Exploring the underlying mechanism of action of a traditional Chinese medicine formula, Youdujing ointment, for cervical cancer treatment

Abstract

Background: A traditional Chinese medicine formula, Youdujing (YDJ) ointment, is widely used for treatment of human papilloma virus-related diseases, such as cervical cancer. However, the underlying mechanisms by which active compounds of YDJ alleviates cervical cancer are still unclear.

Methods: We applied a comprehensive network pharmacology approach to explore the key mechanisms of YDJ by integrating potential target identification, network analysis, and enrichment analysis into classical molecular docking procedures. First, we used network and enrichment analyses to identify potential therapeutic targets. Second, we performed molecular docking to investigate the potential active compounds of YDJ. Finally, we carried out a network-based analysis to unravel potentially effective drug combinations.

Results: Network analysis yielded four potential therapeutic targets: ESR1, NFKB1, TNF, and AKT1. Molecular docking highlighted that these proteins may interact with four potential active compounds of YDJ: E4, Y2, Y20, and Y21. Finally, we found that Y2 or Y21 can act alone or together with E4 to trigger apoptotic cascades via the mitochondrial apoptotic pathway and estrogen receptors.

Conclusion: Our study not only explained why YDJ is effective for cervical cancer treatment, but also lays a strong foundation for future clinical studies based on this traditional medicine.

Graphical abstract

Keywords

Youdujing / cervical cancer / traditional Chinese medicine / network pharmacological / molecular docking / synergy effect

1 INTRODUCTION

Cervical cancer (CC) is one of the most common malignant tumors afflicting women [1,2]. Indeed, it has been estimated that in 2018 approximately 570,000 new cases and 310,000 deaths were due to this disease worldwide [3]. In particular, the incidence of CC in developing countries is much higher than in developed countries because of inadequate screening [3]. Although preventive vaccines against human papilloma virus (HPV) can decrease the incidence and mortality of CC [4], these are neither effective for infected patients [5] nor affordable for patients in developing countries [6]. Therefore, drug-based treatments are becoming the most common method to cure HPV-infected patients in developing countries [7–9]. Although platinum-based chemotherapy has been used to treat CC for years, the survival rate of CC patients did not improve significantly until bevacizumab was developed [1]. However, bevacizumab cannot be considered an ideal drug because of its high price [10], side effects [11] and incomplete effectiveness [12]. For this reason, it is urgent to find an effective, low-cost, and side effect-free treatment for CC.

Recently, clinical use of traditional Chinese medicine (TCM) formula Youdujing (YDJ) was reported to treat multiple diseases including HPV-dependent CC [13,14] because of its moderate therapeutic effects and lesser side effects [15,16]. Interestingly, Xiao et al. [17] found that YDJ can counteract HPV infection and lead to an almost complete recovery of local lesions. However, the mechanisms underlying the effect of YDJ remained unclear due to the following reasons. First, the etiology of CC is so complex that it is difficult to identify the most effective therapeutic targets. Second, since this TCM formula includes hundreds of chemical compounds acting on multiple targets, it is difficult to pinpoint its active compounds and localize the sites where these exert their effects. Third, because there are too many drug combinatorial possibilities [18,19], it is difficult to explore optimal drug combinations exhibiting the best clinical efficacy and the least side effects.

To address these issues, we developed a network pharmacology method to investigate the active compounds of YDJ ointment and their possible combinations by integrating network [18,20] and enrichment analyses [21–29] with molecular docking [19,30]. First, we identified potential therapeutic targets using network and enrichment analyses to explore hub nodes and their related biological processes. Second, we utilized molecular docking to compute the binding energies between the chemical compounds of YDJ and their potential therapeutic targets to predict the active compounds of YDJ. Finally, we used network-based proximity parameters to investigate the optimum drug combinations.

We obtained the following interesting findings. First, network and enrichment analyses yielded four potential therapeutic targets: ESR1, NFKB1, TNF, and AKT1, which were confirmed by manual review of existing literature [31–34]. Second, molecular docking allowed to identify four potential active compounds: E4, Y2, Y20, and Y21. Interestingly, E4, Y2, and Y21 were shown to promote apoptosis [35–37]. Third, after using network-based proximity to estimate the relationships between E4, Y2, and Y21, we considered Y2-E4 and Y21-E4 as potential drug combinations and demonstrated that Y2 or Y21 can act together with E4 to trigger apoptotic cascades via the mitochondrial apoptotic pathway [38] and/or estrogen receptors [39].

In general, this study not only describes the reason why YDJ is effective for CC treatment, but also offers a general procedure to investigate the underlying mechanism of action of TCM formulas.

2 RESULTS

2.1 Step 1: Data collection

The YDJ ointment consists of three herbs, Brucea javanica (L.) Merr., Curcuma phaeocaulis val., and Arnebia euchroma [14], from which we collected 30, 21, and 10 active chemical compounds, respectively, as illustrated in step 1 of Fig. 1. MedChem Studio [40] was used to identify the potential targets of these compounds. After examining duplicate information and removing redundancies, we obtained 320 potential targets, which are detailed in Supplementary Tables S1–S3. Next, we downloaded the sequences of 5003 proteins known to be targeted during CC treatment from the GeneCard database [41]. Finally, we intersected the 320 YDJ potential targets with the 5003 CC-related proteins to obtain 119 candidate targets (Supplementary File 1) at the end of step 1.

