Hepatoprotective Effect of Yiqijianpi Formula (YQJPF) on Liver Failure Through Modulation of Hypoxic and Apoptosis Pathways

Li Tang Nanjing University of Chinese Medicine Feixia Wang NANJING UNIVERSITY OF CHINESE MEDICINE Lingyan Xiao The Second A liated Hospital of Nanjing University of Chinese Medicine Min Shen Nanjing University of Chinese Medicine Siwei Xia Nanjing University of Chinese Medicine Zili Zhang Nanjing University of Chinese Medicine Hanzhongmen Campus Feng Zhang Nanjing University of Chinese Medicine Shizhong Zheng Nanjing University of Chinese Medicine Shanzhong Tan (  20195183@njucm.edu.cn ) Nanjing hospital a liated to Nanjing university of Chinese Medicine


Introduction
Liver failure is the critical condition, the incidence of the disease has an increasing tendency with alcohol abuse or alcoholism, and growing epidemic of obesity, diabetes, which have long plagued the medical profession [[i]] , [[ii]] . Liver failure mainly concludes acute liver failure (ALF), acute-on chronic liver failure (ACLF), or an acute decompensation of an end-stage liver disease, in some cases they have some common pathological process, liver cell damage runs through the entire course of disease development [[iii]]. Among them ACLF is a syndrome of hepatic decompensation [[iv] , [v]] , although there are existed some disputes on its clinical de nition, in general, it is a distinct entity where acute damage to liver function on the basis of chronic liver disease or cirrhosis patient, which could cause acute hepatic decompensation with a high short-term mortality. It seriously threatens the lives of patients with chronic liver disease [[vi] , [vii]] .
However, no effective therapy for liver failure beyond supportive treatment is currently available, as such, the search for a new treatment or medicine against liver failure is still needed.
Traditional Chinese Medicine (TCM) has been widely used for the treatment of liver failure with a long history. Yiqijianpi formula(YQJPF) is used to treat ACLF patients based on the TCM theory of 'Strengthening body resistance', it composes 9 herbs: Astragalus membranaceus (Huang Qi), and Radix pseudostellariae This study we investigated the effect of YQJPF on hepatocyte in vitro and in vivo, and a network pharmacology analysis [[xvii]] was performed to explore the potential molecular mechanisms. (Figure abstract).

YQJPF decoction Preparation
The YQJPF decoction contained 30 g of Astragalus membranaceus, 30 g of Radix pseudostellariae, 30 g of Angelica sinensis, 10g of Fructus Ligustri Lucidi, 15 g of Poria Wol poria extensa, 15 g of Atractylodes macrocephala, 3 g of Pericarpium citri reticulatae, 10g of Scutellaria baicalensis, 10 g of Glycyrrhiza uralensis. All the herbs were obtained from Nanjing Hospital A liated to Nanjing University of Chinese Medicine ( Nanjing, China ). HPLC analysis was performed against a known herbal standard further verifying herb identity and its active components. Each formula was resuspended in deionized water and incubated at 80 °C for 30 min. The supernatant was separated from any insoluble material by centrifugation (3000 g, 20 min) followed by passage through a 0.22 μm lter. The concentration of herbal supernatant was based on the dry weight of herb per unit volume. For the in vitro cell culture studies, 28.8 mg/mL YQJPF extract was prepared, while 2.86 g/mL YQJPF was used for the in vivo animal experiments.

Establish ACLF rat model
Sprague-Dawley rats (males, 180-220g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). The rats were housed at 24±1•C under a 12-h light/dark cycle with free access to food and water. The rats were randomly assigned to two groups, the control group (Control,n=6) and the model group (Model, n=24). The rats in the Model group were intraperitoneally injected with 50% CCl 4 in an olive oil solution at 1mL/100g body weight twice a week for 12th weeks. At the end of the eighth week, the rats in the CCl 4 -treated group were randomly divided into four groups: the ACLF model group (ACLF Model, n=6), the low-dose YQJPF group (YQJPF 14.3g/kg, n=6), the high-dose YQJPF group (YQJPF 28.6g/kg, n=6) and methylprednisolone treatment group (MP 15mg/kg, n=6). The methylprednisolone group was given methylprednisolone 15mg/kg by intravenous injection. The rats in Control group were administered with the corresponding amount of saline. At the end of the 12th week, LPS (100 ug/kg) and D-Gal (400 mg/kg) were used to induce acute liver injury except the Control group. All rats were sacri ced when 24 hours after acute injury, and the liver samples were collected for subsequent research and analysis.

