Extracellular Acidification Activates cAMP Responsive Element Binding Protein via Na+/H+ Exchanger Isoform 1–Mediated Ca2+ Oscillation in Central Nervous System Pericytes
Objective—We have previously shown that Na+/H+ exchanger isoform 1 (NHE1) plays an important role in Ca2+ signaling and cell proliferation in human central nervous system (CNS) pericytes. The aims of the present study were to elucidate how NHE1–induced Ca2+ signaling during acidosis is transformed into cellular responses in CNS pericytes.
Methods and Results—Human CNS pericytes were cultured, and the activation of cAMP responsive element–binding protein (CREB) was evaluated by Western blotting analysis, immunofluorescence, and luciferase assays. In human CNS pericytes, low extracellular Na+ or low pH generated Ca2+ oscillation and subsequently phosphorylated Ca2+/calmodulin-dependent kinase II (CaMKII) and CREB in a time-dependent manner. Focal cerebral ischemia was applied using photothrombotic distal middle cerebral artery occlusion in mice, and the phosphorylation of CREB and the production of interleukin-6 were observed in pericytes migrating into the peri-infarct penumbra during the early phase after ischemic insult.
Conclusion—Our results indicate that extracellular acidosis induces Ca2+ oscillation via NHE1, leading to Ca2+/CaMKII–dependent CREB activation in human CNS pericytes. Acidosis may upregulate a variety of proteins, such as interleukin-6, through the NHE1-Ca2+/CaMKII–CREB pathway in brain pericytes and may thus modulate brain ischemic insult.
Central nervous system (CNS) pericytes form a neurovascular unit with neurons, astrocytes, and vascular endothelial cells. They are essential for various vascular functions in the brain microvasculature, such as angiogenesis, regulation of blood flow, and maintenance of the blood brain barrier. A disruption of the pericytes seems to result in impaired microcirculation, and therefore, to play an important role in a variety of CNS diseases, including cerebral ischemia, Alzheimer disease, and multiple sclerosis.1–6
Acid–base balance disturbance is an early event that occurs in the brain under pathophysiological conditions, such as hypoxia and ischemia. Intracellular and extracellular pH is shown to decrease by 0.8 to 1.2 during ischemia.7–9 Intracellular pH is maintained by the buffering capacity and is actively controlled by transport proteins, such as the Na+/H+ exchanger (NHE), Na+-HCO3− symporter, Cl−-HCO3− exchanger, and Cl−-OH− exchanger.10,11 Among the transporters, NHE1 is ubiquitously expressed in the plasma membrane of virtually all tissues and is a primary regulator of intracellular pH, Na+ concentration, and cell volume.12–15 In addition to acting as a housekeeping protein involved in Na+ and H+ homeostasis, NHE1 has been suggested to regulate cell proliferation, migration, differentiation, apoptosis, organization of the cytoskeleton, and immune responses.14–16 During ischemia, NHE1 seems to be activated and have a role in diverse cellular actions; however, the detailed mechanisms underlying the regulation of these functions by NHE1 are not fully understood.
We have recently shown that NHE1 generates a periodic intracellular Ca2+ increase (Ca2+ oscillation) by releasing Ca2+ from the endoplasmic reticulum into the cytosol by means of its reverse mode (ie, influx of protons and extrusion of Na+ in response to extracellular acidification).17 Ca2+ oscillation is composed of repetitive Ca2+ peaks and probably transmits long-lasting Ca2+-dependent reactions differently from the transient Ca2+ peak. Ca2+ oscillation is suggested to regulate gene transcription via activation of Ca2+-dependent transcriptional factors, such as cAMP responsive element–binding protein (CREB) and nuclear factor of activated T cells,18–20 which have the potential to regulate cellular functions.
