Low-level gestational lead exposure increases retinal progenitor cell proliferation and rod photoreceptor and bipolar cell neurogenesis in mice

Abstract

Background: Gestational lead exposure (GLE) produces novel and persistent rod-mediated electroretinographic (ERG) supernormality in children and adult animals.

Objectives: We used our murine GLE model to test the hypothesis that GLE increases the number of neurons in the rod signaling pathway and to determine the cellular mechanisms underlying the phenotype.

Results: Blood lead concentrations ([BPb]) in controls and after low-, moderate-, and high-dose GLE were ≤ 1, ≤ 10, approximately 25, and approximately 40 µg/dL, respectively, at the end of exposure [postnatal day 10 (PND10)]; by PND30 all [BPb] measures were ≤ 1 µg/dL. Epifluorescent, light, and confocal microscopy studies and Western blots demonstrated that late-born rod photoreceptors and rod and cone bipolar cells (BCs), but not Müller glial cells, increased in a nonmonotonic manner by 16-30% in PND60 GLE offspring. Retinal lamination and the rod:cone BC ratio were not altered. In vivo BrdU (5-bromo-2-deoxyuridine) pulse-labeling and Ki67 labeling of isolated cells from developing mice showed that GLE increased and prolonged retinal progenitor cell proliferation. TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) and confocal studies revealed that GLE did not alter developmental apoptosis or produce retinal injury. BrdU birth-dating and confocal studies confirmed the selective rod and BC increases and showed that the patterns of neurogenesis and gliogenesis were unaltered by GLE.

Conclusions: Our findings suggest two spatiotemporal components mediated by dysregulation of different extrinsic/intrinsic factors: increased and prolonged cell proliferation and increased neuronal (but not glial) cell fate. These findings have relevance for neurotoxicology, pediatrics, public health, risk assessment, and retinal cell biology because they occurred at clinically relevant [BPb] and correspond with the ERG phenotype.

Figure 1

GLE selectively increases the number of rod photoreceptors and BCs in the adult (PND60) mouse retina as shown in light and confocal microscopy studies. (A) Representative DAPI nuclear labeling shows that ONL (rod and cone nuclei), INL (horizontal, bipolar, amacrine, and MGC nuclei), outer plexiform layer (OPL), inner plexiform layer (IPL), and total retinal thickness increased in retinas from animals in the LD, MD, and HD GLE groups. Retinal GCL cellularity and retinal lamination were not different in control and GLE retinas. (B–E) Representative double-labeled confocal microscopy studies reveal rod and BC selectivity of the retinal phenotype. (B) The number of rhodopsin-IR rod nuclei and PKCα-IR rod BC somas—in distal (upper) INL—increased in GLE retinas. (C) The numbers of cone outer segments (OS) immunoreactive for middle- and short-wavelength–sensitive opsin (M-/S-opsin) and horizontal cells immunoreactive for calbindin were not different in control and GLE retinas. (D) The numbers of Chx10-IR BC nuclei and Chx10/PKCα colabeled rod BCs increased in adult GLE retinas. (E) The number of cyclin D3–IR MGCs colabeled with glutamine synthetase (GS), were similar in adult control and GLE retinas. Bars = 20 μm for A and for B–E.

Figure 2

GLE selectively increases the number of rod photoreceptors and BCs in retinas from adult mice. (A) Unbiased stereological analyses of major retinal cell types in adult retinas show that LD, MD, and HD GLE produced selective and significant nonmonotonic increases in the number of rods and BCs, relative to controls. Values are mean ± SE from three nonadjacent sections per retina from four to seven retinas per treatment group, with each retina from a different mouse. (B) Representative Western blots reveal that GLE increased the retinal content of rhodopsin, Chx10, and PKCα but did not change the amount of glutamine synthetase (GS) or amacrine cell ChAT content. GAPDH was used as the protein loading control. *p < 0.05, compared with control. Groups sharing # or † were significantly different within the cell type at p < 0.05.

Figure 3

BrdU birth-dating and confocal studies demonstrate that MD GLE selectively and significantly increased the number of rods and BCs in retinas from adult mice. GLE did not alter the kinetics of neurogenesis and gliogenesis. BrdU-IR and rhodopsin-IR rods (A) and BrdU-IR and Chx10-IR BCs (B) increased in GLE retina, compared with controls. (C) No differences in the number of BrdU-IR and cyclin D3–IR MGCs were observed in control and GLE retinas. Values are mean ± SE of IR cells per 400 μm of central retina and and represent from four to seven retinas at each age per treatment group, with each retina from a different mouse. *p < 0.05, and **p < 0.01 compared with controls.

Figure 4

MD GLE increases and prolongs RPC proliferation in vivo and ex vivo. (A) The M-phase marker PH3 and S-phase marker BrdU are present until PND3 in controls and PND5 in GLE central retina. They colocalize in the distal NBL. At all ages, except PND7, there are more BrdU-IR and PH3-IR cells in the GLE retinas. Abbreviations: inbl, inner NBL; onbl, outer NBL. Bar = 20 μm. Stereological analysis of BrdU (B) and PH3 (C) labeling shows that GLE significantly increased and prolonged RPC proliferation. (D) Dissociated single cells from PND2, PND4, and PND6 control and MD GLE retinas were double labeled with an anti-Ki67 antibody and DRAQ5; results reveal that GLE significantly increased RPCs in an age-dependent manner, similar to that shown in B. In B–D, values are mean ± SE from four to seven retinas at each age per treatment group, with each retina from a different mouse; in B and C, values represent IR cells per 400 μm of central retina. *p < 0.05, and **p < 0.01 compared with control.

Figure 5

MD GLE does not change the amount or pattern of retinal apoptosis during development. (A) TUNEL-positive cells in control and GLE retinas from GD16.5 to PND10. Bar = 100 μm. (B) The number of TUNEL-positive cells increased rapidly in the NBL during development, however, there were no age-dependent differences between control and GLE mice. (C) Although the kinetics of TUNEL-positive cells differed slightly in the retinal GCL of control and GLE retinas, there was no difference in the number of retinal ganglion cells at PND60 (Figures 1A, 2A). (D) There were no differences in the number of TUNEL-positive cells in the ONL or INL between groups. In B–D, values are mean ± SE from four to seven retinas at each age per treatment group, with each retina from a different mouse. *p < 0.05 compared with control.