Nanoemulsions of Hydroxysaor Yellow A for Enhanced Brain Delivery in Cerebral Ischemia Reperfusion-injured mice: Physicochemical and In Vivo Performances

Stroke has always been a disease threatening human life and health worldwide. Here, we synthesized a new type of hyaluronic acid-modied multi-walled carbon nanotube. Then we prepared hydroxysaor yellow A-hydroxypropyl-β-cyclodextrin phospholipid complex water-in-oil nanoemulsion with hyaluronic acid-modied multi-walled carbon nanotubes and chitosan (HC@HMC) for oral treatment of ischemic stroke. First, we measured the intestinal absorption and pharmacokinetics of HC@HMC in rats. We found that the intestinal absorption and the pharmacokinetic behavior of HC@HMC is better than HYA. Then we detected intracerebral concentrations after oral administration of HC@HMC and found that more HYA crossed the blood-brain barrier (BBB). Finally, we evaluated the ecacy of HC@HMC in middle cerebral artery occlusion/reperfusion (MCAO/R)-treated mice. In MCAO/R mice, oral administration of HC@HMC demonstrated a signicant protection against cerebral ischemia-reperfusion injury (CIRI). And we also found HC@HMC may exert a protective effect on cerebral ischemia-reperfusion injury through COX2/PGD2/DPs pathway. These results suggest that oral administration of HC@HMC may be a new strategy for the treatment of stroke.


Introduction
Stroke is a brain disease that generates devastating neurological de cits and is a major public health challenge, given its high mortality and adult disability worldwide [1] . Stroke is divided into hemorrhagic stroke(cerebral blood vessels are ruptured) and ischemic stroke (cerebral blood vessels are blocked), of which ischemic stroke accounts for about 87% of the total incidence of stroke [2] . Ischemic stroke is a kind of cerebrovascular accident caused by the reduction or interruption of local blood supply to brain tissue [3] . Ischemic tissue will appear the corresponding regional neurological de cit symptoms, causing a series of complex biochemical molecular cascade reactions [4] , such as energy metabolism disorders, activation of excitotoxic glutamate signaling pathway, free radical response, in ammatory response and so on, usually to restore perfusion and reoxygenation for the treatment clinically. However, perfusion and reoxygenation will cause cerebral ischemia-reperfusion injury (CIRI) [5] , a kind of brain tissue injury and related dysfunction caused by reperfusion of blood ow after cerebral ischemia, with the characteristics of high incidence, high disability rate and high mortality. Several mechanisms such as excitotoxicity, oxidative stress induced damage, apoptosis and in ammation are involved in the process of ischemiareperfusion [6,7] . Studies have shown that in ammation/immune response plays a key role in the pathogenesis of CIRI and has a dual role in promoting tissue damage and repair. Our previous studies have found that in ammation and cyclooxygenase-2 (COX-2) and its downstream pathways are involved in the mechanism of CIRI [8] , but whether the pathogenic mechanism of hydroxysa or yellow A (HYA) on CIRI involves COX-2/PGD2/DPs pathway has not been reported.
Sa ower is a Chinese medicine that can promote blood circulation and remove blood stasis, which contains a variety of chemical components. Of sa ower yellow pigment, HYA is the highest in content, and has the strongest pharmacological effect. Many research teams have indicated that HYA contains antioxidant [9] , anti-tumor [10] , and anti-ischemia reperfusion injury and other effects [11] . And HYA is widely used in the clinical treatment of dysemia diseases such as myocardial ischemia, cerebral ischemia, coronary heart disease, and cerebral thrombosis [12] . And there is no doubt that HYA is a promising lead drug candidate in designing new multi-targeted therapeutic agents against cardio-cerebrovascular diseases. So far, only the injection of HYA has been widely used in clinical therapy [13,14] .
However, its application was restricted by short half-life and low brain concentration. HYA is a hydrophilic compound with a chalcone glycoside structure, which has poor lipid solubility and low oral bioavailability [15] . So it is di cult for HYA to cross the BBB and reach the focus, and the BBB is still the main barrier for brain disease intervention [16] . In order to achieve the purpose of improving the oral absorption and bioavailability of HYA, HYA has been prepared into solid lipid nanoparticles [17] , a selfdouble-emulsifying drug delivery system [18] , water-in oil microemulsion [15] and other dosage forms. All can improve the oral absorption of HYA, but the shortcomings of low bioavailability was not signi cantly improved, and when used for the treatment of brain diseases, the intracranial drug concentration is still insu cient. Therefore, the oral preparation of HYA for brain disease intervention is urgently needed.
Nanoemulsions are biphasic dispersion of two immiscible liquids: either the water in oil (W/O) or the oil in water (O/W) is stabilized by an amphiphilic surfactant [19] . Nanoemulsions consist of nano-scale droplets, which can slowly release, target accurately and of low toxicity. It is able to increase the contact area with the gastric mucosa, enhance the ability to absorb, get through the BBB, and prevent the hydrolysis of the wrapped substance or the degradation of enzyme [20,21] . It is a promising drug carrier in brain-targeted preparations [22,23] . Desai et al prepared oral Darunavir loaded lipid nanoemulsions, which improved the bioavailability of Darunavir and enhanced brain to absorb [20] . Phospholipids are amphiphilic, with good solubility in fat and bio lm compatibility. Drugs can be combined with phospholipids to prepare phospholipid complexes, thereby improving their bioavailability [24,25] . Results by Liu et al showed that the stability constant of the complex formed with HPCD was enhanced after avanones are complexed with cyclodextrin [26] . Chitosan is a cationic polysaccharide with the function of adhering mucous membrane, which can reversibly open Caco-2 cells to promote drug absorption [27] , Ma et al prepared the HYA-CS complex, the bioavailability of HYA was increased by 4.76 times [28] . Lv et al also found that HYA is a Pglycoprotein substrate analog, which is easily excreted by P-glycoprotein in the intestinal tract, and leads to a strong bile e ux effect and low absorption [18] .
The effect of compound formulations was superior to that of single formulations, so it combined the advantages of multiple carriers to deliver HYA. Therefore, we formed HYA complexing with phospholipid and HPCD under certain conditions, and then prepared the water-in-oil HYA complex nanoemulsion with hyaluronic acid-modi ed multi-walled carbon nanotubes and chitosan (HC@HMC) by titration, hoping to improve the oral bioavailability and the ability to cross the BBB of HYA for treatment of ischemic stroke via oral administration. We measured the intestinal absorption and pharmacokinetics of HC@HMC in rats. And we determined the intracerebral concentration of HC@HMC after oral administration. Finally, we explored the protective effect of HC@HMC on cerebral ischemia-reperfusion in mice and HC@HMC may protect mice against cerebral ischemia-reperfusion injury through COX2/PGD2/DPs pathway.