CREB is a ubiquitously expressed 43-kDa nuclear transcription factor that activates target genes through cAMP responsive elements (CRE).21 CREB is activated by phosphorylation of the serine residue at position 133 (Ser133) of the protein. The phosphorylation is caused by cAMP-dependent protein kinase A (PKA) and by various other signaling pathways, including ribosomal protein S6 kinase (p90RSK), p38 mitogen-activated protein kinase (MAPK), Akt protein kinase, and Ca2+/calmodulin-dependent kinase (CaMK) pathways.22,23
The mechanism by which Ca2+ signaling is transformed into a variety of cellular functions remains unclear. We hypothesized that the NHE1-mediated Ca2+ oscillation may cause activation of CREB and thereby upregulate various proteins in CNS pericytes. The aim of the present study was to elucidate the mechanisms by which acidosis is transformed into cellular responses in brain pericytes. We first show that extracellular acidosis triggers the Ca2+/CaMKII-dependent CREB activation via NHE1-mediated Ca2+ oscillation and that this pathway upregulates various proteins in CNS pericytes.
Materials and Methods
An expanded Methods section is available in the online-only Data Supplement.
Human brain microvascular pericytes, initiated from normal brain cortical tissues, were purchased from Cell Systems Corporation. The cells were cultured as previously described (Figure I in the online-only Data Supplement).17,24
Measurement of Intracellular Ca2+ Concentration
Mouse Focal Brain Ischemia Model
All animal procedures were approved by the Animal Care and Use Review Committee of Kyushu University (17-067-0). Male S129 mice (8–9 weeks old; body weight 25–30 g; Jackson Laboratory) were used for the present study. We induced focal brain ischemia by photochemical occlusion of the distal middle cerebral artery, as described previously.28–30
All data are expressed as mean values±SD. We have defined n as the number of experiments or treated animals. The statistical analysis was performed using a 1-way ANOVA. Post hoc comparisons were made using Fisher protected least significant difference (PLSD) tests. P<0.05 was considered significant.
Low Extracellular pH or Na+ Mediates Ca2+ Oscillation and Phosphorylation of CREB
A decrease in extracellular pH evoked a repetitive spike-shaped increase in intracellular Ca2+ (Ca2+ oscillation) in human CNS pericytes, as was already shown in a previous study (Figure 1A; Video 1 in the online-only Data Supplement).17 A Western blotting analysis showed discrete bands of 43 kDa for CREB protein in the pericytes, corresponding to previously reported sizes. To elucidate the possible involvement of CREB during the activation of NHE1, the phosphorylation of protein was examined after NHE1 was stimulated in reverse mode by low extracellular pH or Na+. Following the time course of the extracellular low pH, CREB phosphorylated at Ser133 was significantly increased in a time-dependent manner compared with untreated cells. The phosphorylated form of activating transcriptional factor 1, a CREB-related protein, was also detected. The peak of the density for phospho-CREB was observed from 30 to 120 minutes after the stimulation (n=3; Figure 1B).
A decrease in extracellular Na+ evoked a similar Ca2+ oscillation to that evoked by low pH (Figure 1C) and increased the phosphorylation of CREB Ser113 time-dependently. The peak of the density for phospho-CREB was also detected from 30 to 120 minutes after the stimulation (n=3; Figure 1D).
NHE1-Mediated Ca2+ Signaling Activates CREB via Ca2+/CaMK
Several pathways are known to phosphorylate CREB at Ser133. Therefore, we investigated which pathway is predominant during its phosphorylation induced by stimulation with NHE1. The phosphorylation of CREB induced by extracellular acidification was inhibited by 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) (50 μmol/L), a Ca2+ chelator, or KN-93 (10 μmol/L), a CaMK inhibitor. Neither wortmannin (0.1 μmol/L), a phosphoinositide 3-kinase inhibitor, U-0126 (20 μmol/L), a MAPK/extracellular signal-regulated protein kinase (MEK) inhibitor, SB-203580 (10 μmol/L), a p38 MAPK inhibitor, nor H-89 (10 μmol/L), a PKA, affected CREB phosphorylation (n=3; Figure 2A).