2.2 Step 2: Identify potential key targets of YDJ involved in cervical cancer

In step 2, network and enrichment analyses were employed to identify key targets of YDJ involved in CC to answer our first scientific question: “How to identify the most effective therapeutic targets for the complicated etiology of CC”. We first used two topological features (degree and BC) to identify hub nodes in the protein-protein interaction (PPI) network; then, we carried out enrichment analysis to identify genes involved in important Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Finally, we intersected these results to pinpoint potential key targets.

2.2.1 Using network analysis to predict hub nodes

First, we input the above-mentioned 119 putative targets (step 2 of Fig. 1) into the STRING database to compute PPI scores, whose threshold was set to 0.9 (the highest confidence). Disconnected nodes were hidden. The generated PPI network included 90 nodes and 203 edges. Next, we examined the topological features illustrated in the section of Materials and Methods (i.e., degree and betweenness centrality, BC) to identify the potential key targets of YDJ (Fig. 2). We obtained 10 hub nodes (Table 1) regarding to the thresholds of degree and BC (Supplementary Table S4). In addition, Fig. 2 shows that the average node degree was 3.41, and the node degree distribution in the PPI network followed a power law with a long tail.

2.2.2 Using enrichment analysis to identify targets

To understand the pharmacological mechanisms of the effect of YDJ on CC, we carried out enrichment analysis of the proteins of the PPI network depicted in Fig. 2 using DAVID v6.8 [42]. Supplementary Table S4 indicates the threshold of false discovery rates (FDR) applied for Gene Ontology (GO) and KEGG enrichment analysis. These enrichment analyses yielded 140 GO biological processes and 73 KEGG pathways, which are detailed in Supplementary Tables S5–S6.

Since previous epidemiological studies have demonstrated that HPV infection [43], smoking [44] and oral contraceptive use [45] are important causes of CC, we considered that drugs acting on these pathways and their corresponding receptors could be effective for CC treatment [46]. Thus, we highlighted GO biological processes (Table 2) and KEGG pathways (Table 3) related to HPV infection [43], smoking [44] and oral contraceptive use [45] from Supplementary Tables S5–S6.

Table 2 shows the GO biological processes predicted to be most related to CC: regulation of reactive oxygen species metabolic process (GO2000377) and response to steroid hormone (GO0048545). Table 3 shows the KEGG pathways predicted to be most related to CC: oxidative phosphorylation (hsa00190), human papillomavirus infection (hsa05165), estrogen signaling pathway (hsa04915), and steroid hormone biosynthesis (hsa00140).

Finally, we intersected the results of network analysis (Table 1) and enrichment analysis (Table 3) to obtain the four key targets listed in Table 4.

2.3 Step 3: Screening of potential active compounds

To answer our second scientific question: “since this TCM formula includes hundreds of chemical compounds acting on multiple targets, how to pinpoint its active compounds and localize the sites where these exert their effects”, we followed the procedure of Chen et al. [51] as illustrated in step 3 of Fig. 1. First, we calculated the binding energy threshold ({{\displaystyle \mathrm{\Delta }G}}_{i,j}^{\text{'}}) between targets i (i = ESR1, TNF, AKT1, and NFKB1) and their intrinsic ligands j (j =XBR [47], 307 [48], XM1 [49] and Compound 5 [50]) by Autodock Vina [30]. The results are listed in Supplementary Table S7.

Second, we calculated the binding energy ({{\displaystyle \mathrm{\Delta }G}}_{i,j}) between targets i and their ligands j (“Ligands” in Table 4) using Autodock Vina [30]. Then, we identified the active compounds with strong binding energies (Fig. 3) by comparing the binding energy of each target-ligand pair ({{\displaystyle \mathrm{\Delta }G}}_{i,j}) with their corresponding thresholds ({{\displaystyle \mathrm{\Delta }G}}_{i,j}^{\text{'}}).

As shown in Fig. 3A, we determined that seven compounds could bind to ESR1, whereas only compounds Y20 and E4 displayed lower binding energy than the intrinsic ligand XBR. Moreover, three compounds could bind to TNF, whereas only compounds Y2 and Y21 exhibited lower binding energy than the intrinsic ligand 307 (Fig. 3B). Conversely, only one compound, E4, could bind to AKT1, but this showed a greater binding energy than the intrinsic ligand XM1 (Fig. 3C). Finally, four compounds could bind to NFKB1, although only compound Y2 displayed lower binding energy than the intrinsic ligand Compound 5 (Fig. 3D). Therefore, we believe that Y2, Y20, Y21, and E4 could be potential active compounds of YDJ (Table 5).

2.4 Step 4: Predict potential drug combinations

To answer our third scientific question: “because there are too many drug combinatorial possibilities [18,19], how to explore optimal drug combinations exhibiting the best clinical efficacy and the least side effects”, we used a network-based method [18] to identify effective drug combinations, as illustrated in step 4 of Fig. 1. First, we determined the potential target disease module based on the function of the above-mentioned compounds (Table 5). Y2, Y21, and E4 were found to be related to apoptosis, while the function of Y20 was unclear. Thus, we removed Y20 from the list of potential active compounds and considered the apoptotic process as the disease module of interest. Second, we computed network-based proximity (S_AB) to measure the relationships between Y2, E4, and Y21. S_Y2-Y21, S_Y2-E4, and S_Y21-E4 were found to equal –0.207, 0.333, and 0.200, respectively.

As described in the section of Materials and Methods, a S_Y2-Y21<0 indicated that the targets of Y2 and Y21 were adjacent to each other in the PPI network (Fig. 4A). A close network proximity of two targets indicates highly similar biological functions. Conversely, E4 targets were separated from those of Y2 or Y21 (Fig. 4B,C); consequently, E4 and Y2 (or Y21) were predicted to be pharmacologically different. Only in this case, compound pairs show statistically significant efficacy during combinatorial therapies [18,56,57]. Therefore, we selected the drug pairs characterized by S_AB>0 and considered Y2-E4 and Y21-E4 as effective drug combinations.

