The protection of LingqiHuangban Granule on human retinal endothelial cells against oxidative stress-induced injury by network pharmacology

Background : LingqiHuangban Granule(LQHBG) is a famous traditional Chinese medicine formula used to manage retinal diseases, as an effective holistic treatment through warming Yang to exert tonifying effects on kidney and invigorating spleen to remove dampness to nourish essence of effect. The study examined protection of LQHBG on oxidative stress-induced injury in human retinal endothelial cells(HRECs) in vitro, determined the potential molecular targets of LQHBG using network pharmacology. Methods : The potential targets of active ingredients in LQHBG were predicted using pharmmapper. Gene Ontology(GO) and Kyoto Encyclopedia of Genes and Genomes(KEGG) pathway enrichment analyses were carried out using Molecule Annotation System. The protein-protein interaction networks were constructed using Cytoscape. LQHBG was administered to rabbits to prepare medicated serum. The apoptosis of HRECs was evaluated by TUNEL and Flow Cytometry(FCM). MDA, SOD, LDH, GSH-Px, and T-AOC were detected. The mRNA expressions of Nrf2, NF-κB and HO-1 were detected, protein expression levels of Nrf2, Bcl-2, NF-κB, HO-1 and caspase-3 were analyzed. Results: TUNEL demonstrated the numbers of apoptotic cells in low-and high-dose LQHBG groups was obviously less than model group(P<0.05). FCM analysis revealed apoptotic rates of HRECs in low-and high-dose LQHBG groups were obviously reduced in a dose-dependent manner(P<0.05). The potential mechanism of LQHBG was the NF-κB pathway identified using PharmMapper. LQHBG significantly decreased MDA, LDH levels and enhanced SOD, GSH-Px and T-AOC generation. LQHBG inhibited upregulation of NF-κB, caspase-3 and enhanced Bcl-2, Nrf2, and HO-1 expression. Conclusion: LQHBG protected HRECs against

protects retinal function and morphology from light-induced retinal damage in rats (Gong et al., 2008).
Network pharmacology is a distinctive new approach based on advances in network biology and polypharmacology, providing a more comprehensive understanding of the targets, mechanism, and pathways behind drug action. PharmMapper provides deeper insights and scientific evidence for TCM and helps identify potential targets of chinese herbs and their underlying mechanisms (Tang and Aittokallio., 2014;Poornima et al., 2016). The Encyclopedia of Traditional Chinese Medicine (ETCM, http://www.nrc.ac.cn:9090/ETCM/) (Xu et al., 2018) was used to identify the candidate target genes of each herb contained in LQHBG. In the present study, investigations based on the pharmacology database, potential signaling pathway were predicted using the network pharmacology method. The current research investigated the protective effects of LQHBG in HRECs under hydrogen peroxide-induced oxidative stress injury and used the serum pharmacology method for drug administration. TUNEL and FCM assays were used to measure apoptosis of HRECs, and MDA, SOD, LDH, GSH-Px, and T-AOC analyses were used to measure oxidative stress levels. We also investigated the underlying mechanisms. The expressions of Nrf2, NF-κB and HO-1mRNAs were identified by PCR, and protein expressions of Nrf2, Bcl-2, NF-κB, HO-1 and caspase-3 were analyzed by Western blot.

Pharmacological Network construction
The Encyclopedia of Traditional Chinese Medicine (ETCM, http://www.nrc.ac.cn:9090/ETCM/) (Xu et al., 2018) was used to identify the candidate target genes of each herb contained in LQHBG. The network was constructed using Cytoscape 3.7.1 (Shannon et al., 2013) with the protein-protein interaction (PPI) data from High-quality INTeractomes (HINT; version 4, http://hint.yulab.org) (Das J et al., 2012), and major genes were screened after the calculation of topological features. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of major genes was conducted using the Database for Annotation, Visualization and Integrated Discovery (DAVID; version 6.8, https://david.ncifcrf.gov/) (Huang Da W et al., 2009). The herb-gene-pathway network was constructed using Cytoscape 3.7.1 to visualize the relationships among herbs, genes, and pathways.

Preparation of LQHBG
LQHBG consists of ten herbs and the mixed proportion of respective herb in LQHBG is illustrated in Table1. All the mentioned herbal names can be checked in accordance with the database of www.theplantlist.org. LQHBG was developed by the First People's Hospital Affiliated with Shanghai Jiaotong University(Shanghai medicine system word: Z05050795).

Reagents
DMEM, fetal bovine serum and trypsin were purchased from GIBCO company (United States).Two dimethyl sulfoxide sulfone (dimethylsulfoxide, DMSO) were purchased from Sigma. An annexin V/propidium iodide (PI) apoptosis detection kit was purchased from R&D company. A terminaldeoxynucleotidyl transferase-mediated deoxyuridine triphosphate-biotinnick end labeling(TUNEL) kit was purchased from America Promega. A Western blotting kit (AC031, AB126, antibody) was purchased from Beyotime Company.

