Identification of new p53 regulatory networks through NGS data analysis

Motivation Around 50% of all human tumours carry point mutations in the p53 tumour suppressor gene, which alter p53 DNA binding specificity. In tumours with p53 wild type, p53 is often rendered functionally inert by the inactivation of its positive modulators or by the activation of negative factors, which block p53 transcriptional activities [1]. We identified a new p53 direct target gene, TRIM8, belonging to the Tripartite Motif (TRIM) protein family, defined by the presence of a RING domain, one or two B-boxes and a Coiled-Coil region. We found that TRIM8 overexpression leads, through a positive feedback loop, to p53 stabilization and p53-mediated suppression of cell proliferation. In order to identify the pathways activated by TRIM8 leading to p53 stabilization we transiently transfected with TRIM8 the HCT116-p53 (wt) cell line, and sequenced the total transcriptome performing a NGS run on a 454 GS FLX platform. Here we report some statistics and the preliminary results of: i) reads mapping on the human genome and analysis of differential expressed genes; ii) functional analysis of differentially expressed genes. Method Total RNA was extracted from HCT116-p53 (wt) cell line 48h after transfection, depleted of rRNA, retro-transcribed, amplified and sequenced by using the pyrosequencer Roche GS FLX Titanium Series. Genome mapping, statistics and differential expression analyses were performed by using the "NGS-Trex" system (NGS Transcriptome profile Explorer) (Mignone F. et al., submitted), a automatic system designed for analyzing Next Generation Sequencing data generated from large-scale transcriptome studies. The overall procedure involves three steps: 1) creation of a project and upload of reads in a multi-fasta format; 2) reads mapping onto the reference genome after setup of appropriate parameters; 3) annotation of mapped reads; 3) data mining by using simple query forms. TRIM8 and FLAG data were submitted to NGS-Trex using default parameters that can briefly summarized as follows: reads were mapped onto human genome (min similarity 90% and min overlap 50 nt) discarding reads mapping onto more than 10 genomic regions. Mapped reads were compared to annotation to assign reads to genes and to identify new splice variants. Differentially expressed genes and splicing events were identified by computing a P-value associated to an hypergeometric distribution. Housekeeping genes were used to normalise reads count before identification of differentially expressed genes. The lists of genes showing a differential expression in the two samples were then analysed by using DAVID v(6.7), an integrated biological knowledgebase and analytic tools (text and pathway-mining tools) for large gene list functional annotation [2,3]. An additional analysis on TRIM8 and FLAG sequence samples was made for the detection and annotation of the ncRNA genome fraction. We used a bioinformatic analysis pipeline, developed by us, which is able to: 1) select ncRNA from the whole read sample; 2) cluster them on the basis on their Sequence Ontology classification [4]; 3) provide statistics of expressed ncRNAs, indicating the number of represented reads for each SO category; 4) extract read collections using SO categorisation. Results The first processing of NGS data with 454 GS FLX analysis pipeline provided a total number of 400,879 reads for the TRIM8 transiently transfected sample, and 468,381 for the control (FLAG). Using the NGS-Trex system we were able to map these reads on the human genome and to get annotation data of sequenced transcripts (a custom made database based on NCBI data was used). Amongst all reads analysed by NGS-Trex, 216,489 were unambiguously mapped for the TRIM8 sample, and 303,030 for FLAG. Crossing mapping data with genome annotations, we obtained a list of 13,506 expressed genes for the TRIM8 sample and a list of 14,864 genes for FLAG. These lists were then normalised and filtered by using statistical modules implemented in NGS-Trex for the analysis of differential expressed genes in the two samples. Only those genes having a differential expression level with a P-value<0.05 were selected. Results of the analysis of differential expressed genes showed that the number of down regulated genes is significantly higher than the number of those up regulated. Functional classification and clustering of these genes, performed by DAVID, assigned them to functional categories and metabolic pathways that are known, or supposed to be, modulated by TRIM8 and p53 activities. Conclusions The results obtained by this preliminary analysis brings new insights on the cell pathways controlled by TRIM8 dependent-p53 activity. The unravelling of TRIM8-p53 signalling pathway will undoubtedly provide an opportunity for therapeutic intervention by revealing novel targets for drug design that will potentially maximize the p53 response and sensitize tumour cells to chemotherapeutic agents. References 1.Riley T, Sontag E, Chen P and Levine A. Transcriptional control of human p53-regulated genes. Nat Rev Mol Cell Biol 2008;9:402-412 2.Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID Bioinformatics Resources. Nature Protoc. 2009;4(1):44-57 3.Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 2009;37(1):1-13 4.Eilbeck et al. The Sequence Ontology: A tool for the unification of genome annotations. Genome Biology (2005) 6:R44