Role of Central Nervous System Periostin in Cerebral Ischemia
Background and Purpose—Although periostin, an extracellular matrix glycoprotein, plays pivotal roles in survival, migration, and regeneration in various cells, its expression and function in the brain are still unknown. Here, we investigated the expression and role of periostin in the ischemic brain.
Methods—Expression of full-length periostin (periostin 1 [Pn1]) and its splicing variant lacking exon 17 (periostin 2 [Pn2]) was examined by real-time reverse transcription polymerase chain reaction (RT-PCR), Western blotting, and immunohistochemical staining in male C57BL6/J mice. The actions of periostin were examined in adult primary neuronal culture and in a transient middle cerebral artery occlusion (tMCAo) model.
Results—Expression of Pn2, but not of Pn1, mRNA was markedly changed after tMCAo. Pn2 mRNA was decreased in the ischemic core at 3 hours after ischemia. At 24 hours, Pn2 mRNA was significantly increased in both the peri-ischemic and ischemic regions. Periostin was mainly observed in neurons in normal brain. However, neuronal expression of periostin was decreased temporarily in the ischemic region, but increased in astrocytes and around endothelial cells at 24 hours after tMCAo. Of importance, intracerebroventricular injection of Pn2 resulted in a significant reduction in infarct volume at 24 hours after tMCAo associated with phosphorylation of Akt. Also, the Pn2-treated mice survived longer until 1 week after tMCAo. Pn2 significantly inhibited neuronal death under hypoxia and stimulated neurite outgrowth.
Conclusions—Here, we demonstrated that periostin was expressed in the brain, and exogenous Pn2 exhibited neuroprotective effects and accelerated neurite outgrowth. Additional studies on periostin may provide new insights into the treatment of ischemic stroke.
Periostin is a 93-kDa secreted N-glycoprotein that modulates cell-matrix interactions and cell functions in the extracellular matrix (ECM).1 Although periostin was originally found in osteoblasts,2 recent studies showed that periostin plays pivotal roles in cell survival under hypoxic conditions,3 migration of cancer cells,4 and proliferation of cardiomyocytes after acute myocardial infarction.5
Interestingly, periostin has homology with fasciclin I, which is expressed in grasshoppers and Drosophila. Fasciclin I is a cell adhesion molecule expressed in the central nervous system (CNS) during embryonic CNS development in Drosophila.6 Also, laser inactivation of fasciclin I disrupts axon adhesion of grasshopper pioneer neurons.7 Thus, periostin might have pivotal roles in the CNS in mammals, although its expression and function have not been clarified in the adult CNS.
One of the unique characteristics of periostin is its variable regions in the C-terminal regions, which contain exons 15 to 23 (Figure 1A). The presence of splicing variants of periostin is becoming the center of the interest, given that various splicing variants of periostin, lacking some of exons 15 to 23, have been speculated to have different functions.8 Among the splicing variants, periostin lacking exon 17 was shown to have different roles in cancer metastasis. For example, Kyutoku et al reported that full-length periostin prevented the progression and metastasis of breast cancer,9 whereas Kim et al reported that overexpression of a periostin variant that lacked exon 17 suppressed lung metastasis in B16-F10 cells.10
From this viewpoint, we examined the expression of full-length periostin (Pn1, Figure 1A) and a splicing variant of periostin that lacked exon 17 (Pn2, Figure 1A) in the adult mouse brain in this study. Also, we studied the actions of periostin using hypoxic cultured neurons and a transient middle cerebral artery (MCA) occlusion model in mice.
Surgical Procedure and Measurement of Infarct Volume
All procedures were approved by the Institutional Animal Care and Use Committee of Osaka University. Experiments were performed in 6- to 8-week-old male C57BL/6 mice (CLEA Japan Inc.). Some mice were exposed to transient MCA occlusion model (online-only Supplemental Methods, http://stroke.ahajournals.org). The variation in infarct volume is presented in online-only Supplemental Figures S4B and S4C. Overall mortality was 14% at 24 hours after ischemia and 67% at 7 days after MCA occlusion. Ischemic damage was evaluated using sections stained with cresyl violet (online-only Supplemental Methods).
Administration of Recombinant Periostin
Recombinant human Pn1 as full-length periostin and mouse Pn2 periostin that lacks exon 17 (R&D Systems) were injected intracerebroventricularly (online-only Supplemental Methods).
Neurological deficit was assessed using a modification of the Bederson neurological scale, as previously described11 (online-only Supplemental Methods).
