Brain ischemic insult induces cofilin rod formation leading to synaptic dysfunction in neurons

Abstract

Ischemic stroke not only induces neuron death in the infarct area but also structural and functional damage of the surviving neurons in the surrounding peri-infarct area. In the present study, we first identified cofilin rod, a pathological rod-like aggregation, formed in neurons of in vivo ischemic stroke animal model and induced neuronal impairment. Cofilin rods formed only on the ipsilateral side of the middle cerebral artery occlusion and reperfusion (MCAO-R) rat brain and showed the highest density in peri-infarct area. Our real-time live cell imaging, immunostaining and patch clamp studies showed that cofilin rod formation in neurons led to dendritic mitochondrial transportation failure, as well as impairment of synaptic structure and functions. Overexpression of LIM kinase or activation of its upstream regulator Rho, suppressed ischemia-induced cofilin rod formation and showed protective effect on synaptic function and structure impairment in both cultured neurons and MCAO-R rat model. In summary, our results demonstrate a novel mechanism of ischemic stroke-induced neuron injury in peri-infarct area and provide a potential target for the protection of neuronal structure and function against brain ischemia insult.

Keywords

Middle cerebral artery occlusion ischemia cofilin rod brain cortex synaptic dysfunction

Introduction

Stroke is one of the major causes of human death and adult disability worldwide.¹ Ischemic stroke, which is caused by acute cerebral blood flow reduction induced by the blood vessel occlusion or cardiac arrest, accounts for about 80% of all stroke cases.^2,3 During ischemic stroke, neuron death was induced after a series of spatial and temporal pathological changes especially in the infarct area where the blood flow was dropped to under 20% of baseline level.^4,5 On the other hand, in the surrounding areas with milder blood flow decline, also known as peri-infarct area, stroke damages the dendritic structure and causes spine loss without killing the neurons.⁶ As a result, those changes disrupt the neuron circuit and impair the function of the brain.⁴ Since actin dynamics is crucial for synaptic plasticity and maintenance of dendritic spine structure,⁷ the abnormal change of cofilin, a major regulator of actin dynamics,⁸ may play an important role of causing dendritic injury of neurons during ischemic stroke.

Cofilin is a small protein (19 kDa) from the actin-depolymerizing factor/cofilin family.⁹ It is prevalent in the human central nervous system and plays an important role during actin dynamics by severing filamentous actin into monomeric actin.¹⁰ The activity of cofilin is modulated by the phosphorylation of its Ser3. Phosphorylation by LIM kinase (LIMK) causes the deactivation of cofilin, while dephosphorylation by slingshot or chronophin turns cofilin into active form and promotes the turnover of actin filaments.^11,12 In cultured neurons, environmental stress including ATP-depletion, oxidative stress, and high-dose glutamate treatment can induce the formation of cofilin rod, a rod-like structure composed of cofilin and actin.^13–15 Cofilin rods disrupt the cytoskeleton, impairs the integrity of synapse, and induces the loss of dendritic spine.^16,17 Evidence suggests that cofilin rods are involved in neurodegenerative diseases such as Alzheimer’s disease (AD) and Huntington’s disease^13,18 and promote the progress of those diseases.¹⁹

Since some of the inducing factors of cofilin rod formation shares with pathologies during ischemic strokes, including energy depletion and excitotoxicity, researchers have assumed that brain ischemia may also cause cofilin rod formation.¹⁹ However, until now, there is no direct in vivo evidence of cofilin rods in ischemic brain. In this study, we introduced middle cerebral artery occlusion and reperfusion (MCAO-R) rat model and oxygen and glucose deprivation (OGD) cell model to investigate the formation of cofilin rods. We discovered massive cofilin rods in the MCAO-R rat brain, which mostly accumulated at the peri-infarct area, and demonstrated that the progress is partially mediated by N-Methyl-D-aspartate (NMDA) receptor. Cofilin rods induced the blockade of dendritic trafficking, and impairment of synaptic structure and function. Furthermore, elevation of cofilin phosphorylation level successfully inhibited cofilin rod formation and showed rescue effect on synaptic structure and function in both cultured neuron and MCAO-R rat brain.

Materials and methods

Establishment of MCAO-R rat model

Male Sprague Dawley (SD) rats (220–240 g body weight, six-to-seven weeks of age, Shanghai SLAC Laboratory Animal Co., Ltd, China) were used for MCAO-R model establishment. Rats were co-housed in standard cages, under an inverted 12–12-h light–dark circle. They were in a pathogen-free environment and with ad libitum access to food and water. To induce transient focal cerebral ischemia, we first anesthetized rats with inject 10% chloral hydrate i.p. (350 mg/kg). PeriFlux System 5000 Laser Doppler Monitor (Perimed AB) was used for monitoring the efficiency of occlusion and reperfusion. A siliconized filament (Beijing Sha Dong Biology Company) was used to block the blood flow of left middle cerebral artery through internal carotid artery after ligating left external carotid artery including pterygopalatine artery. The filament was removed after 2 h and the rat was sacrificed 22 h after the removal of filament for further tests. Sham control rats received the same surgical procedure without the introduction of the filament. All animal procedures were approved by the Animal Care Committee at Fudan University, performed in accordance with the guidelines established by the Chinese Society for Animals Welfare and complied with the ARRIVE guidelines.

Primary neuron culture

Unless stated otherwise, all tissue culture reagents were obtained from ThermoFisher Scientific. Primary rat neurons were prepared from embryonic days 18–19 SD rats. Briefly, after dissection of the hippocampus, the tissue was rinsed in cold Hank’s balanced salt solution and then digested with 0.05% trypsin–EDTA for 20 min at 37℃, followed by trituration with pipettes in the plating media (Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal bovine serum and 10% F12). After rinsing twice, cells were counted and plated onto Fisherbrand® microscope cover glass (12 mm round; Fisher Scientific) put previously in LabServ® 24-well cell culture plate (Fisher Scientific) or 35 mm petri dishes with 20 mm glass bottom well (In Vitro Scientific) precoated with 0.1 mg/mL poly-D-lysine (Sigma-Aldrich). One day after culture, half of the medium was changed into neuronal culture media (neurobasal media containing 2 mM GlutaMAX™-I Supplement and 2% B27) and twice weekly thereafter. Arabinofuranosyl cytidine (2 mM; Sigma-Aldrich) was added six days after plating. All cells were put in an incubator kept at 37℃ with 5% CO₂.