3 DISCUSSION

This research aimed to develop a network pharmacology framework (Fig. 1) to investigate the active compounds of YDJ and their possible combinations for CC treatment. First, we collected available data on CC-related proteins (Supplementary File 1), chemical compounds of YDJ, and their corresponding targets (Supplementary Tables S1–S3).

Second, we identified four potential key targets: ESR1, TNF, AKT1, and NFKB1 by intersecting the results of network analysis (Table 1) and enrichment analysis (Tables 2 and 3). ESR1 was found to be closely associated with CC [58]. As CC can depend on excessive estrogen stimulation, drugs acting on ESR1 could contribute to the treatment and/or prevention of CC [59]. Moreover, James et al. showed that the E6 protein of HPV16 can inhibit TNF-induced apoptosis, thereby contributing to cell immortalization and carcinogenesis [32]. Therefore, the stimulation of TNF-induced apoptosis could have a great impact on the treatment of CC [60,61]. Furthermore, Cui et al. demonstrated that SLUG can suppress the proliferation of CC cells by inhibiting the expression of AKT1 [34]. Finally, NFKB1 plays an important role in inflammation, which can induce accumulation of reactive oxygen species (ROS), thereby leading to an increased risk of CC [33]. In recent years, many studies have demonstrated that natural antioxidants can regulate cellular signal transduction pathways through the activation or inhibition of NFKB1 [33,62,63]. Since these four targets corresponded to the hub nodes in the PPI network (Table 1), we considered them to be potential therapeutic targets.

Third, we analyzed the relationships between the four targets and their corresponding ligands by molecular docking (Table 4) and identified four active compounds of YDJ: Y2, Y20 Y21, and E4 (Fig. 3 and Supplementary Table S7). Y2, Y21, and E4 were found to be related to apoptosis, while the function of Y20 was unclear (Table 5). E4 is a dihydroxyflavanone with versatile pharmacological activities. For example, many studies have demonstrated that E4 can inhibit cancer cell growth by inducing apoptosis [35,55]. In addition, Zhao et al. proved that E4 can interact with the human estrogen receptors ESR1 and ESR2 [64]. Y21 is a triterpene, reported to significantly inhibit cancer cell proliferation and promote apoptosis [37,54]. Finally, Y2 is a steroid saponin, which has been demonstrated to be effective for inducing apoptosis [36,52,53].

Fourth, we used a network-based method to identify two potential drug combinations: Y2-E4 and Y21-E4 (Fig. 4). We predicted that E4 can act together with both Y21 and Y2 to trigger cell apoptosis through mitochondrial apoptotic pathways and estrogen receptors (Fig. 5).

Based on the principle of Sovereign-Minister-Assistant-Envoy (Jun-Chen-Zuo-Shi in Chinese) [65,66], we propose Y2 and Y21, the active compounds of B. javanica (L.) Merr., as the sovereign of the therapeutic process, since they can inhibit NFKB1 and promote TNF-induced apoptosis [60,61]. In addition, we consider E4, the active compound of C. phaeocaulis val., as a minister acting in synergy for the induction of apoptosis. Indeed, E4 can generate increased ROS accumulation to promote the activity of tumor suppressor proteins, thereby triggering apoptotic cascades through mitochondrial apoptotic pathways (Fig. 5) [38,67]. Moreover, E4 may trigger apoptotic cascades via estrogen receptors. In fact, E4 exhibited estrogenic activity with an EC₅₀ value of 67 µM [64] and ESR1 can initiate estrogen-triggered apoptotic pathways [39]. Therefore, we conclude that E4 is a phytoestrogen that can compete with estrogen for binding to ESR1 to promote the activation of apoptotic pathways (Fig. 5).

Here we investigated the potential targets, active compounds, and drug combinations enabling YDJ-mediated CC treatment using a network pharmacology approach. Four proteins related to apoptosis (ESR1, TNF, AKT1, and NFKB1) were highlighted as potential therapeutic targets. Moreover, the compounds E4, Y21, and Y2 exhibited greater binding energies for ESR1, TNF, and NFKB1, respectively, than the respective thresholds. Finally, Y2-E4 and Y21-E4 were denoted as potentially effective drug combinations. Indeed, Y2 (or Y21) can act together with E4 to trigger apoptotic cascades via the mitochondrial apoptotic pathway and/or estrogen receptors. However, these findings should be confirmed by further experiments in vitro and in vivo, whose results can be integrated in the present bioinformatic algorithm [26–29,68] in order to optimize our method in the future.

4 MATERIALS AND METHODS

We developed a network pharmacology method to investigate the active compounds of YDJ ointment and their possible combinations, as shown in Fig. 1. In this process, step 1 was to collect information on the active compounds of YDJ and their CC-related targets from public databases. Then, step 2 was to pinpoint hub genes through network analysis. Next, step 3 was to investigate the related biological processes and signaling pathways by enrichment analysis. Lastly, step 4 was to identify the active compounds by molecular docking and to estimate compound-compound relationships using a network-based approach.

4.1 Step 1: Data collection

4.1.1 Predict YDJ potential targets

The identities of the chemical compounds of YDJ ointment were collected from the TCMIP database [69]. To assess the similarity between these compounds and the respective associated drugs (Fig. 1), we calculated the molecular fingerprint [70] of each compound or drug using RDkit, and then we measured the similarity between the compound and the drug using the Tanimoto score in Eq. (1) [71]. The example of fingerprint is listed in Supplementary File 3. Here, we denote fingerprint as a sequence of bits transformed from the molecular.

The Tanimoto score in Eq. (1), an index of the structural similarity between the chemical compounds of YDJ and known drugs, was used to identify active compounds.

(1)\mathrm{T}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{m}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{s}\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{e}(a,b)={\displaystyle \frac{c}{a+b-c}}

where a equals the amount of bit set to 1 in the molecular fingerprint of the YDJ compound, b is the amount of bit set to 1 in the molecular fingerprint of the known drug in DrugBank [72], and c equals the amount of bit set to 1 in both. Herein, according to the default threshold used by TCMIP [69], we set the Tanimoto score threshold to 0.8. Compounds were considered active, if Tanimoto score was greater than the threshold.

4.1.2 Proteins related to cervical cancer

The candidate target proteins for treatment of CC were obtained from the GeneCard database [41].

4.2 Step 2: Identify potential key targets

4.2.1 Protein-protein interaction network and network analysis

First, we intersected the list of potential targets of YDJ and that of CC-related proteins. Second, we used this intersection to construct a PPI network by STRING (version 11.0) [73]. PPI scores were calculated using Eq. (2).

(2)r_{xy}={\displaystyle \frac{{\displaystyle {\sum }_{i=\mathrm{1}}^n\left(x_i-\overline{x}\right)}\left(y_i-\overline{y}\right)}{\sqrt{{\displaystyle {\sum }_{i=\mathrm{1}}^n{\left(x_i-\overline{x}\right)}^{\mathrm{2}}}}\sqrt{{\displaystyle {\sum }_{i=\mathrm{1}}^n{\left(y_i-\overline{y}\right)}^{\mathrm{2}}}}}},

where x and y represent the gene expression values of different genes, respectively; \bar{x} and \bar{y} represent the mean values of x and y, respectively; and n is the number of experiments. The threshold value of PPI scores was set to 0.9 (the highest confidence). Disconnected nodes were hidden in the PPI network visualized by Cytoscape (version 3.7.2) [74]. Most PPI networks are scale free [75], that is, they are based on a small number of hub proteins (hubs are nodes with high degrees). Hubs represent the most vulnerable points of disease-related protein networks [76], and thus are considered potential therapeutic targets [75].

However, since Lamb et al. demonstrated that the degree might not be the only influential parameter for the identification of crucial proteins from PPI networks [77], we considered BC, described by Eq. (3), as another important parameter for pinpointing therapeutic targets [78,79].

(3)C_i={\displaystyle \sum \frac{{\partial }_{st,i}}{{\partial }_{st}}},

where {{\displaystyle \partial }}_{st} represents the number of shortest paths from node s to node t, and {{\displaystyle \partial }}_{st,i} indicates the number of shortest paths connecting node s and node t through node i. BC measures the importance of node i in mediating the shortest path between node s and node t. Therefore, the network topology structure was based on these two topological features.

4.2.2 Enrichment analysis

In order to elucidate the effect of YDJ on CC, we used the Database for Annotation, Visualization, and Integrated Discovery (DAVID v6.8) [42] to carry out GO [80] and KEGG [21] enrichment analysis. Next, we combined network and enrichment analysis to identify potential targets according to their roles in the network and their biological functions.

4.3 Step 3: Screen potential active compounds

We used Eq. (4) to predict the affinity between a given active compound j and a potential therapeutic target i with Autodock Vina [30].

(4){{\displaystyle \Delta G}}_{i,j}=\sum \left[\begin{array}{l}{{\displaystyle \omega }}_{\mathrm{1}}{{\displaystyle Gauss}}_{\mathrm{1}}(d)+{{\displaystyle \omega }}_{\mathrm{2}}{{\displaystyle Gauss}}_{\mathrm{2}}(d)\\ +{{\displaystyle \omega }}_{\mathrm{3}}\mathrm{Re}pulsion(d)\\ +{{\displaystyle \omega }}_{\mathrm{4}}Hydrophobic(d)+{{\displaystyle \omega }}_{\mathrm{3}}HBond(d)\end{array}\right]

where d is the surface distance, and {{\displaystyle \omega }}_{\mathrm{1}}, {{\displaystyle \omega }}_{\mathrm{2}}, {{\displaystyle \omega }}_{\mathrm{3}}, {{\displaystyle \omega }}_{\mathrm{4}} and {{\displaystyle \omega }}_{\mathrm{5}} are weights.

The 3D chemical structures of the active compounds were retrieved from PubChem. The crystal structures of potential therapeutic targets were obtained from the RCSB Protein Data Bank and visualized using Autodock Tools [81]. Herein, the docking sites of each target were determined by co-crystallization experiments.

4.4 Step 4: Predict potential drug combinations

We used network-based proximity Eq. (5) [18] to estimate compound–compound relationships.

(5)S_{AB}=\langle d_{AB}\rangle -{\displaystyle \frac{\langle d_{AA}\rangle +\langle d_{BB}\rangle }{\mathrm{2}}},

where \langle {{\displaystyle d}}_{AA}\rangle and \langle {{\displaystyle d}}_{BB}\rangle are the shortest distances within the interactome between the targets of each compound, and \langle {{\displaystyle d}}_{AB}\rangle and is the shortest distance between the A-B target pairs. The pseudocodes and workflows for the computation of \langle {{\displaystyle d}}_{AB}\rangle, \langle {{\displaystyle d}}_{AA}\rangle, and \langle {{\displaystyle d}}_{BB}\rangle are listed in Supplementary Table S8. If S_AB<0, the targets of compounds A and B are adjacent to each other, and a close network proximity of two targets indicates highly similar biological functions. In this case, instead of being statistically more effective than monotherapy, these compound pairs exert statistically significant adverse effects [18]. If the S_AB>0, the targets of the two compounds are separated in the network. Therefore, two compounds are pharmacologically different. Only in this situation, these compound pairs show statistically significant efficacy for combinatory therapies [18,56,57].

References

[1] Cohen, P. A., Jhingran, A., Oaknin, A. and Denny, L. (2019) Cervical cancer. Lancet, 393, 169–182

[2] Wu, W., Song, L., Yang, Y., Wang, J., Liu, H. and Zhang, L. (2020) Exploring the dynamics and interplay of human papillomavirus and cervical tumorigenesis by integrating biological data into a mathematical model. BMC Bioinformatics, 21, 152

[3] Bray, F., Ferlay, J., Soerjomataram, I., Siegel, R. L., Torre, L. A. and Jemal, A. (2018) Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, Ca-A Cancer J. Clin., 6, 394–424

[4] Cleveland, A. A., Gargano, J. W., Park, I. U., Griffin, M. R., Niccolai, L. M., Powell, M., Bennett, N. M., Saadeh, K., Pemmaraju, M., Higgins, K., (2020) Cervical adenocarcinoma in situ: Human papillomavirus types and incidence trends in five states, 2008-2015. Int. J. Cancer, 146, 810–818

[5] Roden, R. and Wu, T.-C. (2006) How will HPV vaccines affect cervical cancer? Nat. Rev. Cancer, 6, 753–763

[6] Vaccarella, S., Laversanne, M., Ferlay, J. and Bray, F. (2017) Cervical cancer in Africa, Latin America and the Caribbean and Asia: Regional inequalities and changing trends. Int. J. Cancer, 141, 1997–2001

[7] Toots, M., Ustav, M. Jr, Männik, A., Mumm, K., Tämm, K., Tamm, T., Ustav, E. and Ustav, M. (2017) Identification of several high-risk HPV inhibitors and drug targets with a novel high-throughput screening assay. PLoS Pathog., 13, e1006168

[8] Kolluru, S., Momoh, R., Lin, L., Mallareddy, J. R., and Krstenansky, J. L. (2019) Identification of potential binding pocket on viral oncoprotein HPV16 E6: a promising anti-cancer target for small molecule drug discovery. BMC Mol. Cell. Biol., 20, 30

[9] Celegato, M., Messa, L., Goracci, L., Mercorelli, B., Bertagnin, C., Spyrakis, F., Suarez, I., Cousido-Siah, A., Travé G., Banks, L., (2020) A novel small-molecule inhibitor of the human papillomavirus E6-p53 interaction that reactivates p53 function and blocks cancer cells growth. Cancer Lett., 470, 115–125

[10] Gyawali, B. and Iddawela, M. (2017) Bevacizumab in advanced cervical cancer: issues and challenges for low- and middle-income countries. J. Glob. Oncol., 3, 93–97

[11] Corr, B. R., Breed, C., Sheeder, J., Weisdack, S. and Behbakht, K. (2016) Bevacizumab induced hypertension in gynecologic cancer: Does it resolve after completion of therapy? Gynecol. Oncol. Rep., 17, 65–68

[12] Tewari, K. S., Sill, M. W., Penson, R. T., Huang, H., Ramondetta, L. M., Landrum, L. M., Oaknin, A., Reid, T. J., Leitao, M. M., Michael, H. E., (2017) Bevacizumab for advanced cervical cancer: final overall survival and adverse event analysis of a randomised, controlled, open-label, phase 3 trial (Gynecologic Oncology Group 240). Lancet, 390, 1654–1663

[13] Lin, J., Chen, L., Qiu, X., Zhang, N., Guo, Q., Wang, Y., Wang, M., Gober, H. J., Li, D. and Wang, L. (2017) Traditional Chinese medicine for human papillomavirus (HPV) infections: A systematic review. Biosci. Trends, 11, 267–273

[14] Chen, X., Hu, X., Liu, L., Liang, X. and Xiao, J. (2019) Extracts derived from a traditional Chinese herbal formula triggers necroptosis in ectocervical Ect1/E6E7 cells through activation of RIP1 kinase. J. Ethnopharmacol., 239, 111922

[15] Tang, F., Zhang, Q., Nie, Z., Yao, S. and Chen, B. (2009) Sample preparation for analyzing traditional Chinese medicines. Trends Analyt. Chem., 28, 1253–1262

[16] Gao, H., Wang, Z., Li, Y. and Qian, Z. (2011) Overview of the quality standard research of traditional Chinese medicine. Front. Med., 5, 195–202

[17] Xiao, J., Wu, J., and Yu, B. (2012) Therapeutic efficacy of Youdujing preparation in treating cervical high-risk human papilloma virus infection patients. Chinese journal of integrated traditional and Western medicine, 9, 1212–1215, in Chinese

[18] Cheng, F., Kovács, I. A. and Barabási, A.-L. (2019) Network-based prediction of drug combinations. Nat. Commun., 1, 1–11

[19] Zhang, L., Fu, C., Li, J., Zhao, Z., Hou, Y., Zhou, W. and Fu, A. (2019) Discovery of a ruthenium complex for the theranosis of glioma through targeting the mitochondrial DNA with bioinformatic methods. Int. J. Mol. Sci., 20, 4643

[20] Liu, G.-D., Li, Y.-C., Zhang, W. and Zhang, L. (2020) A brief review of artificial intelligence applications and algorithms for psychiatric disorders. Engineering (Beijing), 6, 462–467

[21] Kanehisa, M. and Goto, S. (2000) KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res., 28, 27–30

[22] Zhang, L., Qiao, M., Gao, H., Hu, B., Tan, H., Zhou, X. and Li, C. M. (2016) Investigation of mechanism of bone regeneration in a porous biodegradable calcium phosphate (CaP) scaffold by a combination of a multi-scale agent-based model and experimental optimization/validation. Nanoscale, 8, 14877–14887

[23] Zhang, L., Liu, Y., Wang, M., Wu, Z., Li, N., Zhang, J. and Yang, C. (2017) EZH2-, CHD4-, and IDH-linked epigenetic perturbation and its association with survival in glioma patients. J. Mol. Cell Biol., 9, 477–488

[24] Zhang, L. and Zhang, S. (2017) Using game theory to investigate the epigenetic control mechanisms of embryo development: Comment on: “Epigenetic game theory: How to compute the epigenetic control of maternal-to-zygotic transition” by Qian Wang et al. Phys. Life Rev., 20, 140–142

[25] Zhang, L., Zheng, C., Li, T., Xing, L., Zeng, H., Li, T., Yang, H., Cao, J., Chen, B. and Zhou, Z. (2017) Building up a robust risk mathematical platform to predict colorectal cancer. Complexity, 2017, 1–14

[26] Zhang, L., Xiao, M., Zhou, J. and Yu, J. (2018) Lineage-associated underrepresented permutations (LAUPs) of mammalian genomic sequences based on a Jellyfish-based LAUPs analysis application (JBLA). Bioinformatics, 34, 3624–3630

[27] Xiao, M., Yang, X., Yu, J., and Zhang, L. (2019) CGIDLA: Developing the Web Server for CpG Island related Density and LAUPs (Lineage-associated Underrepresented Permutations) Study, IEEE/ACM Trans. Comput. Biol. Bioinform. 17, 2148–2154

[28] Zhang, L., Bai, W., Yuan, N. and Du, Z. (2019) Comprehensively benchmarking applications for detecting copy number variation. PLOS Comput. Biol., 15, e1007069

[29] Zhang, L., Liu, G., Kong, M., Li, T., Wu, D., Zhou, X., Yang, C., Xia, L., Yang, Z. and Chen, L. (2019) Revealing dynamic regulations and the related key proteins of myeloma-initiating cells by integrating experimental data into a systems biological model. Bioinformatics, btz542

[30] Trott, O. and Olson, A. J. (2010) AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem, 31, 455–461

[31] Brake, T. and Lambert, P. F. (2005) Estrogen contributes to the onset, persistence, and malignant progression of cervical cancer in a human papillomavirus-transgenic mouse model. Proc. Natl. Acad. Sci. USA, 102, 2490–2495

[32] James, M. A., Lee, J. H. and Klingelhutz, A. J. (2006) Human papillomavirus type 16 E6 activates NF-kappaB, induces cIAP-2 expression, and protects against apoptosis in a PDZ binding motif-dependent manner. J. Virol., 80, 5301–5307

[33] Beevi, S. S., Rasheed, M. H. and Geetha, A. (2007) Evidence of oxidative and nitrosative stress in patients with cervical squamous cell carcinoma. Clin. Chim. Acta, 375, 119–123

[34] Cui, N., Yang, W. T. and Zheng, P. S. (2016) Slug inhibits the proliferation and tumor formation of human cervical cancer cells by up-regulating the p21/p27 proteins and down-regulating the activity of the Wnt/beta-catenin signaling pathway via the trans-suppression Akt1/p-Akt1 expression, Oncotarget. Article, 7, 26152–26167

[35] Rasul, A., Millimouno, F. M., Ali Eltayb, W., Ali, M., Li, J. and Li, X. (2013) Pinocembrin: a novel natural compound with versatile pharmacological and biological activities. BioMed Res. Int., 2013, 379850

[36] Rajavel, T., Mohankumar, R., Archunan, G., Ruckmani, K. and Devi, K. P. (2017) Beta sitosterol and Daucosterol (phytosterols identified in Grewia tiliaefolia) perturbs cell cycle and induces apoptotic cell death in A549 cells. Sci. Rep., 7, 3418

[37] Wang, G., Wang, Y.-Z., Yu, Y., Wang, J.-J., Yin, P.-H. and Xu, K. (2019) Triterpenoids extracted from Rhus chinensis Mill act against colorectal cancer by inhibiting enzymes in glycolysis and glutaminolysis: network analysis and experimental validation. Nutr. Cancer, 72, 293–319

[38] Kumar, M. S., Nair, M., Hema, P., Mohan, J., and Santhoshkumar, T. (2007) Pinocembrin triggers Bax-dependent mitochondrial apoptosis in colon cancer cells. Mol. Carcinog., 3, 231-241. Mol Carcinog., 46, 231–41

[39] Galluzzo, P., Martini, C., Bulzomi, P., Leone, S., Bolli, A., Pallottini, V. and Marino, M. (2009) Quercetin-induced apoptotic cascade in cancer cells: antioxidant versus estrogen receptor α-dependent mechanisms. Mol. Nutr. Food Res., 53, 699–708

[40] Miller, D. W. (2012) Improving potency and ADMET properties using matched molecular pair analysis. Abstracts of Papers of the American Chemical Society

[41] Safran, M., Dalah, I., Alexander, J., Rosen, N., Stein, T., Shmoish, M., Nativ, N., Bahir, I., Doniger, T., Krug, H., (2010) GeneCards Version 3: the human gene integrator. Database, 2010, baq020

[42] Huang, W., Sherman, B. T. and Lempicki, R. A. (2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc., 4, 44–57

[43] Walboomers, J. M. M., Jacobs, M.V., Manos, M.M., Bosch, F.X., Kummer, J.A., Shah, K.V., Snijders, P.J., Peto, J., Meijer, C.J., and Muñoz, N., (1999) Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J. Pathol., 189, 12–19

[44] Moktar, A., Singh, R., Vadhanam, M. V., Ravoori, S., Lillard, J. W., Gairola, C. G. and Gupta, R. C. (2011) Cigarette smoke condensate-induced oxidative DNA damage and its removal in human cervical cancer cells. Int J Oncol, 39, 941–947

[45] Wang, S. S., Zuna, R. E., Wentzensen, N., Dunn, S. T., Sherman, M. E., Gold, M. A., Schiffman, M., Wacholder, S., Allen, R. A., Block, I., (2009) Human papillomavirus cofactors by disease progression and human papillomavirus types in the study to understand cervical cancer early endpoints and determinants. Cancer Epidemiol. Biomarkers Prev., 18, 113–120

[46] Chung, S.-H., Franceschi, S. and Lambert, P. F. (2010) Estrogen and ERalpha: culprits in cervical cancer? Trends Endocrinol. Metab., 21, 504–511

[47] De Savi, C., Bradbury, R. H., Rabow, A. A., Norman, R. A., de Almeida, C., Andrews, D. M., Ballard, P., Buttar, D., Callis, R. J., Currie, G. S., (2015) Optimization of a novel binding motif to (E)-3-(3,5-Difluoro-4-((1R,3R)-2-(2-fluoro-2-methylpropyl)-3-methyl-2,3,4,9-tetrahydro-1H-pyrido[3,4-b]indol-1-yl)phenyl)acrylic Acid (AZD9496), a potent and orally bioavailable selective estrogen receptor downregulator and antagonist. J. Med. Chem., 58, 8128–8140

[48] He, M. M., Smith, A. S., Oslob, J. D., Flanagan, W. M., Braisted, A. C., Whitty, A., Cancilla, M. T., Wang, J., Lugovskoy, A. A., Yoburn, J. C., (2005) Small-molecule inhibition of TNF-α. Science, 310, 1022–1025

[49] Blake, J. F., Kallan, N. C., Xiao, D., Xu, R., Bencsik, J. R., Skelton, N. J., Spencer, K. L., Mitchell, I. S., Woessner, R. D., Gloor, S. L., (2010) Discovery of pyrrolopyrimidine inhibitors of Akt. Bioorg. Med. Chem. Lett., 20, 5607–5612

[50] Park, H., Shin, Y., Choe, H. and Hong, S. (2015) Computational design and discovery of nanomolar inhibitors of IκB kinase β. J. Am. Chem. Soc., 137, 337–348

[51] Chen, L., Du, J., Dai, Q., Zhang, H., Pang, W., and Hu, J., (2014) Prediction of anti-tumor chemical probes of a traditional Chinese medicine formula by HPLC fingerprinting combined with molecular docking. Eur. J. Med. Chem., 83, 294–306

[52] Rajavel, T., Banu Priya, G., Suryanarayanan, V., Singh, S. K. and Pandima Devi, K., (2019) Daucosterol disturbs redox homeostasis and elicits oxidative-stress mediated apoptosis in A549 cells via targeting thioredoxin reductase by a p53 dependent mechanism. Eur J Pharmacol, 855, 112–123

[53] Zhao, C., She, T., Wang, L., Su, Y., Qu, L., Gao, Y., Xu, S., Cai, S. and Shou, C. (2015) Daucosterol inhibits cancer cell proliferation by inducing autophagy through reactive oxygen species-dependent manner. Life Sci., 137, 37–43

[54] Wang, G., Wang, Y.-Z., Yu, Y. and Wang, J.-J. (2019) Inhibitory ASIC2-mediated calcineurin/NFAT against colorectal cancer by triterpenoids extracted from Rhus chinensis Mill. J Ethnopharmacol, 235, 255–267

[55] Zheng, Y., Wang, K., Wu, Y., Chen, Y., Chen, X., Hu, C. W. and Hu, F. (2018) Pinocembrin induces ER stress mediated apoptosis and suppresses autophagy in melanoma cells. Cancer Lett, 431, 31–42

[56] Cokol, M., Chua, H. N., Tasan, M., Mutlu, B., Weinstein, Z. B., Suzuki, Y., Nergiz, M. E., Costanzo, M., Baryshnikova, A., Giaever, G., (2011) Systematic exploration of synergistic drug pairs. Mol. Syst. Biol., 7, 544

[57] Zou, J., Ji, P., Zhao, Y. L., Li, L. L., Wei, Y. Q., Chen, Y. Z. and Yang, S. Y. (2012) Neighbor communities in drug combination networks characterize synergistic effect. Mol. Biosyst., 8, 3185–3196

[58] den Boon, J. A., Pyeon, D., Wang, S. S., Horswill, M., Schiffman, M., Sherman, M., Zuna, R. E., Wang, Z., Hewitt, S. M., Pearson, R., (2015) Molecular transitions from papillomavirus infection to cervical precancer and cancer: Role of stromal estrogen receptor signaling. Proc. Natl. Acad. Sci. USA, 112, E3255–E3264

[59] Chung, S.-H. and Lambert, P. F. (2009) Prevention and treatment of cervical cancer in mice using estrogen receptor antagonists. Proc. Natl. Acad. Sci. USA, 106, 19467–19472

[60] Naka, T., Sugamura, K., Hylander, B. L., Widmer, M. B., Rustum, Y. M. and Repasky, E. A. (2002) Effects of tumor necrosis factor-related apoptosis-inducing ligand alone and in combination with chemotherapeutic agents on patients’ colon tumors grown in SCID mice. Cancer Res, 62, 5800–5806

[61] Chang, K. J.Reid, T., Senzer, N., Swisher, S., Pinto, H., Hanna, N., Chak, A., and Soetikno, R., (2012) Phase I evaluation of TNFerade biologic plus chemoradiotherapy before esophagectomy for locally advanced resectable esophageal cancer. Gastrointes. Endos., 6, 1139–1146. e1132

[62] D’Archivio, M., Santangelo, C., Scazzocchio, B., Varì R., Filesi, C., Masella, R. and Giovannini, C. (2008) Modulatory effects of polyphenols on apoptosis induction: relevance for cancer prevention. Int. J. Mol. Sci., 9, 213–228

[63] Gopalakrishnan, A. and Tony Kong, A. N. (2008) Anticarcinogenesis by dietary phytochemicals: cytoprotection by Nrf2 in normal cells and cytotoxicity by modulation of transcription factors NF-κB and AP-1 in abnormal cancer cells. Food Chem. Toxicol., 46, 1257–1270

[64] Zhao, J., Dasmahapatra, A. K., Khan, S. I. and Khan, I. A. (2008) Anti-aromatase activity of the constituents from damiana (Turnera diffusa). J Ethnopharmacol, 120, 387–393

[65] Wang, L., Zhou, G. B., Liu, P., Song, J. H., Liang, Y., Yan, X. J., Xu, F., Wang, B. S., Mao, J. H., Shen, Z. X., (2008) Dissection of mechanisms of Chinese medicinal formula Realgar-Indigo naturalis as an effective treatment for promyelocytic leukemia. Proc. Natl. Acad. Sci. USA, 105, 4826–4831

[66] Tian, P. (2011) Convergence: Where West meets East. Nature, 480, S84–S86

[67] Chen, Z., Rasul, A., Zhao, C., Millimouno, F. M., Tsuji, I., Yamamura, T., Iqbal, R., Malhi, M., Li, X. and Li, J., (2013) Antiproliferative and apoptotic effects of pinocembrin in human prostate cancer cells. Bangladesh J. Pharmacol., 3, 255–262

[68] Zhang, L., Dai, Z., Yu, J. and Xiao, M. (2020) CpG-island-based annotation and analysis of human housekeeping genes. Brief. Bioinform. bbz134

[69] Xu, H.-Y., Zhang, Y. Q., Liu, Z. M., Chen, T., Lv, C. Y., Tang, S. H., Zhang, X. B., Zhang, W., Li, Z. Y., Zhou, R. R., (2019) ETCM: an encyclopaedia of traditional Chinese medicine. Nucleic Acids Res., 47, D976–D982

[70] Xue, L., Godden, J. W., Stahura, F. L., Bajorath, J. (2003) Design and evaluation of a molecular fingerprint involving the transformation of property descriptor values into a binary classification scheme. J. Chem. Inf. Comput. Sci., 4, 1151–1157

[71] Cereto-Massagué A., Ojeda, M. J., Valls, C., Mulero, M., Garcia-Vallvé S. and Pujadas, G. J. M. (2015) Molecular fingerprint similarity search in virtual screening. Methods, 71, 58–63

[72] Wishart, D. S.,Feunang, Y.D., Guo, A.C., Lo, E.J., Marcu, A., Grant, J.R., Sajed, T., Johnson, D., Li, C., Sayeeda, Z., (2018) DrugBank 5.0: a major update to the DrugBank database for 2018. Nucleic Acids Res., 46(D1), D1074–D1082

[73] Szklarczyk, D., Gable, A. L., Lyon, D., Junge, A., Wyder, S., Huerta-Cepas, J., Simonovic, M., Doncheva, N. T., Morris, J. H., Bork, P., (2019) STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res., 47, D607–D613

[74] Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J. T., Ramage, D., Amin, N., Schwikowski, B. and Ideker, T. (2003) Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res., 13, 2498– 2504

[75] Barabási, A.-L. and Oltvai, Z. N. (2004) Network biology: understanding the cell’s functional organization. Nat. Rev. Genet., 5, 101–113

[76] Albert, R., Jeong, H. and Barabási, A.-L. (2000) Error and attack tolerance of complex networks. Nature. 6794, 378–382

[77] Lamb, J., Crawford, E.D., Peck, D., Modell, J.W., Blat, I.C., Wrobel, M.J., Lerner, J., Brunet, J.P., Subramanian, A., Ross, K.N., (2006) The Connectivity Map: using gene-expression signatures to connect small molecules, genes, and disease. Science, 5795, 1929–1935

[78] Yu, H., Kim, P. M., Sprecher, E., Trifonov, V. and Gerstein, M. (2007) The importance of bottlenecks in protein networks: correlation with gene essentiality and expression dynamics. PLOS Comput. Biol., 3, e59

[79] Goñi, J., Esteban, F. J., de Mendizábal, N. V., Sepulcre, J., Ardanza-Trevijano, S., Agirrezabal, I. and Villoslada, P. (2008) A computational analysis of protein-protein interaction networks in neurodegenerative diseases. BMC Syst. Biol., 2, 52

[80] Consortium, G. O. (2004) The Gene Ontology (GO) database and informatics resource. Nucleic Acids Res., suppl_1, D258–D261

[81] Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew, R. K., Goodsell, D. S. and Olson, A. J. (2009) AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J. Comput. Chem., 30, 2785–2791

RIGHTS & PERMISSIONS

The Author(s) 2021. Published by Higher Education Press