HSPB8 overexpression prevents disruption of blood-brain barrier after intracerebral hemorrhage in rats through Akt/GSK3β/β-catenin signaling pathway

Blood brain barrier (BBB) disruption is a crucial factor contributing to secondary brain injury after intracerebral hemorrhage (ICH). Heat shock protein B8 (HSPB8) has been recently reported to confer neuroprotection against against ischaemic stroke through maintaining BBB integrity. However, the role of HSPB8 in ICH is still elusive. In this study, we found that HSPB8 was upregulated by ICH and extensively expressed in neurovascular structure including endothelial cells and astrocytes. lentivirus intracerebroventricular (i.c.v) injection achieved a widespread and persistent HSPB8 overexpression in brain tissues. HSPB8 overexpression significantly ameliorated neurobehavioral deficits and brain edema at 24 and 72h following ICH. Moreover, HSPB8 overexpression remarkedly inhibited BBB disruption and significantly increase the level of p-Akt, p-GSKβ and intranuclear β-catenin 24h post-ICH. This effect was obviously reversed by Akt specific inhibitor, MK2206. Based on these findings, HSPB8 exerted its protective effect on BBB, at least partly, via Akt/ p-GSKβ/β-catenin pathways.


INTRODUCTION
Spontaneous intracerebral hemorrhage (ICH), caused by the rupture of blood vessels in the brain, is the most devastating type of stroke [1,2]. It is becoming increasingly clear that BBB disruption is one of the critical mechanisms of ICH-induced brain injury. Therefore, search for effective neuroprotective agents ameliorating BBB collapse may provide therapeutic potential for ICH.
As a member of small heat shock proteins (sHSPs) family, HSPB8 is of particular interest in the field of neurological diseases. HSPB8 has been demonstrated to interfere with detrimental processes during neurodegenerative diseases [3,4]. Our previous investigation has reported that HSPB8 effectively conferred protection against cerebral I/R injury by improving BBB stability [5]. However, we do not yet know the possible role of HSPB8 in ICH.
Accumulating evidence has confirmed that the Wingless integrated MMTV (Wnt)/β-catenin is a major pathway important for the stabilization of brain endothelial tight junctions [6,7]. Mounting evidence indicates that Wnt/βcatenin signaling is negatively regulated by the activity of GSK3β [8,9]. In addition, Akt results in GSK3β inhibition through inducing GSK3β phosphorylation at Ser9 [10]. Therefore, Akt/GSK3β/β-catenin signaling may play a positive role in BBB protection. Previous studies have shown that HSPB8 activated PI3k/Akt in different disease models, including brain ischemia, chronic myocardial ischemia and hepatic carcinoma [11][12][13]. However, whether HSPB8 protects against BBB disruption after ICH via regulating Akt/GSK3β /βcatenin signaling pathway is largely unknown. The aim AGING of the present study was to explore the potential role of HSPB8 in BBB maintenance in ICH rat models. We also investigated the possible mechanisms underlying HSPB8 protective effects.

