Use this URL to cite or link to this record in EThOS:
Title: Dissecting melanoma heterogeneity by integrative genomic analysis for tailored anti-cancer therapy
Author: Dugo, Matteo
ISNI:       0000 0004 7661 2564
Awarding Body: Open University
Current Institution: Open University
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
Cutaneous melanoma is a highly aggressive disease resistant to conventional treatment and characterized by poor prognosis. Targeted therapies against MAPK pathway and immune checkpoint inhibitors have dramatically improved survival of metastatic melanoma patients but the extent and duration of response are variable. Classification based on gene expression profiling have so far allowed identification of melanoma subtypes with distinctive biological features and with potential clinical impact. However, clinical translation of molecular subtypes is hampered by inconsistencies among the different classifications. Here, through a harmonized bioinformatic analysis of public transcriptomic datasets, we compared and combined nine published classification systems to derive the consensus transcriptional subtypes of melanoma. Beyond confirming previously reported subtypes, our approach enabled the identification of a novel highly mitotic, chromosomally unstable group of melanomas that recapitulated a transitory state from a proliferative, melanocytic, differentiated phenotype to a more mesenchimal invasive program. We provided evidence that this classification has a prognostic role in metastatic melanoma patients, independently from the levels of tumour immune infiltration. We translated consensus subtypes to in vitro melanoma cell lines and combining them with pharmacological data we highlighted subtype-specific sensitivity to MAPK inhibitors and other drugs. Analysis of baseline gene expression data of metastatic melanoma patients treated with MAPK or immune checkpoint inhibitors showed that the predictive role of consensus subtypes in clinical setting remains to be elucidated. Finally, through the analysis of multi-omics data from the same set of patients, we comprehensively characterized the consensus subtypes at the genomic, transcriptional, and epigenomic levels. Our results showed that melanoma gene expression classifications converged on five biological entities determined by transcriptional and epigenetic events, and with potential implications for prognostication.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral