Rho kinase inhibition protects CA1 cells in organotypic hippocampal slices during in vitro ischemia

BR A IN RE S E A RCH 1 3 16 ( 20 1 0 ) 9 2 – 1 00

available at www.sciencedirect.com

www.elsevier.com/locate/brainres

Research Report

Rho kinase inhibition protects CA1 cells in organotypic hippocampal slices during in vitro ischemia Lennart Gisselsson, Håkan Toresson, Karsten Ruscher, Tadeusz Wieloch⁎ Laboratory for Experimental Brain Research, Wallenberg Neuroscience Center, Lund University, BMC A13, S-22184 Lund, Sweden

A R T I C LE I N FO

AB S T R A C T

Article history:

The actin cytoskeleton is a dynamic superstructure that regulates multiple cellular

Accepted 23 November 2009

functions and that has been implicated in cell death regulation. We investigated whether

Available online 22 December 2009

modulating the neuronal actin cytoskeleton polymerization by Rho-GTPase kinase (ROCK) inhibition influences cell death in hippocampal neuronal cultures and in murine

Keywords:

organotypic hippocampal slice cultures subjected to in vitro ischemia (IVI). During IVI,

Apoptosis

spines on vehicle treated hippocampal neurons collapsed and large dendritic actin

Cell death

aggregates were formed. Following ROCK inhibition by Y27632, the actin aggregates were

Cytoskeleton

markedly smaller while large filopodia extended from the dendritic trunk. Y27632 also

Ischemia

provided strong neuroprotection of hippocampal pyramidal CA1 neurons, which was of

Latrunculin

similar magnitude as protection by NMDA receptor blockade. Likewise, treatment with the

Spine

F-actin depolymerizing agent latrunculin during IVI diminished actin aggregation and mitigated cell death following IVI. We propose that ROCK inhibition protects neurons against ischemic damage by disrupting actin polymerization thereby mitigating NMDA receptor induced toxicity and releasing ATP bound to actin for cellular energy use. We conclude that ROCK inhibitors abrogate multiple detrimental processes and could therefore be useful in stroke therapy. © 2009 Published by Elsevier B.V.

1.

Introduction

Ischemic brain damage is multifactorial with injurious processes differentially activated over time (Endres et al., 2008). It is therefore likely that future therapy against ischemic brain damage will involve several treatment modalities that are directed against multiple detrimental mechanisms. Alternatively, a “hub” mechanism that regulates many detrimental processes could be targeted for neuroprotection by a monopharmacological approach. The actin cytoskeleton is such a “hub” since it is dynamic and regulates complex processes such as cell motility and mobility, and the organization of

⁎ Corresponding author. Fax: +46 46 2220615. E-mail address: [email protected] (T. Wieloch). 0006-8993/$ – see front matter © 2009 Published by Elsevier B.V. doi:10.1016/j.brainres.2009.11.087

organelles and cellular processes including receptor/channel trafficking and activity (Schulz et al., 2004). Normally a high cellular ATP/ADP ratio ensures that actin monomers are stored in complex with ATP keeping the free monomeric actin level low (Nicholson-Dykstra et al., 2005; Pollard and Borisy, 2003). Actin in neuronal spines is therefore a storage of ATP that can be used to compensate for ATP loss during oxygen and glucose deprivation. During ischemia, tissue ATP/ADP ratio decreases and ATP bound to actin is exchanged for ADP thereby releasing ATP for cellular use. Concomitantly, actin polymers dissociate, spines collapse (Gisselsson et al., 2005) and ADP-actin levels increase (Hasbani

BR A I N R ES E A RC H 1 3 1 6 ( 2 01 0 ) 9 2 – 1 00

et al., 2001). The normal turnover of actin is highly regulated, and monomeric globular actin (G-actin) polymerizes to fibrous actin (F-actin) under the control of multiple regulatory proteins and cell signaling processes (Carlisle and Kennedy, 2005; Nicholson-Dykstra et al., 2005; Pollard and Borisy, 2003). The small Rho-GTPases, including RhoA, Rac1 and Cdc42 (Ladwein and Rottner, 2008; Tashiro and Yuste, 2004), and downstream regulatory proteins, such as the Rho kinase (ROCK) (Luo, 2002; Tada and Sheng, 2006), regulate actin polymer structure and hence determine spine shape and function. The disruption of the actin cytoskeleton is neuroprotective. Treatment with cytochalasin D decreases infarct size following MCA occlusion (Laufs et al., 2000), and the F-actin severing protein gelsolin prevents neuronal death and diminishes infarct size (Harms et al., 2004). Also, Rho kinase modulates cell death, and the ROCK inhibitor fasudil is neuroprotective in rodents subjected to global (Satoh et al., 1996) and focal ischemia (Hitomi et al., 2000), though the mechanism of neuroprotection is not clarified. We have established an in vitro model based on the murine organotypic hippocampal slice where oxygen and glucose deprivation, induced in a medium similar to that found in the brain during brain ischemia, selectively damages neurons in the CA1 region (Rytter et al., 2003). In this in vitro ischemia (IVI) model, blockade of the NMDA receptor, removal of extracellular calcium, blockade of the mitochondrial permeability transition pore and a decrease of incubation temperature by 4 degrees provide a marked neuroprotection, while inhibition of caspases or treatment with anti-oxidants or free radical scavengers is ineffective (Cronberg et al., 2004, 2005; Rytter et al., 2005). Using these slice culture preparations and isolated hippocampal neurons from actin-GFP overexpressing mice (Fischer

