Inhibition of autophagy as a therapeutic strategy of iron-induced brain injury after hemorrhage

Abstract

Premenopausal women have better survival than men after intracerebral hemorrhage, which is associated with iron overproduction and autophagy induction. To examine the participation of neuronal autophagy and estrogen receptor α (ERα) in the E₂–mediated protection, PC12 neurons treated with Atg7 (autophagy-related protein 7) siRNA, rapamycin (an autophagy inducer), or Erα siRNA were applied. To study whether autophagy involves in β-estradiol 3-benzoate (E₂)-mediated neuroprotection against iron-induced striatal injury, castration and E₂ capsule implantation were performed at 2 weeks and 24 h, respectively, before ferrous citrate (FC) infusion into the caudate nucleus (CN) of Sprague Dawley male and female rats. Furthermore, the role of neuronal autophagy in the sex difference of FC-induced CN injury was confirmed by using conditional knockout Atg7 in dopamine receptor 2 (DRD2)-containing neurons in mice. The results showed that the suppression of FC-induced autophagy by E₂ was abolished by Erα siRNA preincubation. Atg7 silencing simulates and rapamycin diminishes E₂-mediated neuroprotection against FC-induced neurotoxicity. In vivo, FC induced a lower degree of autophagy, autophagic cell death, injury severity, histological lesion and behavioral deficit in female rats than in males. E₂ implantation decreased the levels of both FC-induced autophagy and injury in ovariectomized rats. Moreover, the sex difference of FC-induced CN injury was diminished in Atg7 knockout mice. Thus, suppression of autophagy by E₂ via ERα contributes to less severity of iron-induced brain injury in females than in male. This finding opens up the prospect for a therapeutic strategy targeting autophagic inhibition for patients suffering from intracerebral iron overload.

Keywords: :

• autophagic cell death
• cerebral hemorrhage
• estradiol
• iron-induced brain injury
• macroautophagy
• estrogen receptor
• Atg7

Introduction

Intracerebral hemorrhage (ICH) accounts for about 10–30% of stroke and the ICH 30-d mortality rate is higher than 50%.¹ In the brain of surviving patients, the overproduction of iron complexes because of the lysis of red blood cells leads to free radical formation, oxidative chain reactions, and thus neurotoxicity/brain injury.¹ Therefore, even after removing the blood clot, delayed onset brain injury remains a major complication in patients suffering from ICH.² Premenopausal women have better survival and functional recovery than men after ICH.^3,4 Female rodents tolerate ICH better than males with less edema, a faster recovery from behavioral deficits, and better outcome.⁵ Moreover, estradiol (E₂) has been shown to confer protection in experimental ICH⁶ and iron-induced brain injury.⁷ However, the underlying mechanism of the sex-dimorphic pathogenesis after iron overload remains unclear.

Autophagy is an essential recycling system for maintaining physiological homeostasis.⁸ However, iron-induced autophagy may contribute to brain injury after ICH in some conditions.⁹ Autophagy is enhanced by oxidative stress, and inhibited by mechanistic target of rapamycin kinase (MTOR).¹⁰ E₂ exhibits an antioxidant effect,¹¹ possibly through the activation of MTOR via ERα.¹² Accordingly, we hypothesize that E₂ protects brain from iron-induced injury in females via inhibiting iron-induced autophagy through an ERα-dependent mechanism.

To test this hypothesis, the participation of neuronal autophagy or ERα in E2-mediated protection was elucidated in vitro. Effects of E₂ on the FC-induced striatal injury and autophagy were also examined in vivo. Spectrin breakdown products (SBDPs), fragments SBDP 145/150, were used as a marker of injury severity in both in vivo and in vitro studies, because α-II spectrin is a structural protein abundant in neurons and is cleaved into fragments by proteases involved in cell death.¹³ During autophagy, a cytosolic form of microtubule-associated protein 1 light chain 3 (LC3-I) is conjugated with phosphatidylethanolamine to form lipidated LC3-II, which aggregates on the autophagosomal membrane, both LC3 aggregation and the level of LC3-II were examined as markers of autophagy. To investigate the involvement of neuronal autophagy in protection conferred by E₂, loss or gain of autophagic function was obtained in vitro by silencing of Atg7 (an essential gene for autophagosome formation) and preincubation with rapamycin (an autophagy inducer), respectively. Furthermore, mice with conditional knockout of Atg7 in DRD2-containing neurons, which are abundant in striatum,¹⁴ were used to confirm the participation of neuronal autophagy in the sex difference of FC-induced injury.