HPLC analysis
LQHBG were analyzed by Agilent HPLC instrument(Agilent, USA) equipped with a Eclipse XDB-C 18 (250mm×4.6 mm5, μm) column and an evaporative light scattering detector(ELSD). The drift tube temperature of the ELSD system was set to 30 o C and the auxiliary gas flow rate was 1.0L/min. The mobile phase consisted of acetonitrile(A) and water(B) at a flow rate of 1.0L/min. A gradient program was used as follows: 0-8min:A-B(20:80), 8-18min: from A-B(20:80) to A-B(30:70)18-20min: A-B(30:70),20-23min: from A-B(30:70) to A-B(20:80). The detection wavelength was set at 334nm and flow rat was maintained at 1.0ml/min. The typical HPLC-ELSD chromatogram of LQHBG was shown in Fig.2, and the contents of acteoside and salvianolic acid in LQHBG were 36.9ug/g and 344ug/g, respectively. Twelve male chinchilla rabbits, body mass (2.0±0.2) kg, were randomly divided into a traditional Chinese medicine group(n=8) and control group(n=4). The traditional Chinese medicine, low-dose and high-dose groups(n=4 rats in each group) received a clinical adult dose (according to the average body weight of 60kg), 10 times and 20 times the daily dose, the daily dose of a 30g /60kg×rabbit body weight (kg)×10 times or 20 times, divided into 8:00am and 8:00pm administration, for 3 consecutive days. During the periods of feeding and fasting (12h), abdominal blood was collected every 1~2h under sterile conditions with an injection of 1% pentobarbital sodium( Sigma, USA), the control group was given the same amount of normal saline, and the blood was collected at the same time for 3 days, and then the rabbits were sacrificed. Blood was stored at 4 o C in the refrigerator overnight and centrifuged at 3000r/min (centrifugal radius 10cm) for 25 min. The supernatant was subjected to 0.22μm filter sterilization. Filtered samples from both groups were heated to 56°C for 30 min and stored at -20°C.

Cell culture
HRECs were purchased from Sciencell(Research Laboratories, Inc., Carlsbad, CA, USA). The cell were cultured in human endothelial medium with 5% fetal bovine serum and 1% endothelial cell growth supplement at 37°C with 5% CO 2 and 95% ambient air. HRECs were inoculated into 5 large dishes at 2 * 10 5 /mL. Disposable culture fluid and PBS were added to the following groups: control group serum, serum-free DMEM/F culture liquid plus H 2 O 2 (500μmol/L), control group serum and H 2 O 2 (500μmol/L), low-dose drug serum plus H 2 O 2 (500μmol/L), highdose drug serum plus H 2 O 2 (500μmol/L). Cells were cultured for 24h, digested and collected with 0.25% trypsin.

Flow cytometry
Cell apoptosis was analyzed with FCM analysis. After treatment, cells were harvested and washed twice with PBS, resuspended in 500μL binding buffer, then stained with Annexin V-FITC and propidium iodide (PI), followed by incubation at room temperature for 15 min in the dark. After washing with binding buffer, the cells were analyzed by FCM. The above experiments were repeated 3 times.

TUNEL detection
HRECs were inoculated to preset coverslips in 24-well plates at a density of 5 * 10 4 /mL. Two wells were used for each treatment group. Cells were grown to a monolayer, the culture fluid was discarded, and cells were washed with PBS. Cells were cultured for 24h, and the coverslips were removed. Apoptotic cell detection was performed according to the TUNEL kit.
The relative levels of target mRNA expression were calculated using the 2 -ΔΔCt method.

MDA level and SOD, GSH-Px, CAT and T-AOC activity
Analysis kits for MDA, SOD, GSH, CAT, and T-AOC were purchased from Beyotime Biotechnology (Jiangsu, China). The GSH-PX, CAT, SOD, MDA, and T-AOC contents were measured using assay kits in strict accordance with the manufacturers' instructions. MDA contents are expressed as nmol/ml protein. SOD contentsare expressed as U/ml protein, LDH contents are expressed as U/L protein, GSH-PX are expressed as U/g protein, T-AOC contents are expressed as g/L protein.

Statistical analysis
Data analysis was performed using SAS6.12. Data were measured using the mean±SD. Multiple sets of data between the single factor analysis of variance. Student's t-test was used for comparisons between two groups. P<0.05 was statistically significant.  (Table S1). We identified 125 major genes (Degree≥6) after the calculation of topological features (Table S2), and analyzed 114 KEGG pathways using DAVID (Table S3). We selected 6 interesting pathways for network construction (Figure 1): TNF signaling pathway, HIF-1 signaling pathway, VEGF signaling pathway, NF-kappa B signaling pathway.

