Identification of MKNK1 and TOP3A as ovarian endometriosis risk-associated genes using integrative genomic analyses and functional experiments

Abstract

The risk of endometriosis (EM), which is a common complex gynaecological disease, is related to genetic predisposition. However, it is unclear how genetic variants confer the risk of EM. Here, via Sherlock integrative analysis, we combined large-scale genome-wide association studies (GWAS) summary statistics on EM (N = 245,494) with a blood-based eQTL dataset (N = 1490) to identify EM risk-related genes. For validation, we leveraged two independent eQTL datasets (N = 769) for integration with the GWAS data. Thus, we prioritised 14 genes, including GIMAP4, TOP3A, and NMNAT3, which showed significant association with susceptibility to EM. We also utilised two independent methods, Multi-marker Analysis of GenoMic Annotation and S-PrediXcan, to further validate the EM risk-associated genes. Moreover, protein-protein interaction network analysis showed the 14 genes were functionally connected. Functional enrichment analyses further demonstrated that these genes were significantly enriched in metabolic and immune-related pathways. Differential gene expression analysis showed that in peripheral blood samples from patients with ovarian EM, TOP3A, MKNK1, SIPA1L2, and NUCB1 were significantly upregulated, while HOXB2, GIMAP5, and MGMT were significantly downregulated compared with their expression levels in samples from the controls. Immunohistochemistry further confirmed the increased expression levels of MKNK1 and TOP3A in the ectopic and eutopic endometrium compared to normal endometrium, while HOBX2 was downregulated in the endometrium of women with ovarian EM. Finally, in ex vivo functional experiments, MKNK1 knockdown inhibited ectopic endometrial stromal cells (EESCs) migration and invasion. TOP3A knockdown inhibited EESCs proliferation, migration, and invasion, while promoting their apoptosis. Convergent lines of evidence suggested that MKNK1 and TOP3A are novel EM risk-related genes.

Keywords: Endometriosis; Expression quantitative trait loci; Genome-wide association study; Integrative genomics analysis; Risk genes.

Graphical abstract

Fig. 1

Schematic of framework. EM, endometriosis. GWAS, genome-wide association study. eQTL, expression quantitative trait loci. MAGMA, Multi-marker Analysis of GenoMic Annotation. GO, Gene Ontology. DGE, differential gene expression. IHC, immunohistochemical. EESCs, ectopic endometrial stromal cells.

Fig. 2

Integrative genomics analyses identify risk genes and pathways for EM. (A) Venn diagram of three identified EM-risk gene sets: Gene set #1, Gene set #2, and Gene set #3 are based on Sherlock integrative genomics analysis by combining Zeller et al. eQTL data (Dataset #3), Dixon et al. eQTL data (Dataset #4), and GTEx blood eQTL data (Dataset #5) with GWAS summary statistics on EM, respectively. (B, C) Venn diagrams of the significantly enriched pathways (B) and GO terms (C) by three identified gene sets. (D, E) The scatter diagrams showing the 19 common significant pathways (D) and 35 common significant GO terms (E) based on Gene set #1. EM, endometriosis. GO, Gene Ontology.

Fig. 3

Consistent evidence supporting the identified EM-risk genes from integrative genomics analyses. (A, B) Venn diagrams showing the overlapping genes of three Sherlock-identified EM-risk gene sets with MAGMA-identified EM-risk genes (Gene set #4; A), and with MAGMA-identified Null trait-related genes (Gene set #5, as a negative control; B). (C–E) Computer-based permutation analyses of 100,000 times for comparison of genes from Gene set #1 with that from Gene set #2 (C), Gene set #3 (D), and Gene set #4 (E). EM, endometriosis. MAGMA, Multi-marker Analysis of GenoMic Annotation.

Fig. 4

PPI-network of 14 EM-risk genes using the GeneMANIA tool. The 14 EM-risk genes were marked with red colour, and the 20 predicted genes were marked with green colour. The underlying molecular interactions between each gene pair were attributed based on the co-expression links (account for 71.51%), shared protein domains (account for 27.25%), and co-localization (account for 1.24%). PPI, protein–protein interaction.

Fig. 5

Differential gene expression and co-expression patterns of EM-risk genes in PBMCs of women with ovarian EM and health controls based on RNA-Seq. (A-G) Violin plots showing significantly different expressed genes between EM and controls for TOP3A (A), MKNK1 (B), SIPA1L2 (C), NUCB1 (D), HOXB2 (E), GIMAP5 (F), and MGMT (G). (H) Co-expression pattern analysis of 14 EM-risk genes between controls and EM. The colour legend showing the degree of correlation coefficients, red represents − 1 and blue represents + 1. PBMC, peripheral blood mononuclear cell. EM, endometriosis.

Fig. 6

Immunoreactivity of MKNK1, TOP3A, HOXB2 and NUCB1 in endometrium from women with and without ovarian endometriosis. The expression and localisation of MKNK1 (A–C), TOP3A (E–G), HOXB2 (I–K), NUCB1 (M–O) in normal, eutopic, and ectopic endometrium evaluated by IHC staining, respectively. The comparisons of IRS across three groups (D, H, L, P). Values are presented as mean± SEM. P-values were determined by Kruskal-Wallis tests followed by multiple comparisons. ∗P < 0.05, ∗∗P < 0.01. and ∗∗∗P < 0.001. IHC, immunohistochemical. IRS, immunoreactive score.

Fig. 7

The role of MKNK1 and TOP3A in proliferation, apoptosis, migration and invasion of EESCs. (A–C) mRNA (A) and protein (B–C) expression levels of MKNK1 and TOP3A in EESCs transfected with siRNA were determined by RT-qPCR (n = 3) and western blot analysis (n = 6), respectively. (D) Proliferation of EESCs transfected with si-MKNK1, si-TOP3A, and si-Ctrl was assessed with CCK-8 assay at 0, 24, 48, 72 and 96 h, n = 6. (E) Representative images and the graphical statistics of apoptosis rate assessed by flow cytometry of EESCs transfected with si-MKNK1, si-TOP3A, and si-Ctrl. n = 6. (F) Representative fields (100 × magnification) and the graphical statistics of Transwell migration and invasion assay of EESCs transfected with si-MKNK1, si-TOP3A, and si-Ctrl. n = 5. Values are presented as mean± SEM. P-values were determined by unpaired two-tailed t test. ∗P < 0.05, and ∗∗∗P < 0.001. EESCs, ectopic endometrial stromal cells. Real-time quantitative PCR (RT-qPCR). Cell counting kit-8, CCK-8.