Transfection

GFP, GFP-cofilin, RFP-cofilin, GFP-mito, and RFP-LIMK plasmids were all synthesized and sequenced at GeneChem China (Shanghai, China). Primary neuronal culture transfection was performed as previously described.²⁰ Briefly, plasmid DNA (for 24-well plate: 1 mg/well; for 35 mm dish: 2 mg/dish) and 9 mg/mL CaCl₂ (Sigma-Aldrich) were gently mixed with 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)-buffered saline. Then, the mixture was added into neuronal culture that had previously transferred into DMEM. After 90 min of incubation, DMEM was replaced by the original medium.

ATP-depletion

For ATP-depletion treatment, 3 mM of 2-deoxyglucose (2-DG) and 5 mM of NaN₃ were added into DMEM medium (ATP-depletion medium). Neuron culture of 14 days in vitro (DIV 14) was transferred into the ATP-depletion medium and placed back into the incubator and maintained at 37℃ with 5% CO₂. After 30 min, the culture was rinsed by fresh DMEM medium twice, transferred back into previous medium and returned to the incubator for 24 h before further test. For Rho activator or ROCK inhibitor treatment, neuron culture was transferred into medium containing 0.25 µg/ml of CN03 or 10 μM Y-27632 (Cytoskeleton) after ATP-depletion.

Live cell imaging

Live cell imaging study was done under an Ultraview VoX live cell scanning unit (PerkinElmer Life Sciences) together with Olympus IX81 microscope. The imaging system was driven by Volocity (Version 5.4, Improvision Ltd) software. Neuron culture was placed in an environmental chamber with 5% CO₂ and 37℃ during the imaging. The culture was firstly scanned by a × 10 0.3 NA objective to find neurons transfected with both RFP-cofilin and GFP-mito. A × 60 1.4 NA objective was then used to trace the movement of mitochondria by GFP. Images were acquired every 15 s for 15 min. Data analysis was performed with Volocity and kymograph was produced by ImageJ software (ImageJ 1.3 Edition, National Institutes of Health, https://imagej.nih.gov/ij/).

OGD

DIV 14 neuron culture was transferred into glucose-free DMEM medium. For glutamate receptor antagonist treatment, 25 µm NMDA receptor antagonist 2-amino-5-phosphonopentanoic acid (AP-5) (Ascent Scientific, UK) and/or 10 µm non-NMDA receptor antagonist 6,7-dinitroquinoxaline-2,3-dione (DNQX) (Ascent Scientific, UK) was added into the glucose-free DMEM medium prior to culture transfer. Then the neuron culture was placed into an anaerobic incubator (Thermo Scientific 1029), which started up previously, reduced the oxygen inside to less than 0.1% and the temperate was kept at 37℃ with 5% CO₂. Ninety minutes after OGD treatment, the culture was taken out and fixed immediately for immunohistochemistry experiment afterwards.

Immunohistochemistry and imaging

Antibodies

Primary antibody against cofilin (1:200; Rabbit polyclonal, ACFL02, Cytoskeleton), MAP2 (1:300; Mouse monoclonal, M4403, Sigma), PSD95 (1:300; Mouse monoclonal, MA1-046, Thermo Scientific) and SV2 (1:150; Mouse monoclonal, 119011, Synaptic Systems) were used. Donkey anti-Mouse IgG (H + L), Alexa Fluor 555/647 (A31570/A31571, Invitrogen) and Donkey anti-Rabbit IgG (H + L), Alexa Fluor 488/594 (A21206/A21207, Invitrogen) at a concentration of 0.1 mg/ml were used as secondary antibodies.

Cultured cells

Neuron cell culture was rinsed with tris-buffered saline (TBS) once and fixed with a solution of 4% paraformaldehyde and 0.1% glutaraldehyde, pH 7.4 for 45 min. Cells were then rinsed three times in TBS by 5 min and permeabilized with methanol pre-chilled to −20℃ for 90 s. Permeabilization solution was removed and washed three times with TBS for 5 min each and blocked with 10% donkey serum diluted with TBS for 2 h. Afterwards, cells were transferred into primary antibody solution diluted with donkey serum blocking solution and incubate overnight at 4℃. The following day, cells were rinsed three times in TBS for 10 min each. Subsequently, cells were incubated with secondary antibodies for 2 h in a dark box. Coverslips were then rinsed three times in TBS for 15 min each and then mounted with Fluoromount (Sigma-Aldrich) on glass slides. All samples were examined by Nikon A1R-A1 Confocal System under ×25/1.10 objective with NIS Elements 4.2 Software (Nikon Instruments) and analyzed with Image-Pro Plus 6.0 (Media Cybernetics).

Brain sections

Brains dissected out from MCAO-R or sham rats were fixed in 4% paraformaldehyde at 4℃ for 24 h, dehydrated with 15% sucrose overnight at 4℃ and further dehydrated with 30% sucrose for two days at 4℃. The brain was then immersed with JUNG tissue freezing medium (Leica Biosystems) and cooled by isopentane in the bath of dry ice and acetone mixture. Coronal sections (15 µm) of cortex were made using a Leica CM1950 cryostat (Leica Biosystems). For staining, sections were washed three times in TBS for 10 min. Sections were permeabilized in 98% methanol for 3 min at −20℃. Sections were washed three times in TBS for 10 min and blocked with 10% donkey serum diluted with TBS for 2 h at room temperature. After blocking, sections were incubated overnight at 4℃ with the primary antibodies. Sections were washed three times in TBS for 5 min each and incubated with secondary antibodies for 2 h in a dark box. Then, sections were washed three times in TBS for 10 min each and then mounted onto glass slides with Fluoromount (Sigma-Aldrich). Staining results were visualized using an Olympus BX51 microscope with DP controller 3.2 software (Olympus) or a Nikon A1R-A1 Confocal System under ×25/1.10 objective with NIS Elements 4.2 Software (Nikon Instruments). Adobe Photoshop CS2 (Adobe Systems Incorporated) was used for merging the images of the same brain section. The images were analyzed with Image-Pro Plus 6.0 (Media Cybernetics, Inc.). Cofilin rod was identified, according to previous studies, as objects with major axis longer than 2 µm, minor axis shorter than 1.5 um and aspect ratio larger than 2.

Electrophysiology

Whole-cell patch clamp recordings were performed using Multiclamp 700B amplifier (Axon Instruments). Patch pipettes were pulled from borosilicate glass by P-97 Flaming/brown micropipette Puller (Sutter Instrument Co.) and fire-polished by MF-830 microforge (NIRISHIGE Group) (4–6 MΩ). The recording chamber was perfused with a bath solution containing 128 mM NaCl, 30 mM glucose, 25 mM HEPES, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 1 μM Tetrodotoxin, pH 7.3 adjusted with NaOH. The pipette solution contained 125 mM KGluconate, 10 mM KCl, 10 mM Tris phosphocreatine, 5 mM EGTA, 10 mM HEPES, 4 mM MgATP, 0.5 mM Na₂GTP, pH 7.3 adjusted with KOH. The membrane potential was clamped at −70 mV. Data were acquired using Clampex software (MDS analytical technologies), sampled rate was set at 5 kHz, with low-pass filtered at 1 kHz, and recorded as long as 15 min. Image of tested neurons was taken by Cool SNAP HQ2 CCD camera (Photometrics) with Metaflour 7.5 Software (MDS analytical technologies) before clamping. Mini excitatory postsynaptic current (mEPSC) data were analyzed by mini analysis program 6.0.3 (Synaptosoft Inc.). Only neurons recorded longer than 5 min with access resistance lower than 30 MΩ and a holding current less than 100 pA were included in data analysis.

Virus injection

SD male rats (140–160 g body weight, 40 days old) were anesthetized with 350 mg/kg chloral hydrate by i.p. injection and mounted in a stereotaxic apparatus. The scalp was removed after sterilization and skull was exposed. The pGC-FU-GFP or pGC FU-GFP-LIMK1 lentivirus (1 × 10⁹ TU/ml) was injected into the cortex by a 10-μl Hamilton syringe with a 26 gauge needle using the injection parameters as follows: AP 0.7 mm, ML −3.0 mm, DP 1.5 mm. The needle was advanced vertically through the burr hole to the depth 0.2 mm deeper than target depth and slowly retracted to the target after 10 min. Two microliters of the lentivirus were infused in 10 min with a steady speed and stayed for 10 more minutes before slowly retracted out. After surgery, animals were allowed to recover for 10–12 days before MCAO.

Statistical analysis

SPSS, version 18.0 (SPSS, Chicago, IL, USA), was used for all statistical analyses and all the data were presented as mean ± SEM. Differences between means were evaluated by two-tailed t-test or one-way ANOVA followed by the Bonferroni or Dunnett post hoc test. Chi-square test was used for nonparametric analysis. P < 0.05 was considered statistically significant.

Results

MCAO-R induces cofilin rod formation in the ipsilateral side of rat brain

To investigate cofilin rod formation during ischemic stroke, we occluded rat left middle cerebral artery for 2 h and then allow the brain to be re-perfused for 22 h. The result of immunostaining showed that cofilin rods formed in the ipsilateral cortex, but not in the contralateral side (Figure 1(a), (c) and (e)).

Figure 1.

Cofilin rods form in the ipsilateral side of MCAO-R rat brain. (a–b) After MCAO-R, MAP2 intensity showed a sudden change in the ischemic side, which was used to identify the ‘border’ of neuronal injury (Cyan dash line). (c) Straighten and magnification of the area around ‘border’ (c, box area in (a, b)) showed formation of cofilin rods corresponding to the decrease of MAP2 signal. (d) The schematic of the definition of non-ischemic area and infarct area according to Bonova et al.²³ for further analysis of cofilin distribution. (e–g) The lateral side of the ‘border’ had the highest cofilin rod (indicated by arrows) burden, which was coincident with the dramatic change of MAP2 signal (n = 5). (f, g) Positive value of the X-axis indicated the lateral side, while the negative value indicated the medial side. **p < 0.01, ***p < 0.001, vs. corresponding area of contralateral side, ##p < 0.01, ###p < 0.001 vs. 0–300 µm lateral to the border. Contra: contralateral side; Peri: peri-infarct area; NI: non-ischemic area. Scale bar: (a, b) 1 mm; (c) 300 µm; (e) 10 µm.

We further counterstained MAP2 with cofilin as an indicator of neuron injury to verify the relationship between cofilin rod formation and ischemic damage.²¹ MAP2 intensity showed a sharp decrease in the ipsilateral cortex (Figure 1(b)) and we defined the location where MAP2 signals reduced by more than a half within 50 µm as the “border” of neuronal damage (Figure 1(c)). For a detailed analysis of cofilin rod distribution, we further divided the ipsilateral side of MCAO-R brain cortex into infarct, peri-infarct and non-ischemic area (Figure 1(d)) according to the previous studies.^22,23 Although the volume of infarction after MCAO-R varied among samples, the border in each experimental rat was always located in the range of the peri-infarct area. It showed that at the area within 300 µm lateral to the border, the density of cofilin rods reached the peak (276.3 ± 29.0/mm², n = 5), which was significantly higher than the medial side of the border (0–300 µm medial to the border: 87.6 ± 11.7/mm², n = 5, p < 0.001), as well as the infarct area (169.8 ± 18.3/mm², n = 5, p < 0.001) (Figure 1(e) and (g)).

Furthermore, immunostaining also showed that cofilin rods were co-localized with MAP2 or on the extension line of the MAP2 signal (Figure 1(e)), indicating cofilin rods induced by ischemic stroke were likely located in the dendrites, similar to what we reported earlier in cultured neurons.¹⁷

Cofilin rods induced by oxygen and glucose deprivation is NMDAR-dependent signaling in cultured neurons

OGD is a commonly used in vitro model to mimic the ischemic stroke in cells. To verify whether cofilin can form rod structure in cultured neurons under ischemic condition, we treated primary cultured neurons with OGD at DIV 14. Immunostaining showed that there were cofilin rods formed after 90 min of OGD treatment with an average of 2.2 ± 0.3 rods/neuron (from 14 cultures of 6 batches), while few rods were observed in the control group neurons (0.1 ± 0.0 rod/neuron from 6 cultures of 6 batches, p < 0.001) (Figure 2(a)). Double staining of cofilin and MAP2 showed that cofilin rods induced by OGD were located in the neuronal dendrites (Figure 2).

Figure 2.

