Nowadays, a widespread cancer disease is leukaemia. This term refers to a group of diseases which involved the uncontrolled proliferation of a hematopoietic stem cell. Leukemias are divided into acute and chronic, which in turn can be subdivided as lymphoids or myeloids depending on the type of blood cell involved. Four main subtypes of leukaemia are known: acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL) and chronic myeloid leukemia (CML). Chronic lymphocytic leukemia is a lymphoproliferative disorder characterized by the progressive accumulation of mature B lymphocytes in the peripheral blood (PB), bone marrow and lymphoid organs. With an incidence of 4.7 new cases per 100,000 person per year, it is the most common leukemia among adults in Western Countries. The incidence increases by age: currently the median age at diagnosis is 72 years. The clinical course of CLL is extremely heterogeneous. Some patients never require treatment and have a high overall survival rate. On the other hand, other patients show an aggressive disease that requires timely treatment, which with high probability is followed by frequent relapses. For this reason, the survival range varies from a few months to more than a decade. Chronic lymphocytic leukemia does not require treatment until the disease progresses. The therapy can include chemotherapy, chemoimmunotherapy or drugs able to target and stop the growth and survival of CLL cells. Thanks to these numerous alternatives, the recommended therapy varies according to the patient’s physiological state. As for the molecular profile, although there is no specific mutation associated with this disease, genetic alterations can be identified in more than 80% CLL cases. The patients can be divided into two subgroups based on the mutational status of the heavy chain variable region of the immunoglobulin genes (IGHV). These encode for a part of the B cell receptor (BCR). Several analyses show that about 50% of these genes have somatic mutations. The IGHV mutational state (mutated vs unmutated) has a prognostic value as patients without mutation have a more aggressive clinical course. Moreover, the most frequent somatic mutations were found to be involved in critical cellular pathways including: DNA damage and cell cycle control (TP53, ATM, BIRC3), mRNA processing (SF3B1), NOTCH signalling (NOTCH1), and inflammatory pathways (MYD88). In addition to somatic mutations, analysis of CLL gene expression profiles has revealed that hundreds of genes are up- or down-regulated. This could be due to epigenetic events such as hyper- or hypo-methylation of the transcription promoter region of these genes, rather than genetic alterations. On top of this, CLL cell interactions with the supportive tissue microenvironment play a critical role in disease pathogenesis. The microenvironment is the set of accessory cells that, within individual organs, support parenchymal cells through complex communication mechanisms. CLL cells recirculate between peripheral blood and secondary lymphoid organs, where they proliferate rapidly. The stromal cells in these tissues attract and retain CLL cells in cooperation with adhesion molecules. Therefore, given the many factors involved, the division of patients into risk classes is very complex and there are currently no staging systems capable of accurately predicting the evolution of the disease. The following project aims to analyze the genetic differences between two classes of CLL patients, A and B; this division was carried out on the basis of the phenotypic characteristics developed in a murine model of chronic lymphocytic leukemia (Eμ-TCL1). Subtype A is considered aggressive. The study takes into consideration expression data from five public dataset consisting of a large number of samples. The samples were assigned to the subtype that best correlated with their expression profile based on a Fisher test. Samples that were not significant for both groups were considered "unclassified". Subsequently, differential expression analyses and enrichment analyses were performed with the GSEA software to highlight genes and cellular processes with a different behaviour between A and B. The results obtained were compared with those of the mouse model to identify which characteristics are preserved inter-species. Then, through the use of the ARACNe algorithm, a network-based approach was applied to identify the relationships between genes. The objective was to identify the regulatory genes, i.e. those that interact with many other genes and are able to influence their behaviour. In particular, this method was used to study the interactions of the most deregulated genes in the murine model within human dataset, to analyze the behavior of genes belonging to the interferon α pathway and to investigate the role of the extracellular matrix in the two classes A and B. The gene network thus constructed were then visualized and analyzed in Cytoscape. Moreover, since it is known that the onset and progression of a tumour is driven by the aberrant activity of oncoproteins, the activity of some regulatory proteins involved in key processes for cell development within human dataset has been studied. This has been carried out through the analysis of regulons, i.e. the levels of expression of a set of genes believed to be directly regulated by a protein; the VIPER algorithm has made it possible to identify regulators whose activity is deregulated between the two classes. In regards with the classification of samples into the two subtypes, the results showed that more than 50% of the samples belonging to the CD19-positive B cell dataset are attributed to one of the two classes. In contrast, this percentage is much lower for dataset with samples consisting of peripheral blood cells. In addition, dataset of the first type show more deregulated genes between the two categories in common with mice than the others. These differences are due to the different levels of gene expression of the two cell types; the reason why dataset consisting of purified B lymphocytes show better results is related to the fact that the murine phenotypes on which the classification was based were derived from the latter cell type. The GSEA enrichment analysis identified several enriched terms within the dataset, out of which five were among the most significant of the mouse model: "HALLMARK INTERFERON GAMMA RESPONSE", "HALLMARK INTERFERON ALPHA RESPONSE", "MODULE 47", "HALLMARK EPITHELIAL MESENCHYMAL TRANSITION" and "HALLMARK MYOGENESIS". Of these, particular attention has been given to the term "HALLMARK INTERFERON ALPHA RESPONSE", related to the interferon α pathway; in the literature it has been reported that an aberrant activation of the signalling of STAT proteins, usually involved in normal cellular processes, gives rise to various pathological events, including oncogenesis, especially in haematological malignancies. The network-based approach used has highlighted how the genes involved in this pathway are highly connected within our dataset, especially JAK1. This gene encodes for a membrane protein that phosphorylates the STAT proteins, which are transcription activators. Moreover, they are up-regulated in B; consequently, variations in the activity of these genes can lead to a different pathology course. Given the important role that the tissue microenvironment plays in the development of CLL, another gene set considered was that of “Module 47”, which contains extracellular matrix (ECM) and collagen genes. These proved to be highly connected and up-regulated in subtype A of leukemia; the high production of these molecules can stimulate the migration of CLL cells within secondary lymphoid organs, where they proliferate, thus leading to the progression of the disease. This result has also been confirmed by the analysis of protein activity: the protein regulator COL1A1 gene is significantly enriched in class A. Furthermore, a network-based approach applied to the study of the most deregulated gene interactions between A and B in the mouse model highlighted the EPHB4 gene. This gene is up-regulated in the aggressive leukemia subtype and encodes for a protein involved in the development of new vessels. This discovery confirms the importance of the microenvironment in this pathology; in fact, the presence of pro-angiogenic factors can increase the migration of cells from the blood to the tissues, where they reproduce. Finally, the analysis of the deregulation of protein activity between the two subtypes of leukemia has highlighted several regulators with different activity. In particular, DPYSL3 was found to be activated in subtype A: this protein was detected in other tumour types and its over-expression was associated with the development and spread of tumour metastases. As a consequence, its activation in the aggressive leukemia subtype is consistent with what reported in the literature. In addition, the deregulation of the ATM protein, which plays a key role in the control of the cell cycle, emerges. The increased activity of the ATM protein in subtype B of leukemia could indicate a greater ability to control the cell cycle in malignant cells, which therefore lead to apoptosis more easily. Hence the lower aggressiveness of this phenotype. Very interesting is the high activity of the DLC1 protein in subtype A of leukemia. This result does not agree with the oncosuppressant role generally attributed to the protein and, consequently, it is assumed that it is mediated by other factors. In conclusion, the high heterogeneity between the two classes of CLL A and B, both in terms of gene expression and protein activity, was also found in the identification of different therapeutic agents that should be able to treat the two pathological phenotypes. Despite the inter-species diversity, this work demonstrates the possibility of transferring a mouse model of CLL to humans. It also highlights the genetic differences between two phenotypes of leukemia, one of which is very aggressive. In particular, it identifies a dysregulation between the genes and the proteins responsible for crucial biological functions for the development of the disease. This shows that at the basis of the heterogeneity of the clinical course there are not only genetic mutations, but many complex gene regulation factors involved; among all, the role of the microenvironment assumes considerable importance.
