Collection and processing of public transcriptomic data

The Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov/) database was explored to obtain transcriptomic data for 6 types of digestive system cancers. A total of 316 samples were obtained, including 8 pairs of cholangiocarcinoma (CHOL), 50 pairs of colorectal cancer (CRC, including both coloo and rectal adenocarcinoma (COAD/READ)), 13 pairs of esophageal carcinoma (ESCA), 50 pairs of liver hepatocellular carcinoma (LIHC), 4 pairs of pancreatic adenocarcinoma (PAAD), and 33 pairs of stomach adenocarcinoma (STAD) as well as their adjacent normal controls (S1 Table). We calculated gene expression levels using the transcripts per kilobase million (TPM) matrix, which was subsequently used for further data processing [10]. A filtering criterion was applied to identify genes with an expression level of 20 or above in at least 180 samples, yielding a dataset of 17,673 genes. Differential gene expression was assessed with DESeq2 using the raw count tables generated by TCGA (The Cancer Genome Atlas homepage.http://cancergenome.nih.gov/abouttcga.). Gene expression fold changes were computed, followed by dimensionality reduction via Uniform Manifold Approximation and Projection (UMAP) and hierarchical clustering.

Rank-rank hypergeometric overlap analysis

To compare the global transcriptomic signatures of each COADREAD group and LIHC group, we applied the RRHO algorithm that is an unbiased and threshold-free method [11]. For each group, genes list were ranked according to the log 2-transformed fold-change generated by DESeq2, from the most up- to the most down-regulated ones. A hypergeometric test was performed to assess the significance of the similarity of gene profile, using a sliding window with step size (i.e., default) for each pair of impacts by using the RRHO2 (v1.0) package. A False Discovery Rate (FDR) correction was applied to adjust for the multiple hypothesis testing. The visualization of the output of this analysis was the RRHO level map, in which the most significant hypergeometric p-value (log 10 transformed and direction-signed) was labeled after computing all possible rank combinations, generating an index of the matrix for the most significant rank combination in each pair of impacts. Heatmaps generated using RRHO2 have top-right (both decreasing) and bottom-left (both increasing) quadrants, representing the concordant mRNA changes, while the top-left and bottom-right quadrants represent discordant overlaps (opposite directional overlap between datasets). For each comparison, the stratified RRHO method was used on log 10 (P-values) with the default step size for each quadrant, and adjusted Benjamini-Yekuteli p-values were calculated to determine significant concordant and discordant overlaps.

Weighted gene co-expression network analysis

The R package “WGCNA” was utilized to construct gene co-expression networks for CRC and LIHC samples [12]. We first created an adjacency matrix of 9978 gene expressions using a soft threshold of 13 based on gene-gene correlation, which was then transformed into a topological overlap matrix (TOM). Hierarchical clustering of genes identified 18 co-expression modules. We calculated module eigengenes (ME) for each co-expression module and correlated these with a binary variable representing the two cancer types, where CRC was assigned a value of 1 and LIHC a value of 0. The Pearson correlation coefficients indicated the degree of association between module expression patterns and the cancer type, with positive and negative values indicating a stronger association with CRC and LIHC, respectively.

Functional enrichment analysis and PPI network analysis

Functional enrichment analysis was conducted using the clusterProfiler package (v4.0.5). Significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were identified using the enrichKEGG function. Gene set enrichment analysis (GSEA) was performed on a gene list sorted by log2 fold change (FC), utilizing the gseKEGG function, with a significance threshold set at P < 0.05 [13].

For protein-protein interaction (PPI) network construction, we used the STRING database (https://string-db.org/). Clustering analysis was performed using MCL clustering with default parameters, and TSV files for clusters 1, 2, and 3 were downloaded. Finally, Cytoscape (version 3.7.1) was employed to visualize the PPI networks [14]. In the visualization of PPI network, we defined connections between proteins at a correlation threshold of 0.7, indicating significant interactions. Node connectivity, based on the number of edges, was visually represented by color-coding: highly connected nodes were marked in red, while those with fewer connections were green.

C-Map analysis

The CMap database is the largest perturbation-driven gene expression dataset, serving as a resource to explore relationships between diseases, genes, and therapeutics (https://clue.io/) [4]. In this study, potential drug candidates were screened through CMap analysis using 300 candidate central genes (150 up-regulated and 150 down-regulated) as input. The selected query parameters were “Gene expression (L1000)” and “Latest.”

