AG-946 Normalizes Glycolysis and Improves Red Cell Indices in a Humanized Sickle Cell Mouse Model

We have previously demonstrated that enzymes of the glycolysis pathway and deoxy hemoglobin (deoxyHb) reversibly bind to the cytoplasmic domain of Band 3 (CdB3). For these studies we generated 3 transgenic mice in which the endogenous mouse Band 3 was replaced with: i) human wild type CdB3 (WT); ii) human CdB3 lacking the deoxyHb binding site (Hb-); iii) human CdB3 with high-affinity binding site for deoxyHb (Hb+). In the humanized WT strain, glycolytic enzymes (including fructose-1,6-biphosphate [FBP], dihydroxyacetone phosphate/ glyceraldehyde-3-phosphate [G3P] and 3-phosphoglycerate/2-phosphoglycerate [PG]) bind to CdB3 in oxygenated conditions and are displaced by the binding of deoxyHb in deoxygenated conditions. Erythrocytes from the mutant lines were insensitive to oxygen concentration (Chu et al. Blood 128, 2016, Zheng et al. JBC 294, 2019, Zhou et al. Sci. Adv. 5, 2019), showing that CdB3 constitutes a molecular switch regulating assembly of glycolytic enzymes on the erythrocyte membrane based on the oxygenation state. To study if this mechanism plays a role in Sickle Cell Disease (SCD), we crossed the humanized Band 3 mouse strains to the Townes SCD mouse model. -Hb_SS mice had a significantly higher percentage of sickled cells and a higher rate of sickling compared to WT_SS animals. ++Hb_SS mice showed a decrease in the percentage of sickled cells and the rate of sickling. We hypothesized that the inability of the glycolytic enzymes to reversibly bind to CdB3 in the -Hb mice inhibited glycolysis. To test this hypothesis, we analyzed a panel of 28 cellular metabolites in each genotype (WT_AA, _AS, and _SS; -Hb_AA, _AS, _SS; and ++Hb_AA, _AS, _SS) using an API 4500 triple quadrupole mass spectrometer (AB Sciex), with a polymeric amino column (apHera by Supelco) with stable isotope spike in controls allowing the absolute quantification of each metabolite. Consistent with the constitutive binding of the terminal glycolytic enzymes to CdB3 in -Hb erythrocytes, glycolysis was inhibited, as evidenced by significant accumulation of the intermediates at the top of the glycolysis pathway, including FBP, G3P and PG (p values all <0.01) compared to WT cells. In the ++Hb mutant where the terminal glycolytic enzymes are constitutively displaced from CdB3, significantly lower levels of FBP, G3P and PG compared to WT cells were observed (p values all <0.01). AG-946 is a novel small-molecule activator of erythrocyte pyruvate kinase that mediates the final step of glycolysis yielding 2 ATP and 2 pyruvate molecules per glucose molecule. We hypothesized that treatment of the -Hb-SS animals treated with AG-946 would reduce the levels of the glycolytic intermediates by accelerating the terminal stage of glycolysis. We designed a pilot study in which -Hb-SS mice were fed control chow for 4 weeks and then randomly separated into a group receiving control chow and a group receiving AG-946 at an approximate dose of 10 mg/kg/day. After 8 weeks on study, blood was collected from treated and untreated animals for analysis of complete blood counts, plasma levels of AG-946, the levels of 28 metabolites, intracellular 2,3-DPG, and the rate and degree of sickling. AG946 was well tolerated, all mice in both arms of the study gained weight over the 8 weeks. At the end of the study, the plasma levels of AG946 in the treated animals ranged from 736 to 2281 nM. In the treated animals we observed a complete normalization of the levels of the glycolytic intermediates and 2,3-DPG while the levels in the untreated group were unchanged. Treated animals showed significant increases in the red blood cell counts (from 4.6 + 0.52 in the control group to 7.07 + 0.21 1012/L in the treated group), hematocrit (from 24.7 + 1.8% to 33 + 1.8%) and hemoglobin (6.9 + 0.8 to 8.9 + 0.3 g/dL; p values all <0.016). We observed significant decreases in the MCV (53.5 + 2.3 to 46.6 + 1.1) and MCH (14.9 + 0.05 to 12.6 + 0.25; p values all <0.01). These differences did not correlate with the plasma levels of AG-946 indicating that levels of 736 nM are sufficient to correct the block in glycolysis and improve the red cell indices. No differences were observed in MCHC or WBC, or in the rate or degree of sickling. We conclude that AG-946 treatment effectively enhances glycolysis in humanized Band 3/SS mice. The normalization of glycolytic intermediates is accompanied by increases in critical red cell indices (RBC, HCT, Hb, and MCV). AG946 may be an effective treatment for sickle cell disease.