93

et al., 1998, 2000; Gisselsson et al., 2005), we aimed to study the importance of F-actin integrity for the survival of hippocampal neurons and spine morphology and dynamics after IVI using the actin depolymerizing agent latrunculin and the ROCK inhibitor Y27632 (Pellegrin and Mellor, 2005).

2.

Results

2.1. Rho kinase inhibition protects against neuronal damage in organotypic tissue cultures In organotypic hippocampal tissue cultures, 12 min of IVI caused a progressive increase in cell death in the CA1 region over the following 48 h of recovery (Fig. 1). The RhoA kinase inhibitor Y27632 (100 μM) applied 60 min before IVI, during IVI and the subsequent recovery phase diminished cell death by approximately 90%. Y27632 (100 μM) alone was not toxic to control slices. The protective effect of Y27632 was of similar magnitude as that of NMDA receptor blockade with MK-801 (20 μM) (Fig. 2A). This treatment also diminished cell death by approximately 85% (Fig. 2B).

2.2.

Effect of Rho kinase inhibition on dendritic spines

To analyze the cellular mechanisms underlying Y27632 protection, we next examined the effect of Rho kinase inhibition on spine morphology in hippocampal neurons. Neurons responded within 15–30 min to exposure to 100 μM Y27632 with dramatic changes in morphology (Fig. 3A). Spines adopted a filopodia-like structure with a gain in length and

Fig. 1 – Effect of Y27632 treatment on cell death following in vitro ischemia in murine hippocampal slice cultures. (A) Representative fluorescence micrographs of PI-stained slice cultures exposed to IVI alone (left column), treated with Y27632 (100 μM) and exposed to IVI (middle column) and control stimulated slice cultures (right column). Images were captured immediately before IVI (0 h), at 24 and 48 h of recovery. (B) Quantification of cell damage presented as mean propidium iodide fluorescence intensity (MFI) in the CA1 region after IVI. White bars represent cultures exposed to 12 min of IVI, light gray bars are cultures exposed Y27632 (100 μM) during IVI and the subsequent recovery period. Black bars are showing the MFI of control slices treated with Y27632 (100 μM). Data were obtained from 4 to 6 slices in each group. Experiments were performed in triplicate and data are expressed as mean ± SEM. Asterisks indicate a significance level of p < 0.005 using two-way ANOVA with Scheffé's post-hoc test.

94

BR A IN RE S E A RCH 1 3 16 ( 20 1 0 ) 9 2 – 1 00

Fig. 2 – The effect of Y27632 and MK-801 treatment on cell death in murine hippocampal slice cultures following IVI. (A) Representative micrographs of PI-stained cultures exposed to IVI for 12 min. The first column represents slices exposed to IVI only; the second column shows slices preincubated with MK-801 (20 μM) for 2 h prior to IVI; the right column shows slices treated with Y27632 (100 μM) and exposed to IVI. Images were captured before IVI (0 h), and at 24 and 48 h of recovery. Data were obtained from 4to 6 slices in each group and experiments were performed in triplicate. Data are presented as mean ± SEM. Asterisks indicate a significance level of p < 0.05 using two-way ANOVA with Scheffé's post-hoc test.

loss of a distinct head. The number of filopodia-like protrusions was 3.0 ± 3.2 in the control and 15.2 ± 4.5 in Y27632 treated neurons, respectively, counted per 50 μm of dendritic length (Fig. 3B). Also, spine length in the non-filopodia-like spines increased from 1.64 ± 0.32 μm to 1.98 ± 0.28 μm (Fig. 3C). After a 24 h exposure to Y27632, filopodia-like spines regained their head size (Fig. 3D), but were still significantly longer than prior to treatment (Fig. 3D). The average length on the whole spine population was increased after 24 h (Fig. 3E).

2.3. Effect of ROCK inhibition on dendritic spines during IVI Evidently the Rho-GTPase regulates formation of dendritic filopodia and the shape of spines in our experimental system as also reported earlier (Li et al., 2000; Nakayama et al., 2000). We next studied how these spines and dendrites were affected by IVI in the presence of Y27632. In vehicle treated cultures, 15 min of IVI caused spine collapse and dendritic swelling at the position of the former spines and with accumulation of large actin aggregates in the dendritic swellings (Fig. 4A). Also, an increased formation of filopodia was observed as described previously (Gisselsson et al., 2005). In cultures treated with Y27632 (100 μM) we found that spines collapsed but the actin aggregates formed were smaller. Hence, the swellings decreased and the number of actin aggregates increased compared to vehicle treated cultures (Fig. 4B), but also filopodia formation increased (Fig. 4C).

2.4.

Effects of latrunculin on cell death and dendritic spines

To assess whether a general F-actin depolymerization also provides neuroprotection, hippocampal slices were treated with latrunculin. We found that latrunculin (5 μM) rapidly

induced morphological changes, and after 20 min of treatment, the signal from GFP-actin was reduced and the size of spine heads diminished, indicating actin depolymerization (Fig. 5A). After 1 h, only remnants of actin filaments were present in spine shafts (Fig. 5B). To investigate whether the spine morphology was changed or if only the fluorescent actin filaments were depleted, embryonic primary neuronal cultures were transfected with EGFP to visualize spine morphology independent of actin filaments. Fig. 5C shows prominent EGFP expressing spine heads that collapse into thin irregular protrusions after latrunculin exposure (Fig. 5D). Spine motility is a sign of turnover of dynamic actin (Gisselsson et al., 2005). In a time-lapse sequence of 30 micrographs, computer generated spine outlines were stacked to illustrate motility (Figs. 6A, B). Two perpendicular lines enclosed in every spine outline measuring the width/length (w/l) were placed and the w/l ratio was calculated. These ratios were then plotted against time (Figs. 6C, D). Numerical values of the variability in spine (w/l) ratio over 3 min time periods were then assessed (Table 1). Using this procedure we found that spine motility significantly decreased by approximately 50% after addition of latrunculin. Next we studied how latrunculin treatment affected cell death after IVI. Slices were incubated with either 0, 0.5, 5, and 50 μM of latrunculin for 1 h, then exposed to 12 min of IVI, still in the presence of latrunculin, and then returned to growth medium without latrunculin added during recovery. Cell death was assessed in the CA1 region at 24 and 48 h after the insult. Latrunculin, dose-dependently decreased cell death in the CA1 region when assessed at 48 h of recovery (Fig. 7A). Cell death decreased by 32% following treatment with 0.5 μM latrunculin, and by 50% following treatment with 50 μM. Five micromolars of latrunculin added prior to IVI, enhanced the spine collapse during IVI (Fig. 7B), and fewer swellings were

BR A I N R ES E A RC H 1 3 1 6 ( 2 01 0 ) 9 2 – 1 00

95

Fig. 3 – Inhibition of the Rho/Rho kinase pathway by Y27632 induces fast filopodia formation and spine elongation in hippocampal neurons. (A) Inverted fluorescence micrographs from a time-lapse series exposed to Y27632 (100 μM) for 30 min. Black arrowheads show filopodia-like transformation of spines. White arrowheads show newly formed filopodia. Scale bar 5 μm. (B) The effect of 100 μM Y27632 on the number of filopodia-like protrusions/50 μm dendrite and (C) the length of protrusions with spine morphology. (D) Dynamics of filopodia formation in a dendritic segment after treatment for 24 h. (E) Mean spine length in a dendritic segment of Y27632 treated neurons at 24 h. Data are presented as means ± SD. The mean values at each time point were compared using Student's t-test. Asterisks indicate significant level of p < 0.01. Protrusions longer than 3.5 μm long and without spine head were defined as filopodia.

formed compared to vehicle treated cells exposed to IVI. The IVI-induced filopodia seen in vehicle treated neurons were abolished. The number of spines/10 μm dendrite was calculated and we found that 92% of all spines were lost in latrunculin treated cultures compared to 64% in vehicle treated IVI exposed cultures (Fig. 7C).

3.

Discussion

We confirm earlier observations showing that dendritic spines of hippocampal neurons collapse during exposure to oxygen and glucose deprivation. Concomitantly, dendritic swellings with aggregated actin are formed and multiple membrane protrusions, filopodia, are induced along the dendrites

(Gisselsson et al., 2005; Hasbani et al., 2001; Murphy et al., 2008). We now show that ROCK inhibition by Y-27632 and actin depolymerization by latrunculin protect CA1 neurons following in vitro ischemia. The neuroprotective effect was associated with a decrease in the dendritic swellings and the associated actin aggregates, and a decrease dynamic F-actin in spine heads of hippocampal neurons. We have earlier shown that neuronal death in the CA1 region of organotypic hippocampal slices induced by IVI strongly depends on NMDA receptor activation, extracellular calcium ions, and activation of the mitochondrial permeability transition pore (MPTP) (Cronberg et al., 2004, 2005; Rytter et al., 2005). Our in vitro systems therefore model the NMDA receptor mediated cell death cascade during ischemia and it is reasonable to assume that the neuroprotection provided by Y-

Fig. 4 – Spine dynamics and dendritic swelling after IVI—influence of ROCK inhibition. (A) Dendritic segment from a GFP-actin expressing hippocampal neuron exposed to 15 min of IVI illustrating dendritic swellings with actin polymer networks with sprouting filopodia. (B) Dendrite from sister culture exposed to 15 min IVI insult in the presence of Y27632. Note, the reduced size of actin containing aggregates and formation of large filopodia. (C) The number of filopodia per 100 μm dendritic segment at the end of the IVI insult. White bar—IVI (control); black bar—IVI and Y27632 (100 μM) treatment during IVI. Data are means of ± SD. Asterisk indicates significant level of p < 0.001 using Student's t-test.

96

BR A IN RE S E A RCH 1 3 16 ( 20 1 0 ) 9 2 – 1 00

Fig. 5 – Actin interacting agents induces morphological changes in dendritic spines. Fluorescence micrographs of a dendritic segment from a GFP-actin expressing neuron before (A) and 1 h after exposure of 5 μM latrunculin (B). Confocal projections of a dendritic segment from a hippocampal neuron transfected with EGFP before (C) and 2 h after the addition of 5 μM latrunculin (D). White filled arrowheads indicate spine heads that collapse into thin protrusions.

27632 and latrunculin interrupts the detrimental activation of the NMDA receptor–calcium–MPTP cascade. This is in concordance with the finding that the ROCK inhibitor fasudil protects against NMDA-induced cell death in retina (Kitaoka et al., 2004) and cortical neurons (Yamashita et al., 2007). Also, ROCK inhibition increases the filopodia formation with a decrease of the size of the PSD, and a decrease in the number of synaptic NMDA receptor clusters and AMPA receptor labeled spines (Allison et al., 1998). Moreover, actin depolymerization inhibits NMDA receptor function (Lei et al., 2001; Rosenmund and

Westbrook, 1993), and decreases trafficking of NMDA receptors to the postsynaptic membrane (Halpain, 2003). Protection by ROCK inhibition and actin depolymerization could also be due to a decrease in the speed of ATP depletion during subsequent IVI. Latrunculin treatment prevents actin polymerization by binding to G-actin, and therefore interrupts the treadmilling of F-actin, and preserves ATP levels (Bernstein and Bamburg, 2003). This is evident in latrunculin treated cells, where we found that depolymerization is not complete and remnant spine shafts remain, indicative of a pool of “stable” F-actin providing the backbone of the spine insensitive to latrunculin action. In contrast, the high turnover dynamic F-actin pool in the spine heads responsible for the observed spine motility, is depleted by latrunculin treatment (Halpain, 2003). The collapse of dendritic spines during ATP depletion, was recently demonstrated by in vivo 2-photon microscopy studies of cortical pyramidal neurons in GFP overexpressing mice subjected to ischemia (Murphy et al., 2008). Since the dynamic F-actin is turning over rapidly under the hydrolysis of ATP, the collapse of spine during ATP depletion may therefore initially be a protective response to preserve energy (Bernstein and Bamburg, 2003). Results in the present study are in line with this notion. Latrunculin treatment decreases spine motility and enhances spine collapse during IVI preventing turnover of dynamic actin with the concomitant expenditure of ATP. Latrunculin reduces ATP consumption in neurons by 50% (Bernstein and Bamburg, 2003). The actin-GFP containing fluorescent aggregates formed in dendrites during IVI are indicative of the formation of new actin containing fibers (Gisselsson et al., 2005) or actin rods

Fig. 6 – Effect of latrunculin on dendritic spine motility. Computer generated outlines of a dendritic segments showing spine motility before (A) and 5 min after (B) addition of latrunculin (5 μM). Width and length of six individual spines before (C) and after (D) latrunculin addition are calculated, normalized and plotted against time.

BR A I N R ES E A RC H 1 3 1 6 ( 2 01 0 ) 9 2 – 1 00

Table 1 – Effect of latrunculin B on spine motility. a.

Latrunculin B (5 μM) Before After

0–3 min

3–6 min

6–9 min

0.75 ± 0.28 0.41 ± 0.15⁎

0.73 ± 0.22 0.42 ± 0.11⁎

0.80 ± 0.29 0.39 ± 0.13⁎

a The difference between the highest and the lowest “shape ratio” values during 3 min at three time intervals were calculated and then the mean values from six spines were calculated. Values for the same spines sets before and after drug was added were compared. (⁎) denotes p < 0.05, paired Student's t-test. Values are means ± SD.

complexed with ADF/cofilin (Bernstein et al., 2006). If the aggregates are of moderate size this process may be protective since the actin/ADF cofilin rods preserve the mitochondrial membrane potential during excitotoxicity (Bernstein et al., 2006). However, large aggregates may contribute to dendritic swellings since uncapped barbed ends of the actin fibers may form clusters with plasma membrane channels and receptor

97

protein complexes (Hebert et al., 2008) causing osmotic swelling due to ion leakage. The small Rho-GTPases stabilize actin filaments (Meng et al., 2002; Maekawa et al., 1999). Hence, ROCK inhibition by Y27632 may increase F-actin severing thereby decreasing the size of the actin aggregates accumulating during IVI. Our data are in line with earlier findings that the ROCK inhibitor fasudil is neuroprotective in rodents subjected to global (Satoh et al., 1996) and focal ischemia (Hitomi et al., 2000). This neuroprotective action has been attributed to improved hemodynamics and prevention of neutrophil infiltration (Satoh et al., 2001). Here, we demonstrate that in addition to the anti-inflammatory effect and its vascular action, ROCK inhibition provides a robust neuroprotection in brain tissue by depressing actin turnover and NMDA receptor toxicity. We conclude that ROCK inhibitors target multiple detrimental processes after brain ischemia such as inflammatory cells, the neurovascular unit, and excitotoxicity, and therefore are suitable agents for therapeutic intervention following brain ischemia.

Fig. 7 – The effect of latrunculin treatment on cell death following in vitro ischemia in murine hippocampal slice cultures. (A) Cultures were incubated for 1 h with 0, 0.5, 5 and 50 μM of latrunculin and then exposed to a 12 min IVI insult at pH 6.8 with corresponding concentration of latrunculin. After the IVI insult the slices were returned to medium without latrunculin. White bars show mean propidium iodide fluorescence intensity (MFI) in cultures exposed to 50 μM latrunculin during the IVI insult, light gray bars are cultures exposed to 5 μM, dark gray show 0.5 μM and black are IVI controls. Data were obtained from four slices in each group. The experiments were repeated in three separate experiments. Data are expressed as mean ± SEM. Asterisks indicate a significance level of p < 0.01 using two-way ANOVA with Scheffé's post-hoc test. (B) Morphological changes in dendritic segments from a GFP-actin expressing hippocampal neuron prior to treatment and exposure to 15 min of in vitro ischemia. IVI induces large dendritic swellings with actin networks and multiple filopodia, while treatment with 5 μM latrunculin before 15 min of IVI decreases the swellings and abolish spines and filopodia. (C) The number of dendritic spines in a 10 mm long dendritic segment after 15 min of IVI.

98

BR A IN RE S E A RCH 1 3 16 ( 20 1 0 ) 9 2 – 1 00

4.

Experimental procedures

4.1.

Primary hippocampal culture preparations

All animal experiments were approved by the Malmö/Lund animal ethics committee. Primary cultures of hippocampal neurons from GFP-actin transgenic mice were prepared as described previously (Gisselsson et al., 2005). Briefly, cells were plated on poly-L-lysin coated cover glasses at a density 20 cells/mm2 and maintained in a CO2 incubator at 37 °C in Neurobasal A medium supplemented with B27, L-glutamine (0.5 mM), penicillin/streptomycin (100 U/ml), HEPES (20 mmol/l) and FGF2 (10 ng/ml). Glial proliferation was suppressed by addition of 5 μM cytosine arabinoside (Calbiochem) into the culture medium. For all experiments, cultures were grown for 7–10 days in vitro before experiments. If not otherwise stated all media and buffers were purchased from Invitrogen (Carlsbad, CA, USA). For confocal microscopy, hippocampi from fetuses (E17) of CD1 mice were dissociated and plated on chambered cover glass wells (LabTek 1.8 cm2/well) (Nunc, Roskilde, Denmark), coated with poly-D-lysine (10 mg/ml) and laminin (5 mg/ml) (Sigma-Aldrich) at a density of 2 × 105 cells/well in 0.5 ml of culture medium containing Neurobasal medium supplemented with B27, 0.5 mM L-glutamine, 25 μM glutamate and penicillin/streptomycin (100 U/ml). The day before transfection (4 DIV) half of the medium was replaced by glutamatefree culture medium. Transfection with the pEGFP-N1 vector (Clontech, Mountain View, CA, USA) was performed using lipofectamine 2000 (Invitrogen). After transfection, cells were fed with glutamate-free culture medium containing 10 mM of 1-beta-D-arabino-furanosylcytosine to suppress glial proliferation. One third of the medium volume was changed every week. Experiments were performed between day 17 and 22 in vitro.

The oxygen content in the perfusion medium was measured continuously by a Clark electrode (Consort Z921, Tumhout, Belgium). In vitro ischemia in hippocampal organotypic tissue cultures was induced as described earlier (Cronberg et al., 2004; Rytter et al., 2003). Cultures were washed in glucose free medium and transferred to an anaerobic incubator (Elektrotek ltd, Keighly, UK) containing 10% H2, 5% CO2 and 85% N2. Inside the incubator, culture were incubated in equilibrated anoxic iCSF medium (as described above) at 35 °C for 12 min.