Flow cytometry
The normal control group exhibited the fewest number of apoptotic cells, and the apoptosis rate was 1.9+0.16% (Figure3A). Serum levels of cell apoptosis were similar in model group and model plus serum group(12.70+1.37% and 12.5+0.94%, respectively) (Figure3B, C). The difference between the two groups was not significant (P > 0.05). Apoptosis was significantly reduced in low-dose drug plus serum group(10.8+0.55%) and high-dose drug plus serum group(8.50+0.39) (Figure3D, E). These two groups were significantly different (P<0.01) from model group. However, the difference between the two groups was also significant (P < 0.05) (Figure3F).

TUNEL staining
TUNEL staining was negative in normal control group, and no yellow HREC nuclei were observed ( Figure 4A). Visible large yellow TUNEL-positive staining was observed in HREC nuclei in model group and model serum group (Figure 4B, C). Small yellow TUNEL-positive staining of HREC nuclei were detected in low-dose drug plus serum group ( Figure 4D). Yellow TUNEL-positive staining in nuclei of HREC was further reduced in high-dose drug plus serum group, which was significantly less than that in model group ( Figure 4E). The difference between the low-dose and high-dose drug plus serum groups was significant (P < 0.05) (Figure4F).

Protein expression of caspase-3, Bcl-2, NF-KB, Nrf2, and HO-1
Protein expression was measures at the relative intensity of gray value ( Figure 6A). Caspase-3 and NF-KB protein were expressed at low levels in HRECs of normal control group. Caspase-3 and NF-KB protein expressions were higher in model and model serum groups than in normal control group. Caspase-3 and NF-KB protein expressions were lower in low-dose drug plus serum group than in model group. The high-dose drug serum group expressed lower than low-dose drug plus serum group( Figure 6D,E)(P < 0.05). Nrf2, Bcl2 and HO-1 proteins were detected in normal control group. The levels of these proteins were significantly lower in model and model serum groups than normal control group. LQHBG induced a dose-dependent increase in Bcl2, Nrf2 and HO-1 protein expression compared to model and model serum groups (P < 0.05) ( Figure 6A,B,C,F).

MDA and LDH levels and SOD, GSH-Px and T-AOC activities
The MDA and LDH levels in model and model plus serum groups were significantly higher than normal control group (Figure 7) (P< 0.05). LQHBG treatment decreased these levels in a dose-dependent manner (p < 0.05) ( Figure 7B,C). SOD, GSH-Px, and T-AOC activities were significantly lower in model and model plus serum groups compared to normal control group. The activities of these enzymes were dose-dependently higher in LQHBGtreated groups than the model group (p < 0.05) ( Figure 7A,D,E).

Discussion
Ischemia is an important pathological mechanism in retinal diseases. Most cell death associated with retinal ischemia is due to programmed cell death (i.e., apoptosis) (Levin andLouhab., 1996, Zhang et al., 2005). The retina is highly dependent on the oxygen supply, and ischemia-induced oxidative stress is a direct cause of retinal damage (Bonne, Muller and Villain., 1998). Reactive oxygen species(ROS) are a variety of molecules and free radicals derived from molecular oxygen, including superoxide, hydroxyl radicals, hydrogen peroxide(H 2 O 2 ), and singlet oxygen(Di Mascio, Murphy and Sies., 1991). These molecules are produced during the ischemiareperfusion process. Free radicals lead to the destruction of DNA molecules and result in energy exhaustion and cell death (Erdelyi et al., 2005). Endothelial-sourced ROS are important physiologically and pathologically in many ischemic retinopathies (Pellegrini-Giampietro et al., 1990) .
MDA and LDH are biomarkers of oxidative stress levels (Anderson, Maude and Nielsen., 1985), and these proteins are often used to evaluate oxidative stress-induced endothelial injury (Chen et al., 2009). Enzymatic antioxidant enzymes, such as SOD, GSH-Px, and T-AOC, play key roles in neutralizing the ROS that are produced during stress conditions, and these enzymes are essential in defending against endothelial damage (Agardh et al., 2006, Lewden et al., 1996. Accumulated ROS during antioxidant responses in endothelial cells, and maintaining a balance is essential for homoeostasis in endothelial cells (Shi et al., 2015). Our results demonstrated that LQHBG significantly inhibited the formation of intracellular ROS in H 2 O 2 -stimulated HRECs. LQHBG notably decreased MDA and LDH levels and increased SOD, GSH-Px, and T-AOC activity levels. These results suggest that the protective effect of LQHBG is associated with the inhibition of intracellular ROS production.
The present study used the serum pharmacological method to investigate the protective effect of LQHBG on oxidative stress in vitro. FCM and TUNEL tests confirmed that LQHBG-medicated serum inhibited HREC apoptosis under oxidative stress. Further studies on apoptotic signal transduction molecules revealed that LQHBG played a protective role via the downregulation of NF-KB, caspase-3 and upregulation of Bcl-2 to inhibit hypoxia-induced HREC apoptosis, which was consistent with the analysis of network pharmacology.
Nuclear factor (erythroid-derived 2)-like 2 (Nrf2) is a critical transcription factor that regulates the antioxidant defense system (de Vries et al., 2008, Enomoto et al., 2001, Alfieri et al., 2011. Nrf2 controls the antioxidant response elements (AREs) located in the promoter regions of its target genes and activates downstream gene expression, including the antioxidant enzymes heme oxygenase 1(HO-1) and NAD(P)H:quinine oxidoreductase 1 (NQO1) (Jeong, Jun and Kong., 2006, Zhao et al., 2014, Apopa, He and Ma., 2008. Our results demonstrated that LQHBG significantly alleviated stress-induced HREC damage via promoting Nrf2 and HO-1 activation (Figure6). The findings of the present study suggest that the protective ability of LQHBG involves activation of Nrf2/ARE/HO-1 signaling pathway and inhibition of NF-kappa B signaling pathway.