Cofilin rod formation induced by OGD is partially mediated by NMDA receptor activation. (a) After OGD treatment for 90 min, massive rods (indicated by arrows) formed in cultured neurons (n = 14 cultures including 1505 neurons), while there were no rods in control group (n = 6 cultures including 1329 neurons). AP-5 reduced the number of cofilin rods induced by OGD (n = 12 cultures including 2667 neurons). (b) Magnification of the box areas in (a). Enhanced MAP2 signal showed that rods were located in dendrites. Weak MAP2 in the area near and distal to rod indicated local dendritic impairment. (c) Quantified results showed that AP-5 or AP-5 + DNQX (n = 13 cultures including 2840 neurons) treatment can partially but significantly suppress cofilin rod formation induced by OGD. On the other hand, DNQX only did not show significant effect (n = 14 cultures including 2597 neurons). (***p < 0.001 vs. normoxia, ##p < 0.01 vs. OGD). Scale bar: (a), 30 µm; (b), 10 µm.

Previous studies have showed that glutamate over release is one of the early pathological changes after ischemic stroke,⁴ while excess doses of glutamate can induce cofilin rod formation.¹⁵ To investigate whether excitotoxicity was involved in ischemia-induced cofilin rod formation, NMDA receptor antagonist AP-5 and/or AMPA receptor antagonist DNQX was added into culture medium just before OGD treatment. With the treatment of 25 μM AP-5, OGD-induced cofilin rods were significantly decreased (1.0 ± 0.1 rod/neuron, from 12 cultures of 7 batches, p < 0.01 in comparison with OGD treatment alone), but non-NMDAR antagonist DNQX (10 μM) did not have significant effect (1.5 ± 0.3 rod/neuron, from 14 cultures of 7 batches, p = 0.485, in comparison with OGD treatment alone). In addition, neurons treated with both AP-5 and DNQX also had significantly fewer rods (0.8 ± 0.1 rod/neuron from 13 cultures of 7 batches) than those in control group (p < 0.01), but had no difference compared with the AP-5 alone group (Figure 2(a) and (c)). These results indicate that excitotoxicity mediated by glutamate NMDA receptors plays a major role in cofilin rod formation after OGD treatment.

Cofilin rod formation induced mitochondrial transport failure during ischemia

It has been reported that ischemia impairs dendritic mitochondrial transport.²⁴ Since our previous study showed that cofilin rod formation blocks dendritic organelle transport,^15,17 we assumed that mitochondrial transport impairment during ischemic condition might be caused by cofilin rods.

To test this hypothesis, we co-transfected RFP-cofilin with mitochondria-specific marker Mito tagged with GFP into cultured neurons. DMEM containing 5 mM NaN₃ and 3 mM 2-DG was used to induce ATP-depletion for 30 min and then replaced by original medium for 24 h of recovery. After the treatment, cofilin rods were detected in 66.3% (53 out of 80) of transfected neurons, while only 12.2% (12 out of 98) of the neurons in the control group formed cofilin rods (p < 0.001) (Figure 3(a) and (b)).

Figure 3.

Cofilin rod caused the transport failure of mitochondria in cultured neurons after ATP depletion or glutamate treatment. (a, b) Live cell images showing both ATP-depletion and glutamate treatment induced significant increase of cofilin rod formation. Scale bar: 50 µm. (c) Live cell image video clips showing mitochondrial transportation in dendrites indicated by GFP-mito signal. Arrowheads indicate static mitochondria and arrows indicate moving mitochondria. Scale bar: 10 µm. (d–h) Kymograph representing the spatial position of mitochondria and cofilin rod overtime. After ATP depletion and recovery or glutamate treatment, mitochondria transportation was stopped in dendrites with cofilin rods (f, h), while the mitochondria in dendrites without cofilin rod (e, g) kept moving just as those in vehicle group dendrites (d). Scale bar: horizontal: 10 µm, vertical: 5 min. (i) Quantified results of the percentage of moving mitochondria. (j) Within all the moving mitochondria, the average maximum velocity of those in dendrites with cofilin rod (n = 12 in ATP-depletion group and 25 in glutamate group) is significantly lower than that in dendrites without cofilin rod (n = 41 in ATP-depletion group and 52 in glutamate group) or in control group dendrites (n = 34). (**p < 0.01, ***p < 0.001 vs. vehicle; ###p < 0.001 vs. ATP-D no rod group; $$$ p < 0.001 vs. Glu no rod group; ATP-D: ATP-depletion; glu: glutamate).

Live cell imaging of mitochondria indicated by GFP-mito showed that after ATP-depletion treatment, only 14.3% (12 out of 84, from 9 dendrites of different neurons) of mitochondria in cofilin rod containing dendrites moved (maximum velocity over 0.05 μM/s) during 15 min of observation, which was significantly lower than in dendrites without cofilin rod (47.7%, 41 out of 86, from 6 dendrites of different neurons, p < 0.001) or in dendrites of neurons from the control group (47.9%, 34 out of 71, from 13 dendrites of different neurons, p < 0.001) (Figure 3(c) to (f) and (i)). Moreover, the velocity of those moving mitochondria in dendrites with cofilin rod was also significantly lower (0.11 ± 0.02 µm/s, n = 12) than that in dendrites without rod burden (0.62 ± 0.06 µm/s, n = 41, p < 0.001) or in that of the medium control group (0.51 ± 0.07 µm/s, n = 34, p < 0.001) (Figure 3(j)).

We also investigated mitochondria transport after excitotoxicity induced cofilin rod formation. Similar to the result of ATP-depletion treatment, transient treatment of 50 μM glutamate for 30 min induced cofilin rod formation in 79.6% transfected neurons (43 out of 54) 24 h after washout (p < 0.001 compared with medium control) (Figure 3(a) and (b)). Furthermore, there were also less moving mitochondria in dendrites with cofilin rods (25.8%, 25 out of 97, from 14 dendrites of different neurons, p < 0.001 vs. control group), but not in those without rods (59.8%, 52 out of 87, from 11 dendrites of different neurons, p = 0.14 vs. control group) (Figure 3(c) and (d), (g) to (i)). The maximum velocity of the moving mitochondria in dendrites with cofilin rods (0.25 ± 0.03 µm/s, n = 25) was also significantly lower than that in dendrites without rods (0.69 ± 0.05 µm/s, n = 52, p < 0.001) or in medium control group (p < 0.01) (Figure 3(j)).

These results provided evidence that dendritic mitochondria trafficking disruption during ischemic condition is caused by cofilin rod formation.

Cofilin rod formation disrupts synaptic structure and function during ischemia

Research has shown that neurons in the peri-infarct area suffer from synaptic damage,⁶ while there are evidence showing cofilin rods induce disruption of cytoskeleton and synaptic loss.^13,17,25 We treated GFP-cofilin-transfected cultured neurons with ATP-depletion medium to investigate the relationship between cofilin rod formation and synaptic impairment under ischemic condition.

