Network pharmacology of black cumin ( Nigella sativa L.) as a candidate of OMAI in colorectal cancer: in silico study

Colorectal cancer is the third most common cancer globally and the second leading cause of cancer‐related deaths. The management of colorectal cancer requires consideration of various factors due to the non‐selectivity of drugs, meaning that highly effective treatment with lower side effects is needed. Black cumin ( Nigella sativa L.) contains thymoquinone and various other metabolites with potential as anticancer effects. The involvement of various genes and the difficulty of drug development have led to a ashift in the drug development paradigm towards plant‐based medicine that is both multicomponent and synergistic in supporting the resulting pharmacological effects. Network pharmacology can predict the synergistic effect of a multicomponent approach. This study aimed to predict the network pharmacology of black cumin as a candidate for OMAI (“Obat Modern Asli Indonesia”, Indonesian‐origin modern medicine) in colorectal cancer. This research was an in silico study using various ethnobotanical databases and software. The results show that seven metabolites in black cumin are correlated with ten surface receptor proteins, 30 intracellular proteins, and mechanisms involving six colorectal cancer signaling pathways. This result indicates that Nigella sativa L. has potential in OMAI and can be a reference for the development of cancer treatment, especially for colorectal cancer.


Introduction
Cancer is a malignant, autonomous, and uncontrolled cell proliferation process that can spread (metastasize) to sur rounding organs. The number of cases and deaths from cancer is predicted to grow rapidly, and the increasing population, age, and lifestyle changes can increase can cer risk. This problem affects the quality of human re sources due to the impact of cancer on various aspects of life (Singer 2018). Colorectal cancer is the third most com mon cancer globally and is the second leading cause of cancerrelated deaths (FlorescuŢenea et al. 2019). Col orectal cancer can occur sporadically (70%), familial clus ters (20%), and hereditary syndromes (10%) such as fa milial adenomatous polyposis (FAP) and hereditary non polyposis colorectal cancer (HNPCC) and are associated with a poor environment and lifestyle (RecioBoiles and Cagir 2021). According to the Ministry of Health of the Republic of Indonesia (2018), the management of colorec tal cancer is multidisciplinary and requires various factors. Local procedures have a higher recurrence rate, surgical therapy is less effective, especially for metastatic cancer, and chemotherapy and monoclonal antibodies have differ ent side effects. This problem is mainly due to the non selectiveness of the drugs used because they also damage healthy cells, so drugs with high effectiveness and low side effects are needed.
Indonesia consists of thousand islands with abun dant biodiversity, so it has enormous potential as a source of new drug discovery (Sutiono et al. 2017). OMAI ("Obat Modern Asli Indonesia", Indonesianorigin modern medicine) are natural and authentic Indonesian medicines that already have scientific evidence of their safety and efficacy, consisting of Standardized Herbal Medicines and Phytopharmaceuticals (National Agency of Drug and Food Control of the Republic of Indonesia 2020). One of the natural ingredients in Indonesia that have the potential as an anticancer is black cumin (Nigella sativa L.). Thymoquinone (TQ), as the main phytochemi cal in Nigella sativa L., can inhibit the growth of colorectal cancer cells, increase cell morphology changes and induce apoptotic (Kooti et al. 2016; Hosseinzadeh et al. 2017. Most diseases involve different groups of genes. On the other hand, drug development costs a lot and takes a long time (Guney et al. 2016). This problem has caused the drug development paradigm to change its focus from conventional drugs that are one drugone target to plant based treatments containing various chemical compounds with various targets (multicomponent -network targets) (Syahrir et al. 2016). The combination of active com pounds in multicomponent drugs can be synergistic to sup port the resulting pharmacological effects. The synergis tic effect on the multicomponent approach can be carried out quickly and gives more promising results through var ious in silico computational methods. One of them is net work pharmacology, which explains the principles of net work theory between compounds and biological systems and from the results in in vitro and in vivo testing for drug development (Yi et al. 2018).
A preliminary study is very important to provide an initial view in reaching the right conclusions and efficient preclinical research. We use several online databases to predict the compound content and the mechanism of ac tion described based on the proteinstructure relationship of the compound through computer prediction. This study aims to predict the network pharmacology of black cumin (Nigella sativa L.) as a candidate for OMAI in colorec tal cancer. Our research can served as a reference for the development of cancer drugs based on target genes to sup port the creation of new medicines or innovations in cancer treatments that are safe and effective.