Al giorno d’oggi una patologia oncologica molto diffusa è la leucemia. Con questo termine ci si riferisce a un gruppo di malattie legate alla proliferazione incontrollata di una cellula staminale emopoietica. Le leucemie si suddividono in acute e croniche, che a loro volta possono essere classificate come linfoidi o mieloidi in base al tipo di cellula del sangue coinvolta. Si conoscono quattro sottotipi principali di leucemia: la leucemia linfoblastica acuta (ALL), la leucemia mieloide acuta (AML), la leucemia linfocitica cronica (CLL) e la leucemia mieloide cronica (CML). La leucemia linfocitica (o linfatica) cronica è un disordine linfoproliferativo caratterizzato dall’accumulo di linfociti B maturi nel sangue periferico, nel midollo osseo e negli organi linfoidi periferici. Con un’incidenza di 4.7 nuovi casi su 100,000 individui ogni anno, la CLL è la leucemia più comune negli adulti nei paesi occidentali. La sua incidenza aumenta con l’aumentare dell’età: attualmente l’età media al momento della diagnosi è di 72 anni. Il decorso clinico di questa patologia è molto eterogeneo; alcuni pazienti non necessitano di particolari cure, soprattutto durante i primi anni, e hanno una probabilità di sopravvivenza elevata, mentre altri manifestano una malattia aggressiva che richiede trattamenti tempestivi che con alte probabilità sono comunque seguiti da frequenti ricadute. Per questo motivo il range di sopravvivenza varia da pochi mesi a più di un decennio. L’inizio della terapia è necessario nel momento in cui la malattia progredisce ed essa può includere chemioterapia, una combinazione di chemioterapia e immunoterapia oppure farmaci in grado di colpire e arrestare la crescita e/o la sopravvivenza delle cellule CLL. Grazie a queste numerose alternative, quindi, la terapia consigliata varia a seconda dello stato fisiologico del paziente. Per quanto riguarda le caratteristiche molecolari, alterazioni genetiche possono essere identificate in più dell’80% dei pazienti. Tuttavia, esse sono molto eterogenee e non è possibile identificare una mutazione specifica associata a questa patologia. I pazienti possono essere suddivisi in due sottogruppi in base alla presenza o all’assenza di mutazioni nella catena pesante delle immunoglobuline di superficie (IGHV), che codificano per una parte del recettore cellule B (BCR). Diverse analisi mostrano che circa nel 50% dei casi questi geni presentano delle mutazioni somatiche. Lo stato mutazionale IGHV (mutato vs non mutato) ha un valore prognostico in quanto i pazienti senza mutazione presentano un decorso clinico più aggressivo. Inoltre, si è visto che le alterazioni più frequenti sono coinvolte in importanti pathway biologici come nel controllo del ciclo cellulare e dei danni al DNA (TP53, ATM, BIRC3), nel processamento dell’mRNA (SF3B1), nel signalling pathway di NOTCH1 (NOTCH1) e nei pathways infiammatori (MYD88). In aggiunta alle mutazioni somatiche, analisi dell’espressione genica delle cellule CLL hanno rivelato che centinaia di geni risultano up- o down-regolati. Questo potrebbe essere dovuto ad eventi epigenetici come la iper- o ipo-metilazione della regione promotrice della trascrizione di questi geni, piuttosto che ad alterazioni genetiche. Infine, si ritiene che il microambiente tissutale abbia un ruolo fondamentale nello sviluppo di questa patologia. Esso è l’insieme delle cellule accessorie che, all’interno dei singoli organi, supportano le cellule parenchimali attraverso complessi meccanismi di comunicazione. Le cellule patologiche ricircolano tra il sangue periferico e gli organi linfoidi secondari, dove proliferano rapidamente. Le cellule stromali presenti in questi tessuti attirano e trattengono le cellule CLL attraverso la secrezione di chemochine e l’intervento di molecole di adesione. Visti i numerosi fattori in gioco, la suddivisione dei pazienti in classi di rischio risulta molto