Journal of the Neurological Sciences 337 (2014) 25–37

Contents lists available at ScienceDirect

Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns

Temporal and regional patterns of Smad activation in the rat hippocampus following global ischemia Takayuki Nakajima ⁎, Masafumi Yanagihara, Hideki Nishii Department of Veterinary Anatomy, Graduate School of Life and Environmental Sciences, Osaka Prefecture University, 1-58 Rinku-Ohraikita, Izumisano, Osaka 598-8531, Japan

a r t i c l e

i n f o

Article history: Received 7 August 2013 Received in revised form 25 October 2013 Accepted 11 November 2013 Available online 19 November 2013 Keywords: Global ischemia Smad Hippocampus Rat Delayed neuronal death Microglia

a b s t r a c t In this study, we examined the temporal and regional patterns of Smad activation in the rat hippocampus following global ischemia. We also examined the association between Smad activation and ischemia-induced pathology in the hippocampus. We found that 1) Smad1, -2, -3, and -5 proteins were detected in the rat hippocampus by means of western blot and immunohistochemistry; 2) after 5 min of ischemia, Smad2 and Smad3 proteins accumulated in the nuclei of pyramidal cells in the CA1 region, which is vulnerable to ischemia; 3) after 3 min of ischemia, which was nonlethal, there was no such nuclear accumulation of Smad2 and Smad3 in the CA1 region; 4) following injection of activin A, nuclear accumulation of Smad2 and Smad3 was induced not only in pyramidal cells of the CA1 region, but also in pyramidal cells of the CA3 region as well as in granule cells of the DG region; 5) activin Ainduced nuclear accumulation of Smad2 and Smad3 neither caused degeneration of hippocampal neurons nor prevented degeneration induced by ischemia. These results suggest that in the hippocampus, ischemiainduced activation of Smad2 and Smad3 is associated with the response to stress but is not related to neuronal survival or death. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Cardiac arrest causes global brain ischemia, often leading to cognitive or memory disturbances. The hippocampus plays an important role in cognition and memory formation, and is well known to be one of the brain regions most vulnerable to ischemia [1]. The pyramidal neurons of the CA1 region of the hippocampus are particularly vulnerable. In contrast, pyramidal neurons in the CA3 area of the hippocampus and granule cells of the dentate gyrus (DG) are both relatively resistant to ischemia. Previous studies using animal models of global ischemia have shown that pyramidal neurons in the CA1 region degenerate after 2 to 4 days of reperfusion [2,3]; this phenomenon is known as delayed neuronal death [2]. Delayed neuronal cell death has been the focus of a great deal of research in the hope of developing preventative interventions for patients with ischemia. Previous studies have demonstrated that the activation of a number of molecules is altered in neurons prior to the onset of ischemia-induced neuronal cell death. These molecules include calpain, a Ca2+-dependent protease [4–6], MAPK/ERK kinase 1, a tyrosine/threonine kinase [7], Akt, a serine/threonine kinase [8,9], and CREB, a transcriptional factor [6,10,11]. Such findings suggest that ischemia disturbs the homeostasis in affected neurons. However, the precise mechanisms that mediate neuronal cell death have yet to be elucidated.

⁎ Corresponding author. Tel.: +81 72 463 5594. E-mail address: [email protected] (T. Nakajima). 0022-510X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jns.2013.11.012

It is well known that the transforming growth factor-β (TGF-β) superfamily of proteins, including TGF-βs, activins, and bone morphogenetic proteins (BMPs), regulates multiple cellular functions in both fetal and adult tissues. TGF-βs inhibit the proliferation of epithelial cells and lymphocytes [12], activins regulate the release of folliclestimulating hormone and the differentiation of erythrocytes [13–15], and BMPs contribute to the differentiation of osteoblasts [16]. Smad proteins are known to transduce the actions of the TGF-β family [17,18]. To date, 8 Smad proteins have been characterized: Smad1, -2, -3, -4, -5, -6, -7, and -8. These Smad proteins are divided into 3 subclasses on the basis of their function and structure: receptor-regulated Smads (R-Smads), the common partner Smad (Co-Smad), and inhibitory Smads (I-Smads). When TGF-βs, activins, and BMPs bind to their specific receptors on the cell surface, the receptors activate R-Smads via carboxy-terminal phosphorylation. R-Smads include Smad1, -2, -3, -5, and -8. Receptors for TGF-βs, activins and BMPs consist of 2 types of transmembrane serine/threonine kinase receptors; type I and type II. Seven type I receptors (activin receptor-like kinases (ALK) 1-7) and 5 type II receptors have been characterized. When type II receptors bind to their ligands, they activate type I receptors, which in turn phosphorylate R-Smads. ALK4, -5, and -7 form complexes with type II receptors that have affinity for TGF-βs and activins, and phosphorylate Smad2 and Smad3, while ALK1, -2, -3, and -6 form complexes with type II receptors with affinity for BMPs, and phosphorylate Smad1, -5, and -8. The only known mammalian Co-Smad is Smad4, which forms a complex with phosphorylated R-Smad before translocating to the nucleus. I-Smads include Smad6 and Smad7, and inhibit the activation

26

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

of R-Smads by inducing degradation of the receptors or by competing with the R-Smads for type I receptor binding. TGF-βs include 3 mammalian isoforms: TGF-β1, -β2, and -β3 [17]. In the adult rat hippocampus, the distributions of TGF-β1, -β2, and -β3 proteins and their mRNAs have been characterized using immunohistochemistry and in situ hybridization. TGF-β1, -β2, and -β3 proteins and their mRNAs were detected in pyramidal cells in the CA1 and CA3 regions, as well as in granule cells in the DG region [19–21]. Following ischemia, the immunoreactivities of TGF-β1, -β2, and -β3 in the hippocampus were found to be altered. The immunoreactivities were decreased in CA1 pyramidal neurons, although they did not change in the CA3 pyramidal neusons or in DG granule cells [19]. In contrast, the immunoreactivities of TGF-β1, -β2, and -β3 in astrocytes were increased after ischemia [19]. There are 3 types of activin isoforms known, including activin A, activin B, and activin AB [18]. Activin A and activin B are homodimers of inhibin βA and inhibin βB subunits, respectively. Activin AB is a heterodimer formed by inhibin βA and inhibin βB subunits. In situ hybridization revealed that inhibin βA subunit mRNA was expressed in the adult CA1, CA3, and DG regions, and that inhibin βB subunit mRNA was expressed in the CA1 and CA3 regions [22]. In addition, studies using a global ischemia model in adult rats and a hypoxic–ischemic brain injury model in infant rats demonstrated that treatment with either TGF-β1 or activin A decreased neuronal death in the CA1 region after ischemia [23,24]. These findings suggest that TGF-β1 and activin A may act as endogenous neuroprotective factors following ischemia. However, it has also been reported that TGF-β and activin induce cell death in an oligodendroglial cell line [25,26]. Furthermore, very little is known regarding the association between Smad activation and the pathology of the ischemic hippocampus. In this study, we examined the temporal and regional patterns of Smad activation in the rat hippocampus following global ischemia, and investigated the association between Smad activation and ischemia-induced pathology in the hippocampus. 2. Materials and methods 2.1. Animal surgery and experimental groups Male Sprague–Dawley rats, each weighing 250–350 g, were subjected to global ischemia or to intracerebroventricular injection of activin A. Rats were divided into 8 groups: (a) 5-min and (b) 3-min ischemia group, rats subjected to 5 min and 3 min of ischemia, respectively; (c) sham group, rats subjected to the same operation without ischemia; (d) intracerebroventricular injection of activin A group, rats subjected to injection of activin A into the bilateral cerebral ventricles; (e) intracerebroventricular injection of vehicle group, rats subjected to injection of vehicle control into the bilateral cerebral ventricles; (f) intracerebroventricular injection of activin A + 5-min ischemia group, rats subjected to injection of activin A into the bilateral cerebral ventricles followed by 5 min of ischemia; (g) intracerebroventricular injection of vehicle + 5-min ischemia group, rats subjected to injection of vehicle into bilateral cerebral ventricles followed by 5 min of ischemia; and (h) intracerebroventricular injection of vehicle + sham group, rats subjected to injection of vehicle into bilateral cerebral ventricles followed by sham operation. The experimental design for each group is summarized in Fig. 1. The number of animals used is shown in Table 1. 2.2. Global ischemia Global ischemia was induced using the four-vessel occlusion method. The surgical procedure and induction of ischemia were performed as has been described previously [5]. Briefly, the bilateral vertebral arteries were permanently occluded by electrocauterization under chloral hydrate anesthesia (i.p., 400 mg/kg). After allowing a 24 h recovery, rats were anesthetized with 1.5% halothane in 30% oxygen and 70% nitrous oxide, and ischemia was induced by occluding the bilateral