Results

Role of autophagy modulation in the E₂-mediated protective effect on FC-induced neurotoxicity in PC12 cells

Rat PC12 cells were differentiated into neurons by NGF induction in vitro to examine whether autophagy participates in the E₂-mediated protection against the neurotoxicity induced by FC. Two days after NGF treatment, the cells extended neuron-like process. After FC exposure, part of the cells underwent round up, while pretreatment of E₂ significantly decreased the number of cells with round-up morphology. The results showed that FC exposure increased levels of LC3-II by 48 ± 6.8% (p < 0.05) and SBDP 145/150 kD by 97 ± 13% (p < 0.01), and preincubation with E₂ (10 nmol/l) decreased the FC-induced autophagy (Fig. 1A) by 45% ± 5.8% (p < 0.01), cleavage of α-II spectrin (Fig. 1B) by 53 ± 7.2% (p < 0.01) and cell death (Fig. 1D) by 40 ± 2.1% (p < 0.01). Moreover, FC induced an increase in the number of TUNEL(+) BECN1 immunoreactive cells reflecting autophagic cell death, whereas E₂ preincubation decreased the incidence of FC-induced autophagic cell death (Fig. 1C).

Figure 1. E₂ pretreatment decreases FC-induced autophagy, injury, autophagic cell death and neurotoxicity in PC12 cells. (A) Autophagy was induced after FC exposure and E₂ significantly decreased FC-induced autophagy. (B) FC induced the neuronal injury and E₂ significantly decreased the FC-induced neuronal injury. (C) E₂ significantly decreased FC-induced autophagic cell death. Cells were stained by TUNEL and BECN1 for DNA damage (green) and autophagy (red), respectively. Nuclei were stained by DAPI. The TUNEL(+) Beclin 1 immunoreactive cells with an intact nucleus in PC12 cells pretreated with or without E₂ before FC exposure depict an index of autophagic cell death as indicated by arrows in the merged view. (D) FC increased the number of dead cells and E₂ significantly increased the number of living cells after FC exposure. Data are expressed as means ± SD (n = 6), *p < 0.05, p** < 0.01.

Atg7 siRNA was used to investigate whether the suppression of autophagy by E₂ might mediate the protection against FC-induced neurotoxicity conferred by E_2. As shown in Figure 2A, the degree of FC-induced LC3 lipidation was decreased by 42 ± 2.3% (p < 0.05), at 48 h after silencing Atg7. Pretreatment of Atg7 siRNA per se simulated the protective effect of E₂ on FC-induced neurotoxicity (Fig. 2B). Pretreatment with the MTOR inhibitor rapamycin, which enhances autophagy, at a concentration of 25 nmol/l for 29 h increased the basal level of LC3-II (Fig. 2C) by 200 ± 2.4% (p < 0.01) and totally eliminated the protective effect of E₂ on FC-induced cytotoxicity (Fig. 2D).

Figure 2. Involvement of autophagy in E₂-mediated neuroprotection against FC-induced neurotoxicity in PC12 cells. (A) Atg7 silencing decreased the levels of both constitutive and FC-induced autophagy. The differentiated PC12 cells were pretreated with 100 nM Atg7 siRNA or nontargeted siRNA for 24 h followed by E₂ treatment for another 24 h. Subsequently, cells were exposed to FC. (B) Atg7 silencing simulated the protective effect of E₂ against the FC-induced cytotoxicity. (C) Preincubation with rapamycin significantly increased the levels of autophagy in each group except for FC-exposure group per se. The differentiated PC12 cells were pretreated with E₂ followed by FC exposure. Five hours before FC exposure, the cells were preincubated with 25 nM rapamycin. (D) Preincubation with rapamycin significantly diminished the protective effect of E₂ against FC-induced neurotoxicity. Data are expressed as means ± SD (n = 6), *p < 0.05, **p < 0.01.

