Decoding the key compounds and mechanism of Shashen Maidong decoction in the treatment of lung cancer

Collection of chemical components and screening of active components in SMD

A total of 523 types of chemical components of SMD were retrieved from TCMSP, ETCM, and SymMap databases (Additional Table 1). Among these components, some had higher concentration than expected in SMD by UPLC-MS/MS method (Table 1) [41], such as rutin, liquiritin, psoralen, xanthotoxin, bergapten, monoammonium glycyrhizinate, ophiopogonin D, methylophiopogonanone A, and methylophiopogonanone B. Previous reports confirmed that high-concentration components of SMD may play important roles in the treatment of lung cancer. By combining the experimental validated high-concentration components (Table 1) [41], and ADMET model predicted components, 284 active components were figured out for subsequent analysis (Additional Table 2, Fig. 2).

Table 1 The UPLC validated high-concentration components (μg/mL). (\(\overline\) ± s, n = 3)Fig. 2figure 2

Distribution of all components and active components. A All components and active components of SMD. B All components and active components of each herb in SMD

Among all active components, 14 components shared two or multiple herbs in SMD (Fig. 3), quercetin, isoquercitrin, rutin, scopoletol, glucuronic acid, 2-pentylfuran, beta-carotene, gynesine, arachic acid, 2-heptanone, linoleic acid, palmitic acid, oleic acid, and octadiene, have been shown to have clear anticancer effects. For example, quercetin has definite anticancer activity and can inhibit a variety of carcinogenic signaling pathways [42]. Rutin promotes the apoptosis of TNF-α-induced A549 human lung cancer cells [43]. Scopoletol can play an anticancer role by triggering apoptosis, blocking the cell cycle, inhibiting cell invasion, and regulating the PI3K/AKT signaling pathway [44]. Quercetin, isoquercitrin, rutin, and scopoletol are also present in G. littoralis, M. alba and G. uralensis, which may be of great importance in the therapeutic mechanism of SMD on lung cancer. These results indicate that shared components may play important therapeutic roles in lung cancer.

Fig. 3figure 3

Common and specific active components in different herbs of SMD. The histogram shows the number of components common to different herbs, and the dark dots represent the original herb of the intersection

Predicting targets of active components in SMD

The 1221 targets of 284 active components were predicted by SwissTargetPrediction, HITPICK, and SEA (Additional Table 3). The active components and their targets were used to construct the CT network. This network contains 1505 nodes and 11,505 interactions (Fig. 4). We analyzed the degree of each component and each target in this network and found that the average degree of the components was 40.51 and the average degree of the targets was 9.42. Ten components with the highest degree were vanillic acid, ferulic acid, dibenzoylmethane, butyl benzoate, lauric acid, nicotinic acid, dibutyl phthalate, salicylic acid, eugenol, n-cis-feruloyltyramine. Most have been reported to be associated with cancer therapy, vanillic acid has antioxidant activity in scavenging free radicals, and thus, it has an effective preventive effect on lung cancer [45]. Lauric acid can be used as a carrier of targeted drugs due to its specific accumulation in lung cancer tissues [46]. The ten targets with the highest degree of the components were MAPT, ESR1, TDP1, ESR2, CYP1B1, ABCG2, ODC1, PTPN1, CYP19A1, and TYR. Most of these genes are reported to be related to the pathogenesis of lung cancer. MAPT has been shown to induce lung cancer cells to gain taxol resistance by activating the PI3K/Akt signaling pathway [47]. ESR1 is a central gene of lung cancer and promotes the occurrence of lung cancer by regulating the p53 signaling pathway and the cell surface receptor signaling pathway [48]. Overexpression of TDP1 is closely related to tumorigenesis, and is a crucial target for tumor treatment, the TDP1 inhibitor can significantly increase the antitumor effect of drugs [49]. These analyses demonstrate that one component can regulate multiple targets and in turn, one target is regulated by multiple components, which reflects the characteristics of the “multicomponents-multitargets” theory in treating complex diseases of TCM.

