Elsholtzia bodinieri Vaniot Ameliorated Acute Lung Injury by NQO1, BCL2 and PTGS2 In Silico and In Vitro Analyses

Acute lung injury (ALI) is a clinical respiratory disease caused by various factors, which lacks effective pharmacotherapy to reduce the mortality rate. Elsholtzia bodinieri Vaniot is an annual herbaceous plant used as a traditional herbal tea and folk medicine. Here we used bioinformatic databases and software to explore and analyze the potential key genes in ALI regulated by E. bodinieri Vaniot, including B cell leukemia/lymphoma 2 (Bcl2), prostaglandin-endoperoxide synthase 2 (Ptgs2) and NAD(P)H dehydrogenase, quinone 1 (Nqo1). In an inflammatory cells model, we verified bioinformatics results, and further mechanistic analysis showed that methanol extract of E. bodinieri Vaniot (EBE) could alleviate oxidative stress by upregulating the expression of NQO1, suppress pyroptosis by upregulating the expression of BCL2, and attenuate inflammation by downregulating the expression of PTGS2. In sum, our results demonstrated that EBE treatment could alleviate oxidative stress, suppress pyroptosis and attenuate inflammation by regulating NQO1, BCL2 and PTGS2 in a cells model, and E. bodinieri Vaniot might be a promising source for functional food or as a therapeutic agent.


Introduction
ALI is clinical respiratory syndrome caused by various factors, such as COVID-19, toxic inhalation pneumonitis, sepsis, severe trauma, and acute pancreatitis [1][2][3]. The prominent characteristics of ALI involves edema of airways with epithelial sloughing, inflammation of lung and edema with hypoxemia [4]. Existing research showed that the development of ALI was closely related to the inflammatory injury, oxidative stress and immune cell infiltration [5][6][7]. The histopathological lesions of ALI in COVID-19 are of wide concern at the moment. Although various pharmacotherapies are available, the outcome and mortality of ALI patients remains poor [8]. Therefore, there is an urgent need to develop treatments to ameliorate lung inflammation and tissue damage.
E. bodinieri Vaniot belongs to the taxonomically diverse group of the family Labiatae, and it is an annual herbaceous plant. The genus Elsholtzia consists of at least 33 species, which are widely distributed in East Asia, Africa, and Europe, and some of which have been used as beverages (herbal tea), domestic folk medicine, spices, cosmetics and so on [9]. E. bodinieri Vaniot is commonly known as "Dongzisu" in China, which grows in the northwest and southwest districts of China, especially in the Yunnan and Guizhou Provinces [10]. Usually, E. bodinieri Vaniot has been used as traditional folk medicine for the treatment of cough, headache, pharyngitis, fever and hepatitis [11]. Previous reports showed that the aerial parts of E. bodinieri Vaniot have been isolated clerodane diterpenoid glycosides, flavonoid glycosides, triterpenoid saponins, sesquiterpene glycosides, and phenolic constituents [10][11][12][13][14][15][16]. 2 of 14 In this work, we use databases and software to explore and analyze potential target genes in ALI, which are regulated by E. bodinieri Vaniot. In order to verify the results of the analysis, we investigate the effects of E. bodinieri Vaniot on potential target genes, and the mechanism in vitro. The overall design of this study is shown in Figure 1. These findings indicated that E. bodinieri Vaniot might be a therapeutic agent for ALI. showed that the aerial parts of E. bodinieri Vaniot have been isolated clerodane diterpenoid glycosides, flavonoid glycosides, triterpenoid saponins, sesquiterpene glycosides, and phenolic constituents [10][11][12][13][14][15][16].
In this work, we use databases and software to explore and analyze potential target genes in ALI, which are regulated by E. bodinieri Vaniot. In order to verify the results of the analysis, we investigate the effects of E. bodinieri Vaniot on potential target genes, and the mechanism in vitro. The overall design of this study is shown in Figure 1. These findings indicated that E. bodinieri Vaniot might be a therapeutic agent for ALI.

