Activation of Signal Transducer and Activator of Transcription-3 by a Peroxisome Proliferator-Activated Receptor Gamma Agonist Contributes to Neuroprotection in the Peri-Infarct Region After Ischemia in Oophorectomized Rats
Background and Purpose—The role of the phosphorylated signal transducer and activator of transcription-3 (p-STAT3) after cerebral ischemia by the peroxisome proliferator-activated receptor γ (PPARγ) agonist pioglitazone (PGZ) remains controversial. Whether the increase in p-STAT3 by estrogen is mediated by the estrogen receptor α is also obscure. We examined the role of p-STAT3, PPARγ, and estrogen receptor α against ischemic brain damage after PGZ treatment.
Methods—Female Wistar rats subjected or not subjected to bilateral oophorectomy were injected with 1.0 or 2.5 mg/kg PGZ 2 days, 1 day, and 1 hour before 90-minute middle cerebral artery occlusion–reperfusion and compared with vehicle-control rats.
Results—The cortical infarct size was larger in ovariectomized than in nonovarietomized rats; it was reduced by PGZ treatment. Inversely with the reduction of the infarct size, PPARγ, and p-STAT3 but not estrogen receptor α in the peri-infarct area were increased in PGZ-treated compared with vehicle-control rats. The increase in PPARγ and p-STAT3 was associated with the transactivation of antiapoptotic and survival genes and the reduction of caspase-3 in this area. Inhibitors of PPARγ or STAT3 abolished the PGZ-induced neuroprotection and the increase in p-STAT3. More importantly, p-STAT3 increased by PGZ was bound to PPARγ and the complex translocated to the nucleus to dock to the response element through p-STAT3.
Conclusions—Our findings suggest that the activation in the peri-infarct region of p-STAT3 and PPARγ by PGZ is essential for neuroprotection after ischemia and that PGZ may be of benefit even in postmenopausal stroke patients.
Stroke has devastating sequelae and is a major cause of death.1 The lack of ovarian steroid hormones renders postmenopausal women more vulnerable to cerebrovascular disease than cycling women. Estrogen exerts various vascular and neuroprotective effects2 and is thought to be mediated by estrogen receptors. The signal transducer and activator of transcription-3 (STAT3) is affected by estrogen3; it is activated through phosphorylation of tyrosine kinases-2, a member of the Janus family, in response to a wide variety of external stimuli, including cytokines, hormones,4 growth factors,5 and epidermal growth factor-α 7 nicotinic-, interleukin-, and erythropoietin receptor pathways.6 Phosphorylated STAT3 (p-STAT3) dimerizes and translocates to the nucleus where it binds to specific DNA regions, the STAT-inducible elements, ultimately leading to an increase in gene transcription.7
Dziennis et al8 demonstrated that estradiol upregulated the antiapoptotic gene Bcl-2 and the survival marker MAP2 through an increase in p-STAT3 in the peri-infarct region, thereby inhibiting cerebral ischemic damage in female rats subjected to oophorectomy. They also suggested that this neuroprotective effect may be mediated by estrogen receptor α (ERα). On the other hand, Tureyen et al9 showed that in male rats, the peroxisome proliferator-activated receptor γ (PPARγ) agonists rosiglitazone and pioglitazone (PGZ) exerted protection against ischemic brain damage and that the prevention of STAT3 phosphorylation by rosiglitazone dampened interleukin-6 signaling in infarct regions. The discrepancy between these studies regarding the role of p-STAT3 points to the diverse effects of p-STAT3 in cerebral ischemia. We hypothesized that the role of p-STAT3 would be different depending on the ischemic region and that the neuroprotection by the PPARγ ligand may be associated with the upregulation of ERα in female animals. To investigate the relationship among ERα, PPARγ, and p-STAT3 as molecular mechanisms underlying neuroprotection after ischemia, we used nonovariectomized (OVX−) and ovariectomized (OVX+) rats treated or not treated with PGZ.
We provide new evidence that PGZ reduces the cortical infarct size in OVX+ rats susceptible to cerebral ischemic damage and that the increase in p-STAT3 by PPARγ activation in the peri-infarct area is associated with the transactivation of neuroprotective genes. Unexpectedly, PGZ did not affect ERα. More notably, the binding p-STAT3 and PPARγ was increased and the resulting complex translocated to the nucleus to dock to the response element through p-STAT3. Under estrogen-deficient conditions, the activation of p-STAT3 and PPARγ by PGZ may play an essential role in neuroprotection in the peri-infarct area.
Materials and Methods
Animals and Agents
All experiments were performed on 10-week-old female Wistar rats (Charles River Laboratories, Japan) weighing 250 to 280 g; they were anesthetized with 2% isoflurane in 30% oxygen and 70% nitrous oxide. Four weeks before the experiments they were subjected to bilateral oophorectomy (OVX+) or sham OVX (OVX−). The OVX− and OVX+ rats subjected to 90-minute middle cerebral artery occlusion–reperfusion (MCAO-R) were divided into 3 groups; 1 group was injected intraperitoneally with 1.0 mg/kg PGZ, the other with 2.5 mg/kg PGZ at 2 days and 1 day and 1 hour before MCAO-R; the third group served as the vehicle control (VC). The effect of PGZ, a gift from the Takeda Pharmaceutical Company, was examined 24 hours post-MCAO-R. It was dissolved in dimethylsulfoxide and diluted ×3 with saline just before the injection of 0.4 mL/kg. The PPARγ antagonist GW9662 and the STAT3 inhibitor cucubitacin1 (JSI-124) were purchased from Cayman Chemicals (Ann Arbor, MI) and Merck Chemicals (Tokyo, Japan), respectively. Like PGZ, they were dissolved in dimethylsulfoxide; GW9662 was injected (4 mg/kg, intraperitoneally) 1 hour after MCAO-R and JSI-124 (0.5 mg/kg, intraperitoneally) 15 minutes after MCAO.
For other detailed descriptions of Materials and Methods, see http://stroke.ahajournals.org.
PGZ Reduced the Cortical Infarct Size in OVX+ Rats Susceptible to Cerebral Ischemic Damage
As shown in Figure 1, at 24 hours after MCAO-R, the cortical infarct volume was significantly larger in OVX+ than OVX− rats (P<0.01). In OVX− and OVX+ rats treated with 1.0 (PGZ 1.0) and 2.5 mg/kg PGZ (PGZ 2.5), the infarct volume in the cortex but not the basal ganglia was reduced in a PGZ dose-dependent manner compared with the VC (PGZ 1.0, P<0.05; PGZ 2.5, P<0.01 versus VC; PGZ 1.0 versus PGZ 2.5, P<0.05), respectively. Together with the reduction in the cortical infarct size, the neurological score at 24 hours after MCAO-R was significantly lower in PGZ 2.5 than VC rats (2.16±0.40 versus 3.00±0.68, P<0.05), indicating the amelioration of neurological dysfunction.
There was no significant difference in the cerebral blood flow, blood glucose levels, and blood pressure among the 4 groups (OVX−, OVX+ rats treated with VC, PGZ 1.0, and PGZ 2.5 rats) during and after MCAO (data not shown). The plasma level of estrogen in OVX+ rats (13.1±3.4 pg/mL; n=24) was significantly lower than in OVX− rats (31.3±14.1 pg/mL; n=8; P<0.01); PGZ treatment did not affect the estrogen level.
Activation of STAT3 and PPARγ Was Associated With Apoptosis Inhibition in the Peri-Infarct Region
Next, we focused on p-STAT3, PPARγ, and ERα, molecules that play a role in cortical neuroprotection.
In the peri-infarct region, immunohistochemical studies showed that the expression of PPARγ (Figure 2A) and ERα (Supplemental Figure S1A) was increased in OVX− rats 24 hours after MCAO-R; it was reduced in OVX+ rats as was the level of p-STAT3 (Figure 2A; Supplemental Figure S1A). On the contralateral side without ischemia, these molecules were low level. PPARγ and p-STAT3 but not ERα were increased in OVX+ rats treated with PGZ 2.5 (Figure 2A; Supplemental Figure S1A–B); p-STAT3 peaked at 24 hours after MCAO-R (Supplemental Figure S2A). PPARγ- and p-STAT3-positive cells were found among endothelial cells and neurons (Figure 2C). As the level of PPARγ and p-STAT3 increased, the level of cleaved caspase-3 decreased (Figure 2A). The PPARγ, p-STAT3, and caspase-3 level in the peri-infarct area (Figures 3A, 3B, and 3D), determined by Western blot analysis, reflected our immunohistochemical results in the same region (Figure 2A); total STAT3 was not different among the groups (Figure 3C).
In the infarct core (Figure 2B), immunopositivity for PPARγ and p-STAT3 was not necessarily correlated with the ischemic damage; the expression of cleaved caspase-3 was not different among the 4 groups (OVX−, OVX+ rats treated with VC, PGZ 1.0, or PGZ 2.5). These results suggested that the increase in p-STAT3 and PPARγ in the peri-infarct region but not in the ischemic core is associated with neuroprotection afforded by PGZ.