After ATP-depletion treatment, cofilin rods formed in nearly all neurons transfected with GFP-cofilin in ATP-depletion group (97.4%, 38 out of 39), while only 60.4% in the vehicle group (26 out of 43, p < 0.001) (Figure 4(a) and (b)). In the ATP-depletion group, neurons with cofilin rods formed 15.8 ± 1.4 rods in average (n = 38), which were also significantly more than that in the vehicle group (9.6 ± 1.3/neuron, n = 26, p < 0.01) (Figure 4(c)). Immunofluorescent staining showed that ATP-depletion significantly reduced dendrite PSD 95 density. Further analysis showed that synapse loss in area with cofilin rods (18.7 ± 5.2% of vehicle group, n = 14) was more severe than that in dendrites without cofilin rods (47.2 ± 5.6% of vehicle group, n = 14, p < 0.01) (Figure 4(a) and (d)).

Figure 4.

Cofilin rod formation mediates ATP-depletion-induced impairment of the structure and function of synapse. (a–d) Confocal pictures showing that ATP-depletion treatment significantly induced severe cofilin rod (arrow) formation in GFP-cofilin-transfected neurons (b–c), and PSD 95 density decreased more in rod area (n = 14) compared with area without rod (n = 14) (d). The magnified confocal image in (a) showing the reduction of PSD 95 signal in rod area. Scale bar: Vehicle and ATP-D row: 50 µm; Magnified images, 10 µm. *p < 0.05; **p < 0.01; ***p < 0.001 compared with vehicle. ##p < 0.01 compared with ATP-D non-rod area. (e–g) After GFP-cofilin transfection, the mEPSCs frequency of neurons with cofilin rods (n = 8), but not of those without rods (n = 8), is significantly lower than control neurons transfected with GFP (n = 10). ATP depletion treatment further almost diminished mEPSCs (n = 11). Scale bar: 50 µm. ***p < 0.001 compared with GFP; #p < 0.05, ##p < 0.01 compared with GFP-cofilin no rod.

We further studied whether cofilin rod formation would influence synaptic function change by patch clamp recording of mEPSCs. The frequency of mEPSCs of GFP-cofilin-transfected neurons without cofilin rod burden (0.81 ± 0.24 Hz, n = 8) did not show significant difference with neurons transfected with GFP (0.67 ± 0.15 Hz, n = 10). On the other hand, mEPSCs frequency of neurons with cofilin rod after GFP-cofilin transfection (0.24 ± 0.06 Hz, n = 8) was significantly lower than those without rods (p < 0.05). Furthermore, after ATP-depletion treatment, the mEPSC frequency of GFP-cofilin-transfected neurons, which had much more cofilin rods, was almost diminished (0.02 ± 0.01 Hz, n = 11, p < 0.01 compared with neurons without cofilin rods) (Figure 4(e) and (f)). However, there was no significant difference among all the groups for the amplitude of the mEPSCs (Figure 4(g)). These results indicated that ATP-D caused synapse loss and synaptic transmission impairment correlated to the cofilin rod formation.

To investigate whether cofilin rod formation also disrupts synaptic structure during ischemia in vivo, MCAO-R rat brains were immunostained with cofilin and presynaptic marker SV2A (Figure 5(a)). The result showed that in the non-ischemic area of the ipsilateral side, where no cofilin rods were observed, the SV2 density (95.7 ± 16.8%, relative to the contralateral side, n = 4, from three rats) had no difference to that of contralateral side (100 ± 9.8%, n = 6, from three rats). On the other hand, in both ischemic core and peri-infarct area, SV2 density was dramatically reduced (8.3 ± 3.1%, n = 4, from three rats, p < 0.001 and 42.8 ± 5.9%, n = 28, from three rats, p < 0.001, respectively) (Figure 5(b)). Furthermore, in the peri-infarct area, SV2 density showed a negative correlation with local cofilin rod density (lg (Den_SV2A) = 1.79–3.17 × 10⁻³ Den_{cofilin rod}, n = 28, from three rats, p < 0.001, r²= 0.757) (Figure 5(c)). This result indicates the reduction of synapses in MCAO-R rat cortex is directly correlated to the cofilin rod formation in the neuronal dendrites.

Figure 5.

The SV2A signal density in the peri-infarct area is negative correlated with the cofilin rod density in MCAO-R rat cortex. (a) Immunostaining confocal images showing SV2A signal density is reduced in the infarct area and the cofilin rods existing peri-infarct area (brain area defined as Figure 1(c)) from the cortex of a 24-h MCAO-R rat. Note that there were no colocalization between cofilin rod and SV2A. Scale bar: 10 µm. (b) Quantified results showed that SV2A density in both peri-infarct area (n = 28) and infarct area (n = 4) is significantly lower than contralateral side (n = 6) and the ipsilateral non-ischemic area (n = 4) (***p < 0.001 vs. contralateral; ##p < 0.01, ###p < 0.001 vs. non-ischemic area). (c) Logarithm value of SV2A signal density in the peri-infarct area is negative linear correlated with the cofilin rod density (p < 0.001, r²= 0.757).

Elevation of cofilin phosphorylation inhibits cofilin rod formation and rescues neuron function under ischemic condition

The formation of cofilin rod has been proved to depend on the dephosphorylation of cofilin.^17,26 Next, we investigated whether the elevation of cofilin phosphorylation level can suppress cofilin rod formation and protect neuronal function from ischemic insult.

Since cofilin is phosphorylated by LIMK, which was regulated by Rho-Rock,²⁷ we administrated either Rho activator CN03 (Cytoskeleton) to enhance the LIMK activity, or Rock inhibitor Y-27632 (Cytoskeleton) to inhibit the LIMK activity before ATP-depletion treatment to see whether they could affect the ATP-D-induced cofilin rod formation. The results showed that compared with vehicle group CN03 (0.25 µg/ml) treatment significantly reduced the percentage of the neurons forming rods from a 96% (26 out of 27) in vehicle group to 69% (18 out of 26) after ATP-depletion (p < 0.01), and the ATP-depletion also induced a significant reduction of the quantity of cofilin rods from 11.4 ± 2.1 rods/neuron (n = 27) to 3.6 ± 0.8 rods/neuron (n = 26, p < 0.01). On the other hand, the treatment with Y-27632 (10 μM) induced more cofilin rods (18.8 ± 2.1 rods/neuron, n = 40, p < 0.05) (Figure 6(a) and (b)). Furthermore, the immunostaining of PSD95 showed that synapses loss induced by ATP depletion (34.1 ± 4.2% of the control group, n = 14, p < 0.001) can also be partially rescued by treatment of CN03 (60.3 ± 7.8% of the control group, n = 7, p < 0.01 vs. ATP depletion group + vehicle group) (Figure 6(d) and (e)).