Knockdown of Erα in E₂-mediated suppression of FC-induced autophagy in PC12 cells

To examine whether ERα is essential for suppression of FC-induced autophagy, the level of LC3-II was examined after knockdown Erα by siRNA. As shown in Figure 3A, the ERα protein level in the Erα siRNA-treated group was lower than that in the nontargeted siRNA-treated group. E₂ suppressed the FC-induced autophagy in nontargeted siRNA-treated group by 53 ± 8.2% (p < 0.01). The suppression of FC-induced LC3-II by E₂ was diminished by Erα siRNA pretreatment (Fig. 3B). The LC3 immunocytochemistry showed a similar trend to the western blot analysis (Fig. 3C). In addition, the protective effect of E₂ on FC-induced cytotoxicity was also diminished (p < 0.01) by Erα siRNA (Fig. 3D).

Figure 3. Involvement of ERα in the E₂-mediated suppression of FC-induced autophagy in PC12 cells. (A) Erα RNA silencing decreased the levels of ERα. The differentiated PC12 cells were pre-treated with Erα siRNA (50 nM) for 24 h, followed by E₂ treatment for an additional 24 h before FC exposure. (B) Erα RNA silencing diminished the suppression of FC-induced LC3 lipidation by E₂. (C) Erα RNA silencing diminished the suppression of LC3 immunoreactivity by E₂. In the nontargeted siRNA treated group, E₂ suppressed the FC-induced LC3 immunoreactivity. (D) Erα silencing diminished the protective effect of E₂ against FC-induced neurotoxicity. Data are expressed as means ± SD (n = 6), *p < 0.05, **p < 0.01.

Sex difference in FC-induced autophagy and autophagic cell death in the rat striatum

To examine the sex difference in FC-induced autophagy, brain injury, behavioral deficit, the levels of LC3 aggregation and LC3-II protein abundance in the CN of both male and female rats were compared. Autophagic cell death was identified by TUNEL(+) BECN1 immunoreactive cells with intact (round) nuclei (Fig. 4B). The results demonstrated that the levels of FC-induced LC3 aggregation, TUNEL(+) BECN1 immunoreactive cells, and lipidated LC3-II were lower in the CN of female rats than they were in males (Fig. 4A–C). In addition, the level of α-II spectrin breakdown products (SBDP 145/150) was lower in female rats than it was in males after FC infusion (Fig. 4D). Moreover, both the FC-induced lesion ratio (Fig. 4E) and forelimb use asymmetry ratio (Fig. 4F) were lower in female rats than it were in males.

Figure 4. Sex differences in FC-induced autophagic signaling in the rat striatum. (A) FC increases more LC3 aggregation around the neuronal nucleus in male than in females. LC3 antibody followed by a secondary antibody conjugated with FITC and NeuN were used to stain LC3 aggregation (green) and neuronal nuclei (red), respectively. Nuclei were stained by DAPI. Arrows indicate LC3 aggregation around the NeuN positive nucleus. (B) FC-induces more autophagic cell death in male than in female rats. Rat striatum was stained using TUNEL and BECN1 for DNA damage (green) and autophagy (red), respectively. The merged views show round nuclei (yellow) depicting TUNEL(+) BECN1 immunoreactive cells with an intact nucleus, which is referred to as autophagic cell death and is indicated by arrows. (C) The basal level and level of FC-induced autophagy were lower in the CN of female than in male rats. (D) The level of FC-induced cleavage of spectrin was lower in the CN of female rats than it was in males. The levels of SBDP 145/150 in the CN of male and female rats with or without infusion of FC are shown as a ratio of SBDP 145/150/spectrin acting as an index of severity of injury. (E) The lesion ratio of FC-induced brain injury was lower in the CN of female rats than it was in males. The hemispheric area of the CN was quantified according to the density of the hematoxylin and eosin-stained tissue section by Image-pro plus software. The lesion ratio was estimated by dividing the hemispheric volume of the CN on the ipsilateral side by that on the contralateral side. (F) FC induces more severe behavioral deficit in male than in female rats. Forelimb use asymmetry ratio depicting an index of behavioral deficit. Data are expressed as means ± SD (n = 6), *p < 0.05, **p < 0.01.