4.4.

Treatment protocols

Latrunculin B was purchased from EMDBiosciences (Gibbstown, NJ, USA), MK801 and Y27632 from Sigma (St Louis, USA). Compounds were dissolved in DMSO. Different dilutions were used to treat cells with 0.5 μM, 5 μM or 50 μM latrunculin. MK801 and Y27632 were used in final concentration of 20 μM and 100 μM. Final concentration of DMSO in the culture medium was below 1%. For controls, only DMSO was added to the culture medium.

4.5.

Spine motility evaluation

Hippocampal organotypic tissue cultures were prepared as described previously (Rytter et al., 2003). Hippocampi were retrieved from 6-day old Balb/c mice and cut into 250 μm thick slices on a MacIllwain tissue chopper and plated individually onto Millicell culture inserts (diameter 12 mm, 0.4 μm MillicellCM, Millipore, Bedford, MA, USA). Cultures were maintained in a CO2 incubator at 35 °C for 3 weeks.

Fluorescent light microscopy was performed using a cooled digital camera (F-view) and an Olympus IX-81 fluorescence microscope. Cells were grown on cover glasses mounted in a closed bath imaging chamber (Model RC-21BR) placed on a heated platform (Model PH2). Temperature was regulated by a heater controller (Model TC-344B). The medium entering the imaging chamber was heated with a in-line solution heater (Model SH-27B). Product models above were from Warner Instrument Corp. Hamden, CT, USA. A perfusion pump, MS-Reglo, (Ismatec SA, Glattbrugg, Switzerland), set to a flow rate of 260 μl/min, perfused the medium. Time-lapse image acquisition, analysis, and image processing were performed utilizing software from Soft Imaging System SIS (AnalySIS, Olympus). To further minimize phototoxicity and bleaching we used specialized GFP filters (Chroma, Rockingham, VT, USA) and natural density filters. Spine motility alteration in mushroom and stubby spines were evaluated in consecutive micrographs captured every 20 s over a time period of 10 min (30 time points). Computer generated outlines of the spines were generated and a shape ratio was calculated (Gisselsson et al. 2005, modified from Fischer et al. 2000).

4.3.

4.6.

4.2.

Hippocampal organotypic tissue cultures

In vitro ischemia models

In vitro ischemia was performed as described earlier (Gisselsson et al., 2005). In brief, primary cell cultures grown on glass cover slips were mounted in a temperature controlled perfused microscope stage (Warner Instrument Corp, Hamden, CT, USA) and were exposed to 15 min of IVI by changing medium to an artificial ischemic cerebrospinal fluid (iCSF) (concentrations in millimolar; 0.3 CaCl2, 70 NaCl, 5.25 NaHCO3, 70 KCl, 1.25 NaH2PO4, 2 MgSO4, 40 sucrose at pH 6.8) using a perfusion pump (MS-Reglo, Ismatec SA, Glattbrugg, Switzerland). The medium was equilibrated with an anoxic gas mixture containing 90% N2, 5% CO2, 5% H2.

Cell death assessment

Quantification of cell death was evaluated by measuring mean fluorescence intensity of PI incorporation into damaged neurons (Rytter et al., 2003). Twenty-four hours before experiments, propidium iodide (PI, final concentration 10 μg/ml) was added to the culture medium. Images were taken from the slices immediately before, 24 h and 48 h after IVI. The fluorescence intensity of a defined area of the CA1 region and a background area was measured and analyzed with Image-Pro Plus 4.0 (Media Cybernetics, Maryland, USA). The mean fluorescence intensity of the CA1 regions was acquired by subtracting the background area.

BR A I N R ES E A RC H 1 3 1 6 ( 2 01 0 ) 9 2 – 1 00

4.7.

Confocal microscopy

Confocal images were captured on an inverted Zeiss LSM510 laser scanning confocal microscope using argon laser (488 nm). Cells were imaged in culture medium on a heating insert P fitted with a humidified incubation chamber (37 °C and 5% CO2) mounted on the microscope.

4.8.

Statistics

Slice culture experiments were performed in triplicate with at least 4 independent slice cultures each. Two-way ANOVA with Scheffé's post-hoc test was used to evaluate differences between groups. For statistical analysis Statview 4.0 (Abacus Concepts Inc, Berkeley, CA, USA) was used. All other statistical analyses are stated in the respective figure legends.

Acknowledgments This study was supported by the Swedish Research Council (grant 8644), the EU 7th work program through the European Stroke Network (201024), The Pia Ståhls Foundation and the Swedish Brain Fund.

