Evidence of perturbations of cell cycle and DNA repair pathways as a consequence of human and murine NF1-haploinsufficiency

Abstract

Background: Neurofibromatosis type 1 (NF1) is a common monogenic tumor-predisposition disorder that arises secondary to mutations in the tumor suppressor gene NF1. Haploinsufficiency of NF1 fosters a permissive tumorigenic environment through changes in signalling between cells, however the intracellular mechanisms for this tumor-promoting effect are less clear. Most primary human NF1+/- cells are a challenge to obtain, however lymphoblastoid cell lines (LCLs) have been collected from large NF1 kindreds. We hypothesized that the genetic effects of NF1-haploinsufficiency may be discerned by comparison of genome-wide transcriptional profiling in somatic, non-tumor cells (LCLs) from NF1-affected and -unaffected individuals. As a cross-species filter for heterogeneity, we compared the results from two human kindreds to whole-genome transcriptional profiling in spleen-derived B lymphocytes from age- and gender-matched Nf1+/- and wild-type mice, and used gene set enrichment analysis (GSEA), Onto-Express, Pathway-Express and MetaCore tools to identify genes perturbed in NF1-haploinsufficiency.

Results: We observed moderate expression of NF1 in human LCLs and of Nf1 in CD19+ mouse B lymphocytes. Using the t test to evaluate individual transcripts, we observed modest expression differences in the transcriptome in NF1-haploinsufficient LCLs and Nf1-haploinsuffiicient mouse B lymphocytes. However, GSEA, Onto-Express, Pathway-Express and MetaCore analyses identified genes that control cell cycle, DNA replication and repair, transcription and translation, and immune response as the most perturbed in NF1-haploinsufficient conditions in both human and mouse.

Conclusions: Haploinsufficiency arises when loss of one allele of a gene is sufficient to give rise to disease. Haploinsufficiency has traditionally been viewed as a passive state. Our observations of perturbed, up-regulated cell cycle and DNA repair pathways may functionally contribute to NF1-haploinsufficiency as an "active state" that ultimately promotes the loss of the wild-type allele.

Figures

Figure 1

Quantitative PCR and western blot of the relative abundance of neurofibromin in lymphoblastoid cell lines. A) Relative abundance of NF1 mRNA in Coriell-18 and ECACC-6 sets as detected by quantitative PCR. P values of respective t tests are shown above the bar plots. B) Western blot of neurofibromin in select lymphoblastoid cell lines (LCLs). TUBB - β-tubulin. C) Relative abundance of neurofibromin in select LCLs. NF1 abundance in sample E2-U was arbitrary set at 100%; NF1 abundance in the remaining LCLs is expressed as percentage of that in E2-U.

Figure 2

Quantitative PCR analysis of relative abundance of Nf1 mRNA in Nf1+/- mice. A) Plot of real-time amplification of Nf1 and Gapdh mRNAs in six wild-type and six Nf1+/- mice. Samples containing no input RNA are shown as "Negative control". B) Relative abundance of Nf1 mRNA from wild-type and Nf1+/- mice. P value of the t test is shown above the bar plot.

Figure 3

Intersection analysis between top-ranking transcripts in human and mouse sets. Top ~5% of all transcripts from analysis of three sample sets on the Illumina platform were chosen for the intersection analysis. A) Coriell-18 and Nf1-Mouse-12 comparison; B) ECACC-6 and Nf1-Mouse-12 comparison; C) Coriell-18 and ECACC-6 comparison; D) Three-way comparison. "Mm" designates Nf1-Mouse-12 set.

Figure 4

Comparison of expression profiles generated by Illumina_Human_WG_v2 and spotted oligonucleotide microarrays. A) Coriell-18 samples were analyzed in parallel on two different types of microarrays: Illumina_Human_WG_v2 ("Illumina") and spotted oligonucleotide microarrays manufactured in the NHGRI core facility ("NHGRI"). Genes were ranked according to nominal P values of the t tests, and top ~5% of genes in each list were chosen for the intersection analysis; B) Bar plot represents fold difference of expression values (NF1-affecteds vs. NF1-unaffecteds) for the overlapping genes. Each overlapping transcript is represented by a pair of bars. Blue bars denote transcripts on Illumina platform; orange bars denote transcripts on the NHGRI platform. Numbers on X-axis represent transcripts in the overlap (114 total).

Figure 5

GSEA analysis of human and mouse expression datasets and ontological categories of LEA gene lists. Up- and down-regulated statistically significant GSEA gene sets were subjected to LEA analysis and resulting genes were grouped in related ontological categories. The pie chart diagrams of up- and down-regulated LEA genes are shown on the left and on the right, respectively. The same ontological categories are represented by the same color on the pie charts. Only the first six most abundant ontological categories are shown for each set. To save space, the following ontological categories were shortened as follows: Translation -Translation/Protein biosynthesis/Ribosome biogenesis; Transcription - Transcription/RNA processing; DNA repair - DNA repair/replication/recombination; Cell cycle - Cell cycle/Mitosis/Cytokinesis; Immune response - Immune/Defense/Antiviral/Inflammatory response.