Mapping Pharmacological Network of Multi-Targeting Litchi Ingredients in Cancer Therapeutics

Considerable pharmacological studies have demonstrated that the extracts and ingredients from different parts (seeds, peels, pulps, and flowers) of Litchi exhibited anticancer effects by affecting the proliferation, apoptosis, autophagy, metastasis, chemotherapy and radiotherapy sensitivity, stemness, metabolism, angiogenesis, and immunity via multiple targeting. However, there is no systematical analysis on the interaction network of “multiple ingredients-multiple targets-multiple pathways” anticancer effects of Litchi. In this study, we summarized the confirmed anticancer ingredients and molecular targets of Litchi based on published articles and applied network pharmacology approach to explore the complex mechanisms underlying these effects from a perspective of system biology. The top ingredients, top targets, and top pathways of each anticancer function were identified using network pharmacology approach. Further intersecting analyses showed that Epigallocatechin gallate (EGCG), Gallic acid, Kaempferol, Luteolin, and Betulinic acid were the top ingredients which might be the key ingredients exerting anticancer function of Litchi, while BAX, BCL2, CASP3, and AKT1 were the top targets which might be the main targets underling the anticancer mechanisms of these top ingredients. These results provided references for further understanding and exploration of Litchi as therapeutics in cancer as well as the application of “Component Formula” based on Litchi’s effective ingredients.


INTRODUCTION
Cancer is one of the most serious public health problems globally. In 2018, approximately 18.1 million new cancer cases and 9.6 million cancer-related deaths occurred in the world (Bray et al., 2018). There is an urgent need for a more effective therapy. Traditional Chinese medicine (TCM) has been used for thousands of years in Asia for its good efficacy and compliance, and this also made it an important supplemental medicine in cancer treatment (Xiang et al., 2019). Comparing with the current "one drug, one target" mode, TCM has the feature of "multiple active ingredients, multiple targets" (Li and Zhang, 2013). Given that cancer is a complex disease which alters a range of cellular and molecular processes, TCM may hold the advantage of targeting multiple cancer-related molecules simultaneously with potential synergistic effects. However, as a result of the feature of "multiple ingredients, multiple targets", herbs can potentially interact with prescription medications like when cancer patients use plant-based regimens with chemotherapy (Yeung et al., 2018;Parvez and Rishi, 2019;Pezzani et al., 2019). Therefore, the potential risk of using TCM as complementary medicine should be considered for maximum safety and efficacy.
Litchi chinensis Sonn (Litchi), a member of Litchi, Sapindaceae family, is a subtropical evergreen plant which has been widely cultivated as an economic cultivar for its delicious taste and rich nutrition fruitage in China, Philippines, Indonesia, and Vietnam (Mitra, 2002;Menzel et al., 2005). In China, Litchi seeds were used as an analgesic agent for the alleviation of neuralgia, orchitis, testicular swelling, hernia, gastralgia, lumbago, abdominal pain, etc. (Lan and Lan, 2011). The decoctions of Chinese herbal formula containing Litchi seeds were used as indigenous remedies for urologic neoplasms including prostate cancer, bladder cancer, and renal carcinoma (Shi, 2004;Wang, 2011c). Moreover, a considerable amount of studies have shown that in addition to Litchi seeds, the extracts and ingredients from other parts (peels, pulps, and flowers) of Litchi can exert multiple pharmacological actions which have the anti-inflammatory (Das et al., 2016), anti-oxidative , anti-bacterial , anti-viral (Gangehei et al., 2010;Xu et al., 2010a), anti-liver injury, and immuneenhancing effects (Noh et al., 2011;Huang et al., 2014a;Yamanishi et al., 2014;Huang et al., 2014b;Huang et al., 2016a;Su et al., 2016;Xiao et al., 2017;Queiroz et al., 2018). Furthermore, there was accumulating evidence indicating that the extracts and compounds from Litchi exhibit anticancer effects by targeting multiple proteins and signal pathways involved in cancer cell proliferation, metastasis, angiogenesis, apoptosis, autophagy, etc. However, current studies are limited to the traditional research method of identifying "single-drug, single-target, and single-pathway", which failed to reflect the "multiple ingredients-multiple targets-multiple pathways" anticancer effects of Litchi. In order to elucidate its multiple modes of action, network pharmacology and bioinformatics were employed in this study as a powerful approach (Zhang et al., 2019a) to systematically analyze the complicated interactions between Litchi ingredients and confirmed targets based on published research results. This study has provided a solid base for the further exploration of its anticancer effects.

