Histological, functional and transcriptomic alterations in the juvenile hippocampus in a mouse model of thyroid hormone resistance.

Background: Proper thyroid hormone signaling via the TRα1 nuclear receptor is required for normal neurodevelopmental processes. The specific downstream mechanisms mediated by TRα1 that impact brain development remain to be investigated.
Methods: In this study, the structure, function and transcriptome of hippocampal tissue in a mouse model expressing the first RTHα mutation discovered in a patient, THRA E403X, were analyzed. RNAscope was used to visualize the spatial and temporal expression of Thra1 mRNA in the hippocampus of WT mice, which is corresponding to THRA1 mRNA in humans. The morphological structure was analyzed by Nissl staining, and the synaptic transmission was analyzed on the basis of long-term potentiation. The Morris water maze test and the zero maze test were used to evaluate the behavior. RNA-seq and quantitative real-time PCR were used to analyze the differentially expressed genes (DEGs) of the hippocampal tissues in the mouse model expressing the Thra E403X mutation.
Results: The juvenile mutant Thra E403X mice presented with delayed neuronal migration, disordered neuronal distribution, and decreased synaptic plasticity. A total of 754 DEGs, including 361 upregulated genes and 393 downregulated genes, were identified by RNA-seq. DEG-enriched Gene Ontology (GO) and KEGG pathways were associated with PI3K-Akt signaling, ECM-receptor interaction, neuroactive ligand-receptor interaction, and a range of immune-related pathways. 25 DEGs were validated by qPCR. Conclusions: The ThraE403X mutation results in histological and functional abnormalities, as well as transcriptomic alterations in the juvenile mouse hippocampus. This study of the ThraE403X mutant offers new insights into the biological cause of RTHα-associated neurological diseases.

Introduction

Thyroid hormones (THs, namely, the prohormone thyroxine T4 and its active form, triiodothyronine T3) play pivotal roles in diverse aspects of life, mainly during development, when they regulate skeletal growth, cardiac function, metabolism, and the nervous system (1, 2, 3). T3 acts predominantly by binding to thyroid hormone receptors (TRα and TRβ), which are encoded by two separate genes, THRA and THRB in humans and Thra and Thrbin mice (4).

Resistance to thyroid hormone α (RTHα) is a rare genetic disease caused by mutations in the THRAgene that was first described in 2012 (5). Using whole-exome sequencing, the patient was found to be carrying a heterozygous de novo mutation (c1207G→T, p.E403X) in THRA and presented intellectual disability in childhood that resembles congenital hypothyroidism (6). Although treatment with THs leads to improvements in certain clinical phenotypes of patients with RTHα, including constipation and problems with nerve conduction, the cognitive deficits typically remain (7). Mounting evidence suggests that as one of the most extensively investigated regions in the brain, the hippocampus mediates cognitive functions, notably, learning, and memory functions. In both animals and humans, the hippocampus is vulnerable to hypothyroidism, regardless of the time it begins or its duration (8, 9, 10).

We generated a novel knock-in mouse model (ThraE403X) with a phenotype that corresponds to that of the nonsense mutation derived from a patient with RTHα. Specifically, the mutant mice presented features of the THRAE403X/+ phenotype, including neurological and motor coordination deficits (11). Unfortunately, the mechanism by which RTHα causes these neurocognitive deficits has not been adequately characterized.

Here, we analyzed the spatiotemporal expression of Thra1 mRNA in the hippocampus of mouse and identified hippocampal morphological alterations manifesting in juvenile mice. Furthermore, we employed RNA-seq technology and bioinformatics analysis to identify dysregulated genes that may be associated with the ThraE403Xmutation. Finally, 25 candidate DEGs were selected for qPCR analysis. These findings represent an essential starting point for the comprehension of crucial molecules and mechanisms of cognitive deficiency caused by RTHα, especially during the early stage of the disease.

Materials and methods

Animal maintenance

ThraE403X mutant mice, and WT littermates were used in the present study. The ThraE403X mouse strain carries the dominant negative E403X mutation in the Thra gene, which has been described previously (11). The mice were housed at 22 ± 2°C for 12 h light:12 h darkness cycle with ad libitum access to water and rodent chow. ThraE403Xmutant mice and WT mice were obtained by crossing heterozygous mutant mice. Animals were tagged and randomly allocated to each group. All animal care and experimental procedures were approved by the Animal Care and Use Committee at China Medical University.