Furthermore, in the peri-infarct regions, the mRNA level of PPARγ, antiapoptotic Bcl-2 and Bcl-xL, and survival-related MAP2 was reduced in OVX+ compared to OVX− rats (Figure 4A–D); there was an inverse correlation with the expression of caspase-3. In OVX+ rats treated with PGZ, these neuroprotective molecules were upregulated and caspase-3 was reduced. The increase in p-STAT3 and PPARγ in this region appeared to contribute to the transcriptional activation of neuroprotective molecules.
Inhibition of p-STAT3 Abolished Neuroprotection Afforded by PGZ
To assess whether the reduction in p-STAT3 affects the neuroprotection afforded by PGZ and renders neurons susceptible to damage, we used the PPARγ antagonist GW9662 and the STAT3 inhibitor JSI-124. Although the infarct size in OVX+ rats treated with GW9662 or JSI-124 was similar to that in nontreated OVX+ rats, GW9662 or JSI-124 abolished the PGZ-induced reduction in the cortical infarct size (Figure 5A; P<0.01). This was associated with a decrease in p-STAT3 and an increase in caspase-3 in the peri-infarct area (Figure 5B). The PGZ-induced increase in the mRNA level of Bcl-2, Bcl-xL, and MAP2 in OVX+ rats was abolished by the inhibitors (Figure 5C), indicating that PGZ-induced neuroprotection was at least partly p-STAT3-dependent.
PPARγ Facilitated the Binding of p-STAT3 to the Response Element in DNA
To verify that there was a causal link between p-STAT3 and PPARγ, we examined the interaction of these molecules. In cell lysates from the peri-infarct area of PGZ 2.5-treated compared with VC rats, p-STAT3 in PPARγ immunoprecipitates were elevated (Figure 6A), suggesting that PGZ promotes the binding of p-STAT3 to PPARγ. Furthermore, in DNA binding assay using nuclear extracts obtained from the peri-infarct area, docking to the response element through p-STAT3 was increased by PGZ treatment (Figure 6B). On the other hand, binding to the response element through PPARγ was not affected by PGZ. These observations suggest that in the peri-infarct area, PPARγ activated by PGZ facilitates the binding of p-STAT3 to specific DNA regions and contributes to the transactivation of neuroprotective molecules.
The incidence of stroke in women dramatically increases after menopause and it further rises with age.10 However, there are few studies on female mammals that mimic the postmenopausal state in women. We provide new evidence that under conditions of hypoestrogenicity, the cortical infarct size was increased and that it was reduced by PGZ treatment. As the size of the infarct area decreased, PGZ increased the level of p-STAT3 and PPARγ but not ΕRα in the peri-infarct area. The PGZ-induced expression of p-STAT3 was abolished by PPARγ and STAT3 inhibitors, resulting in a decrease in the expression of neuroprotective genes and an increase in apoptosis in this area. More importantly, p-STAT3 increased by PGZ treatment was bound to PPARγ and promoted docking to the DNA response element through p-STAT3. These results suggest that the activation of PPARγ is attributable to STAT3 phosphorylation in the peri-infarct area and that the activation of STAT3 and PPARγ by PGZ contributes, at least partly, to neuroprotection after cerebral ischemia under estrogen-deficient conditions.
Several studies found that STAT3 activated by granulocyte colony stimulating factor,11 NRG-1β,12 or secretoneurin13 exerted neuroprotective effects. Although the mechanisms underlying the activation of p-STAT3 may be different, these studies support our hypothesis that the activation of p-STAT3 and PPARγ in the peri-infarct region is essential for neuroprotection against ischemic brain damage.
PPARγ transduces signals as an obligate heterodimer with the retinoid X receptor.14 These nuclear receptors include the common domains: the N-terminal autonomous activation function domain that interacts with transcriptional coregulator proteins and the DNA binding-, the hinge-, and the C-terminal ligand binding domain that contains a ligand-regulated activation function.15 Agonist ligands for PPARγ activate transcription by promoting the recruitment of coactivators.16 In contrast, antagonist ligands contain an extended pendant group that is found in corepressors such as the silencing mediator for retinoid and thyroid hormone and the nuclear receptor corepressor silence transcription.17 In our study, the transcriptional activation of neuroprotective genes by p-STAT3 through PPARγ in the peri-infarct area may reflect the recruitment of coactivators.