METHODS
We collected the anticancer ingredients and targets of Litchi based on original published articles. In order to systematically analyze the complex relationships between these anticancer ingredients and their targets, an interaction network was constructed by network pharmacology approach. All networks maps were visualized and analyzed by Cytoscape 3.2.1 (http:// www.cytoscape.org/). As shown in the ingredient-target network (Figures 1A,2A,3A,4A,and 5), the oval nodes represent ingredients, the rectangle nodes represent targets and each edge linking an ingredient to a target indicates a regulatortarget relationship. In Figures 1A-4A, the targets distributing in the inner orange circle (rectangle) can be modulated by multiple ingredients rather than a single ingredient. The "degree" is an important parameter for the network pharmacology approach, which represents the number of related nodes to a particular node in the network. The greater the degree of a node, the more biologically important it is. Therefore, the top ingredients and targets were screened out by the Network Analyzer in Cytoscape based on the major parameter of "degree". To further explore the core biological processes of the top targets involved, we performed KEGG pathway enrichment analysis (http://www.kegg.jp/) and screened out the top signal pathways based on the P-value. The relationships among top targets, corresponding ingredients and signal pathways were analyzed by combining Cytoscape 3.2.1 with KEGG pathway enrichment analysis. In order to test the reliability of the top ingredient-target interactions and explore the accurate binding modes, we performed molecular docking analysis by using surflex module of Sybyl X2.0. A total score greater than 6 represents good protein-ligand binding. The crystal structures of proteins (targets) were extracted from Protein Data Bank (https://www.rcsb.org/).
anticancer ingredients with their corresponding effects, molecular targets, and cancer types were listed in Table 2. We then discovered that a single component could have a range of targets and different components had overlapping molecular targets, hence they formed a complicated regulatory network. In order to unravel this intricate web of interactions, we applied network pharmacology method to analyze the anticancer effects of Litchi from a perspective of system biology.

Inhibition of Cancer Cell Proliferation
Sustained proliferation is a hallmark of cancer cells, and the restoration of dysregulated signaling pathways has always been a target for cancer treatment. The extracts from Litchi peels, pulps, seeds, leaves have been shown to inhibit the proliferation of a variety of cancer cells (Huang et al., 2015a;Gong et al., 2018;Zhao et al., 2019a). The 13 anti-proliferative compounds identified from Litchi and 100 regulated targets were summarized in Table S1. The detailed analysis of the top active ingredients, corresponding targets, and signal pathways affected was shown in Figure 1.
In total, this ingredient-target network ( Figure 1A) was consisted of 113 nodes (Table S1) and the mean degree of all nodes in the network was 3.080. Overall, 3 out of the 13 anticancer compounds ( Figure 1A) had high degree distributions (kaempferol: degree=39, Epigallocatechin gallate (EGCG): degree=36, gallic acid: degree=22) and all of them modulated more than 20 targets, which marked their pharmacological importance. Notably, those targets have more than one regulator (Table S2). Apart from 1 target that was regulated by 10 ingredients, 4 targets were regulated by over 5 ingredients and 28 targets were regulated by 2-4 ingredients ( Figure 1B). Further, the 4 top targets (MAPK1, CDKN1A, MAPK14, AKT1) were screened out from Figures 1A, B, whose degree values were more than two folds of the median degree of all nodes in the network. This suggested that multiple ingredients could potentially exert synergistic anti-proliferation effects. In particular, the interactions among the above 4 top targets and Litchi ingredients (Table S3) were analyzed in Figure 1C. With the results shown in Figure 1C, we could conclude that there were 11 out of 13 ingredients that could regulate the top targets with anti-proliferative effects. It was also confirmed that the top 4 targets played an important role in the anti-proliferative process. Particularly, kaempferol, EGCG, and gallic acid could regulate all the top targets, and this conclusion was similar to that in Figure  1A where 3 ingredients mentioned above had outstanding pharmacological significance. To further clarify the anticancer mechanism of Litchi ingredients, the pathway enrichment analysis based on above 4 top targets was performed. There were 63 signaling pathways involved in the anti-proliferation effects of Litchi ingredients ( Figure 1C and Table S3), and FoxO, VEGF, Prolactin, ErbB, HIF-1, Toll-like receptor, TNF, Rap1, MAPK, and PI3K-Akt signaling pathways were the top 10 pathways according to their P values ( Figure 1D). All of the 4 top targets were elements of FoxO signaling pathway and 3 out of the top 4 targets were elements of other 9 top pathways. It indicated that these top 10 pathways might be the major signaling pathways that are responsible for the antiproliferation effects of Litchi.

Induction of Cancer Cell Apoptosis and Autophagy
Apart from uncontrollable proliferation, resistance to cell death is another strategy employed by cancer cells to fuel its growth.  Cancer cells have evolved a series of strategies to inhibit cell death while Litchi ingredients have been reported to have proapoptosis and pro-autophagy effects (Hsu et al., 2012a;Emanuele et al., 2018). Hence, we summarized data from literature and constructed the network ( Figure 2A) based on 18 ingredients from Litchi and 138 targets ( Table S4) which related to cell apoptosis and autophagy. The network was consisted of 156 nodes and 283 edges altogether, representing the extensive interactions among 18 ingredients and 138 targets ( Table S4).
Not surprisingly, we found that the mean degree of node was 3.679 based on the topological analysis, suggesting that it was common for ingredients to have multiple targets. By referring to the mean degree, we identified 6 top ingredients with a median degree ≥20, namely luteolin, EGCG, kaempferol, gallic acid, betulinic acid, and chlorogenic acid, with the top 2 having over 40 targets. Hence, we concluded that those top 6 ingredients were likely to be crucial components in promoting apoptosis and autophagy. Further, in order to clearly elucidate if these targets were regulated by multiple ingredients, another analysis was performed in Figure 2B, which showed that there were 3 targets regulated by over 10 ingredients, 9 targets were regulated by 5-10 ingredients and 33 targets were regulated by more than 2 ingredients ( Figure 2B and Table S2). From Figures 2A, B, we next screened out the top 6 targets (BAX, BCL2, CASP3, CASP9, TP53, AKT1) based on their degrees in the ingredienttarget network. As shown in Figure 2C and Table S5, all of the top 6 targets could be regulated by luteolin and EGCG, and this implied that they had multiple anticancer activities. In addition, all the 18 ingredients involving in apoptosis and autophagy interacted with the top targets, which consolidated the importance of these top targets. KEGG enrichment analysis based on these 6 top targets showed that 39 signaling pathways were involved in the effects of inducing cancer cell apoptosis and autophagy ( Figure 2C and Table S5), while p53, Neurotrophin, Sphingolipid, PI3K-Akt, Thyroid hormone, MAPK, VEGF, HIF-1, TNF signaling pathway and Adrenergic signaling in cardiomyocytes were the top 10 pathways ( Figure 2D). Four out of these top 6 targets were elements of p53, Neurotrophin, Sphingolipid, and PI3K-Akt signaling pathways, which indicates that these four signaling pathways might be the major pathways responsible for anticancer effect by inducing apoptosis and autophagy.

Inhibiting Metastasis
Metastasis is another target in cancer therapeutic development due to its lethality . Litchi seed extracts could attenuate migration and invasion capabilities of PC3 and DU145 cells (Guo et al., 2017). Nine anti-metastasis ingredients of Litchi and 99 corresponding targets were listed in Table S6, the interaction network of which was shown in Figure 3A. We found that the mean degree of nodes in the network was 3.296. Then we screened out 4 top ingredients, namely EGCG, gallic acid, luteolin, and PA, with a median ≥20 degrees, which acted on 41, 29, 22, and 21 targets respectively. Therefore these 4 top ingredients identified were likely to be crucial bioactive components to inhibit metastasis. In addition, among the 99 targets, the network showed that MMP2 had the largest number of ingredient-target interactions (degree value of 8), followed by MMP9 (degree value of 7), making them likely to perform antimetastasis functions. The remaining targets with lower degree and less than two folds of the mean degree of all nodes were also included. Then, the targets regulated by multiple ingredients were analyzed with a similar approach for more information. As shown in Figure 3B and Table S2, MMP2 and MMP9 were regulated by 8 and 7 ingredients respectively, followed by another 6 targets regulated by up to 5 ingredients and 26 targets regulated by 2 to 4 ingredients. The "ingredients-top targets-pathways" network ( Figure 3C and Table S7) was constructed for the purpose of confirming the significance of top 2 targets, and this network indicated that as much as 8 ingredients exerted the anti-metastasis function through modulating MMP2 and MMP9. However, the signaling pathways enriched by KEGG based on 2 top targets merely included bladder cancer, estrogen signaling pathway, leukocyte transendothelial migration, proteoglycans in cancer and pathways in cancer. Both the top 2 targets were elements of these 5 pathways (Figures 3C, D and Table S6), which indicated these 5 pathways might be the key anti-metastasis mechanism of Litchi.

