NKX3.1 Expression Contributes to Epithelial-Mesenchymal Transition of Prostate Cancer Cells

Abstract

Studies demonstrate that inflammation synergizes with high-grade aggressive prostate tumor development and ultimately metastatic spread, in which a lot of work has been done in recent years. However, the clear mechanism of inflammation inciting prostate cancer remains largely uncharacterized. Our previous study has shown that the conditioned media (CM)-mediated LNCaP cell migration is partially correlated with the loss of expression of the tumor suppressor NKX3.1. Here, we continue to investigate the inflammation-mediated migration of prostate cancer cells, and the role of NKX3.1 in this process to gain insights into cell migration-related changes comprehensively. Earlier, the model of inflammation in the tumor microenvironment have been optimized by our research group; here, we continue to investigate the time-dependent effect of CM exposure together with NKX3.1 changes, in which we observed that these changes play important roles in gaining heterogeneous epithelial-to-mesenchymal transition (EMT) phenotype. Hence, this is an important parameter of tumor progression; we depleted NKX3.1 expression using the CRISPR/Cas9 system and examined the migrating cell clusters after exposure to inflammatory cytokines. We found that the migrated cells clearly demonstrate reversible loss of E-cadherin expression, which is consistent with subsequent vimentin expression alterations in comparison to control cells. Moreover, the data suggest that the AR-mediated transcriptional program also contributes to mesenchymal-to-epithelial transition (MET) in prostate cancer progression. Furthermore, the quantitative proteomic analysis showed that migrated subpopulations from the same cell line presented different phenotypes in which the proteins overexpressed are involved in cell metabolism and RNA processing. According to KEGG pathway analysis, the ABC transporters were found to be the most significant. Thus, the dynamic process of cellular migration favors diverse genetic compositions under changing tumor microenvironments. The different levels of invasiveness are supported by shifting the cells in between these EMT and MET phenotypes.

Figure 1

Migrated cells display heterogeneity in a model of the inflammatory microenvironment. (A) Representative light microscopy images of migrated LNCaP subpopulations; subpopulations varied depending on the duration of CM treatment (scale bar: 200 μm). (B) Cells undergoing the suspension cell stage, named nonadhesive (NA). (C) Hierarchical scheme of the obtained subpopulations; NA refers to the group of cells that has passed through the nonadhesive stage. (D) Investigation of proteins involved in cell migration and proliferation by western blotting analysis in selected subpopulations.

Figure 2

Specific invasive phenotype of chosen cells represented by EMT-related markers. (A) Loss of E-cadherin and overexpression of proliferative p-H3^(S10) in 12h_5(NA) cells. Primary LNCaP cells were chosen as a control. (B) 12h_5(NA) cells simultaneously gained N-cadherin expression. Metastatic PC-3 cells were used as a positive control. (C) Other subpopulations from 12 h CM-treated and migrated cells together with migrated control cells showed either enhanced or decreased expressions of N-cadherin, representing the heterogeneity of migrated cells. (D) Immunofluorescence images of 12h_5(NA) and 6h_(NA) subpopulations stained with the mesenchymal marker vimentin (green channel) and DAPI (blue channel).

Figure 3

Alteration of EMT factors in NKX3.1-silenced cells. (A/D) Schematic diagram of inflammation-related migration assay in NKX3.1-silenced LNCaP cells and (G) obtained subpopulations. (B) Verification of NKX3.1 silencing by western blotting (up) and quantification of NKX3.1 expression compared to the control (down). (C) Immunofluorescence images of NKX3.1-silenced cells are shown in green GFP expression. Some cells still showed red NKX3.1 positivity. (E) Effect of CM in NKX3.1-silenced cells verified by western blotting. (F) Light microscopy images of migrated cells showing an aggressive phenotype compared to the control. Scale bars: 200 μm. (H/I) Control of E-cadherin, vimentin, and NKX3.1 expression in migrated subpopulations.

Figure 4

Investigation of the heterogeneous phenotype of subpopulations. (A) Colony-forming abilities of selected subpopulations. The graphical representation of the number of colonies at different sizes is shown. Each color represents a different size [pixel] of counted colonies. Analysis was performed using Image J software, and data are means ± SD of three independent experiments (p < 0.001). (B) Epigenetic changes investigated by SIRT and DNMT3a. (C) Expressions of metastasis-related BMP and MMP gene family members in subpopulations.

Figure 5

(A) Differentially expressed proteins in each subpopulation identified by the volcano plot. Genes that are nonspecific are marked black, only p-value-specific are marked blue, only FC-specific are marked green, and those that are significant according to the p-value and the FC threshold are marked red. (B) Biological replicate data plotted as a scatter plot matrix; the difference between subpopulations is represented by the Pearson correlation coefficient (R).

Figure 6

Differential protein analysis and GO enrichment analysis of the quantitative proteome. (A) Nonsupervised hierarchical clustering of the top 985 most variant proteins. (B) Principal component analysis (PCA). (C) Gene ontology (GO) term enrichment analysis of differentially expressed proteins. (D) GO results by Cnetplot. (E) KEGG pathway enrichment analysis. (F) Venn diagram showing the distribution of the 985 identified proteins in each subpopulation.