Systems-level effects of ectopic galectin-7 reconstitution in cervical cancer and its microenvironment

Abstract

Background: Galectin-7 (Gal-7) is negatively regulated in cervical cancer, and appears to be a link between the apoptotic response triggered by cancer and the anti-tumoral activity of the immune system. Our understanding of how cervical cancer cells and their molecular networks adapt in response to the expression of Gal-7 remains limited.

Methods: Meta-analysis of Gal-7 expression was conducted in three cervical cancer cohort studies and TCGA. In silico prediction and bisulfite sequencing were performed to inquire epigenetic alterations. To study the effect of Gal-7 on cervical cancer, we ectopically re-expressed it in the HeLa and SiHa cervical cancer cell lines, and analyzed their transcriptome and SILAC-based proteome. We also examined the tumor and microenvironment host cell transcriptomes after xenotransplantation into immunocompromised mice. Differences between samples were assessed with the Kruskall-Wallis, Dunn's Multiple Comparison and T tests. Kaplan-Meier and log-rank tests were used to determine overall survival.

Results: Gal-7 was constantly downregulated in our meta-analysis (p < 0.0001). Tumors with combined high Gal-7 and low galectin-1 expression (p = 0.0001) presented significantly better prognoses (p = 0.005). In silico and bisulfite sequencing assays showed de novo methylation in the Gal-7 promoter and first intron. Cells re-expressing Gal-7 showed a high apoptosis ratio (p < 0.05) and their xenografts displayed strong growth retardation (p < 0.001). Multiple gene modules and transcriptional regulators were modulated in response to Gal-7 reconstitution, both in cervical cancer cells and their microenvironments (FDR < 0.05 %). Most of these genes and modules were associated with tissue morphogenesis, metabolism, transport, chemokine activity, and immune response. These functional modules could exert the same effects in vitro and in vivo, even despite different compositions between HeLa and SiHa samples.

Conclusions: Gal-7 re-expression affects the regulation of molecular networks in cervical cancer that are involved in diverse cancer hallmarks, such as metabolism, growth control, invasion and evasion of apoptosis. The effect of Gal-7 extends to the microenvironment, where networks involved in its configuration and in immune surveillance are particularly affected.

Keywords: Cervical cancer; Differential network analysis; Galectin-7; Microenvironment crosstalk.

Fig. 1

Study design and galectin expression in cervical cancer. a Pipeline of the complete experimental approach. b Gene expression profiles of galectin family members in clinical samples (normal cervix and squamous cell carcinoma, SCC) and CaCx cell lines obtained from the Scotto cohort. c Gal-7 transcription in normal cervical tissue, high grade squamous intraepithelial lesions (HSIL) and squamous cell carcinomas (SCC) from the Zhai cohort. d qPCR and Western blot analysis of Gal-7 in primary keratinocytes (PK), HPV 16 E6, E7 and E6/E7 immortalized human keratinocytes, and CaCx cells (CaSki, SiHa, HeLa). e Analysis of Gal-7 expression in clinical samples (derived from HPV16 positive SCCs) and normal tissue. (^BKruskall-Wallis Test, P < 0.0001, Dunn’s Multiple Comparison Test P < 0.05; ^ET Test, P < 0.05, *** means highly significant, ** moderately significant)

Fig. 2

Analysis of Gal-1/-7 expression and survival in the TCGA-CESC cohort. a Hierarchical clustering of Gal-7 and Gal-1 expression in a 185-patient panel of CESC from the TCGA. b Gal-7 and Gal-1 expression in TCGA-CESC patients. c Kaplan-Meier curve for 5000 days of overall survival in the CESC panel of TCGA. Censored events were marked with vertical black lines. d Inverse correlation of Gal-7 expression and methylation. e Immunohistological staining of Gal-1 and Gal-7 in normal and cervical cancer tissue sections (immunohistochemistry images were taken from the Human Protein Atlas Project). Gal-1 was detected using the HPA000646 antibody. Gal-7 was detected using the HPA001549 antibody. (^BMann Whitney test P < 0.0001 Gaussian Approximation two-tailed P-value; *** means highly significant. ^Clog-rank test; p = 0.005. ^D P < 0.0001, Spearman r, two tailed P value, alpha = 0.05)

