B4GALT1 Is a New Candidate to Maintain the Stemness of Lung Cancer Stem Cells

Abstract

Background: According to the cancer stem cells (CSCs) hypothesis, a population of cancer cells with stem cell properties is responsible for tumor propagation, drug resistance, and disease recurrence. Study of the mechanisms responsible for lung CSCs propagation is expected to provide better understanding of cancer biology and new opportunities for therapy.

Methods: The Lung Adenocarcinoma (LUAD) NCI-H460 cell line was grown either as 2D or as 3D cultures. Transcriptomic and genome-wide chromatin accessibility studies of 2D vs. 3D cultures were carried out using RNA-sequencing and Assay for Transposase Accessible Chromatin with high-throughput sequencing (ATAC-seq), respectively. Reverse transcription polymerase chain reaction (RT-PCR) was also carried out on RNA extracted from primary cultures derived from malignant pleural effusions to validate RNA-seq results.

Results: RNA-seq and ATAC-seq data disentangled transcriptional and genome accessibility variability of 3D vs. 2D cultures in NCI-H460 cells. The examination of genomic landscape of genes upregulated in 3D vs. 2D cultures led to the identification of 2D cultures led to the identification of Beta-1,4-galactosyltranferase 1 (B4GALT1) as the top candidate. B4GALT1 as the top candidate. B4GALT1 was validated as a stemness factor, since its silencing caused strong inhibition of 3D spheroid formation.

Conclusion: Combined transcriptomic and chromatin accessibility study of 3D vs. 2D LUAD cultures led to the identification of B4GALT1 as a new factor involved in the propagation and maintenance of LUAD CSCs.

Keywords: ATAC-seq; B4GALT1; CSCs; LUAD; NSCLC; RNA-seq; cancer stem cells; genome-wide; lung cancer; transcriptome.

Figure 1

Graphical overview of the investigation. (A) 2D and 3D cell cultures were obtained from the stable NCI–H460 cell line obtained from the malignant pleural effusion of a lung adenocarcinoma (LUAD) patient. Total RNA was extracted and subjected to RNA-seq. Nuclei were isolated and processed to perform ATAC-seq. (B) Computational workflow developed to identify differentially expressed genes, pathways and biological processes in 3D compared to 2D cell cultures.

Figure 2

Analysis of 2D vs. 3D culture transcriptomes. (A) Volcano plot of over-expressed or under-expressed genes in 3D transcriptome vs. 2D. (blue: significantly under-expressed transcripts; red: significantly over-expressed transcripts). (B) Functional enrichment for upregulated genes in 3D cultures.

Figure 3

Analysis of DNA accessibility in 3D vs. 2D cultures. (A) ATAC-seq signal enrichment of normalized read counts at differential accessible sites in 3D vs. 2D (red = open; blue = closed). (B) Heatmap of differential chromatin accessibility sites (N = 404) showing the cell condition specificity of 2D and 3D ATAC-seq peaks. (C) Regulatory landscape of B4GALT1 locus. Peaks represent Assay for Transposase Accessible Chromatin with high-throughput sequencing (ATAC-seq) raw signals. (D) Reverse transcription polymerase chain reaction (RT-PCR) validation level of B4GALT1 upregulated in 3D cells compared to 2D culture in NCI-H460 (stable lung cell line) and BBIRE T-238, BBIRE T248 (primary lung cell lines). H3 reference gene have been used for normalization. Bars represent the mean of three biological replicates and technical replicates with their corresponding standard error of the mean (SEM). *** p < 0.0002; * p < 0.02; ** p < 0.0085.

Figure 3

Analysis of DNA accessibility in 3D vs. 2D cultures. (A) ATAC-seq signal enrichment of normalized read counts at differential accessible sites in 3D vs. 2D (red = open; blue = closed). (B) Heatmap of differential chromatin accessibility sites (N = 404) showing the cell condition specificity of 2D and 3D ATAC-seq peaks. (C) Regulatory landscape of B4GALT1 locus. Peaks represent Assay for Transposase Accessible Chromatin with high-throughput sequencing (ATAC-seq) raw signals. (D) Reverse transcription polymerase chain reaction (RT-PCR) validation level of B4GALT1 upregulated in 3D cells compared to 2D culture in NCI-H460 (stable lung cell line) and BBIRE T-238, BBIRE T248 (primary lung cell lines). H3 reference gene have been used for normalization. Bars represent the mean of three biological replicates and technical replicates with their corresponding standard error of the mean (SEM). *** p < 0.0002; * p < 0.02; ** p < 0.0085.

Figure 4

B4GALT1 gene expression and survival analysis in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) patients. (A) On the left panel is shown a Box Plot of B4GALT1 expression in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). Each dot represents a patient. (red: cancer tissues; grey: healthy tissues). On the right panel is shown a survival curves depict the B4GALT1 prognostic value in LUAD (N = 239 high expression tissues +239 low expression tissues) and LUSC cohort (N = 241 high expression +241 low expression). Comparison of survival curves was performed using a log-rank (Mantel–Cox) test. HR = Hazard ration. Dotted lines represent the 95% of Confidence Interval. (B) Kaplan–Meier curves depict the cumulative prognostic value of B4GALT1 and SCD1 gene expressions in LUAD (N = 250 high expression tissues + 254 low expression tissues) and LUSC (N = 153 high expression tissues + 342 low expression tissues) Abbreviations: LUAD, lung adenocarcinoma; LUSC lung squamous cell carcinoma; num, number; T, tumor; N, normal.

Figure 4

B4GALT1 gene expression and survival analysis in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) patients. (A) On the left panel is shown a Box Plot of B4GALT1 expression in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). Each dot represents a patient. (red: cancer tissues; grey: healthy tissues). On the right panel is shown a survival curves depict the B4GALT1 prognostic value in LUAD (N = 239 high expression tissues +239 low expression tissues) and LUSC cohort (N = 241 high expression +241 low expression). Comparison of survival curves was performed using a log-rank (Mantel–Cox) test. HR = Hazard ration. Dotted lines represent the 95% of Confidence Interval. (B) Kaplan–Meier curves depict the cumulative prognostic value of B4GALT1 and SCD1 gene expressions in LUAD (N = 250 high expression tissues + 254 low expression tissues) and LUSC (N = 153 high expression tissues + 342 low expression tissues) Abbreviations: LUAD, lung adenocarcinoma; LUSC lung squamous cell carcinoma; num, number; T, tumor; N, normal.

Figure 5

Knockdown of B4GALT1 RNA decreases the 3D structure formation and the expression of stemness markers. (A) Representative images of silencing of B4GALT1 reduction of 3D structure formation potential compared to the scramble of H460 cells. Scale bar = 100 µm. (B) Graphs show that silencing of B4GALT1 in 3D induces a decrease of volume and number of 3D spheroids. Number and volume of the 3D cells counted in each well after four days of culture. (C) Validation of B4GALT1 silencing in 3D cells, the results show a strongly decreases of stemness markers mRNA levels, such as Oct4, Sox2, Nanog, and SCD1. Expression of each gene was normalized to that of H3. (D) ALDH activity decrease substantially in 3D siB4GALT1 vs. Scramble cells. Experiments were performed in triplicate, and the background interference and the blank value were subtracted from the absorbance of the samples. In the bar plots, the mean ± standard error of mean (SEM) was shown from at least three independent experiments * p< 0.05, **p< 0.005, **** p< 0.0001 (vs. scramble).