Increased HSPB8 protein expression in rats after ICH
We first measured the protein expression of HSPB8 in perihematomal brain tissue after ICH at different time points. In comparison to the sham group, the western blot results showed that HSPB8 protein level in the perihematomal tissues was significantly increased in the rats suffered from ICH injury from 6h post-ICH and peaked at 24h after ICH ( Figure 1A, 1B). Moreover, double immunofluorescence staining showed that HSPB8 was sparsely expressed and co-localized with CD31 (endothelial), and GFAP (astrocytes)in sham group. HSPB8 immunoreactivity on neurovascular structures was increased around the hematoma 24 h after ICH ( Figure 1C, 1D).
Effect of intracerebroventricular (i.c.v) injection of pLV-HSPB8 on the expression of HSPB8 in brain A lentiviral delivery of HSPB8 gene into cerebroventricular was performed to investigate the role of HSPB8 in ICH-induced brain injury. Two weeks following i.c.v intraventricular injection of pLV-HSPB8 lentivirus, widespread GFP fluorescence was observed throughout the CNS, including the cortex, hippocampus and striatum (Figure 2A). The data demonstrated that widespread delivery of the transgene was achieved using pLV vectors. To estimate level of HSPB8 in brain, we performed mRNA and western-blot. The results revealed a dramatic increase in HSPB8 gene and protein expression in the brain tissues of the experimental ICH+pLV-HSPB8 group compared with ICH+pLV-GFP group ( Figure 2B, 2C). Furthermore, brain sections of animals treated with ICH+pLV-GFP showed low immunopositive HSPB8-elements, while in ICH+pLV-HSPB8 treated group, HSPB8 immunostaining was strongly expressed throughout the brain ( Figure 2D). This implied that lentivirus achieved widespread infection within brain tissues with high AGING efficiency and allowed a widespread and persistent HSPB8 overexpression.

HSPB8 overexpression improved behavioral deficits and restrained cerebral edema at 24 and 72 h after ICH
Next, we investigated the effects of HSPB8 overexpression on the ICH-induced brain injury. ICH rats exhibited significant neurobehavioral deficits at modified Garcia test, performances of the forelimb placement and corner turn test at 24h after surgery when compared with sham animals. Lentivirus induced-HSPB8 overexpression markedly improved these neurobehavioral functions compared to ICH+ pLV-GFP group at 24h after ICH ( Figure 3A-3C). At 24h after ICH, a significant increase in brain water content was observed in the ipsilateral cortex and basal ganglia in ICH+ pLV-GFP group compared with the sham group. There were no differences in brain water content observed between all groups in contralateral cortex, contralateral basal ganglia or in cerebellum. HSPB8 overexpression significantly reduced the water content in the ipsilateral brain basal ganglia and cortex at 24h post-ICH compared with the vehicle control ( Figure 3D, 3E). The performance of ICH+ pLV-HSPB8 group on Modified Garcia, Forelimb placement, and Corner turn test was obviously improved at 72h after ICH, compared with ICH+ pLV-GFP group. Consistently, HSPB8 overexpression reduced brain water content in the ipsilateral basal ganglia significantly.
HSPB8 overexpression rescued the blood-brain barrier impairment after ICH BBB functionality at 24h following ICH was evaluated by Evans Blue (EB) extravasation. As shown in Figure  4A, ICH caused a significant increased dye extravasation in ipsilateral hemisphere compared with the sham group. HSPB8 overexpression markedly attenuated the EB leakage compared with ICH+ pLV-GFP group (p < 0.05). We observed the ultrastructure of BBB in the perihematomal brain tissues by TEM. In normal condition, basal laminas were found to be continuous and integrate; the endothelial tight junctions (TJs) appeared as electron-dense linear structures. At AGING 24h after ICH, the TJs of endothelial cells were seriously collapsed, which appeared as reduced electron density and endothelium detachment. HSPB8 overexpression notably reversed these damages ( Figure 4B).

HSPB8 increased the TJPs expression after ICH
The expression of occludin and claudin-5 was assessed by western blot and immunofluorescence staining. We observed that the level of occludin and claudin-5 was significantly decreased at 24h after ICH in comparison with the sham group. HSPB8 rescued the TJPs loss in perihematomal area ( Figure 5A-5C). Consistently, immunofluorescence staining revealed that ICH induced occludin and claudin-5 junctional discontinuity compared with the control group. HSPB8 overexpression attenuated ICH-induced degradation of occludin and claudin-5 ( Figure 5D-5E).