common carotid arteries with aneurysm clips. Sham-operated animals were treated similarly to those subjected to ischemia, except for the occlusion of the common carotid arteries. Body temperature was maintained at 37.0 ± 0.5 °C with a rectal thermistor and heat lamp until rats fully recovered from the anesthesia. Variability of results was minimized by excluding rats that failed to show complete loss of the righting reflex and bilateral pupil dilation during ischemia. Rats that stopped breathing during ischemia were also excluded. All procedures were approved by the Animal Experiment Committee of the Osaka Prefecture University. All surgeries and subsequent experiments including histology, immunohistochemistry, western blot, and RT-PCR were performed by operators blinded to the treatment group. 2.3. Intracerebroventricular injection One microgram of recombinant human activin A (Peprotech, Rocky Hill, NJ) was dissolved in 5 μl of vehicle (0.1% bovine serum albumin diluted in 0.01 M PBS (pH 7.2)). Rats were anesthetized with 1.5% halothane in 30% oxygen and 70% nitrous oxide, and placed into a stereotactic apparatus (Muromachi, Tokyo, Japan). Recombinant human activin A was delivered into the bilateral cerebral ventricle using a microinjector (Muromachi, Tokyo, Japan) (0.5 mm posterior to bregma, 1.5 mm lateral to the midline, 5 mm ventral from the cranium). A total volume of 5 μl was administered at a rate of 1 μl/min. An equal volume of vehicle infusion served as the vehicle control.

2.4. Histological assessment Animals were deeply anesthetized with chloral hydrate and perfused transcardially with physiological saline containing 10 U/ml heparin sulfate, followed by 4% paraformaldehyde-0.1 M phosphate buffer (PB) (pH 7.4). Brains were quickly removed from the skull and postfixed for 20 h at 4 °C, then dehydrated and embedded in paraffin. Coronal sections (3.5 μm thick) containing areas of the hippocampus were cut between –3.3 mm and –3.8 mm from bregma and stained with hematoxylin/eosin. The number of neurons considered to have survived ischemia in the bilateral CA1 regions was counted. Counting of neurons was performed as described previously [5]. Briefly, neurons were considered to have survived if they had an intact round or oval nucleus. Four sections from each animal were analyzed, with the sections separated from one another by at least 20 μm. 2.5. RT-PCR Brains were removed, rinsed in ice-cold physiological saline, and sliced at 2-mm intervals using Brain Matrices (BrainScience-Idea, Osaka, Japan). The brain sections were placed onto chilled plates, and the hippocampus was removed from the slices under a dissection microscope for mRNA expression analysis of Smad1, -2, -3, -5, and -8. For the analysis of inhibin βA, -βB, TGF-β1, -β2, -β3, Smad6, and Smad7 mRNA expression, the sliced hippocampi were further divided into 2 regions; the CA1 region and CA3/DG regions. These tissues were then snap-frozen in liquid nitrogen, and stored at –80 °C until further processing. Total RNA was isolated from the frozen tissue samples using TRIZOL reagent (Invitrogen Japan K.K., Tokyo) according to the manufacturer's instructions. Reverse transcription of 1 μg of the total RNA was performed using M-MLV reverse transcriptase (Promega, Madison, WI) and an oligo (dT)18 primer at 42 °C for 50 min. cDNA was amplified using HybriPol™ DNA Polymerase (Bioline Ltd. London, UK), with the annealing temperatures, and PCR cycles dependent on the target genes. Primers, annealing temperatures and PCR cycles used are listed in Table 2. The identities of the RT-PCR products were confirmed by direct sequencing of PCR products using an automated ABI 3130 Genetic Analyzer (Life Technologies, Tokyo, Japan).

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

: Ischemia or sham-operation

27

: Recovery

: Reperfusion

(1) 5 min of ischemia or sham group 7 days 4 days 3 days 2 days 1 days 10 h 3h 5 min

Fixation for histology

Fixation for histology

Tissue homogenization for WB Fixation for histology Tissue homogenization for WB Fixation for IHC or tissue homogenization for WB or RT-PCR Fixation for IHC or tissue homogenization for WB or RT-PCR

(2) 3 min of ischemia or sham group 10 h 3 min

Fixation for IHC or tissue homogenization for WB

(3) Intracerebroventricular injection (I.C.V.I.) of activin A or vehicle group 6h 7 days

4h

Fixation for histology

Fixation for IHC Fixation for IHC I.C.V.I. of activin A or vehicle

(4) Intracerebroventricular injection (I.C.V.I.) of activin A or vehicle + 5 min of ischemia or sham group 4h

5 min

7 days

I.C.V.I. of activin A or vehicle

Fixation for histology

Fig. 1. Experimental protocols for histological assessment, western blot (WB), RT-PCR, and immunohistochemistry (IHC).

2.6. Preparation of subcellular fractions After slicing the brain at 2-mm intervals using Brain Matrices, the sections were placed onto chilled plates, and the hippocampal CA1 and CA3/DG regions were removed from the slices under a microscope, snap-frozen in liquid nitrogen, and stored at −80 °C. The frozen tissue was minced using ophthalmic scissors in 1:10 (w/v) ice-cold buffer A [10 mM Tris–HCl (pH 7.6), 0.25 M sucrose, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM ethylenediamine tetraacetic acid (EDTA), 1 mM ethylene glycol tetraacetic acid (EGTA), 50 mM sodium fluoride (NaF), 0.5 mM

dithiothreitol (DTT), 10% glycerol, 2 mM sodium pyrophosphate, 1 mM sodium orthovanadate (NaVO4), 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 1.25 mg/ml pepstatin A, and 10 mg/ml leupeptin, 2.5 mg/ml aprotinin]. The minced sample tissues were rotated at 4 °C for 20 min, and then centrifuged at 18,000 × g for 20 min at 4 °C. After centrifugation, the supernatant sample (S1) was collected. The residual pellet was re-suspended in Buffer B [20 mM Tris–HCl (pH 7.6), 0.25 M sucrose, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM EDTA, 1 mM EGTA, 50 mM NaF, 0.5 mM DTT, 10% glycerol, 2 mM sodium pyrophosphate, 1 mM NaVO4, 0.5 mM PMSF, 1.25 mg/ml pepstatin A, 10 mg/ml

28

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

Table 1 The number of animals used in the present study. Groups

Histological assessment 2 days§

Naïve Sham 5 min of ischemia 3 min of ischemia I.C.V.I. of activin A I.C.V.I. of vehicle I.C.V.I. of vehicle + sham I.C.V.I. of vehicle + 5 min of ischemia I.C.V.I. of activin A + 5 min of ischemia

Western blot

4 days§

7 days§

4

4 5 (1)

3

RT-PCR

3 h§

10 h§

1 day§

3 day§

4 5 (1)