Role of E₂ in the sex difference in FC-induced autophagy and injury in the rat striatum

Our previous result showed that E₂ plays a critical role in mediating the sex dimorphism in FC-induced neurological deficit and brain lesion examined by histology using fixed samples.⁷ In the present study, to get better correlation between level of autophagy and injury severity, the level of lipidated LC3 and the ratio of SBDP 145/150/spectrin (instead of lesion ratio) were analyzed simultaneously using the same tissue samples in each groups. As shown in Figure 5A, intrastriatal infusion of FC induced a 33 ± 3.2% (p < 0.05) higher level of LC3-II in the CN of male rats than in females. The result of striatal injury showed that the level of SBDP 145/150 in FC-infused female rats was 40 ± 11.5% (p < 0.01) lower than that in males. Subcutaneous implantation of E₂ 24 h prior to infusion of FC decreased the levels of the FC-induced autophagy in castrated females by 39 ± 9.6% (p < 0.05) but only slightly decreased it in castrated males. Furthermore, E₂ implantation decreased by 46 ± 6.4% (p < 0.05) and 42.9 ± 8.3% (p < 0.05) the FC-induced cleavage of α-II spectrin in female and male castrated rats in vivo, respectively (Fig. 5B).

Figure 5. Effect of E₂ on FC-induced autophagy and injury in the CN of castrated rats. (A) The level of FC-induced autophagy was lower in the CN of ovariectomized females (♀ with a slash) than it was in castrated males (♂ with a slash), and E₂ implantation significantly decreased the levels of FC-induced autophagy in ovariectomized females. (B) The severity of FC-induced injury was lower in the CN of female rats than it was in males and E₂ implantation significantly decreased the ratios of SBDP 145/150/spectrin in both female and male castrated rats. Data are expressed as means ± SD (n = 6), *p < 0.05, **p < 0.01.

Effect of Atg7 conditional knockout in DRD2 neurons on the sex difference in FC-induced injury in the CN

To confirm the participation of autophagy in the sex difference in FC-induced injury, the behavior deficit and lesion ratio in CN of Atg7^+/–Drd2-Cre mice (heterozygote conditional knockout Atg7 in DRD2 neurons) were compared with those in wild-type Atg7^F/F mice. The Atg7 knockout efficiency in DRD2 neurons was checked after immunohistochemical staining of both ATG7 and LC3 in DRD2 neurons. As shown in Figure 6A, in Atg7 ^F/F mice, ATG7 immunoreactivity was observed in DRD2 neurons in CN, while, in CN of Atg7^+/–Drd2-Cre mice, ATG7 immunoreactivity in DRD2 neurons was diminished. The level of LC3 aggregation was higher in the CN of Atg7^F/F male mice than in females and conditional knockout of Atg7 significantly decreased the level of LC3 aggregation in both male and female mice (Fig. 6B). Moreover, both the behavioral deficit and lesion ratio showed a significant sex difference in wild-type Atg7^F/F mice, whereas no sex difference in either behavioral deficit or lesion ratio was observed in Atg7^+/–Drd2-Cre mice. Moreover, in the male group, both the FC-induced behavioral deficit and lesion ratio in the CN of Atg7^+/–Drd2-Cre mice were significantly lower than those in Atg7^F/F mice (Fig. 6C and D, respectively).

