In this study we report that Rcan1-4 protein and mRNA levels are increased after brain ischemia/reperfusion injury in vivo. Lack of Rcan1 is associated with larger infarct volume and higher expression of inflammation associated genes. Rcan1 expression after I/R injury occurs mainly in astroglial cells, correlating with the increased Rcan1-4 mRNA and protein expression observed in murine astrocytes subjected to hypoxia plus glucose deprivation. Consistent with the effect of Rcan1 deletion in the I/R in vivo model, overexpression of exogenous Rcan1-4 inhibits the production of the inflammatory marker Cox-2, while lack of Rcan1 augments this marker. These results support a protective role for Rcan1 during the inflammatory brain response during stroke.
Our immunoblot analyses indicate that the Rcan1 isoform induced after I/R is Rcan1-4 and not Rcan1-1 (Figure 1). Our data also show that the increased expression occurs mainly in GFAP-positive cells around the infarcted area. Using a different focal brain ischemia model, a recent report showed upregulation of Rcan1 protein in both the neural and glial compartments around the infarcted area . This apparent discrepancy cannot be easily explained; however, in our experiments, in which staining was compared with the equivalent contralateral area, differential Rcan1 labeling was detected only in GFAP-positive cells. For our analysis of Rcan1 cellular localization in the rat MCAO I/R model we used a polyclonal antibody that recognizes both Rcan1-1 and Rcan1-4. While this antibody clearly distinguishes the two isoforms on immunoblots, from their different molecular masses, it cannot differentiate them by immunofluorescence staining. In our analyses, the Rcan1 antibody also stained NeuN-positive cells in the infarcted area, although this staining was no stronger than that observed in the contralateral area of sham-operated animals (data not shown). Thus increased Rcan1 reactivity was restricted to GFAP-positive cells, in agreement with the GFAP-positive staining seen by Cho et al. . Moreover, the immunoblots indicate that the induced expression is due to Rcan1-4, with no significant change in Rcan1-1 (Figure 1). We therefore conclude that at the times examined the increased Rcan1 immunostaining in brain slices is due mostly to the expression of Rcan1-4 in glial cells surrounding the infarcted tissue. Confirmation of this must await the availability of Rcan1-1 and Rcan1-4 specific antibodies for immunohistochemistry analysis.
Treatment of astrocytes with PIo induced the appearance of a slower migrating band for NFATc3 protein (Figure 3A). Calcineurin activity has been classically estimated from the NFAT phosphorylation status [34, 35]. We and others have shown that, in response to PIo, other NFAT proteins such as NFATc2 shift to a faster migrating band that has been considered the dephosphorylated form [13, 36]. We also showed previously that NFATc1, c2, c3 and c4 are present in astrocyte cultures, with NFATc3 the most abundant member . Pretreatment of these cells with CsA retards the gel mobility of all NFAT members analyzed. The unique response of NFATc3 to PIo induction has been reported previously by ourselves in astrocytes and by Urso et al. in Jurkat cells. These authors detected a slower migrating NFATc3 band at the same time (1 h) after exposure to PIo as we observe in astrocytes. This slower migrating form of NFATc3 might be a phosphorylated form generated by the action of PIo-induced kinases. Further experiments would be needed to demonstrate this hypothesis. To confirm that PIo induction in astrocytes efficiently translocates NFATc3 protein to the nucleus, we analyzed nuclear and cytosolic fractions (data not shown), detecting NFATc3 protein in the nuclear fraction of astrocytes treated with PIo for 1 h. When cells are pretreated with CsA, slower migrating NFATc3 forms were generated that run above 190 kDa; these NFATc3 forms accumulate in the cytosolic fraction and are not observed in the nuclear extract. Notably, non-stimulated astrocytes present NFATc3 proteins with intermediate mobilities on western blot analysis.
An anti-neuroinflammatory action of Rcan1 is consistent with inhibition of the CN signaling pathway. CN is a central inflammatory regulator in many cell types. In the brain, CN inhibitors such as CsA and FK506 have been shown to reduce tissue damage in response to stroke (reviewed in ). In line with our findings, Rcan1-4 expression has been shown to attenuate inflammatory and angiogenic responses [14, 15, 38], and a protective effect has also been proposed against oxidative stress . Furthermore, targeted overexpression of constitutively active CN A in astrocytes has been reported to protect against brain inflammatory injury, as measured by negative regulation of inflammation hallmarks such as lipopolysaccharide-inducible Cox-2 and inducible NO synthase (iNOS) . However, this study did not examine the expression of Rcan1-4, which is likely activated in response to the excess CN activity. Further studies to define the mechanism of endogenous modulation of CN activity will be important to determine how Rcan1 and CN influence the final outcome of the neuroinflammatory process.
Our current results with Rcan1 KO mice are compatible with Rcan1 downregulating the CN/NFAT signaling pathway. This notion is further supported by the appearance of slower migrating forms of endogenous NFATc3 in astrocytes overexpressing exogenous Rcan1-4 protein (Figures 4 and 5); these bands are very similar to the high molecular NFATc3 bands observed in CsA-treated cells (Figure 4). We suggest that these slower mobility forms might be hyperphosphorylated NFATc3 proteins. Furthermore, in cells overexpressing Rcan1-4 there is a partial inhibition of PIo-induced expression of the NFAT-dependent gene Cox-2 (Figure 5). The level of inhibition is similar to that obtained with CsA, as we previously described . The potential of Rcan1 proteins to inhibit CN signaling either by an effect on CN phosphatase activity directly or by an action on the CN/NFAT signaling pathway via other means has been the focus of much controversy in the field. Rcan1 was originally identified as a negative regulator of CN activity [18, 41–44]. However, other reports have shown that Rcan1 might not always repress CN activity. For example, heart tissue extracts form Rcan1 KO mice [45, 46] and extracts from rcn1-/- yeast (the Rcan1 homologue in yeast) , display reduced CN activity, leading these authors to conclude that Rcan1 proteins might facilitate CN activity. Biochemical analysis has shown that the specificity of signaling kinases and phosphatases toward their targets is usually mediated by docking interactions with substrates and regulatory proteins, and this is the case with Rcan1 and CN. Martinez-Martinez et al. recently showed that the inhibitory action of Rcan1 on calcineurin-NFAT signaling results not only from the inhibition of CN phosphatase activity, but also from competition between NFAT and Rcan1 for binding to the same docking site on calcineurin . This competition for docking sites has been corroborated by another group . Therefore the inhibition of the CN-dependent pathway might not be by direct inhibition of the activity of the phosphatase CN. We previously showed that CN is able to bind to endogenous Rcan1 proteins in astrocytes (see ), indicating that a similar mechanism might be operating in astrocytes. As we show in Figure 6B, Rcan1 loss of function does not affect CN enzymatic activity in brain cortices, indicating that in our system the inhibitory action of Rcan1 on CN-NFAT signaling is not due to direct inhibition of CN phosphatase activity.
The appearance of reactive gliosis in the periphery of the infarcted tissue is one of the most striking changes occurring in the brain after ischemia . Furthermore, expression of constitutively active CN in astrocyte cultures has been shown to mimic the phenotype and gene transcription of activated astrocytes . Although the exact role of this pathway in astrocyte activation is not completely understood, these findings suggest a scenario in which Ca2+-initiated CN activation is instrumental in the mounting of astrocyte inflammatory responses, with Rcan1-4 expression subsequently induced as an auto regulatory mechanism to prevent uncontrolled gliosis. It will therefore be of interest to study the effect of targeted glial-CN activation on the response to brain injury.
Although the precise mechanism of glial Rcan1-4 upregulation during brain ischemia is unclear, our results support an important contribution via hypoxia-induced CN-NFAT signaling. Little is known about the induction of CN-NFAT signaling by hypoxia, although transcriptional regulation has been reported in pulmonary arterial smooth muscle cells in response to chronic hypoxia  and in neuronal cells as a response to CN activation . Rcan1-4 expression is a reliable marker of NFAT-dependent transcription, and our in vitro model of hypoxia plus glucose deprivation shows rapid (1 h) induction of Rcan1-4 expression and NFATc3 protein dephosphorylation. We have observed a clear inhibition of NFATc3 dephosphorylation by the CN inhibitor CsA (data not shown). We can therefore conclude that HGD activates the CN-NFAT signaling pathway in glial cells. However, other possible CN activators include proinflammatory cytokines such as IL-1β, TNFα, and IL-6, which are detected in the ischemic cortex 1 h after MCAO . Indeed, IL-1β has been reported to activate the CN-NFAT pathway in highly enriched astrocytes . In our hands, IL-1β only weakly induced Rcan1-4 expression in astrocytes, and we conclude that the initial Rcan1-4 expression in primary cultures is likely due to the calcium surge associated with the primary hypoxia. Consistent with an early, direct response to hypoxia, we see Rcan1-4 after 1 h in the in vitro astrocyte HGD model. However, the in vivo profile of Rcan1-4 expression in response to I/R injury is slower, with accumulation evident at 5 h and maximal expression detected at 24 h, the last time point examined. This profile probably reflects the complex multicellular environment of the intact brain, possibly including a contribution from the second wave of inflammatory cascades. The impact of other signaling pathways on Rcan1-4 expression during brain I/R injury is a possibility that should be pursued further.