La Sclerosi Multipla è una malattia con decorso clinico molto eterogeneo e, in considerazione dell’elevato numero di terapie ad oggi disponibili, vi è una grande necessità di individuare dei fattori prognostici che possano guidare la gestione clinica dei indagato i parametri clinici, genomici ed immunologici associati all’attività infiammatoria di malattia. A tal fine abbiamo incluso due coorti di pazienti con Sclerosi Multipla: una coorte “Estesa” di ~1,000 individui trattati con un farmaco di prima linea per cui erano disponibili dati clinici e genetici, ed una coorte “Centrale” di ~200 pazienti per cui i dati genetici, trascrittomici e di repertorio immunitario sono stati generati da campioni pre-trattamento. I pazienti sono quindi stati osservati per 4 anni e classificati secondo il criterio di assenza di attività di malattia e sulla base del tempo alla prima ricaduta. E’ quindi stata testata l’associazione tra le singole omiche e questi obiettivi clinici ed infine i diversi livelli di informazione sono stati integrati per costruire un modello predittivo di attività di malattia. I risultati hanno confermato che un esordio più giovanile, una durata di malattia più corta ed il sesso femminile sono associati ad una maggiore attività infiammatoria di malattia durante l’osservazione. Inoltre, lo studio genetico ha evidenziato alcuni segnali suggestivi di associazione nelle vicinanze di geni implicati nel sistema della coagulazione, nella modulazione della permeabilità vasale nonché in processi antiossidanti e nella regolazione della differenziazione degli oligodendrociti e della remielinizzazione. Questi risultati vengono corroborati dalle analisi di trascrittomica e di pathway che hanno selezionato processi biologici implicati non solo in funzioni immunitarie ma soprattutto nella regolazione dell’omeostasi e morte cellulare ed in meccanismi di neurodegenerazione. Inoltre, sembra che un maggior grado di diversità immunologica correli con un maggior rischio di riattivazione di malattia. In ultimo, abbiamo applicato algortimi di “machine learning” per integrare i dati clinici, genetici ed immunologici al fine di predire l’attività di malattia: i migliori esiti in termini predittivi sono stati ottenuti utilizzando le sole informazioni cliniche mentre l’aggiunta di dati molecolari ha portato solo un modesto incremento di predittività. In conclusione, i risultati della presente tesi dimostrano che dati genetici, di trascrittomica e di repertorio immunitario possono aiutare a decifrare i processo biologici alla base della patogenesi e della evoluzione della Sclerosi Multipla ma il potere predittivo di modelli che integrano questi diverse informazioni è attualmente insufficiente per l’applicazione in ambito clinico, probabilmente anche in considerazione del numero limitato di pazienti inclusi nello studio in confronto alle migliaia di variabili considerate in tali modelli.

Multiple Sclerosis has a highly heterogeneous clinical course and, given the large number of available therapies, there is a strong need to identify prognostic markers of disease activity that can guide patients’ management towards a more personalized approach. In the present thesis we investigated clinical, genomic and immunological parameters associated with inflammatory activity. To this aim, we enrolled two cohort of multiple sclerosis patients: an “Extended” cohort of ~1,000 subjects that started a first-line drug, with available clinical and genetic data, and a “Core” dataset of ~200 patients with genetic, transcriptomic and immune repertoire information obtained before treatment. The patients were observed for 4 years and classified according to the no evidence of disease activity status and the time to first relapse criteria. We then tested the relationship between the different –omics data and disease outcomes and we integrated the different layers of information into a prognostic model of disease activity. Our results confirmed that a younger age at onset, a shorter disease duration and female gender strongly correlate with higher inflammatory activity during follow-up. Besides, the genetic study highlighted some suggestive associations near to genes implicated in the regulation of coagulation system and vascular permeability as well as in antioxidant processes, regulation of oligodendrocyte differentiation and remyelination. These findings were strengthened by the results of the transcriptomic and pathway analyses that prioritized biological paths involved not only in immune functions but largely in cell homeostasis and death as well as neurodegeneration. Moreover, we also found that a higher immunological diversity seems to correlate with MS reactivation during follow-up. Finally, we applied machine learning algorithms to integrate clinical, genetic and immunological data and found that the best predictive performance was obtained using clinical information, while the addition of molecular data only slightly improved the prediction of disease activity. In conclusions, our findings demonstrated that genetic, transcriptomic and immune repertoire data can help in deciphering biological processes underlying Multiple Sclerosis pathophysiology and clinical expression but the predictive power of models integrating the different layers of information is currently not enough for application in clinical practice, probably because of the limited sample size of the studied cohort compared also to the number of thousands of features that are tested in the predictive model.

Characterization of the genomic profile and immune repertoire in Multiple Sclerosis patients to predict disease activity / Laura Ferrè - : . , 2022 Jul 08. ((34. ciclo, Anno Accademico 2020/2021.

Characterization of the genomic profile and immune repertoire in Multiple Sclerosis patients to predict disease activity

FERRÈ, LAURA
2022

Abstract

Multiple Sclerosis has a highly heterogeneous clinical course and, given the large number of available therapies, there is a strong need to identify prognostic markers of disease activity that can guide patients’ management towards a more personalized approach. In the present thesis we investigated clinical, genomic and immunological parameters associated with inflammatory activity. To this aim, we enrolled two cohort of multiple sclerosis patients: an “Extended” cohort of ~1,000 subjects that started a first-line drug, with available clinical and genetic data, and a “Core” dataset of ~200 patients with genetic, transcriptomic and immune repertoire information obtained before treatment. The patients were observed for 4 years and classified according to the no evidence of disease activity status and the time to first relapse criteria. We then tested the relationship between the different –omics data and disease outcomes and we integrated the different layers of information into a prognostic model of disease activity. Our results confirmed that a younger age at onset, a shorter disease duration and female gender strongly correlate with higher inflammatory activity during follow-up. Besides, the genetic study highlighted some suggestive associations near to genes implicated in the regulation of coagulation system and vascular permeability as well as in antioxidant processes, regulation of oligodendrocyte differentiation and remyelination. These findings were strengthened by the results of the transcriptomic and pathway analyses that prioritized biological paths involved not only in immune functions but largely in cell homeostasis and death as well as neurodegeneration. Moreover, we also found that a higher immunological diversity seems to correlate with MS reactivation during follow-up. Finally, we applied machine learning algorithms to integrate clinical, genetic and immunological data and found that the best predictive performance was obtained using clinical information, while the addition of molecular data only slightly improved the prediction of disease activity. In conclusions, our findings demonstrated that genetic, transcriptomic and immune repertoire data can help in deciphering biological processes underlying Multiple Sclerosis pathophysiology and clinical expression but the predictive power of models integrating the different layers of information is currently not enough for application in clinical practice, probably because of the limited sample size of the studied cohort compared also to the number of thousands of features that are tested in the predictive model.
Characterization of the genomic profile and immune repertoire in Multiple Sclerosis patients to predict disease activity / Laura Ferrè - : . , 2022 Jul 08. ((34. ciclo, Anno Accademico 2020/2021.
Doctoral Thesis
File in questo prodotto:
File Dimensione Formato  
PHD_THESIS_FERRE.pdf

embargo fino al 08/07/2023

Descrizione: TESI DEFINITIVA
Tipologia: Tesi di dottorato
Dimensione 4.05 MB
Formato Adobe PDF
4.05 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11768/133077
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact