Iron Binding to Specific Lobes of Transferrin Differentially Affects Erythropoiesis and Hepcidin Signaling

Introduction: Transferrin (Tf) plays a role in iron homeostasis as both an iron cargo molecule and as a signaling molecule. Tf is comprised of two homologous lobes (N and C), each of which can bind one iron molecule. As such Tf circulates as four forms: diferric, monoferric C, monoferric N and apo. Upon iron binding, clefts in each lobe close resulting in a conformational shift in the structure of the molecule and altering its affinity for the transferrin receptors. The most abundant molecules in the circulation at normal transferrin saturations are the monoferric forms. The relative abundance of circulating monoferric N in healthy individuals is greater than C, and greater with increased iron status. The N and C lobes differ in several physical properties, including binding affinities for iron. Whether these differences have functional consequences has been a subject of controversy. In vitro data suggest that iron is released initially from the C-lobe during delivery from diferric Tf to the erythron. These observations led to the proposal that iron occupancy of the N lobe might serve as an indicator of erythroid iron sufficiency, and as a signal to regulate dietary iron absorption. We thus hypothesized that iron homeostasis would be influenced by the presence of iron on the transferrin C-lobe vs the N-lobe.

Methods: To test this hypothesis, we generated mice producing exclusively monoferric N or C transferrins by mutating two essential tyrosines in the iron binding site of each of the lobes to phenylalanine (Y114 and Y207 in the N-blocked (N-bl) mouse and Y448 and Y537 in the C-blocked (C-bl) mouse). Heterozygous mice were crossed and offspring analyzed for hematological parameters, tissue iron concentrations, and expression by RT-PCR of selected genes relevant to iron homeostasis.

Results: The N-bl and C-bl Tf mutant mice each demonstrate hepatic iron loading; however, N-bl mice demonstrated higher non-heme liver iron concentrations than C-bl Tf mutant mice (1542 +/- 89.48 vs 1089 +/- 110.6 ug Fe/g dry weight, respectively; p <0.0001) compared to wild-type (261.8 +/- 22.17 ug Fe/g dry weight; p < 0.0001) at 2 months of age. Liver hepcidin (Hamp1) expression was decreased in both Tf mutants and was significantly lower in the N-bl mutant mice compared to wild-type (1.18 +/- 0.15). Bmp6 expression was not different across strains. Bone marrow Erfe expression was increased ~4-fold in each strain relative to wild-type (p < 0.05). Both strains demonstrate a mild microcytic anemia with hemoglobin concentrations of 12.11 +/- 0.17 and 12.87 +/- 0.24 and MCV values of 42.42 +/- 0.15 and 40.24 +/- 0.13 for N-bl and C-bl mice, respectively, compared to wild type hemoglobin concentrations of 14.98 +/- 0.11 and MCV values of 53.16 +/- 0.11 ( p < 0.01 across all categories). Red blood counts are significantly higher in C-bl mutant mice (9.98 +/- 0.12) than in N-bl mutant mice (9.16 +/- 0.09; p <0.001) or controls (8.91 +/- 0.06; p < 0.0001). In contrast, serum erythropoietin (Epo) is significantly higher in N-bl (503 +/- 50.18; p < 0.0001) compared to C-bl (273 +/- 29.5) and wild-type mice (221.6 +/- 10.61; p < 0.0001), resulting in a RBC:Epo ratio of 0.019 +/- 0.003 for N-bl mice, 0.04 +/- 0.003 for C-bl mice and 0.04 +/- 0.001 for wild-type mice (p <0.0001 for N-bl compared to C-bl and wt).

Conclusions: N-bl and C-bl Tf mutant mice demonstrate a hemochromatosis phenotype consistent with decreased hepcidin signaling; however, severity differs. Both mouse models demonstrate iron restrictive erythropoiesis; however, erythrocyte count is suppressed relative to erythropoietin levels only in the N-bl mutant mice. The N and C lobes of transferrin thus demonstrate functionally distinct properties. We speculate that iron bound to the N-lobe of transferrin augments the signaling properties of transferrin which regulate hepcidin expression and erythropoiesis.