MicroRNA-99a and 100 mediated upregulation of FOXA1 in bladder cancer

Abstract

Urothelial cell carcinoma of the bladder (UCC) is a common disease often characterized by FGFR3 dysregulation. Whilst upregulation of this oncogene occurs most frequently in low-grade non-invasive tumors, recent data reveal increased FGFR3 expression characterizes a common sub-type of invasive UCC sharing molecular similarities with breast cancer. These similarities include upregulation of the FOXA1 transcription factor and reduced expression of microRNAs-99a/100. We have previously identified direct regulation of FGFR3 by these two microRNAs and now search for further targets. Using a microarray meta-database we find potential FOXA1 regulation by microRNAs-99a/100. We confirm direct targeting of the FOXA1 3'UTR by microRNAs-99a/100 and also potential indirect regulation through microRNA-485-5p/SOX5/JUN-D/FOXL1 and microRNA-486/FOXO1a. In 292 benign and malignant urothelial samples, we find an inverse correlation between the expression of FOXA1 and microRNAs-99a/100 (r=-0.33 to -0.43, p<0.05). As for FGFR3 in UCC, tumors with high FOXA1 expression have lower rates of progression than those with low expression (Log rank p=0.009). Using global gene expression and CpG methylation profiling we find genotypic consequences of FOXA1 upregulation in UCC. Genetic changes are associated with regional hypomethylation, occur near FOXA1 binding sites, and mirror gene expression changes previously reported in FGFR3 mutant-UCC. These include gene silencing through aberrant hypermethylation (e.g. IGFBP3) and affect genes characterizing breast cancer sub-types (e.g. ERBB2). In conclusion, we have identified microRNAs-99a/100 mediate a direct relationship between FGFR3 and FOXA1 and potentially facilitate cross talk between these pathways in UCC.

Figure 1. Regulation of FOXA1 expression by microRNAs-99a/100 in bladder cancer

(a). Western blotting for protein expression reveals upregulation of FGFR3, phosphorylation of ERK1/2 and FOXA1 following transfection of NHU cells with anti-miRs to miRs-99a/100 but not the scrambled RNA control. Image densitometry reveals 3-fold increases in expression for FOXA1 with anti-miRs-99a/100. (b). A reporter construct assay reveals 1.3/1.4 fold increases in luciferase expression in NHU cells treated with anti-miRs-99a/100, when normalized to the scrambled control. Expression profiling using QrtPCR in 220 benign and malignant urothelial samples revealed (c). overlap of upregulation for FOXA1 and FGFR3 in UCC, which is greatest in low-grade NMI cancers, and (d). lower rates of progression to more advanced disease for tumors with high FGFR3 and high FOXA1 when compared to those with low expression (the difference in greatest in low stage and low tumors (supplementary figure 3)). (e). Genes with upregulation following FOXA1 transfection in T24/EJ cells are located closer to known FOXA1 binding sites (in HepG2 cells) than those with no change in expression or those with reduced expression (down regulation).

Figure 2. Symmetry of genotype between FOXA1 transfected cells, FGFR3 mutant tumors and in sporadic UCC

Gene expression changes following FOXA1 transfection in T24/EJ cells share many similarities with those found in FGFR3 mutant bladder UCC. In (a). we stratify changes in normalized gene expression following FOXA1 transfection within T24/EJ by those seen in FGFR3 mutant UCC. Genes upregulated in FGFR3 mutant tumors (labeled as Increased) have significantly higher expression in FOXA1 transfected cells than those with no change in FGFR3 mutants. The difference is largest for genes with associated DNA hypomethylation. No significant difference is seen for genes with decreased expression or DNA hypermethylation. (b). Area proportional Venn diagrams reveal the overlap for genes in FGFR3 mutant UCC and those upregulated following FOXA1 transfection or with increases in DNA hypomethylation following FOXA1 transfection (left). Overlap was also seen for genes downregulated in FGFR3 mutant UCC, and those down regulated or hypermethylated following FOXA1 transfection. We identified a 156 gene cohort with reciprocal, symmetrical expression in FOXA1 transfected cells and FGFR3 mutant UCC and analyzed this in a publically deposited microarray dataset [32]. In (c). we plot the fold change of these 156 genes as a ratio for low-grade NMI cancers relative to high-grade NMI cancer (orange triangle) and invasive cancers (blue). In (d). we plot the correlation coefficient vales (r) between expression of FGFR3 and FOXA1 in the 256 sporadic UCC [32]. The 156 genes in (c) and (d) are ordered according to expected fold changes (bar: green is loss of expression, red is increased expression) as seen in FOXA1 transfected cells. Significance of difference in fold change (low grade NMI versus more aggressive cancers) is shown as *p<0.05 and **p<0.01.