Tesi etd-04082025-230103
Link copiato negli appunti
Tipo di tesi
Dottorato
Autore
ROBOTTI, MARZIA
URN
etd-04082025-230103
Titolo
Molecular characterization of Ovarian Cancer, studying salivary microRNA expression profiles
Settore scientifico disciplinare
BIO/11
Corso di studi
Istituto di Scienze della Vita - PHD IN MEDICINA TRASLAZIONALE
Commissione
relatore Prof.ssa ANGELONI, DEBORA
Parole chiave
- Biomarker
- Breast Cancer
- Machine Learning
- microRNA
- Ovarian Cancer
- Saliva
- Screening.
Data inizio appello
18/07/2025;
Disponibilità
parziale
Riassunto analitico
Ovarian cancer (OC) is a highly lethal and heterogeneous malignancy, ranking as the fifth leading cause of cancer-related death among women in developed countries. While rare germ cell and stromal tumors primarily affect women under 40 years of age, over 90% of cases are epithelial in origin, classified into five major histotypes intricating the disease classification. Late-stage diagnosis significantly contributes to the high morbidity and mortality of OC, because of the absence of distinct and specific OC symptoms, together with and the lack of effective early diagnostic methods.
In this context, microRNAs (miRNAs) have emerged as a promising class of biomarkers (BMs) due to their ability to regulate gene expression. These small non-coding RNA molecules, ranging in size from 18 to 25 nucleotides, have been the focus of intensive research as potential diagnostic and monitoring tools for many diseases, including cancer.
Among the body fluids where miRNAs can be detected, saliva stands out as a promising diagnostic biofluid due to its ability to reflect systemic pathophysiological conditions. This makes it an attractive candidate source for the future of non-invasive, cost-effective screening procedures for systemic diseases, including cancer.
In this study, due to the molecular complexity of OC and the lack of diagnostic tools for early diagnosis, we aimed to identify a specific salivary miRNome signature able to discriminate OC patients from healthy subjects (HS) and breast cancer patients (BC).
A total of 100 saliva samples were collected from the three groups of patients, and a first comprehensive analysis of the miRNome expression profile was performed on six pools of saliva (two per each group). This preliminary phase was conducted with Microarray Affymetrix GeneChip 4.0.
The results, as proof of concept of the study, demonstrated that there were several miRNAs exhibiting differential expression in at least one group. Indeed, a pair-wise comparison of the three conditions revealed 25 miRNAs at different levels in OC patients compared to the HS with statistical significance. Subsequent analyses focused on five candidate miRNAs: four were found to be significantly elevated and one significantly lower in OC saliva samples. These were then subjected to a second validation phase involving qRT-PCR. This analysis was performed on individual samples from all participants to confirm the five miRNA profiles in saliva, and microarray results were significantly confirmed by qRT-PCR.
The following stage of the research was based on selecting one microRNA from the group of validated ones for detailed studies and functional analysis. The study was designed to investigate a miRNA showing significantly higher level in OC saliva samples compared to HS and BC, as determined by microarray and qRT-PCR results.
The function this small nc-RNA molecule is still completely unknown in biological systems, and there are no validated targets in miRTarbase with strong evidence. Thus, the investigation into its biological role started with the analysis of target genes in the available repositories.
MiRTarbase provided a list of 179 target genes that had been weakly validated through CLIP-SEQ analysis or NGS. Through bioinformatic tools, we predicted pathways in which our miRNA of interest may be involved and then which were, among miRNA’s targets, the hub-genes enriching the highest number of pathways and showing a strong degree of connectivity. By this workflow, ACVR2B resulted to be the most relevant mRNA for further analysis.
To identify the most suitable model for functional studies, the expression levels of the miRNA and its putative target were initially verified using qRT-PCR in three distinct cell lines (MCF-7, MDA-MB-231 and HCT-116). These cell lines were selected as models for gain-of-function and loss-of-function assays, as well as for luciferase assays.
Subsequently, to identify a specific salivary miRNA signature, a more sensitive high-throughput analysis, as the RNA-sequencing, was performed on each of the 100 saliva samples. A machine learning (ML) method was then established to enhance the accuracy and interpretation of data, focused on multiclass identification of OC, BC and HS. A predictive method was customized, based on a hybrid features selection and a classification pipeline designed to train the model for the classification task. A subset of miRNAs, that achieves high value of precision, recall and F1-score for each class, was obtained. These results, as a preliminary study, highlight the potential of salivary miRNA profiles as promising diagnostic BMs for OC.
At the same time, a further validation phase on an independent patient cohort is required to expand our understanding in this area and to confirm the best pattern of molecules for screening purposes.
In this context, microRNAs (miRNAs) have emerged as a promising class of biomarkers (BMs) due to their ability to regulate gene expression. These small non-coding RNA molecules, ranging in size from 18 to 25 nucleotides, have been the focus of intensive research as potential diagnostic and monitoring tools for many diseases, including cancer.
Among the body fluids where miRNAs can be detected, saliva stands out as a promising diagnostic biofluid due to its ability to reflect systemic pathophysiological conditions. This makes it an attractive candidate source for the future of non-invasive, cost-effective screening procedures for systemic diseases, including cancer.
In this study, due to the molecular complexity of OC and the lack of diagnostic tools for early diagnosis, we aimed to identify a specific salivary miRNome signature able to discriminate OC patients from healthy subjects (HS) and breast cancer patients (BC).
A total of 100 saliva samples were collected from the three groups of patients, and a first comprehensive analysis of the miRNome expression profile was performed on six pools of saliva (two per each group). This preliminary phase was conducted with Microarray Affymetrix GeneChip 4.0.
The results, as proof of concept of the study, demonstrated that there were several miRNAs exhibiting differential expression in at least one group. Indeed, a pair-wise comparison of the three conditions revealed 25 miRNAs at different levels in OC patients compared to the HS with statistical significance. Subsequent analyses focused on five candidate miRNAs: four were found to be significantly elevated and one significantly lower in OC saliva samples. These were then subjected to a second validation phase involving qRT-PCR. This analysis was performed on individual samples from all participants to confirm the five miRNA profiles in saliva, and microarray results were significantly confirmed by qRT-PCR.
The following stage of the research was based on selecting one microRNA from the group of validated ones for detailed studies and functional analysis. The study was designed to investigate a miRNA showing significantly higher level in OC saliva samples compared to HS and BC, as determined by microarray and qRT-PCR results.
The function this small nc-RNA molecule is still completely unknown in biological systems, and there are no validated targets in miRTarbase with strong evidence. Thus, the investigation into its biological role started with the analysis of target genes in the available repositories.
MiRTarbase provided a list of 179 target genes that had been weakly validated through CLIP-SEQ analysis or NGS. Through bioinformatic tools, we predicted pathways in which our miRNA of interest may be involved and then which were, among miRNA’s targets, the hub-genes enriching the highest number of pathways and showing a strong degree of connectivity. By this workflow, ACVR2B resulted to be the most relevant mRNA for further analysis.
To identify the most suitable model for functional studies, the expression levels of the miRNA and its putative target were initially verified using qRT-PCR in three distinct cell lines (MCF-7, MDA-MB-231 and HCT-116). These cell lines were selected as models for gain-of-function and loss-of-function assays, as well as for luciferase assays.
Subsequently, to identify a specific salivary miRNA signature, a more sensitive high-throughput analysis, as the RNA-sequencing, was performed on each of the 100 saliva samples. A machine learning (ML) method was then established to enhance the accuracy and interpretation of data, focused on multiclass identification of OC, BC and HS. A predictive method was customized, based on a hybrid features selection and a classification pipeline designed to train the model for the classification task. A subset of miRNAs, that achieves high value of precision, recall and F1-score for each class, was obtained. These results, as a preliminary study, highlight the potential of salivary miRNA profiles as promising diagnostic BMs for OC.
At the same time, a further validation phase on an independent patient cohort is required to expand our understanding in this area and to confirm the best pattern of molecules for screening purposes.
File
| Nome file | Dimensione |
|---|---|
Ci sono 1 file riservati su richiesta dell'autore. |
|