Erythroferrone exacerbates iron overload and ineffective extramedullary erythropoiesis in a mouse model of β-thalassemia

Abstract

β-thalassemia is characterized by chronic hepcidin suppression and iron overload, even in patients who have not undergone transfusion. The HbbTh3/+ (Th3/+) mouse model of nontransfusion-dependent β-thalassemia (NTDBT) partially recapitulates the human phenotype but lacks chronic hepcidin suppression, progressive iron accumulation into adulthood, or the interindividual variation of the rate of iron loading observed in patients. Erythroferrone (ERFE) is an erythroid regulator that suppresses hepcidin during increased erythropoiesis. ERFE concentrations in the sera of patients with NTDBT correlate negatively with hepcidin levels but vary over a broad range, possibly explaining the variability of iron overload in patients. To analyze the effect of high ERFE concentrations on hepcidin and iron overload in NTDBT, we crossed Th3/+ mice with erythroid ERFE-overexpressing transgenic mice. Th3/ERFE-transgenic mice suffered high perinatal mortality, but embryos at E18.5 showed similar viability, appearance, and anemia effects as Th3/+ mice. Compared with Th3/+ littermates, adult Th3/ERFE mice had similarly severe anemia but manifested greater suppression of serum hepcidin and increased iron accumulation in the liver, kidney, and spleen. The Th3/ERFE mice had much higher concentrations of serum ERFE than either parental strain, a finding attributable to both a higher number of erythroblasts and higher production of ERFE by each erythroblast.Th3/+ and Th3/ERFE mice had similar red blood cell count and shortened erythrocyte lifespan, but Th3/ERFE mice had an increased number of erythroid precursors in their larger spleens, indicative of aggravated ineffective extramedullary erythropoiesis. Thus, high ERFE concentrations increase the severity of nontransfusional iron overload and ineffective erythropoiesis in thalassemic mice but do not substantially affect anemia or hemolysis.

Graphical abstract

Figure 1.

Impaired postnatal survival of T-E(M) mice. Th3/+ (Th3) mice were bred with ERFE-overexpressing Line-M (E[M]) mice. (A) Expected and observed percentages of generated offspring with specified genotypes at either 3 weeks after birth or in utero at embryonic day E18.5. Total numbers are in parentheses next to percentages. (B-H) Pups were harvested at E18.5, and blood and fetal livers were collected for analysis. Blood: (B) RBC counts, (C) Hb level, (D) hematocrit, (E) MCV. Fetal liver: (F) Erfe mRNA, (G) Hamp mRNA, and (H) nonheme iron concentrations. Statistics: (A) significant differences from expected proportions were assessed using χ² test (∗∗∗∗P < .0001). (B-H) P values were assessed using two-tailed unpaired t test (∗∗∗∗P < .0001) between the Th3 and T- E(M) groups only.

Figure 2.

ERFE overexpression does not alter blood erythrocyte parameters in adult thalassemic mice. Mice generated from Th3 × E(M) breedings were analyzed at 16 weeks. Measurements in males are depicted in closed circles and females in open circles. (A-C) Erythroferrone levels were assessed by (A) BM Erfe mRNA, (B) spleen Erfe mRNA levels, (C) serum ERFE. (D) Serum EPO. (E-J) Erythrocyte parameters: (E) Hb levels, (F) RBC count, and (G) hematocrit. (H-J) Body and organ weights: (H) total body weight, (I,J) liver and spleen weights as a percentage of total body weight. P values were assessed by two-tailed unpaired t test (∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001) between the Th3 and T-E(M) groups only and (A-C) also between E(M) and T-E(M).

Figure 3.

ERFE overexpression exacerbates iron overload in thalassemic mice. Mice generated from Th3 × E(M) breedings were analyzed at 16 weeks of age. Males are depicted as closed circles and females as open circles. (A-E) Nonheme iron concentrations were assessed in (A) serum, (B) liver, (C) kidney, and (D) spleen: (E) total splenic nonheme iron, (F) iron visualized in formalin-fixed paraffin-embedded sections sections of the kidney (top), liver (middle) and spleen (bottom) from male 16-week-old mice using diaminobenzidine-enhanced Perls’ stain (kidney and liver) and Perls’ stain (spleen). Original magnification ×10 (F). Scale bars represent 200 μm (F). (A-F) P values were assessed by two-tailed unpaired t test (∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001) between the Th3 and T-E(M) groups only.

Figure 4.

ERFE overexpression further suppresses hepcidin levels in adult thalassemic mice. Mice generated from Th3 × E(M) breedings were analyzed at 16 weeks. Data for males are depicted in closed circles and for females in open circles. Liver hepcidin Hamp mRNA (A) and serum hepcidin protein concentration (B) are shown and also expressed as a ratio to liver nonheme iron concentration (LIC) to determine the appropriateness of hepcidin levels relative to iron status (C,D). P values were assessed by two-tailed unpaired t test (∗∗P < .01; ∗∗∗∗P < .0001) between the Th3 and T-E(M) groups only.

Figure 5.

Increased density of erythroid cells in erythropoietic organs of Th3 and T-E(M) mice contributes to enhanced ERFE production. Mice generated from Th3 × E(M) breeding were analyzed at 16 weeks. Males are depicted as closed circles and females as open circles. Erythroid precursor cell density was estimated in (A-C) BM and (D-F) spleen (extramedullary) by measuring (A,D) Gypa mRNA and (B,E) Tfrc mRNA. (C,F) Correlation between Gypa mRNA and Erfe mRNA in (C) BM and (D) spleen of E(M) and T-E(M) mice. (C,F) P values and R² were determined using Pearson correlation analysis.

Figure 6.

ERFE overexpression augments ineffective extramedullary erythropoiesis in older adult thalassemic mice. (A-B) Mice generated from Th3 × E(M) breeding were injected retroorbitally with NHS-Sulfo-Biotin 4 times in 48 hours, and then ∼50 μL weekly cheek bleeds were performed to obtain peripheral blood for analysis. Erythroid cells were analyzed via flow cytometry using PE-conjugated–anti-TER119 to identify erythroid lineage, APC-conjugated streptavidin to detect biotinylation, and TO-1 to detect RNA-containing reticulocytes. (A) Percentage of biotinylated erythrocytes over the course of 5 weeks (color scheme same as that in panels B-F). (B) Percent reticulocytes at terminal bleed defined as TER119⁺; TO-1⁺ cells relative to TER119⁺ cells. (C-F) Erythroid cells in BMs and spleens were harvested from mice at ∼37 to 45 weeks. Dead cells, and CD45⁺ cells, were excluded by 7-AAD and CD45 staining, and live erythroid cells from BM (C) or spleen (D) were counted via flow cytometry as live TER119⁺ cells. (E) Spleen weight as the percentage of body weight (F) Total splenic erythroid cells indices adjusted for body size were estimated by multiplying the values in panel D with the relative spleen size in panel E. (B-E) P values were assessed using two-tailed unpaired t test (∗ P < .05; ∗∗ P < .005) between the Th3 and T-E(M) groups only. N = 3 for each group.