Figure 6.

Elevation of cofilin phosphorylation can suppress cofilin rod formation and prevent synaptic transmission impairment. (a–c) Rho activator CN03 suppressed cofilin rod formation after ATP depletion, while ROCK inhibitor Y-27632 induced more rods. Scale bar: 30 µm. *p < 0.05, **p < 0.01 compare with vehicle; ##p < 0.01, ###p < 0.001 compared with CN03. (d, e) After ATP depletion, the PSD 95 density of neurons treated with CN03 (n = 7) was higher than vehicle group (n = 14). Scale bar: upper row: 30 µm, lower row: 5 µm. **p < 0.01 ***p < 0.001 compared with ctrl; ##p < 0.01 compared with ATP-depletion + vehicle. (f–i) LIMK overexpression completely prevented cofilin rod formation induced by ATP-depletion in GFP-cofilin-transfected neurons. Furthermore, the frequency of mEPSCs of neurons transfected with both cofilin and LIMK (n = 14) was significantly higher than neurons transfected only with cofilin after ATP-depletion (n = 11). However, there is no significant difference between the amplitude of mEPSCs between the groups. *p < 0.05, ***p < 0.001. The data of cofilin transfection group are the same as in Figure 4(b), (f) and (g). Scale bar: 50 µm.

Next, we co-transfected RFP-LIMK together with GFP-cofilin into the cultured neurons to overexpress LIMK and enhance cofilin phosphorylation. As expected, the overexpression of LIMK completely inhibited ATP-depletion-induced cofilin rod formation (0%, 0 out of 24, p < 0.001 compared with neurons transfected with GFP-cofilin alone) (Figure 6(f) and (g)).

Furthermore, we recorded the mEPSCs of these double-transfected neurons after ATP depletion treatment. The result showed that compared with neurons transfected only with GFP-cofilin (0.02 ± 0.04 Hz, n = 11) neurons transfected with both RFP-LIMK and GFP-cofilin had a significantly higher mEPSCs frequency (0.25 ± 0.10 Hz, n = 14) than that of the GFP-cofilin-transfected neurons (Figure 6(f) and (h)).

To test whether elevating cofilin phosphorylation level can also suppress cofilin rod formation and have neuronal protective effect during ischemic stroke in vivo, we injected pGC-FU-GFP-LIMK containing lentivirus, or pGC-FU-GFP containing lentivirus as control, into the peri-infarct area of ipsilateral rat cortex 10–12 days before MCAO surgery (Figure 7(a)). Twenty-two hours after reperfusion, the cofilin rod burden of the LIMK-transfected area was significantly lower than surrounding non-transfected area (35.5 ± 13.3%, n = 4), which had a stronger inhibition effect than that transfected with GFP (88.5 ± 8.9%, n = 4, p < 0.05 compared with LIMK group) (Figure 7(b) to (e)). Furthermore, LIMK overexpression also showed protective effect on synapses, indicated by the enhanced SV2 density (199.2 ± 66.4% of the surrounding peri-infarct area, n = 4) after MCAO-R, but GFP transfection failed to show same effect (72.0 ± 19.4% of the surrounding peri-infarct area, n = 4, p < 0.05 compared with LIMK group) (Figure7(d) and (f)).

Figure 7.

LIMK overexpression can inhibit cofilin rod formation and protect synaptic structure after MCAO-R. (a) The schematic of the position of virus injection, which is in the peri-infarct area close to predicted neuronal injury ‘border’. Green area is the approximate area transfected with virus. (b–d) After MCAO-R, in the area with LIMK transfection (green GFP signal), there were significantly less cofilin rods (red) and more SV2A (cyan) signals than surrounding non-transfected area. However, GFP alone transfection did not show same effect. Scale bar: (b): 100 µm; (d): 10 µm. (e, f) The density of cofilin rod and SV2A was analyzed pair wisely at the border of virus-transfected area as arrows in (b, c). n = 4 in each group. *p < 0.05.

Taken together, these results indicate that suppression of cofilin rod formation by enhancing LIMK activity can rescue, at least partially, both the synapse loss and the synaptic transmission dysfunction induced by ischemic insult both in vitro and in vivo.

Discussion

This study is the first to identify the formation of cofilin rod in the brain following ischemic stroke in an animal model. Ischemia-induced rod formation was partially mediated by glutamate excitotoxicity. Cofilin rods disrupted dendrite trafficking-induced loss of synapse and impairment of synaptic function in the neurons with ischemic insult. Finally, we demonstrated that elevation of cofilin phosphorylation level with the activation of LIMK could suppress ischemia-induced cofilin rod formation and prevent the structure and functional damage during ischemia insult.

Since ischemic-like conditions such as ATP-depletion have been proved to induce cofilin rod formation in cultured neurons,¹³ some researchers have suggested cofilin rod formation during ischemic stroke.¹⁹ And this study for the first time provided direct in vivo evidence of cofilin rod formation using MCAO-R rat model. These cofilin rods formed in the ipsilateral side of cortex and their location was mostly located in the area with significant MAP2 signal decrement, an indication of neuronal damage.²¹ On the other hand, there was also a small amount of cofilin rods at the medial side of the “border” without significant MAP2 signal change, suggesting that the cofilin rod formation might be a sensitive indicator of ischemic insult. Furthermore, the place with the highest cofilin rod density was not the infarct area, but at the lateral side of the “border” might be due to cofilin rod being unstable under extracellular condition,¹⁴ and part of rods in the infarct area were dissolved after the neuronal death. We have also found that the existence of cofilin rods in the ischemic brain area is immediately after 2 h of occlusion, although much less than those with 22 h of reperfusion (data not show). These results indicated that cofilin rod is an early and sensitive pathological structure during ischemic stroke but not a result of cell death.