Synthesis of MWCNT-HA
Firstly, synthesis carboxylated multi-walled carbon nanotubes (MWCNT-COOH), the mixed acid solution (the volume ratio of concentrated sulfuric acid and concentrated nitric acid is 3:1) and 30% H 2 O 2 solution (10 mL) were added into the MWCNT (100 mg). After 3-hour ultrasonic reaction, it was diluted with a large amount of distilled water, using 0.22 µm microporous lter membrane (mixed cellulose grease) to suck and lter, and then washed with distilled water to neutral, then put the ltered solid into the oven at a constant temperature, 80°C for drying, that was MWCNT-COOH.
Secondly, synthesis of aminated HA-NH2(HA), HA (200 mg) was dissolved in formamide (10 mL) at 50°C, EDC (520 mg) and NHS (310 mg) dissolving in 10 mL and 5 mL formamide respectively, and the activation reaction was performed for 30 min. 2 mL of ethylenediamine was dissolved in 10 mL of formamide. Under ice bath conditions, the activated HA solution was slowly dropped into the ethylenediamine solution, controlling the dripping rate (after 60 min of dripping). The reaction solution was warmed to room temperature, and reaction was continued for 3 h. Then added a large amount of prechilled acetone precipitate, the precipitate was reconstituted with water, with a 0.45 µm lter membrane, transfer to a dialysis bag (MWCO=3500) for 48 h, and change the water every 6 h, freeze-dried to get HA-NH2.
Finally, MWCNT-COOH is combined with HA (MWCNT-HA), The amidation reaction of HA-NH2 and MWCNT-COOH under the conditions of EDC and NHS. Brie y, formamide (30 mL) was added into MWCNT-COOH (80 mg), transferred to a round-bottom ask after 30 minutes of sonication under ice bath conditions, rinsed with formamide (20 mL), and the formamide was transferred to a round bottom ask. EDC (305 mg) and NHS (182 mg) were dissolved in 5 mL formamide, and then transferred to a ask, activated at room temperature for 30 min. Then, 10 mL formamide with HA-NH2 (160 mg) and triethylamine (180 µL) was quickly added dropwise to the activated MWCNT-COOH, and reacted for 24 h. Pre-cooled excess acetone (4 times of the reaction solution) was added in an ice bath, cooled, crystallized, and left to precipitate MWCNT-HA, then suck and lter with 0.22 µm organic membrane, wash the precipitate with acetone, reconstitute with ultrapure water, using a dialysis bag (MWCO=12000 kDa) to dialyze in water, and after 48h freeze-drying to obtain the product MWCNT-HA.