Liver histopathological analysis
Rat liver tissues were collected after sacri ce, then weighed before sectioned into small pieces, xed them in 4% paraformaldehyde and subjected to para n embedding for haematoxylin and eosin (H&E) staining.

Biochemistry analysis for liver injury
The contents of AST and ALT in liver tissues were measured according to the instructions of the corresponding Alanine aminotransferase Assay Kit and Aspartate aminotransferase Assay Kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

Cell Culture
Human L0 2 cell line was brought from Cell Bank of Chinese Academy of Sciences (Shanghai, China). Cells were cultured in DMEM with 10% FBS, 1% antibiotics, and incubated in a 5% CO2 and 95% air humidi ed atmosphere at 37 °C.

Cell apoptosis detection
Cell apoptosis was detected by the Apoptotic/Necrotic Cell Detection Kit (BestBio, Shanghai, China). The blue uorescence density was measured by uorescence microscope (Nikon, Tokyo, Japan).

Analysis of YQJPF active compounds and disease targets
The active compounds in YQJPF were obtained from two distinctive TCM databases, namely Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, http://lsp.nwu.edu.cn/tcmsp.php, ver. 2.3) and Traditional Chinese Medicine Integrated Database (TCMID, http://www.megabionet.org/tcmid/). The pharmacokinetic and pharmacodynamic parameters should meet the following criteria: oral bioavailability (OB) ≥30% and drug-likeness (DL) ≥ 0.18 in TCMSP. Because Glycyrrhiza uralensis was used as an adjuvant in the YQJPF, the candidate compounds in it should meet the criteria of OB ≥50% and DL≥0.18. In addition, combined with HPLC, Chinese National Knowledge Infrastructure (CNKI), Pubmed database, some chemical components that failed to be meet the criteria, but reported to have biological and pharmacological effects were also included as candidate active components for further analysis.
The corresponding targets of these compounds were searched in the TCMSP and Swiss Target Prediction (http://www.swisstargetprediction.ch/) databases. The gene symbols were obtained from the UniProtKB database (http://www.uniprot.org) with the selected species as Homo sapiens. After deletion of redundant items, the nal herb-active compounds-targets network graphs were constructed and visualized using Cytoscape v3.7.1.
Liver failure targets were predicted and screened using the GEO database, two gene expression pro les, GSE14668 and GSE38941, which contained microarrays of liver tissue samples from liver failure patients and normal liver donor tissue samples, were chosen for further analysis. The data in these two pro les were obtained from the GPL570 [HG-U133_Plus_2] (Affymetrix Human Genome U133A array) platform and were initially analyzed with GEO2R. A foldchange was regarded as signi cant with P<0.05 and with threshold values>1.0 or <−1.0. The co-expression of differentially expressed genes (DEGs) were identi ed to establish the liver failure-related targets data.
2.8.2 Compound-Target-Disease Network Construction and Enrichment of KEGG/GO Pathways Venny 2.1.04 (Venny, 2018) was used to screen for common targets between YQJPF and disease-related targets. Then we inputted the intersected gene symbols into String (Search Tool for the Retrieval of Interacting Genes/Proteins, Version: 10.5, http://string-db.org/), which was used to search for and predict protein interactions. The analysis results were imported into Cytoscape 3.7.1 to analyze and draw the Protein-Protein Interaction (PPI) Network. Hub targets were ranked by degree values to construct a hub gene targets network and a compound-target network for YQJPF on liver failure.
In order to further explain the potential mechanism of YQJPF Decoction in treating liver failure. Gene ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway were performed using Metascape.

Immunohistochemistry
Liver tissues were xed with 10% paraformaldehyde for 12h and then embedded in para n. Embedded para n sections were de-waxed in xylene and rehydrated in ethanol. Antigen retrieval was performed in 0.01M citrate buffer (pH 6.0) using a pressure cooker followed by incubation for 3 min. Samples were then washed thrice with PBS and xed in 95% ethanol for 30 min. Immunohistochemical staining was performed using antibodies against HIF-1α, VEGF-A as previously described.