HSPB8 overexpression alleviated hemorrhagic inflammatory injury 24h post-ICH
BBB compromise has been regarded as a key event in ICH-induced SBI through an orchestrated series of events to facilitate passage of the leukocyte into brain tissue, which subsequently exacerbates neuroinflammatory response [17]. Those inflammatory processes determine the severity and prognosis of ICH. Therefore, we then examined the effect of HSPB8 overexpression on the leukocyte infiltration at 24h after ICH. As depicted in Figure 6A, immunofluorescence staining of 24h after ICH revealed that significant presence of MPO+ neutrophils in the perihematomal tissue, which markedly attenuated by HSPB8 overexpression. To evaluate the neuroinflammatory response, proinflammatory cytokines, including tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), in perihematomal tissue were quantified by ELISA. As indicated in Figure 6B, 6C, HSPB8 suppressed ICHinduced proinflammatory cytokines secretion. Consistently, relative mRNA transcript levels of inflammatory cytokines in perihematomal regions increased markedly 24h after ICH, compared with sham group. HSPB8 overexpression attenuated the elevated mRNA level of TNF-α and IL-1β after ICH ( Figure 6D, 6E). These results indicated that HSPB8 suppressed the inflammatory responses following ICH.

HSPB8 activated the Akt/GSK3β/β-catenin pathway following ICH
To determine Akt activation, the phosphorylation of Akt at S473 was measured. In the in vivo ICH model, the level of total Akt did not change, whereas the p-Akt (S473) in the perihematomal area was decreased at 24h after ICH compared with sham group. HSPB8 overexpression reversed ICH-induced reduction of p-Akt ( Figure 7A, 7C).

Figure 3. HSPB8 overexpression significantly reduced brain injury and brain edema at 24 and 72h after ICH. (A-C) Modified
Garcia test, corner turn test, and limb placement test at 24h and 72h post-ICH in sham, ICH+ pLV-GFP or ICH+pLV-HSPB8 groups (n=6, mean±SEM). (D) The brain water content at 24h following surgery in sham, vehicle, and pLV-HSPB8 treatment groups. (E) The brain water content at 72h following surgery in different groups. Brain sections were divided into five parts: ipsilateral basal ganglia, contralateral basal ganglia, ipsilateral cortex, contralateral cortex, and cerebellum (n=6, mean±SEM). **P <0 .01 versus indicated group; *P <0.05 versus indicated group. n=number of animals per group.
AGING Likewise, levels of p-GSK3β(ser9) were significantly higher in HSPB8-treated group than in the vehicletreated group at 24h after ICH ( Figure 7A, 7D). Moreover, compared with sham group, plasm-protein β-catenin was increased and nucleoprotein β-catenin was decreased in the ICH+ pLV-GFP group, while HSPB8 overexpression significantly increased nuclear accumulation of β-catenin and led to a corresponding decrease of β-catenin in the cytoplasm ( Figure 8A-8D).