7 4 3

4 4

4 4

5 4

Immunohistochemistry

3 h§

10 h§

3 3

3 4 (1)

3 h§

4

4 h§

6 h§

10 h§

3 2 3

3 3 4 5 (1) 4

3 4 3 2 2

2 2

Sham: sham-operation. §: Time of recovery after ischemia or after injection of activin A or vehicle. I.C.V.I.: intracerebroventricular injection. (): the number of rats excluded because of experiencing breathing arrest or failing to show complete losses of righting reflex and bilateral pupil dilation during ischemia.

leupeptin, 2.5 mg/ml aprotinin], rotated for 5 min at 4 °C, and centrifuged at 18,000 ×g for 20 min at 4 °C. After discarding the supernatant, the residual pellet was re-suspended in Buffer C [20 mM Tris–HCl (pH 7.6), 0.42 M NaCl, 0.25 M sucrose, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM EDTA, 50 mM NaF, 0.5 mM DTT, 10% glycerol, 2 mM sodium pyrophosphate, 1 mM NaVO4, 0.5 mM PMSF, 1.25 mg/ml pepstatin A, 10 mg/ml leupeptin, 2.5 mg/ml aprotinin], rotated for 20 min at 4 °C, and centrifuged at 18,000 × g for 20 min at 4 °C. The resultant supernatant sample (S2) was collected, and the residual pellet was resuspended in Buffer B, followed by rotation for 5 min at 4 °C and centrifugation at 18,000 ×g for 20 min at 4 °C. After discarding the supernatant, the residual pellet was re-suspended in Buffer D [20 mM Tris–HCl (pH 6.8), 0.15 M NaCl, 0.25 M sucrose, 1.5 mM MgCl2, 0.5 mM EDTA, 1 mM EGTA, 50 mM NaF, 0.5 mM DTT, 10% glycerol, 1% Nonidet P-40, 0.1% sodium dodecyl sulfate (SDS), 2 mM sodium pyrophosphate, 1 mM NaVO4, 0.5 mM PMSF, 1.25 mg/ml pepstatin A, 10 mg/ml leupeptin, 2.5 mg/ml aprotinin], followed by rotation for 20 min at 4 °C and centrifugation at 18,000 ×g for 20 min at 4 °C. After centrifugation, the supernatant sample (S3) was collected. TGF-β-treated HeLa cells were used as a positive control for the detection of Smads. After collection

of TGF-β treated HeLa cells, subcellular fractions were prepared using the same methods described above. The protein concentration of the samples was determined using the Bradford method with a bovine serum albumin standard. 2.7. Western blot Equal amounts of protein (30 μg) were loaded on a 10% sodium dodecyl sulfate-polyacrylamide gel for electrophoresis (SDS-PAGE). After separation by electrophoresis, proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Bio-Rad, Hercules, CA). Blotting membranes were blocked with 5% non-fat milk in TBST [10 mM Tris–HCl (pH 7.4), 0.15 M NaCl, and 0.05% Tween 20] at room temperature for 1 h, and then incubated overnight at 4 °C with a primary antibody solution containing anti-Smad1 (1:4000) (M03; Abnova, Taiwan), anti-Smad2 (1:2500) (Ab-220; Signalway Antibody, Pearland, TX), anti-Smad3 (1:4000) (EP568Y; Epitomics, Burlingame, CA), antiSmad5 (1:3000) (EP619Y; Epitomics), anti-Smad8 (1:500) (R-64; Santa Cruz Biotechnology, Santa Cruz, CA), or anti-β-actin (1:7000) (A5441; Sigma-Aldrich, St. Louis, MO). Membranes were washed 3

Table 2 List of primers used in the present study. Primer

Direction

Sequence

Annealing temperature (°C)

PCR cycles

Product (bp)

Smad1

Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse

5′-AACCATGGGTTCGAGACAGT-3′ 5′-ACCGACGAAATAGGGTTGTG-3′ 5′-AGGGGAGTGCGCTTGTATTA-3′ 5′-TGATAAACGGCCTCAAAACC-3′ 5′-TGACAGTGCTATTTTCGTCCAGTCT-3′ 5′-CGATCCCTTTACTCCCAGTGTCT-3′ 5′-TGTACTATGAACTGAACAACCGTGT-3′ 5′-CCAACGTAGTATAGATGGACACCTT-3′ 5′-GATCTGTCCGATTCTACATTGTCTT-3′ 5′-AGGTAGGTCGTAGAAGATGCTGAC-3′ 5′-TCAGATTCCCAACTTCTTCTGGAGCC-3′ 5′-TGTGAAGATGACCTCCAGCCAGCAC-3′ 5′-TGAAGTTGTCTATGAACTGACGAAG-3′ 5′-AAGACACTGAAGAAATAGGGTTGTG-3′ 5′-AGGGCTGGAAGAGGAAAAAG-3′ 5′-ATGATAGCCAGAGGGAGCAA-3′ 5′-TCTCTAACGAAGGCAACCAGA-3′ 5′-TTCTCCACCACATTCCACCT-3′ 5′-TGAGTGGCTGTCTTTTGACG-3′ 5′-TTCTCTGTGGAGCTGAAGCA-3′ 5′-CACGCCTCTCTTGTTTCCTC-3′ 5′-CACATGTTTTCTGGGGCTTT-3′ 5′-GGGAGTTGCTGGAAGAGATG-3′ 5′-GGACACATTGAAACGGAAAA-3′ 5′-GGCATCCTGACCCTGAAGTA-3′ 5′-TCTCAGCTGTGGTGGTGAAG-3′

57

32

203

57

33

221

57

34

375

53

34

196

60

35

210

60

33

203

57

32

197

57

34

201

55

35

162

60

32

293

57

27

296

57

42

181

58

27

449

Smad2 Smad3 Smad5 Smad6 Smad7 Smad8 Inhibin βA Inhibin βB TGF-β1 TGF-β2 TGF-β3 β-Actin

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

29

A

A Naïve

5 min ischemia

1000

CA1

Sham

500 100

CA3

50μ m

C Sham

50μ m 2 days

4 days

7 days

Number of neurons in CA1/mm

B

* * 140 120 100 80 60 40 20 0

Smad1 Smad2 Smad3 Smad5 β-actin S1 S2

*

S3

Naïve

S1 S2

S3 S1 S2

10 h

Sham

S3

S1 S2

S3

3 day

Reperfusion after 5 min of ischemia

CA1

Time of reperfusion after 5 min of ischemia

TGFβ-treated Hela cells

DG

B

C Calpain-II CREB

Fig. 2. Histological changes in the hippocampus after 5 min of ischemia. (A) Photographs of hematoxylin/eosin-stained sections from naïve rats, sham-operated rats, and rats subjected to 5 min of ischemia followed by 7 days of reperfusion. (B) Photographs of hematoxylin/eosin-stained sections from sham-operated rats and rats subjected to 5 min of ischemia followed by 2, 4 and 7 days of reperfusion. (C) Changes in the number of pyramidal cells per 1-mm length were tested using a two-tailed unpaired Student's ttest. Data are expressed as the mean ± standard deviation (SD). *Statistically significant difference from rats subjected to sham-operation (P b 0.05); n = 4. DG: dentate gyrus. Sham: sham-operation.

times with TBST for 10 min each wash, and then incubated for 1 h at room temperature with horseradish peroxidase (HRP)-conjugated secondary antibody (NA934; GE Healthcare, Buckinghamshire, UK). After washing with TBST, membranes were processed with the Phototope HRP western blot detection system (#7003; Cell Signaling Technology, Beverly, MA). Image J software was used to scan and quantify the density of the detected bands. Western blot analysis was repeated twice for the experimental group, and the mean values were used in the statistical analysis.