Real-Time Reverse Transcription Polymerase Chain Reaction
The cerebral cortex was collected using a punch (FST No.18035–80). RNA of the punched brain was isolated using a QIAGEN RNeasy Lipid TissueMini Kit (Qiagen), according to the manufacturer's recommendations. Each quantitative PCR analysis was performed using an ABI Prism 7700 Sequence Detection System (Applied Biosystems, Applera Co.) with SYBR green staining of DNA double strands. Primer pairs were shown in the online-only Supplemental Methods. Each mRNA value (relative quantification) was normalized by glyceraldehyde-3-phosphate dehydrogenase (GAPDH). RNase Cocktail (Applied Biosystems) was used to degrade RNA.
To examine the neuroprotective effects of periostin in hypoxic conditions, cultured neurons, which were pretreated with Pn2 (10 μg/mL) or bovine serum albumin (10 μg/mL), were exposed to hypoxia. To examine the effects of periostin on neurite outgrowth, neurons were cultured in medium containing Pn1, Pn2, or bovine serum albumin (online-only Supplemental Methods).
Immunohistochemical staining was performed in frozen sections (online-only Supplemental Methods). The images were examined with a Nikon A1 confocal scanning laser microscope, and images were analyzed using NIS Elements software (Nikon). All parameters were set in a similar manner when the signal intensity was compared.
Tissue samples were lysed in radio immunoprecipitation assay buffer containing 150 mmol/L NaF, 2 mmol/L sodium orthovanadate, and protease inhibitors (protease inhibitor mixture; Roche Applied Science). Protein of the total lysate (20 μg) was loaded and blotted. Primary antibodies against periostin (1:25; goat polyclonal; Santa Cruz, SC-49480), phospho-Akt (Ser473; 1:1000, Cell Signaling Technology, #9271) and total Akt (1:1000, Cell Signaling, #9272) were used.
All values are expressed as mean±SEM. Multiple comparisons were performed by ANOVA followed by Dunnett's Test. Two groups were compared by unpaired t test. Survival rates were evaluated by log-rank test. Differences were considered significant at P<0.05.
Expression of Periostin in Adult Brain
To examine the expression of periostin, we first studied the expression of periostin mRNA in the adult mouse brain. Among the various variants of periostin,12 we analyzed full-length periostin (Pn1) and a spliced variant form of periostin (Pn2) that lacked exon 17. Using specific primers (online-only Supplemental Figure S1), mRNA expression of Pn1 and Pn2 was successfully detected in extracts from normal cerebral cortex (Figure 1B, 1C). The relative ratio of Pn1/GAPDH and Pn2/GAPDH mRNA calculated by the relative quantification was 0.46±0.11 and 0.47±0.11, respectively. Consistent with real-time reverse transcription polymerase chain reaction (RT-PCR), Western blotting using S-15 antibody against an epitope (amino acids 110–160) of periostin to recognize both Pn1 and Pn2 showed a 90-kDa band, which was identical to the size of periostin, in the cerebral cortex and hippocampus (Figure 1D). Although a 65-kDa band was also detected, this 65-kDa band might have been the spliced form of periostin via proteolytic processing, given that previous reports also exhibited the same findings.13,14 Immunohistochemical study showed that periostin was widely expressed in the brain (online-only Supplemental Figure S2A), and its expression was observed in MAP2-positive cells (mature neurons), whereas GFAP-positive (astrocytes), Iba-1-positive (microglia), and CD-31-positive (endothelial cells) cells did not express periostin (Figure 2). In the hippocampus, periostin was expressed in MAP2-positive cells and partially in GFAP-positive cells (online-only Supplemental Figure S2C), but not in Iba-1-positive or CD-31-positive cells (data not shown). In the striatum, some cells expressed periostin, and their signals were weak as compared with those in the cerebral cortex (online-only Supplemental Figure S2B). Overall, periostin was expressed mainly in neurons in the adult mouse brain.
Change in Expression of Periostin in Ischemic Brain
To clarify the functional roles of periostin in pathological conditions, we analyzed the temporal profile of mRNA expression in the cerebral cortex after transient MCA occlusion. Although the expression of Pn1 and Pn2 was changed by ischemia, the change in Pn2 mRNA was greater (Figure 3A). At 3 hours after MCA occlusion, the expression of Pn1 and Pn2 mRNA was significantly decreased in the ischemic core. At 24 hours, the expression of Pn2 mRNA was significantly increased in both the peri-ischemic and ischemic regions, whereas the expression of Pn1 mRNA was not changed (Figure 3A). Immunohistochemical staining showed that the expression of periostin was very weak in the ischemic region (region I in Figure 3B and online-only Supplemental Figure S3A) at 3 hours after MCA occlusion. Astrocytes, microglia, and endothelial cells did not express periostin at this time (data not shown). At 24 hours after ischemia, expression of periostin emerged in the infarct region (online-only Supplemental Figure S3B), and the expression was observed around CD-31-positive cells (Figure 3C). In the peri-ischemic region, periostin was expressed in astrocytes (Figure 3D, online-only Supplemental Figure S3B) and neurons, whereas microglia did not express periostin (data not shown). The present study demonstrated that the expression of periostin was decreased in the ischemic region at 3 hours, but was increased in the peri-ischemic and ischemic regions at 24 hours after MCA occlusion, suggesting that periostin might play roles in neuroprotection and/or the healing process of the ischemic brain.
