All procedures and assays were approved by the Institutional Animal Care and Use Committee of Jining Medical University.

We screened the active components according to Oral bioavailability (OB) values and drug-likeness (DL) utilizing Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP) (Ru et al., 2014). The ingredients greater or equal to 0.1 (Drug likeness, DL) and 20% (OB) were included according to TCMSP parameter information and criteria.

The targets of each active ingredient in HYJJ were retrieved from TSMSP database and transformed into gene symbols of Homo sapiens species by the UniProt knowledge database (www.uniport.org).

We collected IBD-related genes from five sources with the keyword “ulcerative colitis,” including GeneCards (Rebhan et al., 1997; Safran et al., 2010), DrugBank (Wishart et al., 2018), Online Mendelian Inheritance in Man (OMIM) (Hamosh et al., 2002), PharmGkb (Whirl-Carrillo et al., 2012), and Statistics of Therapeutic Target Database (TTD) (Wang et al., 2020) (Supplementary Table S1).

We intersected the targeted genes of HYJJ decoction with disease genes (Supplementary Table S4), and constructed a Venn diagram. By Cytoscape software (Su et al., 2014) Herb–Ingredient–Target (HIT) interaction network was visualized.

STRING database (Search Tool for Retrieval of Interacting Genes/Proteins) (http://string-db.org/) (Szklarczyk et al., 2019) was used to construct a protein-protein interaction (PPI) network based on the obtained HIT gene symbols. Full PPI network with both functional and physical protein associations and without disconnected nodes was constructed with a confidence score of ≥0.7 for significance.

Network nodes represent proteins, and edges represent protein-protein associations that proteins jointly contribute to a shared function, but this does not mean they are physically binding each other. The thicker of edge means the higher edge confidence.

Hub genes with a high-degree node were identified by Cytocape plugin CytoNCA according to betweenness centrality (BC), closeness centrality (CC), degree centrality (DC), eigenvector centrality (EC), local average connectivity (LAC) and network centrality (NC) (Tang et al., 2015). These features were calculated twice to establish the final sub-network consisting of hub-genes with higher than the median values of the features.

GO (Ashburner et al., 2000) and KEGG enrichment analysis (Kanehisa and Goto, 2000) of genes were implemented by Bioconductor (R). GO terms with adjusted p-value below 0.05 were considered as significantly enriched genes.

A murine model of ulcerative colitis was established by feeding 4% DSS water to C57 male mice weighing 20–25 g, six mice in a group with or without HYJJ administration. The severity of experimental colitis was defined according to epithelial deconstruction, shortened colon length, and fecal blood. As shown in Figure 1A, HE staining showed that HYJJ markedly alleviated the epithelial structural collapse in UC mouse models. Exacerbated fecal album and fecal blood in DSS group were apparently rescued after HYJJ treatment (Figures 1B,C). Consistently, HYJJ therapy successfully maintained the colon length of UC mouse models. Collectively, in vivo assay indicated that HYJJ could alleviate UC progression.

W In a bid to ascertain the pharmacological mechanism of HYJJ, we screened the active components of HYJJ decoction in the TCMSP database and identified 188 active compounds targeting 2,165 genes. ACD yields 21 components targeting 92 genes, ZR yields five components targeting 54 genes, PCM. yields 22 components targeting 256 genes, GF yields 92 components targeting 1769 genes, AMK yields seven components targeting 23 genes, CR yields five components targeting 95 genes, SCF yields eight components targeting 30 genes, PCW yields 15 components targeting 30 genes, PB yields 13 components targeting 175 genes, CC yields 10 components targeting 116 genes (Supplementary Table S1).

We screened five databases for UC-relevant genes and obtained 5,811 genes after deleting redundant items (Figure 2A, Supplementary Table S2). Venn diagram analysis further showed that HYJJ shared 218 putative targets with UC (Figure 2B). The overlapping gene symbols were detailed in Supplementary Table S3.

To illustrate the pharmacological mechanism of HYJJ, we constructed a HIT network utilizing Cytoscape (Figure 2C). The yellow nodes represent UC genes and the blue nodes represent HYJJ molecules, while colorful dots within the component indicate different herbs, and the edges mean the interaction between UC and components. As illustrated, most herbs not only worked synergistically targeting the same genes, but shared some components, such as GR and ZR, PCM and SCF as well as PB. Furthermore, we explored the role of each herb in UC treatment. The pie chart showed the target genes of each herb in UC (Figure 3A), and GO and KEGG analysis was performed with these genes (Figure 3B). As the primary source of active components, GF and PCM target genes were similarily enriched in responses to lipopolysaccharide and bacterial as well as metal ions, steroid hormone pathway; CR and PB, as well as ZR influenced responses to oxidative stress and metal ions, while PB specifically regulated adrenergic and catecholamine activity and CR was involved in endocrine resistance; Of note was the anti-neoplastic role of the herbs mentioned above, indicating the potential therapeutic capacity of HYJJ in cancer handling. CR was responsible for responses to fatty acid and acid chemicals as well as the regulation of Adenosine 5‘-monophosphate (AMP)-activated protein kinase (AMPK) and peroxisome proliferators-activated receptors (PPAR) signaling pathways that exert an anti-inflammatory role in UC (Arafa et al., 2020; Xu et al., 2017). Other seemingly minor but vital herbs were ACD, SCF, PSW and AMK. Interestingly, ACD, AMK as well as PSW associated with muscle contraction and vasoconstriction, while AMK and PSW as well as SCF not only worked with neurological processes, including neuroactive ligand-receptor interaction and neurotransmitter receptor activity, but improved digestion processes, such as protein digestion and absorption, and salivary secretion. From the enrichment analysis angle, HYJJ decoction could suppress inflammation, orchestrate hormone activity and restore gut function, thereby holistically combating UC.

We construct a full PPI network with a PPI enrichment p-value (<1.0 e-16) based on the STRING database (Figure 4A) and extracted 29 hub-genes (Figure 4B). The median values of BC, CC, DC, EC, LAC, and NC in the left subnetwork were 76.52928591, 0.4048583, 10, 0.030461067, 5.333333333, and 6.5, respectively. The median values in the right subnetwork (hub genes) were 15.67041167, 0.591562335, 19.5, 0.1026759405, 11.64171123, 13.451165665, respectively. 29 hub genes with higher than the median values of the features were identified (Figure 4B). In line with the individual function of each herb, KEGG analysis demonstrated that the hub-genes in the HIT network were enriched in infection, cancer and endocrine processes.

Network pharmacological analysis not only indicated the anti-inflammatory effect of HYJJ, but also pointed out the anti-neoplastic potential of this decoction. Given that UC has been regarded as an independent risk factor of UC-CRC, we treated DLD-1 cells with HYJJ and extracted RNA for RNA-sequencing. As a result, 784 DEGs were obtained between HYJJ group and the control group. Among these DEGs, 643 genes were down-regulated and 141 were up-regulated (Figure 5A). Based on the STRING database, PPI network composed of down-regulated or up-regulated genes were constructed (Figure B, and C). We then performed GO and pathway enrichment analysis. The top enrichments of up-regulated genes in CC were membrane part, intrinsic component as well as integral component of membrane; In BP were regulation of cell response to heat and xenobiotic stimulus; In MF were molecular function regulator and enzyme regulator activity. Of note, the top enrichment of down-regulated genes in CC was cell periphery, plasma membrane and extracellular region, in BP were response to stimulus, regulation of multicellular organismal process as well as inflammatory response, in MF were receptor binding and cytokine activity (Figure 5E). KEGG analysis showed that HYJJ treatment facilitated cell adhesion molecules, salivary and pancreatic secretion, glutamatergic and dopaminergic synapse, fat digestion and absorption, all of which were mentioned in network analysis about the individual role of AMK and PSW as well as SCF, these seemingly minor components of HYJJ. Not surprisingly, HYJJ down-regulated genes were enriched in inflammation-related pathways, such as cytokine-cytokine receptor interaction, TNF signaling pathway as well as IBD, NOD-like and Toll-like receptor signaling pathways (Figure 5F), fully corroborating the anti-inflammatory as well as the anti-cancer therapeutic role of HYJJ.

Intestinal organoids (IOs) culture is an ideal tool to mimic the stem cell niche and investigate the stemness of ISCs under pathological settings. The ex-vivo IBD cellular model was established by treating IOs with TNFα (20 ng/ml, 24 h), and then HYJJ was administered for another 24 h. TNF is interconnected with ROS, dysregulation of which is a hallmark of cancers and inflammatory diseases (Blaser et al., 2016). Given the inhibition of TNF signaling pathway by HYJJ shown in RNA-seq results, and that excessive reactive oxygen species (ROS) enables stem cells to escape from a long-lived quiescent state-a slow-cycling state and enter into a proliferative state undergoing differentiation or apoptosis (Tothova et al., 2007; Zhang et al., 2018), we examined ROS level as well as apoptosis of IOs. As illustrated in Figure 6A, TNFα increased ROS level, an effect that was reversed by HYJJ. The swollen shape and dark lumen of IOs in the TNFα group were also rescued by HYJJ administration (30 ng/ml, 24 h). Moreover, the number of multi-buds IOs was considerably decreased after TNFα treatment, which was successfully reversed by HYJJ (Figure 6B). Western blotting results showed the increased apoptosis in the presence of TNFα was suppressed by HYJJ, as evidence by the protein expression of BAC and BCL-2 (Figure 6C). Altogether, HYJJ facilitates the re-establishment of the ISCs niche in an ex-vivo model of ulcerative colitis.

Mφs, antigen-presenting cells with high plasticity linking the innate and adaptive immune system, transit to the pro-inflammatory phenotype (M1 Mφs) and secret interleukins to amplify the inflammation processes, such as IL-2 and IL-12, as well as IL-10. Nevertheless, Mφs facilitate wound-healing when they act as guardians of the homeostasis and present M2 phenotype.

ELISA confirmed that HYJJ administration inhibited the levels of IL-2, IL-10 as well as IL-12 in the serum collected from mice undergoing experimental colitis (Figure 7A). Moreover, We established a co-culture system by Mφs and IEC-6 cells to mimic the inflammatory microenvironment around epithelial cells. Firstly, we isolated peritoneal Mφs from HYJJ-DSS mice and co-cultured these Mφs with IEC-6 cells for 24 h. A scratch assay showed that Mφs from DSS mice hampered the migration capacity of IEC-6 cells, but no detrimental role of HYJJ-DSS Mφs was observed (Figure 7B). Altogether, HYJJ might inhibit M1 Mφs transition and hence suppressing the pro-inflammatory interleukins levels in DSS-induced colitis mice.