Fig. 4figure 4

Pharmaceutical component-target interaction network. The red nodes represent pharmaceutical components, and the blue nodes refer to the targeted genes. The size of the nodes represents their degree

Construction of the pathogenic genes weighted network

The pathogenic gene data set was retrieved and downloaded from the DisGeNET database and 2973 pathogenic genes with the number of supporting publications greater than or equal to the mean value were retained (Additional Table 4). The number of supporting publications reflects the correlation of genes associated with lung cancer. A higher number of supporting publications suggests that genes have more associations with lung cancer. After counting the number of supporting publications for 2973 pathogenic genes (Fig. 5A), we found that more than half of these pathogenic genes have only one supporting reference and 33 pathogenic genes with more than 40 supporting references. The top ten genes with the highest number of supporting publications are EGFR, TP53, KRAS, ALK, GSTM1, CDKN2A, CYP1A1, ERBB2, BCL2, and MET. The epidermal growth factor receptor (EGFR) is a transmembrane glycoprotein of the ErbB family of tyrosine kinase receptors, and activated mutations of EGFR are a remarkable feature of lung cancer [50]. EGFR-activating mutations are highly sensitive to tyrosine kinase inhibitor gefitinib, so gefitinib is commonly used in clinical targeted treatment of NSCLC [51]. TP53-induced glycolytic phosphatase (TIGAR) is a key regulator of glycolysis and apoptosis, which can protect cells from oxidative stress-induced apoptosis and provide the necessary conditions for the survival of cancer cells [52]. Mutations exist widely in many types of lung cancer [53, 54]. KRAS is a potential oncogene and has been reported to have a high mutation rate, which makes cancer cells escape apoptosis-induced cell death [55]. The rearrangement of anaplastic lymphoma kinase (ALK) plays an important role in promoting the occurrence and development of lung cancer [56], thus it becomes an important clinical targeted treatment of ALK-positive NSCLC [57]. To explore whether it is reliable to measure the importance of gene function by the number of relevant supporting publications, we constructed a KEGG and GO analysis of pathogenic genes of lung cancer (Fig. 5B). The results showed that there is a positive correlation between supporting publications and the functional pathways involved by these genes, as well as GO terms. Genes with more supporting publications are associated with a highest number of pathways involved.

Fig. 5figure 5

The number of supporting publications and involved pathways of pathogenic genes in lung cancer. A The number and distribution of supporting publications; B Average number of pathways and GO terms related to pathogenic gene in a distinct interval of supporting publications

Key functional network selection and validation

In formulas that treat complex diseases, some components play major therapeutic roles, some play auxiliary roles, and others play antagonistic roles. The group of components with major therapeutic roles is usually considered KFC. The KFC and their targets form key functional network, embedded within the complex CT network. How to detect the key functional network and KFC is the basis for optimizing TCM formulations. Additionally, the process of drug treatment of diseases is a continuous process of drug interventions through protein–protein interactions. Based on this information, we integrated the CT network and pathogenic gene weighted network to construct the comprehensive CTP network (Additional Fig. 1). Then we designed a new node importance calculation method to capture the key functional network. The nodes larger than the median important values of all nodes in the network were retained, and these nodes and their interactions were defined as the key functional network.

To test the reliability of the node importance calculation method, we first performed GO enrichment analysis on targeted genes and pathogenic genes of lung cancer and considered the intersection of GO terms as the effective GO terms to serve as a reference for further comparison. Comparing our proposed node importance detection method with the other traditional methods (Fig. 6A), such as Radiality, Closeness to center, Degree, Neighborhood Connectivity Clustering coefficient and Average Shortest Path length, we found that our method covered up to 97.66% of effective GO terms, which is higher than Radiality 96.02%, Degree 95.74%, Neighborhood Connectivity 86.66%, and the Clustering Coefficient 77.54%. Figure 6B shows that our model also covered the most pathways, thus it indicated that the key functional network detection model we designed could retain the key intervention information.

Fig. 6figure 6

Validation of key functional networks. A The Venn diagrams display the number of elements of the seven models that overlap with effective GO terms. The red cycle represents the effective GO terms, and the blue cycle represents the GO terms predicted by different models. B Comparison of our proposed models with other traditional models on the coverage of enriched GO terms and pathways

Find key functional components

After extracting the key functional network, we used the CDR model to deduce which components could retain the key functional network information to the maximum extent. Finally, we obtained 82 components that defined the KFC (Additional Table 5, Fig. 7). In the KFC, the CDR of the first 8 components reached 50% target coverage, and the 82 components reached 90% target coverage. Among these components, vanillic acid has been reported to have antioxidant activity in scavenging free radicals and had a significant and effective preventative role in B(a)P-induced lung cancer [45, 46]. Lauric acid is a good carrier of targeted drugs for lung cancer [46] and can also inhibit the expression of carcinogenic miRNA and significantly up-regulate the expression of some cancer-inhibiting miRNA in KB cells and HepG2 cells [58]. Salicylic acid can maintain the stability of the genome and plays a key role in reducing the risk of cancer. These findings indicate that the lack of salicylic acid will lead to the delay of DNA excision and repair mechanisms, the accumulation of single-strand and double-strand breaks, cell cycle arrest, damage from apoptosis, and will increase the susceptibility to cancer development [59].

