Binding sites for ETS family of transcription factors dominate the promoter regions of differentially expressed genes in abdominal aortic aneurysms

Abstract

Background: Previously, we identified 3274 distinct differentially expressed genes in abdominal aortic aneurysm (AAA) tissue compared with nonaneurysmal controls. As transcriptional control is responsible for these expression changes, we sought to find common transcriptional elements in the promoter regions of the differentially expressed genes.

Methods and results: We analyzed the up- and downregulated gene sets with Whole Genome rVISTA to determine the transcription factor (TF) binding sites overrepresented in the 5-kb promoter regions of the 3274 genes. The downregulated gene set yielded 144 TF binding sites that were overrepresented in the subset when compared with the entire genome. In contrast, the upregulated gene set yielded only 13 distinct overrepresented TF binding sites. Interestingly, as classified by TRANSFAC, 8 of the 13 TFs binding to these regions belong to the ETS family. Additionally, nuclear factor kB and its subunits p50 and p65 showed enrichment. Immunohistochemical analyses of 10 TFs from the upregulated set showed 9 to be present in AAA tissue. Based on gene ontology analysis of biological process categories of the upregulated target genes of enriched TFs, 10 TFs had enrichment in immune system process among their target genes.

Conclusions: Our genome-wide analysis provides further evidence of ETS and nuclear factor kB involvement in AAA. Additionally, our results provide novel insight for future studies aiming to dissect the pathogenesis of AAA and have uncovered potential therapeutic targets for AAA prevention.

Figure 1

Study design. Overall design of the enrichment analysis of transcription factor binding sites (TFBSs) in 5 kb promoter regions upstream from transcription start sites among differentially expressed genes in abdominal aortic aneurysm (AAA). The binomial distribution was used for statistical analyses with a p-value cutoff of 0.006. FDR, False-Discovery Rate.

Figure 2

Sequence logo representations for the transcription factor binding sites (TFBS) enriched in the upregulated gene set. Standard sequence logo representations of the binding site sequence matrices are shown in the column TFBS. Sites are identified, and grouped by family, according to the TRANSFAC^© database. For each site the transcription factor is identified by protein name, gene symbol and gene ID. Also shown are the number of sites in the gene set, in the genome and the P value for the comparison. Symbols: *, transcription factors, also called binding proteins; †, gene symbol and gene identifier obtained from the National Center for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov/); ‡, negative log base 10 of the P value from the binomial distribution; §, No transcription factor identified in human; and ‖, factors also bind as a heterodimer to the respective site. Whole Genome rVISTA reported monomeric binding sites for NFKB p65 and NFKB p50 in addition to the NFKB site.

Figure 3

Schematic of the 5 kb promoter region in MMP9. Upper portion of the figure shows conserved, aligned binding sites in the 5 kb promoter region of MMP9 for the enriched TFBSs in the upregulated gene set analysis. In the lower portion of the figure the y-axis demonstrates the conservation (%) between mouse and human, and the x-axis represents the location on the human chromosome. The genomic sequence is in kilobase pairs (k). Conserved regions above the level of 70%/100 bp are highlighted under the curve, with pink indicating a conserved noncoding region; and blue, a conserved exon. UTR, untranslated region; CNS, conserved non-coding sequence; LINE, long interspersed nuclear element; SINE, short interspersed nuclear element. Figure modified from Whole Genome rVISTA output.

Figure 4

Immunohistochemical staining for transcription factors with enriched binding sites. Immunostaining was performed in both AAA tissue (first and third column) and non-aneurysmal control aorta (second and fourth column) for PEA3 (A,B), ELF1 (C,D), ETS2 (E,F), ETS1 (G,H), NFKB p65 (I,J), NFKB p50 (K,L), GABP alpha (M,N), AML1 (O,P), and STAT1 (Q,R). Positive staining is seen for all proteins in both AAA tissue and non-aneurysmal control aorta with the exception of PEA3. Bars=20 µm.

Figure 5

Box and whisker plots of mRNA levels for transcription factors with statistically overrepresented binding sites in the promoters of the upregulated gene set. Results are based on microarray expression profiling which have been described previously. RNA expression levels in signal intensity units are indicated on the ordinate-axis in logarithmic scale. Open boxes represent AAA patients (n = 7) and shaded boxes represent abdominal aorta samples from non-aneurysmal autopsy samples (n = 7). Thick horizontal bars in the boxes indicate median values, boxes indicate interquartile range, whiskers indicate range of non-outlier values, circles indicate outliers less than 3 interquartile range units. Gene symbols available from the National Center for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov/) were used. Asterisks indicate statistically significant results after FDR correction. ND, not detected, indicates that PEA3 (ETV4) was scored as not expressed since the signals were not statistically different from background noise at the 99% confidence level. For numerical values of the expression data, see Table S6.

Figure 6

Functional classification of target genes for ELF1 (A), AML1 (RUNX1, B) and ISGF-3 (STAT1, C). DAGs were generated using WebGestalt, and re-drawn in Adobe Illustrator. BP categories with enrichment at P<0.001 using the hypergeometric test are shown in red, whereas parent nodes with P>=0.001 are shown in black. Immune-related categories are shown in yellow boxes. For a complete listing of these target genes and their GO annotations, see Tables S5, S8 and S9.

Figure 6

Functional classification of target genes for ELF1 (A), AML1 (RUNX1, B) and ISGF-3 (STAT1, C). DAGs were generated using WebGestalt, and re-drawn in Adobe Illustrator. BP categories with enrichment at P<0.001 using the hypergeometric test are shown in red, whereas parent nodes with P>=0.001 are shown in black. Immune-related categories are shown in yellow boxes. For a complete listing of these target genes and their GO annotations, see Tables S5, S8 and S9.