A transient window of hypothyroidism alters neural progenitor cells and results in abnormal brain development

Abstract

Cortical heterotopias are clusters of ectopic neurons in the brain and are associated with neurodevelopmental disorders like epilepsy and learning disabilities. We have previously characterized the robust penetrance of a heterotopia in a rat model, induced by thyroid hormone (TH) disruption during gestation. However, the specific mechanism by which maternal TH insufficiency results in this birth defect remains unknown. Here we first determined the developmental window susceptible to endocrine disruption and describe a cellular mechanism responsible for heterotopia formation. We show that five days of maternal goitrogen treatment (10 ppm propylthiouracil) during the perinatal period (GD19-PN2) induces a periventricular heterotopia in 100% of the offspring. Beginning in the early postnatal brain, neurons begin to aggregate near the ventricles of treated animals. In parallel, transcriptional and architectural changes of this region were observed including decreased Sonic hedgehog (Shh) expression, abnormal cell adhesion, and altered radial glia morphology. As the ventricular epithelium is juxtaposed to two sources of brain THs, the cerebrospinal fluid and vasculature, this progenitor niche may be especially susceptible to TH disruption. This work highlights the spatiotemporal vulnerabilities of the developing brain and demonstrates that a transient period of TH perturbation is sufficient to induce a congenital abnormality.

Figure 1

Perinatal exposure to an anti-thyroid agent is sufficient for cortical heterotopia formation. (a) The heterotopia is a permanent malformation. Light micrographs depict coronal brain sections from adult male rats born to either control (euthyroid, left) or PTU treated dams (hypothyroid, right). The heterotopia is visualized by Neuronal Nuclei (NeuN) immunostaining (red arrow). (b) The experimental design for determining the developmental period necessary for heterotopia formation. Timed pregnant dams were treated with 10 ppm PTU via the drinking water over one of three windows (GD9-GD14, GD14-GD19, and GD19-PN2). A positive control group was also included (dams treated GD9-PN2). Control animals received plain drinking water only. Offspring were analyzed on PN14 for heterotopia presence and severity. (c) Representative sections on posterior forebrain on PN14. Only animals from the positive control (GD9-PN2) and perinatal (GD19-PN2) treatments possessed this birth defect. The numbers on each panel represent the number of animals screened with a heterotopia; one male and one female pup was analyzed per litter. Scale bar represents 200 µm. (d) As there is no sex difference in heterotopia formation, the calculated heterotopia volume from the male and female littermates were averaged before analysis. The dashed line represents the minimum criterion for a neuron cluster to be considered a heterotopia (volume ≥ 0.006³ mm). Pups born to dams treated from GD9-PN2 and GD19-PN2 are significantly different than controls. (e) Analysis of the number of hemispheres (both right and left) in which ectopic neurons were observed in the heterotopia forming region across a 1.8 mm anterior-posterior interval of the brain. In panels (d) and (e) error bars represent ± SEM and asterisks p < 0.05. N = 3–6 litters analyzed.

Figure 2

Maternal PTU treatment during the perinatal period did not induce overt developmental toxicity. (a) Body weight in dams treated from GD19-PN2 with 10 ppm PTU was not significantly altered as compared to controls. (b) Litter size (recorded on PN2) was similarly unaffected. (c) Male and (d) female pup body weight was slightly, but not significantly, reduced during the first two postnatal weeks. In all analyses error bars represent ± SEM and asterisks p < 0.05. N = 5–6 litters per treatment analyzed.

Figure 3

Five days of maternal PTU treatment reduces serum T4 and T3 in both dams and pups. (a) Dams treated with PTU from GD19-PN2 exhibited a significant decrease in serum T4 and (b) T3 on PN2. (c) Pup serum T4 was significantly reduced in pups on PN0, PN2, and PN6. (d) Pup serum T3 was also significantly reduced on PN2; however, serum T3 was significantly elevated in treated pups as compared to controls on PN6. PN0 pup serum was sampled, but T3 concentrations were below the lower limit of quantification (LLOQ). (e) Pup brain T4 was significantly reduced in treated animals on PN0, 2 and 6. (f) Pup brain T3 concentrations, in comparison, were reduced in treated pups on PN0 and 2. However on PN6, T3 levels were not significantly different from controls. In all panels dashed lines denote the LLOQ for serum (0.1 ng/ml for T4 and 10 ng/dl for T3) and brain tissue (0.1 ng/g for both T4/T3). In all analyses error bars represent ± SEM and asterisks p < 0.05. N = 5–7 analyzed.

