Histamine H3 receptors aggravate cerebral ischaemic injury by histamine-independent mechanisms

The role of the histamine H3 receptor (H3R) in cerebral ischaemia/reperfusion (I/R) injury remains unknown. Here we show that H3R expression is upregulated after I/R in two mouse models. H3R antagonists and H3R knockout attenuate I/R injury, which is reversed by an H3R-selective agonist. Interestingly, H1R and H2R antagonists, a histidine decarboxylase (HDC) inhibitor and HDC knockout all fail to compromise the protection by H3R blockade. H3R blockade inhibits mTOR phosphorylation and reinforces autophagy. The neuroprotection by H3R antagonism is reversed by 3-methyladenine and siRNA for Atg7, and is diminished in Atg5−/− mouse embryonic fibroblasts. Furthermore, the peptide Tat-H3RCT414-436, which blocks CLIC4 binding with H3Rs, or siRNA for CLIC4, further increases I/R-induced autophagy and protects against I/R injury. Therefore, H3R promotes I/R injury while its antagonism protects against ischaemic injury via histamine-independent mechanisms that involve suppressing H3R/CLIC4 binding-activated autophagy, suggesting that H3R inhibition is a therapeutic target for cerebral ischaemia.

I schaemic stroke is one of the leading causes of death worldwide 1 . Few therapies are effective other than tissue plasminogen activator, which promotes reperfusion and improves the long-term clinical outcome 2 with limitations 3 . Despite the fact that many neuroprotective agents, including AMPA antagonists, N-methyl-D-aspartate antagonists and 5-hydroxytryptamine 1A agonists, have been developed in attempts to cure ischaemic stroke, the results have been unimpressive for a variety of reasons. Therefore, more effort is needed to elucidate the mechanisms underlying ischaemia and further explore new neuroprotective strategies.
Histamine is an endogenous neurotransmitter in the brain 4 . To date, four subtypes of receptors have been identified: H1R, H2R, H3R and H4R, of which H1R-H3R are found in brain [5][6][7] . H3R (histamine H3 receptor) is a presynaptic autoreceptor that regulates histamine release from histaminergic neurons via negative feedback 8,9 , as well as a heteroreceptor on nonhistaminergic neurons that regulates the release of many other neurotransmitters [10][11][12][13] . H3R is a G-protein-coupled receptor that activates G i/o proteins to inhibit adenylyl cyclase activity and modulate phospholipase A2 and mitogen-activated protein kinase activity 14 . In primary cultured rat cortical neurons, H3R activates the Akt/glycogen synthase kinase 3b (GSK-3b) axis both in a constitutive and an agonist-dependent manner 15 . Dysregulation of the downstream signalling of H3R in the Akt/GSK-3b pathway is linked to several prevalent pathological conditions 16 . H3R antagonism blocks neuropathic pain in rats 17,18 , electrically induced convulsions 19 and amygdaloid-kindled seizures 20 . In addition, our previous data showed that H3R antagonism protects against N-methyl-D-aspartate-induced neurotoxicity in cultured cortical neurons 21 . H3R antagonism also inhibits ischaemiainduced oxidative stress 22 . Therefore, it has been proposed that H3R blockade is generally neuroprotective. However, few studies have investigated its role in ischaemia and the underlying mechanisms.
After ischaemia/reperfusion (I/R) injury, most cells in the penumbra undergo apoptosis 23 . Recently, investigations showed that autophagy also participates in the pathological process of ischaemia 24,25 . Autophagy is a major cellular process for the degradation of long-lived proteins and cytoplasmic organelles in eukaryotic cells 26 . Studies show that inhibition of GSK-3b, a downstream kinase of H3R, stimulates mammalian target of rapamycin (mTOR) signalling and thus inhibits autophagy 27 . Although the role of autophagy in ischaemia remains controversial, recent reports indicate that it is neuroprotective in moderate injury 24,[26][27][28] , and these investigations have given rise to the proposal that the injury can be rescued by inducing autophagy 28 .
Therefore, the present study is designed to investigate the role of H3R in I/R injury in vivo and in vitro. We find that H3R antagonism and H3R deletion protect against I/R injury in a histamine-independent manner. H3R antagonists reinforces I/R-induced autophagy and the neuroprotective effect of H3R antagonism is reversed by autophagy blockade. Further investigations show that H3R antagonism decreases the binding of H3R to chloride intracellular channel 4 (CLIC4), and thus further increases autophagy. Therefore, H3R aggravates ischaemic brain injury by histamine-independent mechanisms, and its antagonism may provide a novel neuroprotective strategy against cerebral ischaemia by regulating autophagy.
