Inhibition of Purine Nucleoside Phosphorylase Is a Promising Anti-Sickling Approach for Sickle Cell Disease

Introduction: Red blood cells (RBC) have the highest level per cell volume of purine nucleoside phosphorylase (PNP), a key enzyme in purine metabolism. PNP converts guanosine to guanine, inosine to hypoxanthine, and 2'-deoxyguanosine to guanine, ultimately leading to the production of pro-oxidative and vasculotoxic hypoxanthine and xanthine. We reported increased PNP activity and accelerated purine metabolism in patients with sickle cell disease (SCD) (Bilan et al, 2018). Recently, we confirmed the presence of dysregulated purine metabolism in transgenic sickle Townes mice (SS) and showed that the PNP inhibitor 8-aminoguanosine (8-AG) is safe, rebalances purine metabolism and reduces organ damage in SS mice (Tofovic et al. 2022). However, the mechanism for PNP's beneficial effects in SCD is unknown. Herein, we tested our hypothesis that modulating purine metabolism by inhibiting PNP activity could lead to reduced sickling and hemolysis. Methods: We analyzed PNP protein levels by ELISA in banked plasma samples from individuals with HbSS SCD (n=23) at steady state as well as race, age and gender matched HbAA controls (n=22) enrolled in a prior clinical trial (NCT02946905). We also collected blood prospectively from patients with SCD at steady state (n=3) and incubated it with vehicle, 8-AG or the potent PNP inhibitor forodesine (positive control), and then determined the point of sickling (PoS; the partial pressure of oxygen at which there is a 5% decrease in maximum RBC deformability) and assessed RBC deformability during normoxia (maximum elongation index EImax) and hypoxia (minimum elongation index EImin) using automated ektacytometry (Lorrca Oxygenscan) whereby a decrease in PoS or increase in RBC deformability suggests decreased RBC sickling. Next, to demonstrate the in vivo effect of PNP inhibition, we treated Townes SS and AA mice with vehicle or 8-AG in their drinking water and assessed PoS and hemolysis. Results: Plasma PNP levels were markedly elevated in the patients with SCD compared to controls (3198 ± 2154 vs. 1365 ± 985 pg/mL, P<0.0001). The PNP levels of both groups negatively correlated with RBC counts (r=-0.56, p=0.0004) and positively with platelets counts (r=0.42, p=0.012). Exposure of patient's blood to 8-AG and forodesine reduced PoS compared to vehicle (pO2 39.55 ± 8.28% vs. 35.47 ± 8.86% vs. 43.37 ± 8.17%). The EImax was increased with 8-AG (0.4 ± 0.11) and forodesine (0.40 ± 0.11) compared with vehicle (0.37 ± 0.13). While 8-AG (0.26 ± 0.07) increased EImin, forodesine (0.22±0.05) decreased EImin compared to vehicle (0.23 ± 0.07). Like patients with SCD, SS mice had higher PNP activity in the RBC lysate compared to AA control mice at baseline (0.7 ± 0.3 vs. 0.4 ± 0.1 x 10-3 pmol/min/cell, p=0.0016). After two weeks of 8-AG treatment, the PoS was reduced in the 8-AG-treated-SS mice from 48.08 ± 8.75% to 42.48 ± 10.92%, the EImax increased from 0.38 ± 0.08 to 0.41 ± 0.09, and the EImin decreased from 0.17 ± 0.12 to 0.15 ± 0.10. There was a parallel improvement in hemolytic markers (reduced plasma free hemoglobin, AST, and hemoglobinuria) in the 8-AG-treated-SS mice and no adverse effects on other blood counts. Additionally, platelet count increased, suggesting reduced splenic platelet pooling. Conclusion: This is the first study to show that pharmacological modulation of PNP activity could be beneficial in SCD by improving RBC deformability and thereby reducing hemolysis. We suggest that PNP inhibition could be a novel anti-sickling strategy for the treatment of SCD.