Sensitizing Chemotherapy and Radiotherapy
Chemotherapy and radiotherapy are two of the most common cancer treatments. Despite their clinical efficacy in clearing cancer cells, therapeutic resistance often inevitably occurs. Another reported effect of Litchi was that it sensitized chemotherapy and radiotherapy. Here we identified 12 compounds from Litchi and 106 corresponding molecular targets responsible for this function (Table S8), with the detailed interactions of the top ingredients, targets and signal pathways shown in the Figure 4. From Figure 4A, we screened out 5 top ingredients with a median degree ≥20, including luteolin, EGCG, kaempferol, gallic acid, and betulinic acid, which linked to as much as 35, 34, 25, 22, and 21 targets respectively. Not surprisingly, the mode of "multi-ingredients, multi-targets" was confirmed again by identifying CASP3, BAX, and BCL2 as the top targets, which had the degree values of 9, 8, 6 respectively, which were more than two folds of the median degree of all nodes in the network. In addition, there were another 32 targets regulated by more than 2 ingredients ( Figure 4B and Table S2), which implied that Litchi ingredients could overcome chemo-and radio-resistance through a "multi-compounds, multi-targets" mode with potential synergistic effects. The "ingredients-top targetspathways" network (Table S9) confirmed the importance of CASP3, BAX, and BCL2 further. In Figure 4C, 10 out of 12 ingredients that were involved in sensitizing chemotherapy and radiotherapy exerted anticancer activity through regulating the 3 top targets. Moreover, KEGG enrichment analysis of top 3   targets showed that 15 signaling pathways were involved in the chemotherapy and radiotherapy sensitization ( Figure 4C and Table S9). All of the top 3 targets were elements of Amyotrophic lateral sclerosis (ALS), Colorectal cancer, Apoptosis, Hepatitis B, Tuberculosis and pathways in cancer, and 2 out of the top 3 targets were elements of p53 signaling pathway, Toxoplasmosis, Sphingolipid, and Neurotrophin signaling pathway, which indicates that the 10 pathways mentioned above might be responsible for the anticancer effect of Litchi on chemotherapy and radiotherapy sensitization ( Figure 4D).

Other Anticancer Effects
Apart from the four effects exerted by Litchi ingredients for the major anticancer functions as listed above, several other targets were also found to be involved in the suppression of cancer stemness, metabolism, and angiogenesis, while also in the enhancement of immunity as listed in Table S10. However, the experiments validations on the anticancer effect of Litchi ingredients from these four aspects were very limited. Therefore, we only constructed a simple ingredient-target network map ( Figure 5). The results showed that these mechanisms involved a total of 10 active ingredients, among which 5 belonged to the top ingredients from the previous screening including betulinic acid, EGCG, luteolin, gallic acid, and kaempferol, which further illustrated their importance. At the same time, we suggest that the remaining 5 ingredients (chlorogenic acid, (-)-Epicatechin-3gallate (ECG), naringenin, cyanidin-3-glucoside, lupeol) and their detailed mechanisms need to be further explored.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/ Supplementary Material.

AUTHOR CONTRIBUTIONS
HG, SC, and ZS designed this work. SC, YH, and YC drafted the manuscript. HG, YH, and DZ performed the network pharmacology analysis. QL made the figures. All authors read and approved the final version.