ETD system

Electronic theses and dissertations repository

 

Tesi etd-04042014-181248


Thesis type
Tesi di dottorato di ricerca
Author
FRANCESCHI, SARA
URN
etd-04042014-181248
Title
PREDICTION OF DISEASE OUTCOME IN GLIOBLASTOMA PATIENTS BY MOLECULAR ANALYSIS
Settore scientifico disciplinare
MED/06
Corso di studi
SCIENZE BIOLOGICHE E MOLECOLARI
Commissione
relatore Prof. Bevilacqua, Generoso
tutor Dott.ssa Zavaglia, Katia
Parole chiave
  • progression free survival
  • prognosis
  • overall survival
  • outcome
  • nanostring
  • molecular markers
  • molecular alterations
  • glioblastoma
  • ffpe
  • bevacizumab
  • recurrence
  • rna-seq
  • TCGA subtypes
Data inizio appello
11/04/2014;
Consultabilità
completa
Riassunto analitico
Malignant gliomas, especially glioblastoma (GBM), represent the most common primary malignant tumors of the adult central nervous system. Surgical resection, radiotherapy and adjuvant chemotherapy are currently the standard of care for GBM. However, despite progress in recent years, due to the infiltrative and dispersive nature of the tumor, recurrence rate remains high and typically results in very poor prognosis. From a practical perspective, controlling growth and dispersal of the recurrence may have a greater impact on disease-free survival. Our study had the intent to provide novel information on GBM tumor aggressive behavior by connecting multiple molecular markers, that could play an important role in tumor progression, with the outcome of the patients meant as the time of recurrence, time of survival after a specific therapy and overall survival. In this work we collected a total of 205 formalin fixed paraffin embedded (FFPE) first-surgery glioma samples from two different populations, Italian and American.

Italian glioma samples were molecularly characterized investigating 15 of the most studied candidate molecular markers involved both in glioma tumor progression and radiotherapy response : IDH1 and IDH2 mutations, SNPs (XRCC1, XRCC3, RAD51, GSTP1), EGFR and ERBB2 amplification, deletion of PTEN, TP53 and CDK2NA genes, presence of the active mutant form of EGFR (EGFRvIII), MGMT promoter methylation and 1p19q codeletion. To better understand the dynamics of genomic alterations associated with GBM recurrence in more detail we selected 19 GBM samples with specific recurrence free survival outcome; by using real-time PCR we profiled the expression of 84 genes important for cell-cell and cell-matrix interactions, the expression of 84 miRNAs known to alter their expression during nervous system-related carcinogenesis and a copy number variation analysis of 23 genes reported to undergo frequent genomic alterations in human glioma tumor DNA. In order to have even more selected samples with differences in time of first recurrence we chosen 6 GBM; on those samples we performed a whole transcriptome sequencing analyses (RNA-seq) by using Ion Proton System technology.

American GBM samples were analyzed using Nanostring technology. We used a customized gene panel to evaluate the expression levels of 151 genes involved in glioblastoma and classified our samples into one of the four TCGA molecular classes: Classical, Mesenchymal, Neural and Proneural.

We found strong correlation between expression levels of different genes related to cellular migration and invasion and patient outcome. Those genes had also high power to predict patient outcome on the basis of the combination of their expression levels. By using RNA-seq, we found several genes with statistically significant expression level differences in GBM patient with distinct outcomes, and we were able to identify specific molecular functions networks into those groups of tumors. Moreover we were able to identify specific gene fusion transcripts with high probability of a “driver” behavior in the oncogenic process; we also detected exclusively gene fusion transcripts in GBM patient with distinct outcomes.
Furthermore we found strong correlation between TCGA molecular classification and patient outcome and we selected characteristic genes with expression levels related both to the molecular TCGA subclass and the survival.

In conclusion our study shows how the integration of miRNA and mRNA espression and genomic alterations can define a GBM molecular profile related to patient outcome. Investigation of diagnostic and prognostic value of these molecular markers could be useful to allow better predictions of GBM patients prognosis and response to therapies.
File