complessa e ad oggi non esistono sistemi di staging in grado di predire in modo accurato l’evoluzione della malattia. Il seguente progetto ha lo scopo di individuare e analizzare le differenze genetiche tra due classi di leucemia linfocitica cronica, A e B, le quali sono state determinate sulla base delle caratteristiche fenotipiche sviluppate in un modello murino di CLL (Eμ-TCL1). Il sottotipo A è considerato aggressivo. Lo studio prende in considerazione i dati di espressione di cinque dataset pubblici costituiti da un elevato numero di campioni. Questi ultimi sono stati assegnati al sottotipo che meglio correlava con il loro profilo di espressione sulla base di un test di Fisher. I campioni che sono risultati non significativi per entrambi i gruppi sono stati considerati “non classificati”. Successivamente, sono state svolte analisi di espressione differenziale e di arricchimento con il software GSEA per evidenziare i geni e i processi cellulari con un comportamento diverso tra A e B. I risultati così ottenuti sono stati confrontati con quelli del modello murino per individuare le caratteristiche che si conservano inter-specie. In seguito, attraverso l’utilizzo dell’algoritmo ARACNe è stato applicato un approccio network-based per individuare le relazioni che intercorrono tra i geni. L’obiettivo era quello di identificare i geni regolatori, ovvero quelli che interagiscono con molti altri geni e che sono in grado di influenzarne il comportamento. In particolare, questo metodo è stato usato per studiare le interazioni dei geni maggiormente deregolati nel topo all’interno dei dataset umani, per analizzare il comportamento dei geni appartenenti al pathway dell’interferone α e per indagare il ruolo della matrice extracellulare nelle due classi A e B. Le reti geniche così costruite sono state poi visualizzate e analizzate in Cytoscape. Inoltre, siccome è noto che l’insorgenza e la progressione di un tumore sono guidati dall’attività aberrante delle oncoproteine, si è studiata l’attività di alcune proteine regolatrici coinvolte in processi chiave per lo sviluppo cellulare all’interno dei dataset umani. Questo è stato possibile attraverso l’analisi del regulon, ovvero dei livelli di espressione di un insieme di geni che si ritiene siano direttamente regolati da una proteina; l’algoritmo VIPER ha permesso di individuare i regolatori la cui attività risulta deregolata tra le due classi. Per quanto riguarda la classificazione dei campioni nei due sottotipi, i risultati hanno mostrato che più del 50% dei campioni appartenenti ai dataset di cellule B CD19-positive vengono attribuiti ad una delle due classi. Invece, questa percentuale è molto più bassa per i dataset con campioni costituiti da cellule del sangue periferico. Inoltre, i dataset del primo tipo mostrano un maggior numero di geni deregolati tra le due categorie in comune con i topi rispetto agli altri. Queste differenze sono evidentemente dovute alla diversità intrinseca tra i livelli di espressione genica dei due tipi cellulari; il motivo per cui i dataset costituiti da linfociti B purificati presentano risultati migliori è legato al fatto che i fenotipi murini su cui si è basata la classificazione sono stati ricavati proprio da quest’ultima tipologia cellulare. L’analisi di arricchimento GSEA ha individuato numerosi termini arricchiti all’interno dei dataset, tra cui cinque sono risultati anche tra i più significativi del modello murino: “HALLMARK INTERFERON GAMMA RESPONSE”, “HALLMARK INTERFERON ALPHA RESPONSE”, “MODULE 47”, “HALLMARK EPITHELIAL MESENCHYMAL TRANSITION” E “HALLMARK MYOGENESIS”. Di questi, particolare attenzione è stata riservata al termine “HALLMARK INTERFERON ALPHA RESPONSE”, relativo al pathway dell’interferone α; in letteratura è stato infatti riportato che un’attivazione aberrante del signalling delle proteine STAT, solitamente coinvolte in normali processi