NFkB disrupts tissue polarity in 3D by preventing integration of microenvironmental signals

Abstract

The microenvironment of cells controls their phenotype, and thereby the architecture of the emerging multicellular structure or tissue. We have reported more than a dozen microenvironmental factors whose signaling must be integrated in order to effect an organized, functional tissue morphology. However, the factors that prevent integration of signaling pathways that merge form and function are still largely unknown. We have identified nuclear factor kappa B (NFkB) as a transcriptional regulator that disrupts important microenvironmental cues necessary for tissue organization. We compared the gene expression of organized and disorganized epithelial cells of the HMT-3522 breast cancer progression series: the non-malignant S1 cells that form polarized spheres ('acini'), the malignant T4-2 cells that form large tumor-like clusters, and the 'phenotypically reverted' T4-2 cells that polarize as a result of correction of the microenvironmental signaling. We identified 180 genes that display an increased expression in disorganized compared to polarized structures. Network, GSEA and transcription factor binding site analyses suggested that NFkB is a common activator for the 180 genes. NFkB was found to be activated in disorganized breast cancer cells, and inhibition of microenvironmental signaling via EGFR, beta1 integrin, MMPs, or their downstream signals suppressed its activation. The postulated role of NFkB was experimentally verified: Blocking the NFkB pathway with a specific chemical inhibitor or shRNA induced polarization and inhibited invasion of breast cancer cells in 3D cultures. These results may explain why NFkB holds promise as a target for therapeutic intervention: Its inhibition can reverse the oncogenic signaling involved in breast cancer progression and integrate the essential microenvironmental control of tissue architecture.

Figure 1. A cluster of genes revealed by microarray analysis is associated with tissue polarity in HMT3522 cell lines

A) Immunofluorescence analysis of basal marker alpha6-integrin in S1, T4-2, and T4-2 cells phenotypically reverted in 3D culture. S1 and reverted T4-2 cells formed polarized spheroids, but T4-2 cell formed disorganized structures. Blue:DAPI, green: alpha6 integrin. Scale: 25μm B) Scheme showing the pathways that were blocked to reverse the malignant phenotypes of T4-2 cells in 3D culture for the microarray analysis. C) Unsupervised clustering of the 180 genes upregulated in T4-2 cells as compared to S1 cells and reverted T4-2 cells. D) Gene Ontology terms (biological processes, cell components, molecular functions) and KEGG pathways overrepresented in the disorganization gene signature.

Figure 2. GSEA and network analyses reveal NFkB as a potential regulator of disorganization genes

A) Gene set enrichment analysis [21] finds the gene set ‘positive regulation of IKK and NFkB cascade’ to be enriched among genes that show a high expression in unorganized cells (1:'negatively correlated') and low expression in organized cells in the microarrays (0:'positively correlated') (p=0.09). B) A regulatory network consisting of the 180 genes of the disorganization signature as well as connecting nodes that interact with at least two disorganization genes was identified using Cytoscape. Red nodes: disorganization genes, light blue nodes: connecting genes, green edges: interactions found in NCBI, dark blue node: p65.

Figure 3. The NFkB pathway is activated in disorganized T4-2 cells compared to polarized S1 and reverted T4 cells

A) Immunoblotting analysis of p65 expression in S1, T4-2, and reverted T4-2 cells in 3D culture. Laminin A/C was used as a loading control. T4-2 cells were reverted by tyrphostin (Tyr), an EGFR inhibitor; LY294002 (LY), a PI3K inhibitor; GM6001 (GM), a MMP inhibitor; 2DG, a glucose metabolism inhibitor; VPA, a HDAC inhibitor; AIIB2, a β1 integrin blocking antibody, MAB225 (MAB), an EGFR blocking antibody. B) Immunoblotting analysis of phosphorylated p65 in S1, T4-2, and T4-2 cells reverted by tyrphostin. The blot results were quantified by AlphaEaseFC software, and expressed as relative levels of phosphorylated p65 to total p65. C) Immunoblotting of nuclear protein extracted from S1, T4-2, and T4-2 cells reverted with tyrphostin. D) Luciferase analysis of NFkB transcription activities in T4-2 and reverted T4-2 cells. Graph displays average ± SEM; n=4.

Figure 4. Inhibition of the NFkB pathway reverses the malignant phenotypes of T4-2 cells in 3D culture

A) Phase and immunofluorescence images of T4-2 cells. Cells were either treated with NFkB inhibitor (Wedelolactone, 10 μM) and cultured in 3D for four days orinfected with p65 shRNA lentivirus before plating in 3D culture. Blocking the NFkB pathway reprograms the cells to form polarized spheroids structure. Green: alpha6-integrin. Scale: 25μm. B) Immunoblotting analysis of beta1-integrin, EGFR and p65 expression in control and p65 shRNA-expressing T4-2 cells. C) Transwell invasion analysis of control and p65 shRNA-expressing T4-2 cells. Graph displays average ± SEM; n=3. D) Quantification of mRNA levels of NFkB target genes by real-time PCR. Control and p65 shRNA-expressing T4-2 cells were cultured in 3D for 4 days before RNA extraction.

Figure 5. Expression of disorganization genes and activation of the NF-κB pathway are associated with aggressive 3D phenotypes of multiple breast cancer cell lines

A) Unsupervised hierarchical clustering of breast cancer cell lines in 3D culture using the 180 disorganization genes. Each row represents a gene; each column represents a cell line. B) Bar graph of RelB mRNA levels in basal and luminal types of breast cancer cell lines. RelB mRNA expression is assessed by Affymetrix microarray and measured as log2 (probe intensities). C) Phase and immunofluorescence images of BT549 cells. Cells were either treated with Wedelolactone in 3D culture for 4 days or infected with RelB shRNA retrovirus before being seeded in 3D culture. D) Transwell invasion analysis of BT549 cells. The cells were treated with NF-κB inhibitor (Wedelolactone, 10 μM) or transfected with RelB shRNA vector.

Figure 6. A scheme showing the role of NF-κB as potential integrator of various microenvironmental cues and regulator of multicellular organization