Screening Differentially Expressed Genes and Enrichment Analysis in ALI
In the GSE17355 dataset, a total of 871 differentially expressed genes (DEGs) in ALI were identified, including 379 upregulated genes and 492 downregulated genes ( Figure   Figure 1. The bioinformatics and experiments workflow of E. bodinieri Vaniot ameliorate acute lung injury.

Screening Differentially Expressed Genes and Enrichment Analysis in ALI
In the GSE17355 dataset, a total of 871 differentially expressed genes (DEGs) in ALI were identified, including 379 upregulated genes and 492 downregulated genes (Figure 2A). In the GSE1871 dataset, a total of 5016 DEGs in ALI were identified, including 2484 upregulated genes and 2532 downregulated genes ( Figure S1A). In the GSE2411 dataset, a total of 1053 DEGs in ALI were identified, including 703 upregulated genes and 350 downregulated genes ( Figure S1B). We obtained 288 downregulated overlapping genes ( Figure 2B) and 621 upregulated overlapping genes ( Figure 2C); these genes varied across the three datasets. Next, the downregulated overlapping genes were identified by Gene Ontology (GO) analysis, and the top nine enriched terms were shown ( Figure 2D). Meanwhile, upregulated overlapping genes were explored by GO analysis, the top nine enriched terms were also shown ( Figure 2E). Finally, the different Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways of downregulated and upregulated overlapping genes were analyzed. The results showed the top 16 KEGG pathways involving downregulated overlapping genes, which included hepatocellular carcinoma, fluid shear stress and atherosclerosis, parathyroid hormone synthesis, secretion and action and so on ( Figure 2F). The top ten KEGG pathways involving upregulated overlapping genes included TNF signaling pathway, IL-17 signaling pathway, NF-kappa B signaling pathway and so on ( Figure 2G).

Exploration and Identification of Key Genes in ALI Regulated by E. bodinieri Vaniot
We retrieved 97 compounds of E. bodinieri Vaniot from CAS scifinder and PubChem databases (Table S1). In the Swiss Target Prediction database, 637 potential molecular targets (Table S2) of 97 compounds were collected ( Figure 3A). In the DisGeNET database, 93 ALI-related genes were obtained (Table S3). Then, we screened 33 hub genes from upregulated genes, ALI-related genes, potential target genes of E. bodinieri Vaniot, and downregulated genes ( Figure 3B). The GO network of hub genes was investigated by Metascape database, and Cytoscape software was utilized to construct a GO network that consisted of 163 nodes and 1967 edges ( Figure 3C). Furthermore, based on potential target proteins and corresponding signaling pathways, a key genes-compounds network was constructed to further clarify the possible modes of action ( Figure 3D). We analyzed the difference of Ptgs2 and Bcl2 between normal and ALI lung tissues at the transcription level by GSE1871, GSE17355, and GSE2411 datasets. The expression level of Ptgs2 was significantly higher (p < 0.05, p < 0.01) in ALI than normal lung tissues ( Figure 3E,F). Differently, Bcl2 expression was markedly lower (p < 0.01) in ALI than in normal lung tissues ( Figure 3G).

EBE Alleviated Oxidative Stress of Cells Model by NQO1
To investigate the effect of EBE in vitro, gradient concentrations of EBE were determined at different times via Cell Counting Kit-8 (CCK-8). Then, 50 µg/mL EBE (EBE high concentration, EBEH) and 12.5 µg/mL EBE (EBE low concentration, EBEL) were used to explore effects on cells model ( Figure 4A). We induced an inflammatory cells model by lipopolysaccharide (LPS) in RAW 264.7, NQO1 expression was shapely downregulated (p < 0.05) in cells model, EBEH and EBEL treatment both remarkably rescued NQO1 expression (p < 0.01, p < 0.05) ( Figure 4B). Accordingly, reactive oxygen species (ROS), malondialdehyde (MDA) and cell viability in cells model were evaluated. In cells model, ROS content was marked accumulated (p < 0.01), comparatively, ROS content both significantly reduced (all p < 0.001) with EBEH and EBEL treatment ( Figure 4C,D). Identically, the MDA level was remarkably swelled (p < 0.001), EBEH and EBEL treatment significantly diminished the MDA level (all p < 0.01) ( Figure 4E). Consequently, cell viability was remarkably decreased (p < 0.01) in the cells model, EBEH and EBEL treatment all significantly increased (p < 0.05, p < 0.001) cell viability ( Figure 4F).

EBE Inhibited Pyroptosis of Cells Model by BCL2
To examine whether EBE contributes to the inhibition of pyroptosis in ALI by BCL2, the protein expression levels of BCL2, CASPASE9, CASPASE3 and gasdermin D (GS-DMD) were evaluated in vitro. BCL2 expression was downregulated (p < 0.05), EBEH and EBEL treatment significantly upregulated BCL2 levels (p < 0.001, p < 0.01) in cells model ( Figure 5A). The CASPASE9 level was increased (p < 0.05), EBEH and EBEL treatment all decreased (all p < 0.01) it in cells model ( Figure 5B). Accordingly, the CASPASE3 level was increased (p < 0.05), EBEH and EBEL treatment all decreased (all p < 0.05) in the cells model ( Figure 5C). Further, we observed that the CASPASE3-cleaved level was increased (p < 0.05) by LPS, and were reversed (p < 0.001, p < 0.01) by EBEH and EBEL treatment in cells model ( Figure 5D). Earlier studies demonstrated that the activation of the executioner CASPASE3 (CASPASE3-cleaved) promoted microglial pyroptosis by active GSDMD [17,18]. Next, we observed that the N-terminal fragments of GSDMD (GSDMD-N) level was significantly increased (p < 0.05), EBEH and EBEL treatment all decreased (all p < 0.01) the GSDMD-N level in the cells model ( Figure 5E). As result, the number of living cells were remarkable decreased (p < 0.01) in the model, EBEL treatment significant restored (p < 0.05) numbers of living cells ( Figure 5F).
In addition, transmission electron microscopy (TEM) results showed that mitochondrial swelling, mitochondrial cristae collapse, membrane pore formation and heterochromatin marginalizing in cells model, EBEH and EBEL treatment partially alleviated the abnormal changes of mitochondrial morphology, nuclear membrane and heterochromatin ( Figure 5G). Collectively, these results showed that EBE suppressed pyroptosis of the cells model by upregulating the expression of BCL2.

EBE Attenuated Inflammation of Cells Model by PTGS2
To determine whether EBE plays a basic role in the amelioration of inflammation in ALI by PTGS2, inflammatory factors and prostaglandin E2 (PGE2) were investigated in the cells model. PTGS2 expression was raised significantly (p < 0.01) in the cells model, but EBEH and EBEL treatment both reduced it (all, p < 0.01) ( Figure 6A). Similarly, interleukin 6 (IL-6), interleukin 1 beta (IL-1β) and nitric oxide (NO) levels all were increased (p < 0.05, p < 0.01, p < 0.05) in the cells model; comparatively, EBEH and EBEL treatment significantly decreased IL-6 and IL-1β (p < 0.001, p < 0.01), but only EBEH treatment could markedly decrease NO (p < 0.001) ( Figure 6B-D). Moreover, PGE2 level was increased (p < 0.05), EBEL treatment decreased (p < 0.05) it in the cells model ( Figure 6E). These data indicated that EBE attenuated inflammation of the cells model by downregulating the expression of PTGS2.