cellulari, dà luogo a vari eventi patologici, tra cui l’oncogenesi, specialmente nelle malignità ematologiche. L’approccio network-based utilizzato ha evidenziato come i geni coinvolti in questo pathway siano altamente connessi all’interno dei nostri dataset, specialmente JAK1. Quest’ultimo codifica per una proteina di membrana che ha il compito di fosforilare le proteine STAT, le quali sono attivatori della trascrizione. Inoltre, essi sono risultati upregolati in B; di conseguenza, variazioni nell’attività di questi geni possono portare ad un diverso decorso della patologia. Dato l’importante ruolo che il microambiente tissutale gioca nello sviluppo della CLL, un altro gene set preso in considerazione è stato quello del Modulo 47, che contiene geni relativi alla matrice extracellulare (ECM) e al collagene. Questi si sono rivelati altamente connessi e upregolati nel sottotipo A di leucemia; l’elevata produzione di queste molecole può stimolare la migrazione delle cellule CLL all’interno degli organi linfoidi secondari, dove tendono a proliferare, portando così ad una progressione della patologia. Questo risultato è stato confermato anche dall’analisi dell’attività proteica: la proteina regolatrice COL1A1 risulta arricchita significativamente nella classe A. Inoltre, un approccio network-based applicato allo studio delle interazioni dei geni maggiormente deregolati tra A e B nel modello murino ha evidenziato il gene EPHB4. Quest’ultimo è upregolato nel sottotipo aggressivo di leucemia e codifica per una proteina coinvolta nello sviluppo di nuovi vasi. Questa scoperta conferma ulteriormente l’importanza del ruolo del microambiente in questa patologia; infatti, la presenza di fattori pro-angiogenici può aumentare la migrazione delle cellule dal sangue ai tessuti, dove esse si riproducono. Infine, l’analisi della deregolazione dell’attività proteica ha evidenziato numerosi regolatori con un’attività differente tra le due classi di leucemia. In particolare, DPYSL3 è risultata attivata nel sottotipo A: questa proteina è stata individuata in altre tipologie tumorali e la sua sovra espressione è stata associata allo sviluppo e alla diffusione di metastasi tumorali. Di conseguenza, la sua attivazione nel sottotipo aggressivo di leucemia risulta concorde con quanto riportato in letteratura. Inoltre, anche l’attività della proteina ATM, che ha un ruolo chiave nel controllo del ciclo cellulare, risulta deregolata. In particolare, ha un’attività maggiore nel sottotipo B di leucemia e questo potrebbe indicare una migliore capacità di controllo del ciclo cellulare nelle cellule maligne, che quindi tendono all’apoptosi più facilmente. Da qui la minore aggressività di questo fenotipo. Molto interessante risulta l’elevata attività della proteina DLC1 nel sottotipo A di leucemia. Questo risultato non concorda con il ruolo di oncosoppressore generalmente attribuito alla proteina e, di conseguenza, si ipotizza che esso sia mediato da altri fattori. In conclusione, l’elevata eterogeneità tra le due classi di CLL A e B, sia a livello di espressione genica che di attività proteica, è stata anche riscontrata nell’individuazione di diversi agenti terapici che dovrebbero essere in grado di trattare i due fenotipi patologici. Nonostante la diversità inter-specie, questo lavoro dimostra la possibilità di trasferire un modello murino di CLL sull’uomo. Inoltre, mette in evidenza le differenze genetiche tra due fenotipi di leucemia, di cui uno molto aggressivo. In particolare, individua una deregolazione tra i geni e le proteine responsabili di funzioni biologiche cruciali per lo sviluppo della patologia. Questo dimostra che alla base dell’eterogeneità del decorso clinico non ci sono esclusivamente mutazioni genetiche, ma sono coinvolti numerosi e complessi fattori di regolazione genica; tra tutti, il ruolo del microambiente assume una notevole importanza.