Secondary metabolites of black cumin (Nigella sativa L.)
The metabolites used in this study were obtained from the KNApSAcK Family database (http://www.KNApSA cKfamily.com/) and Dr. Duke's Phytochemical and Eth nobotanical Databases (https://phytochem.nal.usda.gov/p hytochem/search). The chemical structure of black cumin metabolites was validated by using PubChem database (ht tps://pubchem.ncbi.nlm.nih.gov) and were generated by Marvin Sketch software.

Prediction of biological activity and protein target of black cumin (Nigella sativa L.) metabolites
The Way2Drug PASS Online (http://way2drug.com/passo nline/) and Swiss Target Prediction (http://www.swisstar getprediction.ch/) were used to predict the biological ac tivity and target proteins of black cumin metabolites in colorectal cancer. All accessible targets were restricted to Homo sapiens.

Metabolites-target-pathway network construction
Visualization of network pharmacology the black cumin (N. sativa L.) was carried out using Cytoscape v.3.8.2 soft ware.

Methods
Prediction of the metabolites contained in black cumin was carried out through the KNApSAcK Family and Dr. Duke's Phytochemical and Ethnobotanical, then gener ated by Marvin Sketch software. The metabolites' bio logical activity and target proteins were predicted through Way2Drug PASS Online and Swiss Target Prediction. GeneCards and KEGG Pathway were used to determine the involvement of target proteins with colorectal can cer. The metabolites mechanism of action was pre dicted through STRING, thus referring to biological ac tivity in the KEGG Pathway. The interaction between the metabolites and their target proteins was predicted through STITCH and visualized through the Cytoscape software to describe a network pharmacology. The research flow chart can be seen in Figure 1.