Preparation of HC@HMC
HYA, HPCD and PC (molar ratio 1: 2.33: 2.47) are dissolved in a clean and dry round-bottom ask in an appropriate amount of absolute ethanol and stirred for 3.5h (50°C) in the dark. Rotary evaporation removing the absolute ethanol to obtain HYA compound hydroxysa or yellow A complex (HYAC); After HYAC was completely dissolved in the mixed solution of GMC, RH40, and PEG400, the chitosan solution (2mg·mL −1 ) was added dropwise. The solution will undergo a process from turbid to clear and transparent. Finally, a certain amount of novel hyaluronic acid-modi ed multi-walled carbon nanotubes (HMC), F188 and F407 mixture was added, and stirring was continued until the dissolution was complete, forming a clear, uniformly dispersed viscosity thick liquid, that was hydroxysa or yellow Ahydroxypropyl-β-cyclodextrin phospholipid complex water-in-oil nanoemulsion with hyaluronic acidmodi ed multi-walled carbon nanotubes and chitosan modi ed (HC@HMC).

Characterization of HC@HMC
Malvern laser particle size potentiometer (Zetasizer Nano zs90; UK) was used to measure the particle size and Zeta potential of HC@HMC; conductivity meter (DDS-307A; Shanghai, China) was used to measure conductivity and the pH of HC@HMC.

Absorption in the Intestine
The in situ absorption of HYA or HC@HMC in the intestinal system of rats was investigated by the unidirectional perfusion method [29] . All rats were raised under controlled conditions and fasted more than 12 h before drug was administrated. Rats were anesthetized by intraperitoneal injection of 20% uratan solution (7 mL·kg −1 ). Then the rats were xed and a midline abdominal incision was made, found the duodenum, jejunum, ileum and colon, respectively, cut a small cannula at the upper end of the four intestine segments, ligated and xed at the lower end, the wound covered with clean gauze soaked with saline. The contents were rinsed with physiological saline at a constant temperature (37°C) and drained, then equilibrated with Krebs-Ringer solution at a ow rate of 0.25 mL·min −1 for 15 min. Finally, HYA circulating liquid or HC@HMC circulating liquid (the circulating liquid was placed in a 100 r·min −1 magnetic stirrer) was injected and stirred at a ow rate of 0.25 mL·min −1 . After 1 h, we respectively collected the perfusate out ow at the outlets of the four intestinal segments. After the experiment, the four intestine segments were cut out, and the length (L) and radius (R) of each intestine segment were measured and recorded. The samples were sonicated with methanol for 20 min, and the continuous ltrate ltered by 0.45 µm microporous membrane was collected. For the 100 µL intestinal juice sample, 10 µL 50 µg·mL −1 RT and 50 µL perchloric acid solution (1 mol·L −1 ) were added. After vortexing, centrifugation at 12000 rpm for 10 min, the supernatant was analyzed by HPLC.
The measured HYA content in the intestinal perfusion sample was substituted into the following three formulas to calculate the absorption rate constant (Ka), effective permeability coe cient (Peff) and absorption percentage (PA) of HYA and HC@HMC respectively.
In this formula X 0 represents the total mass of the initial drug; X t represents the total mass of the drug at time t; C 0 represents the initial drug concentration; X in represents the total mass of the drug entering the perfusate; X out represents the total mass of the drug owing out of the perfusate; t represents perfusion time; Q stands for ow velocity; R and L are the inner diameter and length of each intestinal segment of perfusion. For the K a , P eff and PA, non-parametric statistical Wilcoxon signed rank sum test was used to compare the signi cant difference between HYA and HC@HMC (P<0.05).

