Iron control of erythroid development by a novel aconitase-associated regulatory pathway

Abstract

Human red cell differentiation requires the action of erythropoietin on committed progenitor cells. In iron deficiency, committed erythroid progenitors lose responsiveness to erythropoietin, resulting in hypoplastic anemia. To address the basis for iron regulation of erythropoiesis, we established primary hematopoietic cultures with transferrin saturation levels that restricted erythropoiesis but permitted granulopoiesis and megakaryopoiesis. Experiments in this system identified as a critical regulatory element the aconitases, multifunctional iron-sulfur cluster proteins that metabolize citrate to isocitrate. Iron restriction suppressed mitochondrial and cytosolic aconitase activity in erythroid but not granulocytic or megakaryocytic progenitors. An active site aconitase inhibitor, fluorocitrate, blocked erythroid differentiation in a manner similar to iron deprivation. Exogenous isocitrate abrogated the erythroid iron restriction response in vitro and reversed anemia progression in iron-deprived mice. The mechanism for aconitase regulation of erythropoiesis most probably involves both production of metabolic intermediates and modulation of erythropoietin signaling. One relevant signaling pathway appeared to involve protein kinase Calpha/beta, or possibly protein kinase Cdelta, whose activities were regulated by iron, isocitrate, and erythropoietin.

Figure 1

A primary human model system for iron restriction of erythropoiesis. (A) Effect of transferrin saturation (Sat) on growth and viability of hematopoietic lineages. CD34⁺ cells underwent unilineage culture in serum-free media. Results show mean ± SEM for 3 to 5 independent experiments. (B) Inhibition of early erythroid differentiation by iron restriction. Erythroid cultures with 100% or 15% transferrin saturation underwent flow cytometry (fluorescence-activated cell sorting [FACS]) with gating on viable cells. CD34/CD36 expression was analyzed on day 4. Indicated are percentages of cells in various quadrants. (C) Minimal apoptosis associated with erythroid iron restriction. Cells from panel B were analyzed by FACS for 7-amino-actinomycin D (7-AAD) and annexin V staining with gating on all cells. Whole cell lysates from day 5 cultures underwent immunoblotting (IB) for caspase 3 (bottom; arrow indicates cleavage product). (D) Minimal effect of erythroid iron restriction on cell-cycle parameters. Day 5 cultures stained with propidium iodide underwent FACS. Indicated are percentages of cells in various fractions, including subdiploid (Sub).

Figure 2

Implication of aconitase activity in iron regulation of erythropoiesis. (A) Erythroid-specific aconitase inactivation during iron restriction. Enzymography (Acon Activity) was performed on extracts from day 4 cultures in erythroid (Ery), megakaryocytic (Mk), or granulocytic (Gran) media with 100% or 15% transferrin saturation; controls consisted of K562 cells expressing shRNAs targeting ACO2 or ACO1. Expression of mitochondrial aconitase protein in samples (bottom panel, IB: ACO2). (B) Summary of 3 independent experiments as in panel A. The y-axis is the ratio of scanned signals obtained with 15% transferrin saturation to those obtained with 100% transferrin saturation, expressed as mean percentage ± SEM. **P < .001 for Ery ACO1 changes and 0.03 for Ery ACO2 changes. Mk and Gran signals showed no significant changes. (C) Direct aconitase inhibition impairs erythroid differentiation in a manner similar to iron restriction. Erythroid cultures with 100% transferrin saturation were supplemented with fluorocitrate (FC) and underwent FACS assessment of differentiation (top panels) and viability (forward [FSC] and side scatter [SSC] in bottom panels) on day 5. Where indicated, 10mM citrate was also included in culture. (D) Summary of 3 independent experiments as in panel C, using 200μM fluorocitrate.

Figure 3

Isocitrate abrogation of the erythroid iron-restriction checkpoint. (A) Isocitrate (IC) reverses defects in differentiation and viability. Five-day erythroid cultures with 15% transferrin saturation included trisodium isocitrate (IC) at 5mM. FACS analysis of GPA and CD41 expression with gating on viable fraction. (B) Summary of 3 independent experiments as in panel A. (C) Isocitrate reverses structural changes associated with iron restriction. Light microscopy of Wright-stained cytospins from cultures in panel A (400× magnification). Images were acquired with the use of an Olympus BX51 microscope equipped with an Olympus DP70 digital camera. The objective lens consisted of Uplan Fl 40×/0.75 NA. Image acquisition and processing used Adobe Photoshop, CS3/10.0 and CS2/9.0, respectively. (D) Isocitrate enhances hemoglobinization. Erythroid cultures with 100% or 15% transferrin saturation were supplemented with 20mM isocitrate. Photograph of day 5 cell pellets. (E) Isocitrate augments globin chain expression. Whole cell lysates from panel D underwent immunoblotting. (F) IRE-binding activity of IRP1 and IRP2: influences of transferrin saturation and isocitrate. RNA gel-shift assays were performed on extracts from day 4 CD34⁺ cultures in erythroid medium with 100% or 15% transferrin saturation and 20mM isocitrate. (G) IRP2 stabilization by iron deprivation: minimal effect of isocitrate. Cellular extracts from panel F underwent immunoblotting. FSC indicates forward scatter; and SSC, side scatter.