Neuroprotection by H3R antagonists is histamine independent. As a presynaptic receptor on histaminergic neurons, H3R suppresses histamine synthesis and releases in a negative feedback manner 8 . Thus, inhibition of H3R by thioperamide leads to synaptic histamine release 29 . Therefore, we asked whether the protection conferred by H3R inhibition is histamine dependent, and found that the neuroprotection by thioperamide was not reversed by pyrilamine (10 mg kg À 1 ) or cimetidine (10 mg kg À 1 ), H1R and H2R antagonists in WT mice (Fig. 3a,b). Moreover, a-fluoromethylhistidine (a-FMH; 50 mg kg À 1 i.p., administered 2 h before ischaemia), a selective histamine synthesis inhibitor by irreversibly inhibiting histidine decarboxylase (HDC), did not reverse the neuroprotection of thioperamide as shown by the neurological deficit score (Fig. 3c), as well as the infarct volume in WT mice (Fig. 3a,b). a-FMH is reported to maximally decrease histamine activity 2 h after administration and this lasts for 4 days in mice 30 . More importantly, in HDC À / À mice, the protection by thioperamide remained, as revealed by the reduced neurological score and infarct volume (Po0.01; Fig. 3).

CLIC4 is involved in H3R antagonism-enhanced autophagy.
It has been reported that H3R may function by binding with CLIC4 (Maeda et al. 32 ), which is thought to regulate autophagy 33 . Thus, we speculated that thioperamide regulates autophagy in the context of OGD/R in a CLIC4-related manner. To test this hypothesis, we assessed CLIC4 expression and the influence of thioperamide on the interaction between H3R and CLIC4. After OGD/R, both H3R and CLIC4 total protein were upregulated, and the interaction of H3R with CLIC4 significantly increased (Fig. 7a). Co-immunoprecipitation analysis showed that thioperamide did not change the protein expression of H3R or CLIC4, but it inhibited the interaction of H3R with CLIC4, which was reversed by the H3R agonist immepip (Fig. 7a). To further clarify the role of CLIC4 in the protection by thioperamide, the peptide Tat-H3R CT414-436 (10 À 6 mol l À 1 at reperfusion) was used to block the H3R-CLIC4 binding in neurons ( Supplementary Fig. 5). Tat-H3R CT414-436 preserved the viability under OGD/R (increased from 64.92 ± 6.92 to 95.54 ± 2.37% of control, Po0.01, Fig. 7b). Moreover, Tat-H3R CT414-436 led to dephosphorylation in the Akt/GSK-3b/mTOR/P70S6K signalling pathway, upregulating LC3-II expression and increasing LC3positive autophagic vacuoles (Fig. 7c,d). In addition, siRNA for CLIC4 increased the viability under OGD/R (from 55.70 ± 1.59 to 82.26±2.93% of control, Po0.001, Fig. 7g), and increased the LC3-II expression as well as the cumulative autophagy by dephosphorylating the Akt/GSK-3b/mTOR/P70S6K signalling pathway ( Fig. 7e-g).

Discussion
The H3R is found predominantly in the brain, which implies a role in brain disorders. In the present study, we demonstrated for the first time that H3R antagonism protects against I/R injury in vivo and OGD/R injury in vitro. Both H3R antagonists and H3R-knockout decreased the ischaemic damage. We also found that H3R antagonism disturbed the binding of H3R with CLIC4, which may subsequently protect cells against OGD/R injury via Akt/GSK-3b signalling. The present findings shed light on H3R and its signalling pathways in the context of ischaemic brain injury, and suggest that H3R is a potential target in therapy for cerebral ischaemia. At present, the role of H3Rs in cerebral ischaemia has not been fully elucidated. We showed that OGD/R significantly increased H3R expression, and H3R transfection increased the vulnerability of HEK293 cells to serum deprivation ( Supplementary Fig. 1). Conversely, three H3R antagonists (thioperamide, clobenpropit and A331440) protected against I/R-induced focal brain ischaemia. The protection by thioperamide was significantly reversed by the selective H3R agonist  (a) 2,3,5-Triphenyltetrazolium chloride (TTC)-stained brain sections from WT and HDC À / À mice showing the infarct area in those receiving saline, the H3R antagonist thioperamide (THIO), the H1R antagonist pyrilamine (PYRI; 10 mg kg À 1 , i.p., at reperfusion), the H2R antagonist cimetidine (CIME; 10 mg kg À 1 , i.p., at reperfusion) and a-FMH (50 mg kg À 1 , i.p., 2 h before ischemia) after 24 h reperfusion (n ¼ 6 per condition; scale bar, 5 mm). (b) Infarct volumes and (c) neurological deficit scores (*Po0.05, **Po0.01, ***Po0.001 with analysis of variances followed by the Bonferroni/Dunn post hoc test). Data are presented as mean ± s.e.m. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4334 ARTICLE immepip, suggesting that H3R aggravates ischaemic injury. We further used H3R À / À mice which, interestingly, showed ameliorated ischaemic brain injury compared with their WT littermates, and the protective effects of H3R antagonists were absent. It is likely that H3R antagonists protect ischaemic neurons by directly blocking H3Rs. Therefore, our data for the first time showed that H3R aggravates ischaemic injury and its inhibition enhances neuronal survival.