2.8. Immunohistochemical staining While deeply anesthetized with sodium pentobarbital, animals were perfused transcardially with physiological saline containing 10 U/ml heparin sulfate, followed by 4% paraformaldehyde—0.1 M PB (pH 7.4). Brains were quickly removed from the skull, and post-fixed for 20 h at 4 °C. For cryoprotection, the brains were then immersed overnight at 4 °C in 30% sucrose—0.1 M PB (pH 7.4). Coronal brain sections were cut at a thickness of 7 μm on a cryostat. Immunohistochemical staining with an avidin–biotin peroxidase complex (ABC) was performed using the Vectastain Elite ABC kit (Vector laboratories, Burlingame, CA). Briefly, sections were incubated in 0.3% H2O2–methanol for 20 min at room temperature. They were then incubated in 3% normal goat serum for 1 h at 32 °C and overnight at 4 °C in a primary antibody solution containing anti-Smad1 (1:500) (M03; Abnova), anti-Smad2 (1:2500) (Ab-220; Signalway Antibody), anti-Smad3 (1:4000) (EP568Y; Epitomics), anti-Smad5 (1:3000)

NR2A S1 S2

CA1

S3 S1

S2

S3

CA3/DG

Fig. 3. Expression of R-Smad mRNA and localization of R-Smad proteins in the rat hippocampus. (A) RT-PCR amplification of R-Smads containing Smad1, -2, -3, -5, and -8 gene transcripts in the hippocampus. (B) Western blot analysis of R-Smads containing Smad1, -2, -3, -5, and -8 in the hippocampal CA1 region. TGF-β-treated HeLa cells were used as a positive control. (C) Assessment of prepared subcellular fractions in the hippocampal CA1 region and CA3/DG regions by western blot analysis. Sham: sham-operation. NR2A: NMDA receptor 2A.

(EP619Y; Epitomics) or anti-Smad8 (1:500) (R-64; Santa Cruz Biotechnology, Santa Cruz, CA). Sections were next incubated with biotinylated anti-mouse or anti-rabbit goat IgG (1:600) (Vector laboratories) for 1 h at 32 °C. Color development was performed using ImmPACT™ DAB peroxidase substrate (Vector laboratories). Double-labeled immunofluorescent staining was performed in order to evaluate the type of cells in which Smad2 and Smad3 were localized. Sections were incubated in 3% normal goat serum for 1 h at 32 °C and overnight at 4 °C with primary antibody containing antiSmad2 (1:200) (Ab-220; Signalway Antibody), anti-Smad3 (1:400) (EP568Y; Epitomics), anti-βIII tubulin (1:200) (MAB1637; Millipore, Temecula, CA), anti-glial fibrillary acidic protein (GFAP) (1:2000) (G6171; Sigma-Aldrich, St. Louis, MO), or anti-2′, 3′-cyclic nucleotide 3′-phosphodiesterase (CNPase) (1:400) (MAB326; Millipore). Sections were then incubated with DyLigh 549-labeled anti-mouse or DyLigh 488-labeled anti-rabbit goat IgG (1:600) (Jackson ImmunoResearch Laboratory Inc., West Grove, PA) for 1 h at 32 °C.

2.9. Statistical analysis Microsoft Excel 2001 was used for all statistical analysis. Differences between sham-operated and ischemia groups were analyzed using unpaired Student's t-tests. Differences among the groups were analyzed using one-way ANOVA followed by Tukey's multiple comparison test. The threshold for statistical significance was set at P b 0.05.

30

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

βIII-tubulin

Smad2

A

Smad2/βIII-tubulin

Naïve

SO SP SR

Sham

SO SP

5 min ischemia

SR

B

SO SP 50 μ m

SR

βIII-tubulin

Smad3

Smad3/βIII-tubulin

Naïve

SO SP SR

Sham

SO SP

5 min ischemia

SR

SO SP 50 μ m

SR

Fig. 4. Double-labeled immunofluorescent staining for Smad2 (green)/βIII-tubulin (red) (A) and Smad3 (green)/βIII-tubulin (red) (B) in the hippocampal CA1 region of naïve rats, shamoperated rats, and rats subjected to 5 min of ischemia followed by 10 h of reperfusion. After 5 min of ischemia, immunoreactivity for Smad2 and Smad3 was observed in the nuclei of pyramidal cells. Sham: sham-operation. SO: stratum oriens, SP: stratum pyramidale, SR: stratum radiatum.

3. Results 3.1. Ischemia-induced neuronal cell death in the CA1 region Transient global ischemia for 5 min selectively induced neuronal cell death in the CA1 region (Fig. 2A). A portion of the pyramidal cells began to die 2 days after ischemia, and most were dead within 7 days of ischemia (Fig. 2B). 3.2. Expression of Smad1, -2, -3, -5, and -8 mRNA in the hippocampus To examine the expression levels of Smad1, -2, -3, -5, and -8 mRNAs in the rat hippocampus, we performed RT-PCR. The PCR products of

Smad1, -2, -3, and -5 mRNA were reliably detected at the expected size, but Smad8 mRNA was detected at a much lower level than Smad1, -2, -3, and -5 (Fig. 3A). 3.3. Ischemia-induced changes in the localization of Smad1, -2, -3, and -5 in the hippocampus Using western blot analysis, ischemia-induced changes in the expression patterns of Smad1, -2, -3, -5, and -8 proteins were examined in the samples divided into the S1, S2, and S3 fractions derived from the CA1 and the CA3/DG regions. The bands for Smad1, -2, -3, and -5 proteins were detected predominantly in the S1 and S2 fractions prepared from the CA1 (Fig. 3B) and CA3/DG (data not shown) regions of

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

31

Time of reperfusion after 5 min of ischemia Sham

Naïve

10 h

3h

Smad2

SO SP SR

Smad3

SO SP SR

Cresylviolet

SO SP 50 μm

SR

Fig. 5. Photographs of hippocampal CA1 sections stained with the avidin–biotin peroxidase complex (ABC) method or with cresyl violet, taken from naïve rats, sham-operated rats, and rats subjected to 5 min of ischemia followed by 3 h and 10 h of reperfusion. Upper and middle panels show Smad2 and Smad3 immunoreactivity, respectively. Sham: sham-operation. SO: stratum oriens, SP: stratum pyramidale, SR: stratum radiatum.

naïve, sham-operated rats, as well as in rats subjected to 5 min of ischemia. In the S2 fraction from the CA1 region of rats subjected to 5 min of ischemia, the bands for Smad2 and Smad3 proteins were of a higher density (Fig. 3B). However, in the S2 fraction from the CA3/DG regions of rats subjected to 5 min of ischemia, these bands did not change (data not shown). There was enriched expression of Calpain-II, CREB and NMDA receptor 2A in the S1, S2, and S3 fractions, respectively (Fig. 3C). Calpain-II is a Ca2+-dependent protease, and is predominantly localized to the cytosol. CREB is a transcriptional factor, and is predominantly localized to the nucleus. NMDA receptor 2A is a glutamate receptor, and is predominantly localized to the neuronal cell membrane. Therefore, the increase in band density for Smad2 and Smad3 proteins in the S2 fraction implies the nuclear accumulation of Smad2 and

Smad3 in the CA1 region following ischemia. We were unable to detect the band for Smad8 protein under our experimental conditions. In order to examine the ischemia-induced morphological changes in the localization of Smad2 and Smad3 within the hippocampus, we performed immunohistochemistry. The localization of Smad2 and Smad3 immunoreactivity was altered predominantly in the CA1 region (Figs. 4A and B, and 5). Smad2- and Smad3-immunopositive cells in the CA1 region were primarily pyramidal cells, as determined by the neuronal location and morphology revealed by immunofluorescent staining for βIII-tubulin, a neuronal marker (Fig. 4A and B). In naïve and sham-operated rats, immunoreactivity for Smad2 and Smad3 was detected in the cytoplasm of pyramidal cells. In rats subjected to 5 min of ischemia followed by 10 h of reperfusion, immunoreactivity

A Naïve

Sham

5 min ischemia

40 μm

B

Iba-1

Smad2/Iba-1

5 min ischemia

Naïve

Smad2

50 μm

Fig. 6. Photographs of Smad2-immunopositive microglia in the hippocampus from naïve rats, sham-operated rats, and rats subjected to 5 min of ischemia followed by 10 h of reperfusion. (A) Immunohistochemical staining using the avidin–biotin peroxidase complex (ABC) method. Immunoreactivity for Smad2 was detected in small-sized cells throughout the hippocampus (arrows). (B) Double-labeled immunofluorescent staining with anti-Smad2 and anti-Iba-1 antibodies. White arrows indicate representative cells showing immunoreactivity for Smad2 and Iba-1. Sham: sham-operation.

