Identification of distinct molecular signatures in BRAF V600E and BRAF wt cutaneous melanomas

The purpose of this study was to investigate the presence of a gene expression signature in BRAF V600E melanomas compared to wild type ones, all derived from sun-exposed sites. Microdissected tissues from excisional biopsies of 18 cutaneous melanomas were analyzed to detect the presence of BRAF and NRAS mutations and to profile whole genome expression by means of oligonucleotide microarrays. Class comparison methods were used to select differentially expressed genes between wild type and mutated lesions. Real Time RT-PCR and immunohistochemistry were applied to validate differences at the mRNA and protein levels on an independent cohort of samples. BRAF mutations were evidenced in 67% of melanomas. All of them consisted of the oncogenic change V600E (and the mutation event was independent of Clark's level). Data indicate that in V600E melanomas there is an over-expression of cancer stem cell markers and an upregulation of important oncogenes, like KRAS. This, together with the downregulation of genes involved in oxidative UV stress response and immuno system defense, confers an advantage to V600E melanomas compared to wt lesions. Moreover, the downregulation of topoisomerase I and CDKN2A results in an increased replicative potential associated with a decrease in senescence markers and a diminished DNA damage response. As far as the wild-type lesions, we interestingly pointed out the overexpression of PML, PIK3CA and the downregulation of two tumour suppressor genes (BRCA1 and TP73) relevant to DNA repair. A total of 18 excisional biopsies from primary radial growth phase malignant melanomas (n = 5) and primary vertical growth phase malignant melanomas (n = 13) were examined. Fresh melanoma specimens were examined after removal by a pathologist and a representative portion of the lesion was immediately put in RNA later (RNA Stabilization Reagent – Ambion Inc, Austin TX) for further molecular processing. Lesions were classified according to TNM. All tumor samples and clinical data were collected according to the IRCC ethical commitee’ s approval. Informed consent was also obtained for each patient. All the samples were homogenized (Ultra – TurraxT8, Ika – Werke, Staufen, Germany) with 600 ul of RNA lysis buffer plus 1% 2-mercaptoethanol, and total RNA was isolated by using RNeasy Mini Kit (Qiagen, Dusseldorf, Germany). TotRNA quality was checked by means of RNA 6000 pico chip assays (Agilent Technologies, Palo Alto, CA) run on the Agilent 2100 bioanalyzer. Paraffin-embedded tissues derived from each lesion were subjected to laser microdissection by means of optical microscope or laser microdissector (Leica AS-LMS, Leica Microsystems, Wetzlar, Germany) and DNA was extracted from microdissected cells by using QIAamp Kit (Qiagen) as described by Venesio et al. (2008). BRAF exon 11 and 15 and NRAS exons 1 and 2 were amplified by PCR using primers and conditions as previously reported by Venesio et al. (2008). PCR products were purified by ExoSAP-IT (USB Corporation, Cleveland, Ohio) and sequenced using Big Dye Terminator V3.1 Cycle Sequencing Kit and Prism Model 3730 DNA Analyzer (Applied Biosystems, Foster City, CA). Sequence analysis was based on BRAF and NRAS sequences (GenBank accession nos. NM_00433 and NM_002524 for BRAF and NRAS, respectively). RNA from each sample and from the human universal reference (BD™ Human Universal Reference Total RNA, Clontech, Palo Alto, CA) was amplified by means of the Amino Allyl MessageAmp I aRNA Kit (Ambion) to obtain amino allyl antisense RNA (aaRNA) following the method developed by Eberwine and coworkers (1990). Two rounds of amplification were carried out to obtain the necessary quantity of aaRNA for labeling. Briefly: mRNA was reverse transcribed into cDNA single strand; after the second strand synthesis (in the second round of amplification), cDNA was in vitro transcribed in aaRNA including amino allyl modified nucleotides (aaUTP). Both dsDNA and aaRNA underwent a purification step using columns provided with the kit. Labeling was performed using NHS ester Cy3 or Cy5 dyes (Amersham Biosciences, Buckinghamshire UK) which are able to react with the modified RNA. mRNA quality was checked by means of RNA 6000 nano chip assays (Agilent Technologies). At least 5 ug of mRNA for each sample were labeled and purified with columns. Equal amounts (0.75 ug) of labeled specimens from sample and reference were put together, fragmented and hybridized to oligonucleotide glass arrays representing 41K human unique genes and transcripts (Human Whole Genome Oligo Microarray Array, Agilent Technologies). All steps were performed using the In Situ Hybridization kit-plus (Agilent Technologies) and following the 60-mer oligo microarray processing protocol (Agilent Technologies). Then, slides were washed with the SSPE wash procedure and scanned with the dual-laser microarray scanner Agilent G2505B. For each sample, a dye-swap replicate was performed. Images were analyzed using Feature Extraction software (Agilent Technologies) version 7.6. Output files containing feature and background intensities and the related statistical parameters for red and green signals were then loaded into the Resolver SE System (Rosetta Biosoftware, Seattle, WA) together with the scan images and the Agilent Human Whole Genome pattern file. Data processing and normalization were performed as previously described (Scatolini et al. 2010). To compare class expression patterns between BRAF mutated samples and wild-type ones t-test and empirical Bayes statistics (LIMMA package) for two class comparison were applied. Transcripts with significant modulation, B value greater than 2 for empirical Bayes statistic, were further considered. A B value greater than 2 means that the gene has more than 88% probability of being differentially expressed by assuming that 1% of the genes are expected to be differentially expressed. Moreover, we applied to the ratio experiment a one-way ANOVA available within the Resolver System with Multiple Test Correction method (Benjamini Hochberg FDR). Three classes were considered for the ANOVA (wt, ++, +++). For all the transcripts with significant modulation (p-value<0.01), we looked for those with increasing or decreasing expression and, in particular, with an absolute log ratio difference >0.7 between the wt class and the +++ one. In order to underline the enrichment of Biological Processes-level 5 (BP5) of the Gene Ontology (GO), we used the functional annotation tool available within DAVID website, using the Entrez Gene ID of modulated transcripts (p-value< 0.01). Several databases were interrogated to find out and compare functional information on single genes and on groups of differentially expressed genes. In particular, NCBI database (http://www.ncbi.nlm.nih.gov/) and UCSC Genome Bioinformatics website were used (http://genome.ucsc.edu/). Quantitative RT polymerase chain reaction (qPCR) was performed, as previously described (Scatolini M et al. 2010), to validate a subset of differently expressed transcripts between BRAF mutated and wt samples, identified by microarray analysis.