Results and Discussion
Prediction of metabolites contained in black cumin through KNApSAcK Family and Dr. Duke's Phytochemi cal and Ethnobotanical yielded ten metabolites in the seed portion and five metabolites obtained in the Nigella sativa L. seed oil portion (Table 1). According to Corso et al. (2020), KNApSAcK is a comprehensive database contain ing information on the relationship between metabolites, biological activities, and species. Dr. Duke's Phytochemi cal and Ethnobotanical database has the advantages of free access at no cost and the existence of supporting refer ences that correlate with ethnomedicinal data which shows high significance (Savithramma et al. 2016). Black cumin seeds have an important history in traditional medicine practices, especially in South and Southeast Asia, Arabia, Africa and the Mediterranean, treating several diseases in cluding cancer (Khan et al. 2017).
Prediction of the structure of the selected metabolites through Marvin Sketch is shown in Figure 2. The structure of metabolites varied from simple organic compounds, longchain lipids, to glucosides.
The biological potential of the metabolites was pre dicted through the Way2Drug PASS Online database ac companied by the value of the activity probability (Pa) or the likelihood of "becoming active" (Juan et al. 2020). Way2Drug PASS Online provides predictions of biologi cal activity of more than 4,000 types of biological activity with an average of 95% accuracy based on the structure of organic compounds, both existing and new, enabling screening of unpromising compounds at the earliest stage of a study (Filimonov et al. 2014). All metabolites con tained in black cumin, except nigellidine, have biologi cal potential associated with cancer through various ac tivities with Pa>0.7, which means that the probability of finding activity in the experiment is quite high and the pre dicted metabolites are likely to be close to the structure  analogs drugs of known ( Figure 3). The highest predic tion was found in the quercetin 3(6""feruloylglucosyl) (1>2)galactosyl(1>2)glucoside with a probability of 0.997 on free radical scavenger activity and 0.996 on chemopreventive and antimutagenic activity. Their other activities include HIF1A expression inhibitor, JAK2 ex pression inhibitor, TP53 expression enhancer, apoptotic agonist, transcription factor stimulant, antineoplastic, anti inflammatory, anticarcinogenic, etc. The results of target protein prediction through the Swiss Target Prediction database showed that the metabo lite contents in black cumin have many possibilities of tar geting various proteins from various classes in the body, including those involved in cancer. The proteins analyzed through the GeneCards database resulted a probability>0 (Table 2). This value is derived from the target score range of 0-1 which describes the similarity of the prediction to be true. The greater the probability, the greater the chance of prediction accuracy (Gfeller et al. 2014). The userfriendly graphical interface makes Swiss Target Prediction often used in similar studies (Daina et al. 2019). According to Kononenko et al. (2014), GeneCards also features a key in terface that enables a better user experience, including im proved data and product information consolidation, retain ing legacy content and functionality, and easy navigation to other sites. Prediction of target proteins involved in can cer in these metabolites amounted to 192 target proteins, mostly from the kinase class and several other classes such as oxidoreductases, nuclear receptors, hydrolases, etc.
Prediction of colorectal cancer signaling pathways was obtained from KEGG Pathway. The construction of the interaction network between the target proteins was carried out through STRING and focused on proteins with a minimum interaction score>0.400 (medium confidence) from the range 0-1. This value indicates a biologically sig nificant relationship (interaction) compared to the accom panying evidence. The higher the interaction score, the more biologically meaningful the interaction is. KEGG is a metabolic pathway database containing 16,568 com pounds and provides a reference pathway for predicting metabolic pathways in an organism (Altman et al. 2013). The high coverage, ease of use, and consistent scoring sys tem, as well as additional access features via APIs and applications make STRING widely used in bioinformat ics studies to analyze proteinprotein interactions (Szklar czyk et al. 2019). The results of the intersection analysis showed that the ten target proteins produced are surface receptor proteins (PIK3CA, TGFBR1, EGFR, MAPK1, MAPK3, MAPK8, AKT1, GSK3B, CASP3, MAPK10) that were directly involved in six colorectal cancer signal ing pathways (Table 3).
The construction of the interaction network between the black cumin metabolites and its target protein through STITCH was viewed with a confidence view (default) where the thickness of the edges depicted different lev els of confidence based on the interaction analysis scores showing differences in the strength of the interaction data support (Figure 4). The score indicates a medium confi dence level (>0.400) based on the binding affinity or inhi bition constant (K i ), EC 50 , or IC 50 . The thicker the inter action line, the stronger the bond between the two inter actors. STITCH, a database of proteinchemical interac tions derived from various sources, namely prediction of genomic content, experiments, coexpression, text mining, and knowledge of preexisting protein complexes make it widely used in various kinds of related research (Szklar czyk et al. 2016). The number of intracellular proteins that appear suggests the presence of signal transduction, where the response to changes in chemical signals (extracellular) is converted into intracellular signals.
Based on the interaction analysis, the black cumin metabolites that interact directly with surface receptor pro teins were thymoquinone, thymol, myristicin, and oleic acid (Table 4). In addition, there were also four other metabolites in black cumin that do not interact directly with surface receptor proteins but were related to the pre vious four metabolites, namely linolenic acid, pcymene, nigellimine, and carvone ( Figure 3).
The network pharmacology was visualized with Cy toscape as in Figure 5. The black cumin metabolites in teracted with ten surface receptor proteins and 30 intracel lular proteins, and the mechanism involved six colorectal cancer signaling pathways (Ras signaling pathway, TGFβ signaling pathway, ERK/MAPK signaling pathway, apop totic, PI3KAKT signaling pathway, and Wnt signaling pathway). According to Li et al. (2017), Cytoscape is a friendly and open bioinformatics platform, which demon strates outstanding performance in both virtualization and manipulation of biological networks, including due to the rich functionality of access to other applications or related databases .