Matching results were filtered to include “Compound- trt_cp”, “Knock Down- trt_kd” and “Over Expression- trt_sh” with a screening criterion of a score ≤ −1.2 in CRC cell lines (HT29, HCT116, SW480, and NCIH508) and LIHC cell lines (HEPG2 and HUH7). Resuts of “Over Expression- trt_sh” were neglected later due to limited output. Positive scores indicated synergistic effects, while negative scores indicated antagonistic effects in cancer contexts.

Statistical analysis

The P-value was calculated using the unpaired two-tailed Student’s t test, and P < 0.05 was considered statistically significant. Hierarchical clustering analysis was performed using pheatmap (v1.0.12) package with Pearson correlation. For RNA-seq analysis, significant differentially expressed genes (DEGs) with fold change > 2 and Benjamini-Hochberg (BH)–adjusted P-value < 0.01 were identified by DESeq2 and adjusted for litter. The clusterProfiler package was used to determine potential KEGG. Terms with Benjamini-Hochberg corrected P-value < 0.05 were defined as significantly enriched terms and pathways. All the statistical analyses were done using available packages in R (v4.1.3).

Cell culture

The human colon cancer cell line HT29 and the human liver cancer cell line HepG2 (both purchased from ATCC in the United States) were cultured in DMEM medium (C11995500BT, Gibco, USA) supplemented with 10% fetal bovine serum (FSP500, ExCell Bio, China) and 1% penicillin-streptomycin (C3420-0100, Vivacell, China). The cells were incubated at 37°C in 5% carbon dioxide.

Cell viability assay

The effects of Amonafide and BX795 on the viability of HT29 and HepG2 cells were evaluated using the CCK8 kit (MI00615A, MISHU, China). Amonafide (HY-10982) and BX795 (HY-10514) were purchased from Shanghai MedChemExpress Bio-Technology Co., Ltd. The drugs were diluted with DMEM medium to achieve the desired concentrations. HT29 and HepG2 cells were seeded in 96-well plates at a density of 1 × 10^4 cells per well and cultured in a 5% CO2 incubator at 37°C for 24 hours to allow cell adhesion. The culture medium was then replaced with DMEM medium containing various concentrations of Amonafide and BX795 (0, 5, 10, 20 μM). After 12 hours, each well was treated with 10 μL of CCK8 reagent and incubated for 1 hour. Absorbance at 450 nm was measured using an enzyme-linked immunosorbent assay (ELISA) reader (SpectraMax iD5, Shanghai, USA). The same procedure was repeated for 24, 36, and 48 hours.

Colony formation

HepG2 cells were seeded in 6-well plates at a density of 3 × 10^3 cells per well, and HT29 cells were seeded at a density of 8 × 10^3 cells per well, with three replicates for each group. The cells were incubated at 37°C and 5% CO2 for 24 hours to allow cell adhesion, and the culture medium was discarded. Amonafide and BX795 were prepared in DMEM containing 10% serum at various concentrations (0, 0.1, 1, 5 μM), and the cells were cultured for 48 hours, with the medium being changed once. After visible colonies formed, the cells were washed with PBS and fixed at room temperature for 20 minutes in 4% tissue cell fixative solution. After PBS washing, they were stained with 0.5% crystal violet solution for 15 minutes, washed again with PBS, dried, and observed under a microscope. Colonies were counted using ImageJ software.

Cell scratch assay

HT29 and HepG2 cells were seeded in six-well plates with three replicates per group. Once the cells reached 80–90% confluence, the culture medium was removed. Amonafide and BX795 were prepared in DMEM containing 1% fetal bovine serum, and the cells were cultured at different concentrations (0, 1, 5, 10 μM). A scratch was made using a sterile 200 μL pipette tip, and PBS was used to remove cell debris. Images were taken with an inverted microscope to record the scratch location. The wounds were then incubated at 37°C in 5% CO2 for 48 hours, and photographs were taken with an optical inverted microscope connected to a real-time imaging system. Images from three separate experiments were recorded for each group, and cell migration distances were analyzed using ImageJ.