Pharmacokinetic of HC@HMC
Twelve male SD rats weighing (230±20) g were randomly divided into 2 groups and fasted for 12 h, gavaged with HYA solution or HC@HMC (equivalent to HYA 6 mg·kg −1 ). After administration, 0.5 mL blood was collected from the orbit of rats at different time points. Each sample was immediately transferred to heparin-in ltrated centrifuge tubes, after centrifugation at 6000 r·min −1 for 10 min, the supernatant 100 µL was taken and added 10 µL 50 µg·mL −1 RT and 50 µL 6% perchloric acid solution into, then after vortexed and centrifuged at 12000 r·min −1 for 10 min, and the supernatant was measured by HPLC. Pharmacokinetic parameters were analyzed by DAS 2.1.1 using non-compartmental analysis, then analysis of variance and double one-sided t-test were carried out.

LC/MS Analysis of HYA in Brain
Eighteen mice were randomly divided into six groups. After intravenous administration of HYA solution (40 mg·kg −1 ) or oral administration of HC@HMC (160 mg·kg −1 ), the mice were sacri ced 0.5 h, 1 h or 1.5 h later, and brain tissues were taken and weighed. 500 µL of ultrapure water was added into the homogenizer and homogenized the brain tissue; centrifuged the homogenate (12000 r·min −1 , 10 min, 4°C); then 100 µL of supernatant was taken, 10 µL of internal standard (RT) solution and 490 µL of methanol were added into, centrifuged in the same way after mixing, the supernatant was used for sample injection.  Middle cerebral artery occlusion/reperfusion model Ischemic stroke was produced by using the MCAO/R method as previously reported [30] . The mice were anesthetized by intraperitoneal injection of sodium pentobarbital (40mg·kg −1 ). The right common carotid artery (CCA), external carotid artery (ECA), and the internal carotid artery (ICA) were exposed under a surgical microscope. The whole of the ECA and CCA were ligated. A nylon suture (Jialing, Guangzhou, China) was inserted from the CCA into the right side of ICA in a depth of 10±0.5 mm, to occlude the origin of the middle cerebral artery (MCA). After 1-hour occlusion, reperfusion was achieved by withdrawing the suture to restore blood supply to the MCA territory. The sham-operated mice underwent the same surgery, but without the suture inserted. Body temperature was maintained 37±0. (n=10). Mice were given HYA (Intravenous administration) and HC@HMC (Intragastric administration) 5 days before surgery, once a day. The brains of mice were removed at 24h after MCAO/R operation. The recommended clinical dose of HYA is 100 mg (approximately 2 mg·kg −1 ) once a day for common cases or twice a day for severe cardiac patients [31] . Therefore, to mimic the current clinical treatment regimen, we chose to treat the ischemic mice with HYA at a dose of 20 mg·kg −1 . According to the absolute bioavailability, the dose of HC@HMC was 80 mg·kg −1 .

Neurological Scoring
The neurological de cit evaluation was conducted after 24h-reperfusion according to the Zea Longa' method as previously described (Longa et al. 1989) [32] . The criteria were used as follows: 0, normal, no neurological de cit; 1, mild neurological de cit, failure to completely extend the right forelimb; 2, moderate neurological de cit, twisting to the contralateral side; 3, severe neurological de cit, falling to the left; 4, no ability of spontaneous motor activity. Animals that died during the process were not included in the statistics.

Rota Rod Test
Motor coordination was tested by using rota-rod tread-mill test [33] . The rota rod apparatus (UGO BASILE S.R.L, Italy) consisted of a rotating rod (75 mm diameter), on which the mice were allowed to hold. After twice daily training for 2 days (at a speed of 10 rpm to 40 rpm in 5 minutes), mice were tested 2 times as preoperative value on the third day. After 24 hours' reperfusion, mice were tested 2 times as postoperative value. The time for each mouse to remain on the rotating rod was recorded. The maximum time allowed

Hematoxylin/eosin (HE) Staining
Hematoxylin and eosin (HE) staining was used to observe pathological histological damage in the cerebral cortex and hippocampus. After 24-h reperfusion, the mice were anesthetized with sodium pentobarbital and perfused with PBS, and then perfused with 4% paraformaldehyde. After that, the brains were dehydrated by graded ethanol, embedded in para n, and cut into 5-µm-thick sections. Finally, the sections were stained with HE reagents and observed by light microscopy.

