Maternal enrichment during pregnancy accelerates retinal development of the fetus

Abstract

The influence of maternal environment on fetal development is largely unexplored, the available evidence concerns only the deleterious effects elicited by prenatal stress. Here we investigated the influence of prenatal enrichment on the early development of the visual system in the fetus. We studied the anatomical development of the rat retina, by analyzing the migration of neural progenitors and the process of retinal ganglion cell death, which exerts a key role in sculpturing the developing retinal system at perinatal ages. The number of apoptotic cells in the retinal ganglion cell layer was analyzed using two distinct methods: the presence of pyknotic nuclei stained for cresyl violet and the appearance of DNA fragmentation (Tunel method). We report that environmental enrichment of the mother during pregnancy affects the structural maturation of the retina, accelerating the migration of neural progenitors and the dynamics of natural cell death. These effects seem to be under the control of insulin-like growth factor-I: its levels, higher in enriched pregnant rats and in their milk, are increased also in their offspring, its neutralization abolishes the action of maternal enrichment on retinal development and chronic insulin-like growth factor-I injection to standard-reared females mimics the effects of enrichment in the fetuses. Thus, the development of the visual system is sensitive to environmental stimulation during prenatal life. These findings could have a bearing in orienting clinical research in the field of prenatal therapy.

Figure 1. (A) Accelerated natural cell death in the RGC layer of EC rats.

RGC layer apoptotic cell number in EC and SC rats, analyzed at the indicated ages with the Tunel method (top) and with cresyl violet staining of whole-mount retinas (bottom). With both methods, two-ways ANOVA showed an effect of age (p<0.001) and housing condition (p<0.05) and a significant age×housing condition interaction (p<0.001). Mann-Whitney rank sum test with Bonferroni correction revealed a difference between EC and SC at E18, E20 and P1 (p<0.001) for the tunel method, and at E18 and P1 (p = 0.002) for cresyl violet staining. (B) RGC number was not different between SC and EC adult rats either as estimated by calculating the 50% of total cell number in the RGC layer (A), or by subtracting the number of displaced amacrine cells remaining in the RGC layer 30 days after ipsilateral optic nerve transection from the number of cells counted in contralateral retinas (B) (p = 0.77 and 0.28, t-test). (C) Micrographs of RGC layer of P1 whole mount retinas labeled with B4 isolectin. No qualitative difference was detected in the shape and intensity of microglial cells between SC and EC pups. Scale bar: 20 µm. Graph: microglial cell number in the RGC layer of SC and EC rats. Mann-Whitney rank sum test showed no difference between the two groups (p = 0.429). Bars indicate s.e.m.

Figure 2. Anticipated migration of retinal neural progenitors in EC fetuses.

(A, top) Micrographs of EC and SC retinal sections immunostained for double-cortin (DCX) at E15 and E18. DCX expression was increased in EC rats at E15. (A, bottom) Quantitative analysis of DCX immunofluorescence intensity in the outer retinal layers of SC and EC rats. Two-ways ANOVA showed a statistical interaction between animal age and housing condition (p = 0.02). A Pairwise Multiple Comparison Procedure (Holm-Sidak method) revealed that the two groups were statistically different at E15 (p = 0.03), but not at E18 (p = 0.19). (B, top) Micrographs of EC and SC retinal sections immunostained for double-cortin (DCX) at E15, acquired at 5× (left, scale bar: 100 µm) or 20× (right, scale bar: 50 µm) magnification to count the number of migrating cells. (B, bottom) Number of cells stained for DCX in the outer retinal layers of SC and EC rats. The number of DCX-labeled cells was higher in EC than in SC embryos (Mann-Whitney rank sum Test, p<0.05). Bars indicate s.e.m.

Figure 3. Identification of specific cell types involved in the accelerated migration of retinal cells in EC fetuses.