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

3.4. RT-PCR analysis of ischemia-induced changes in the mRNA expression of inhibins and TGF-βs in the hippocampus Smad2 and Smad3 are downstream molecules of activin and TGF-β signaling [17,18]. To examine the expression levels of inhibin βA, inhibin βB, TGF-β1, TGF-β2, and TGF-β3 mRNA in the hippocampus after 5 min of ischemia, we performed RT-PCR. In samples from both the CA1 and the CA3/DG regions, RT-PCR products of inhibin βA, TGF-β1, and TGF-β2 mRNA were detected (Fig. 8). Among these 3 genes, only inhibin βA mRNA was significantly up-regulated after ischemia in both the CA1 and the CA3/DG regions, suggesting the induction of activin A in these 2 regions. 3.5. Activin-induced changes in the localization of Smad2 and Smad3 in the hippocampus

A Smad2 Smad3 β-actin

Reperfusion after 5 min of ischemia

B Relative band density of Smad2/β-actin (% of naïve)

for Smad2 and Smad3 was detected predominantly in the nucleus of pyramidal cells. As revealed by cresyl violet-stained sections, there was no significant change in the morphology of pyramidal cells at 10 h after 5 min of ischemia (Fig. 5). In pyramidal cells of the CA3 region and granule cells of the DG region, immunoreactivity for Smad2 and Smad3 was also detected, but there was little change in the pattern of their immunoreactivity after ischemia (data not shown). Smad2 immunoreactivity was also detected in small-sized cells throughout the hippocampus. Since these small-sized cells were also immunopositive for Iba-1, a marker for microglia, these cells were considered to be microglia (Fig. 6A and B). Smad2 immunoreactivity was detected in a portion of Iba-1-immunopositive microglia. After 5 min of ischemia, the number of Smad2-immunopositive microglia tended to increase (Fig. 6A). Neither Smad2 nor Smad3 immunoreactivity was detected in GFAP- or CNPase-immunopositive cells (data not shown). GFAP and CNPase are makers for astrocytes and oligodendrocytes, respectively. Therefore, these results suggest that Smad2 and Smad3 are not present in these cells. The time-course of nuclear accumulation of Smad2 and Smad3 in the CA1 region was also investigated by western blot analysis (Fig. 7). Nuclear accumulation of Smad2 and Smad3 was enhanced as early as 3 h after ischemia, and was significantly increased at 10 h after ischemia, compared to both naïve and sham-operated rats.

1600 1200 800 400 0

Reperfusion after 5 min of ischemia

C Relative band density of Smad3/β-actin (% of naïve)

32

800 600 400 200 0

Reperfusion after 5 min of ischemia

It has been reported that ALK4 and ALK7, activin type I receptors, and ActR-II, activin type II receptors, are expressed in neurons of the CA1, CA3, and DG regions of the hippocampus [27–29]. To examine whether the nuclear accumulation of Smad2 and Smad3 is specific to the CA1 region, we performed immunohistochemistry using hippocampal sections from animals subjected to intracerebroventricular injection of recombinant human activin A. We used hippocampal sections harvested at 4 and 6 h after the injection of activin A. As early as 4 h after the injection of activin A, immunoreactivity for Smad2 and Smad3 was observed in the nucleus of pyramidal cells of the CA1 and CA3 regions, and in granule cells of the DG region (Fig. 9). Nuclear immunoreactivity for Smad2 and Smad3 was also observed at 6 h after the injection of activin A (data not shown). Under our experimental conditions, nuclear accumulation of Smad2 and Smad3 was observed in the hippocampus over a widespread area, ranging at least from 3.8 to 5.8 mm caudal to Bregma. These results suggest that pyramidal cells of the CA1 and CA3 regions, as well as granule cells of the DG region have the potential to accumulate Smad2 and Smad3 in their nuclei.

Fig. 7. Level of nuclear accumulation of Smad2 and Smad3 in the hippocampal CA1 region. (A) Western blot analysis of Smad2 and Smad3 in the CA1 region after 5 min of ischemia. Quantitative analysis of Smad2 nuclear accumulation (B) and Smad3 nuclear accumulation (C). Data are expressed as the mean ± standard deviation (SD). The changes in the density of bands among different groups were tested by one-way ANOVA, followed by Tukey's multiple comparison test. *Statistically significant difference from rats subjected to 5 min of ischemia followed by 10 h of reperfusion (P b 0.05) n = 4. Sham: sham-operation.*

3.6. RT-PCR analysis of the expression level of I-Smad mRNA in the hippocampus after ischemia

3.7. Effect of Smad activation on the hippocampus

As shown in Fig. 9, activin A induced the nuclear accumulation of Smad2 and Smad3 in the CA3/DG regions as well as the CA1 region. In contrast, ischemia induced the nuclear accumulation of Smad2 and Smad3 in pyramidal neurons of the CA1 region, but not in neurons of

the CA3/DG regions. Therefore, we hypothesized that the inhibition of Smad signaling may be more influential in the CA3/DG regions than in the CA1 region. I-Smads play a key role in the regulation of Smad signaling activation [30]. To examine whether the expression level of Smad6 and Smad7 is higher in the CA3/DG regions compared to the CA1 region, we performed RT-PCR. The band density of PCR products of Smad7 mRNA did not increase in samples from the CA3/DG regions in comparison with those of the CA1 region (Fig. 10). Under our conditions, the PCR product of Smad6 was not detectable.

Since the nuclear accumulation of Smad2 and Smad3 occurred primarily in the CA1 region after ischemia, we hypothesized that the activation of Smad might be an inducible factor contributing to neuronal degeneration. To test this hypothesis, we examined the histological changes in the hippocampus 7 days after activin A injection. Activin A

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

33

A CA1

CA3/DG

Inhibin βA

Inhibin βA

TGF β1

TGF β1

TGF β2

TGF β2

β actin

β actin Reperfusion after 5 min of ischemia

Reperfusion after 5 min of ischemia

B 1200

CA3/DG

∗

∗

∗

900 600 300 0

Reperfusion after 5 min of ischemia

Inhibin βA/β-actinratio (% of naïve)

Inhibin βA/β-actinratio (% of naïve)

CA1

3000

∗

∗

∗

2000 1000 0

Reperfusion after 5 min of ischemia

Fig. 8. RT-PCR analysis for inhibin βA, TGF-β1, and TGF-β2 mRNA in the CA1 region and CA3/DG regions after 5 min of ischemia (A). Quantitative analysis of band intensity of PCR products (B). Data are expressed as the mean ± standard deviation (SD). The changes in the density of bands among different groups were tested by one-way ANOVA, followed by Tukey's multiple comparison test. *Statistically significant difference from rats subjected to 5 min of ischemia followed by 3 h of reperfusion (P b 0.05) n = 3. Sham: sham-operation.

injection did not induce any notable change in the histology of the CA1, CA3, or DG regions (data not shown). Next, to clarify the role of Smad activation on the hippocampus after ischemia, we compared the extent of ischemia-induced neuronal degeneration in the CA1 region between rats that received injection of activin A and vehicle controls. There was no significant difference in the number of surviving neurons in the CA1 region between rats that received the injection of activin A and those that received vehicle control (Fig. 11). 3.8. Localization of Smad2 and Smad3 in the hippocampus after non-lethal ischemia We examined whether nuclear accumulation of Smad2 and Smad3 occurs in the CA1 region after 3 min of ischemia, i.e., non-lethal ischemia [5]. In the hippocampus of animals subjected to 3 min of ischemia followed by 10 h of reperfusion, no nuclear accumulation of Smad2 and Smad3 was observed (Fig. 12A). Western blot analysis also showed no significant difference in the nuclear localization of Smad2 and Smad3 between naïve, sham-operated rats and rats subjected to 3 min of ischemia (Fig. 12B, C, and D). 4. Discussion The major findings of the present study are as follows: 1) Smad1, -2, -3, and -5 proteins were detected in the rat hippocampus by western blot and immunohistochemistry; 2) among Smad1, -2, -3, and -5 proteins, Smad2 and Smad3 proteins accumulated into the nucleus of pyramidal cells in the CA1 region, which is vulnerable to damage after 5 min of ischemia; 3) 3 min of ischemia, i.e., non-lethal ischemia, did not induce the nuclear accumulation of Smad2 or Smad3 in the CA1 region; 4) after injection of activin A, nuclear accumulation of Smad2 and Smad3 was induced not only in pyramidal cells of the CA1 region, but also in pyramidal cells of the CA3 region as well as in granule cells of

the DG region; 5) activin A-induced nuclear accumulation of Smad2 and Smad3 neither caused degeneration of hippocampal neurons nor prevented degeneration induced by ischemia. Based on these results, ischemia-induced nuclear accumulation of Smad2 and Smad3 may be associated with the response to ischemic stress, but may have little relevance to either neuronal survival or death in the ischemic hippocampus. In the present study, we were able to amplify the PCR products for Smad1, -2, -3, -5, and -8 mRNA using hippocampal samples, indicating the expression of these genes in the hippocampus. In contrast, we were able to detect Smad1, -2, -3, and -5 proteins but not Smad8 protein in the hippocampus by western blot analysis. This may reflect a low expression level of Smad8 protein in the hippocampus. When TGF-βs, activins, and BMPs bind to their specific receptors on the cell surface, the receptors activate R-Smads via carboxy-terminal phosphorylation. Once phosphorylated, Smad1, -2, -3, and -5 are activated and translocated into the nucleus [17]. Among Smad1, -2, -3, and -5, only Smad2 and Smad3 were found to accumulate in the nucleus after 5 min of ischemia. This suggests that 5 min of ischemia is sufficient to induce the Smad2/3 signaling cascade in the hippocampus. Smad2 and Smad3 are downstream signaling molecules of activins and TGF-βs [17,18,30]. Here, we confirmed that the RT-PCR product of inhibin βA mRNA was increased not only in samples from the CA1 region subjected to 5 min of ischemia, but also in samples from the CA3/DG regions subjected to 5 min of ischemia. This finding suggests that activin A has an effect not only on the CA1 region, but also on the CA3/DG regions. In contrast, intracerebroventricular injection of activin A induced the nuclear accumulation of Smad2 and Smad3 in the CA3/DG regions as well as the CA1 region, suggesting that dentate granule cells as well as CA1 and CA3 pyramidal cells have the potential to accumulate Smad2 and Smad3 proteins in the nucleus. After binding of activin A to activin type II receptors, activin type I receptors interact with the activin A/activin type II receptor complex, leading to the phosphorylation of Smad2 and Smad3. It has been reported that ALK4 and ALK7, activin type I receptors,

34

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

A

Smad3

Cresylviolet

Activin

DG

Vehicle

Activin

CA3

Vehicle

Activin

CA1

Vehicle

Smad2

B

20 μm

CA1

CA3

DG

Bregma: -3.3 mm

Bregma: -5.8 mm Fig. 9. Nuclear accumulation of Smad2 and Smad3 immunoreactivity in hippocampus subjected to intracerebroventricular injection of activin A. (A) Immunoreactivity for Smad2 and Smad3 in CA1, CA3, and DG regions of hippocampi from animals injected with activin A or vehicle control. Right column shows cresyl violet-stained sections. (B) Schematic representation of the distribution of activin A sensitive neurons in the CA1 (left column), CA3 (middle column), and DG (right column) regions of the hippocampus. Neurons with Smad2 or Smad3immunopositive nuclei are represented by red dots. Upper and lower panels show coronal sections at 3.3 and 5.8 mm posterior to bregma, respectively.

and ActR-II, type II receptors, are expressed in neurons of the CA1, CA3, and DG regions of the hippocampus [27–29]. Smad6 and Smad7 inhibit the activation of R-Smad by inducing degradation of the receptors, or by competing with the R-Smad for type I receptor binding [17,30]. Here, we confirmed that the RT-PCR products of Smad6 and Smad7 mRNA were not increased in the CA3/DG regions after ischemia. Therefore, it is unlikely that the inhibitory system for the activation of Smad2 and Smad3 has a greater effect in the CA3/DG regions than in the CA1 region. Further studies are required to clarify why the response of Smad signaling in the CA3/DG regions is different from the response in the CA1 region. In the present study, we did not provide direct evidence that the nuclear accumulation of Smad2 and Smad3 in the ischemic hippocampus is caused by activin A, and are therefore unable to conclude that activin

A mediates the nuclear accumulation of Smad2 and Smad3 in the ischemic hippocampus. Moreover, we did not find that the expression level of TGF-β mRNA was significantly up-regulated in the ischemic hippocampus. However, this result does not rule out the possibility that ischemia-induced nuclear accumulation of Smad2 and Smad3 is mediated by TGF-β. After synthesis in cells, TGF-β is secreted as part of an inactive complex. Bioactive TGF-β is liberated from this inactive complex by proteolytic cleavage or acidosis [31]. The nuclear accumulation of Smad2 and Smad3 may be caused by the action of bioactive TGF-β in the CA1 region after ischemia. In any case, the factors influencing the activation of Smad2 and Smad3 in the CA1 region are currently unclear. In the brain, inhibin βA subunits, which are constituents of activin A, are extensively expressed [22,32]. Several functions of activin A in the brain have been proposed, with a number of studies suggesting that

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

A

A

CA1

Reperfusion after 5 min of ischemia

3 min ischemia

Reperfusion after 5 min of ischemia

Sham

β actin

CA3/DG

B

20 μm

B

150

Smad2 Smad3 β-actin

100

50

Ischemia

C Reperfusion after 5 min of ischemia CA1

Reperfusion after 5 min of ischemia CA3/DG

Fig. 10. RT-PCR analysis for Smad7 mRNA in the CA1 region and the CA3/DG regions after 5 min of ischemia (A). Quantitative analysis of band intensity of PCR products (B). Data are expressed as the mean ± standard deviation (SD), n = 3.

150 100 50 0

150 100 50 0

Fig. 12. The localization of Smad2 and Smad3 in the CA1 region after 3 min of ischemia followed by 10 h of reperfusion. (A) Photographs of immunohistochemically stained sections from naïve rats, sham-operated rats, and rats subjected to 3 min of ischemia followed by 10 h of reperfusion. (B) Western blot analysis of Smad2 and Smad3 in the hippocampus after 3 min of ischemia. Samples from rats subjected to 5 min of ischemia followed by 10 h of reperfusion were used as positive controls. Quantitative analysis of Smad2 nuclear accumulation (C) and Smad3 nuclear accumulation (D). Data are expressed as the mean ± standard deviation (SD), n = 3. Sham: sham-operation.

activin A-containing neurons in the hypothalamus are involved in modulating the secretion of adrenocorticotropic hormone or of luteinizing hormone [33,34]. Intracerebroventricular injection of activin A is known to modulate synaptic transmission in the hippocampus [35].