Real-Time Polymerase Chain Reaction
Total RNA was isolated from the mice cerebral cortex by Trizol reagent (Vazyme, Nanjing, China) according to the manufacturer's protocol. Reverse transcription of mRNA was performed using HiScript Q Select RT SuperMix (Vazyme, Nanjing, China). To detect the amount of COX2 mRNA and DP1 mRNA and DP2 mRNA. SYBR Green II (Biomake, USA) incorporation method was applied with ATGB being an internal control of mRNA. The primer sequences are reported in Table 1.

Western Blotting
Mouse cortical samples were lysed with RIPA containing 1% phenylmethanesulfonyl uoride (PMSF) (100 mM). The sample lysis buffer was collected into an EP tube and centrifuged at 12,000 rpm and at 4°C for 15min. Then the precipitate was discarded and the protein concentration was determined using the BCA Protein Assay Kit (Beijing, China). We performed Western blotting as described previously. In short, proteins were separated and transferred onto a polyvinylidene uoride (PVDF) membrane (Millipore, USA). The membrane was incubated with primary antibodies overnight at 4°C. After washing, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (Proteintech, Wuhan, China) for 1h at room temperature. At last, the immunoblots were developed by ECL western blotting detection reagent (Millipore, USA). The primary antibodies against COX2 (dilution 1:1000, Abcam, USA) and DP1 (dilution 1:1000, Abcam, USA) and were purchased from Abcam. DP2 (dilution 1:500, Ai nity, China) were purchased from Ai nity. and Tublin (dilution 1:5000, Proteintech, China) were purchased from Proteintech.

Enzyme-linked Immunosorbent Assay
The levels of PGD2, IL-1 , and TNF-α were detected by following the protocol of a commercial ELISA kit (MeiMian, Jiangsu, China). Brie y, the sample, the standard, and a HRP-labeled antibody were added to microwells coated with the antibody, incubated at 37°C and washed thoroughly. The color was developed with a TMB substrate, which was converted to blue by peroxidase catalysis and nally converted to yellow by that action of acid. The intensity of the color was positively correlated with the expression of the target protein in the sample. The optical density value was measured with a microplate reader at a wavelength of 450 nm, and the sample concentration was calculated.

Statistical Evaluation
All data were reported as mean ± standard deviation. Statistical signi cance was determined by t test to compare two groups or one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test for multiple comparisons. Statistical analysis was performed by GraphPad Prism 6 (GraphPad Software, USA). P value of less than 0.05 was regarded as statistically signi cant.

Characterization of HC@HMC
The average Zeta potential of HC@HMC was about 0 mV. It can be seen from the appearance (Graphical abstract) that HC@HMC is a clear and uniformly dispersed liquid, which can preliminarily explain the successful preparation of HC@HMC.

HC@HMC Improved the Oral Absorption of HYA Intestinal Perfusion in Rats
Compared with HYA, HC@HMC showed higher absorption in total-intestinal tracts (duodenum, jejunum, ileum and colon). HYA and HC@HMC are mainly absorbed in the duodenum and colon. The determined Ka, Peff and absorption percentage (PA) values comparison between HYA and HC@HMC the same gastro-intestinal areas were further described as follows (Fig. 1A).

Pharmacokinetic of HC@HMC
The main pharmacokinetic parameters of the non-compartmental model are shown in the Fig. 1. It can be seen from Figure 1B that within 0-4h, the blood concentration of HC@HMC is signi cantly higher than that of HYA. Due to the low solubility and poor permeability of HYA, the maximum blood concentration (C max ) of HYA is only 1.25 ± 0.33 mg·L −1 , while the C max of HC@HMC is as high as 20.14 ± 4.23 mg·L −1 , which is 16.11 times that of HYA. In addition, the AUC 0−72h , T max and MRT of HC@HMC are higher than those of HYA, and all parameters except T max and MRT 0−96h are signi cantly different. The AUC 0−96 h and C max of HC@HMC are about 10 times and 16 times of HYA, respectively, indicating that HC@HMC can increase the absorption of HYA in rats. Besides, the Cl of HYA is more than double that of HC@HMC, indicating that HC@HMC can increase the residence time in rats to promote absorption. The pharmacokinetic parameters of HC@HMC calculated by the non-compartment model is superior to HYA (Fig. 1C). So, the pharmacokinetic behavior of HC@HMC is superior to HYA. HYA and HC@HMC were found to not be bio-equivalent (Tab. 2). The oral bio-availability of HC@HMC compared with HYA was 999.63%.