(A, top) Micrographs of EC and SC retinal sections immunostained for ISLET-1 (a marker for ganglion and cholinergic amacrine cells) at E15, acquired at 5× (left, scale bar: 100 µm) or 20× (right, scale bar: 50 µm). (A, bottom) Number of cells stained for ISLET-1 in the outer retinal layers of SC and EC rats. The number of ISLET-1-labeled cells was higher in EC than in SC embryos (t-test, p<0.001). Bars indicate s.e.m. (B) Micrographs of EC and SC retinal sections immunostained for calbindin (a marker for horizontal cells) at E15, acquired at 20× (scale bar: 50 µm). The number of calbindin-labeled cells did not differ between EC and SC embryos (t-test, p = 0.253). (C) Micrographs of EC and SC retinal sections co-immunostained for DCX (red) and ISLET-1 (green) at E15, acquired at 60x. Scale bar: 10 µm.

Figure 4. (A) Increased IGF-I concentration in the maternal milk.

RIA determination of IGF-I concentration in the milk of SC and EC suckling pups: two-ways ANOVA showed a significant age×housing condition interaction (p<0.05). Post-hoc Tukey test revealed a difference at P1 (p<0.05), but not at P10 (p = 0.258) between EC and SC groups. Bars indicate s.e.m. (B–C) Enhanced IGF-I expression in the RGC layer of EC rats. (B) Micrographs of EC and SC retinal sections immunostained for IGF-I at different ages. Scale bar: 50 µm (C) Quantitative analysis of IGF-I immunofluorescence intensity in the RGC layer of SC and EC rats. Two-ways ANOVA showed an effect of age (p<0.001) and housing condition (p<0.001). t-test with Bonferroni correction revealed a statistical difference between EC and SC groups at E15 (p = 0.009) and E18 (p<0.01).

Figure 5. IGF-I is the mediator of maternal enrichment effects on retinal development in the fetus.

Number of DCX (A) and of ISLET-1 (B) positive cells in the outer retinal layers of EC, anti-IGF-I EC, SC and IGF-I SC rats at E15. For both (A) and (B), one-way ANOVA showed an effect of the treatment (p<0.05). A difference was found between EC and SC, between EC and EC anti-IGF-I and between SC and SC IGF-I groups (p<0.05, Post-hoc Tukey test). Neither EC anti-IGF-I and SC groups nor EC and SC IGF-I groups were instead found to differ between each other. (C) Quantitative analysis of IGF-I immunofluorescence intensity in the RGC layer of EC, anti-IGF-I EC, SC and IGF-I SC rats at E18. (D) Pyknotic cell number of EC, anti-IGF-I EC, SC and IGF-I SC rats at E18. After treatment with anti-IGF-I, levels of IGF-I expression and number of pyknotic profiles in the RGC layer of EC fetuses were lowered to those of SC rats while, after chronic IGF-I protein infusion, levels of IGF-I expression and the number of pyknotic profiles in the RGC layer of SC fetuses were enhanced up to those of EC rats. For both (C) and (D), one-way ANOVA showed an effect of the housing treatment (p<0.001). A difference was found between EC and SC, between EC and EC anti-IGF-I and between SC and SC IGF-I groups (p<0.05, Post-hoc Tukey test). Neither EC anti-IGF-I and SC groups nor EC and SC IGF-I groups were instead found to differ between each other. The bars indicate s.e.m.

Figure 6. (A) Increased IGF-I levels in the cerebellum of EC rats.

Coronal sections through the cerebellum: IGF-I immunoreactivity is low in cerebellar cells of SC rats, while the cerebellar cells of EC rats show a strong IGF-I staining. Scale bar: 50 µm. Graph: quantitative analysis of the pixel intensity of IGF-I immunofluorescence reactivity showed higher levels in the cerebellum of EC (black) compared with SC (grey) rats at P1 (Mann-Whitney rank sum test, p<0.05). The bars indicate s.e.m. (B) Prenatal enrichment increases body weight. The weight of EC fetuses was 10% greater at E18 (n = 19 for EC and n = 15 for SC), and 8% greater at P1 (n = 64 for EC and n = 49 for SC). Two ways ANOVA of rat weights for different environmental conditions and ages showed a significant effect of age (p<0.001) and environmental housing condition (p<0.001). Mann-Whitney rank sum test with Bonferroni correction revealed a significant increase in EC animals body weight compared with SC rats at both E18 and P1 (p<0.001 in both cases).