Open hepatic artery flow with portal vein clamping protects against bile duct injury compared to pringles maneuver

Abstract Background Conventional hepatic artery and portal vein clamping strategies can prevent blood loss and ischemia-reperfusion liver injury, and such preventative measures are the key to successful liver surgery. However, ischemic-induced damage to cholangiocytes is rarely considered. Here, we aimed to investigate the effect of different hepatic inflow interruption methods on bile duct injury. Methods Forty rats were randomly grouped as sham, Pringle maneuver (PM) and hepatic arterial blood flow open (HAFO) groups. We evaluated liver histology and function in liver sections, and biliary histology, cholangiocyte apoptosis and proliferation, cytokine production, and bile composition. RNA sequencing is performed to explore possible molecular mechanisms. The Blood-biliary barrier permeability and tight junctions were analyzed by HRP injection, immunofluorescence staining and analysis of ZO-1 expression by immunoblotting. Results HAFO significantly attenuated ischemia-induced liver injury and decreased ALT, ALP, TBIL, and DBIL levels in serum. The histopathological observations showed that bile duct injury in the PM group was more serious than that in the HAFO group. The numbers of apoptotic biliary epithelial cells in HAFO-treated rat bile ducta were lower than those in the PM group. RNA-seq showed that tight junctions may be related to the mechanism underlying the protective effect of HAFO, as shown by the reduced HRP levels and increased ZO-1 and claudin-1/3 expression in the HAFO group compared to the PM group. Conclusion Compared with PM, HAFO alleviated the ischemic injury to the biliary system, which was characterized by decreased biliary epithelial cell apoptosis, reduced inflammatory responses and decreased blood-biliary-barrier permeability.


Introduction
Liver resection is the main approach for the treatment of benign and malignant liver lesions. The control of blood loss is an important consideration during liver surgery and major blood loss is usually related to postoperative morbidity and mortality [1][2][3]. The Pringle maneuver, which involves continuously or intermittently clamping the hepatic artery (HA) and portal vein (PV), has been widely used to reduce blood loss during liver transection [4,5]. However, occlusion, especially continuous and long-term occlusion, of the inflow to the liver inevitably results in liver ischemia followed by metabolic, inflammatory, microvascular events, and even liver failure [6,7]. The intermittent pringle maneuver has been accepted as a technique that can be used to balance the control of blood loss and protection of liver function by alternately interrupting and opening blood inflow to the liver. Long-term occlusion of the PV and HA is frequently indicated to prevent major bleeding during complicated liver surgery. The balance of the control of blood loss and protection of liver function is much more significant and difficult to maintain in complicated cases.
The liver has a dual supply of blood from the HA and PV. Cholangiocytes solely or mainly rely on the blood of the peribiliary vascular plexus that originates from the hepatic artery [8], while hepatocytes can be supplied by blood from the PV. The resistance of cholangiocytes to ischemia is remarkably inferior to hepatocytes [9]. Biliary complications after liver transplantation, which are related to warm and cold ischemia of the bile duct, seriously impair graft function and even lead to graft loss [10,11]. Considering the close relationship between cholangiocytes and hepatocytes, it is reasonable to hypothesize that the status of the liver that are subjected to ischemia is more reliant on the cholangiocytes rather than hepatocytes. Previous studies have shown that compared to the Pringle maneuver (PM), continuous occlusion of the portal vein alone and keeping the hepatic artery flow open (HAFO) can substantially attenuate ischemic injury to the liver without increasing blood loss [12]. However, no studies have focused on the effect of HAFO and PM on bile duct damage.

Objectives
The present study was designed to investigate whether HAFO exerts a potential protective effect on cholangiocytes and liver function compared with the conventional Pringle maneuver.

Animals
Male Sprague Dawley rats (180-220 g, 8-10 weeks old) were obtained from the Animal Experiment Center of Kunming Medical University. All rats were housed under conditions of 55%-60% humidity, 22-24 C, 12 h light/12 h dark cycle, and the rats were provided with commercial rodent chow and water ad libitum. The Animal Care and Use Committee of Kunming Medical University approved the study protocol (License: kmmu2021058). Euthanasia of animals was carried out according to the Chinese Code for the Care and Use of Animals for Scientific Purposes 8th edition.