Intracranial Concentration of HYA in Mice
To study whether our preparation can penetrate the BBB, we used liquid chromatograph-mass spectrometer (LC/MS) to determine the intracranial drug concentration in mice after administration. The results show that both HC@HMC and HYA can penetrate the BBB, and HC@HMC penetrates the BBB more. 1 h after oral administration of HC@HMC, HC@HMC reached the highest concentration 23.46 ng·mL −1 , which was 1.15 times that of HYA (Fig. 2). It is speculated that the water-in-oil nanoemulsion improves the liposolubility, so it penetrates more through the BBB. Combining the blood drug concentration curve and the intracranial drug concentration curve, the change trend and peak time of HC@HMC are almost the same (Fig. 1B). This result shows that HC@HMC improves the absorption of intestine rst, and then more easily penetrates the BBB.

Effect of HC@HMC on Cerebral Injury Caused by MCAO/R in Mice
To determine whether HC@HMC had effects on CIRI, the TTC assay was used to detect the brain infarct size. Representative images of TTC-stained brain sections are shown in Fig. 3A. Obviously, compared with the sham group, the infarct volume was signi cantly increased in MCAO/R group. Compared with MCAO/R group, 80mg·kg −1 HC@HMC group and 160mg·kg −1 HC@HMC group and 20mg·kg −1 HYA group signi cantly decreased the infarct volume and neurological score (Fig. 3B, 3C). Compared with the sham group, the residence time of mice on the rotating rod was signi cantly reduced. Compared with MCAO/R group, 80 mg·kg −1 HC@HMC group and 160 mg·kg −1 HC@HMC group and 20 mg·kg −1 HYA group signi cantly improved the motor ability and coordination ability of mice after surgery (Fig. 3D).
To determine whether the brain had histopathological changes, HE staining was conducted. In the sham group, nerve cells were arranged tightly in order, and most cells were round or oval with large cell bodies.
HC@HMC Reduces In ammation in MCAO/R Treated-mice.
We examine whether HC@HMC had effects on COX2, DP1, DP2 expression. Compared with the sham group, the expressions of COX2, DP1, DP2 mRNA and protein were signi cantly increased in the cortex of MCAO/R-treated mice. Compared with the model group, HC@HMC markedly decreased the expression of COX2, DP1, DP2 mRNA and protein in the cortex of MCAO/R-treated mice, and HYA had the same effect.
To determine whether related products and in ammatory factors had changed, ELISA assay was used to test PGD2, TNF-α, IL-1β content in the mouse cortex. Compared with the sham group, the mice cortical PGD2, TNF-α, IL-1β content were signi cantly increased in the model group. Compared with the model group, HC@HMC signi cantly blunted the increase of mice cortical PGD2, TNF-α and IL-1β content, and HYA had the same effect (Fig. 5H,5I,5J).