Figure 4

Participation of a nonmetabolic isocitrate signaling pathway. (A) Cellular ATP, NADH, and NADPH levels in day 5 erythroid cultures. Shown are means ± SEMs for 3 independent experiments. (B) Viability effects of L- versus D-isocitrate in erythroid iron restriction. Erythroid cultures received 1mM, 5mM, and 10mM isocitrate (IC) enantiomers, with FACS analysis on day 5. Graph summarizes 3 independent experiments with the use of 10mM isocitrate. **P = .012 for D-IC + 15% versus 15% and 0.017 for L-IC + 15% versus 15%. No significant difference was seen for D-IC + 15% versus L-IC + 15%. (C) Effects of L- versus D-isocitrate on erythroid differentiation. FACS analysis of GPA and CD41 expression in cultures from panel B, with gating on viable fraction. **P = .04 for D-IC + 15% versus 15%; *P = .05 for L-IC + 15% versus 15%. In addition, P = .05 for D-IC + 15% versus L-IC + 15%.

Figure 5

Crosstalk of iron and isocitrate with Epo signaling: PKCα/β as common target. (A-B) Epo levels influence isocitrate rescue. FACS analysis of 5 day cultures with indicated doses of Epo plus 25 ng/mL SCF; 100% or 15% transferrin, and 20mM isocitrate (IC) were included as indicated. Graph summarizes results of 3 independent experiments. (C) Iron restriction and Epo deprivation both induce PKCα/β hyperactivation. Epo levels influence the capacity of isocitrate to reverse the hyperactivation. Whole cell lysates from 4 day erythroid cultures underwent immunoblotting. (D) Kinetics of PKCα/β hyperactivation associated with erythroid iron restriction. (E-F) Complete reversal of viability defects by pan-PKC inhibitor. Erythroid cultures with 100% or 15% transferrin were treated with 0.5μM BIM, with FACS analysis on day 5. Graph represents 3 independent experiments. (G) Partial reversal of differentiation defects with selective PKC inhibitor. Cultures as in panel E were treated with either 0.5μM BIM or with 0.5μM Gö6983. Graphs summarize 3 independent experiments. DN indicates GPA⁻ CD41⁻ double-negative cells. **P = .02 for DN percentage in BIM + 15% versus 15%. **P = .01 for GPA percentage in Gö6983 + 15% versus 15%.

Figure 6

Influence of isocitrate on erythropoiesis in vivo. Isocitrate (IC) inhibits anemia progression during iron deprivation. Peripheral red blood cell (RBC) number and hematocrit (HCT) in mice on iron-deficient diets. Treatment with isocitrate (dashed lines) versus either normal saline (NS) or α-ketoglutarate (αKG; solid lines) was initiated on days indicated by arrows. Each point represents mean ± SEM for 6 mice/group. **Experiment 1 day 54, isocitrate versus saline, P values are .007 and .001 for RBC number and HCT. ^##Experiment 2 day 45, isocitrate versus saline, P values are .015 and .018 for RBC number and HCT. †Comparison of isocitrate with α-ketoglutarate, day 56, P values are .044 and .046 for RBC number and HCT.

Figure 7

Model for erythropoietic regulation by an iron-aconitase-isocitrate pathway. In the absence of iron restriction (A; > 15% Transferrin Saturation) aconitase enzymes possess intact iron-sulfur clusters and function mainly in the conversion of citrate to isocitrate by the Metabolic Flux Pathway. In the presence of iron restriction (B; 15% Transferrin Saturation), destabilization of the aconitase iron-sulfur clusters induces assembly of a repressive signalosome which may act in part through PKC hyperactivation. In addition, diminished metabolic flux may compromise heme production and lead to shunting of citrate by activated ATP-citrate-lyase (P-ACL) to oxaloacetate (OAA) and acetyl-CoA. Isocitrate rescue (C; 15% Transferrin Saturation with Exogenous Isocitrate) may prevent assembly of a repressive signalosome by stabilization of aconitase iron-sulfur clusters and by binding to isocitrate dehydrogenase (IDH) enzymes. In addition, exogenous D- but not L-isocitrate may support heme biosynthesis. However, exogenous isocitrate does not prevent shunting of citrate by activated ATP-citrate-lyase.