Flow cytometry

A cell cycle detection kit (C6031S, UElandy, China) was used to assess the effects of Amonafide and BX795 on the cell cycle of HT29 and HepG2 cells. Cells were cultured in 60 mm dishes and treated with Amonafide and BX795 at different concentrations (0, 10 μM) for 24 hours. The cells were collected, centrifuged, and washed with 1 mL of pre-cooled PBS. Subsequently, 1 mL of 75% ethanol was added, and the cells were fixed overnight at −20ºC. After centrifugation to remove the supernatant, the cells were washed with pre-cooled PBS. Appropriate propidium iodide staining solution was added, and cells were incubated at room temperature in the dark for 15–30 minutes. Cell cycle analysis was conducted using a flow cytometer (Beckman, USA).

Apoptosis assessment

Apoptosis was assessed using the Annexin V-FITC/PI staining kit (FA101, TransGen Biotech, China). HT29 and HepG2 cells were cultured in 60 mm dishes and treated with Amonafide and BX795 at different concentrations (0, 10 μM) for 24 hours. Cells were collected, centrifuged, washed with pre-cooled PBS, and resuspended in 100 μL of pre-cooled 1 × Annexin V Binding Buffer. Cells were then incubated with 5 μL of Annexin V-FITC and 5UL-PI in the dark for 15 minutes at room temperature. After adding 400 μL of pre-cooled 1 × Annexin V Binding Buffer, the samples were placed on ice and analyzed by flow cytometry.

Assessment of gene expression alterations across six digestive system cancers

Gene expression analysis reveals the molecular signatures of cancers, potentially highlighting shared therapeutic targets and biomarkers. In order to delineate common and distinct molecular features, we collected paired transcriptome data (cancerous and adjacent non-cancerous tissues) for six digestive system cancer types, amounting to 316 samples (Methods, S1 Table). After dimensionality reduction with Uniform Manifold Approximation and Projection (UMAP), the samples from six cancer types of the digestive system did not exhibit significant differences and were uniformly distributed, indicating that they share similar features in their gene expression profiles (Fig 1A). In line with the UMAP results, hierarchical clustering analysis also showed no distinct clusters among samples from various cancer types, presenting a homogeneous distribution (Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Gene Expression Correlation and Dimensionality Reduction Analysis in Digestive System Cancers.

(a) Uniform Manifold Approximation and Projection (UMAP) analysis visualizes the distribution of fold change (FC) in pairwise samples. (b) Heatmap depicts the Pearson correlation coefficients between gene expression profiles of different cancer samples. 8 cholangiocarcinoma (CHOL), 50 colorectal cancer (CRC), 13 esophageal carcinoma (ESCA), 50 liver hepatocellular carcinoma (LIHC), 4 pancreatic adenocarcinoma (PAAD) and 33 stomach adenocarcinoma (STAD) samples were used.

https://doi.org/10.1371/journal.pone.0325700.g001

Further analysis focused on identifying differentially expressed genes (DEGs) involved in distinguishing between six types of cancers, and included the calculation of intersections among different combinations of DEGs (S1A Fig, S2 Table). The combination of cholangiocarcinoma (CHOL) and liver hepatocellular carcinoma (LIHC) exhibited the highest number of shared DEGs, with 738 common genes, while colorectal cancer (CRC) and LIHC shared 358 DEGs, ranking sixth. The combination of all six cancer types showed the least number of shared DEGs, totaling only 4 genes. As the number of cancer combinations increased, the number of shared DEGs decreased. Thus, while there is overall similarity in gene expression profiles among various cancer types, specific combinations of cancers exhibit higher similarity in characteristic DEGs.

To assess the functional pathway similarities of DEGs across different cancer types, we also evaluated the intersections of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways among various combinations of DEGs (S1B Fig, S3 Table). The combination of CHOL and CRC demonstrated the highest number of shared KEGG pathways, with 26 common pathways, while colorectal cancer and liver hepatocellular carcinoma shared 16 pathways. The combination of all six cancer types exhibited the fewest shared pathways, with only 4 common pathways, including Cell Cycle and Cytokine-cytokine receptor interaction.