Statistical Analysis
Data were presented as the mean ± SD. The signi cance of results was determined based on one-way ANOVA analysis using Prism 8.0.1 (Graphpad, San Diego, CA, USA). P< 0.05 was considered signi cant. All experiments were repeated at least three times.

YQJPF preserved liver failure in vivo
Establish ACLF model on SD rat as previously described [[i]]. The pathological changes in liver tissues were evaluated with H&E staining. The livers were rosy and smooth with intact lobule structure in the Control group. However, the livers in the ACLF Model group became small and hard accompanied by necrosis, with small nodules on their surface (Fig. 1A). As shown in Fig. 1B, the liver tissue of rats in the Control group was well-organized and the liver cells were intact. In the ACLF Model group, the liver tissue structure was disordered, and the brous tissue was obviously proliferated, accompanied by a large number of in ammatory cell in ltration, pseudolobule formation, large and sub-large hepatocyte necrosis, vacuolation, dilation and congestion of hepatic sinusoids. The liver tissue brosis in the YQJPF group and Methylprednisolone group was improved, the number of in ammatory cell in ltration and pseudolobule formation were reduced compared with the Model group, so was the hepatocyte necrosis and vacuolation.
As shown in Masson staining (Fig. 1C), in the Control group, only a small amount of collagen staining was seen on the blood vessel wall. The collagen deposition in the ACLF Model group was signi cantly increased. Collagen bers were found to form laments or even thick cords around the sinuses. The structure of liver lobules was disordered and brous tissue proliferation was also seen. After YQJPF and methylprednisolone treatment, collagen deposition and collagen volume fraction were signi cantly improved compared with the ACLF Model group .
Sirius staining results showed that the liver tissues in the Model group were severely brotic (Fig. 1D), and there was a large amount of red collagen-like brous material deposited around the liver tissue portal area. Compared with the Model group, the red ber-like material around the liver tissue portal area, and the bold duct of the treatment group were signi cantly reduced after YQJPF and methylprednisolone treatment.
Levels of AST and ALT were detected to assess the damage of hepatocytes. Liver tissue levels of ALT and AST were increased signi cantly in the Model group. The levels of these enzymes were decreased in YQJPF treatment group and methylprednisolone group. Thus, YQJPF exerted a protective effect on liver damage in ACLF.

YQJPF impress Hepatocyte proliferation and apoptosis in Liver Injury
In order to further explain the mechanism of YQJPF, we selected L0 2 human hepatocyte cell line for further analysis, LPS (4 μg/ml) induced to establish a liver injury model, and then YQJPF (10μg/ml, 20μg/ml, 40μg/ml) intervent for 24 hours to verify the effect. Cell Counting Kit-8 analysis showed that YQJPF could dose-dependently attenuates cell viability ( Fig. 2A). Apoptosis morphology was detected with uorescence microscopy, it implied that LPS induce apoptosis in hepatocytes and YQJPF could inhibit apoptosis (Fig. 2B). In the LPS-induced Model group, the expression of Bax was increased, while the production of Bcl-2 was dramatically reduced compared to that in the Control group. YQJPF activates Bcl2, reduces BAX, and shifts the BAX/Bcl2 ratio in a anti-apoptotic direction (Fig. 2D). CCK-8 analysis indicated that YQJPF 20μg/ml have the analogous effect as apoptosis inhibitor Z-VAD-FMK to restore cell viability (Fig.  2C). Overall YQJPF could promote hepatocyte proliferation, and in uence the apoptosis.

Screening of Active Components in YQJPF
The ngerprint of the water extract of YQJPF was performed by high performance liquid chromatography (HPLC) (Fig.1A).  Table S1).

The Potential Targets of Active Compounds in YQJPF
The active compound targets were searched via the TCMSP and Swiss Target Prediction databases for each chemical component. The targets were transformed using the UniProt knowledge database, data were merged to obtain gene symbols. After eliminating the redundant targets, a total of 573 known therapeutic targets were collected from 135 compounds (Tables S3-S11). For further analysis, a Compound-Target network was visualized using Cytoscape.
The network contains 715 nodes and 2979 edges. Among them 135 nodes represented the active components, and 573 nodes represent the corresponding targets of the ingredients (Fig. 3B). The Compound-Target network showed that a herb could interact with multiple components, and a compound could also interact with several targets, which coincided with the synergistic effect theory of multi-components and multi-targets of traditional Chinese medicine formula. The degree (connection strength) was re ected by the node's sizes, according to the degree value, Fig. 3C and Table S12-13 showed the top 10 active components and targets.

Acquisition of Therapeutic Gene Targets for Liver Failure
To investigate the therapeutic gene targets in liver failure, we analyzed the expression pro les data from GEO database. We applied the GEO2R online analysis tool with default parameters to screen the differentially expressed genes (DEGs) in two GEO series (GSE14668, GSE38941 ), using adjusted P value < 0.05 and logFC ≤ −1 or logFC ≥ 1 as the cut-off criteria. Two GEO series contains 18 normal and 25 liver failure samples. Fig. 4A-4B showed the up-regulated genes and down-regulated genes. Venn diagram identi ed 2961 common DEGs in two GEO series. 21 DEGs with inconsistent trends had been excluded. Overall, 2940 DEGs were obtained, including 1760 up-regulated and 1180 down-regulated genes (Fig. 4C,Table S14-16).

Potential Targets of YQJPF Decoction on Liver Failure
Venn diagram showed that 2940 gene symbols for disease and 573 gene symbols for drugs had 163 overlap. That was, 163 gene symbols would mostly likely to be the therapeutic targets for liver failure treatment by YQJPF Decoction (Fig. 5A, Table S17).

Construction of PPI and Compound-Target-Disease network for potential therapeutic targets
The 163 potential therapeutic targets were uploaded to the STRING database, which provided the information on predicted interaction. Furthermore, we imported the above data into Cytoscape 3.7.1 to calculate the characteristics of the network and construct the protein-protein interaction network (PPI network) (Fig. 5B). In the PPI network, targets with higher degree played an important role in the protein-protein correlation. In Fig. 5C, ranked by degree value, 18 targets were collected to be the hub targets, namely VEGFA, EGFR, STAT3, CXCL8, ESR1, CCND1, PTPRC, MMP2, PECAM1, AR, SPP1, EDN1, CRP, F2, TIMP1, IRS1, HGF, CAV1, more details shown in Table S18. Then we built a Compound-Target-Disease network of complex information based on interactions between the drug (YQJPF), active compounds, and disease (ACLF) using Cytoscape to undertake visual analysis (Fig. 5D). In this network quercetin, resveratrol, syringaresinol diglucoside, luteolin, etc had higher degree value than other active compounds, connected to more than seven genes, so they would be the main active compounds to treat liver failure, more details shown in Table 1.

Analyses of Enrichment of the KEGG/GO Pathways
Gene ontology (GO), KEGG pathway enrichment analysis were performed in Metascape (http://metascape.org), with P Value <0.01; Enrichment > 1.5. GO enrichment analysis of the 163 potential therapeutic targets was performed for identifying the relevant biological functions of YQJPF against liver failure. The top 10 signi cantly enriched terms with a greater number of involved targets in biological process (BP) , molecular unction (MF) and cellular component (CC) categories were shown in Fig. 5E, which indicated that YQJPF may regulate oxidoreductase, nuclear receptor, get involved in wound healing, regeneration, metabolic process and so on. To explore the potential pathways of YQJPF on liver failure, the pathway enrichment of the 163 potential therapeutic targets was performed. The top 20 signi cantly enriched pathways were shown in Fig. 5F. Excluded pathways not related to liver disease or cancer-related pathways. PI3K/AKT, p53, FoxO, HIF-1, AMPK signaling pathways were the prominently enriched signaling pathways according to the gene count and P value (Table   S20), which associated with in ammation, hypoxia, metabolism and proliferation. VEGF-A showed the highest degree value in hub targets. According to KEGG pathway enrichment analysis ( Table 2, Table S20), apoptosis related genes BAX, Bcl-2 were closely related to the above pathways.
According to the network pharmacology analysis results, PI3K-AKT, HIF-were the primary pathways, VEGF-A was identi ed as the primary drug candidate targets for the treatment of YQJPF on liver failure, and it was a downstream target marker of HIF-1 , hypoxia-related cell injury plays an important role in acute and chronic liver failure [[ii]]. HIF-1 and VEGF-A increase under hypoxic conditions, act as a mediator for cell adaptation to a harmful environment. So we further investigated the mechanism of YQJPF on relieving liver injury by examining the protein levels in CCl 4 and LPS/D-gal induced ACLF model rats. YQJPF upregulated the expression of PI3K, AKT, HIF-1 , VEGF-A. Besides, YQJPF increased HIF-1 and VEGF-A expression by immunohistochemistry analysis, which indicated that hypoxic pathway signi cantly enhanced after YQJPF treatment. Moreover The expression of BCL-2, and BAX followed the same trend as the in vitro study results. Taken together, these ndings may indicate that the mechanism of YQJPF on ACLF may through PI3K/AKT signaling pathway, modulate the the expression of HIF-1 , VEGF-A, and apoptosis associated BCL-2, BAX. YQJPF could ameliorates liver injury through in uencing hypoxia damage and apotosis (Fig. 6A).