Exogenous Pn2, but Not Pn1, Attenuated Ischemic Injury In Vivo
Given that the expression of periostin was decreased in the ischemic region at 3 hours and increased in the peri-ischemic region at 24 hours after reperfusion, we hypothesized that exogenous supplementation of periostin in the acute stage of ischemia would prevent neuronal damage after ischemia-reperfusion. Pn1 or Pn2 was injected intracerebroventricularly at 30 minutes before ischemia and 5 minutes after reperfusion (1.5 hours after ischemia). First, we examined the effect of 10 μg/mL of periostin because this concentration of periostin was reported to have effects in cardiomyocytes, fibroblasts,15 and 293T cells.16 Although intracerebroventricular injection of Pn1 did not affect the infarct volume, treatment with Pn2 significantly reduced infarct volume at 24 hours after reperfusion, with decreased neurological deficit (Figure 4A, 4B). Injection of Pn1 or Pn2 before ischemia had no influence on cerebral blood flow during ischemia and reperfusion (online-only Supplemental Figure S4A). Next, we examined the 1 μg/mL of Pn2 because another article showed that 1 μg/mL of periostin exhibited effects in cancer cells.17 However, 1 μg/mL of Pn2 did not show any effect on infarct volume (Figure 4A). Of importance, Pn2 (10 μg/mL) injected at 1.5 hours after ischemia (5 minutes after reperfusion), but not at 3.5 hours after ischemia (2 hours after reperfusion), resulted in a significant reduction in the infarct volume (Figure 4C). Finally, we checked whether the effects of injection of Pn2 lasted for 7 days. When mice were treated with Pn2 (10 μg/mL) at 1.5 hours after MCA occlusion, they survived significantly longer (Figure 4D). Given that previous studies showed that periostin activated the p-Akt/PI-3 kinase pathway through activation of the integrin receptor, we examined expression of p-Akt in the injured hemisphere using Western blot analysis (Figure 5). Although the expression of p-Akt was slightly increased at 75 minutes after reperfusion, its expression was much higher in mice treated with Pn2 as compared with PBS-treated mice. Level of p-Akt was the same in mice treated with Pn1 and vehicle-treated mice. These data suggest that Pn2 prevented neuronal death through activation of the Akt signaling pathway.
Pn2 Promoted Neuroprotection and Neurite Outgrowth in Cultured Adult Neurons
To confirm the neuroprotective action of Pn2, we employed primary adult neuronal culture. Addition of recombinant Pn2 (10 μg/mL) to cultured neurons under hypoxia significantly inhibited cell death, whereas a hypoxic condition strongly promoted cell death (Figure 6A, 6B); this indicates that Pn2 might directly act on neurons and protect them from ischemic injury in vivo.
Finally, we checked whether periostin had effects on neurite outgrowth, given that other extracellular matrix glycoproteins, such as fibronectin and tenascin-C, accelerated neurite outgrowth after ischemic injury18 or spinal cord injury.19 When cultured neurons were treated with Pn2 after plating, the average length of the maximum neurite at 48 hours was significantly longer than that of neurons treated with bovine serum albumin (10 μg/mL) or with Pn1 (10 μg/mL; Figure 6C,D). These data indicate that Pn2 has the potential to promote neurite outgrowth.
Recently, the role of ECM in the CNS has been a focus of interest, because glycoproteins in ECM, such as fibronectin and tenascin-C, have various effects in neuronal protection, neurite outgrowth, and the migration of neural stem cells.20–24 However, periostin has not yet been studied in the brain. In the present study, we demonstrated that periostin was expressed in neurons in the brain, and Pn2 showed a marked change of expression after cerebral ischemia. Exogenous Pn2 revealed neuroprotective effects in the ischemic brain as well as in cultured neurons, and Pn2 also promoted neurite outgrowth in cultured neurons. Because periostin is also an important glycoprotein in other tissues, such as the heart and cancers,25,26 we speculate that it also has a critical role in the brain. Given that the expression of Pn1 mRNA did not dynamically change after MCA occlusion, and exogenous Pn1 showed no effect on p-Akt expression in the ischemic brain, it follows that Pn1 and Pn2 might have different roles in the brain.