Comprehensive C-Map analysis

To further determine which combinations of cancers share more similar characteristics and help decipher common features across multiple tumors, we conducted Connectivity Map (C-Map) matching analysis. The top 150 upregulated and downregulated DEGs from each of the six tumor types were input into CMap for small molecule matching based on similar characteristics. All the cell lines were applied in this process. Matching scores for the same drug treated by the respective cancer type cell lines were obtained by combining the median scores into a unique score. The matching drugs across different cancer types were overlapped in pairs to generate a list of co-acting drugs (see Method). The results indicate that the liver-intestinal interaction involves the highest number of small molecules, totaling 1147 (Fig 2A), and the highest absolute values of small molecule matching scores. It is important to note that due to the lack of CMap data for CHOL and ESCA cell lines, these two cancer types were excluded from the analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Post-CMap Matching Analysis of Compounds and Genes.

(a) Bar diagram representation of the intersection of CMap compounds matched to four types of cancer, quantifying the shared compounds among two, three, and four cancer types. (b) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment pathways of CMap knockdown genes matched to differentially expressed genes (DEGs) in CRC. (c) KEGG enrichment pathways of CMap knockdown genes matched to DEGs in LIHC.

https://doi.org/10.1371/journal.pone.0325700.g002

The CMap analyses also revealed genes that upon knockdown (KD) led to transcriptional signitures similar to the input DEG set of CRC or LIHC. Based on a criterion of norm_cs less than 1.2, we obtained 588 and 579 genes for CRC and LIHC, respectively (S4 Table). KEGG analysis indicated that enrichment of these two list of genes showed very similar pathways, expecially PI3K-Akt and MAPK signaling pathways (Fig 2B and C, S2 Fig). Therefore, we consider the liver-intestinal interaction to be more closely associated and have chosen it as the focus for subsequent research. Building upon the DEG analysis and C-Map-based drug repurposing findings from six digestive tract cancers, we have narrowed our focus to an in-depth investigation of CRC and LIHC.

Identification of shared core gene expression modules between CRC and LIHC

Next, we employed an alternative analytical strategy to further explore the commonalities in the genetic pathways between CRC and LIHC. Initially, we utilized the Rank Rank Hypergeometric Overlap (RRHO) algorithm to compare the global gene expression profiles of CRC and LIHC, which contains 17673 filtered genes (Methods). The heatmap revealed that the gene expression signatures were predominantly concentrated in the first and third quadrants, indicating a generally concordant expression trend (Fig 3A). Among these genes with similar expression trends, we identified 5,415 upregulated and 4,564 downregulated genes, totaling 9,978 genes (p < 0.05), which were then subjected to Weighted Gene Co-expression Network Analysis (WGCNA).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Identification of Shared Core Gene Expression Modules between CRC and Liver Cancer.

(a) Rank-rank hypergeometric overlap (RRHO) heatmaps reveal significant overlaps in gene expression changes between CRC and liver cancer (P value < 0.05), with white arrows indicating up- or downregulation. (b) Weighted Gene Co-expression Network Analysis (WGCNA) heatmap displays co-expression modules. Pearson’s correlation coefficients and their respective p-values are indicated. (c) Protein-protein interaction (PPI) network enriched for the Cell Cycle pathway within cluster 1, highlighting genes with high (red) and low (green) correlations. (d) PPI network for Ubiquitin mediated proteolysis within cluster 2, with color coding for correlation strength. (e) PPI network for RNA translation within cluster 3, using the same color coding scheme as in panels (c) and (d).

https://doi.org/10.1371/journal.pone.0325700.g003

The WGCNA identified 18 distinct each co-expression modules, for each of which we calculated module eigengenes (ME) and correlated them with a binary variable representing the two cancer types. We then ranked the co-expression modules based on their correlation coefficients, aiming to identify modules with high relevance to both CRC and LIHC. The results highlighted the turquoise and blue modules, which were enriched with 2,929 and 1,189 genes, respectively. Notably, these two modules exhibited a strong correlation with the gene expression patterns of both cancer types (Fig 3B). Pathway analysis of these gene sets revealed nine common pathways (p < 0.05): Cell Cycle, Nucleocytoplasmic Transport, Ubiquitin Mediated Proteolysis, ATP-dependent Chromatin Remodeling, DNA Replication, RNA Polymerase, Base Excision Repair, Non-homologous End-joining, and Endocytosis (S5 Table). A total of 328 genes were found to be enriched in these pathways, suggesting their potential as core similar gene expression modules in these two types of cancer (S5 Table).

