Most cancer deaths are caused by the ability of cancer cells to metastasize, i.e. colonise organs distant to the site from the primary tumour. Although much remains to be learned, the epithelial to mesenchymal transition (EMT)-inducing transcription factor ZEB1 is considered a key regulator of the metastatic cascade and it is thus of great importance to dissect its mechanisms of action. A genome-wide analysis of ZEB1 binding sites performed by Brabletz’s group in MDA-MB-231, an aggressive breast cancer cell line, revealed that less than half of ZEB1 sites are located in promoter region, while the rest are in distant and thus potential enhancer regions. The present thesis aims to further characterize the mechanism of action of ZEB1 by the generation of a genome-wide map of ZEB1-dependent changes in chromatin status. To this aim, histone ChIP-seq of five different chromatin marks was performed in MDA-MB-231, both upon ZEB1 knockdown and in control conditions. Relying on the ChromHMM software, based on a multivariate hidden Markov model six chromatin states associated with the histone marks were inferred and then annotated on the basis of the learned emission probabilities for the histone marks as well as available literature. The model allowed to systematically characterize both active and inactive chromatin states and highlighted promoter regions as well as putative enhancers. Model validation on well-characterised ZEB1 target genes and integration of our results with the previous analysis by Brabletz’s group of ZEB1 binding to activated and repressed genes showed that the model was able to capture coherent changes in chromatin status of promoter regions. Furthermore, the same model confirmed changes in chromatin status also in one experimentally validated ZEB1 enhancer region. Thus, we feel that the model can be reliably used also to identify distant regulatory regions on which ZEB1 act. Taken together, our results provide the basis to dissect ZEB1 transcriptional control in enhancer regions and can contribute to the development of novel prognostic and diagnostic tools in metastatic cancer.
Analisi bioinformatiche integrate di dati ChIP-seq in una linea cellulare metastatica di cancro al seno. La maggior parte delle morti per cancro sono causate dalla capacità delle cellule tumorali di metastatizzare, cioè di colonizzare organi distanti dalla sede del tumore primario. Sebbene ci sia ancora molto da capire, il fattore di trascrizione ZEB1, che porta alla transizione delle cellule da epiteliali a mesenchimali (EMT), è considerato un regolatore chiave della cascata metastatica ed è quindi di grande importanza esaminarne i suoi meccanismi d'azione. Un'analisi a livello dell’intero genoma dei siti di legame di ZEB1, eseguita dal gruppo del Prof. Brabletz in MDA-MB-231, una linea aggressiva di cellule tumorali al seno, ha rivelato che meno della metà dei siti di ZEB1 si trovano nella regione del promotore, mentre il resto si trova in potenziali regioni enhancer. La presente tesi mira a caratterizzare ulteriormente il meccanismo d’azione di ZEB1 attraverso la generazione di una mappa genomica dei cambiamenti di stato della cromatina ZEB1-dipendenti. A questo scopo, histone ChIP-seq per cinque differenti marker cromatinici sono state eseguite in MDA-MB-231, sia con ZEB1 knockdown che in condizioni di controllo. Utilizzando il software ChromHMM, basato su un hidden Markov model (HMM) multivariato, sono stati successivamente dedotti sei stati di cromatina associati ai marker istonici e sono stati quindi annotati sulla base delle relative probabilità di emissione e della letteratura disponibile. Il modello ha permesso di caratterizzare sistematicamente stati di cromatina attiva e inattiva e ha evidenziato le regioni promotrici e i potenziali enhancer. La validazione del modello su ben caratterizzati geni bersaglio di ZEB1 e l'integrazione dei risultati con le precedenti analisi del gruppo Brabletz sul legame di ZEB1 ai geni attivati e repressi hanno dimostrato che il modello è in grado di catturare, per le regioni promotrici, cambiamenti coerenti nello stato della cromatina. Inoltre, lo stesso modello ha confermato cambiamenti in una regione enhancer di ZEB1 validata sperimentalmente. Pertanto, riteniamo che il modello possa essere utilizzato in modo affidabile anche per identificare le regioni regolatorie di tipo enhancer sulle quali ZEB1 agisce. Nel loro insieme, i nostri risultati forniscono la base per analizzare il controllo trascrizionale di ZEB1 nelle regioni enhancer e possono contribuire allo sviluppo di nuovi strumenti prognostici e diagnostici nel cancro metastatico.
Integrative bioinformatics analyses of ChIP-seq data in a metastatic breast cancer cell line
GRAZIANO, GIULIA
2018/2019
Abstract
Most cancer deaths are caused by the ability of cancer cells to metastasize, i.e. colonise organs distant to the site from the primary tumour. Although much remains to be learned, the epithelial to mesenchymal transition (EMT)-inducing transcription factor ZEB1 is considered a key regulator of the metastatic cascade and it is thus of great importance to dissect its mechanisms of action. A genome-wide analysis of ZEB1 binding sites performed by Brabletz’s group in MDA-MB-231, an aggressive breast cancer cell line, revealed that less than half of ZEB1 sites are located in promoter region, while the rest are in distant and thus potential enhancer regions. The present thesis aims to further characterize the mechanism of action of ZEB1 by the generation of a genome-wide map of ZEB1-dependent changes in chromatin status. To this aim, histone ChIP-seq of five different chromatin marks was performed in MDA-MB-231, both upon ZEB1 knockdown and in control conditions. Relying on the ChromHMM software, based on a multivariate hidden Markov model six chromatin states associated with the histone marks were inferred and then annotated on the basis of the learned emission probabilities for the histone marks as well as available literature. The model allowed to systematically characterize both active and inactive chromatin states and highlighted promoter regions as well as putative enhancers. Model validation on well-characterised ZEB1 target genes and integration of our results with the previous analysis by Brabletz’s group of ZEB1 binding to activated and repressed genes showed that the model was able to capture coherent changes in chromatin status of promoter regions. Furthermore, the same model confirmed changes in chromatin status also in one experimentally validated ZEB1 enhancer region. Thus, we feel that the model can be reliably used also to identify distant regulatory regions on which ZEB1 act. Taken together, our results provide the basis to dissect ZEB1 transcriptional control in enhancer regions and can contribute to the development of novel prognostic and diagnostic tools in metastatic cancer.È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: unitesi@unipv.it.
https://hdl.handle.net/20.500.14239/20955