Similarly, both BAPTA (50 μmol/L) and KN-93 (10 μmol/L) inhibited the CREB phosphorylation induced by a decrease in extracellular Na+ (n=3; Figure 2B). We have already shown that a decrease in extracellular pH or Na+ evoked Ca2+ oscillation via NHE1. Therefore, a decrease in extracellular pH and extracellular Na+ may activate CREB via the Ca2+/CaMK-dependent pathway.
Because activated CaMK translocates into nuclei and phosphorylates CREB, we further investigated the immunofluorescence staining with anti–phospho-CREB antibody. Immunostaining showed that phospho-CREB was localized in the nuclei of pericytes, when extracellular pH or Na+ was decreased (Figure 2C). KN-93 (10 μmol/L) again inhibited CREB phosphorylation, which is consistent with the results of immunoblotting. In contrast, the phosphorylation of CREB by forskolin (10 μmol/L), which activates adenylate cyclase and produces cAMP, was not affected by KN-93.
We used reverse transcription polymerase chain reaction analysis to examine the expression of CaMK in human brain CNS pericytes. Reverse transcription polymerase chain reaction revealed that transcripts of CaMKI, CaMKIIγ, and CaMKIIΔ were expressed in the pericytes, although we obtained no evidence of CaMKIIα, CaMKIIβ, and CaMKIV expressions (Figure 3A). CaMKII was phosphorylated at threonine residue 286 (Thr286) by extracellular acidosis or removal of extracellular Na+. The peak of the phospho-CaMKII density was observed from 15 to 60 minutes after the stimulation (n=2; Figure 3B), and this activation of CaMKII was inhibited by BAPTA (50 μmol/L) and KN-93 (10 μmol/L) (n=4; Figure 3C).
Activation of CREB Induces CRE-Dependent mRNA Transcription
To confirm the functional activation of CREB during extracellular acidosis, pericytes were transfected with CRE-luc plasmid, and the luciferase assay was performed. The activity of CRE-luc was increased by treatment with extracellular acidosis and was significantly inhibited by KN-93 (10 μmol/L) (n=4; Figure 4A). To investigate the role of NHE1 in Ca2+ signaling and Ca2+/CaMK-dependent CREB activation, anti-NHE1 small interfering RNA duplexes were cotransfected into these cells. The increase in CRE-luc activity by extracellular acidosis was attenuated by knockdown of NHE1 (n=4). Therefore, a decrease in extracellular pH may induce NHE1-mediated Ca2+ oscillation and CRE-dependent transcriptional activity via the Ca2+/CaMK pathway.
Next, we further analyzed the effect of extracellular acidosis on gene expression by DNA microarray analysis. Incubation of cells under acidic conditions resulted in significant changes in numerous genes (Table II in the online-only Data Supplement), and several of these, such as matrix metalloproteinase 10, interleukin (IL) 6, gremlin 2, and myeloid/lymphoid or mixed-lineage leukemia 5, have CRE with TATA boxes in upstream sites.
To confirm the role of CREB activation on gene expression, we focused on the expression of mRNA for IL-6, one of the downstream targets of CREB, and evaluated it by quantitative real-time polymerase chain reaction. A decrease in extracellular pH or Na+ increased the expression of IL-6 mRNA in a time-dependent manner (n=3; Figure 4B). The upregulation of IL-6 was inhibited by KN-93 (10 μmol/L) (n=3; Figure 4C). In addition, the knockdown of NHE1 by RNA interference (RNAi) attenuated the upregulation of IL-6 (n=3; Figure 4D). Therefore, the activation of CREB, mediated by NHE1-regulated Ca2+ oscillation, may induce the upregulation of IL-6 in CNS pericytes.
CREB Is Activated in CNS Pericytes of the Ischemic Penumbra
We investigated the time-course change in the expression of phospho-CREB in cells after focal cerebral ischemia using the photothrombotic distal middle cerebral artery occlusion model in S129 mice (8–9 weeks old; body weight 25–30 g) to clarify whether CREB is phosphorylated in pericytes during ischemia in vivo. Approximately 40 mm3 of cerebral hemisphere resulted in infarction in the middle cerebral artery occlusion models (Figure IIA in the online-only Data Supplement). Platelet-derived growth factor receptor (PDGFR)-β–positive cells, which were relatively small and circular, accumulated in the area surrounding the ischemic lesion 1 day after cerebral ischemia (Figure 5A). These cells were situated around the endothelial cells and also expressed desmin, thus indicating that they had the characteristics of pericytes (Figure IIB in the online-only Data Supplement). The majority of cells expressed phospho-CREB at this phase, whereas PDGFR-β–positive cells were scarce in the contralateral hemisphere, and no phosphorylation of CREB was detected (Figure 5B). During the time course after ischemic insult, the number of PFGFR-β–positive cells significantly increased in the peri-infarct area (n=3 mice; Figure 5C; Figure IIC in the online-only Data Supplement). However, the appearance of these cells changed, with the development of larger cells that were circumflex, and encircled the microvascular lumens with long projections. In the later phase, the number of cells that had colocalized PDGFR-β and phospho-CREB was decreased, whereas NHE1 was persistently coexpressed with PDGFR-β in these cells. Because PFGFR-β–positive cells are thought to be pericytes or vascular smooth muscle cells, phosphorylation of CREB might be associated with time-specific cellular functions of CNS pericytes in the microvasculature in the ischemic penumbra.
We further investigated whether pericytes expressed IL-6 after ischemic insult. The expression of IL-6 mRNA was increased in ischemic hemisphere (n=2; Figure 6A). The number of IL-6–positive cells increased in the ischemic penumbra, and these cells included not only inflammatory cells or neurons but also PDGFR-β–positive cells (Figure 6B). Finally, we used PDGFR-β+/− mice to clarify the role of pericytes (Figure IIIA in the online-only Data Supplement). The expression of IL-6 mRNA in the penumbra area tended to be reduced in the PDGFR-β+/− mice compared with wild-type mice (n=7; Figure IIIB in the online-only Data Supplement).
In the present study, we have shown that a decrease in extracellular pH or Na+ induced Ca2+ oscillation through NHE1 in the reverse mode and led to subsequent phosphorylation of CREB. We have also shown that the phosphorylation of intranuclear CREB was mediated by Ca2+/CaMKII, that the expression of IL-6 mRNA was increased after the activation of CREB in human CNS pericytes, and that the phosphorylation of CREB and the expression of IL-6 were observed in the pericytes that accumulated in the ischemic penumbra. We have previously demonstrated that low extracellular pH or Na+ induced Ca2+ oscillation by releasing Ca2+ from the endoplasmic reticulum via NHE1.17 Therefore, we propose the existence of a novel signaling pathway during acidosis, wherein extracellular acidosis upregulates various proteins according to the following mechanisms: (1) acidosis induces the operation of NHE1 in the reverse mode and causes Ca2+ oscillation; (2) repetitive cytosolic Ca2+ increases mediate the phosphorylation of CaMKII; (3) phosphorylated CaMKII translocates into nuclei and phosphorylates CREB; and (4) phosphorylated CREB binds the CRE region and initiates gene transcription, leading to a variety of cellular responses in human CNS pericytes (Figure IV in the online-only Data Supplement).
The phosphorylation of CREB is required to activate the protein and promote gene transcription. Although its phosphorylation is known to be regulated by various pathways, the mechanism underlying CREB activation by acidosis has not been undefined. It has been suggested that extracellular acidification increases the intracellular cAMP level and phosphorylated CREB via activation of proton-sensitive G-protein–coupled receptors.31,32 Several reports have demonstrated that the intracellular Ca2+ level is important for phosphorylation of CREB through CaMK.23,33 In the present study, the activation of CREB was confirmed by Western blotting analyses, immunofluorescence, luciferase assays, and real-time polymerase chain reaction. Extracellular acidification-induced CREB phosphorylation was inhibited by a Ca2+ chelator and a CaMK inhibitor, whereas the CREB phosphorylation was not affected by a variety of inhibitors that block the other pathways known to activate CREB, including phosphoinositide 3-kinase, MEK, p38 MAPK, and PKA. These results indicate that the cAMP/PKA pathway may not be the primary mechanism for the activation of CREB but that Ca2+ signaling and the CaMKII pathway together induce the subsequent activation of CREB. Therefore, CREB phosphorylation during acidosis may be mainly related to Ca2+ increase and CaMKII activation in CNS pericytes.
