Defective Nucleotide Catabolism Defines a Subset of Cancers Sensitive to Purine Nucleoside Phosphorylase Inhibition

Small molecule inhibitors of purine nucleoside phosphorylase (PNP) have been explored as a treatment strategy for leukemia and lymphoma, however, the determinants of response to this class of drugs are incompletely understood. PNP inhibitors impair cell proliferation by preventing catabolism of the nucleoside deoxyguanosine (dG) which induces toxic imbalances amongst intracellular deoxyribonucleotide triphosphate (dNTP) pools following its phosphorylation and trapping by nucleoside kinases. We hypothesized that differential nucleoside uptake or catabolism defines cancer cell lines as either sensitive or resistant to PNP inhibition. Among cancer cell lines we found that T-cell acute lymphoblastic leukemia (T-ALL) cells are uniquely and acutely sensitive to PNP inhibition whereas the B-cell leukemia and solid tumor models are completely resistant. We determined that although the nucleoside scavenging kinase deoxycytidine kinase (dCK) was active in all cells tested, PNP inhibitors only induced dGTP pool increases in sensitive models. By evaluating the expression of key genes involved in nucleotide scavenging, biosynthesis, and phosphohydrolysis in a panel of sensitive and resistant cell lines we found that the dNTP phosphohydrolase SAM histidine aspartate containing protein 1 (SAMHD1) was exclusively expressed in resistant models. Using CRISPR/Cas9 SAMHD1 knockout cell lines, we verified that PNP inhibitor sensitivity is a function of SAMHD1 expression and determined that the pharmacological inhibition of dCK or genetic restoration of SAMHD1 conferred resistance to PNP inhibition. Importantly, we determined that low expression of SAMHD1 is not limited to T-ALL as subset of established and primary solid tumors models are SAMHD1-deficient. These solid tumor models were consistently acutely sensitive to PNP inhibitors which indicates that the utility of PNP inhibitors extends beyond hematological malignancies. Additionally, we found that deoxycytidine (dC) can limit the anti-proliferative effects of PNP inhibitors by competing with dG for phosphorylation by dCK but this effect can be overcome by expression of dC-catabolizing gene cytidine deaminase (CDA). Collectively, these results indicate that SAMHD1, dCK and CDA are critical biomarkers that must be used to stratify patients in clinical trials evaluating pharmacological PNP inhibition.