Conclusions
Previous studies have demonstrated that LQHB obviously protects the function and morphology of the retina from light-induced retinal damage in rats (Gong et al., 2008). LQHB has been used to treat retinal diseases in patients for over 17 years with great efficacy. Patient visual capacity is improved via the promotion of circulation in the choroids and retina, which prevents cell death, and reverses the inability to absorb and circulate vital nutrients for healthy vision. Conclusively, drug-containing LQHBG serum protects HRECs from oxidative stressinduced injury via reduced apoptosis and improved antioxidant capacity, which may involve activation of Nrf2/ARE/HO-1 signaling pathway and inhibition of NF-kappa B signaling pathway.

Consent to publish
No human was involved in this study.

Availability of data and materials
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

Funding
This study was supported by the Tibet Natural Science Foundation of China (No. XZ2017ZR-ZYZ09; to QHQ) and National Natural Science Foundation of China (No. 81300779;to ZZZ); National Key R&D Program of China2016YFC09048002019YFC0840607; to QHQ and ZZand National Science and Technology Major Project of China2017ZX09304010;to QHQ and ZZin animal experiments, cell experiments, analysis, interpretation of data, and in writing the manuscript.. Figure 1 Signaling pathways for network construction.

Figure 2
Chromatogram of LQHBG and standard chemicals mixture analyzed by HPLC.1. Acteoside ; 2. salvianolic acid.   mRNA levels in model and blank serum group were strongly reduced compared with normal control group; LQHBG dose-dependently attenuated the reduction of HO-1, Nrf2 expression. Histograms represent the mean+SEM of HO-1, Nrf2mRNA levels in all groups.*: p <0.01 versus normal control group; #: p < 0.01 versus model group.

Figure 6
LQHBG reduced release of caspase3, NF-KB, and attenuated the reduction of Nrf2, HO-1, Bcl-2 protein expressions in HRECs. (A,D,E): Western-blot analysis shows significant activation of caspase3, NF-KB in model group and blank serum model group; LQHBG treatment dose-dependently reduced these effects*: p< 0.01 versus normal control group ; #: p<0.01 versus model group. (A,B,C,F): Histograms represent the mean+SEM of Nrf2, HO-1, Bcl-2 protein levels in retinas of all groups. The HO-1, Nrf2, Bcl-2 protein levels in model and blank serum group were strongly reduced compared with normal control group; LQHBG dose-dependently attenuated the reduction of HO-1, Nrf2, Bcl-2 protein expression. Histograms represent the mean+SEM of HO-1, Nrf2, Bcl-2 protein levels in all groups.*: p < 0.01 versus normal control group; #: p < 0.01 versus model group.*: p < 0.01 versus normal control group ; #: p < 0.01 versus model group. and blank serum model group became significantly higher than normal control group(p < 0.05); MDA, LDH levels in low dose and high dose drug plus model groups were higher than normal control group, but were significantly lower than model group(p<0.05). LQHBG dose-dependently attenuated the activation of MDA, LDH levels.*: p<0.05 versus normal control group; #: p< 0.05 versus model group. (B, C, D): Activities of SOD, GSH-Px, and T-AOC in model group and model plus serum group became significantly lower than those of normal control group (p < 0.05); Activities in low dose and high dose drug plus model groups were significantly higher than those of model group(p < 0.05). LQHBG dose-dependently attenuated the reduction of SOD, GSH-Px and T-AOC levels.*: p < 0.05 versus normal control group; #: p< 0.05 versus model group.

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