Temporal manipulation of transferrin-receptor-1-dependent iron uptake identifies a sensitive period in mouse hippocampal neurodevelopment

Abstract

Iron is a necessary substrate for neuronal function throughout the lifespan, but particularly during development. Early life iron deficiency (ID) in humans (late gestation through 2-3 yr) results in persistent cognitive and behavioral abnormalities despite iron repletion. Animal models of early life ID generated using maternal dietary iron restriction also demonstrate persistent learning and memory deficits, suggesting a critical requirement for iron during hippocampal development. Precise definition of the temporal window for this requirement has been elusive due to anemia and total body and brain ID inherent to previous dietary restriction models. To circumvent these confounds, we developed transgenic mice that express tetracycline transactivator regulated, dominant negative transferrin receptor (DNTfR1) in hippocampal neurons, disrupting TfR1 mediated iron uptake specifically in CA1 pyramidal neurons. Normal iron status was restored by doxycycline administration. We manipulated the duration of ID using this inducible model to examine long-term effects of early ID on Morris water maze learning, CA1 apical dendrite structure, and defining factors of critical periods including parvalbmin (PV) expression, perineuronal nets (PNN), and brain-derived neurotrophic factor (BDNF) expression. Ongoing ID impaired spatial memory and resulted in disorganized apical dendrite structure accompanied by altered PV and PNN expression and reduced BDNF levels. Iron repletion at P21, near the end of hippocampal dendritogenesis, restored spatial memory, dendrite structure, and critical period markers in adult mice. However, mice that remained hippocampally iron deficient until P42 continued to have spatial memory deficits, impaired CA1 apical dendrite structure, and persistent alterations in PV and PNN expression and reduced BDNF despite iron repletion. Together, these findings demonstrate that hippocampal iron availability is necessary between P21 and P42 for development of normal spatial learning and memory, and that these effects may reflect disruption of critical period closure by early life ID.

FIGURE 1. Generation of DNTfR1 transgene

(A) Schematic of DNTfR1. (B) DNTfR1 transgene contains a g1946a point mutation resulting in R649H substitution in the RGD Tf binding domain (Dubljevic et al., 1999). (C) tTA drives DNTfR1 expression from the TRE CMV promoter.

FIGURE 2. Loss of Tf binding capability in CHO cells expressing DNTfR1

Panels A–C) CHO cells transfected with high levels of eGFP-TfR (A, arrows) showing strong Tf binding (B, arrows). Merged image (C, arrows) showing extensive overlap of TfR-Tf binding in yellow. Panels D–F) CHO cells transfected with high levels of eGFP-DNTfR1 (D, arrows) showing minimal or no Tf binding (E, arrows). Trace Tf-binding is due to a small amount of residual endogenous TfR1 expression in CHO cells. Merged image (F, arrows) shows loss of TfR-Tf binding.

FIGURE 3. DNTfR1 expression selectively and reversibly inhibits iron uptake

in vivo. Perls’ iron staining in adult hippocampus and cortex from (A) WT^nodox (B) DN^nodox (C) WT^P21dox and (D) DN^P21dox mice; inset, magnification of CA1 pyramidal cell layer indicated by arrowhead, scale bar=200 μm. (E) Quantification of staining intensity. (F) Relative expression of DNTfR1 mRNA. Data are mean ± SEM; **p<0.01, ***p<0.001

FIGURE 4. Hippocampal ID impairs spatial learning

(A–C) Average daily escape latency for (A) WT^nodox and DN^nodox, (B) WT^P21dox and DN^P21dox, and (C) WT^P42dox and DN^P42dox animals. (D–F) Average daily percentage of time spent swimming in platform quadrant during training trials for (D) WT^nodox and DN^nodox, (E) WT^P21dox and DN^P21dox, and (F) WT^P42dox and DN^P42dox animals. (G–I) Average daily percentage of time spent swimming in the perimeter of MWM during training trials for (G) WT^nodox and DN^nodox, (H) WT^P21dox and DN^P21dox, and (I) WT^P42dox and DN^P42dox animals. (J–L) Percentage of time spent target quadrant during single daily MWM probe trail for (J) WT^nodox and DN^nodox, (K) WT^P21dox and DN^P21dox, and (L) WT^P42dox and DN^P42dox animals. Data are mean±SEM (n=10–14), # indicates p<0.05 effect of genotype by two-way ANOVA; + indicates p<0.05 effect of training by two-way ANOVA; *p<0.05 and ***p<0.001 indicate Bonferroni posthoc significance.

FIGURE 5. Hippocampal ID impairs CA1 apical dendrite morphology

CA1 apical dendrite morphology in (A) WT^nodox (B) DN^nodox (C) WT^P21dox (D) DN^P21dox (E) WT^P42dox and (F) DN^P42doxanimals. Note the abnormal dendrite morphology (arrows) in the groups represented in Panels B and F and the lack of such abnormalities in Panel D. Scale bar=200 μm

FIGURE 6. Timing of iron repletion differentially alters critical period markers and BDNF-V mRNA

Mean number of PV positive cells, PNNs and percent of PV positive cells with PNNs in CA1 of WT, DN^nodox and DN^P21dox mice at P30 (A, B) and in WT, DN^nodox, DN^P21dox, and DN^P42dox mice at P70 (C, D). (E) Relative expression of BDNF-V mRNA at P70. Data are mean ± SEM; *p<0.05, **p<0.01