H3R is a presynaptic autoreceptor that inhibits histamine release via negative feedback, and H3R antagonism leads to increased histamine release. H1R and H2R are postsynaptic receptors responsible for protective histaminergic signalling 34,35 . Thus, it has been hypothesized that the beneficial effects of H3R antagonism might involve the augmentation of central histaminergic activity 34,36,37 . To our surprise, neither the H1R antagonist pyrilamine nor the H2R antagonist cimetidine   compromised the protection by thioperamide. We demonstrated that thioperamide-induced histamine release was not sufficient to rescue ischaemic brain injury even in the presence of metoprine, which inhibits histamine degradation 38 . So, it is likely that increased histamine does not underlie the protection by H3R antagonists. To provide more solid evidence, we showed that both a-FMH, a selective HDC inhibitor that decreases histamine synthesis 30 (Supplementary Fig. 6), and HDC knockout failed to reverse the protection by thioperamide. In addition, it has been reported that H3R À / À mice show a significantly decreased brain histamine level 39 . Nevertheless, the ameliorated I/R injury was still evident in the present data, which further supported our notion that histamine is not involved in the protection by H3R antagonism. Taken together, these results suggested that a new non-histaminergic mechanism exists in the protection against ischaemic brain injury. Studies have revealed that brain H3Rs couple with G i/o proteins by which protein kinases such as cyclic AMP-dependent protein kinase A and GSK-3b are modulated 15,40 . The present work found that LiCl (a GSK-3b inhibitor) completely reversed the protection by thioperamide, indicating that activation of GSK-3b may be necessary for the protection by H3R antagonism. It has been reported that I/R-induced activation of Akt, an upstream kinase of GSK-3b, may ultimately lead to neuronal injury 41 , supporting our results that Akt/GSK-3b phosphorylation was deleterious to neuronal survival. Our results are also supported by a recent finding that dephosphorylation of GSK-3b protects against prolonged myocardial ischaemic injury 27 . In addition, it was previously reported that an H3R agonist protects neurons against neuroexcitotoxicity by reversing the dephosphorylation of the Akt/GSK-3b pathway 42 , whereas we found that Akt/GSK-3b was significantly phosphorylated in OGD/R injury (Fig. 4a), which suggests that different mechanisms may be involved in the two cell-injury models. Thioperamide may be protective in targeting I/R-induced cell injury that involves phosphorylated Akt/GSK-3b.
Interestingly, the activation of GSK-3b by H3R inhibition subsequently inhibits mTORC1 activity, as revealed by decreased phosphorylation of p70S6K 43 . The mTORC1 signalling plays a key role in autophagy induction, and we found that both H3R inhibition and knockout enhanced autophagy in I/R. Additional data indicated that autophagy occurred primarily in neurons but neither in astrocytes nor in microglia. In addition, autophagy inhibition by 3-MA, Atg7 silencing and Atg5 knockout all reversed the protection by thioperamide against I/R injury, suggesting that autophagy reinforced by H3R inhibition underlies the neuroprotection. Our previous research also showed that the protective role of autophagy during I/R may be attributable to autophagy-related mitochondrial clearance and the inhibition of downstream apoptosis 44 . Moreover, our previous study  ARTICLE showed that autophagy plays distinct roles in ischaemia and the reperfusion phase. In particular, we demonstrated that autophagy protects against reperfusion-induced neuronal injury 44 . Here we found that thioperamide only protected against I/R injury but not that induced by ischaemia alone. These findings reinforced the idea that protective autophagy is only activated during reperfusion. Therefore, our results revealed a novel protective mechanism of H3Rs. It is noteworthy that the H3R is also a heteroreceptor involved in the regulation of several neurotransmitters including glutamate, dopamine, 5-hydroxytryptamine and g-aminobutyric acid 45 . These neurotransmitters are known to participate in the pathogenesis of cerebral ischaemia, and it has been suggested that H3R antagonism ameliorates excitotoxicity 21 . Besides, an anti-inflammatory effect of H3R antagonism has also been suggested in brain ischaemia 46 . Therefore, the involvement of other mechanisms cannot be excluded.