Conclusion
Generally speaking, drugs for the treatment of stroke could pass through the BBB to be meaningful.
Studies have shown that the gastrointestinal absorption of HYA is poor, and oral bioavailability is low [15] . So the main application of HYA is intravenous administration. Therefore, there is an urgent need for powerful nanomaterials to overcome the problems of poor blood stability and low BBB penetration e ciency of HYA.
Here, we formed HYA complexing with phospholipid and HPCD under certain conditions, and then prepared the water-in-oil HYA complex nanoemulsion with hyaluronic acid-modi ed multi-walled carbon nanotubes and chitosan (HC@HMC) by titration. The results showed that HC@HMC signi cantly improved the absorption of HYA, thereby increasing its bioavailability, which may reduce its e ux by Pglycoprotein and improve the absorption of HYA. Because polyethylene glycol 400 acts as a Pglycoprotein (P-gp) inhibitor, which promotes the absorption of P-gp substrate HYA. We rstly prepared HYA with phospholipids and HPCD into a complex. The phospholipids could increase cell membrane uidity, which may help open the tight junctions between cells [34,35] , resulting in the enhanced permeability of hydrophilic drugs. Chitosan can also enhance drug permeability across the BBB by affecting the tight junction [36] . HPCD can improve the stability of HYA and reduce the degradation of HYA [26] , improving its bioavailability and activity. HC@HMC can signi cantly increase the C max of HYA and prolong MRT, thereby signi cantly increasing its AUC and achieving the purpose of improving the bioavailability of HYA.
In addition to increasing drug plasma concentrations, effective BBB penetration is another major challenge in stroke treatment. Combined with LC/MS results, after oral administration of HC@HMC 1 h, the highest concentration of HC@HMC was 23.46 ng/ml, which was 1.15 times of HYA. Therefore, oral administration of HC@HMC can achieve intravenous drug concentrations of HYA. We prepared HYA into oral preparations, which can improve the compliance of patients in clinical practice. The water-in-oil nanoemulsion is an effective strategy for oral delivery of highly water-soluble and low-permeability drugs, and an effective strategy of crossing the BBB.
Finally, we investigated the therapeutic effect of HC@HMC in MCAO/R-treated mice. Li sun et al found that injecting HYA (1, 5, 10 mg·kg −1 ) into the tail vein 30 minutes before surgery protects rats from focal cerebral CIRI through inhibited I/R-induced protein oxidation and nitration [37] . In the present study, by using the middle cerebral artery occlusion (MCAO) model, Lu Yu et al found that 8 mg·kg −1 and 16 mg·kg −1 HYA administration by common carotid artery (CCA) injection improved impaired cognitive function in Morris water maze(MWM) and passive avoidance tasks, suggesting that HYA treatment in a certain concentration can improve cognitive impairment in MCAO rats [38] . The recommended clinical dose of HYA is 100 mg (approximately 2 mg·kg −1 ) once a day for common cases or twice a day for severe cardiac patients [31] . Therefore, to mimic the current clinical treatment regimen, we chose to treat the Prior to the study of mechanism, 3 doses (40, 80 and 120 mg·kg −1 ) of HC@HMC were assessed according to neurological function score. The low-dose group showed no improvement of the neurological function scores. There were obvious improvements of the function scores in the two highdose groups. However, there was no signi cant inter-group difference. So we choose the 80 mg·kg −1 group to study the mechanism of HC@HMC. From our results, compared with the sham group, the expressions of COX2, DP1 and DP2 in the model group were signi cantly increased. In ammation plays an important role in the occurrence and development of central nervous system diseases [39] . COX2 is a key rate-limiting enzyme in the in ammatory pathway [40] . PGD2 was downstream products of COX2, which was abundant prostaglandins in brain [41] . Studies have found that some harmful effects of PGD2 are mainly regulated by DP2 receptors or through the PPARγ pathway [42] . Compared to the model group.
HC@HMC 80mg·kg −1 group can signi cantly reduce the expression of COX2, DP1 and DP2. This is the rst study that demonstrated the protective effect of HC@HMC on reducing ischemia-reperfusion cerebral injury through COX2/PGD2/DPs pathway. HYA has the same effect, which shows that HC@HMC and HYA have the same mechanism. However, it is not clear whether HYA exerts its protective effect through the drug substance or its metabolites, and further research is needed. Studies have found that HYA has 8 main metabolites in rats, but we do not know which one plays an important role [43] . I believe that there will be more high-quality studies to con rm it in the future, which will provide more possibilities for the treatment and prognosis of ischemic stroke risk groups.
In summary, we developed an effective strategy to survive HYA through the BBB with good stability, which ameliorated cerebral ischemia-reperfusion injury in mice. These results indicate that our HC@HMC nanomedicine has good clinical therapeutic potential for stroke owing to ease of formulation, stability and BBB permeability.

Declarations
Ethics approval and consent to participate This experiment has been approved by the Ethics Committee of Chongqing Medical University. All experiments in this study were consistent with the National Institute of Health Guide for the Care and Use of Laboratory Animals.

Consent for publication
Not applicable.

Availability of data and material
All data generated or analysed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests.

FUNDING
This work was nancially supported by the Chongqing Education Committee (CYS20212), Chongqing Science and Technology Committee (cstc2017shmsA130028), and China Postdoctoral Science  showing kinetic characteristics. The data were shown as mean ± standard deviation. n= 6, *P<0.05 or **P<0.01 indicated signi cant differences or very signi cant differences between free HYA and HC@HMC.

Figure 2
Intracranial concentration of HYA in mice. Line graph of drug concentration at 0.5h, 1h, and 1.5h. Data are presented as the mean ± SD, n=3.