Colorectal cancer (CRC) is the third most commonly diagnosed cancer worldwide and the second leading cause of cancer-related deaths. While chemotherapy and targeted therapies have improved outcomes, significant challenges remain, including drug resistance and severe side effects that limit treatment effectiveness.
Natural products from traditional Chinese medicine (TCM) offer a promising alternative because they often act on multiple molecular targets simultaneously, have favorable safety profiles compared to synthetic drugs, and have been used medicinally for centuries. Among these, Salvia chinensis Benth, known in Chinese as Shijianchuan (SJC), has been recognized for its anticancer properties in various cancers including breast, liver, and pancreatic cancer.
SJC is a medicinal plant rich in flavonoids, phenolic acids, and alkaloids, compound classes known for their anti-inflammatory and anticancer effects. However, despite promising results in other cancer types, a systematic investigation of how SJC works against colorectal cancer specifically had not been conducted.
This study aimed to close that gap by using a modern integrated approach: first mapping exactly which compounds are present in SJC, then testing those compounds against colorectal cancer cells, and finally using advanced computational and genetic analysis to identify the specific molecular targets responsible for the anticancer effects.
The first step was to characterize the complete chemical profile of the SJC aqueous extract using UPLC-MS/MS (ultra-performance liquid chromatography combined with tandem mass spectrometry). This powerful analytical technique can identify compounds based on their molecular weight and fragmentation patterns, even in complex herbal mixtures.
SJC was prepared as a decoction to mimic traditional use: 20 grams of dried whole plant were boiled in water for one hour, then concentrated using a rotary evaporator and freeze-dried to yield a standardized powder. This approach ensures the analysis reflects the biologically active form of the herb as it would be consumed medicinally.
The analysis identified 60 natural compounds in the SJC extract, spanning multiple chemical classes including flavonoids, phenolic acids, and organic acids. Nine of these were quantified precisely using validated targeted methods with certified reference standards. The most abundant compound identified was citric acid at 15.7 micrograms per gram, followed by caffeic acid (6.3 ug/g), rutin (5.9 ug/g), and chlorogenic acid (1.1 ug/g). Naringenin, which proved important in subsequent experiments, was present at 0.11 ug/g.
This comprehensive chemical profiling is a critical first step in natural products research because it provides a detailed inventory of what is actually present in the extract - a prerequisite for understanding which specific compounds are responsible for observed biological effects and for ensuring reproducibility across studies.
Testing SJC extract against HCT-116 colorectal cancer cells in laboratory culture demonstrated clear anticancer activity. The extract inhibited cell proliferation with an IC50 (the concentration needed to kill half the cells) of approximately 400 micrograms per milliliter, indicating moderate but significant growth-suppressing effects.
Cell migration was assessed using two complementary methods. In the wound-healing assay, a scratch was made in a layer of cells and the rate of gap closure was measured. SJC treatment at 300 and 500 ug/mL reduced migration rates from 56% (untreated) to 42% and 24% respectively, showing a dose-dependent reduction in the ability of cancer cells to move and spread.
The Transwell migration assay measured how many cells could migrate through a porous membrane toward nutrients, simulating how cancer cells invade surrounding tissue. SJC treatment significantly reduced the number of migrating cells compared to untreated controls, further confirming anti-invasive activity.
Apoptosis (programmed cell death) was measured by flow cytometry. SJC at concentrations of 300, 500, and 800 ug/mL induced apoptosis in 26%, 33%, and 39% of cells respectively, compared to only 9% in untreated cells. These results confirm that SJC acts through multiple mechanisms: slowing cancer cell growth, blocking migration, and triggering cell death.
To understand how SJC works at the molecular level, the researchers performed RNA sequencing (transcriptomics) on HCT-116 cells treated with SJC. This technology measures the activity of every gene in the cell simultaneously. Treatment with SJC caused 1,787 genes to change their expression levels, with 1,183 genes being turned up and 604 being turned down.
These drug-responsive genes were then cross-referenced with a large public database of colorectal cancer patient samples using Weighted Gene Co-expression Network Analysis (WGCNA). This statistical method identifies groups of genes that tend to be active together and are clinically associated with cancer. The intersection of SJC-responsive genes and clinically relevant cancer genes produced 162 high-confidence candidates.
Three independent machine learning algorithms - Random Forest, LASSO regression, and XGBoost - were then applied to these 162 candidates to rank them by importance. Random Forest uses an ensemble of decision trees to rank genes by how well they distinguish cancer from normal tissue (achieving 98.1% accuracy). LASSO regression identifies the smallest set of genes that explain the most variation. XGBoost uses gradient boosting to rank features by their contribution to predictive power.
The intersection of the top 20 genes from all three machine learning models identified four candidate target genes that were consistently ranked as important: ENC1, KLF4, CXCL8, and KCTD9. This multi-algorithm consensus approach reduces the risk of false positives that would arise from using any single method.
Among the four candidate targets, further analysis using a protein-protein interaction (PPI) network prioritized CXCL8 as the single most important target. The PPI network maps how proteins interact with each other in cells, and proteins at the center of these networks tend to have the broadest biological impact. CXCL8 emerged as the only gene that appeared both among the top 30 most highly connected hub genes and among the four machine learning-identified candidates.
CXCL8, also known as interleukin-8 (IL-8), is a signaling protein called a chemokine. In healthy tissue, it helps recruit immune cells to fight infection. In cancer, however, CXCL8 is frequently overproduced by tumor cells and promotes cancer progression by stimulating cell proliferation, triggering new blood vessel formation to feed the tumor, and suppressing immune attack through the PI3K/Akt/NF-kappaB signaling pathway.
Analysis of cancer patient data from The Cancer Genome Atlas (TCGA) confirmed that CXCL8 is significantly overexpressed in colorectal tumor tissue compared to normal colon tissue. Gene Set Enrichment Analysis (GSEA) showed that high CXCL8 expression in CRC patients correlates with activation of multiple cancer-related pathways including cell cycle progression, p53 signaling, extracellular matrix remodeling, and metabolic reprogramming.
The identification of CXCL8 as a key target is biologically significant because it connects SJC activity to a well-characterized oncogenic pathway. CXCL8 overexpression has been linked to poor prognosis and chemotherapy resistance in multiple cancer types, making it an attractive therapeutic target.
With CXCL8 identified as the key target, the next question was which specific compound in SJC is responsible for inhibiting it. Molecular docking simulations were used to computationally test how strongly each of the 60 SJC compounds would bind to the CXCL8 protein. The compound with the strongest predicted binding affinity was naringenin, with a binding energy of -6.7 kcal/mol.
Naringenin is a flavanone, a subtype of flavonoid, found naturally in citrus fruits and various plants. It is a well-studied bioactive compound known for anti-inflammatory, antioxidant, and anticancer properties in multiple cancer types. The molecular docking results suggest naringenin physically fits into the binding site of CXCL8 in a way that would disrupt its normal signaling activity.
Laboratory validation confirmed the computational prediction: RT-qPCR analysis showed that treating HCT-116 cells with naringenin significantly reduced CXCL8 gene expression. Cell viability assays further demonstrated that naringenin inhibited cancer cell growth in a dose-dependent manner, mirroring the effects of the whole SJC extract.
Cell cycle analysis showed that naringenin at 300 micromolar concentration increased the proportion of cells arrested in S and G2/M phases from 37% to 49%, indicating it blocks cells from completing division. Apoptosis rates increased progressively from 8.6% in untreated cells to 57.8% at 800 micromolar, confirming naringenin as a potent inducer of programmed cancer cell death.
The study revealed that SJC likely fights colorectal cancer through a multi-target, multi-compound strategy - consistent with the general principle of traditional Chinese medicine. While naringenin and CXCL8 emerged as the primary compound-target pair, the other three identified candidate genes (ENC1, KLF4, KCTD9) also play established roles in CRC, suggesting the full extract may act through additional mechanisms simultaneously.
ENC1 is overexpressed in CRC and promotes cancer progression through the beta-catenin and JAK2-STAT5-AKT signaling pathways. KLF4 functions as a tumor suppressor by reducing beta-catenin levels and is typically downregulated in colorectal cancers. KCTD9 is also downregulated in CRC and negatively regulates Wnt/beta-catenin signaling. The fact that SJC treatment altered the expression of all these genes suggests it disrupts multiple cancer-promoting networks.
The pathway enrichment analysis showing that high CXCL8 expression activates drug metabolism pathways has a particularly interesting implication: CXCL8-high tumors may be more resistant to standard chemotherapy. If so, targeting CXCL8 with naringenin could potentially restore chemotherapy sensitivity, suggesting a synergistic application of SJC alongside conventional treatment.
The researchers acknowledge important limitations: the study used only one CRC cell line (HCT-116), the compound identifications are tentative without nuclear magnetic resonance confirmation, and molecular docking results are predictive rather than proven. Future studies in additional cell lines and animal models are needed to confirm these findings before any clinical applications can be considered.
This study established the first comprehensive chemical profile of Salvia chinensis Benth using UPLC-MS/MS, identifying 60 compounds with nine precisely quantified. The aqueous extract demonstrated clear anticancer activity against HCT-116 colorectal cancer cells, inhibiting proliferation, migration, invasion, and inducing apoptosis.
Through an integrated approach combining transcriptomics, WGCNA, multiple machine learning algorithms, PPI network analysis, and molecular docking, the study identified CXCL8 as the priority molecular target and naringenin as the primary bioactive compound mediating SJC anti-CRC activity. This represents a significant advance in understanding the molecular pharmacology of a traditional Chinese medicine.
The integrated strategy used in this study, combining chemical profiling with genomic analysis and computational modeling, demonstrates a powerful template for systematically investigating natural products. This approach can be applied to other traditional medicines to uncover their active components and molecular mechanisms without requiring the isolation of each compound individually.
These preliminary findings establish a solid scientific foundation for future work including validation in additional cell lines, testing in animal models of colorectal cancer, investigation of synergistic interactions among SJC compounds, and potentially the development of naringenin-based therapeutics for colorectal cancer treatment.