A

D 200

Relative band density of Smad3/β−actin (% of naive)

0

Relative band density of Smad2/β-actin (% of naive)

Smad7/β-actin ratio (% of naïve in CA1 region)

Smad3

Smad2

Naïve

Smad7

35

B

*

Vehicle/ischemia

Activin/ischemia

Number of neurons in CA/mm

Vehicle/sham 140

*

120 100 80 60 40

n.s.

20 0

5 min of ischemia

50 μ m

Fig. 11. The effect of activin A on ischemia-induced neuronal degeneration in the CA1 region. (A) Photographs of hematoxylin/eosin-stained sections from sham-operated rats with vehicle injection, rats subjected to 5 min of ischemia after vehicle injection, and rats subjected to 5 min of ischemia after activin A injection. (B) The changes in the number of pyramidal neurons among the different groups were tested by one-way ANOVA, followed by Tukey's multiple comparison test. *Statistically significant difference from sham-operated rats with vehicle injection (P b 0.05) n = 4. Data are expressed as the mean ± standard deviation (SD). Sham: sham-operation. n.s.: no-significance.

36

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37

Moreover, an in vitro study has suggested that activin A is associated with the modulation of dendritic spine morphology, which is important for synaptic plasticity [36]. These findings suggest that activin A plays an important role in physiological brain functions. With regard to pathological conditions, increased expression of inhibin βA subunit mRNA has been demonstrated by several studies using different brain injury models, including kainic acid-induced excitotoxicity [37,38], brain infarction [39], and hypoxic–ischemic encephalopathy [40]. The administration of activin A has been shown to protect against neuronal degeneration in these brain injury models [24,41], and activin A is therefore believed to be an endogenous neuroprotectant. However, the molecular mechanisms underlying the neuroprotective effects of activin A remain to be elucidated. In contrast to activin A, few reports have focused on the function of Smad signaling in brain injury. In the current study, the nuclear accumulation of Smad2 and Smad3 did not prevent ischemia-induced neuronal degeneration in the hippocampal CA1 region, suggesting that the activation of these Smad proteins has little relevance to neuroprotection against ischemia. This study does not rule out the possibility of a neuroprotective effect of activin A; instead we focused on the neuroprotective effect of Smad activation against ischemic neuronal cell death. In the present study, activin A (1 μg of recombinant activin A per unilateral cerebral ventricle) was injected into bilateral cerebral ventricles at 4 h prior to the ischemia, which was sufficient to induce the accumulation of Smad2 and Smad3 into the nucleus. It remains possible that the failure to attenuate neuronal degeneration is due to an inadequate dose and duration of treatment. In the current study, the nuclear accumulation of Smad2 and Smad3 was observed in the CA1 region, which is known to be vulnerable to ischemia, but not in the CA3/DG regions, which are regions more resistant to ischemia. Moreover, 3 min of ischemia did not induce the nuclear accumulation of Smad2 and Smad3 in the CA1 region. These findings suggest that Smad2 and Smad3 signaling is associated with neuronal degeneration. However, it is apparent that Smad2 and Smad3 signaling was not involved in the induction of neuronal degeneration, given the present result showing that activin A-induced Smad2 and Smad3 signaling did not induce neuronal degeneration. Therefore, it is likely that the nuclear accumulation of Smad2 and Smad3 reflects the response to stress in the ischemic hippocampus. A previous study using cultured microglia showed that TGF-β1 prevented microglial chemotaxis toward β-amyloid [42], while another study showed that TGF-β1 was associated with the expression of several microglial genes [43]. These studies also showed that Smad2 was activated in microglia following treatment with TGF-β1. However, the functional significance of Smad signaling in microglia remains unclear. In our study, Smad2 immunoreactivity was also present in hippocampal microglia, under both normal and ischemic condition. This is the first report showing the localization of Smad2 in microglia of the normal and ischemic hippocampus. Taken together, our results suggest that Smad2 plays a role in the physiological and pathological function of microglia, although its precise function in this capacity remains unclear. In the present study, Smad2 immunoreactivity was detected in a portion of Iba-1-immunopositive microglia, the number of which tended to increase following ischemia. These findings suggest that microglia expressing Smad2 form a discrete functional subset from those not expressing Smad2. From a functional viewpoint, microglia can be divided into 2 phenotypes; pro-inflammatory (M1) and anti-inflammatory (M2) [44]. M1 microglia release destructive pro-inflammatory mediators, while M2 microglia release anti-inflammatory mediators. To distinguish between these 2 microglial phenotypes, several marker proteins can be used [45,46]. Interleukin (IL)-6, the inducible form of nitric oxide synthase (iNOS), and tumor necrosis factor-α (TNF-α) can all be used as markers for M1 microglia. On the other hand, IL-10, arginase-1, Ym1/2 lectin, and vascular endothelial growth factor (VEGF) can be used as markers for M2 microglia. Therefore, double-labeled immunofluorescent staining using antibodies for these marker proteins as well as Smad2 may help us to better understand the function of Smad2-immunopositive microglia.

