Abstract
Abstract 5119
Diffuse Large B cell Lymphoma (DLBCL) is a clinically and biologically heterogeneous lymphoid malignancy, where complete response is achieved in up to 60% of patients, though over half of all patients relapse (Hunt and Reichard, 2007). The prognostic significance of the expression of FOXP1 transcription factor in DLBCL has been highly controversial with some studies describing immunohistochemical FOXP1 expression as a poor prognostic marker (Banham et. all. 2005). Recent transcript analysis has identified over 10 FOXP1 isoforms where some exhibit COOH- or NH3- truncations and also the absence of regulatory domains. These truncated isoforms are highly expressed in ABC-DLBCL (Brown et. al. 2008) and it is suggested that truncated isoforms may aberrantly regulate FOXP1 target genes. Although there have been many studies debating the prognostic significance of FOXP1 overexpression in DLBCL, there has been little examination of the function of FOXP1 in DLBCL, specifically the genes targeted by FOXP1, and how the function of FOXP1 isoforms may contribute to a more aggressive phenotype.
Expression constructs for 4 FOXP1 isoforms (isoforms 1, 2, 3 and 8) were kindly provided by Dr. Philip Brown and Dr. Alison Banham. Stable cell lines were established by transfecting the BJAB cell line with either FOXP1 isoform 1 or the empty pBKCMV construct, and selection with G418 antibiotic. Total RNA was extracted from both cell lines and 4 DLBCL tumours (with low to high FOXP1 expression), and gene expression was analysed using Illumina HT12v4 microarray. Genes with >2-fold changes in expression between pBKCMV and Isoform 1 stable cell line were investigated. VisANT analysis and Gene set enrichment was performed on differentially expressed genes and candidates were validated by qRT-PCR. Promoter analysis using MatInspector revealed forkhead-binding sites in the promoter region of candidate genes, and dual luciferase reporter gene assays were performed to ascertain promoter regulation by FOXP1 isoforms. Western blot was also performed to validate translational changes in expression.
In the gene expression microarray analysis of FOXP1 isoform 1 expression compared to baseline levels, 271 genes were identified with >2-fold differential expression. High versus low FOXP1 expressing DLBCL patients were also compared, where it was found that 2472 genes were differentially expressed over the 2-fold level. Significantly, comparison of differentially expressed genes in both the cell line and patient samples revealed all 271 genes differentially expressed in the cell line overlapped with the patient genes, indicating these results are translatable to FOXP1 function in DLBCL tumours.
Pathway analysis was performed on the differentially expressed genes, and unexpectedly there were no significant enrichment of curated pathways. Alternatively visANT was used to examine interactive networks of the 271 candidate genes, and an integrative analysis using the FOXP1 microarray expression data was used to enrich the interactions. This analysis revealed critical interactions that could not be revealed by pathway analysis, and also identified genes that are likely FOXP1 targets.
The promoter regions of 5 genes were investigated using reporter gene assays, and FOXP1 isoforms 3 and 8 were found to be significantly stronger transcriptional repressors compared to isoforms 1 and 2. For example, one of the genes identified, heterogeneous nuclear ribonucleoprotein U (HNRNPU) was highly expressed in the presence of isoform 1, however in the presence of isoform 3 was greatly downregulated. Although this gene has not previously been associated with DLBCL, HNRNPU is a potent regulator of snRNP's thereby having global effects on mRNA production and protein synthesis (Xiao et. al. 2012 Cell Vol. 45(5) 656–668). We are currently investigating the biological significance of other novel FOXP1 target genes, and how the truncated isoforms are involved in the differential regulation of these genes.
We have identified 271 genes that are differentially expressed in both a cell line model and DLBCL patient tissues with high versus low FOXP1 expression. These are novel genes in DLBCL pathogenesis that also show differential expression in the presence of truncated isoforms, indicating truncated isoforms may have aberrant functions and may contribute to a more pathogenic phenotype.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.