Mutations in the Hemoglobin Binding Domain of Band 3 Perturb Cellular Metabolism and Alter the Sickling of SS Erythrocytes

Under oxygenated conditions, 4 glycolytic enzymes that perform the terminal steps of glycolysis (phospho-fructoKinase [PFK], lactate dehydrogenase [LDH], aldolase [ALD] and glygeraldehyde 3 phosphate dehydrogenase [GAPDH]) bind to the cytoplasmic domain of band 3. Under deoxy conditions deoxy hemoglobin (Hb) is bound to band 3 and PFK, LDH, ALD and GAPDH are displaced (Campanella et al. PNAS 102, 2005; Blood 112, 2008). We generated transgenic mice in which the sequence encoding the first 35 amino acids of the wild type human band 3 cytoplasmic domain replaced the endogenous mouse band 3 sequences in the Slc4a1 gene, a mutant line in which human amino acids 12-21 were deleted removing the deoxy Hb binding site (-Hb) and a third line in which amino acids 1-11 were deleted creating a high affinity binding site for deoxyHb (++Hb). Erythrocytes from the mutant lines were insensitive to Oxygen concentration resulting in changes in oxygen dependent deformability and other physical properties compared to the wild type line (Chu et al. Blood 128, 2016, Zheng et al. JBC 294, 2019, Zhou et al. Sci. Adv. 5, 2019).

We crossed our humanized band 3 mouse strains to the Townes Sickle Cell Disease (SCD) mouse model, maintaining both the human βA and βS alleles to generate human AA, AS and SS mice homozygous for each of the human band 3 cytoplasmic domain sequences. Using an assay in which SS red cells in phosphate buffer are deoxygenated to 6% oxygen over time (Dunkelberger et al., J. Phys. Chem. B 122, 2018), we observed that -Hb band 3/SS mice showed an accelerated rate of sickle cell formation and a higher percent of sickled cells than wild type band 3/SS mice (p<0.01). Conversely, ++Hb band 3/SS mice showed an inhibition of both the rate of sickling and the precent of sickled cells compared to wild type band 3/SS mice (p<0.05).

We hypothesized that the inability of the glycolytic enzymes to reversibly bind to band 3 in the mutant mice were responsible for the differences in sickling. To test this hypothesis, we analyzed a panel of 28 cellular metabolites in 12 mice (6 female, 6 male) of each genotype: wild type band 3/AA, -AS and -SS, -Hb band 3/AA, -AS, -SS and ++Hb/AA, -AS, -SS. The metabolites were quantified by LC-MS/MS using an API 4500 triple quadrupole mass spectrometer (AB Sciex), with chromatographic resolution enabled on a polymeric amino column (apHera by Supelco) under alkaline mobile phase conditions (pH ~9.3). Stable isotope dilution and 8pt calibration curves allowed the absolute quantification of each metabolite.

Consistent with the constitutive binding of the terminal glycolytic enzymes to band 3 in -Hb erythrocytes, glycolysis was inhibited after the phosphoenol pyruvate step, as evidenced by significant accumulation of the intermediates at top of the glycolysis pathway, including fructose 1,6 biphosphate (FBP; p<0.01), dihydroxyacetone phosphate/ glyceraldehyde-3-phosphate (G3P; p<0.01), and 3-phosphoglycerate/2-phosphoglycerate (PG; p<0.01). In the ++Hb mutant where the terminal glycolytic enzymes are constitutively displaced from band 3, significantly lower levels of FBP, G3P and PG were observed (p<0.01). The levels of these metabolites in wild type band 3/SS erythrocytes were intermediate between the two mutant strains.

We hypothesized that the accumulation of FBP, G3P and PG contributed to the increased rate of sickling in the -Hb band 3/SS mice. To test this, we incubated wild type band 3/SS cells with either FBP or PG. Both intermediates increased the rate of sickle cell formation and percentage of sickled cells in a dose dependent fashion with no alteration in any RBC indices including MCV and osmotic fragility. We next hypothesized that reduction of the levels of glycolytic intermediates would have an antisickling effect. To test this, we incubated wild type band 3/SS cells with 2,3 diphosphoglycerol (DPG), which is a potent inhibitor of glycolysis. We found that DPG treatment led to a dose dependent decrease in the rate of sickle cell formation and percentage of sickled cells, again with no alteration in any RBC indices including MCV and osmotic fragility.

We conclude that the accumulation of glycolytic intermediates leads to increased sickle cell formation. We propose that reduction in the levels of glycolytic intermediates either by accelerating the terminal stages of glycolysis or by redirection to the pentose phosphate pathway may offer a means to treat SCD.