Studio dell'eterogeneità della leucemia linfocitica cronica mediante metodi di biologia dei sistemi
PONTIGGIA, GIULIA
2018/2019
Abstract
Nowadays, a widespread cancer disease is leukaemia. This term refers to a group of diseases which involved the uncontrolled proliferation of a hematopoietic stem cell. Leukemias are divided into acute and chronic, which in turn can be subdivided as lymphoids or myeloids depending on the type of blood cell involved. Four main subtypes of leukaemia are known: acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL) and chronic myeloid leukemia (CML). Chronic lymphocytic leukemia is a lymphoproliferative disorder characterized by the progressive accumulation of mature B lymphocytes in the peripheral blood (PB), bone marrow and lymphoid organs. With an incidence of 4.7 new cases per 100,000 person per year, it is the most common leukemia among adults in Western Countries. The incidence increases by age: currently the median age at diagnosis is 72 years. The clinical course of CLL is extremely heterogeneous. Some patients never require treatment and have a high overall survival rate. On the other hand, other patients show an aggressive disease that requires timely treatment, which with high probability is followed by frequent relapses. For this reason, the survival range varies from a few months to more than a decade. Chronic lymphocytic leukemia does not require treatment until the disease progresses. The therapy can include chemotherapy, chemoimmunotherapy or drugs able to target and stop the growth and survival of CLL cells. Thanks to these numerous alternatives, the recommended therapy varies according to the patient’s physiological state. As for the molecular profile, although there is no specific mutation associated with this disease, genetic alterations can be identified in more than 80% CLL cases. The patients can be divided into two subgroups based on the mutational status of the heavy chain variable region of the immunoglobulin genes (IGHV). These encode for a part of the B cell receptor (BCR). Several analyses show that about 50% of these genes have somatic mutations. The IGHV mutational state (mutated vs unmutated) has a prognostic value as patients without mutation have a more aggressive clinical course. Moreover, the most frequent somatic mutations were found to be involved in critical cellular pathways including: DNA damage and cell cycle control (TP53, ATM, BIRC3), mRNA processing (SF3B1), NOTCH signalling (NOTCH1), and inflammatory pathways (MYD88). In addition to somatic mutations, analysis of CLL gene expression profiles has revealed that hundreds of genes are up- or down-regulated. This could be due to epigenetic events such as hyper- or hypo-methylation of the transcription promoter region of these genes, rather than genetic alterations. On top of this, CLL cell interactions with the supportive tissue microenvironment play a critical role in disease pathogenesis. The microenvironment is the set of accessory cells that, within individual organs, support parenchymal cells through complex communication mechanisms. CLL cells recirculate between peripheral blood and secondary lymphoid organs, where they proliferate rapidly. The stromal cells in these tissues attract and retain CLL cells in cooperation with adhesion molecules. Therefore, given the many factors involved, the division of patients into risk classes is very complex and there are currently no staging systems capable of accurately predicting the evolution of the disease. The following project aims to analyze the genetic differences between two classes of CLL patients, A and B; this division was carried out on the basis of the phenotypic characteristics developed in a murine model of chronic lymphocytic leukemia (Eμ-TCL1). Subtype A is considered aggressive. The study takes into consideration expression data from five public dataset consisting of a large number of samples. The samples were assigned to the subtype that best correlated with their expression profile based on a Fisher test. Samples that were not significant for both groups were considered "unclassified". Subsequently, differential expression analyses and enrichment analyses were performed with the GSEA software to highlight genes and cellular processes with a different behaviour between A and B. The results obtained were compared with those of the mouse model to identify which characteristics are preserved inter-species. Then, through the use of the ARACNe algorithm, a network-based approach was applied to identify the relationships between genes. The objective was to identify the regulatory genes, i.e. those that interact with many other genes and are able to influence their behaviour. In particular, this method was used to study the interactions of the most deregulated genes in the murine model within human dataset, to analyze the behavior of genes belonging to the interferon α pathway and to investigate the role of the extracellular matrix in the two classes A and B. The gene network thus constructed were then visualized and analyzed in Cytoscape. Moreover, since it is known that the onset and progression of a tumour is driven by the aberrant activity of oncoproteins, the activity of some regulatory proteins involved in key processes for cell development within human dataset has been studied. This has been carried out through the analysis