Figure 6.Atg7 conditional knockout in DRD2 neurons abolished the sex difference in FC-induced CN injury. Knockout of Atg7 decreased the ATG7 immunoreactivity (A) and LC3 aggregation (B) in DRD2 neurons in both male and female mice. The brain tissues containing the CN from male or female mice with (Atg7^+/–D rd2-Cre) or without (Atg7^F/F) Atg7 conditional knockout were sampled 2 d after FC infusion and sectioned at 10 μm thickness. ATG7 antibody or LC3 antibody followed by a secondary antibody conjugated to FITC and DRD2 antibody followed by a second antibody conjugated to rhodamine were used to stain for ATG7 or LC3 (green) and DRD2 neurons (red), respectively. Simultaneously, nuclei were stained with DAPI. The nuclei surrounded by cytosolic ATG7 or LC3 in DRD2 neuron depict ATG7-containing or autophagic DRD2 neurons as indicated using arrows. (C) Knockout of Atg7 significantly decreased the FC-induced behavioral deficit in male mice and diminished the sex difference in FC-induced behavioral deficit. Forelimb use asymmetry ratio depicting an index of behavioral deficit. (D) Knockout of Atg7 significantly decreased the level of FC-induced striatal injury in male mice and diminished the sex difference in FC-induced striatal lesion ratio. Data are expressed as means ± SD (n = 6). *p < 0.05, **p < 0.01 compared with the sham control sex-matched group. ^##p < 0.01 compared with the male sham.

Discussion

The present study yielded several interesting findings. First, the protective effect of E₂ on FC-induced neurotoxicity in PC12 cells was simulated by silencing of Atg7, and was diminished by rapamycin preincubation. Second, knockdown of Erα diminished the suppression of FC-induced autophagy by E₂. Third, levels of autophagy and injury in the CN were lower in female rats than in males after FC infusion. Fourth, E₂ replacement significantly decreased levels of FC-induced autophagy in ovariectomized rats. Finally, conditional knockout of Atg7 in DRD2 neurons diminished the sex difference of both FC-induced lesion ratio and behavioral deficit in the CN. These results suggest that suppression of FC-induced autophagy through an ERα dependent pathway might participate in the E₂-mediated neuroprotection on FC-induced brain injury in female rats. To our knowledge, this is the first demonstration that autophagy participates in the E₂-mediated neuroprotection against iron-induced brain injury. This finding opens up the prospect for a therapeutic strategy targeting autophagic inhibition for patients suffering from ICH or neurodegeneration caused by iron overload.

Autophagy is essential for maintaining homeostasis and cell survival. However, autophagic cell death has been recognized in conditions of oxidative stress.¹⁵ A previous report indicated that treatment with rapamycin potentiated oxidative stress-induced cell death.¹⁶ Another report indicated that inhibition of autophagy switched the mode of the cell death from apoptosis to necrosis in neonatal hypoxia-ischemia-induced brain injury.¹⁷ Therefore, whether autophagy plays a beneficial or harmful role in different kinds of brain injury remains controversial. A recent report indicated that iron-induced autophagy may contribute to the brain injury after ICH.⁹ However, no direct evidence has been provided yet. The present study demonstrated that most of the TUNEL(+) nuclei in PC12 neurons were surrounded by a high level of cytosolic BECN1 immunoreactivity after FC exposure (Fig. 1C). Preincubation with rapamycin enhanced the neurotoxicity of FC (Fig. 2C). The in vivo result of FC-induced CN injury also showed that most of the nuclei of the TUNEL(+) cells expressing BECN1 were round (Fig. 4B), which is observed in autophagic cell death.¹⁸ In addition, the higher level of FC-induced autophagy (Fig. 4C) is associated with greater severity of brain injury in male rats than in females (Fig. 4D–F). Moreover, in vivo conditional knockout of Atg7 in DRD2 neurons significantly decreased the FC-induced striatal injury in males (Fig. 6C and D). Taken together, these data suggest that autophagic cell death may contribute to the brain injury caused by iron overload, particularly in males.

