Thyroid hormone action controls multiple components of cell junctions at the ventricular zone in the newborn rat brain

Abstract

Thyroid hormone (TH) action controls brain development in a spatiotemporal manner. Previously, we demonstrated that perinatal hypothyroidism led to formation of a periventricular heterotopia in developing rats. This heterotopia occurs in the posterior telencephalon, and its formation was preceded by loss of radial glia cell polarity. As radial glia mediate cell migration and originate in a progenitor cell niche called the ventricular zone (VZ), we hypothesized that TH action may control cell signaling in this region. Here we addressed this hypothesis by employing laser capture microdissection and RNA-Seq to evaluate the VZ during a known period of TH sensitivity. Pregnant rats were exposed to a low dose of propylthiouracil (PTU, 0.0003%) through the drinking water during pregnancy and lactation. Dam and pup THs were quantified postnatally and RNA-Seq of the VZ performed in neonates. The PTU exposure resulted in a modest increase in maternal thyroid stimulating hormone and reduced thyroxine (T4). Exposed neonates exhibited hypothyroidism and T4 and triiodothyronine (T3) were also reduced in the telencephalon. RNA-Seq identified 358 differentially expressed genes in microdissected VZ cells of hypothyroid neonates as compared to controls (q-values ≤0.05). Pathway analyses showed processes like maintenance of the extracellular matrix and cytoskeleton, cell adhesion, and cell migration were significantly affected by hypothyroidism. Immunofluorescence also demonstrated that collagen IV, F-actin, radial glia, and adhesion proteins were reduced in the VZ. Immunohistochemistry of integrin αvβ3 and isoforms of both thyroid receptors (TRα/TRβ) showed highly overlapping expression patterns, including enrichment in the VZ. Taken together, our results show that TH action targets multiple components of cell junctions in the VZ, and this may be mediated by both genomic and nongenomic mechanisms. Surprisingly, this work also suggests that the blood-brain and blood-cerebrospinal fluid barriers may also be affected in hypothyroid newborns.

Keywords: brain development; cell adhesion; cell junctions; cell migration; hypothyroidism; radial glia; thyroid hormone action; ventricular zone.

Figure 1

Serum thyroid hormones in dams and pups. (A) Dam serum T4 is reduced in dams on PN14, whereas (B) total T3 was not significantly different. (C) TSH was increased by 72% in PTU exposed dams on PN14. (D) Pup serum T4 and (E) T3 were reduced in pups on PN2. (F) TSH was increased in PN2 neonates. In all presented data, N=8 control and N=7 PTU exposed dams or litters were assayed. For all pup data, each biological replicate represents serum pooled from multiple littermates of mixed sex. The data were then analyzed by t-test and each graph shows the value of each biological replicate and the SEM; ****p ≤ 0.0001, *p ≤ 0.05, ns, not significant.

Figure 2

Telencephalon thyroid hormones in PN2 pups. (A) Brain T4 was significantly reduced on PN2, as was (B) T3. (C) Reverse T3 (rT3) was reduced in PTU exposed animals when comparing group means, but this was not statistically significant. In all presented data, N=8 control and N=7 PTU exposed female pups were assayed (1 pup/litter). The data were then analyzed by t-test and each graph shows the value of each biological replicate and the SEM; **p ≤ 0.001, *p ≤ 0.05, ns, not significant.

Figure 3

Laser capture microdissection and RNA-Seq of the neonatal ventricular zone (VZ). (A) Fluorescent immunostaining in control animals demonstrates that the ventricular epithelium (V.E.) of the posterior telencephalon is highly enriched for SOX2, demonstrating that this is the VZ (ventricular zone, progenitor cell niche) on postnatal day 2 (PN2). (B) Example of laser capture microdissection of the VZ. (C) Volcano plot of resulting RNA-Seq data obtained from microdissected VZ from PN2 male pups, with downregulated (pink) and upregulated (yellow) genes highlighted (q ≤ 0.05). (D) These differentially expressed genes (DEGs) were then visualized by a heatmap, which shows clear differences between the control (euthyroid) and PTU exposed (hypothyroid) neonates. Each column of the heatmap represents a biological replicate (N=7 control and N, 6 PTU exposed pups).

Figure 4

Statistically significant gene ontology (GO) analyses of the identified DEGs, as visualized by bubble plots. (A) Cell component GO reveals that the most statistically significant categories include intracellular, cytoplasm, and cell junction (q=0.00013 for each); this shows that most protein products DEGs will localize to these cellular compartments. (B) Biological processes analysis shows that cell component organization, nervous system development, cytoskeletal organization, and supramolecular fiber organization are amongst the most statistically significant processes (q≤0.0018), while more specific categories like cytoskeletal organization, biological adhesion, actin filament-based processes, were also identified. In both panels all categories listed are supported by q≤0.05.