Figure 4

The heterotopia is periventricular. (a) Representative posterior forebrain sections in pups born to dams treated from GD19-PN2 (treated) and control animals. Coronal sections are stained with NeuN to detect mature neurons, and counterstained with Nissl. The cortical heterotopia reproducibly forms directly medial to the ventricular epithelium (V.E.), and superior to the subiculum (Sb). On PN6 aggregates of cells are observed in the treated animal in the heterotopia forming region (Hfr, black arrow); by PN8 NeuN+ cells begin to condense in the treated animal, representing the early neuronal heterotopia (Ht, black arrow). On PN14, the periventricular heterotopia is readily observable. (b) When PTU exposed dams are pulsed with the thymidine analog EdU on GD18–19, EdU labeling is detected in the pup heterotopia (Ht) on PN14 (white arrow, treated animal). Other cells born on GD18 and 19 contribute to the layer II of the neocortex (Ctx-II), the hippocampus (Hp), and the ventricular epithelium (V.E) of both treated and control animals. Scale bar represents 500 µm. (c) A portion of these late-born cells contributing to the heterotopia are neurons, as observed by co-labeling of EdU and NeuN. Scale bars represent 50 µm.

Figure 5

Sonic hedgehog (Shh), a direct TH target, is reduced in the postnatal brain following perinatal PTU exposure. (a) On PN2, a stage where T4/T3 are reduced in brain tissue, several genes associated with thyroid signaling are significantly downregulated. This includes Shh, a direct T3 target implicated in brain morphogenesis and stem cell maintenance. (a’) On PN6, Shh is also downregulated. However, eight more genes were differentially expressed in the posterior forebrain including sprouty related EVH domain 1 (Spred1). Downregulation of Spred1 in the perinatal mouse brain is associated with periventricular heterotopia formation. Asterisks in the gene expression data represent p < 0.001. N = 6–7 were analyzed at both stages. (b) SHH protein is normally expressed within the early postnatal ventricular epithelium (V.E.), hippocampus, and choroid plexus (C.P.) in control animals on PN2; scale bars represent 500 µm. (c) As we detected a downregulation in Casp3 gene expression in the treated brain (see a’), we assayed for potential differences in cleaved-CASP3 by immunohistochemistry on PN6. PTU treated pups exhibited a significant decrease (p < 0.05) in the percentage of labeled cleaved-CASP3 cells in the heterotopia forming region. N = 6 were assayed. (d) No differences in the percent labeling of Ki67 positive cells were detected in the ventricular epithelium on PN2 or (d’) on PN6. In all panels error bars represent ± SEM and N = 4–7 were analyzed.

Figure 6

Radial glial progenitors and the ventricular neuroepithelium are altered in TH insufficient pups. (a) Radial glial progenitor cells in control and PTU treated animals on PN6. In control animals, the normal apico-basal polarity of these cells is observed. However, in the animals treated with PTU from GD19-PN2, radial glial cells appear truncated and highly disorganized in the heterotopia forming region (Hfr), the cortex (Ctx), and near the subiculum (Sb). The extent of these abnormalities is further observed at higher magnification in the Hfr. (b) Adherens junctions were visualized by N-Cadherin expression on PN2 and PN6 in both control and treated animals. On PN2 the slight changes in the apical localization of N-Cadherin are observed; misexpression of this marker is more pronounced on PN6. Confocal images were taken at the ventricle (Vn). (c) In control animals the stem/progenitor cell marker SOX2 is prominently expressed in the posterior ventricular epithelium on the day of birth (PN0). Visualization of blood vessel via PECAM-1 in this same section demonstrates the extensive vascular network that is normally present in the neonatal rat brain. High magnification images highlight the enrichment of the vasculature within the SOX2+ and Vimentin+ population. These cells are also in proximity to the Vn, which contains cerebrospinal fluid (CSF). Together, this progenitor niche resides at the intersection of TH transport (vasculature and CSF) in neonates. All scale bars represent 50 µm and N = 3–4 were analyzed.

Figure 7

Hypothesized working model of heterotopia formation induced by developmental TH insufficiency. In the euthyroid neonate cells of the lateral ventricular neuroepithelium are a tightly organized progenitor population, which includes radial glial (lilac) and maturing neuroependymal cells (dark purple). Radial glial cells physically anchor to the apical surface of the ventricular zone via adherens junctions and possess cilia that extend into the cerebrospinal fluid (CSF). The CSF contains thyroid hormones (pictured here as TH). In addition to residing near the CSF, the neuroepithelium is also highly vascularized (red), which represents a second source of THs. In the hypothyroid pup decreased THs, via the CSF and the vasculature, lead to abnormal adherens junctions and loss of radial glia polarity. The marked changes in the radial glia scaffolding leads to abnormal neuronal migration and consequently, the periventricular heterotopia.