgenic mice showed significantly diminished cell migration upon PDGF treatment in the presence of Dox (Figure 7B), suggesting that the in vivo expression level of DLP1-K38A in these transgenic mice is sufficient to decrease VSMC migration.

Figure 7

Expression of DLP1-K38A in mice decreases intimal hyperplasia. (A) DLP1 immunostaining of cross-section of femoral arteries from Dox-induced mice shows transgene expression in DLP1-K38A transgenic mice (A: control and A′: DLP1-K38A mouse). (B) VSMCs isolated from the transgenic mice showed significantly reduced PDGF-induced cell migration. Cell migration was assessed by Boyden chamber assays with VSMCs cultured in the presence of Dox (6 μM). n = 21, ****P < 0.0001. (CE) DLP1-K38A expression prevents the ROS increase in response to wire injury of the femoral artery. The ROS level was assessed by DHE in frozen sections of the femoral arteries at 2 weeks post injury. Wire injury increased ROS level in the femoral arteries from control mice (C and C′), whereas injured arteries from transgenic mice showed normal levels of ROS (D and D′). Quantification of ethidium fluorescence indicates that DLP1-K38A expression abolishes ROS production. n = 6, **P < 0.01. (FI) Decreasing mitochondrial fission is effective in diminishing intimal hyperplasia in response to arterial injury. Intimal hyperplasia was examined by CME staining at 4 weeks post injury. Wire-injured arteries showed a large increase in neointimal areas, when compared with sham-operated arteries from control (F and F′). Injured arteries from DLP1-K38A transgenic mice exhibited a significantly reduced neointima (G and G′). EEL: external elastic lamina; IEL: internal elastic lamina; L: lumen. The layer between EEL and IEL is the media and the layer encroaching the lumen from IEL is neointima. Morphometric analyses of intimal area and medial area reveal a significant decrease in both intimal area and intima/media ratio in DLP1-K38A transgenic mice (H and I). Note an increase in medial area in DLP1-K38A (H). n = 8, *P < 0.05, **P < 0.01.

We administered wire injury to the vessel wall in femoral arteries of control and DLP1-K38A mice, and then examined ROS levels and the intimal hyperplasia. We detected an increased level of ROS in the frozen sections of injured femoral arteries from control at 2 weeks post injury (Figure 7C, C′, and E), consistent with previous reports.47 In contrast, injured arteries from transgenic mice showed the normal level of ROS (Figure 7D, D′, and E), indicating that DLP1-K38A expression prevents the ROS increase in developing neointima in vivo. Intimal hyperplasia was examined in the femoral arteries 4 weeks post injury. Cross-sections of the injured arteries from control showed vastly increased neointimal areas and decreased lumens compared with sham-operated arteries (Figure 7F and F′). Remarkably, however, the injured arteries of DLP1-K38A-expressing mice revealed a significant reduction in neointima formation (Figure 7G and G′). Calculated values of both the intimal area and intima/media ratio were substantially lower in DLP1-K38A mice compared with control (Figure 7H and I). We also observed an increase of the medial area in DLP1-K38A mice (Figure 7H), suggesting that DLP1-K38A expression in these mice may inhibit VSMC migration to a greater extent than cell proliferation following arterial injury. These in vivo results demonstrated that decreasing mitochondrial fission diminishes intimal hyperplasia by inhibiting VSMC migration. The in vivo efficacy of DLP1-K38A expression in decreasing neointima formation demonstrates that targeting mitochondrial fission can be an effective strategy for limiting intimal hyperplasia in restenosis and the progression of atherosclerotic lesions.

4. Discussion

In this study, we found that chemotactic stimulation of VSMCs increased mitochondrial fission, which was accompanied by enhanced mitochondrial energetics and ROS levels. Importantly, we showed for the first time that the inhibition of mitochondrial fission reduces PDGF-induced VSMC migration and wire injury-induced intimal hyperplasia. Our studies further provided the mechanisms by which chemotactic stimulation activates mitochondrial fission and of how inhibition of mitochondrial fission prevents cell migration.

Previous studies showed that altering mitochondrial morphology affects cell migration. Increased cell migration was observed with mitochondrial fragmentation, whereas mitochondrial elongation decreased cell migration in both metastatic breast cancer cells and chemokine-activated lymphocytes.19,20 We also tested whether the formation of short, small mitochondria is generally associated with migratory cells independently of PDGF stimulation. In wound healing assays with serum-starved MEFs in the absence of PDGF, cells migrated into the wounded area were observed at 24 h after wound scratch. We found that mitochondria in migratory cells in the wounded area were shorter, whereas those in non-migrating cells in unwounded area were tubular (see Supplementary material online, Figure S1), suggesting the correlation between mitochondrial shortening and cell migration.

We found that PDGF stimulation increases DLP1 phosphorylation at serine 616 and promotes mitochondrial binding of DLP1 to activate mitochondrial fission. Several kinases including CDK1, CDK5, PKCδ, and ERK1/2 have been reported to phosphorylate DLP1 at serine 616, resulting in increased mitochondrial fission.2730 It is currently unclear which kinase phosphorylates DLP1 in PDGF stimulation. Our previous report suggests that ERK1/2 is involved in high glucose-induced mitochondrial fission.30 High glucose stimulation increases mitochondrial activity, raising the possibility that ERK1/2-mediated DLP1 phosphorylation and the resulting mitochondrial fission are associated with energetic enhancement of mitochondria. Our current study found that respiration increases promptly upon addition of PDGF, which correlates well with rapid DLP1 phosphorylation and mitochondrial fission, suggesting that PDGF-mediated signalling activates both mitochondrial fission and mitochondrial energetics.