Interestingly, a recent study showed that periostin promoted incorporation of tenascin-C into the ECM and organized the meshwork architecture of the ECM because of adjacent domains banded to tenascin-C and other ECM proteins, acting as a bridge between tenascin-C and other ECM proteins.27 Tenascin-C is also reported to be linked with lecticans1 and to form perineuronal nets,28 which protect neurons from damage, such as oxidative stress.29 From this viewpoint, it is possible that the linkage of tenascin-C to other ECM proteins was accelerated by exogenous periostin, which augmented the formation of a perineuronal net to protect neurons.
Integrin receptors (αvβ1, αvβ3, or αvβ5) have been proposed as the receptor for periostin in the heart and cancer cells.25,26 Periostin activated the Akt signaling pathway through αvβ3 integrin to increase cellular survival,26 and αv, β1, β3, and β5 integrins to induce the proliferation of differentiated cardiomyocytes. Importantly, in neurons, integrin αv promoted neuronal protection and neurite outgrowth,30 and integrin β1 protected neurons through the Akt/PI3K signaling pathway,31 as well as controlled the migration of neuroblasts.32 Phosphorylation of Akt by Pn2 might explain the effects of Pn2 on neurons, probably through activation of integrin receptors in neurons.
The mechanisms of the different effects among splicing variants have not yet been reported. Given that no increase in p-Akt level was observed in Pn1-treated mice, one of the mechanisms might be related to the different actions of splicing variants on the p-Akt signaling pathway. One possibility is that the different splicing variants might activate a different type of integrin receptors. Another possibility is that the exon 17 in periostin might have an inhibitory effect on integrin-binding affinity, considering that alternatively spliced segments in fibronectin modulated integrin-binding affinity of fibronectin.33 Lack of exon 17 in Pn2 might upregulate integrin-binding affinity. Additional studies on the signaling pathways of splicing variants are necessary for clarification of the mechanisms.
Recently, periostin was reported to induce matrix metalloproteinase-2 (MMP-2) in atherosclerotic and rheumatic cardiac valves.34 In the ischemic brain, MMP-2 is one of the key factors promoting the breakdown of extracellular matrix, resulting in opening of the blood-brain barrier.35 In the present study, we speculate that MMP-2 might not have been increased because of the better outcome in Pn2-treated mice, although additional studies are necessary to clarify the relationship between Pn2 and MMP-2.
One limitation of this study is that administration of Pn2 at 3.5 hours after MCA occlusion did not show an effect on infarct volume. This result suggests the existence of a therapeutic time window of treatment with Pn2. Additional studies are necessary to examine the possibility of extending the therapeutic time window by changing the route of injection or injection dose. Another limitation is that differential induction of Pn1 or Pn2 in the astrocytes and around endothelial cells in ischemic brain was not clarified in the present study. Specific antibodies that recognize Pn1 or Pn2 should be developed to address this issue.
Overall, the present study demonstrated that periostin was expressed in the brain and its expression changed after cerebral ischemia. Also, exogenous Pn2, but not Pn1, promoted neuroprotection and neurite outgrowth. Although we focused on Pn1 and Pn2 in the present study, additional studies to clarify the function of periostins, including other splicing variants,8 might provide new insights into the treatment of ischemic stroke.
Sources of Funding
This work was supported by Takeda Science Foundation and a Japan Heart Foundation/Novartis Grant for Research Award on Molecular and Cellular Cardiology, 2011.
Mohammad Allahtavakoli has joined to this study as a fellow of Japan's Matsumae International Foundation.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.636662/-/DC1.
- Received August 21, 2011.
- Accepted December 12, 2011.
- © 2012 American Heart Association, Inc.
- Järveläinen H,
- Sainio A,
- Koulu M,
- Wight TN,
- Penttinen R
- Takeshita S,
- Kikuno R,
- Tezuka K,
- Amann E
- Gillan L,
- Matei D,
- Fishman DA,
- Gerbin CS,
- Karlan BY,
- Chang DD
- McAllister L,
- Goodman CS,
- Zinn K
- Kim BY,
- Olzmann JA,
- Choi SI,
- Ahn SY,
- Kim TI,
- Cho HS,
- et al
- Katsuragi N,
- Morishita R,
- Nakamura N,
- Ochiai T,
- Taniyama Y,
- Hasegawa Y,
- et al
- Yan W,
- Shao R
- Kudo Y,
- Ogawa I,
- Kitajima S,
- Kitagawa M,
- Kawai H,
- Gaffney PM,
- et al
- Rigato F,
- Garwood J,
- Calco V,
- Heck N,
- Faivre-Sarrailh C,
- Faissner A
- Kii I,
- Nishiyama T,
- Li M,
- Matsumoto K,
- Saito M,
- Amizuka N,
- et al
- Belvindrah R,
- Hankel S,
- Walker J,
- Patton BL,
- Müller U
- Manabe R,
- Ohe N,
- Maeda T,
- Fukuda T,
- Sekiguchi K