Conflict of interest The authors have no conflict of interest. References [1] Roine RO, Kajaste S, Kaste M. Neuropsychological sequelae of cardiac arrest. JAMA 1993;269:237–42. [2] Kirino T. Delayed neuronal death in the gerbil hippocampus following ischemia. Brain Res 1982;239:57–69. [3] Pulsienelli WA, Brierley JB, Plum F. Temporal profile of neuronal damage in a model of transient forebrain ischemia. Ann Neurol 1982;11:491–8. [4] Fukuda S, Harada K, Kunimatsu M, Sakabe T, Yoshida K. Postischemic reperfusion induces α-fodrin proteolysis by m-calpain in the synaptosome and nucleus in rat brain. J Neurochem 1998;70:2526–32. [5] Nakajima T, Ochi S, Oda C, Ishii M, Ogawa K. Ischemic preconditioning attenuates of ischemia-induced degradation of spectrin and tau: implications for ischemic tolerance. Neurol Sci 2011;32:229–39. [6] Nakajima T, Wakasa T, Okuma Y, Nomura Y, Kuwabara M, Kawahara K. Dual inhibition of protein phosphatase-1/2A and calpain rescues nerve growth factordifferentiated PC12 cells from oxygen-glucose deprivation-induced cell death. J Neurosci Res 2006;83:459–68. [7] Shamloo M, Rytter A, Wieloch T. Activation of the extracellular signal-regulated protein kinase cascade in the hippocampal CA1 region in a rat model of global cerebral ischemic preconditioning. Neuroscience 1999;93:81–8. [8] Nakajima T, Iwabuchi S, Miyazaki H, Okuma Y, Kuwabara M, Nomura Y, et al. Preconditioning prevents ischemia-induced neuronal death through persistent Akt activation in the penumbra region of the rat brain. J Vet Med Sci 2004;66:521–7. [9] Yano S, Morioka M, Fukunaga K, Kawano T, Hara T, Kai Y, et al. Activation of Akt/protein kinase B contributes to induction of ischemic tolerance in the CA1 subfield of gerbil hippocampus. J Cereb Blood Flow Metab 2001;21:351–60. [10] Hara T, Hamada J-I, Yano S, Morioka M, Kai Y, Ushino Y. CREB is required for acquisition of ischemic tolerance in gerbil hippocampal CA1. J Neurochem 2003;86:805–14. [11] Nakajima T, Iwabuchi S, Miyazaki H, Okuma Y, Inanami O, Kuwabara M, et al. Relationship between the activation of cyclic AMP responsive element binding protein and ischemic tolerance in the penumbra region of rat cerebral cortex. Neurosci Lett 2002;331:13–6. [12] Roberts AB, Sporn MB. Physiological actions and clinical applications of transforming growth factor-β (TGF-β). Growth Factors 1993;8:1–9. [13] Bilezikjian LM, Blount AL, Leal AM, Donaldson CJ, Fischer WH, Vale WW. Autocrine/paracrine regulation of pituitary function by activin, inhibin and follistatin. Mol Cell Endocrinol 2004;225:29–36. [14] Ling N, Ying SY, Ueno N, Shimasaki S, Esch F, Hotta M, et al. Pituitary FSH is released by a heterodimer of the β-subunits from the two forms of inhibin. Nature 1986;321:779–82. [15] Shiozaki M, Kosaka M, Eto Y. Activin A: a commitment factor in erythroid differentiation. Biochem Biophys Res Commun 1998;242:631–5. [16] Schmitt JM, Hwang K, Winn SR, Hollinger JO. Bone morphogenetic proteins: an update on basic biology and clinical relevance. J Orthop Res 1999;17:269–78. [17] Derynck R, Zhang YE. Smad-dependent and Smad-independent pathways in TGF-β family signalling. Nature 2003;425:577–84. [18] Tsuchida K, Nakatani M, Hitachi K, Uezumi A, Sunada Y, Ageta H, et al. Activin signaling as an emerging target for therapeutic interventions. Cell Commun Signal 2009;7:15. [19] Knuckey NW, Finch P, Palm DE, Primiano MJ, Johanson CE, Flanders KC, et al. Differential neuronal and astrocytic expression of transforming growth factor β isoforms in rat hippocampus following transient forebrain ischemia. Mol Brain Res 1996;40:1–14. [20] Unsicker K, Flanders KC, Cissel DS, Lafyatis R, Sporn MB. Transforming growth factor β isoforms in the adult rat central and peripheral nervous system. Neuroscience 1991;44:613–25. [21] Vincze C, Pál G, Wappler EA, Szabó ER, Nagy ZG, Lovas G, et al. Distribution of mRNAs encoding transforming growth factors-β1, -2, and -3 in the intact rat brain and after experimentally induced focal ischemia. J Comp Neurol 2010;518:3752–70. [22] Roberts VJ, Barth SL, Meunier H, Vale W. Hybridization histochemical and immunohistochemical localization of inhibin/activin subunits and messenger ribonucleic acids in the rat brain. J Comp Neurol 1996;364:473–93. [23] Henrich-Noack P, Prehn JH, Krieglstein J. TGF-β1 protects hippocampal neurons against degeneration caused by transient global ischemia. Dose–response relationship and potential neuroprotective mechanisms. Stroke 1996;27:1609–14. [24] Wu DD, Lai M, Hughes PE, Sirimanne E, Gluckman PD, Williams CE. Expression of the activin axis and neuronal rescue effects of recombinant activin A following hypoxic– ischemic brain injury in the infant rat. Brain Res 1999;835:369–78. [25] Schulz R, Vogel T, Dressel R, Krieglstein K. TGF-β superfamily members, Activin A and TGF-β1, induce apoptosis in oligodendrocytes by different pathways. Cell Tissue Res 2008;334:327–38. [26] Schuster N, Bender H, Rössler OG, Philippi A, Dünker N, Thiel G, et al. Transforming growth factor-β and tumor necrosis factor-α cooperate to induce apoptosis in the oligodendroglial cell line OLI-neu. J Neurosci Res 2003;73:324–33. [27] Bengtsson H, Söderström S, Ebendal T. Expression of activin receptors type I and II only partially overlaps in the nervous system. Neuroreport 1995;7:113–6. [28] Morita N, Takumi T, Kiyama H. Distinct localization of two serine–threonine kinase receptors for activin and TGF-β in the rat brain and down-regulation of type I activin receptor during peripheral nerve regeneration. Mol Brain Res 1996;42:263–71.

T. Nakajima et al. / Journal of the Neurological Sciences 337 (2014) 25–37 [29] Tsuchida K, Sawchenko PE, Nishikawa S, Vale WW. Molecular cloning of a novel type I receptor serine/threonine kinase for the TGF β superfamily from rat brain. Mol Cell Neurosci 1996;7:467–78. [30] Pangas SA, Woodruff TK. Activin signal transduction pathways. Trends Endocrinol Metab 2000;11:309–14. [31] Khalil N. TGF-β: from latent to active. Microbes Infect 1999;1:1255–63. [32] Sawchenko PE, Plotsky PM, Pfeiffer SW, Cunningham Jr ET, Vaughan J, Rivier J, et al. Inhibin β in central neural pathways involved in the control of oxytocin secretion. Nature 1988;334:615–7. [33] MacConell LA, Widger AE, Barth-Hall S, Roberts VJ. Expression of activin and follistatin in the rat hypothalamus: anatomical association with gonadotropinreleasing hormone neurons and possible role of central activin in the regulation of luteinizing hormone release. Endocrine 1998;9:233–41. [34] Plotsky PM, Kjaer A, Sutton SW, Sawchenko PE, Vale W. Central activin administration modulates corticotropin-releasing hormone and adrenocorticotropin secretion. Endocrinology 1991;128:2520–5. [35] Ageta H, Ikegami S, Miura M, Masuda M, Migishima R, Hino T, et al. Activin plays a key role in the maintenance of long-term memory and late-LTP. Learn Mem 2010;17:176–85. [36] Shoji-Kasai Y, Ageta H, Hasegawa Y, Tsuchida K, Sugino H, Inokuchi K. Activin increases the number of synaptic contacts and the length of dendritic spine necks by modulating spinal actin dynamics. J Cell Sci 2007;120:3830–7. [37] Tretter YP, Hertel M, Munz B, ten Bruggencate G, Werner S, Alzheimer C. Induction of activin A is essential for the neuroprotective action of basic fibroblast growth factor in vivo. Nat Med 2000;6:812–5.

37

[38] Tretter YP, Munz B, Hübner G, ten Bruggencate G, Werner S, Alzheimer C. Strong induction of activin expression after hippocampal lesion. Neuroreport 1996;7:1819–23. [39] Mukerji SS, Katsman EA, Wilber C, Haner NA, Selman WR, Hall AK. Activin is a neuronal survival factor that is rapidly increased after transient cerebral ischemia and hypoxia in mice. J Cereb Blood Flow Metab 2007;27:1161–72. [40] Lai M, Sirimanne E, Williams CE, Gluckman PD. Sequential patterns of inhibin subunit gene expression following hypoxic–ischemic injury in the rat brain. Neuroscience 1996;70:1013–24. [41] Mukerji SS, Rainey RN, Rhodes JL, Hall AK. Delayed activin A administration attenuates tissue death after transient focal cerebral ischemia and is associated with decreased stress-responsive kinase activation. J Neurochem 2009;111:1138–48. [42] Huang WC, Yen FC, Shie FS, Pan CM, Shiao YJ, Yang CN, et al. TGF-β1 blockade of microglial chemotaxis toward Aβ aggregates involves SMAD signaling and downregulation of CCL5. J Neuroinflammation 2010;7:28. [43] Spittau B, Wullkopf L, Zhou X, Rilka J, Pfeifer D, Krieglstein K. Endogenous transforming growth factor-β promotes quiescence of primary microglia in vitro. Glia 2013;61:287–300. [44] Olah M, Biber K, Vinet J, Boddeke HW. Microglia phenotype diversity. CNS Neurol Disord Drug Targets 2011;10:108–18. [45] Crain JM, Nikodemova M, Watters JJ. Microglia express distinct M1 and M2 phenotypic markers in the postnatal and adult central nervous system in male and female mice. J Neurosci Res 2013;91:1143–51. [46] Hu X, Li P, Guo Y, Wang H, Leak RK, Chen S, et al. Microglia/macrophage polarization dynamics reveal novel mechanism of injury expansion after focal cerebral ischemia. Stroke 2012;43:3063–70.