“Stroke has been recognized as a sexually dimorphic disease.¹⁹ Premenopausal women have a better survival rate than men after ICH.^3,4 Animal study showed that brain edema was significantly severe in female as compared with that in male rats.²⁰ Hormones (e.g., E₂) and sex have been shown to affect outcome after ICH in various animal models of ICH.^6,20,21 In the FC-infused animal model, the severity of the brain injury in males was higher than that in female rats and E₂ treatment reduced the severity of brain edema in male rats²¹ and decreased the lesion ratio of brain injury and the behavioral deficit in both male and female castrated rats.⁷ These results imply that there are sex differences in the mechanisms whereby neurons respond to injury and E₂ may play a role in sex dimorphism of brain injury after ICH. However, previous in vitro studies have shown that neurons derived from male rat are more sensitive to excitotoxicity than neurons derived from female rat.²² These in vitro studies implicate that there is an innate sexual dimorphism independent of circulating sex steroids. The present in vitro study, using PC12 cells but not neurons from male and female rats, showed that E₂ significantly decreased the ratio of SBDP 145/150/spectrin (Fig. 1B) and increased the number of living cells after FC exposure (Fig. 1D). The in vivo study showed that the severity of FC-induced injury was lower in the CN of female rats than males and E₂ implantation significantly decreased the ratio of SBDP 145/150/spectrin in castrated male to the level of castrated female rats (Fig. 5B). Although the innate sex dimorphism can not be excluded based on the current results, E₂ may play a critical role in the sex dimorphism of FC-induced brain injury.”

Sex difference in neuronal autophagy has been first reported by Du L et al., who demonstrated that neurons from males more readily undergo autophagy and die during starvation, whereas neurons from females survive longer.²³ The result of the present in vitro study showed that E₂ decreases the level of FC-induced autophagy. In addition, the protection of E₂ against FC-induced neurotoxicity is simulated by Atg7 siRNA preincubation and is eliminated by rapamycin preincubation (Fig. 2B and D). The present results further show that the level of FC-induced autophagy is lower in female rats than in males and E₂ replacement significantly suppresses the FC-induced autophagy in ovariectomized females (Fig. 5B) suggesting that the suppression of FC-induced autophagy by E₂ contributes to the less severity of brain injury in females. These data further imply that autophagy might be another factor involved in the sexual dimorphism of FC-induced brain injury. Greater therapeutic effort should be emphasized in male patients to suppress the autophagic cell death caused by iron overload.

It is well known that E₂ exerts protective actions against various forms of neuronal injury including iron-induced brain injury in rats.⁷ Multiple pathways may transmit the neuroprotective effect of E_2.²⁴ However, the mechanism underlying how E₂ modulates autophagy remains unknown. Previous findings have shown that levels of both mRNA and protein of ERα, but not ERβ, in the CN are higher in female rats than in males.⁷ The present result showed that E₂ implantation significantly decreased the ratio of SBDP 145/150/spectrin in both male and female group (Fig. 5B) implicating an ERα-independent protective mechanism of E₂ in males. While E₂ replacement significantly inhibited the FC-induced autophagy in females, but not in males (Fig. 5A) suggesting a sex dimorphic effect of E₂ on autophagy. Furthermore, knockdown of ERα diminished the E₂-mediated suppression of FC-induced autophagy in PC12 cells (Fig. 3A–C). These results support the hypothesis that E₂ suppresses FC-induced autophagy via an ERα-dependent pathway_. However, further investigation is needed to explore how E₂/ERα suppresses FC-induced autophagy.

In conclusion, suppression of FC-induced autophagy by E₂ via ERα contributes to the less brain injury caused by iron overload in female rats than in males. Our results provide important information for developing sex-specific therapeutic strategies to prevent brain dysfunction caused by hemorrhage and/or by neurodegenerative disease associated with iron overload.

Materials and Methods

PC12 cells

PC12 cells (4 × 10⁵/well) were plated in a 6-well dish pre-coated with 0.1 mg/ml poly-d-lysine (Sigma-Aldrich, P1024) and grown in phenol red free RPMI medium (Gibco, 11835030) containing 5% heat-inactivated fetal bovine serum for 3 d. The active subunit of NGF (2.5S) (4 nM) (Invitrogen, 13257-019) was added to the cultured medium for two days to differentiate PC12 cells into neurons. The culture medium was then replaced by the medium containing 10 nmol/l E₂ (Sigma-Aldrich, E2758). After E₂ pretreatment for 24 h, cells were treated with FC (2.4 nmol/μl; Sigma-Aldrich) for 24 h before harvested.

Immunostaining and TUNEL staining

The paraffin-embedded brain tissues were cut into 10 μm serial sections. PC12 cells were harvested, fixed with 4% paraformaldehyde, and adhered on slides by cytocentrifugation. Rabbit anti-LC3B antibody (1:50; Sigma-Aldrich, L7543) and mouse anti-BECN1 antibody (1:50; BD Biosciences, 612112) were used for detections of LC3 aggregation and BECN1 immunoreactivity, respectively. NeuN antibody (1:100; Millipore Corporation, MAB377) was used to detect neuron. DAPI (Sigma-Aldrich) and TUNEL solution containing FITC-dUTP (In Situ Cell Death Detection Kit, Roche Diagnostics, 12 156 792 001) were used for nucleus labeling and detection of DNA fragmentation, respectively. Autophagic cell death was identified by the TUNEL(+) BECN1 immunoreactive positive staining cells with intact (round) nuclei.²⁵ Mouse monoclonal ERα antibody (1: 25, Leica Biosystems, 6008295) was used to detect ERα. FITC-conjugated AffiniPure goat anti-rabbit IgG (1:200, Jackson ImmunoResearch Laboratories, 111-095-003) and Rhodamine-conjugated AffiniPure goat anti-mouse IgG (1:50, Jackson ImmunoResearch Laboratories, 115-025-003) were used to recognize the primary antibodies. Images were acquired using a fluorescence microscope or an Olympus FV1000 confocal laser scanning microscope (Olympus, Fluoview FV1000).

Western blot analysis

An equal amount of protein from each sample was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto a PVDF membrane, incubated with antibody against LC3 antibody (1:1000), α-II spectrin antibody (1:1000, Santa Cruz Laboratories, sc-7465), ERα (1: 25) or ATG7 antibody (1:1000, Biosensis, R-161-100) followed by a second antibody, and then visualized by ECL chemiluminescence (PerkinElmer, NEL 105). The amount of LC3-II (lipidated LC3) normalized by β-actin depeicts the level of autophagy.²⁶ The ratio of SBDP 145/150/spectrin acts as an index of severity of injury.

Atg7 and Erα RNA interference

The mixture of rat Atg7 siRNA contained four sequences: 3′-CAAAGUUAACAGUCGGUGUUU-5′, 5′-PACACCGACUGUUAACUUUGUU-3′; 3′-AGUGAAUGCCAGCGGGUUCUU-5′, 5′-PGAACCCGUGGCAUUCACUUU-3′; 3′-GUGGAGGAACUCAUCGAUAUU-5′, 5′-PUAUCGAUGAGUUCCUCCAGUU-3′; and 3′-CCCAGAAGAAGUUGAAGGAUU-5′, 5′-PUCGUUCAACUUCUUCUGGGUU-3′ (Dharmacon, NM_001012097). The sequences of Erα siRNA were 3′-AGUUGAUCAUAUGAACCAGCU-5′, 5′-CUGGUUCAUAUGAUCAACUGG-3′. A 5 μl (20 μmol/l) Atg7 siRNA or a 5 μl (10 μmol/l) Erα siRNA was mixed with 5 μl transfection reagent for 15 min, and then diluted with 990 μl OPTI-MEM medium. Dharmacon siCONTROL non-targeted siRNA (D-001210-01-05) was used as a negative control.

Enhancement of autophagy by rapamycin preincubation

To enhance autophagic machinery, 5 h before FC exposure, the PC12 cells were preincubated with the equivalent of 25 nM rapamycin (LC Laboratories, R-5000), which was encapsulated in a polycaprolactone-grafted chondroitin sulfate amphiphilic polymer by dialysis to improve its water solubility.