Furthermore, using OGD model of neuron culture, we demonstrated that cofilin rod formation under ischemic condition is partially mediated by excitotoxicity, a key pathology during the early ischemic stroke,²⁸ and NMDA receptor played the major role during this process. The overactivation of NMDAR will cause the elevation of intracellular Ca²⁺, induce cofilin dephosphorylation through calcineurin-slingshot pathway,²⁹ which may finally cause the formation of cofilin rod.³⁰

Previous studies showed that the influx of Ca²⁺ mediates impairment of mitochondria transportation induced by excitotoxicity.^31,32 However, compared with the transient change of [Ca²⁺]_i concentration, the transportation failure remains for a long period.^15,24 This study showed that the long-lasting impairment might be caused by the formation of cofilin rod. These blockade effect of cofilin rods might be caused by direct occlude of the dendrite,³³ entrapment of mitochondria,³⁴ and/or disorganization of microtubule network.¹⁷

We further showed that under ischemic condition, cofilin rod damages synaptic structure and impairs its function. Since cofilin plays an important role in maintaining the structure of dendritic spine and the induction of LTP,^35,36 the synaptic impairment might be a result of disruption of cofilin function by over-dephosphorylation and/or the recruit of cofilin into rods.^14,17,26 On the other hand, local energy failure caused by the disruption of mitochondria transportation may also be involved in synaptic loss.³⁷

As mentioned above, cofilin dephosphorylation plays an important role of cofilin rod formation. We showed that overexpression of LIMK or upregulation of its activity by activation of its upstream regulator Rho³⁸ suppressed the cofilin rod formation induced by cofilin overexpression and ischemic insult not only in cultured neurons but also in MCAO-R rat brain. More importantly, our results further demonstrated that LIMK overexpression had a rescue effect on synaptic structure and function in both in vitro and in vivo models. Recently, a study showed that inhibition of cofilin dephosphorylation by calcineurin inhibitor FK-506 can protect neuron from OGD-induced cell death.³⁹ Thus, cofilin phosphorylation enhancement might be a novel way of neuron protection after ischemic stroke.

Functional recovery and rehabilitation are a crucial point of ischemic stroke treatment. Recent clinical approaches are gradually reducing the “door-to-needle” time of ischemic stroke patients.^6,20 Timely reperfusion may prevent the peri-infarct zone join into the infarct area and avoid cell death,⁴⁰ but the neuron injury such as damage of axon and synapse loss is still unavoidable.⁵ Therefore, the protection of synaptic structure during ischemia and reperfusion afterward becomes crucial for better prognosis.⁵ Cofilin have been proved to be crucial for receptor transportation, neurite outgrowth as well as neuronal migration,^41–43 and elevation of cofilin phosphorylation has been proved to facilitate recovery after ischemic insult.⁴⁴ Thus, the maintenance or restoration of cofilin function by preventing rod formation or dissolve existing rods might be beneficial for rebuilding synaptic connection after ischemic stroke. However, there are researchers indicating that cofilin rod may have protective effect in the super-acute phase of ischemia by saving the last remaining energy,⁴⁵ and Rho inhibitor has been proved to have neuron protective effect in multiple studies.^46,47 Therefore, the time point and the method of intervention may be critical and need further investigation for cofilin rod-targeted therapy.

In summary, we investigated cofilin rod formation in MCAO-R rats and studied its pathologic effect. Our result also showed that the suppression of cofilin rod formation by inhibition of cofilin dephosphorylation might be a novel target for ischemic stroke treatment by preserving the synaptic structure and function, better recovery of neuronal network and improving the prognosis.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grants from the Nature Science Foundation of China (31471027, 31771188, 91332202) to YW and from Shanghai Ninth People’s Hospital (JY2011A18) to LS.

Acknowledgements

We thank Bevis Wang for helping editing the English language.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

BenC, LS, LC and YW designed the study and developed the methodology; LS, BinC, HX and YZ performed MCAO-R surgery procedures; BenC, BinC, HG and HX performed immunohistology staining studies; BenC and GW cultured primary neuron cells; BenC and MJ performed live cell imaging studies; BenC, and YH performed electrophysiology studies; BenC, LS, BinC and HX performed data analysis; BinC and LC performed statistic analysis. BenC, LS, XL and YW wrote the manuscript. All authors approved the version to be published.

References

Donnan

Fisher

Macleod

, et al. Stroke. Lancet 2008; 371: 1612–1623.

Dalkara

Moskowitz

. Mechanisms, challenges and opportunities in stroke. Nat Rev Neurosci 2003; 4: 399–415.

Heiss

. The concept of the penumbra: can it be translated to stroke management? Int J Stroke 2010; 5: 290–295.

Dirnagl

Iadecola

Moskowitz

. Pathobiology of ischaemic stroke: an integrated view. Trends Neurosci 1999; 22: 391–397.

Murphy

Corbett

. Plasticity during stroke recovery: from synapse to behaviour. Nat Rev Neurosci 2009; 10: 861–872.

Kamal

Sheng

Xian

, et al. Delays in door-to-needle times and their impact on treatment time and outcomes in get with the guidelines – stroke. Stroke 2017; 48: 946–954.

Andrianantoandro

Pollard

. Mechanism of actin filament turnover by severing and nucleation at different concentrations of ADF/cofilin. Mol Cell 2006; 24: 13–23.

DesMarais

Ghosh

Eddy

, et al. Cofilin takes the lead. J Cell Sci 2005; 118: 19–26.

Sarmiere

Bamburg

. Regulation of the neuronal actin cytoskeleton by ADF/cofilin. J Neurobiol 2004; 58: 103–117.

10.

Maciver

Hussey

. The ADF/cofilin family: actin-remodeling proteins. Genome Biol 2002; 3: reviews3007.

11.

Niwa

Nagata-Ohashi

Takeichi

, et al. Control of actin reorganization by Slingshot, a family of phosphatases that dephosphorylate ADF/cofilin. Cell 2002; 108: 233–246.

12.

Gohla

Birkenfeld

Bokoch

. Chronophin, a novel HAD-type serine protein phosphatase, regulates cofilin-dependent actin dynamics. Nat Cell Biol 2005; 7: 21–29.

13.

Minamide

Striegl

Boyle

, et al. Neurodegenerative stimuli induce persistent ADF/cofilin-actin rods that disrupt distal neurite function. Nat Cell Biol 2000; 2: 628–636.

14.

Minamide

Maiti

Boyle

, et al. Isolation and characterization of cytoplasmic cofilin-actin rods. J Biol Chem 2010; 285: 5450–5460.

15.

Chen

Jiang

Zhou

, et al. Both NMDA and non-NMDA receptors mediate glutamate stimulation induced cofilin rod formation in cultured hippocampal neurons. Brain Res 2012; 1486: 1–13.