To further explore the mechanisms by which H3R blockade enhances autophagy, we investigated the involvement of histamine in the regulation of autophagy by H3R. Interestingly, histamine had no effect on mTOR phosphorylation and autophagy in OGD/R ( Supplementary Fig. 7), which provided further evidence for our previous notion that H3R antagonism protects against ischaemic injury in a histamine-independent manner. It has been suggested that CLIC4 is an H3R-interacting protein, and Zhong et al. 33 have recently reported that CLIC4 is involved in the regulation of autophagy in glioma cells 32 . Therefore, we proposed that the inhibition of H3R suppresses H3R binding with CLIC4 and subsequently induce autophagy. We found that CLIC4 expression was significantly upregulated and its binding with H3R was also increased in OGD/R neurons, suggesting that CLIC4 might be involved in the cell injury in response to stress 47 . More interestingly, H3R inhibition by thioperamide disturbed the H3R-CLIC4 interaction without altering their expression. We also determined the localization of H3R in mouse brain; it was shown that H3Rs predominantly locate in neurons in cortex, hippocampus and stratum ( Supplementary Fig. 8). This effect of thioperamide was reversed by the H3R agonist immepip, indicating that CLIC4-H3R binding may, at least partly, inhibit the autophagy regulated by H3R antagonism. To clarify this, we used the peptide Tat-H3R CT414-436 , which competitively blocks the CLIC4-H3R interaction by binding with CLIC4. We found that Tat-H3R CT414-436 conferred protection to an extent similar to H3R blockade in OGD/R (Fig. 7a). In addition, Tat-H3R CT414-436 inhibited Akt, activated GSK-3b, inhibited mTOR/P70S6K signalling and ultimately upregulated the autophagy level in the OGD/R period. Furthermore, siRNA for CLIC4 also exhibited protection through autophagy induced by mTOR inhibition (Fig. 7). Therefore, these data strongly suggested that H3R blockade disturbs CLIC4-H3R binding and then activates autophagy to protect cells against I/R injury. Previous investigations indicated that suppression of CLIC4 may overactivate Beclin1 and thus activates autophagy, which implies it could be an underlying mechanism 33 . Further studies are needed to address this issue.
In conclusion, the present study showed for the first time that H3R enhances I/R-induced neuronal injury. H3R inhibition confers neuroprotection against ischaemia through histamineindependent mechanisms. H3R antagonism inhibits CLIC4-H3R binding and subsequently further activates I/R-induced autophagy, which protects against ischaemic injury (Fig. 8). The present study provides new prospects for H3R antagonists in therapeutic intervention for cerebral ischaemia.
In brief, a 6-0 nylon monofilament suture, blunted at the tip and coated with 1% poly-L-lysine, was advanced B10 mm into the internal carotid to occlude the origin of the MCA. Reperfusion was allowed after 1 h by monofilament removal. Body temperature was maintained at 37°C with a heat lamp (FHC, Bowdoinham, ME, USA) during surgery and for 2 h after the start of reperfusion. CBF was determined in the territory of the MCA by laser Doppler flowmetry (Moor Instruments Ltd.). A flexible fibre-optic probe was affixed to the skull over the cortex supplied by the proximal part of the right MCA (2 mm caudal to the bregma and 6 mm lateral to the midline). Animals with o80% reduction in CBF in the core of the MCA territory were excluded from the study. Neurologic deficit scores were evaluated at 24 h of reperfusion as follows: 0, no deficit; 1, flexion of the contralateral forelimb on lifting of the whole animal by the tail; 2, circling to the contralateral side; 3, falling to the contralateral side; and 4, no spontaneous motor activity 49,50 .
Infarct volume measurement. Infarct volume was determined 24 h after reperfusion. The brains were quickly removed, sectioned coronally at 2-mm intervals and stained by immersion in the vital dye 2,3,5-triphenyltetrazolium hydrochloride (0.25%) at 37°C for 30 min. The extents of the normal and infarcted areas were analysed using ImageJ (National Institutes of Health, Bethesda, MD, USA) and determined by the indirect method, which corrects for oedema (contralateral hemisphere volume minus non-ischaemic ipsilateral hemisphere volume). The percentage of the corrected infarct volume was calculated by dividing the infarct volume by the total contralateral hemispheric volume, and this ratio was then multiplied by 100.