Animal grouping
All rats were randomly divided into five groups, and the sample size was calculated by PASS (version 15

Surgical procedures
Rats were fasted overnight before operation and had free access to water. Anesthesia was carried out the intraperitoneal injection of 10% chloral hydrate at a dose of 3 mL/kg. All operational steps were taken with microsurgery instruments (Shanghai Medical Instruments, Shanghai, China). In all groups, after a midline laparotomy, the ligaments around the liver were divided, the hepatic portal triad was removed, and collateral vessels to the hepatic hilum were disconnected. The HA and PV were meticulously freed individually. In the sham group, no further steps were performed except for the closing abdomen. In the CPM group, the hepatic portal triad was occluded for 45 min with two microvascular clamps. In the IPM group, the hepatic portal triad was clamped for 15 min and then reperfused for 5 min, for a total of 3 cycles. In the CHAFO group, after being meticulously with microsurgical instruments, the PV was clamped with a microvascular clamp and the HA were kept open for 45 min. In the IHAFO group, the portal vein were clamped for 45 min, while the hepatic artery was clamped for 5 min and re-open for 15 min, and these steps were alternate for 45 min. After completing the steps relevant to at the HA and PV in CPM, IPM, CHAFO and IHAFO groups, the liver perfusion was restored and the abdomen was closed.
Immediately after the operation, the rats were treated with a subcutaneous injection of cefuroxime sodium (16 mg/kg) and buprenorphine (0.1 mg/kg) in a total of 1.5 mL of normal saline solution. The rat were allowed to recover for 60 min in a special intensive care unit cage with warmed air (30-35 C) and an oxygen supply. Then, the rats were move to a normal cage, and access to water and food was provide ad libitum. Twenty-four hours after the operation, a relaparotomy was performed. A polyethylene tube (PE-10, 0.28 mm inner diameter, 0.64 mm outer diameter, American Health & Medical Supply International Corp. Co. Ltd., NY, USA) was implanted into the common bile duct near the duodenum to collect extra drained bile.
Euthanasia of animals was carried out according to the Chinese Code for the Care and Use of Animals for Scientific Purposes 8th edition. All the animals were immediately euthanized by carotid exsanguination just after the loss of consciousness (2 min). Blood was collected in EDTA tubes (4 ml/rat), centrifuged at 4 C to obtain serum plasma and stored at À80 C. The extrahepatic bile ducts were harvested proximal to the hilar and 0.5 cm above the cannula with a micro scissor. A portion of the ductal tissues was placed in liquid nitrogen for sequencing, and the remaining ductal tissues were place in 2.5% glutaraldehyde in cacodylate buffer and 1% OsO 4 for electron microscopy. The middle lobe liver was fixed in 4% neutral buffered formaldehyde for histological studies, and the others part were placed in liquid nitrogen for molecular examination.

The array of cytokines in bile
The concentrations of Interleukin-6 (IL-6), Tumor necrosis factor-a (TNF-a) and Monocyte Chemoattractant Protein-1 (MCP-1) in the bile that was collected from a draining tube were detected through a QuantibodyV R multiplex ELISA array (QAR-CYT-2; RayBiotech, Norcross, GA, USA).

Assessment of blood-biliary-barrier
The permeability of the blood-biliary barrier was evaluated by horseradish peroxidase (HRP) (Sigma-Aldrich, St. Louis, MO, USA) followed by techniques described by Takakuwae [13]. Briefly, sterile water containing 1000 U of HRP was injected into the inferior vena cava 30 min after bile duct cannulation. Bile was collected for an additional 10 min, and HRP activity was assayed by using the Amplex TM Red Hydrogen Peroxide/Peroxidase Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA).