16.

Jang

Han

Lee

, et al. Cofilin expression induces cofilin-actin rod formation and disrupts synaptic structure and function in aplysia synapses. Proc Natl Acad Sci USA 2005; 102: 16072–16077.

17.

Cichon

Sun

Chen

, et al. Cofilin aggregation blocks intracellular trafficking and induces synaptic loss in hippocampal neurons. J Biol Chem 2012; 287: 3919–3929.

18.

Munsie

Caron

Atwal

, et al. Mutant huntingtin causes defective actin remodeling during stress: defining a new role for transglutaminase 2 in neurodegenerative disease. Hum Mol Genet 2011; 20: 1937–1951.

19.

Bamburg

Bernstein

Davis

, et al. ADF/Cofilin-actin rods in neurodegenerative diseases. Curr Alzheimer Res 2010; 7: 241–250.

20.

Kamal

Holodinsky

Stephenson

, et al. Improving door-to-needle times for acute ischemic stroke: effect of rapid patient registration, moving directly to computed tomography, and giving alteplase at the computed tomography scanner. Circ Cardiovasc Qual Outcomes 2017; 10: pii: e003242.

21.

Dawson

Hallenbeck

. Acute focal ischemia-induced alterations in MAP2 immunostaining: description of temporal changes and utilization as a marker for volumetric assessment of acute brain injury. J Cereb Blood Flow Metab 1996; 16: 170–174.

22.

Ashwal

Tone

Tian

, et al. Core and penumbral nitric oxide synthase activity during cerebral ischemia and reperfusion. Stroke 1998; 29: 1037–1046. discussion 1047.

23.

Bonova

Burda

Danielisova

, et al. Development of a pattern in biochemical parameters in the core and penumbra during infarct evolution after transient MCAO in rats. Neurochem Int 2013; 62: 8–14.

24.

Chang

Reynolds

. Mitochondrial trafficking and morphology in healthy and injured neurons. Prog Neurobiol 2006; 80: 241–268.

25.

Chen

Wang

. Cofilin rod formation in neurons impairs neuronal structure and function. CNS Neurol Disord Drug Targets 2015; 14: 554–560.

26.

Huang

Minamide

Bamburg

, et al. Chronophin mediates an ATP-sensing mechanism for cofilin dephosphorylation and neuronal cofilin-actin rod formation. Dev Cell 2008; 15: 691–703.

27.

Chen

Gehler

Shaw

, et al. Cdc42 participates in the regulation of ADF/cofilin and retinal growth cone filopodia by brain derived neurotrophic factor. J Neurobiol 2006; 66: 103–114.

28.

Kostandy

. The role of glutamate in neuronal ischemic injury: the role of spark in fire. Neurol Sci 2012; 33: 223–237.

29.

Wang

Shibasaki

Mizuno

. Calcium signal-induced cofilin dephosphorylation is mediated by Slingshot via calcineurin. J Biol Chem 2005; 280: 12683–12689.

30.

Moriyama

Iida

Yahara

. Phosphorylation of Ser-3 of cofilin regulates its essential function on actin. Genes Cells 1996; 1: 73–86.

31.

Rintoul

Filiano

Brocard

, et al. Glutamate decreases mitochondrial size and movement in primary forebrain neurons. J Neurosci 2003; 23: 7881–7888.

32.

Chang

Honick

Reynolds

. Mitochondrial trafficking to synapses in cultured primary cortical neurons. J Neurosci 2006; 26: 7035–7045.

33.

Bamburg

Wiggan

. ADF/cofilin and actin dynamics in disease. Trends Cell Biol 2002; 12: 598–605.

34.

Davis

Maloney

Minamide

, et al. Mapping cofilin-actin rods in stressed hippocampal slices and the role of cdc42 in amyloid-beta-induced rods. J Alzheimers Dis 2009; 18: 35–50.

35.

Shi

Pontrello

DeFea

, et al. Focal adhesion kinase acts downstream of EphB receptors to maintain mature dendritic spines by regulating cofilin activity. J Neurosci 2009; 29: 8129–8142.

36.

Chen

Rex

Casale

, et al. Changes in synaptic morphology accompany actin signaling during LTP. J Neurosci 2007; 27: 5363–5372.

37.

Okamoto

Hayashi

, et al. The importance of dendritic mitochondria in the morphogenesis and plasticity of spines and synapses. Cell 2004; 119: 873–887.

38.

Bernard

. Lim kinases, regulators of actin dynamics. Int J Biochem Cell Biol 2007; 39: 1071–1076.

39.

Madineni

Alhadidi

Shah

. Cofilin inhibition restores neuronal cell death in oxygen-glucose deprivation model of ischemia. Mol Neurobiol 2016; 53: 867–878.

40.

Kaplan

Brint

Tanabe

, et al. Temporal thresholds for neocortical infarction in rats subjected to reversible focal cerebral ischemia. Stroke 1991; 22: 1032–1039.

41.

Lee

Fan

, et al. ADF/cofilin-mediated actin dynamics regulate AMPA receptor trafficking during synaptic plasticity. Nat Neurosci 2010; 13: 1208–1215.

42.

Bellenchi

Gurniak

Perlas

, et al. N-cofilin is associated with neuronal migration disorders and cell cycle control in the cerebral cortex. Genes Dev 2007; 21: 2347–2357.

43.

Meberg

Ono

Minamide

, et al. Actin depolymerizing factor and cofilin phosphorylation dynamics: response to signals that regulate neurite extension. Cell Motil Cytoskeleton 1998; 39: 172–190.

44.

Catano

Echeverry

, et al. Urokinase-type plasminogen activator promotes dendritic spine recovery and improves neurological outcome following ischemic stroke. J Neurosci 2014; 34: 14219–14232.

45.

Bernstein

Chen

Boyle

, et al. Formation of actin-ADF/cofilin rods transiently retards decline of mitochondrial potential and ATP in stressed neurons. Am J Physiol Cell Physiol 2006; 291: C828–839.

46.

Rikitake

Kim

Huang

, et al. Inhibition of Rho kinase (ROCK) leads to increased cerebral blood flow and stroke protection. Stroke 2005; 36: 2251–2257.

47.

Jeon

Jeong

Park

, et al. The Rho-kinase (ROCK) inhibitor Y-27632 protects against excitotoxicity-induced neuronal death in vivo and in vitro. Neurotox Res 2013; 23: 238–248.