Cell culture and transfection. For primary neuronal cell culture, pregnant Sprague-Dawley rats or C57BL6J mice were anaesthetized by i.p. injection of chloral hydrate (400 mg kg À 1 ), and the cortex was isolated from embryos (16 days) for primary cortical neuron cultures. Cells (800-1,000 cells mm 2 ) were seeded on coverslips coated with 30 mg ml À 1 poly-D-lysine. Cells were placed in fresh serumfree Neurobasal medium (21103, Gibco) plus 2% B27 and fed every 4 days with fresh medium and used after 7 days (DIV7). In addition, HEK293 cells (from the ATCC) and MEFs (kind gifts from Professor Mizushima) were cultured in DMEM with 10% fetal bovine serum (Gibco). The full-length coding domain sequence of mouse H3R was constructed into plasmid pcDNA3.1 and transfected into HEK293 cells by Lipofectamine 2,000 reagent (Invitrogen) 24 h before experiments.
Oxygen-glucose deprivation. OGD was induced in DMEM without glucose and saturated with 5% CO 2 /95% N 2 for 2 h for neurons, 9 h for MEFs. After OGD, 24 h reperfusion with high-glucose DMEM (4.5 g l À 1 ) in normoxia was carried out. For serum starvation, HEK293 cells transfected with mouse H3R complementary DNA or vehicle plasmids were exposed to DMEM without serum for 24 h.
Apoptosis determined by TUNEL assay. Apoptotic cells were determined by TUNEL assay (Roche) and the total cell number was counted after DAPI (4 0 ,6-diamidino-2-phenylindole) staining. For mouse brain samples, coronal cryosections from at least five animals in each group were obtained. Five random fields in the penumbral areas of each section were observed. Cultured cells had been seeded on poly-L-lysine-treated coverslips, and at least three coverslips from each group were included in one experiment. Five random fields were observed on each coverslip, and the experiments were repeated independently three times. The results were expressed as the percentage of TUNEL þ /DAPI þ cells in the sections.
Co-immunoprecipitation and western blot. Mice were anaesthetized by i.p. injection of chloral hydrate (400 mg kg À 1 ), killed 24 h after tMCAO or sham operation, the brain was quickly removed and was immediately put in À 40°C for 5 min. To dissect out the ischaemic penumbra from the ipsilateral hemisphere, the frozen brains were put on a block of pre-chilled metal. A longitudinal cut from top to bottom was made B1 mm from the midline through the ipsilateral hemisphere. Then, a transverse diagonal cut at approximately the '2 o'clock' position was made to exclude the ischaemic core region 51 . The separated tissue was lysed in ice-cold lysis buffer containing (in mmol l À Membranes were washed three times in TBST buffer and incubated with the appropriate secondary antibodies (Odyssey, LI-COR, 1:5,000 dilution) for 2 h. Images were acquired with the Odyssey infrared imaging system and analysed as specified in the Odyssey software manual. The results were expressed as the target protein/GAPDH ratio and then normalized to the values measured in the sham or control groups in vivo and in vitro (presented as 100%). The lanes marked 'input' were loaded with 10% of the starting material used for immunoprecipitation.
Histamine content measurement by HPLC. Samples were de-proteinized with 0.4 mol l À 1 perchloric acid and centrifuged at 15,000 g for 20 min at 4°C. Then, the supernatant was removed and filtered through a 0.22-mm polyvinylidene difluoride membrane. Analysis of histamine in the supernatant was performed by HPLC (ESA, Chelmsford, MA, USA). The system consists of model 582 pump and four channel CoulArray electrochemical detector. After reacting with the derivate ophthalaldehyde, analytes were separated on a 3-mm, 3 Â 50 mm 2 Capcell Pak MG C18 column (Shiseido, Japan). A two-component gradient elution system was used, with component A of the mobile phase being 100 mmol l À 1 Na 2 HPO 4 , 13% acetonitrile and 22% methanol, pH 6.8, and component B being similar to A except with 5.6% acetonitrile and 9.4% methanol. A gradient elution profile was used as follows: 0-3.5 min, isocratic 100% B; 3.5-20 min, linear ramp to 0% B; 20-22 min, isocratic 0% B; 22-23 min, linear ramp to 100% B; 23-30 min, isocratic 100% B. The flow rate was set to 0.75 ml min À 1 . The temperature of the column was maintained at 38°C. The data were acquired and analysed using CoulArray software.