Histological analysis, immunohistochemistry and immunofluorescence staining
Liver tissues were fixed with 10% formalin for 24 h at room temperature, then embedded in paraffin. Sections were cut into 4 lm thickness and prepared for hematoxylin&eosin, immunohistochemistry and immunofluorescence staining. The detailed procedure was carried out as previously described [14]. Briefly, the sections were dewaxed and hydrated by xylene and alcohol (70%, 90%, 100%, v/v) at room temperature. The sections were then stained with hematoxylin for 5 min and then eosin for 1 min (Beyotime, China) at room temperature. The hematoxylin-eosin (H&E)stained paraffin-embedded liver sections were observed in light microscopy (DM6000B, Leica, Germany). The bile duct was histologically assessed by employing the method of bile duct injury severity score (BDISS) that was introduced by Genken et al. [15]. Two independent expert pathologists who were unaware of group assignments performed the scoring.
Immunohistochemistry was performed in 4 lm thick sections. The sections were deparaffinized and endogenous peroxidase activity was blocked by a 30-min incubation in methanolic hydrogen peroxide (2.5%). Later, the endogenous biotin was blocked by a biotin blocking system (MXB Biotechnologies, China) according to the manufacturer's instructions. The sections were then hydrated in graded alcohol and rinsed in 1Â PBS (pH 7.4) before applying the selected primary antibody. Sections were incubated overnight at 4 C with primary antibodies including rabbit polyclonal anti-Ki67 (1:200, 27309, Proteintech, China), rabbit polyclonal anti-caspase 3 (1:200, 19677, Proteintech, China). The following day, samples were rinsed with PBS for 5 min, then followed the manufacturer's instructions of using Maxvision TM 2 HRP-Polymer anti-Mouse/Rabbit IHC Kit (MXB Biotechnologies, China). Negative controls were performed for all immunoreactions to confirm the specificity of immunoreaction. At least 10 different portal areas (from 3 different sections) were evaluated. The intrahepatic bile duct was evaluated using light microscopy (DM6000B, Leica, Germany).

Transmission electron microscope (TEM) evaluation
The fixed extrahepatic bile duct was infiltrated with acetonearaldite and embedded in Araldite. The ductal tissue was cut into 80 nm thin slices and treated with uranyl acetate and lead nitrate. The mitochondria of ductal epithelial cells and tight conjunctions between them were observed under the transmission electron microscope (JEOL Ltd., Japan). TEM evaluation was carried out as Fickert described [16].

Observation of apoptosis of biliary epithelial cell
Formaldehyde-fixed paraffin-embedded liver sections of 4 lm thickness were made for observation of apoptosis of biliary epithelial cells. A quantitative terminal deoxynucleotidyl transferase dUTP-mediated nick-end labeling (TUNEL) assay was used to detect DNA fragmentation in apoptotic cells (TUNEL, #S7100, Merck Millipore, Billerica, MA, USA). The average number of TUNEL positive cells per field was counted in five random, non-overlapping fields at 200Â magnification using Image J image analysis software. The number of TUNEL-positive biliary epithelial cells per field was expressed as a percentage of total biliary epithelial cells.

Transcriptome sequencing and bioinformatics analysis
RNA sequencing of nine rats extrahepatic bile ducts (3 cases in each group) and the data analyses were performed by Wuhan Benagen Tech Solution Co., Ltd. (China). The sequencing data can be provided by the corresponding author upon reasonable request.

Real-time PCR
Total RNA was extracted from 100 mg liver tissue using an RNA extraction kit (DP451, Tiangen, China) and reverse transcribed into complementary DNA using a cDNA Synthesis Kit (KR118, Tiangen, China). Real-time quantitative polymerase chain reaction (RT-qPCR) was performed using SYBR Green I Mix (FP215, Tiangen, China) in the ABI PRISM7500 Sequence Detection System (Applied Biosystems, Life Technologies Corp., CA, USA). The reaction condition for PCR was set as follows: 15 min at 95 C predegeneration, 40 cycles of 95 C (10 s) and 60 C (20 s), then melting/Dissociation Curve Stage. b-actin was used as a reference gene. The relative mRNA expression was calculated by the inverse log of DDCt. All experiments were performed in triplicate. Primer sequences (see Table S1) were synthesized by Invitrogen.

Statistical analysis
Data are presented as mean ± standard deviation. Statistical analysis was performed using SPSS 22.0 software (SPSS, Chicago, IL) and GraphPad Prism 7 software (GraphPad, San Diego, CA, USA). Data between multiple groups is calculated by two-way analysis of variance. The comparison between the pairs is implemented by the Bonferroni test. p < 0.05 was considered statistically significant.

Effects of various hepatic blood inflow occlusion on liver function
The ALT, AST, TBIL, DBIL, ALP, and GGT levels significantly increased in all the groups except for the sham group. The rise increase in the six items was most notable in the CPM group, and these markers were slightly elevated in the other groups ( Figure 1). In IHAFO and CHAFO groups, the degree of increase was the mildest and there was no significant difference between the two groups ( Figure 1(A-F)). We also analyzed hepatocellular necrosis and we observed more cell necrosis in the CPM group, but not in the other groups compared with the sham group. No significant differences were found in the mean values of vascular congestion, and pyknotic nuclei among the four groups (Figure 1(G)). These findings suggest that the animals in the IPM, CHAFO and IHAFO groups are experienced a less ischemic injury to the liver than the CPM-treated rats.