Various interactions between p-STAT3 and PPARγ agonists have been documented in multiple myeloma cells.18 In neurons, the administration of rosiglitazone increased the translocation of PPARγ to the nucleus.19 PPARγ expression after focal ischemia was increased, especially in the peri-infarct area, but surprisingly, the DNA-binding activity of PPARγ was reduced.19,20 These earlier studies support our findings that activation of PPARγ by PGZ in the peri-infarct area increased complex formation with p-STAT3 and that the complex translocated to the nucleus to bind the DNA response element through p-STAT3 but not PPARγ. Furthermore, inhibitors of PPARγ or p-STAT3 abolished the transcriptional activation of neuroprotective genes by PGZ in the peri-infarct area. Taken together, the activation of PPARγ by PGZ may facilitate the transactivation of neuroprotective genes by p-STAT3. To resolve the discrepancy between our and earlier studies by Tureyen et al9 regarding p-STAT3, we examined p-STAT3 level in preliminary studies (Supplemental Figure S2B). Compared with the ischemic core, p-STAT3 was increased in the peri-infarct region in male rats as well as in OVX+ rats by PGZ treatment. Lo21 suggested the penumbra as an area of damaged brain tissue that continues to be viable after focal ischemia and posited that the presence of a penumbra may render therapeutic salvage theoretically possible. Although the role of p-STAT3 may be different depending on the brain area, the PGZ-induced increase of p-STAT3 in the peri-infarct region may play a crucial role for neuroprotection.
Estradiol exerted profound neuroprotective actions in a model of cerebral ischemic injury2 and its protective actions involved the neuronal antiapoptotic pathway and alterations in the expression of multiple genes in an ERα-dependent manner.22 Based on the suggestion by Dziennis et al8 that the estradiol-induced increase in p-STAT3 in the peri-infarct area is mediated by ERα, we examined whether the elevation of p-STAT3 by the PPARγ ligand is associated with ERα. In the presence of estrogen, ERα was increased in the peri-infarct area, as were p-STAT3 and PPARγ, compared with the nonischemic region and the hypoestrogenic condition. However, unexpectedly, ERα in PGZ-treated OVX+ rats was not affected despite the increase in p-STAT3 after MCAO. This phenomenon was also observed in PGZ-treated OVX− rats (data not shown). Elsewhere, we demonstrated that in rats, the activation of ERα by an angiotensin type 1 receptor blocker, olmesartan, was neuroprotective against ischemic brain damage23; olmesartan, on the other hand, did not affect PPARγ (data not shown). The activation of PPARγ by the soy phytoestrogen genistein at high doses downregulated its estrogen transcriptional activity, whereas the activation of ERα at low-dose genistein downregulated PPARγ transcriptional activity.24 Wang et al25 found that estrogen-activated ERα did not associate directly with STAT3 in multiple myeloma cells and Houston et al26 reported that the PPARγ ligand mediated the inhibition of ER-responsive gene transactivation and ER-induced protein expression. Although ER and PPAR signaling pathways may be regulated differently in different cell types, the negative crosstalk between ERα and PPARγ supports our findings that the activation of p-STAT3 by PGZ is ERα-independent.
Although we cannot rule out other mechanisms, our findings provide new insights that activation of STAT3 by PPARγ in the salvageable peri-infarct region may be essential for the prevention of neuronal cell death and PGZ appears to play a role in the transcriptional activation of neuroprotective genes by p-STAT3 through PPARγ in this area even under ER-deficient conditions. Therefore, PGZ may provide neuroprotection against cerebral ischemic damage. To assess the usefulness of PGZ for stroke prevention based on a risk–benefit assessment, additional clinical studies are required.
Sources of Funding
This study was supported in part by a grant from the Ministry of Education, Culture, Sports, Science & Technology of Japan and in part by Takeda pharmaceutical Co Ltd.
We are indebted to Professor Masaaki Uno, Department of Neurosurgery, Kawasaki Medical School, for helpful comments.
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.618926/-/DC1.
- Received March 5, 2011.
- Revision received September 6, 2011.
- Accepted September 26, 2011.
- © 2012 American Heart Association, Inc.
- Hankey GJ,
- Warlow CP
- Suzuki S,
- Brown CM,
- Dela Cruz CD,
- Yang E,
- Bridwell DA,
- Wise PM
- De Domenico I,
- Zhang TY,
- Koening CL,
- Branch RW,
- London N,
- Lo E,
- et al
- Dziennis S,
- Jia T,
- Rønnekleiv OK,
- Hurn PD,
- Alkayed NJ
- Tureyen K,
- Kapadia R,
- Bowen KK,
- Satriotomo I,
- Liang J,
- Feinstein DL,
- et al
- Wong LH,
- Saffery R,
- Anderson MA,
- Earle E,
- Quach JM,
- Stafford AJ,
- et al
- Rau SW,
- Dubal DB,
- Böttner M,
- Gerhold LM,
- Wise PM
- Shimada K,
- Kitazato KT,
- Kinouchi T,
- Yagi K,
- Tada Y,
- Satomi J,
- et al
- Dang ZC,
- Audinot V,
- Papapoulos SE,
- Boutin JA,
- Löwik CW
- Houston KD,
- Copland JA,
- Broaddus RR,
- Gottardis MM,
- Fischer SM,
- Walker CL