Fig. 3

Methylation analysis and re-expression of Gal-7 in CaCx cell lines. a qPCR and Western blot of Gal-7 four days after 5′-Azacyditine (Aza) treatment. Actin was used as the loading and transfer control. b Above: schematic overview of the Gal-7 gene: the locations of CpG sites, the transcriptional start site (TSS) and the first exon are indicated. Below: percentage of methylation after bisulfite pyrosequencing analysis. c Western blot after reconstitution of Gal-7 in CaCx cells and mock infected control cells. Tubulin was used as the loading and transfer control. d Colony formation assay of Gal-7+ CaCx cells (Gal-7+, red) versus mock-infected cells (Gal-7-, black). e Induction of apoptosis due to loss of the mitochondrial transmembrane potential after incubating Gal-7+ and Gal-7- CaCx cells with increasing amounts of H14-1. (^CKruskall-Wallis P < 0.0001, Dunn’s Multiple Comparison Test P < 0.05; ^{E, F G} T Test, P < 0.05, ** means moderated significant)

Fig. 4

Changes in gene and protein expression in Gal-7+ CaCx cell lines. a Experimental strategy for differential gene expression analysis in Gal-7+ CaCx cell lines. b SILAC proteome experiments for the profiling of Gal-7+ cell lines. c-f Differentially expressed transcripts in HeLa (c) and SiHa (e) Gal-7+ cell lines in vitro with respect to Gal-7- controls. The X axis is the P value in log10 scale; the Y axis is the fold change. Differentially expressed proteins between HeLa (d) and SiHa (f) Gal-7+ cell lines in vitro are shown aside. The X axis represents the score in log10 scale (the identification score was obtained from the average score reported by triplicate identification experiments) and the Y axis is the fold change (FC). g, h Functional modules affected in Gal-7+ HeLa (g) and SiHa (h) cells (red: up-regulation, blue: down-regulation)

Fig. 5

The microenvironment induces changes in Gal-7+ tumors. a Experimental strategy. b and c Growth of Gal-7+ and Gal-7- Hela and SiHa tumors. d and e Incidence of tumors in injected mice. f, g Gene profiling analyses in Gal-7+ and Gal-7- HeLa and SiHa tumors. X axis: P value in log10 scale, Y axis: fold change (FC). h, i Functional modules and differentially expressed transcripts in Gal-7+ HeLa. h and SiHa (i) tumors in vivo. (^{B C D E} ***p < 0.001; **p < 0.01; *p < 0.05 Student’s t test). (red: up-regulation, blue: down-regulation)

Fig. 6

Differential network analysis of Gal-7+ CaCx cells and tumors. a Integrative analysis strategy. b Common candidates that were differentially regulated at the transcriptional or translational level compared against their Gal-7- controls. Numbers indicate fold changes. c, d Differential network analysis in Gal-7+ HeLa and SiHa cells and tumors. Color depth represents the fold change with respect to the control. Red: up-regulation, blue: down-regulation. Major transcription factors (MTF) are shown in the center. The colored lines derived from MTFs show the transcriptional relationship. The shape of the symbols indicates the following: square for transcript, circle for protein, hexagon for TF, diamond for miRNA, and triangle for lncRNA. Wide, green shape edges represent that the information for this molecule was obtained in vitro, while the thin, black edges mean in vivo molecule information

Fig. 7

Differential gene profiling of the Gal-7+ tumor associated microenvironment. a Experimental strategy. b, c Differential gene profiles of the of Gal-7+ HeLa and SiHa TAM with respect to controls. X axis: P value in log10 scale; Y axis: fold change (FC). d, e Modules and differentially expressed transcripts in Gal-7+ HeLa (d) and SiHa (e) TAM (up-regulation in red, down-regulation in blue)

Fig. 8

Differential network analysis of the Gal-7+ tumor associated microenvironment. Differential network analysis in Gal-7+ HeLa (a) and SiHa (b) TAM. Color depth represents the fold change. MTFs are shown in the center. The colored lines derived from MTFs show the transcriptional relationship. The shape of the symbols indicates the following: square for transcript, hexagon for TF, and triangle for lncRNA (up-regulation in red, down-regulation in blue)

Fig. 9

Bidirectional crosstalk between Gal-7+ tumors and their associated microenvironments. The expression of genes related to stromal cells, cancer associated fibroblasts, and immune cells results in an aggressive crosstalk that could potentially decide the fate of the tumor: proliferation and immune evasion, or immune destruction. The role of the common proteins in the Gal-7 network remains partially unknown. The color of the participant transcripts and proteins represents differential expression (up-regulation in red, down-regulation in blue)