We then combined the shared 358 DEGs between CRC and LIHC (S2 Table) and these 328 genes for protein-protein interaction (PPI) analysis using the STRING database. The gene clusters were categorized into three primary groups, each comprising 86, 37, and 36 proteins, respectively. Cluster 1 was primarily enriched in the Cell Cycle pathway, with 49 out of 86 proteins; Cluster 2 in Ubiquitin Mediated Proteolysis pathway, with 33 out of 37 proteins; and Cluster 3 in RNA Translation pathway, with 32 out of 36 proteins (S6 Table). We also searched in four databases intOGEN, OncokB, Cancer Gene Census, and Network of Cancer Genes& Healthy Drivers (NCG) to identify driver genes of these two cancers (S7 Table) and matached with these DEGs. Using Cytoscape, we then visualized the core proteins of these clusters and annotated the driver genes in the liver and intestine. In the Cell Cycle protein network (Fig 3C), CCNE1 and CHEK1 emerged as core proteins, which were also identified as driver genes in both the CRC and LIHC. Similarly, the core proteins of NXF1 and NCBP2 in the Ubiquitin Mediated Proteolysis pathway (Fig 3D), and RPS27A as a core protein in RNA Translation pathway (Fig 3E), were also drive genes of the two cancers. These pathways may hold the potential for targeted therapeutic intervention that leverages the shared molecular mechanisms of CRC and LIHC.

Selection of repositioned key drug candidates commonly applicable for both CRC and LIHC

Following the identification of shared differentially expressed genes (DEGs) and functional pathways in liver-intestine tumors, we initiated a stringent drug candidate screening process aimed at identifying potential treatments for both CRC and LIHC. This process was informed by preliminary screening results, leveraging data from the Connectivity Map (CMap) database, which encompasses a broad spectrum of cell lines (S3 Fig).

Our approach involved a refined selection of compounds based on their expression profiles in specific cell types. For LIHC analysis, we focused on HEPG2 and HUH7 cell lines, identifying 5,972 compounds. In the case of CRC, we considered HCT116, HT29, SW480, and NCIH508 cell lines, yielding 11,282 compounds. The intersection of these drug lists revealed 1,147 compounds common to both CRC and LIHC.

Given the variability in scores associated with these drugs for the two different cancers, we calculated the median value of their scores and re-ranked the drugs accordingly. This step was crucial for prioritizing compounds with a median score of ≤ −1.2, which indicated a significant potential for therapeutic intervention. Subsequently, we filtered our list to include only those compounds with documented mechanisms of action (MOA), and conducted an in-depth investigation of each compound, which encompassed the examination of compound names, drug types, and therapeutic indications from FDA, CMap, DrugBank, and PubMed databases. Further, information of driver genes from four databases intOGEN, OncokB, Cancer Gene Census, and Network of Cancer Genes& Healthy Drivers (NCG) were also included (S7 Table). The candidates were chosen for their potential to modulate the expression of shared DEGs and their involvement in key pathways, highlighting their repositioning potential for the treatment of liver and colorectal cancers. The selection process was meticulous, ensuring that the candidates not only aligned with our multi-omics analysis but also possessed a well-documented safety profile and demonstrated therapeutic relevance. Through this rigorous screening process, we established a prioritized list of core drugs, with Amonafide and BX795 emerging as the top two candidates (S8 Table). These drugs were selected for their potential to target the molecular underpinnings of CRC and LIHC, offering a promising avenue for further research and development.

Inhibition of CRC and LIHC cell proliferation and migration by Amonafide and BX795

In our pursuit to identify drugs capable of modulating the behavior of CRC and LIHC cells, we focused on the top two candidates in the list, i.e., Amonafide and BX795. To validate the inhibitory effects of Amonafide and BX795 on the proliferation of CRC (HT-29) and LIHC (HepG2), these cells were evaluated using the Cell Counting Kit-8 (CCK-8) assay. The results indicated a significant reduction in cell viability following treatment with both drugs, with a dose-dependent decrease in proliferation rates (Fig 4A). This was corroborated by colony formation assays, which revealed a marked reduction in the number of colonies formed by treated cells compared to controls (Fig 4B, S4 Fig), underscoring the drugs’ ability to suppress cell clonogenic potential.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Inhibitory Effects of Amonafide and BX795 on Cell Viability and Clonogenicity in CRC and LIHC Cells.

(a) CCK8 assay shows the time-dependent impact of Amonafide on HT29 cell viability. (b) CCK8 assay reveals the effect of BX795 on HT29 cell viability over time. (c) Amonafide’s influence on HepG2 cell viability as determined by CCK8 assay. (d) BX795’s effect on HepG2 cell viability as measured by CCK8 assay. (e) Clonogenic ability of HT29 cells treated with Amonafide assessed by colony formation assay. (f) Colony formation assay for HT29 cells treated with BX795 to determine clonogenic potential. (g) Clonogenicity of HepG2 cells under Amonafide treatment measured by colony formation assay. (h) Colony formation assay shows the effect of BX795 on HepG2 cell clonogenicity. Data represents at least three independent experiments (n ≥ 3). Statistical significance is denoted by * for P < 0.05, ** for P < 0.01, *** for P < 0.001, and **** for P < 0.0001.

https://doi.org/10.1371/journal.pone.0325700.g004

Futher, the impact of Amonafide and BX795 on cell migration was assessed using a scratch assay. The results demonstrated a significant impairment in the migration capacity of both HT-29 and HepG2 cells post-treatment, with a noticeable reduction in the distance migrated relative to control cells (Fig 2). This suggests that both drugs can impede the invasive capabilities of these cells.

Validation of effects on cell cycle and apoptosis

Flow cytometry analysis was employed to examine the effects of Amonafide and BX795 on the cell cycle progression of HT-29 and HepG2 cells. Amonafide treatment led to an accumulation of HT-29 cells in the S phase, with a 1.88-fold increase in the proportion of cells in this phase. Conversely, BX795 induced a significant increase in the proportion of cells in the G2/M phase, with a 1.75-fold elevation in HT-29 cells and a 4.17-fold increase in HepG2 cells (Fig 5A and B). These findings indicate that Amonafide and BX795 can disrupt cell cycle progression at specific checkpoints. Further, annexin V-FITC/PI staining was utilized to assess apoptosis induction. Both Amonafide and BX795 significantly increased the proportion of apoptotic cells compared to the control group, with BX795 showing a particularly pronounced effect on inducing apoptosis in both cell types (Fig 6A and B). The data collectively demonstrate that Amonafide and BX795 possess potent inhibitory effects on the growth, proliferation, and invasion of CRC and LIHC cells. These drugs not only impede cell cycle progression and migration but also promote apoptosis, highlighting their therapeutic potential in the context of liver and intestinal malignancies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Impact of Amonafide and BX795 on Cell Migration in HT29 and HepG2 Cells.

(a) Cell scratch assay displays Amonafide’s effect on HT29 cell migration; left panel shows experimental results, and right panel shows statistical analysis of migration distance. (b) Cell scratch assay illustrates the impact of BX795 on HT29 cell migration; includes both experimental results and migration distance statistics. (c) Amonafide’s influence on HepG2 cell migration as detected by cell scratch assay; presents both experimental outcome and distance quantification. (d) Cell scratch assay results and statistical analysis for HepG2 cell migration under BX795 treatment. Each experiment was repeated at least three times (n ≥ 3). Statistical significance indicated by * for P < 0.05, ** for P < 0.01, *** for P < 0.001, and **** for P < 0.0001.

https://doi.org/10.1371/journal.pone.0325700.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Modulation of Cell Cycle and Induction of Apoptosis by Amonafide and BX795.

(a) The effect of Amonafide on the cell cycle progression of HT29 cells as measured by flow cytometry. (b) Flow cytometry analysis of HT29 cell cycle perturbation by BX795. (c) Amonafide’s impact on HepG2 cell cycle as detected by flow cytometry. (d) Analysis of HepG2 cell cycle effects due to BX795 treatment using flow cytometry. (e) Apoptosis rate in HT29 cells as measured by Annexin V-FITC/PI flow cytometry. (f) Flow cytometric assessment of HT29 cell apoptosis induced by BX795. (g) Determination of HepG2 cell apoptosis rate by flow cytometry. (h) Impact of BX795 on HepG2 cell apoptosis as measured by flow cytometry. Experiments were conducted in triplicate (n ≥ 3). * denotes P < 0.05, ** for P < 0.01, *** for P < 0.001, and **** for P < 0.0001.

https://doi.org/10.1371/journal.pone.0325700.g006