Characterising KMT2A and MLLT10 Rearranged Acute Lymphoblastic Leukaemia

Abstract

KMT2A-rearrangements (KMT2Ar) and MLLT10-rearrangements (MLLT10r) are recurrently identified oncogenic lesions in acute lymphoblastic leukaemia (ALL). KMT2Ar induce aggressive, high-risk ALL in patients of all ages, including >70% of infant ALL diagnoses. MLLT10r, most commonly PICALM-MLLT10 and DDX3XMLLT10, account for approximately 10% of new T-cell ALL (T-ALL) diagnoses, and are biologically and clinically poorly characterised. KMT2Ar and MLLT10r leukaemic blasts share the important feature of HOXA cluster transcriptional dysregulation, suggesting a similar aetiology and subsequently similar therapeutic vulnerabilities. However, there is presently a lack of pre-clinical in vitro and in vivo models that accurately recapitulate the aggressive ALL induced by KMT2Ar and MLLT10r, limiting the ability to understand the biological aetiology of these subtypes and screen for efficacious targeted therapies. This thesis describes novel in vitro models of KMT2A-AFF1, PICALM-MLLT10 and DDX3X-MLLT10 T-ALL, and KMT2A-AFF1 B-ALL. These models demonstrated the ability of KMT2A-AFF1 to induce oncogenic changes in murine MOHITO cells. Specifically, KMT2A-AFF1 expression induced T-ALL characterised by an immature immunophenotype, increased cellular proliferation and upregulation of HOXA cluster genes, reflecting the aggressive disease observed in patients. For the first time, an in vivo transplantation model of DDX3X-MLLT10 and PICALMMLLT10 T-ALL is also presented, generated by the expression of these fusions in CD4-CD8- p19Arf-null (Arf-/-) thymocytes. When transplanted into sub lethally irradiated immunocompromised NSG mice, MLLT10r-expressing cells consistently engrafted to induce T-ALL. Differences were noted in the immunophenotype and dissemination patterns of engrafted cells between the two fusion cohorts, indicating that the 5’ fusion partner plays an important role in MLLT10r. The mutational landscape of non-infant KMT2Ar B-ALL and MLLT10r ALL were investigated by interrogating patient leukaemic blasts for the presence of single nucleotide variants (SNVs) and copy number alterations (CNAs). An understanding of the complete genomic landscape of leukaemic blasts is required, in order to generate pre-clinical models that fully recapitulate ALL. Analysis revealed that non-infant KMT2Ar B-ALL lacks typical leukaemia-associated alterations that are common in other genomic subtypes of B-ALL, demonstrating a silent mutational landscape that parallels KMT2Ar infant cohorts. However, known and novel SNVs within epigenetic regulatory genes were identified in KMT2Ar B-ALL, that were less common or absent in the BCR-ABL1 comparator B-ALL cohort, implying that epigenetic regulators may serve a functional role specifically in KMT2Ar B-ALL. Recurrent alterations in leukaemia-associated genes were identified in MLLT10r T-ALL cases, suggesting that cooperative lesions play a role in MLLT10r leukaemogenesis, unlike in KMT2Ar. Therapeutic efficacy of the novel curaxin CBL0137 in infant KMT2Ar ALL has recently been reported. Pre-clinical sensitivity studies in in vitro models demonstrated that functional p53 is critical in the mechanism of action of CBL0137. CBL0137 demonstrated efficacy across multiple genomic subtypes of acute leukaemia, including KMT2Ar, only in the presence of wild-type TP53. Findings support that CBL0137 has clinical promise in p53 wild-type ALL, including KMT2Ar ALL due to the low prevalence of TP53 alterations across all ages in this subtype. In summary, this thesis provides critical insight into the complex biology of KMT2Ar and MLLT10r ALL, and presents the first described models of KMT2Ar and MLLT10r T-ALL. Together, these findings contribute to improving our understanding of the biology of these high-risk subtypes, and the described models permit the pre-clinical exploration of targeted therapies in a T-ALL context, to ultimately improve outcomes for ALL patients.