REFERENCES

Allison, D.W., Gelfand, V.I., Spector, I., Craig, A.M., 1998. Role of actin in anchoring postsynaptic receptors in cultured hippocampal neurons: differential attachment of NMDA versus AMPA receptors. J. Neurosci. 18 (7), 2423–2436. Bernstein, B.W., Bamburg, J.R., 2003. Actin-ATP hydrolysis is a major energy drain for neurons. J. Neurosci. 23 (1), 1–6. Bernstein, B.W., Chen, H., Boyle, J.A., Bamburg, J.R., 2006. Formation of actin-ADF/cofilin rods transiently retards decline of mitochondrial potential and ATP in stressed neurons. Am. J. Physiol., Cell Physiol. 291 (5), C828–C839. Carlisle, H.J., Kennedy, M.B., 2005. Spine architecture and synaptic plasticity. Trends Neurosci. 28 (4), 182–187. Cronberg, T., Rytter, A., Asztely, F., Soder, A., Wieloch, T., 2004. Glucose but not lactate in combination with acidosis aggravates ischemic neuronal death in vitro. Stroke 35 (3), 753–757. Cronberg, T., Rytter, A., Wieloch, T., 2005. Chelation of intracellular calcium reduces cell death after hyperglycemic in vitro ischemia in murine hippocampal slice cultures. Brain Res. 1049 (1), 120–127. Endres, M., Engelhardt, B., Koistinaho, J., Lindvall, O., Meairs, S., Mohr, J.P., Planas, A., Rothwell, N., Schwaninger, M., Schwab, M.E., Vivien, D., Wieloch, T., Dirnagl, U., 2008. Improving outcome after stroke: overcoming the translational roadblock. Cerebrovasc. Dis. 25, 268–278. Fischer, M., Kaech, S., Knutti, D., Matus, A., 1998. Rapid actin-based plasticity in dendritic spines. Neuron 20 (5), 847–854. Fischer, M., Kaech, S., Wagner, U., Brinkhaus, H., Matus, A., 2000. Glutamate receptors regulate actin-based plasticity in dendritic spines. Nat. Neurosci. 3 (9), 887–894. Gisselsson, L.L., Matus, A., Wieloch, T., 2005. Actin redistribution underlies the sparing effect of mild hypothermia on dendritic spine morphology after in vitro ischemia. J. Cereb. Blood Flow Metab. 25 (10), 1346–1355.

99

Halpain, S., 2003. Actin in a supporting role. Nat. Neurosci. 6 (2), 101–102. Harms, C., Bosel, J., Lautenschlager, M., Harms, U., Braun, J.S., Hortnagl, H., Dirnagl, U., Kwiatkowski, D.J., Fink, K., Endres, M., 2004. Neuronal gelsolin prevents apoptosis by enhancing actin depolymerization. Mol. Cell. Neurosci. 25 (1), 69–82. Hasbani, M.J., Schlief, M.L., Fisher, D.A., Goldberg, M.P., 2001. Dendritic spines lost during glutamate receptor activation reemerge at original sites of synaptic contact. J. Neurosci. 21 (7), 2393–2403. Hebert, M., Potin, S., Sebbagh, M., Bertoglio, J., Breard, J., Hamelin, J., 2008. Rho-ROCK-dependent ezrin-radixin-moesin phosphorylation regulates Fas-mediated apoptosis in Jurkat cells. J. Immunol. 181 (9), 5963–5973. Hitomi, A., Satoh, S., Ikegaki, I., Suzuki, Y., Shibuya, M., Asano, T., 2000. Hemorheological abnormalities in experimental cerebral ischemia and effects of protein kinase inhibitor on blood fluidity. Life Sci. 67 (16), 1929–1939. Kitaoka, Y., Kitaoka, Y., Kumai, T., Lam, T.T., Kuribayashi, K., Isenoumi, K., Munemasa, Y., Motoki, M., Kobayashi, S., Ueno, S., 2004. Involvement of RhoA and possible neuroprotective effect of fasudil, a Rho kinase inhibitor, in NMDA-induced neurotoxicity in the rat retina. Brain Res. 1018 (1), 111–118. Ladwein, M., Rottner, K., 2008. On the Rho'd: the regulation of membrane protrusions by Rho-GTPases. FEBS Lett. 582 (14), 2066–2074. Laufs, U., Endres, M., Stagliano, N., Amin-Hanjani, S., Chui, D.S., Yang, S.X., Simoncini, T., Yamada, M., Rabkin, E., Allen, P.G., et al., 2000. Neuroprotection mediated by changes in the endothelial actin cytoskeleton. J. Clin. Invest. 106 (1), 15–24. Lei, S., Czerwinska, E., Czerwinski, W., Walsh, M.P., MacDonald, J.F., 2001. Regulation of NMDA receptor activity by F-actin and myosin light chain kinase. J. Neurosci. 21 (21), 8464–8472. Li, Z., Van, A.L., Cline, H.T., 2000. Rho GTPases regulate distinct aspects of dendritic arbor growth in Xenopus central neurons in vivo. Nat. Neurosci. 3, 217–225. Luo, L., 2002. Actin cytoskeleton regulation in neuronal morphogenesis and structural plasticity. Annu. Rev. Cell Dev. Biol. 18, 601–635. Maekawa, M., Ishizaki, T., Boku, S., Watanabe, N., Fujita, A., Iwamatsu, A., Obinata, T., Ohashi, K., Mizuno, K., Narumiya, S., 1999. Signaling from Rho to the actin cytoskeleton through protein kinases ROCK and LIM-kinase. Science 285 (5429), 895–898. Meng, Y., Zhang, Y., Tregoubov, V., Janus, C., Cruz, L., Jackson, M., Lu, W.Y., MacDonald, J.F., Wang, J.Y., Falls, D.L., et al., 2002. Abnormal spine morphology and enhanced LTP in LIMK-1 knockout mice. Neuron 35 (1), 121–133. Murphy, T.H., Li, P., Betts, K., Liu, R., 2008. Two-photon imaging of stroke onset in vivo reveals that NMDA-receptor independent ischemic depolarization is the major cause of rapid reversible damage to dendrites and spines. J. Neurosci. 28 (7), 1756–1772. Nakayama, A.Y., Harms, M.B., Luo, L., 2000. Small GTPases Rac and Rho in the maintenance of dendritic spines and branches in hippocampal pyramidal neurons. J. Neurosci. 20, 5329–5338. Nicholson-Dykstra, S., Higgs, H.N., Harris, E.S., 2005. Actin dynamics: growth from dendritic branches. Curr. Biol. 15 (9), R346–R357. Pellegrin, S., Mellor, H., 2005. The Rho family GTPase Rif induces filopodia through mDia2. Curr. Biol. 15 (2), 129–133. Pollard, T.D., Borisy, G.G., 2003. Cellular motility driven by assembly and disassembly of actin filaments. Cell 112 (4), 453–465. Rosenmund, C., Westbrook, G.L., 1993. Calcium-induced actin depolymerization reduces NMDA channel activity. Neuron 10 (5), 805–814.

100

BR A IN RE S E A RCH 1 3 16 ( 20 1 0 ) 9 2 – 1 00

Rytter, A., Cronberg, T., Asztely, F., Nemali, S., Wieloch, T., 2003. Mouse hippocampal organotypic tissue cultures exposed to in vitro “ischemia” show selective and delayed CA1 damage that is aggravated by glucose. J. Cereb. Blood Flow Metab. 23 (1), 23–33. Rytter, A., Cardoso, C.M., Johansson, P., Cronberg, T., Hansson, M.J., Mattiasson, G., Elmer, E., Wieloch, T., 2005. The temperature dependence and involvement of mitochondria permeability transition and caspase activation in damage to organotypic hippocampal slices following in vitro ischemia. J. Neurochem. 95 (4), 1108–1117. Satoh, S., Ikegaki, I., Suzuki, Y., Asano, T., Shibuya, M., Hidaka, H., 1996. Neuroprotective properties of a protein kinase inhibitor against ischaemia-induced neuronal damage in rats and gerbils. Br. J. Pharmacol. 118 (7), 1592–1596. Satoh, S., Utsunomiya, T., Tsurui, K., Kobayashi, T., Ikegaki, I., Sasaki, Y., Asano, T., 2001. Pharmacological profile of hydroxy fasudil as a selective rho kinase inhibitor on ischemic brain damage. Life Sci. 69 (12), 1441–1453.

Schulz, T.W., Nakagawa, T., Licznerski, P., Pawlak, V., Kolleker, A., Rozov, A., Kim, J., Dittgen, T., Kohr, G., Sheng, M., Seeburg, P.H., Osten, P., 2004. Actin/Alpha-actinin-dependent transport of AMPA receptors in dendritic spines: role of the PDZ-LIM protein RIL. J. Neurosci. 24, 8584–8594. Tada, T., Sheng, M., 2006. Molecular mechanisms of dendritic spine morphogenesis. Curr. Opin. Neurobiol. 16 (1), 95–101. Tashiro, A., Yuste, R., 2004. Regulation of dendritic spine motility and stability by Rac1 and Rho kinase: evidence for two forms of spine motility. Mol. Cell. Neurosci. 26 (3), 429–440. Yamashita, K., Kotani, Y., Nakajima, Y., Shimazawa, M., Yoshimura, S., Nakashima, S., Iwama, T., Hara, H., 2007. Fasudil, a Rho kinase (ROCK) inhibitor, protects against ischemic neuronal damage in vitro and in vivo by acting directly on neurons. Brain Res. 1154, 215–224.