Activating pyruvate kinase improves red blood cell integrity by reducing band 3 tyrosine phosphorylation

Abstract

In a phase 1 study (NCT04000165), we established proof of concept for activating pyruvate kinase (PK) in sickle cell disease (SCD) as a viable antisickling therapy. AG-348 (mitapivat), a PK activator, increased adenosine triphosphate (ATP) and decreased 2,3-diphosphoglycerate levels while patients were on treatment, in line with the mechanism of the drug. We noted that the increased hemoglobin (Hb) persisted for 4 weeks after stopping AG-348 until the end of study (EOS). Here, we investigated the pathways modulated by activating PK that may contribute to the improved red blood cell (RBC) survival after AG-348 cessation. We evaluated frozen whole blood samples taken at multiple time points from patients in the phase 1 study, from which RBC ghosts were isolated and analyzed by western blotting for tyrosine phosphorylation of band 3 (Tyr-p-bd3), ankyrin-1, and intact (active) protein tyrosine phosphatase 1B (PTP1B) levels. We observed a significant dose-dependent decrease in mean Tyr-p-bd3 from baseline in the patients, accompanied by an increase in the levels of membrane-associated ankyrin-1 and intact PTP1B, all of which returned to near baseline by EOS. Because PTP1B is cleaved (inactivated) by intracellular Ca2+-dependent calpain, we next measured the effect of AG-348 on ATP production and calpain activity and the plasma membrane Ca2+ ATPase pump-mediated efflux kinetics in HbAA and HbSS erythrocytes. AG-348 treatment increased ATP levels, decreased calpain activity, and increased Ca2+ efflux. Altogether, our data indicate that ATP increase is a key mechanism underlying the increase in hemoglobin levels upon PK activation in SCD. This trial was registered at www.clinicaltrials.gov as #NCT04000165.

Graphical abstract

Figure 1.

Effect of PK activators on Tyr-p-bd3. (A) In vivo effect of AG-348 on Tyr-p-bd3, PTP1B, and ankyrin-1. The levels of Tyr-p-bd3, intact PTP1B, and membrane-associated ankyrin-1 were quantified by western blotting in frozen blood samples from 15 patients (Tyr-p-bd3 and PTP1B) or 11 patients (ankyrin-1) treated twice daily with 5-, 20-, 50-, and 100-mg AG-348. ∗P ˂ .05; ∗∗P ˂ .01; ∗∗∗P ˂ .001. (B-E) Ex vivo effect of AG-348 on Tyr-p-bd3. HbSS RBCs were treated with DMSO (Con) or different concentrations (1, 3, 10, and 30 μM) of AG-348 in DMSO for 4 hours (B-C) or 30-μM AG-348 for different times (1, 2, 4, and 6 hours) (D-E). The level of Tyr-p-bd3 was analyzed by western blotting and quantified by densitometry analysis. The mean reduction on Tyr-p-bd3 of RBC treated with 1-, 3-, 10-, and 30-μM AG-348 were 1%, 22%, 35%, and 50%, respectively, and with 30-μM AG-348 for 2, 4, and 6 hours were 6%, 28%, and 59%, respectively. ∗P ˂ .05; ∗∗P ˂ .01. (F-I) Ex vivo effect of AG-946 on Tyr-p-bd3. HbSS RBCs were treated with DMSO (Con) or different concentrations (1, 3, 10, and 30 μM) of AG-946 in DMSO for 4 hours (F-G) or 30-μM AG-946 for different times (1, 2, 4, and 6 hours) and the level of Tyr-p-bd3 analyzed by western blotting and quantified by densitometry analysis (H-I). The mean reduction on Tyr-p-bd3 of RBC treated with 1-, 3-, 10-, and 30-μM AG-946 were 3%, 24%, 41%, and 56%, respectively, and with 30-μM AG-946 for 4 and 6 hours were 29% and 49%, respectively. ∗P ˂ .05; ∗∗ P ˂ .01. Three biological replicates (blood samples from 3 different HbSS donors) were used in all ex vivo assays, except in panels H and I, which had 4 biological replicates. ∗P ˂ .05; ∗∗P ˂ .01.

Figure 2.

Effect of AG-348 on the association of band 3 with ankyrin-1. (A-B) HbSS RBCs were treated with DMSO (Con) or 30-μM AG-348 in DMSO for different times (1, 2, 4, and 6 hours), and the level of membrane-associated ankyrin-1 was analyzed by western blotting and quantified by densitometry. The mean increase on ankyrin-1 of RBC treated with 30-μM AG-348 for 4 and 6 hours were 31% and 38%, respectively. ∗P ˂ .05. (C-D) HbSS RBCs were treated with DMSO (Con) or different concentrations (1, 3, 10, and 30 μM) of AG-348 in DMSO for 8 hours, and the interaction between ankyrin-1 and band 3 was measured by immunoprecipitation and quantified by densitometry analysis. The mean increase on the interaction between ankyrin-1 and band 3 of RBC treated with 1-, 3-, 10-, and 30-μM AG-348 were 5%, 38%, 45%, and 61%, respectively. ∗P ˂ .05. Three biological replicates of HbSS donors were used in all ex vivo assays.