CaMKII, a member of the Ca2+/calmodulin-dependent protein kinase family, is a multimeric enzyme composed of 4 homologous subunits derived from 4 different genes and has been implicated in the regulation of cell cycle and gene transcription. CaMKIIγ and CaMKIIΔ mRNA were expressed in human CNS pericytes, and we found the phosphorylation of CaMKII to precede the activation of CREB after the exposure of cells to extracellular acidosis. The binding of Ca2+/calmodulin to the regulatory domain causes conformational changes, autophosphorylates Thr286, and activates CaMKII, and thereafter produces substantial Ca2+/calmodulin-independent autonomous activity. Because the phosphorylation of CaMKII induced by the decrease in extracellular pH or Na+ was inhibited by both a Ca2+ chelator and a CaMK inhibitor, the activation of CaMKII is probably related to Ca2+ oscillation mediated by NHE1. The resting concentration of cytosolic Ca2+ is maintained at a low level, typically between 10−7 and 10−6 mol/L. Because excessive cytosolic Ca2+ is toxic, Ca2+ oscillation may, therefore, be suitable for mediating Ca2+-dependent long-lasting signaling. It should be noted that CaMKII is able to respond to the frequency of Ca2+ oscillation by providing distinct amounts of kinase activity.34 In this respect, the Ca2+ oscillation resulting from acidosis may be responsible for regulating the kinase activity, and the effect on CaMKII/CREB signaling seems to be more sustained.
Gene expression is regulated by several transcriptional factors, including CREB. Our results showed that diverse mRNA transcripts were upregulated by extracellular acidosis in CNS pericytes. More than 100 genes, including IL-6, have so far been reported to contain either palindromic (TGACGTCA) or single (CGTCA) motif CREs for CREB binding in their promoters.35 In the present study, IL-6 gene expression was markedly upregulated by extracellular acidosis in human CNS pericytes. In addition, the upregulation of IL-6 was attenuated by inhibition of Ca2+ signaling and CaMK. IL-6 is known to play a role in cerebral ischemia not only as an inflammatory mediator, which contributes to both the injury and repair processes, but also as a neurotrophic factor.36,37 The IL-6 receptor has a gp130 subunit, which is common to neurotrophic cytokine receptors, such as the receptor for leukemia inhibitory factor and ciliary neurotrophic factor.38 The expression of IL-6 was previously identified in neurons, microglia, and macrophage in the ischemic penumbra,39,40 and the direct injection of IL-6 into the brain after ischemia can reduce the ischemic brain injury.41 Furthermore, IL-6 stimulates the proliferation and migration of cerebral endothelial cells.42 Our results indicate the capability of pericytes to express IL-6. Taken together, the upregulation of IL-6 may contribute to neuroprotection in cerebral ischemia.
