Mutational profiling of genes associated with melanoma progression: a single-center study of primary tumours and matched metastases

Cutaneous melanoma is an aggressive cancer originating from normal melanocytes through the accumulation and selection of somatic mutations in driver genes. Four molecular subtypes BRAF, NRAS, NF1, and triple wild type melanomas have been defined based on driver mutations in these genes. The study of molecular heterogeneity between primary tumors and paired metastases is essential for understanding tumor evolution. Most published studies investigated primaries and metastases from different patients and in studies where primaries and metastases from the same patient were considered, the analysis was limited to BRAF and NRAS mutations. The primary aim of this study was to identify the molecular profiles of primary melanomas and matched metastases and estimate their clinical and prognostic significance. Moreover, we evaluated intra-patient and inter-patient heterogeneity of molecular profiles, that could be associated to mechanisms of progression. We included both primary melanomas and paired metastatic tissues from 44 patients, for a total of 44 primary tumors and 63 metastases. Sequencing analysis was performed through the IonTorrent methodology with a custom panel of 18 melanoma related genes. Overall, 43% of primary tumors exhibited BRAF driver mutations, 27% NRAS, 11% NF1, and 20% were triple wild type. Among metastases, the frequency of molecular subtypes was 30% with BRAF mutations, 14% NRAS, 14% NF1 and 43% triple wild type. The wild type status resulted significantly increased in metastatic tissues compared to primary tumors (p=0.035). BRAF 2 were associated with the onset of tumors at a young age (median 57.6, range 25-73, p=0.036), with the presence of ulceration (p=0.003), with higher number of mitosis (median 6 mitosis/mm2, range 0-23, p=0.041) and with a significantly increased risk to metastasize to the skin, independently of other prognostic factors (OR= 6.72, 95%CI 1.39-67.43, p= 0.028) compared to BRAF-wild type melanomas. Regarding NRAS, the presence of mutations was associated with later onset (median=72.5, range 45-84, p=0.004. NF1-mutated melanomas were associated with more advanced stages at diagnoses (p=0.050). The triple wild type subgroup was strongly associated with the superficial spreading histotype (p<0.001) and with melanoma diagnosed at earlier stages (stage I/II, p=0.024). A high frequency of mutations in TACC1, PREX2, EPHA3, and TERT promoter were found both in primary tumors and metastases. Mutations in PREX2 and RAC1 were significantly associated with higher Breslow thickness (p=0.002; p=0.006 respectively). RAC1 variants were also associated with the nodular (p=0.012) and ulcerated (p=0.062) histologic features. When we considered the four main molecular subgroups, intra-patient discordance was observed in 53.3% of patients. Regarding the other genes, we observed a significant discordance of molecular status in STK19 (p=0.007), ARID2 (p=0.021), PPP6C (p=0.035), TACC1 (p=0.046), GRIN2A (p=0.052). Using Multiple Correspondence Analysis (MCA), we were able to identify two distinct clusters of patients based on the mutational status of the 18 genes analyzed. One cluster was characterized by primary lesions with TP53, PPP6C, ARID2, GRIN2A, GNA11, WNT1, and DDX3X mutations, while patients with triple wild type primary melanomas clustered in the second group. Moreover, MCA analysis also demonstrated that mutations in PPP6C, GRIN2A, ARID2, and GNA11 tend to co-occur. In conclusion, our findings identified interesting associations of clinical and prognostic features with the mutational status of melanoma-related genes and discovered a high rate of molecular heterogeneity between primary and metastatic tumors in the same patients. We defined two distinct clusters of melanoma lesions and identified specific molecular profiles of these clusters. Further investigations on large number of cases are required to confirm our data and fully comprehend the significance of these genetic patterns.