PubChem ID Compound name
Apoptotic is a mechanism that leads to cell death in response to internal and external signals, including the re lease of cytochrome c from mitochondrial caspase family proteins (CASP). Initiator CASPs (CASP8, 9, and 10) activate effector CASPs (CASP3, 6, and 7), which medi ate apoptotic through proteolytic cleavage of thousands of proteins. Various pathways have become therapeutic tar gets in oncology that induce apoptotic through the apop totic core pathway, namely the kinase signaling pathway involving AKT, ERK, RAS, RAF, MEK or mTOR. Inhi bition of growth factor receptors, including EGFR, HER2, other members of the ErbB family, MET or NTRK, also causes apoptotic (Carneiro and ElDeiry 2020).
The PI3KAKT signaling pathway is widely activated in all types of cancer known to promote cell growth and survival, inhibit apoptotic, and control metabolism. This pathway is often hyperactivated in cancer due to mutations or deletions in negative regulators. The PI3K pathway is activated by a range of stimuli, including growth fac tors, such as receptor tyrosine kinases (RTKs) including epidermal growth factor receptors (EGFR) and platelet derived growth factor receptors (PDGFR) (Slattery et al. 2018; Barata andOliveira 2019).
Genetic variation in PIK3CA and AKT1 is associated with a strongly increased risk of colon cancer (Slattery et al. 2018). Chromosomal instability in 85% of invasive colorectal cancers accumulates hereditary mutations such as APC, TP53, SMAD4, KRAS, and the catalyticsubunit of PI3K (PIK3CA), which is an adenoma -carcinoma se quence (Vogelstein et al. 2013). The oleic acid in black cumin is predicted to carry a chemical signal received by the surface receptor protein PIK3CA, thereby activating the PI3KAKT signaling pathway as indicated by an inter action score of 0.900 (Table 4).
AKT (protein kinase B or PKB), a component of in tracellular insulin signaling activated by PI3K, is involved in regulating key factors related to cell growth, survival, and proliferation. AKT inhibits glycogen synthase kinase 3α/β, thereby promoting cell viability and proliferation; activates complex I rapamycin (mTORC, consisting of mTORC1 [mTOR,Raptor,mLST8,and PRAS40,Rap tor mLST8,and PRAS40] and mTORC2 [mTOR,Ric tor,mSIN1,and MLST8]) and the ribosomal S6 kinase RPS6KB1 and promotes transcription of the family fork   Barata and Oliveira 2019).
In addition to being activated by oleic acid, the PI3K AKT signaling pathway is also transduced by the TGFβ ligand via the noncanonical TGFβ signaling pathway, MAPK and ROCK signaling pathway pathways. The TGFβ (transforming growth factorbeta) signaling path way is a very important pathway in colorectal cancer tu morigenesis. In normal cells, this pathway plays a role in suppressing growth and tumorigenesis. Still, can cer cells regulate development, proliferation, differentia tion, apoptotic, tissue homeostasis and promote epithelial mesenchymal transition, invasion, and metastasis (TGFβ paradox) (Jung et al. 2017). In addition to the non canonical pathway, there is also a canonical pathway in volving type 1 BMP and TGFβ receptors, activin, and the substrates SMAD1/5/8, SMAD2/3 (RSMAD), and SMAD4, which can bind transcription factors as partners in regulating transcription (Hao et al. 2019). TGFβ sig naling pathway can also induce epithelial to mesenchymal transition (EMT) in cancer cells leading to poor progno sis and minimal chemotherapy benefits (Okita et al. 2018; Sveen et al. 2018. In this regard, Idrus et al. (2019) re vealed that black cumin (Nigella sativa L.) and thymo quinone are associated with EMT, which is regulated by activation of transcription factors, promoting wound heal ing, and reducing the reduction of wound healing tissue inflammation, and preventing organ fibrosis through reg ulation of the EMT process.
The MAPK/ERK (mitogenactivated protein kinase/ extracellular signalregulated kinase) signaling pathway transduced by the noncanonical TGFβ pathway is at the core of a signaling network involved in regulating cell growth, development and division (Koveitypour et al. 2019). Most of the target proteins of black cumin metabo lites are associated with the MAPK/ERK signaling path way in colorectal cancer through surface receptor proteins MAPK1, 3, 8, and 10, TGFBR1, AKT1, EGFR, and CASP3. Guo et al. (2020) stated that the ERK1/2 signaling pathway is involved in cell survival after intestinal injury, and inhibition of this pathway may promote apoptotic of gutinjured cells.
In addition to the TGFβ signaling pathway, the MAPK/ERK signaling pathway is activated by the Ras signaling pathway involving RAF-MEK-ERK/MAPK, which transmits signals downstream and results in the tran scription of genes involved in controlling several cellular mechanisms. Ras communicates external cellular signals to the nucleus, and its altered activation leads to inappro priate cellular activities, including increased cell growth, differentiation, survival, and cancer. Ras activation al lows interaction with several downstream effectors or in tracellular proteins (MAPK3, AKT1, MAPK1, EGFR, PIK3CA, MAPK8, MAPK10) (Santarpia et al. 2012).
Another colorectal cancer signaling pathway is Wnt, in which one of the major regulators or surface receptor proteins involved is GSK3 which is also involved in the PI3KAKT signaling pathway. Increased levels of GSK3 present demonstrate the tumourpromoting action of GSK 3 in certain tumour types and/or by the antiproliferative effect of GSK3 inhibitors, such as in colon and pancreatic cancers (TejedaMuñoz and RoblesFlores 2015).
Our research provides a pharmacological network pre diction of black cumin (Nigella sativa L.) as a candi date for OMAI in colorectal cancer through an in sil ico test. Our results support Hsu et al. (2017), which stated that high doses of thymoquinone showed antipro liferative effects on human colonic adenocarcinoma LoVo cells through downregulation of downregulation of p PI3K, pAkt, pGSK3B and catenin, as well as COX 2 and prostaglandin E2 (PGE2). Thymoquinone also showed efficacy and promoted autophagy and apoptotic in irinotecanresistant LoVo colon cancer cells (CPOI11 R) by inducing mitochondrial outer membrane permeabil ity, activating JNK and p38, and inhibiting the NFκβ, ERK1/2, and PI3K pathways (Chen et al. 2017). In creased phosphorylation of mitogenactivated protein ki nase (MAP) p38 (Woo et al. 2013), inhibition of MEK ERK1/2 signaling (ElBaba et al. 2014), and blocking of the PI3KAkt signaling pathway are also other mech anisms of thymoquinone as anticancer (Dirican et al. 2015). On the other hand, thymoquinone also has in teractions with other metabolites also contained in black cumin, namely thymol, nigellimine, and pcymene. This proves that plantbased treatment with a multicomponent approach allows synergistic interactions between active compounds that support the resulting pharmacological ef fects (Syahrir et al. 2016).
Some of the advantages of the in silico method are re ducing the number of compounds or molecules and the in creased speed of research of most compound and protein interactions through database searches and reducing the use of experimental animals in predictive studies. This in silico prediction can also save time on developing ligand based drugs and can be a starting point before conducting in vitro and in vivo tests, as well as proving the oppor tunity for black cumin to become a candidate for OMAI. However, the weakness of this method also requires con sideration, including a large amount of data analyzed. It requires a high level of accuracy and has the risk of bias depending on the database used to predict. Some of the ac tive metabolites in black cumin that are not in the database can also increase the chance of error in predictions and re duce the accuracy of research results.

Conclusions
Black cumin (Nigella sativa L.) has a pharmacology net work that describes the correlation between the metabo lites contained there, targets or surface receptor and in tracellular proteins, and colorectal cancer signaling path ways. This indicates that black cumin (Nigella sativa L.) has the potential as OMAI and can be a reference in the de velopment of cancer treatment, especially colorectal can cer.