Figure 3.

Effect of AG-348 on protein phosphatase, PTP1B. (A-B) Ex vivo effect of AG-348 and R406 on Tyr-p-bd3. HbSS RBCs were treated with DMSO (Con) or different concentrations (3, 10, and 30 μM) of R406, AG-348, or a combination of R406 and AG-348 for 6 hours, and the level of Tyr-p-bd3 was analyzed by western blotting and quantified by densitometry analysis. The mean reductions on Tyr-p-bd3 of RBCs treated with 10- and 30-μM R406, AG-348, or a combination compared with vehicle control were 15% and 36%, 6% and 31%, 39% and 66%, respectively. ∗∗P ˂ .01; ∗∗∗P ˂ .001. (C-D) Effect of AG-348 on RBC PTP1B. HbSS RBCs were treated with DMSO (Con) or 30-μM AG-348 in DMSO for different times (1, 2, 4, and 6 hours), and the level of membrane-associated intact PTP1B was analyzed by western blotting and quantified by densitometry analysis. The mean increase on PTP1B of RBCs treated with 30-μM AG-348 for 4 and 6 hours were 25% and 43%, respectively. ∗P ˂ .05. (E) Immunoprecipitation with anti-PTP1B antibody clearly showed interaction between PTP1B and band 3. (F-G) HbSS RBCs were treated with DMSO (Con) or different concentrations (1, 3, 10, and 30 μM) of AG-348 for 8 hours, and the interaction between PTP1B and band 3 was measured by immunoprecipitation and quantified by densitometry analysis. The mean increase on the interaction between PTP1B and band 3 of RBC treated with 1-, 3-, 10-, and 30-μM AG-348 were 31%, 35%, 70%, and 121%, respectively. ∗P ˂ .05. Three biological replicates of HbSS donors were used in all ex vivo assays. IgG, immunoglobulin G.

Figure 4.

Effect of AG-348 on calpain activity, ATP production, and Ca²⁺ efflux. (A-B) PTP1B in HbAA vs HbSS RBCs. The levels of membrane-associated intact PTP1B in RBCs from 6 pairs of healthy controls (AA1-6) and HbSS patients (SS1-6) were analyzed by western blotting and quantified by densitometry. The relative levels of PTP1B to actin in AA1-6 RBCs were 0.42, 0.84, 0.81, 0.63, 0.39, and 0.75, respectively, with an average of 0.64; and in SS1-6 RBCs were 0.28, 0.63, 0.37, 0.21, 0.51, and 0.20, respectively, with an average of 0.37. ∗P ˂ .05. (C) Effect of AG-348 on calpain activity. HbSS RBCs were treated with DMSO (Con) or different concentrations (1, 3, 10, and 30 μM) of AG-348 in DMSO overnight, and the activity of calpain was measured using Calpain Activity Assay Kit. Mean reductions in calpain activity of RBCs treated with 1-, 3-, 10-, and 30-μM AG-348 were 9%, 15%, 24%, and 26%, respectively. ∗P ˂ .05. (D-E) Effect of AG-348 on ATP production. HbSS RBCs were treated with DMSO (Con) or 30-μM AG-348 in DMSO for different times (2, 4, 6, and 8 hours) or different concentrations (1, 3, 10, and 30 μM) of AG-348 in DMSO for 8 hours; ATP level was analyzed using CellTiter-Glo 2.0 Cell Viability Assay kit. The mean increase in ATP levels of RBCs treated with 30-μM AG-348 for 2, 4, 6, and 8 hours were 2%, 5%, 11%, and 21%, respectively, and with 1-, 3-, 10-, and 30-μM AG-348 for 8 hours were 2%, 5%, 14%, and 26%, respectively. Three biological replicates of HbSS donors were used in the study here. ∗P ˂ .05. (F-H) Effect of AG-348 on Ca²⁺ efflux. (F) Ca²⁺ efflux kinetics were measured using extracellular Fluo-8. Symbols represent mean ± standard error of the mean (SEM) from triplicate measurements from single HbAA and HbSS donors (dark blue and red circles, respectively). Notice the slower efflux with the SS donor. (G) Mean ± SEM efflux halftimes from 4 donors, measured similar to panel F. The slower efflux observed from SS erythrocytes did not reach statistical significance. (H) Mean ± SEM efflux halftimes for indicated cells after a 6-hour treatment with 30-μM AG-348 or matched DMSO control (blue and red bars, respectively, for each genotype). RBCs from 4 pairs of ethnic-matched healthy controls (HbAA) and HbSS patients were used in each assay, and the results represent a mean from the 4 separate assays.