In situ hybridization

RNAscope, a new ultrasensitive in situ hybridization method (12), was used to detect Thra1mRNAs (NM_178060.4) in the hippocampus of mouse. The probes for Thra1 (nine synthetic oligonucleotides complementary to the nucleotide sequence 1836–2336 of Thra1) were provided by Advanced Cell Diagnostics (#531731, ACD, San Francisco, CA, USA), which only detect Thra1 and do not cross detect Thra2. The probes for the positive and negative controls were designed to detect the mouse housekeeping gene Ppib (#313911, ACD) and the bacterial gene DapB (#310043, ACD), respectively. Thra1mRNA was analyzed at embryonic day (E) 10.5, E14.5, E16.5, and 18.5 and postnatal day (P) 3, P7, P14, and P56. All mice were anesthetized with 3% pentobarbital sodium (2 mL/kg) and fixed by cardiac perfusion with PBS followed by 10% neutral buffered formalin before decapitation. Formalin-perfused, fixed mouse embryo (E10.5 and E14.5) and brain (E16.5, E18.5, P3, P7, P14, and P56) were dehydrated in gradient ethanol, transparent in xylene, embedded in paraffin and cut at a setting of 5 µm in the sagittal plane. The procedure for RNAscope hybridization was performed with an RNAscope® 2.5 HD Detection Kit (#322300, ACD) (12). The RNAscope is a DAB staining. All sections were observed with a light microscope (BX53, Olympus), and images were captured with software (CellSens Standard 3.2, Olympus).

Nissl staining

For Nissl staining experiments, male WT and male mutant mice at P0, P7, and P21 were used. The brain and sliced sections were processed as described in the manufacturer’s instructions for performing RNAscope hybridization. But all the brain tissues in this part were cut at a setting of 5 µm in the coronal plane. Nissl staining of brain slices was performed with 1% toluidine blue (Sigma-Aldrich Co.), an aniline dye that preferentially stains RNA and DNA, according to the standard procedure. All the sections were observed by the light microscope (BX53, Olympus) and the images were captured with a software (CellSens Standard 3.2, Olympus).

Electrophysiology

f-EPSPs of the mice were measured with the MED64 planar microelectrode matrix recording system (Alpha Med Science, Osaka, Japan). Acute hippocampal slices prepared for electrophysiological recordings were obtained from 3-week-old homozygous mutant and WT littermates that had been anesthetized with isoflurane and decapitated.

The fresh brain tissue was cut into 300-μm thick slices, removed, and placed into an oxygenated artificial cerebrospinal fluid containing 7.25 g/L NaCl, 0.22 g/L KCl, 2.18 g/L NaHCO3, 0.12 g/L MgSO4, 0.17 g/L KH2PO4, 1.8 g/L glucose, and 0.22 g/L CaCl2 (pH 7.35–7.45). The basic stimulus was set to induce 30% of the f-EPSP maximum, and a stable 30-min baseline was recorded. LTP was provoked by high-frequency stimulation (HFS) that elicited 50% of the maximal response. We stimulated the tissue twice with HFS at 10 s interval. Recordings were made at baseline intensity for an additional hour after LTP induction. f-EPSP amplitude and slope ratios were analyzed as the LTP quantification.

Animal behaviour

Morris water maze test (MWM)

MWM is a widely used behavioral procedure for the assessment of spatial learning and memory in small rodents. The MWM test apparatus consists of a 110 cm-diameter-wide black circular pool filled with opaque tap water maintained at 22°C with added non-toxic white dye. A circular hidden escape platform (10 cm in diameter) just beneath the surface was placed within the pool surrounded by four elevated visual navigation cues (colored geometric images) above each pool quadrant to provide spatial orientation (Shanghai XinRuan Information Technology Co., Ltd). The experiment was carried out according to the procedure of the mouse water maze experiment (13). EnthoVision XT Animal Trajectory Tracking System (Noldus, Netherlands) was used to determine the escape latency to find the platform (MWM training) and times of crossing of target quadrant and platform area in the spatial probe trial (MWM testing).

Elevated zero maze test

Anxiety-like behavior was measured using the elevated zero maze (EZM) (Shanghai XinRuan Information Technology Co., Ltd). It was a ring-shaped apparatus, elevated 40 cm from the floor, consisted with a circular platform (outer diameter 45 cm, width 5 cm). The circular platform was divided into four quadrants of equal lengths with two open arms (with a 0.5-cm-high curb to prevent falls) and two equal closed arms (surrounded by an 11-cm wall from the surface of the maze). Each mouse was placed in one closed arm of the maze facing one of the open arms, and 5 min of free exploration was recorded by the EnthoVision XT Animal Trajectory Tracking System (Noldus, Netherlands). The following events: the total distance of traveling, the percentage of time spent in the open arms, the number of entries into the open arms (an open arm entry is defined as the mouse’s four paws in an open arm), the latency to visit the open arms (the time of the mouse’s first visit to an open arm), the percentage of freezing time (the freezing time was defined as the mouse not moving or immobility), the number of rearing (rising on the hind paws) in the closed arms, and the number of head dipping (scanning over the side of the maze) in the open arms.