of regulons, i.e. the levels of expression of a set of genes believed to be directly regulated by a protein; the VIPER algorithm has made it possible to identify regulators whose activity is deregulated between the two classes. In regards with the classification of samples into the two subtypes, the results showed that more than 50% of the samples belonging to the CD19-positive B cell dataset are attributed to one of the two classes. In contrast, this percentage is much lower for dataset with samples consisting of peripheral blood cells. In addition, dataset of the first type show more deregulated genes between the two categories in common with mice than the others. These differences are due to the different levels of gene expression of the two cell types; the reason why dataset consisting of purified B lymphocytes show better results is related to the fact that the murine phenotypes on which the classification was based were derived from the latter cell type. The GSEA enrichment analysis identified several enriched terms within the dataset, out of which five were among the most significant of the mouse model: "HALLMARK INTERFERON GAMMA RESPONSE", "HALLMARK INTERFERON ALPHA RESPONSE", "MODULE 47", "HALLMARK EPITHELIAL MESENCHYMAL TRANSITION" and "HALLMARK MYOGENESIS". Of these, particular attention has been given to the term "HALLMARK INTERFERON ALPHA RESPONSE", related to the interferon α pathway; in the literature it has been reported that an aberrant activation of the signalling of STAT proteins, usually involved in normal cellular processes, gives rise to various pathological events, including oncogenesis, especially in haematological malignancies. The network-based approach used has highlighted how the genes involved in this pathway are highly connected within our dataset, especially JAK1. This gene encodes for a membrane protein that phosphorylates the STAT proteins, which are transcription activators. Moreover, they are up-regulated in B; consequently, variations in the activity of these genes can lead to a different pathology course. Given the important role that the tissue microenvironment plays in the development of CLL, another gene set considered was that of “Module 47”, which contains extracellular matrix (ECM) and collagen genes. These proved to be highly connected and up-regulated in subtype A of leukemia; the high production of these molecules can stimulate the migration of CLL cells within secondary lymphoid organs, where they proliferate, thus leading to the progression of the disease. This result has also been confirmed by the analysis of protein activity: the protein regulator COL1A1 gene is significantly enriched in class A. Furthermore, a network-based approach applied to the study of the most deregulated gene interactions between A and B in the mouse model highlighted the EPHB4 gene. This gene is up-regulated in the aggressive leukemia subtype and encodes for a protein involved in the development of new vessels. This discovery confirms the importance of the microenvironment in this pathology; in fact, the presence of pro-angiogenic factors can increase the migration of cells from the blood to the tissues, where they reproduce. Finally, the analysis of the deregulation of protein activity between the two subtypes of leukemia has highlighted several regulators with different activity. In particular, DPYSL3 was found to be activated in subtype A: this protein was detected in other tumour types and its over-expression was associated with the development and spread of tumour metastases. As a consequence, its activation in the aggressive leukemia subtype is consistent with what reported in the literature. In addition, the deregulation of the ATM protein, which plays a key role in the control of the cell cycle, emerges. The increased activity of the ATM protein in subtype B of leukemia could indicate a greater ability to control the cell cycle in malignant cells, which therefore lead to apoptosis more easily. Hence the lower aggressiveness of this phenotype. Very interesting is the high activity of the DLC1 protein in subtype A of leukemia. This result does not agree with the oncosuppressant role generally attributed to the protein and, consequently, it is assumed that it is mediated by other factors. In conclusion, the high heterogeneity between the two classes of CLL A and B, both in terms of gene expression and protein activity, was also found in the identification of different therapeutic agents that should be able to treat the two pathological phenotypes. Despite the inter-species diversity, this work demonstrates the possibility of transferring a mouse model of CLL to humans. It also highlights the genetic differences between two phenotypes of leukemia, one of which is very aggressive. In particular, it identifies a dysregulation between the genes and the proteins responsible for crucial biological functions for the development of the disease. This shows that at the basis of the heterogeneity of the clinical course there are not only genetic mutations, but many complex gene regulation factors involved; among all, the role of the microenvironment assumes considerable importance.File | Dimensione | Formato | |
---|---|---|---|
2020_06_Pontiggia.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
3.35 MB
Formato
Adobe PDF
|
3.35 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/165525