Recent studies suggest that pericytes are essential for blood vessel stabilization and maturation in the ischemic brain.3,4,43 It has also been reported that pericytes express cell markers similar to mesenchymal stem cells44 and may play a role in determining the vascular niche.3,45,46 We have shown the neuroprotective roles of pericytes in brain ischemia.47,48 The present study demonstrated that PDGFR-β–positive cells accumulated in the peri-infarct area and encircled capillary lumens during the early phase of ischemia. Furthermore, we have found that several PDGFR-β–positive cells were colocalized with phosphorylated CREB and IL-6 at the early phase of ischemic insult and that the expression of IL-6 in the penumbra area tended to be reduced in the PDGFR-β+/− mice. After brain ischemia, the number of PFGFR-β–positive cells increased during the subacute phase; however, the number of the cells that were positive for phospho-CREB decreased. These results may suggest that immature pericytes are recruited and proliferate locally and subsequently contribute to the maturation of microvascular formulation by interacting with endothelial cells in newly formed capillary lumen.3,4 Phospho-CREB is probably important for the initial responses after ischemia, such as migration, differentiation, and proliferation of the cells, because CREB was transiently phosphorylated in these cells during the early phase of ischemia. CREB in pericytes may contribute to angiogenesis, neurogenesis, and neuroprotection by initiating these processes after brain ischemia.3,4,47,48
In the ischemic penumbra, the decrease in the glucose supply is mild in comparison with the oxygen supply. Because intracellular protons cause mitochondrial disorder, oxidative glucose metabolism is impaired. As a result, lactic acid is produced, and acidosis occurs in the tissue. Our immunohistochemical results are consistent with the notion that extracellular acidosis in the penumbra induced Ca2+ oscillation and subsequently caused CREB activation during the early stage of the ischemic event. However, during cerebral ischemia, various pathological conditions other than acidosis may affect intracellular Ca2+ signaling. Consequently, the Ca2+ signaling evoked by other mechanisms may also be involved in the activation of CREB in CNS pericytes. For instance, hydrogen peroxide evokes a Ca2+ increase by releasing it from the intracellular Ca2+ stores in human CNS pericytes.24,26 In addition, other vasoactive substances produced under pathological conditions can also evoke Ca2+ signaling,26 which may thus lead to CREB phosphorylation in the pericytes. Further study is, therefore, required to clarify the association between NHE1 and CREB in the pericytes during cerebral ischemia.
There are strengths and limitations associated with this study. The cells derived from the human brain microvasculature may be the strength. However, these data were obtained from an in vitro study, and the signaling pathways and their functional roles in brain pericytes must be elucidated in vivo. Further study using genetically engineered mice in which these pathways in the pericytes are specifically disrupted is thus required to elucidate the functional importance of this signaling cascade in pericytes after stroke and brain ischemia.
In conclusion, extracellular acidosis induces Ca2+ oscillation via NHE1 and Ca2+/CaMKII-dependent CREB activation in human CNS pericytes. The pericytes may increase a variety of proteins, including IL-6, through the CREB pathway, thus contributing to the neuroprotective roles in cerebral ischemia.
We thank Hideko Noguchi (Kyushu University Hospital) and Naoko Kasahara (Department of Medicine and Clinical Science, Kyushu University) for their excellent technical assistance. We also thank Brian Quinn (Official Medical Editor for the Japan Society of Internal Medicine and the Japan Surgical Society) for English editing.
Sources of Funding
This study was supported by the Coordination, Support and Training Program for Translational Research, a Grant-in-Aid for Scientific Research (C) (No. 19590992 and No. 22590937) from The Japanese Ministry of Education, Culture, Sports, Science and Technology, and a grant from Mitsubishi Pharma Research Foundation.
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.112.254946/-/DC1.
- Received June 8, 2011.
- Accepted July 31, 2012.
- © 2012 American Heart Association, Inc.
- Shepro D,
- Morel NM
- Guillemin GJ,
- Brew BJ
- Lipton P
- Karmazyn M,
- Gan XT,
- Humphreys RA,
- Yoshida H,
- Kusumoto K
- Slepkov ER,
- Rainey JK,
- Sykes BD,
- Fliegel L
- Nakamura K,
- Kamouchi M,
- Kitazono T,
- Kuroda J,
- Matsuo R,
- Hagiwara N,
- Ishikawa E,
- Ooboshi H,
- Ibayashi S,
- Iida M
- Tomida T,
- Hirose K,
- Takizawa A,
- Shibasaki F,
- Iino M
- Ichiki T
- Sheng M,
- Thompson MA,
- Greenberg ME
- Yao H,
- Sugimori H,
- Fukuda K,
- Takada J,
- Ooboshi H,
- Kitazono T,
- Ibayashi S,
- Iida M
- De Koninck P,
- Schulman H