The role of cytosolic 5'-nucleotidase II (NT5C2) in drug resistance and relapse of acute lymphoblastic leukemia

• Main Author: Tzoneva, Gannie Valentinova
• Language: English
• Published: 2016
• Subjects: Drug resistance; Lymphoblastic leukemia; Oncology
• Online Access: https://doi.org/10.7916/D85Q4W77

Acute lymphoblastic leukemia (ALL) is an aggressive hematological cancer which arises from the malignant transformation of B-cell or T-cell progenitors. Despite recent pioneering improvements in intensified combination chemotherapy, 20% of pediatric and 50% of adult ALL patients present with primary drug-resistant leukemia or develop relapse. Treatment of refractory and relapsed ALL has remained a significant clinical challenge with survival rates following relapse of only 40%, highlighting the need to understand the mechanisms which drive drug resistance and relapse of ALL. Through extensive sequencing analyses of matched diagnostic, remission and relapsed DNA samples from patients with B-precursor ALL (B-ALL) and T-cell ALL (T-ALL) we have identified recurrent relapse-specific gain-of-function mutations in the cytosolic 5'-nucleotidase II (NT5C2) gene in 25% of relapsed T-ALLs and 6% of relapsed B-ALLs. NT5C2 is a highly conserved, ubiquitously expressed enzyme which regulates intracellular purine nucleotide levels by dephosphorylating purine monophosphates. NT5C2 also dephosphorylates key metabolites in the activation of purine analog prodrugs such as 6-mercaptopurine and 6-thioguanine which are routinely used in the treatment of ALL, allowing purine analog nucleosides to be readily exported out of the cell. Here we show that mutant NT5C2 proteins have increased 5’-nucleotidase activity and confer resistance to 6-mercaptopurine and 6-thioguanine chemotherapy when expressed in leukemic cells. Consistently, NT5C2 mutations correlate with early relapse and relapse while under therapy. We present a novel T-ALL conditional inducible knock-in mouse model of the highly recurrent NT5C2 R367Q mutation and show that expression of one Nt5c2 R367Q allele from the endogenous locus in primary T-ALL lymphoblasts induces overt resistance and disease progression under therapy with 6-mercaptopurine in vivo, while surprisingly conferring reduced growth and decreased leukemia initiating activity in the absence of chemotherapy. Metabolically we show that the observed loss of fitness in Nt5c2 R367Q tumors can be explained by a severe depletion of endogenous purine monophosphate metabolites as a result of increased Nt5c2 5’-nucleotidase activity. Consistently, using ultra-sensitive mutation analyses we show that relapse-associated NT5C2 mutations are not detectable at initial disease presentation, indicating that NT5C2-mutant tumor cells are negatively selected by clonal competition in the early stages of disease development and only positively selected under prolonged 6-mercaptopruine chemotherapy which is the backbone treatment for ALL following remission. Our findings present the first known example of chemotherapy resistance and disease progression driven by a tumor clone with decreased leukemia initiating activity, highlighting the intense selective pressure of chemotherapy in the clonal evolution of tumors from diagnosis to relapse. Through extensive biochemical and structural characterizations of recombinant NT5C2 mutant proteins, we have grouped relapse-specific NT5C2 activating mutations into 3 different classes, each conferring unique enzymatic behavior in basal conditions and in response to allosteric activation, and each with unique structural features which mediate increased 5’-nucleotidase activity. Moreover, we identify a novel auto-regulatory switch-off mechanism of the NT5C2 enzyme involving movement of an unstructured flexible loop, and present the first crystal structure view of the NT5C2 C-terminal acidic tail, implicating it as an auto-inhibitory brake to the allosteric activation of the enzyme. The presence of multiple mutational mechanisms of activating such a highly conserved enzyme, especially in light of the inherent loss of fitness to the tumor cells, indicates a strong convergent evolution towards activating NT5C2. This is supported by our discovery that patients can harbor multiple leukemic clones with NT5C2 mutations at relapse. Overall our findings highlight NT5C2 as a major driver of drug resistance and relapse of ALL and pinpoint metabolic susceptibilities which could be exploited therapeutically to target NT5C2-mutant tumors in the future. Our in-depth structural and enzymatic knowledge of mutant NT5C2 proteins will serve as an essential tool in the rational drug development of novel NT5C2 inhibitors with increased specificity and selectivity for mutant NT5C2, while our novel Nt5c2 R367Q knock-in mouse model will serve as a platform for the pre-clinical testing of both NT5C2 inhibitors and alternative compounds selective for Nt5c2-mutant leukemias which can be used for prevention and treatment of relapsed ALL.