RNA sequencing

The gene expression profiling by RNA-seq and qPCR were performed in the hippocampus of four 3-week-old homozygous mutant mice and four WT control littermates. Total RNA was extracted following the standard RNAiso Plus protocol (Takara Bio, Japan), then qualified and quantified using the Agilent 2100 Bioanalyzer (Thermo Fisher Scientific). The first step in the workflow involves purifying mRNA molecules from total RNA using oligo(dT)-attached magnetic beads. Following purification, mRNA was fragmented into small pieces using a fragmentation reagent. First-strand cDNA was generated using random hexamer-primed RT, then was followed by second-strand cDNA synthesis. The synthesized cDNA was subjected to end-repair and then was 3’ adenylated. Adaptors were ligated to the ends of these 3’ adenylated cDNA fragments. The products are then purified and enriched with PCR amplification. We then quantified the PCR yield by Qubit and pooled samples together to make a single strand DNA circle (ssDNA circle), which gave the final library. DNA nanoballs (DNBs) were generated with the ssDNA circle by rolling circle replication (RCR) to enlarge the fluorescent signals at the sequencing process. The DNBs were loaded into the patterned nanoarrays and single-end read of 50 bp were read through on the BGISEQ-500 platform (BGI, Shenzhen, China) (14).

Identification and functional annotation of the DEGs

Reads from each sample were aligned to the mouse genome (GRCm38) using HISAT2 (15). We detected DEGs with DEGseq following the manufacturer’s instructions (16). Fold-change ≥ 2 and adjusted P value ≤ 0.001 were set as the significance thresholds. Gene ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of DEGs were both performed using pHYPER, a function in R. A false discovery rate (FDR) no greater than 0.05 was defined as significantly enriched.

Validation of the RNA-seq results by qPCR

RNA was reverse transcribed into cDNA using PrimeScript™ RT Master Mix (RR036A, Takara Bio). Two-step quantitative real-time PCR (qPCR) was performed in triplicate using an oligo-dT RT primer and SYBR Premix Ex Taq II (RR820A, Takara Bio) with a Light Cycler 480 (Roche). The results were calculated using the comparative Ct method formula, 2−ΔΔCt, and the results were normalized against the expression of the housekeeping gene Gapdh. Statistical significance was calculated by t test using GraphPad Prism 8 software.

Results

Spatial and temporal expression of Thra1 mRNA in the mouse hippocampus

We analyzed Thra1 mRNA expression levels in the mouse hippocampus by RNAscope at E10.5, E14.5, E16.5, E18.5, P3, P7, P14, and P56. Validation of the RNAscope method was performed using negative and positive control probes (shown in Supplementary Fig. 1, see section on supplementary materials given at the end of this article). Thra1 mRNA was detected as early as E10.5 in the neuroepithelium of cerebral vesicles (shown in Fig. 1A; a and b). With the development of the hippocampus, Thra1 mRNA was observed in the hippocampal primordium at E14.5 (shown in Fig. 1A; c and d) and in the newly developed Ammon's horn of the hippocampus at E16.5 (shown in Fig. 1A; e and f). From E16.5 to P14, Thra1mRNA was expressed in all the layers of the Ammon’s horn in the hippocampus (shown in Fig. 1A; f, h and 1B; A2, A3, A4, A5, B2, B3, B4, B5, C2, C3, C4, and C5), but decreased at P56 (shown in Fig. 1B; D2, D3, D4, and D5).

Figure 1

Spatial and temporal expression of Thra1mRNA in mouse hippocampus by RNAscope. (A) RNAscope hybridization demonstrates the dynamic change of Thra1 mRNAs in mouse hippocampus during the embryonic period. Formalin-perfused, fixed mouse embryo (E10.5 and E14.5) and brain (E16.5, E18.5, P3, P7, P14, and P56) were dehydrated in gradient ethanol, transparent in xylene, embedded in paraffin, and cut at a setting of 5 µm in the sagittal plane. The cytoarchitecture of the hippocampus in b, d, f (100×) is corresponding to the red squares in the sagittal plane of brain (a, c, e, 10×), respectively. The cytoarchitecture of the hippocampus in h and i (100×) is corresponding to the red squares in the sagittal plane of brain (g, 10×), with h denoting CA1 and i denoting DG area. The red arrow shows the Thra1 mRNA hybridized to the probes. Scale bar: 200 μm for a, 1 mm for c, 100 μm for e, g, and 10 μm for b, d, f, h, i. (B) RNAscope hybridization demonstrates the dynamic change of Thra1 mRNAs in mouse hippocampus during the postnatal period. A1, B1, C1, and D1 show the overall structure of the hippocampus at P3, P7, P14, and P56. A2-A8, B2-B8, C2-C8, and D2-D8 show structure of the sublayers in the hippocampus at P3, P7, P14, and P56, respectively. Taking the hippocampal structure of P56 as an example, the names and boundaries of the layers of hippocampal CA and DG are marked in D1. The red arrow shows the Thra1 mRNA hybridized to the probes. E, embryonic day; P, postnatal day; CA1, cornu ammonis area 1; DG, dentate gyrus; so, stratum oriens; sp, stratum pyramidale; sr, stratum radiatum; slm, stratum lacunosum moleculare; mo, moleculare layer; sg, stratum granulosum; po, polymorph layer.

The development of the DG areas of the hippocampus occurred later than that of Ammon's horn of the hippocampus, which includes the outer arm and the inner arm. Thra1mRNA was expressed in the outer arm of DG areas as early as E18.5 (shown in Fig. 1A; i). From P3 to P14, Thra1mRNAs were expressed in all layers of the DG areas (shown in Fig. 1B; A6, A7, A8, B6, B7, B8, C6, C7, and C8), but slightly decreased at P56 (shown in Fig. 1B; D6, D7, and D8).

Histological alterations in the hippocampus

The cytoarchitecture of the hippocampus were shown in Fig. 2A, at P0 (A–C), P7 (D–F), and P21 (G–I). Fig. 2 (A3-I3 and A4-I4) shows the cytoarchitecture of the DG and CA1 areas of the hippocampus, respectively, which correspond to the yellow box area in A1-I1.

Figure 2

Representative photomicrographs of Nissl-stained coronal sections of the hippocampus. (A) The cytoarchitecture of the hippocampus (A1-I1, 4×; A2-I2, 10×) at postnatal day (P) 0 (A–C), P7 (D–F), and P21 (G–I). A3-I3 and A4-I4 showing the cytoarchitecture of the DG and CA1 area of hippocampus, respectively, which is corresponding to the yellow rectangle area in A1-I1. The yellow arrow denotes subventricular zone. Scale bar: 200 μm for A1-I1; 100 μm for A2-I2; and 20 μm for A3-I3 and A4-I4. All the brain tissues were cut at a setting of 5 µm in the coronal plane, and n  = 3 for each genotype male mice at P0, P7, and P21, respectively. (B) Quantitation of the hippocampal sublayer thickness and the cell numbers in the pyramidal cell layer of the CA1 and granular cell layer in the DG at postnatal day 21. The maximal coronal section at the hippocampal level (corresponding to −2.92 mm from Bregma in adult mice (Paxinos brain atlas)) was used to analyze the morphology of hippocampus. The measurement of the layer thickness and the cell number were performed in the region where the highest point of the outer arm of the DG area, and the corresponding CA1 area. Two slides of each mouse were selected for the measurement and cell count of each sublayer in the CA1 or DG region of the hippocampus, average as the final result of each mouse. Three male mice were used for this experiment for each genotype. The number of cells in CA1 pyramidal cell layer and DG granule cell layer were corrected for the area of the counting region: cell counts/mm2 at 40× objective. Data were presented as the mean ± s.e.m. Statistical analysis was calculated by one-way ANOVA. *Compared with Thra+/+, P < 0.05; #Compared with ThraE403X/+, P < 0.05.

In the CA1 area, at P0, the neurons migrated from the ventricular zone to the CA1 area in the Thra+/+mice (shown in Fig. 2A; A4), whereas a large number of neurons accumulated in the ventricular zone in the mutant (ThraE403X/+ and ThraE403X/E403X) mice (shown in Fig. 2A; B4 and C4). At P7, the neurons in the Thra+/+ mice had completed migration, and the borders between the sp, sr, and slm were clear (shown in Fig. 2A; D4). Compared with those of the Thra+/+ mice, the neurons migrated from the ventricular zone to the stratum pyramidale of CA1 area in the mutant (ThraE403X/+ and ThraE403X/E403X) mice, in a manner similar to that of the Thra+/+ mice at P0, indicating that Thra mutation results in delayed migration of neurons in CA1. The borders between the four layers were indistinguishable in the ThraE403X/E403X mutant mice (shown in Fig. 2A; F4), whereas the boundaries of each layer in the ThraE403X/+ mice was clear, although there were still more cells in the stratum oriens that have not migrated to the stratum pyramidale of CA1 area (shown in Fig. 2A; E4). At P21, the neurons in the Thra+/+mice were arranged in order (shown in Fig. 2A; G4), whereas the arrangement of the neurons was loose and disorganized in the ThraE403X/E403Xmice (shown in Fig. 2A; I4). The thickness of the stratum oriens was higher in the Thra+/+mice than in the ThraE403X/+mice and ThraE403X/E403Xmice (shown in Fig. 2B; a). The total thickness of the CA1 region was thinner in the mutant (ThraE403X/+ and ThraE403X/E403X) mice than Thra+/+ mice (shown in Fig. 2B; c). The cell numbers in the pyramidal cell layer of CA1 area in the ThraE403X/E403X mice were less than those in the Thra+/+ mice and ThraE403X/+ mice (shown in Fig. 2B; d).

In the DG region, the neuronal cells in the three genotypes mice were migrating at P0 (shown in Fig. 2A; A3, B3, and C3), and no significant differences were found. At P7, the neurons in the Thra+/+ and ThraE403X/+mice had finished migration, and the borders between the layers were clear (shown in Fig. 2A; D3 and E3), whereas the borders between the three layers were indistinguishable in the ThraE403X/E403Xmice (shown in Fig. 2A; F3). At P21, the neurons in the Thra+/+ mice were arranged in order (shown in Fig. 2A; G3), whereas the arrangement of the neurons was disorganized and dense in the ThraE403X/E403X and ThraE403X/+mice (shown in Fig. 2A; H3 and I3). The total thickness of the DG region was thinner in ThraE403X/E403X mice than Thra+/+ mice (shown in Fig. 2B; c), whereas no significant difference in the thickness of each layer or the cell numbers of the granular cell layer in DG area were found between the WT and the ThraE403Xmutant mice at P21 (shown in Fig. 2B; b and d).

Impaired LTP in the hippocampus

Spatial learning and memory require synaptic transmission, which is characterized as the LTP of f-EPSPs. We examined LTP in acute hippocampal slices of juvenile mice at P21. The LTP results were evaluated by measuring baseline f-EPSP% after HFS. The percentage of the f-EPSP slope and f-EPSP amplitude in the ThraE403X/ E403X group was significantly lower than that in the Thra+/+ group (shown in Fig. 3, P < 0.05). The impairment was 60.9% in the fEPSP slope and 49.7% in the fEPSP amplitude.

Figure 3

Impaired long-term potentiation. Long-term potentiation (LTP) was performed in the male ThraE403X/E403X mice (n  = 7) and male WT mice (n = 5) at postnatal day 21. LTP was induced by HFS, measured as an increase in f-EPSP slope and amplitude, expressed as a percentage of the baseline of the f-EPSP slope and amplitude after HFS in mice. f-EPSP slope and amplitude were reduced in the ThraE403X/E403X group. Data are expressed as the mean ± s.e.m. *Compared with the Thra+/+, P < 0.05. Statistical analysis was calculated by the Student’s t test. f-EPSP, feld excitatory postsynaptic potential; HFS, high-frequency stimulation; TET, tetanic high-frequency stimulation.

Behavioral abnormalities

In contrast to the Thra+/+ mice, the time of escape latency to find the platform in the ThraE403X/+ mice was longer than that in the Thra+/+ mice during the period of training (Fig. 4A; a). After 4 days of training, when the platform was removed, the times of crossing both the target quadrant and the platform area were also reduced in the ThraE403X/+mice than in the Thra+/+mice (Fig. 4A; b).

Figure 4

Behavioral abnormalities of hippocampal origin. (A) Learning and memory performance in Morris Water Maze test. Morris water maze (MWM) test was performed in the ThraE403X/+mice and Thra+/+mice at 16 weeks. n  = 11 for male ThraE403X/+ mice and male Thra+/+mice, respectively. (A) (a) Time of escape latency. (b) Times of crossing of target quadrant and platform area in the spatial probe trial. Data are presented as mean ± s.e.m..Statistical analysis was calculated by the Student’s t test. Compared with the Thra+/+mice: *P < 0.05; **P< 0.01; ***P< 0.001. (B) Anxiety-like behavior in Elevated zero maze test. Elevated Zero Maze Test was performed in the ThraE403X/+ mice and Thra+/+ mice at 16 weeks. n  =11 for male ThraE403X/+mice and male Thra+/+mice, respectively. (B) (a) Total distance of traveling. (b) Latency to visit the open arms. (c) Percentage of freezing time. (d) Number of entries into the open arms. (e) Percentage of time spent in the open arms. (f) Number of rearing and head dipping events. Data are expressed as the mean ± s.e.m. Statistical analysis was calculated by the Student’s t test. Compared with the Thra+/+ mice: *P < 0.05; **P< 0.01; ***P< 0.001.

Mice naturally prefer the enclosed spaces, which makes them feel safe, but the anxious mice spend more time in the enclosed arms. The Elevated zero maze test showed that the total distance of traveling during the 5 min trial was shorter and the time of latency to visit the open arms was longer in the ThraE403X/+ mice than in the Thra+/+mice (Fig. 4B; a and b). The time percentage of both the immobility and not moving were higher in the ThraE403X/+mice than in the Thra+/+mice (Fig. 4B; c). Moreover, the number of either entries into the open arms or the head dipping in the open arms was less in the ThraE403X/+ mice than in the Thra+/+ mice (Fig. 4B; d and f. Although the ThraE403X/+ mice spent less time in the open arms and had less numbers of the rearing behavior in the closed areas, no significant difference was found between the WT and mutant mice (Fig. 4B; e and f).

Transcriptome changes

An average of approximately 24.11 M reads per sample were generated. The average mapping ratio with the reference genome was 95.9%, and the average mapping ratio with genes was 77.78%. A total of 17,994 genes were detected. The raw NGS data were submitted to the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA). The SRA accession number is PRJNA727394. A comparison of hippocampal gene expression between homozygous mutant mice and WT mice led to the identification of 754 DEGs, including 361 upregulated genes and 393 downregulated genes (shown in Fig. 5A). Predictably, a hierarchical clustering heat map based on these DEGs revealed relatively distinct separation between the homozygous mutant and WT mice (shown in Fig. 5B).

Figure 5

Transcriptomic changes induced by ThraE403X. (A) Volcano plot for differentially expressed gene mRNAs. (B) Hierarchical clustering analysis for differentially expressed gene mRNAs. Pheatmap is the abbreviation of pretty heatmap (a function of R). (C) Gene ontology analysis for differentially expressed gene mRNAs. (D) KEGG pathway analysis for differentially expressed gene mRNAs. The lower q-value indicates the more significant enrichment. Point size indicates DEG number (The bigger the ball, the more numbers the detected genes). Rich Factor refers to the value of enrichment factor, which is the quotient of foreground value (the number of DEGs) and background value (total gene amount). The larger the value, the more significant enrichment. Eight samples were codified as follows: HOM_3W, hippocampus sample from male ThraE403X/E403Xmice at 3-week-old (n  = 4); WT_3 W, hippocampus sample from male Thra+/+ mice at 3-week-old (n  =4).

GO and KEGG pathway analyses

The DEGs were subjected to GO analysis and KEGG pathway enrichment analysis to understand the gene biological functions. Notably, metabolic process, developmental process, immune system process within the biological process category, such as synapse within the cellular component category, binding, nucleic acid binding transcription factor activity, transcription factor activity, and protein binding within the molecular function category were the enriched GO terms. In particular, ion channel-related genes are enriched for ‘membrane’ GO term. Pathways including the PI3K-Akt signaling pathway, ECM−receptor interaction, and neuroactive ligand−receptor interaction were found to be enriched pathway terms. Several immune-associated pathways, such as complement and coagulation cascades, the NF-kappa B signaling pathway, the cytokine-cytokine receptor interaction pathway, and primary immunodeficiency, were also represented (shown in Fig. 5C and D).

qPCR analysis of gene expression

To verify the results of the RNA-seq analyses, genes reported to be regulated by T3, genes associated with neurodevelopment, and genes that enriched for GO terms or KEGG pathways were selected for qPCR analysis. Validation by qPCR analyses revealed that 25 genes were significantly dysregulated in ThraE403X homozygous mutant mice compared to WT mice (shown in Fig. 6, P< 0.05). For example, Itga7, Kdr, and Il2rb for PI3K-Akt signaling pathway. Col6a1, Gp1ba, Itga10, Itga11, and Spink10 for ECM–receptor interaction. C3, Ccl17, Cxcl5,and Ciita for immune- and inflammatory-associated pathways. The qPCR results showed changes similar to those identified through the gene expression analyses.

Figure 6

Validation by qPCR analyses. Relative mRNA levels of 25 DEGs were evaluated in Thra+/+(n  = 4) and ThraE403X/ E403X mice (n  = 4). The results were calculated using the comparative Ct method formula 2−ΔΔCt method and normalized against the housekeeping gene Gapdh. Data were presented as the mean ± s.e.m.Thra+/+ control was set to 1. Statistical analysis was calculated by the Student’s t test. *P<  0.05; **P  <  0.01; ***P < 0.001.

Discussion

In this study, we investigated the spatial and temporal expression of Thra1mRNA continuously in the embryonic and the postnatal hippocampus of mice using an ultrasensitive in situ hybridization method (RNAscope). Thra1mRNA was highly expressed during the period from the emergence of the cerebral vesicle neuroepithelium to the development of the hippocampus, and the expression was slightly decreased in adulthood, which indicates TRα1 is important not only for the developing hippocampus but also for the adult hippocampus. The previous study by Wallis et al used a mouse strain that expresses TRα1-GFP protein allowing examination of TRα1 expression in the brain. Their results showed that TRα1 was expressed in the stratum pyramidale and in the granular cell layer and hilar region of the dentate gyrus, as well as scattered in the stratum oriens and the stratum radiatum in the adult hippocampus (17), which is consistent with our results. Both her and our study found TRα1 was expressed in subgranular zone (SGZ), where the neural stem cells originated from. It indicated that TRα1 would play an important role in the regeneration of adult dentate gyrus.

Thyroid hormones have been implicated in multiple processes related to brain formation in mammals, such as neuronal progenitor proliferation, neuronal migration, and neural maturation (18). Thyroid hormone deficiency results in cell death (reduction in cell number) in the hippocampus and that the volume of the CA1 pyramidal cell layer is significantly smaller in hypothyroid groups than in the control groups (19, 20). Consistent with the previous studies, the present study also showed that ThraE403Xmutation results in the obvious cytoarchitectual differences. First, ThraE403X mutation results in the severely delayed migration of neurons in both CA1 and DG areas at P0 and P7. Second, at P21, ThraE403X mutation results in the thinner total thickness of both CA1 and DG region, as well as less cells in pyramidal cell layer of CA1 area. Third, at P21, ThraE403X/E403X mutation results in the disordered neuronal distribution in the hippocampus.

ThraE403X/+ heterozygous mice faithfully recapitulate clinical features of human RTHα, such as delayed postnatal growth and development, neurological and motor coordination deficits, and anemia (11). In the present study, we evaluated the hippocampal behavior in the 16-week-old heterozygous mutant mice, including the Morris water maze test related to learning and memory and the Elevated zero maze test related to emotion regulation. The heterozygous mutant mice showed the impaired learning and memory, as well as anxiety. Since the ThraE403X/E403Xmice die before 30 days of age, and usually the behavioral experiments are used for the adult mice, we selected LTP to evaluate the learning and memory of the 3-week-old ThraE403X/E403X mice. Our study showed that the ThraE403X/E403X mutation results in the impaired LTP in the hippocampus, which is similar to those showing hypothyroidism-induced plasticity impairment (21). Our results suggested that the impaired LTP seen in hypothyroid brains would be mediated by unliganded TRα1. The ThraE403X mouse model might be useful for studying the deficits of learning and memory, as well as anxiety in RTHα.

The hippocampal transcriptomic alterations caused by ThraE403X/E403X mutation pointed to some KEGG and GO pathways were enriched, such as PI3K-Akt signaling, ECM−receptor interaction, neuroactive ligand−receptor interaction, immunology and inflammation, and membrane ion channel.

PI3K-Akt pathway

The PI3K-Akt signaling pathway plays a significant role in neuronal health, by affecting synapse formation and maintenance, and contributes to LTP and synaptic plasticity (22, 23). Several rapid physiological effects of thyroid hormone on mammalian cells in vitro have been shown to be mediated by phosphatidylinositol 3-kinase (PI3K). Importantly, mutant TRβ was previously shown to activate PI3K via protein–protein interactions in TRβPV/PV mice (24). Whether mutant TRα1 can activate PI3K via protein-protein interaction, which deserves further investigation.

ECM and neuroactive ligand–receptor interaction pathway

ECM proteins and receptors play vital roles as guidance molecules during CNS development and have been implicated in the maintenance of stable neuronal connections and in the regulation of synaptic plasticity (25, 26). The involvement of ECM proteins would explain the impaired LTP of ThraE403X/E403X mice in the present study.

Immune- and inflammatory-associated pathway

Interestingly, many immune and inflammatory genes have been found to be differentially expressed in the ThraE403X/E403Xmice. These DEGs, including C3, C8b, Ccl17, Cd4, Cd40, Gata3, and Cxcl5, were found to be enriched in a number of immune-related pathways. Neuroinflammation has been reported to be associated with the abnormal brain development (27, 28, 29, 30). Our study indicated that immune and inflammatory pathway would be involved in RTHα-associated neurological diseases.

Membrane ion channel pathway

Ion channels are ubiquitous across several cell types, including progenitor cells, migrating neurons, and differentiated neurons. The precise control of ionic flux (calcium, sodium, and potassium) contributes to the developmental processes such as neural proliferation, migration, and differentiation (31). The previous studies revealed that thyroid hormone regulates dopaminergic neuron development from ventral midbrain neural stem cells by modulating TRPC1-mediated calcium signaling (32). In the present study, some ion channel-related genes were enriched for ‘membrane’ in Gene Ontology, which included Trpm5, Catsperg2, Asic4, Htr3a, Gabrq, Glra2, Kcnk1, Ano3, Pkd2l1, Cntn2, Trpv4, Kcnj10, and Kcne2. The membrane ion channel hypotheses would be a new direction for further work on the pathogenesis of RTHα.

The DEGs identified in the present study cover a wide range of physiological and biochemical processes, indicating the complicated effects of TRα1 action in the brain. Calml4 protects against spatial learning impairment in a mouse model of Alzheimer’s disease (33). Enpp2 affects cell adhesion and cell positioning in neuronal progenitors located in the ventricular zone of the cerebral cortex (34). Enpp2-knockout mice display neural tube defects and neurite outgrowth deficits (35). Arc is critical for long-term potentiation and depression of synaptic transmission, homeostatic synaptic scaling, and adaptive functions such as long-term memory formation (36). Previous studies have demonstrated that Nrgn (RC3), a direct target of T3, participates in postsynaptic events such as those affecting LTP (19, 37). Prolactin (PRL) acts through its membrane receptor, which is widely distributed throughout the body (38). This hormone was previously thought to be involved solely in female reproduction. However, it can also regulate neurogenesis in both the subventricular zone (SVZ) and the hippocampus (39). Moreover, PRL may exert neuroprotective effects in the hippocampus of adult animals exposed to chronic stress or subjected to hippocampal infusion of kainic acid (39). In this study, genes involved in different aspects of metabolism were also identified, such as Aldh1a1 (retinol metabolism) and Gls2 (glutamine metabolism). The determination of the mechanism that connect these genes and TRs requires further experiments. The mRNA levels of Nrgn, Dio3, and Sema3cwere significantly decreased in ThraE403X/ E403X mice, which was consistent with previously reported evidence indicating that they are T3-regulated genes (40, 41, 42). Recently, it was reported that both thyroid hormone deficiency and unliganded TRα1 caused a decrease in the number of parvalbumin (PV)-expressing neurons, as a subpopulation of gamma-aminobutyric acid (GABA) interneurons, in the cortex and hippocampus (43, 44, 45, 46). In agreement with their results, our RNA-seq data showed that some genes, such as Flywch2, Pvalb, and Syt2,which are preferentially expressed in PV + neurons were downregulated in ThraE403X/E403X mice compared with WT mice. The results from mice with both thyroid hormone deficiency and unliganded TRα1 mutation suggested that GABAergic neurons would be the direct targets of TH/TRα1 during brain development.

In summary, the ThraE403X mutation results in histological and functional abnormalities, as well as dysregulated genes in the mouse hippocampus. This study on the ThraE403X mutation offers new insights into the potential biological causes of RTHα-associated neurological diseases.

Declaration of interest

Weiping Teng is an Editorial Board Member of the European Thyroid Journal. He was not involved in the editorial or review process of this paper, on which he is a listed authors. The other authors declare that they have no other competing or financial interests.

Funding

This study was supported by Chinese National Natural Science Foundation Grants 81970681 and 81570711 to Xiaochun Teng.

Author contribution statement

XiaochunTeng, Zhongyan Shan, and Weiping Teng designed and supervised the study. Pingping Dang performed the RNAscope, Nissl staining, Elevated zero maze test, and prepared the hippocampus samples for RNA-seq. Yingxin Fang wrote the manuscript, analyzed the RNA-seq data, and performed the qPCR. Ranran Wang measured the long-term potentiation. Yue Liang performed the Morris water maze test. Yue Liang, Defa Zhao, Yue Xi, Dan Zhang, and Wei Wang were responsible for mice feeding, breeding, and gene identification. XiaochunTeng revised the manuscript. All authors discussed the data and edited the manuscript.