Effect of various hepatic blood inflow occlusion on bile duct injury
The ducts were affected and representative images of each group were shown in Figure 2(A). Severe bile duct injury occurred in the PM-treated animals and was weak in HAFOtreated animals in the ischemia injury. The number of edematous, necrotic and deciduous biliary epithelial cells in the HAFO groups was significantly decreased compared to the PM group at 24 h (Figure 2(A)). In the CPM group, necrotic cholangiocytes and disrupted tight junctions were observed under TEM. Compared with those in the HAFO group, more extensive and more serious mitochondria swelling in cholangiocytes and tight junction rupture were observed in the IPM group (Figure 2(B)). The BDISS scores of bile duct histology of the HAFO group were significantly lower than those of the PM group (p < .05, Figure 2(C)). In addition to the histological scoring, LDH leakage into bile was measured as a biomarker of biliary injury. PM induced a transient increase in bile LDH levels in the following 24 h of reperfusion, and no significant difference was found between the HAFO and sham groups (Figure 2(D)). Overall, more serious ischemic injury to the bile duct were noted in the PM (either CPM or IPM) groups than in the HAFO (either IHAFO or CHAFO) groups.

Effect of various hepatic blood inflow occlusion on cholangiocyte apoptosis and proliferation
TUNEL staining was used to investigate the apoptosis and necrosis of the liver and cholangiocytes (Figure 3(A)). In all the groups, only a small amount of hepatocyte apoptosis was observed in the CPM groups, and no obvious hepatocyte necrosis was found in the other groups. However, in the bile duct tissue, each group had a different number of TUNELpositive cholangiocytes. Cholangiocyte apoptosis in the PM groups was significantly more than that in the HAFO group. There was no significant difference between the CHAFO and IHAFO groups (Figure 3(D)). Similarly, the number of active caspase-3-positive bile duct cells in the HAFO group was much lower than that in the PM groups at 24 h after the operation (p < .05) (Figure 3(B,E)). Compared to the Sham group, PM and HAFO increased the proliferation of cholangiocytes as evidenced by increased Ki67 expression in bile duct sections. There was a significant increase in the number of proliferating cholangiocytes in the bile duct of PM rats compared with that of HAFO rats. (Figure 3(C,F)).
The changes in histology, apoptosis, immunohistochemical staining, and TME indicated that preservation of arterial blood supply could substantially reduce bile duct ischemia injury.

Bile composition and inflammation in rats subjected to various hepatic blood inflow occlusion
There were no changes in bile flow, total bile salt, or phospholipid output following occlusion in the different groups ( Figure 4(A-C)). To determine the inflammatory response of the liver and bile ducts to different types of hepatic blood inflow occlusion, cholangiocytes producing cytokines such as MCP-1, IL-6, and TNF-a levels in bile were analyzed. After treatment, bile concentrations of MCP1 and TNF-a were significantly increased, while IL-6 decreased (Figure 4(D-F)).

Blood-biliary-barrier permeability and tight junctions in groups subjected to various hepatic blood inflow occlusion
To investigate whether BBB was dysfunctional, we used HRP to measure the blood-biliary barrier permeability. Ischemia caused biliary injury with an increased level of HRP in the PM group, but the HRP levels did not increase in the HAFO group compared with the Sham group ( Figure 5(A)). Immunofluorescence co-staining for CK 19 and ZO-1 showed that ZO-1 was present at the junctions of neighboring cholangiocytes in the sham group that did not develop biliary injury ( Figure 5(B)). Animals with different biliary injuries in the other groups showed markedly altered tight junction morphology with an irregular ZO-1 staining pattern that was not confined to the junctions of neighboring cholangiocytes ( Figure 5(B)). Further, we evaluated the change in tight junction-related factors. We found that the mRNA and protein levels of tight junction (TJ)-related genes (Claudin-1, Claudin-3 and ZO-1) were significantly lower in the PM group than in the sham and HAFO groups ( Figure 5(C)). The data suggest that the HAFO has less impact on tight junctions, which contribute to biliary permeability and bile duct injury.