Discussion
We retrieved 637 potential target genes regulated by E. bodinieri Vaniot in Swiss Target Prediction and 93 ALI-related genes in the DisGeNET database. These constituted the compilation of genes in ALI regulated by E. bodinieri Vaniot. GO and KEGG analysis were used to explore genes functions and signaling pathways. We analyzed different expression of Ptgs2 and Bcl2 between normal and ALI lung tissues at the transcription level by GSE1871, GSE17355, and GSE2411 datasets. Nqo1 was a potential target gene regulated by terpenoids in EBE. As a key ROS scavenger, NQO1 contributes to alleviating oxidative stress [19][20][21]. Considering the above findings, we investigated Bcl2, Ptgs2 and Nqo1 as potential key genes in ALI regulated by E. bodinieri Vaniot. We verified these results by experiments in vitro.
Intensive oxidative stress is regarded as an important process to the pathogenesis of ALI [22]. Oxidative stress causes the damage and enzymolysis of cellular structural molecules, and finally induces apoptosis, necrosis, or other death modes [23]. We found that EBE treatment upregulated NQO1, decreased ROS and MDA content, and increased cell viability in the cells model. These findings indicated that EBE could alleviate oxidative stress of the cells model by upregulating NQO1 expression.
Pyroptosis, a programmed cell death pathway, is also known as inflammatory caspasedependent cell death [24]. Pyroptosis can be triggered by noninfectious stimuli and microbes [25] and plays central regulatory roles in the progression of numerous diseases, such as ALI [26], Alzheimer's disease [27], tumor [28], and inflammatory diseases [24]. We found that EBE contributed to the inhibition of pyroptosis and restoring numbers of living cells in the cells model. Mechanistically, the administration of EBE upregulated BCL2 expression, decreased CASPASE9, CASPASE3, CASPASE3-cleaved, and GSDMD-N levels in vitro. BCL2 promotes cellular survival by preserving the mitochondrial outer membrane, and inactivates the pyroptotic execution program [29,30]. We demonstrated that EBE treatment alleviated the abnormal changes of mitochondrial morphology, nuclear membrane and heterochromatin in cells model by TEM. We analyzed that Bcl2 was the potential target gene regulated by EBE via bioinformatics methods. These experiments confirmed that EBE inhibited pyroptosis of cells models by BCL2.
The robust inflammatory response is another important pathological mechanism of ALI [31]. PTGS2 is induced to increase prostaglandins in response to inflammation and immune responses [32]. Previous research has suggested that the expression of PTGS2 was increased in ALI, and suppressing PTGS2 attenuated ALI induced by LPS [32][33][34]. Our results showed that administration of EBE downregulated PTGS2 expression, and decreased IL-6, IL-1β, NO and PGE2 levels in the cells model. These data verified that Ptgs2 was a potential target gene regulated by EBE, and EBE attenuated inflammation of cells model by PTGS2. We also detected the alteration of TNFα, an important cytokine, but it was not significant in EBE treatment group (not exhibited), additionally, IFNγ is related to the innate and adaptive immune systems; maybe we will consider the effects of E. bodinieri Vaniot on IFNγ in animal models in our following work.
RAW 264.7 is a leukemic monocyte/macrophage cell line of mouse. Previous studies showed that RAW 264.7 was stimulated with LPS to establish a steadily inflammatory model [31]. In order to verify the bioinformatic results of mouse, RAW 264.7 was selected as an inflammatory cells model. However, lung endothelial barrier dysfunction is also an important pathological characteristic of ALI. To further explore the effects of E. bodinieri Vaniot on ALI lung tissue, we will continue working on alveolar epithelial cell lines and animal models.
Failure to resolve the ALI worsening will increase the poor prognosis and mortality of patients [35]. Although a large number of ALI research has been conducted, no effective drug treatments that significantly decrease the mortality have been reported [36]. We aim to explore natural products from E. bodinieri Vaniot that could ameliorate ALI, and serve as a functional food for adjuvant therapy or as a therapeutic agent.

ALI Gene Expression Data Acquisition and Processing
The mRNA expression data of ALI and normal lung tissue were obtained from the GEO database (http://www.ncbi.nlm.nih.gov/geo/). We collected 9 mouse ALI samples and 3 matched control samples in GSE17355 microarray data [37], 6 mouse ALI samples and 6 control samples in GSE2411 [38], and 3 mouse lung samples treated with LPS and 3 corresponded controls in GSE1871 [39]. In order to obtain DEGs of ALI, we used the Limma R package [40] to analyze GSE17355, GSE2411 and GSE1871 datasets based on a criterion of p < 0.05 after adjustment by the false discovery rate (FDR). The clusterProfiler, org.Mm.eg.db, enrichplot, and ggplot2 R packages were used to determine the GO biological processes analysis of overlapping DEGs. To uncover the potential biological functions and interactive networks of overlapping DEGs, KEGG pathway (https://www.genome.jp/kegg/) enrichment analysis was performed using the clusterProfiler R package.

Exploration and Identification of Key Genes
Compounds of E. bodinieri Vaniot were retrieved in CAS SciFinder database (https://accounts.cas.org/products/) [41] and PubChem database (https://pubchem.ncbi. nlm.nih.gov/) [42]. The associated target genes of E. bodinieri Vaniot compounds were predicted and obtained from Swiss Target Prediction database (http://www.swisstargetprediction. ch/) [43]. ALI-related genes were obtained in DisGeNET database (https://www.disgenet. org/) [44]. We screened hub genes from upregulated genes, ALI-related genes, potential target genes of E. bodinieri Vaniot and downregulated genes by VennDiagram R package. GO network of hub genes was investigated by Metascape database (https: //metascape.org/) [45], and Cytoscape software (Version: 3.8.0) was utilized to construct a GO network. The genes-compounds network was constructed with Cytoscape software 3.8.0 based on potential target proteins and corresponding signaling pathways. RAW264.7 cells were seeded at a density of 2.65 × 10 5 cells/well in 6-well cell culture plates, and were stimulated with LPS (1 µg/mL) for 24 h to establish a steadily inflammatory model [31]. According to the instructions, LPS (L8880, Solarbio, Beijing, China) 1 ng is equivalent to 0.5 EU (tachypleus amebocyte lysate method) and 10 EU (chromogenic method). The dose of LPS is 1 µg/mL, which is equivalent to 500 EU/mL (tachypleus amebocyte lysate method) and 10,000 EU/mL (chromogenic method). Our previous studies showed that nine principal compounds were identified from EBE by UHPLC-ESI-HRMS/MS analysis. According to the structural characteristics, the identified compounds included 6 flavanone glycosides (3, 4, 5, 6, 7 and 9) and 3 phenolic acids (1, 2, and 8) ( Table 1). The cells were divided into four groups: LPS was added to the cells after incubation with 50 µg/mL EBE (EBEH) for 24 h; LPS was added to the cells after incubation with 12.5 µg/mL EBE (EBEL) for 24 h; LPS was added to the cells without EBE (Model); nothing was added to the cells without EBE (Control). The viability of RAW 264.7 cells was performed by CCK-8 assay (CK04, DOJINDO, Kumamoto, Japan).

Detection of Inflammatory Cytokines and Oxidative Stress Related Factors
The levels of NO (S0021S, Beyotime, Shanghai, China) and PGE2 (H099-1, Nanjing Jiancheng, Nanjing, China) were assessed by the appropriate kits according to the manufacturer's instructions. The contents of MDA (S0131S, Beyotime, Shanghai, China) and ROS (S0033S, Beyotime, Shanghai, China) were determined according to the instructions of the kits. Data were read on a microplate reader (Biotek, Santa Clara, CA, USA). ROS fluorescent probe were photographed under a fluorescence microscope (Olympus, Tokyo, Japan) and analyzed using Image-Pro Plus 6.0 software.

Transmission Electron Microscopy
The cells were pelleted by centrifugation at 3000 rpm and then fixed in 2.5% glutaraldehyde cold fixing solution for 2 h. The cell mass was sliced to small pieces and pre-fixed in 2.5% glutaraldehyde PBS solution on ice for 2 h. The primarily fixed samples were further sliced to a volume no larger than 2 mm, rinsed with PBS for three times and then post-fixed in 1% osmium tetroxide for 1 h. After dehydration in an ascending gradient ethanol solution, the samples were finally embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with TEM (JEOL, Tokyo, Japan).

Statistical Analysis
All results are presented as the mean ± SEM. A two-tailed unpaired Student's t-test (for two group comparison) and a one-way ANOVA followed by Turkey's post-hoc test (for multi-group comparison) were performed using GraphPad Prism v.8.0.1; p value < 0.05 was considered statistically significant. All experiments were repeated independently at least three times with similar results.

Conclusions
In conclusion, we used bioinformatics methods to explore and analyze potential target genes in ALI regulated by E. bodinieri Vaniot. We confirmed the effects of E. bodinieri Vaniot on cells model of ALI by determining oxidative stress, pyroptosis and inflammation in vitro. This study provides a potential natural source of ameliorating ALI for functional food and pharmaceutical applications.