Discussion
Acute-on-chronic liver failure (ACLF) is a major worldwide medical problem, the worldwide reported mortality of ACLF according to the European Association for the Study of the Liver-Chronic Liver Failure (EASL-CLIF) Consortium de nition ranges between 30% to 50%. Medical management of ACLF consists of early recognition, treatment of the precipitating event and supportive care, which includes antibacterial therapy, HBV-speci c therapy, immunomodulation, extracorporeal liver support systems and liver transplantation [1] . However, there is still no signi cant increase in the availability of new drugs for improving liver injury, and alternative therapies are required to develop approaches to treat ACLF, such as TCM.
YQJPF is a traditional medicine formula, which proven to have a good therapeutic effect in clinical practice. In vivo experiments had con rmed that YQJPF can alleviate liver necrosis and brosis. Meanwhile YQJPF could promote proliferation and inhibit apoptosis in vitro. In order to get a better understanding of mechanism, a series of control experiments were conducted. In network pharmacology, there were 135 active components among the 9 herbs in YQJPF, including quercetin, resveratrol, kaempferol, luteolin, which considered to be the major components (Fig. 3B) (Fig. 6). YQJPF administration activates Bcl2, reduces BAX, and shifts the BAX/Bcl2 ratio in a anti-apoptotic direction, in order to protect cells from apoptosis, similar results were also demonstrated in vitro studies. Previous studies have shown that HIF-1α is subjected to regulation by the PI3K/Akt signaling pathway, especially under hypoxia condition, PI3K/HIF pathway plays a important role in cardio-protection and neuro-protection. This relationship has also been shown in the pathogenesis of several forms of liver disease, a number of mechanisms have been proposed for the protective effect of HIF-1. HIF-1 has been found to protect cells from hypoxic injury by promoting nutrient and O 2 transport via inducing the expression of downstream proteins such as VEGF [[xxiii]], which promote angiogenesis [[xxiv]]. After administration with YQJPF, the expressions of HIF-1 and VEGF-A were increased, which were further con rmed by immunostaining and Western blot analysis. Therefore, the anti-liver failure effect of YQJPF may involve in the regulation of the HIF-1 /VEGF-A and apoptosis signaling pathways. These results suggest that YQJPF is bene cial for alleviating liver failure, may regulate hypoxic liver injury through PI3K/AKT-HIF1α dependent apoptosis pathway. More experiments are necessary to verify the mechanisms of YQJPF on hepatocyte injury.

Conclusion
In summary, we evaluated active compounds and possible targets in YQJPF to treat liver failure by network pharmacology and transcriptomics analysis.
Moreover we establised a CCl 4 and LPS/D-gal induced ACLF rat model and LPS induced hepatocytes model to verify the effect and mechanism in vivo and in vitro. These results showed that YQJPF could treat liver failure in rats by regulating hepatocyte apoptosis and hypoxic injury through PI3K/AKT pathway. Our more in-depth work will focus on verifying compound's targets and pathways.      Hepatoprotective effect of Yiqijianpi formula (YQJPF) on liver failure through modulation of hypoxic and apoptosis pathway. By combining the network pharmacological analysis and our results, we hypothesized that YQJPF activates the PI3K/AKT signaling pathway and modulates the expression of HIF-1 and BCL-2 family proteins, thereby alleviate liver hypoxia injury and apoptosis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.