Transcriptome sequencing analysis
We performed RNA-sequencing analysis to identify the differentially expressed genes (DEGs) in isolated bile duct tissue (extrahepatic bile duct) ( Figure 6(A)). A total of 1882 DEGs were identified between the CPM and sham groups, while 2133 DEGs were identified between CHAFO and sham group. Among the 1362 DEGs between the CPM and CHAFO groups ( Figure 6(B,C)), 50 co-DEGs were identified (Fold Change > 1.5, p < .05; Table S2). KEGG results showed that 22 pathways were enriched in the sham, CPM and CHAFO groups. CPM and CHAFO had significant differences in 26 GO terms and 22 metabolic pathways, such as Retinol metabolism, Ascorbate and aldarate metabolism, Steroid hormone biosynthesis, Histidine metabolism, Arachidonic acid metabolism, Glycine, serine and threonine metabolism, and Tryptophan metabolism ( Figure 6(D), Table S3).

Discussion
The liver receives blood supply from both the PV and HA, but the bile duct predominantly relies on blood from the peribiliary vascular plexus originates from the HA [17]. Cholangiocytes are highly susceptible to ischemia due to the clamping of the HA. Although a variety of strategies have been developed to reduce the impact of I/R on liver injury, such as PM (continuous or intermittent), the impact on bile duct injury has not been evaluated. In this study, we compared the effects of PM and HAFO on intrahepatic bile duct damage in rats. Our data show that PM (and to a lesser extent HAFO) induced functional damage of bile ducts as demonstrated by increased apoptosis of bile ducts in liver sections; changes in bile composition, and increased bloodbiliary-barrier permeability. Further analysis of transcriptome RNA-sequencing revealed that the differentially expressed genes between the PM and HAFO groups were mainly enriched in immunity, endothelial adhesion and tight junctions. Immunoblotting analysis showed that PM-induced damage to bile ducts was coupled with decreased expression of ZO-1, claudin-1, and claudin-3 in bile ducts compared with sham rats.
In the study, IPM and HAFO caused less damage to the liver than CPM, which is consistent with the previous study [12]. Despite the lack of large-scale clinical application, previous studies have confirmed that continuously blocking PV and preserving HA blood flow can alleviate liver injury related to IR, and does not significantly increase blood loss during hepatectomy in a rat model [12,18]. However, no attention has been given to damage to the bile duct in these studies. Our study demonstrates that compared to PMtreated rats, HAFO-treated rats have significantly mild ischemia-related cholangiocyte injury. Additionally, the results of our investigation showed that the degree of liver injury was not markably different between the rats that were treated with the two differently HAFO methods (continuous or intermittent). The mechanism underlying liver ischemia injury is complex and involves multiple factors. Cholangiocyte that are subjected to Ischemia may be involved in the pathogenesis of biliary complications after liver transplantation [19]. In our study, H&E, TUNEL and TEM images showed that compared with PM, HAFO can significantly reduce cholangiocyte ischemia damage and apoptosis.
In the current study, the liver injury seemed to be alleviated as the bile duct damage is reduced. This phenomenon may be complex and multifactorial. First, cholangiocytes are the primary target of injury after hepatic arterial ischemia and cholangiocytes are more susceptible to the interruption of arterial flow than hepatocytes [20]. Second, inflammation of the bile duct results in hepatocyte damage through a complex process. The retention of bile acid and bilirubin is toxic to liver cells, inhibits the respiratory chain, and reduces ATP synthesis [21]. The toxicity caused by the retention of bile acid can also change the permeability of the mitochondrial membrane and cause the mitochondria to lose their function [22]. Bile duct damage leads to increased bile duct permeability, bile exudation, inflammatory cell infiltration, and inflammation of the portal vein [23]. Our results showed that cholangiocyte damage may be the initial manifestation and initial cause of liver injury within a certain clamp time.
The composition of bile did not significantly change in this experiment, which may be due to the lower toxicity of bile in rats than in humans [24,25]. MCP-1 was proven to be related to cholangitis and cholestatic inflammation [26,27]. TNF-a and IL-6 are the key factors for cholestasis-induced cell death, and fibrosis [22,23]. Our study also supported the findings of Cursio et al. that ischemic injury interacts with cholangitis and further aggravates bile duct damage [28].
There were significant differences in the bile HRP levels among the sham, PM and HAFO groups, with the highest level in the CPM group and the lowest level in the sham group. HRP enters the bile through two different ways, paracellular and transcytosis in the former case, HRP passes through the tight junctions of cells. ZO-1 staining also showed that it was most irregular in the CPM group among the five groups, and there was no significant difference between the IPM and HAFO groups. These findings indicated that HAFO may reduce the damage to the blood-biliary barrier, and blocking the hepatic artery in a short time may not result in obvious damage to the blood-biliary-barrier. Tight junctions form a physical seal between cells, preventing water and solutes from passing through the paracellular spaces, and disruptions of cholangiocyte tight junction function disorder contributed to blood-biliary-barrier dysfunction [29,30]. We further verified three genes from RNAseq data, ZO-1, claudin-1 and claudin-3. Among these, claudin-3 may be closely related to bile duct injury. A study reported that claudin-3 is absent in the normal liver, but it is expressed in the biliary, such as in intrahepatic and extrahepatic carcinomas and papillary neoplasms [31]. Thus, the fact that claudin-3 expression was detected by western blotting can be considered a reflection of the tight junctions between cholangiocytes. Our results showed that ZO-1, claudin-1 and claudin-3 were down-regulated in PM and HAFO-treated rats compared with sham rats, and this downregulation was more obvious in the PM group. No significant differences in the three TJ-related protein levels were observed between the continuous-or intermittent-HAFO. This implies that HAFO may maintain the blood-biliary-barrier function and reduce bile duct I/R injury by regulating TJ-related proteins. In short, cholangiocyte apoptosis, cholestasis, impaired blood biliary barrier and inflammation interact with each other and contribute to bile duct damage. Compared with PM, HAFO can effectively exert a better mitigation effect.
Finally, we performed next-generation transcriptome sequencing to explore the mechanism underlying bile duct injury due to various forms of hepatic blood inflow occlusion.
A large number of differential expressed genes were identified and most of them significantly enriched in metabolism pathways, such as Tryptophan metabolism, Pyruvate metabolism, Arginine and proline metabolism, Ascorbate and aldarate metabolism, et al. Retinol metabolism [32] and Tryptophan Metabolism [14] has been reported to be associated with Cholestasis, may be induced or aggravated the bile duct injury. Several studies showed biliary obstruction and subsequent internal biliary bile change may affect cytochrome P450 isozymes [33,34]. Consistently, our study also identified the changes in the Drug metabolism-cytochrome P450 pathway.

Limitations
There are several potential limitations in our study: (1) The size samples was small. (2) The duration of bile duct injury was short, and further work is required to determine the long-term effects of HAFO administration. (3) It is hard to separate intrahepatic cholangiocytes from other liver cells, thus extrahepatic cholangiocytes were used for transcriptional analyses. This difference may introduce bias in the transcriptome information.

Conclusions
In summary, within the context of certain blood flow occlusion, hepatic artery flow opening can substantially decrease bile duct injury compared to the conventional Pringle maneuver. In this experiment, blocking hepatic arterial blood flow for a short time had a slightly negative effect on cholangiocytes. Hepatic artery flow opening attenuated ischemic injury to cholangiocytes by attenuating cholestasis, mitigating injury to cholangiocytes' tight junction and the blood-biliary-barrier, reducing inflammation-related cytokine production, and affecting the expression of specific signaling pathway-related molecules.
These findings may provide guidance for future clinical studies. It appears that clamping the portal vein while preserving hepatic artery flow provides better bile duct cytoprotection than total hepatic blood inflow occlusion, and does not exacerbate liver injury. This approach may reduce the risk of adverse effects caused by bile duct damage, including infection and blocking stasis after major liver surgery (such as liver transplantation or resection). These findings have clinical significance because they are easy and safe to translate into practice.

Author contributions
All authors contributed to this study and participated in the writing and drafting process as well as the critical review the manuscript. YB and LD conceived the work; BPD, WM designed the experiments; ZSL, CPL performed the experiments and analyzed the data; ZSL, YB and LD significantly contributed to the writing process. All authors have given the final approval of the version to be published and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethical approval
The Animal Care and Use Committee of Kunming Medical University approved the study protocol (License: kmmu2021058). Euthanasia of animals was carried out according to the Chinese Code for the Care and Use of Animals for Scientific Purposes 8th edition.