Animals

Eighty-four 12-week-old male and female Sprague Dawley rats (350–420 g; LASC; Charles River Technology) were used. No animals were excluded after random grouping. Experimental procedures were approved by the Kaohsiung Medical University Committee for the Use of Experimental Animals. Castration was performed under anesthesia with pentobarbital (35 mg/kg body weight, intraperitoneally) 2 weeks before E₂ implantation. A Silastic tube (2 mm inner diameter, 30 mm in length) containing 0.38 mmol/l E₂ (Sigma-Aldrich, E8515) was implanted subcutaneously 24 h before FC infusion. Ten microliters of FC (1 nmol/μl) was infused into the right CN (coordinates: 0.2 mm anterior, 3.5 mm lateral, and 5.5 mm ventral to bregma) at a rate of 10 μl/min using a microinfusion pump (VC50; CMA Microdialysis). Ten microliters of saline was infused in the sham control.

Forelimb use asymmetry test

Each individual rat was placed in a transparent cylinder (25 cm in diameter and 30 cm in height) in the dark, and the use of ipsilateral limbs (I), contralateral limbs (C), or simultaneous use of both forelimbs (B) was observed for a 5 min period. The test was randomized, blinded, and repeated twice in each rat. The forelimb use asymmetry ratio was calculated using the following equation: [I/(I + C + B)] − [C/(I + C + B)].²⁷

Histological examination of the brain injury

After hematoxylin and eosin (H&E) staining, the brain lesion in every fifteenth section of the CN was analyzed using Image-pro Plus software (Universal Imaging Corp.) to determine the staining intensity of the CN area. The total hemispheric volume was integrated from the hemispheric areas of 20 sections. The lesion ratio was estimated by dividing the hemispheric volume of the CN on the ipsilateral side by that on the contralateral side.

Conditional knockout of Atg7 in DRD2 neurons

The Atg7-conditional knockout mice (Atg7 flox/flox; Atg7^F/F) (a gift from Dr. Masaaki Komatsu, Laboratory of Frontier Science, Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan.) were cross-bred with transgenic mice expressing Cre recombinase under the control of a Drd2 promoter (Drd2-Cre) (purchased from Mutant Mouse Regional Resource Centers) to produce mice deficient for Atg7 specifically in the DRD2-containing neurons, which is abundant in CN. The F1 heterozygote littermates (Atg7^+/–Drd2-Cre) were used in this study. Atg7 ^F/F male and female mice were used as control groups. Three microliters of FC (1 nmol/μl) was infused into the right CN (coordinates: 0.2 mm anterior, 2.5 mm lateral, and 3.5 mm ventral to bregma) at a rate of 2 μl/min. Two days after FC infusion, the brain tissue containing CN was sampled for the evaluation of histological lesion after behavioral functional test. The knockout efficiency of Atg7 in DRD2 neurons was checked by immunohistochemical staining.

Statistical analysis

FC-induced injury was compared between brains from males and females using a two-way ANOVA followed by a Scheffé post hoc test. Data for the effects of E₂, Atg7 siRNA, rapamycin, or Erα siRNA on FC-cytotoxicity or autophagy were analyzed using a multi-way ANOVA to determine the effect of each factor and the interaction between two factors. Significance was accepted at p < 0.05.

Abbreviations:
Atg7	=	gene of autophagy-related protein 7
BECN1	=	Beclin 1
CN	=	caudate nucleus
DRD2	=	dopamine receptor 2
ERα	=	estrogen receptor α
FC	=	ferrous citrate
ICH	=	intracerebral hemorrhage
LC3	=	microtubule-associated protein 1 light chain 3
SBDP 145/150	=	spectrin breakdown products with molecular weight of 145 or 150

Acknowledgment

This work was supported by research grants from the National Science Council of Taiwan (99-2320-B-037-024-MY3 and 99-2314-B-037-025 to C.H.; 99-2314-B-037-025 to C.-L.L.).

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.