Figure 5

Ingenuity pathway analysis of significant cellular pathways. In this diagram unique genes identified are the outside of the circle, and connected to statistically significant processes (p<0.001). Note the overlap of the biological processes highlighted in purple boxes, which are correlated to dynamics of the cytoskeleton and cytoplasm, cell adhesion, cell morphology, and cell migration.

Figure 6

Developmental hypothyroidism affects collagen IV (COL IV) and filamentous actin (F-actin) in PN2 neonates (A) Representative regions of the posterior telencephalon. The white bonding boxes represent the inferior and superior portions of the VZ, and represent where the images were captured. Note that the choroid plexus (C.P., blood-cerebrospinal fluid barrier) is in proximity to the VZ and thus often imaged simultaneously. (B) Collagen IV (COL IV) is an extracellular matrix protein that is critical to basement membrane maintenance in cerebral endothelial and in some epithelial cells, and was implicated by RNA-Seq. In control pups, COL IV is enriched in blood vessels (B.V.) of the inferior VZ, as well as the vascular component of the choroid plexus. Less pronounced staining is also observed in the more luminal area of the VZ. (C) In PTU exposed pups, there was a clear disorganization of COL IV in the blood vessels, the choroid plexus, and in the inferior VZ parenchyma (see arrows). (D) COL IV is expressed in the blood vessels and choroid plexus of the superior VZ. (E) In PTU exposed pups COL IV also appears reduced in the superior VZ. (F) Visualization of F-actin by phalloidin shows strong signal throughout the brain, including in the blood vessels of the inferior VZ. F-actin was also highly enriched surrounding the nuclei of cells (blue DAPI staining) in the VZ, as expected. (G) In the PTU exposed animals, staining appeared globally reduced, which was especially notable in the parenchymal cells surrounding the VZ as well as in blood vessels (see arrows). (H) F-actin is highly expressed in the superior VZ and its associated blood vessels in control animals. (I) The PTU-associated changes in F-actin are apparent in the VZ and the surrounding parenchna. In all panels VZ, ventricular zone; C.P., choroid plexus; Hp, hippocampus; B.V., blood vessel; Vn, ventricle and scale bars are 50 µm.

Figure 7

Radial glia cells and cell adhesion are affected in the newborn VZ. (A) Vimentin immunohistochemistry highlights radial glia cells. In control animals radial glia endfeet anchor these cells to the ventricular zone (VZ), as seen at high magnification (see asterisk). (B) In PTU exposed pups, there is a loss of the apicobasal polarity of radial glia, which is especially prominent at high magnification (see white arrows). Note that the high magnification image was rotated so that the VZ was viewed horizontally. (C) Visualization of adherens junctions by N-cadherin. N-cadherin expression is observed in the VZ, notably along the apical border which faces cerebrospinal fluid. (D) In PTU exposed pups on PN2 there is a disorganization of N-Cadherin, especailly along the apical border (see arrows). (E) Examination of tight junctions by claudin 5 (CLD5) and endothelial cell by PECAM-1. CLD5 is normally expressed in endothelial cells and is a critical compoenent of the blood-brain and blood-cerebrospinal fluid barriers. It is also expressed along the apical surface of the VZ, similar to N-Cadherin. (F) In hypothyroid (PTU exposed) neontates, there was a pronounced reduction and disorganization in CLD5 staining in the choroid plexus (see arrows). There was also subtle changes in CLD5 along the apical surface of the VZ. Interestingly, PECAM-1 was also affected in the choroid plexus, and at the more basal region of the VZ as it transitions to the SVZ (see arrows). In all panels VZ, ventricular zone; C.P., choroid plexus; Hp, hippocampus; B.V., blood vessel; Vn, ventricle, and scale bars are 50 µm.

Figure 8

Mediators of TH action in PN2 pups. All receptors are visualized by diaminobenzidine (DAB) staining; at low magnification DAB is used alone and high magnification images were taken from sections counterstained with Cresyl violet. (A) Integrin α_Vβ3 is expressed in the VZ of control and (B) PTU exposed pups. Signal was also detected in the hippocampus and choroid plexus. High magnification images of the inferior VZ (see bounding box in panel A) demonstrates enriched expression of integrin αVβ3 along the apical border of the VZ and in the choroid plexus (see arrows). (C) Thyroid receptor alpha 2 (TRα2) was detected in the VZ of both euthyroid control pups and (D) PTU exposed pups. High magnification images show TRα2 staining throughout the VZ and in the choroid plexus (arrows). (E) Thyroid receptor beta 1/2 (TRβ1/2) was also expressed in the VZ of both control and (F) PTU exposed pups, in similar brain compartments as integrin α_Vβ3 and TRα2. At high magnification TRβ1/2 is also expressed in throughout the VZ and in choroid plexus cells (arrows). In all panels VZ, ventricular zone; C.P., choroid plexus; Hp, hippocampus; B.V., blood vessel; Vn, ventricle and scale bars are 50 µm. Low magnification images are stained with DAB alone, and high magnification images counterstained with Cresyl violet.