DISCUSSION
The aim of the first experiment was to examine the time course and spatial expression in perihematomal area after ICH. In our study, HSPB8 expression was significantly upregulated and peaked at 24h after ICH. Double immunofluorescence staining revealed that HSPB8 immunoreactivity was found sparsely in perihematoma GFAP-positive astrocytes and CD31positive endothelial cells. ICH significantly increased expression of HSPB8 in these BBB elements. HSPB8 co-localized with endothelial cells and astrocytes and was upregulated after ICH, indicating that HSPB8 may play a significant role in regulating BBB permeability after ICH.
To investigate the role of HSPB8 after the induction of ICH, we successfully overexpressed HSPB8 in rat local brain tissues with lentivirus system through i.c.v injection. pLV-HSPB8 administration significantly improved neurobehavioral functions and reduced brain edema after ICH.
Composed of highly specialized brain endothelial cells (ECs), BBB maintains the hemostasis of neuronal microenvironment [15]. BBB disruption is a crucial initiating contributor to ICH-induced SBI [16]. In the present study, therefore, the mechanism underlying the effect of HSPB8 after ICH was focused on BBB protection. Lentivirus i.c.v injection induced HSPB8 overexpression attenuated EB extravasation induced by ICH. In accordance with this result, TEM revealed that HSPB8 overexpression reversed the discontinuity of TJs to some extent. TJs consist of several transmembrane proteins and membraneassociated cytoplasmic proteins [17,18]. Mounting evidence support that loss of TJPs results in BBB hyperpermeability and subsequent SBI after ICH [19]. In our study, HSPB8 overexpression obviously protected against reduction and disorganization of TJPs, including claudin-5 and occludin. TJ disorganization leads to leukocyte infiltration and thus triggers inflammation [20]. Inflammation cascade ultimately promotes neuronal cell death and leads to neurological deterioration. MPO, an indicator of AGING neutrophil infiltration [21], was employed in the present study. In our study, HSPB8 overexpression mitigated neutrophil infiltration as measured by immunofluorescence, and downregulated the expression of TNF-α, IL-1β, measured by ELISA and western-blot at 24h after ICH. Thus, our study provided evidence that HSPB8 efficiently preserved BBB integrity and alleviated neuroinflammation after ICH.
Akt is a serine/threonine kinase and a primary downstream effector of phosphatidylinositol -3 kinase (PI3K). Akt coordinates a constellation of intracellular signals and is involved in multiple biological functions, including cell proliferation and survival. After phosphorylation, Akt phosphorylates glycogen synthase kinase-3β(GSK3β) at the S9 residue, which suppresses GSK3β activity [22]. Inactivated GSK-3β leads to β-catenin translocation to the nucleus [23], and thus protects against the disruption of BBB through controlling the expression of multiple downstream targets such as occludin and claudin-5 [24,25]. In this study, we demonstrated  AGING that HSPB8 overexpression led to a significant increase in phosphorylation of Akt and GSK3β, as well as promoted cytoplasm-nucleus translocation of βcatenin at 24h following ICH. Therefore, activating Akt/GSK-3β/β-catenin signaling pathway may be involved in the positive role of HSPB8 after ICH. MK2206, a specific inhibitor of Akt, was then chosen to evaluate the precise mechanism underlying the HSPB8's protective effect. MK2206 inhibited the expression of p-Akt, p-GSK-3β and nuclear translocation of β-catenin. Additionally, MK2206 significantly reversed the suppressive effect of HSPB8 on neuronal injury and BBB disruption, and the restorative effect of HSPB8 on claudin-5 and occludin at 24h after ICH. The results demonstrated that HSPB8 alleviated ICH-induced brain injury and BBB disruption through Akt/ GSK3β /β-catenin signaling pathway. However, there are some limitations in the current study. Akt signaling is modulated by a complex network of regulatory proteins during neurological disease processes, often involving crosstalk with scenarios of molecular pathways. Therefore, further studies need to be conducted to elucidate the precise mechanism of HSPB8 in Akt/GSK3β/β-catenin signaling modulation after ICH.
In conclusion, our study demonstrated that HSPB8 overexpression ameliorated brain damage in the early injury stages after ICH via maintaining BBB integrity. Activation of Akt/GSK3β/β-catenin signaling pathway underlies the protective effects of HSPB8. Therefore, HSPB8-based therapies hold a great promise against ICH progression.

Experimental design
The study was divided into three parts as sketched in Figure 10. In experiment 1, 36 rats were randomly divided into six groups (n=6 for each group): sham, 3h, 6h, 12h, 24h and 72h after ICH, for western blot. 8 additional rats in the sham and 24 h groups(n=4) were used for double immunofluorescence staining of HSPB8 combined with specific markers for astrocytes, and microvessels.

Animals
Male Sprague-Dawley (SD) rats (4-6 weeks old) purchased from Shanghai Laboratory Animal Center were used in the experiment. All experimental procedures were authorized by the Institutional Animal Care and Use Committee of The Second Xiangya Hospital, Central South University. The ethical approval reference number is 2017-204. The rats were acclimated in the animal center and fed ad libitum. All procedures were carried out during day time (light cycle for animals).

Intracerebroventricular (i.c.v) injection
Three days before ICH induction, rats were anesthetized, and then i.c.v injection was performed as previously described [26]. Briefly,1 hole was drilled (Stereotaxic coordinates:1.5 mm anteroposterior to the bregma, 1.8 mm mediolateral to the midline). pLV-HSPB8 or pLV-GFP (control) was then delivered into the right lateral cerebral ventricle through a hypodermic AGING needle connected to a Hamilton microliter syringe (5μl, at the rate of 0.5 μl/min). The needle was remained in place for an additional 5 min after injection and then removed slowly.

ICH model
Animal model of ICH was conducted based on a previous study. After anesthesia with 10% chloral hydrate (3 ml/kg, intraperitoneal injection) was used to anesthetize the rats. Under deep anesthesia, the animals were fixed on a rat brain stereotactic apparatus. Stereotaxic injection of collagenase type VII-S (0.3 U in 2 μl sterile normal saline) was infused into right striatum at the following coordinates: 0.2 mm anterior, 3.5 mm right lateral, and 6 mm ventral to the bregma. The needle was withdrawn after 10 minutes. The shamoperated animals received an operation with intracerebrally injection with equal volume of sterile saline.

Neurological deficit scoring evaluation
Neurological functions of the animals were examined and scored blindly at 24h and 72h after ICH. Garcia test consisted of seven parameters: spontaneous activity, symmetry of limb movement, vibrissae touch, axial sensation, lateral turning, forelimb outstretch, and climbing. In the corner turn test, rats approached into a 30° corner. In order to exit, they had to turn around the corner, either turning to the left or right. The choice of left turns for a total of 10 trials was recorded. In the forelimb placing test, the contralesional vibrissae of the rat were swept, and the number of quick placement of the ipsilateral forelimb to the vibrissa stimulation was recorded. The test was repeated 10 times and the percent of ipsilateral forelimb placement was calculated.

Measurement of brain water content
Brain water content was measured via a wet/dry weight method, according to previous literature. Briefly, at 24h and 72h after ICH, the brains of rats were removed immediately after anesthetization and separated into five parts: ipsilateral and contralateral cortex, ipsilateral and contralateral basal ganglia, and cerebellum. Each part was weighed immediately to obtain wet weight (ww). After being dried at 100 °C for 72 h, dry weight (DW) of brain slices was obtained. Brain water content (BWC) was calculated as: [(WW−DW)/(WW)] × 100%.

Evans blue extravasation
Evans blue (EB) extravasation was used to measure the integrity BBB permeability as previously described [27]. Briefly, 2% evans blue dye (4 ml/kg) was injected via intraperitoneal route. After 2 hour circulation, rats were transcardially perfused with 0.1 M cold PBS. Afterwards, brain samples were weighed, homogenized in PBS, sonicated and then centrifuged (15,000g, 30 min). 0.5 mL of the supernatant was then homogenized in an equal volume of 60% trichloroacetic acid. The absorbency of the supernatant at 620 nm was measured using a spectrophotometer. The tissue content of EB was quantified according to a standard curve.

ELISA
The rats were killed 24h after ICH, expression of inflammation markers, including TNF-α and IL-1β, from perihematomal tissue extracts was measured by respective ELISA kits (R&D Systems), according to manufacturer's instructions.

Real-time PCR (RT-PCR)
Total RNA from perihematomal brain tissues was extracted using Trizol reagent (Invitrogen, Life Technologies). Total RNA was reverse transcribed to generate first-strand cDNA using a PrimeScript RT reagent kit (Invitrogen). The real-time PCR primers used in this study are listed in Table 1. Afterwards, RT-PCR was performed in a 10 μl final reaction volume using the miScriptSYBR®green PCR Kit (Qiagen, USA). The mRNA level of each target gene was normalized to GADPH mRNA. Relative levels of gene expression were calculated using 2−ΔΔCt.