Identification of diagnostic, prognostic and susceptibility biomarkers in cutaneous malignant melanoma

Resumen

Melanoma is a malignant tumor that originates from the transformation of melanocytes,which are the cells responsible for melanin synthesis. Patients with early-stage melanoma (lessthan 1 mm thick) and no lymph node involvement have a high likelihood of cure after completeexcision of the tumor by surgery. However, there is no effective treatment for metastaticmelanoma, which makes it a potentially lethal form of cancer. For this reason, despite thedecades-long study, the treatment of disseminated melanoma is one of the biggest problems inclinical oncology, and the steady increase in the incidence of this neoplasm worldwide an issueof serious concern.The aim of our work is to identify a set of markers of susceptibility or that can be used inthe diagnosis and/or prognosis of malignant melanoma. To this end, our strategy work beginswith the identification of proteins differentially expressed in melanoma cell lines and primarycultures of melanocytes by two-dimensional electrophoresis and mass spectrometry. Aftervalidation of protein levels by Western blot, we analyzed the transcriptional levels ofcorresponding genes by RT-qPCR and looked for genetic and epigenetic alterations.As from the differentially expressed proteins in combination with the informationobtained from a previously arrays study done in our laboratory, we selected and genotyped384 SNPs distributed in 85 different genes. This second part of the project involved the designof a 3- phase case-control study to determine whether any of the selected SNPs is associatedwith the risk of developing malignant melanoma.Our results confirmed the differential expression of 10 proteins in at least 5 of the 10melanoma cell lines tested, and show that the proteins PIRIN and PYCARD could be proteinmarkers of melanoma because all tested tumor cell lines have a reduction to almost undetectableamounts of both proteins compared to primary cultures of melanocytes. Likewise, we found asignificant correlation between the amount of protein and transcriptional levels of their encodinggenes ANXA5, PIR, PYCARD and RUVBL2, suggesting a deregulation at the transcriptionallevel. However, it seems that the deregulation is at post-translational level in the case of theproteins GDI2, HSP27, MAD1, PEA-15, RCC2 and Peroxiredoxin-2, because mRNA levels donot reflect the variation in the amount of these proteins.Then, we combined the results obtained by STRING v9.0 program and the informationgathered after a literature search to build interaction networks between proteins. These networkswere used to select new genes whose transcriptional activity was analyzed by RT-qPCR. Theresults showed a repression in the transcriptional activity of genes PIR, PYCARD, PEBP1 andMYC, and over-expression of NFKB2, BCL3, IL1B and SNAI1 in all melanoma lines tested,suggesting that these genes may constitute markers for melanoma diagnosis. Furthermore,melanoma cell lines with invasive capacity exhibit changes in gene expression patterns whichcan also be used as prognostic markers. Thus, we have found that melanoma cells with invasivecapacity show a reduction in the transcription of the genes PIR, M-MITF and BCL2 while overexpressAR and HGF.Furthermore, we analyzed whether methylation epigenetic factor could be either adiagnostic or prognostic marker in melanoma. After treatment ADN with sodium bisulfite, wefound no methylation in the promoter region of the genes PEBP1 and PIR in none of the cellculture tested. However, methylation does seem to cause transcriptional repression PYCARDgene in melanoma lines, and this epigenetic marker may be used to differentiate between neviand melanoma.We have also analyzed the presence of alteration in the coding nucleotide sequence of thegenes of interest, and the results show that the genes ANXA5, GDI2, HSBP1, ILKAP, MAD1L1,PEA15, PIR, PRDX2, PYCARD, RCC2 and RUVBL2 exhibited no mutations in their codingnucleotide sequence. However, we identified the missense SNP rs1050625 located in PEBP1gene. The derived allele of this SNP originates the mutation p.His145Tyr which seems to have adeleterious effect in PEBP-1 protein in silico predictions. In addition, according to informationfrom the 1000 Genomes database it is not described any TT homozygote among 1,093individuals genotyped worldwide. For this reason, we suggest that the homozygous genotypeTT in the SNP rs1050625 may be an exclusive condition of tumor lines not present in thegenome of the individual.Finally, to identify markers of susceptibility to melanoma, we designed an assay of SNPsassociation by Illumina and Kaspar genotyping technologies. Our results showed that the SNPrs6854854 located in ANXA5 provides protection against melanoma in the population of theBasque Country (464 patients with melanoma and 400 controls), with an OR of 0.541 (95% CI0.371-0.791 and p-value = 0.0013). This protection could be subject to environmentalconditions as the SNP has a population structure. We also identified the SNP rs6431588 locatedin ILKAP as a marker of risk associated with predisposition to develop malignant melanoma,with an OR of 1.29 (95% CI 1.12-1.48 and p-value = 0.0004) after genotyping 1,883 patientswith melanoma and 1,358 healthy individuals. The SNP rs6431588 is also under a populationstructure and the value of Tajima¿s D obtained for European population move away from thezero in a positive fashion, which suggests the existence of a phenomenon of balancing selection