Fig. 7figure 7

The CDR model of active components in lung cancer

Effects of key functional components on the viability of A549 cells

Simple random sampling is a method of probability sampling based on chance events, and it can mitigate selection bias [60]. It’s the incorporation of randomization that provides unpredictability in treatment assignments. Based on this selective strategy, protocatechuic acid (SM-32), paeonol (SM-257), and caffeic acid (SM-229) in 82 KFC were selected to determine the effects on the viability of A549 cells using the MTT assay. Protocatechuic acid (SM-32) originates from G. uralensis. Paeonol (SM-257) originates from M. alba. Caffeic acid (SM-229) originates from G. littoralis. After 24 h of incubation, the cell viabilities of A549 cells were 95.07 ± 8.30%, 85.31 ± 6.08%, 60.65 ± 4.01%, 48.25 ± 10.39%, and 39.12 ± 5.88% after exposure to protocatechuic acid at concentrations of 1, 5, 10, 20, and 40 μM, respectively (Fig. 8A). After exposure to paeonol at concentrations of 25, 50, 100, 200 and 400 μM, the cell viability was 97.34 ± 6.58%, 94.33 ± 9.72%, 89.77 ± 6.94%, 75.19 ± 13.70% and 66.79 ± 11.76% (Fig. 8B). When cells were treated with 25, 50, 100, 200, and 400 μM caffeic acid, the cell viabilities were 95.60 ± 7.97%, 93.90 ± 7.26%, 90.27 ± 6.18%, 79.91 ± 5.54%, and 60.89 ± 5.39% (Fig. 8C). The results show that 5–40 μM protocatechuic acid, 100–400 μM paeonol or caffeic acid exerted significant inhibitory activity on the proliferation of A549 cells.

Fig. 8figure 8

Inhibitory Effects of protocatechuic Acid A, paeonol B and caffeic Acid C on the proliferation of A549 cells at 24 h. Data are represented as mean ± SEM (n = 6). *P < 0.05, **P < 0.01, *** P < 0.001 versus the control group. The 2D structures are obtained from PubChem

Possible therapeutic mechanism

ClusterProfiler was used to perform the functional enrichment analysis of the KFC targeted genes, we obtained pathways with a P-value < 0.05. Among these pathways, the small lung cancer and NSCLC pathways are highly correlated with the pathogenesis of lung cancer. Increasing evidence confirms that the downstream gene regulation of the PI3K/Akt signaling pathway (hsa04151) can be changed by targeting the GPCR receptor family through PI3K and Akt. Furthermore, activation of the PI3K/Akt signaling pathway can inhibit apoptosis, promote gene transcription, and cell proliferation, accelerate the cell cycle process, and promote angiogenesis by regulating various downstream activating factors [61,62,63]. The MAPK signaling pathway (hsa04010) also proved to be vital to the occurrence and development of tumor, and activation of the MAPK signaling pathway may lead to increased proliferation, migration, and invasion of tumor cells [64].

To further explore the synergistic effects of KFC targets in different pathways, we combined the enrichment pathways as a comprehensive pathway. The comprehensive pathway included small lung cancer (hsa05222), NSCLC (hsa05223), the PI3K/Akt signaling pathway (hsa04151), and the MAPK signaling pathway (hsa04010) (Fig. 9, Additional Fig. 2). In the combined pathways, some genes products sharing multiple pathways are named as cross-talk gene products, and main cascade targeting module merged with CDR-predicted comprehensive pathways. Some cross-talk gene products appear intermediately in merged pathways, including Ras, PKC, Raf1, MEK, ERK, CDK4/6, CyclinD1, AKT, IKK, NF-κB, and NUR77. In this module, Ras and AKT frequently regulates downstream receptors, which were also predicted as cross-talk gene products, such as Raf1, MEK, IKK, NF-κB, and so on. The downstream receptors of cross-talk gene products were reported relevant to proliferation and cell survival in other pathways. Myc and POLK, which have indirect effects on increased survival and cell cycle progression, are still worthy of attention. With the prediction of CDR model and the analysis of merged pathways, KFC could act on these targets and deprive indispensable conditions for the proliferation and long-term survival of cancer cells. In this way, therapeutic mechanism could be achieved possibly.

Fig. 9figure 9

Main cascade targeting module merged by CDR-predicted cascade pathways. The red units denote the KFC targeted cross-talk genes shared by multiple pathways. The blue units indicate that the KFC targeted genes exist only in one pathway. The white units are annotations or nontargeted proteins

留言 (0)

沒有登入
gif