Our findings that PDGF increases both mitochondrial fission and energetics in a similar time frame suggest that there may be a common factor regulating both elements. PDGF stimulation induces Ca2+ release from the endoplasmic reticulum and triggers Ca2+ influx across the plasma membrane to increase the cytosolic Ca2+ concentration.48 An increase in intracellular Ca2+ has been shown to promote mitochondrial fission.22,33,49 Ca2+ in mitochondria also activates dehydrogenases of the TCA cycle as well as ATP synthase.5054 It is possible that the PDGF-evoked Ca2+ increase activates mitochondrial fission through ERK1/2 in the cytosol and, concomitantly, enhances oxidative phosphorylation in mitochondria. Indeed, a PDGF-induced respiration increase was completely abolished by chelating extracellular Ca2+, suggesting its dependency on Ca2+ influx (see Supplementary material online, Figure S2). In this scenario, increased fission may serve to facilitate substrate uptake into the mitochondria by increasing the overall mitochondrial surface area inside the cell. In addition, change in mitochondrial shape through increased fission may alter the organization of respiratory complexes to optimize electron flux for energetic activation.

To evaluate the effect of PDGF on cell migration and to rule out its effect on cell proliferation, we limited our evaluation of the PDGF effects up to 6 h post treatment. Interestingly, PDGF-mediated cell proliferation was also associated with increased mitochondrial fission. It was shown that prolonged incubation of VSMCs in PDGF (24–48 h) increased mitochondrial respiration55 and mitochondrial fragmentation,56 which were observed in association with increased cell proliferation. Treatment of VSMCs with a pharmacological inhibitor of DLP1, mdivi-1, blocked the proliferative effect of PDGF in VSMCs.56 Similarly, the mdivi-1 attenuated chronic hypoxia-induced pulmonary artery hypertension by reducing pulmonary arterial SMC proliferation.57 Although the signalling and downstream effects of PDGF on cell proliferation are different from those on cell migration, these observations, along with our results and the aforementioned observations in lymphocytes and metastatic cancer cells, collectively indicate that increased mitochondrial fission is associated with an augmentation of mitochondrial energetics.

Our study demonstrated that inhibition of mitochondrial fission prevents VSMC migration, consistent with previous reports in other cell types.19,20 Mechanistically, we found that the limited cell migration in fission inhibition was due to a decrease in respiration coupling efficiency. We showed that PDGF induces cell migration by increasing mitochondrial energetics and ROS production. Mitochondrial hyperpolarization promotes ROS generation by the respiratory chain. Under these conditions, uncoupling decreases membrane potential and thus ROS production. Because respiration uncoupling decreases not only the efficiency of ATP synthesis but also the ROS levels, an increase of uncoupling would be an underlying mechanism by which fission deficiency limits PDGF-induced cell migration. Previous studies also showed that inhibition of mitochondrial fission is associated with respiration uncoupling. We reported that inhibition of mitochondrial fission blocked glucose-stimulated insulin secretion in pancreatic β-cells by increasing proton leak.18 Furthermore, DLP1 silencing in glucose-infused hypothalamic tissue also increased proton leak and decreased mitochondrial ROS level along with ROS-mediated downstream metabolic signalling.43

The studies discussed thus far indicate that chemotactic or glucose stimulation enhances mitochondrial energetics with the formation of short, small mitochondria, whereas the long large mitochondria in fission-deficient cells exhibit the decreased respiration coupling. In contrast, there are other reports showing that long mitochondria are energetically active, whereas fragmented mitochondria are dysfunctional, destined to autophagy or apoptosis.17,58,59 However, these studies did not directly analyse respiration leak or coupling. Nevertheless, it was suggested that small and fragmented mitochondria are energetically uncoupled, whereas elongated mitochondria are associated with well-coupled respiration.60 Mitochondrial morphology and energetics may respond differently to stimulations or insults of various nature. Although it is possible that there may be an overarching principle that governs mitochondrial morphology and its associated energetics, further study is necessary before any generalizations are made.

Using transgenic mice expressing DLP1-K38A, we showed that decreasing mitochondrial fission in vivo limits the development of arterial injury-induced intimal hyperplasia. We previously showed that these transgenic mice are also protected from hyperglycaemia-induced oxidative stress by exhibiting decreased ROS production.24 Similarly, this present study demonstrated that DLP1-K38A expression prevented the ROS increase in response to arterial injury, and thereby reduced neointima formation, presumably through increased mitochondrial proton leak. Intimal hyperplasia is associated with restenosis, atherosclerosis, hypertension, and other cardiovascular complications. Our studies show that controlling mitochondrial fission can limit intimal hyperplasia by decreasing cell migration, thereby offering a novel therapeutic strategy to ameliorate the pathological progression of related cardiovascular diseases.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Conflict of interest: none declared.


This study was supported by Georgia Regents University Institutional research fund and the National Institute of Health grants DK061991 to Y.Y. and GM089853 to H.S.


View Abstract