I. Développement d’une lignée de souris transgéniques surexprimant, spécifiquement dans les neurones, un récepteur TRβ1 muté pour sa liaison à l’hormone.

Table des matières

Plusieurs laboratoires ont construit de nombreux modèles animaux pour étudier le rôle spécifique des récepteurs thyroïdiens. En 1998, aucun de ces modèles ne permettait de comprendre le rôle tenu par ces récepteurs dans le développement du SNC. En effets, les différentes constructions n’ont pas permis de retrouver les anomalies du développement du cerveau observées dans l’hypothyroïdie congénitale. Ces anomalies pourraient être dues à un effet indirect par altération du système périphérique, provoquée par le manque d’hormone. D’autre part, ces anomalies touchent le développement de tous les types cellulaires majoritaires constituant le cerveau, que se soit les astrocytes, les oligodendrocytes comme les neurones. Il est donc difficile d’attribuer les pertes de fonctions cérébrales rapportées dans l’hypothyroïdie à un disfonctionnement de l’un ou l’autre type cellulaire. Nous avons choisi de créer une nouvelle lignée de souris transgéniques afin de définir le rôle du récepteur TRβ1 dans le développement neuronal. La stratégie choisie à été de sur-exprimer un récepteur TRβ1 muté pour sa liaison à l’hormone afin d’induire une compétition avec le récepteur endogène. L’abolition des fonctions du récepteur thyroïdien, dépendantes des HT, est plus préjudiciable que la destruction complète du récepteur et permet donc de rechercher ses fonctions plus efficacement. Le récepteur TRβ1 humain (G345R) a été choisi parce qu’il engendre des anomalies trés sévères chez les patients porteurs de cette mutation, maladie appelée résistance aux HT décrite précédemment dans l’introduction. Le promoteur des neurofilaments légers permet de cibler spécifiquement le SNC et plus particulièrement les neurones. Deux lignées de souris transgéniques distinctes produisant des quantités différentes de transgène ont ainsi vu le jour.

Le mécanisme moléculaire qui est à l’origine des effets prononcés des hormones thyroïdiennes sur le cerveau en développement est inconnu. Leur action est principalement mediée par trois récepteurs thyroïdiens majeurs : TRα1, TRβ1 et TRβ2. Pour déterminer les fonctions du récepteur TRβ dans le développement du cerveau, nous avons construit deux lignées de souris transgéniques, qui expriment un récepteur TRβ1 muté pour sa liaison à l'hormone, placé sous le contrôle du promoteur humain des neurofilaments légers pour exprimer le transgène dans le système nerveux central. Nous montrons que le transgène est spécifiquement exprimé dans les neurones et plus précisement, les cellules de Purkinje du cervelet. Durant la seconde semaine de vie, le cervelet des souris transgéniques présente toutes les anomalies morphologiques observées dans l’hypothyroïdie congénitale, comme l’altération de la foliation et de la maturation des cellules de Purkinje, le retard dans la prolifération et la migration des cellules granulaires de la couche granulaire externe et l’augmentation de l’apoptose dans la couche granulaire interne autant que l’expression anormale de certains gènes. Nos résultats suggèrent que l’expression d’un TRβ1 muté dans les cellules de Purkinje est suffisante pour récapituler les anomalies morphologiques observées dans le cervelet des rongeurs hypothyroïdiques.

Ma contribution à ces travaux constitue la majorité du contenu de cette publication, en dehors de la construction qui a été réalisée par Louise Cosette et des coupes de cerveau ainsi qu’une partie des colorations au crésyl-violet qui ont été effectuées par Julie Martel.

Cette publication a été rédigée par le Dr jack Puymirat et moi-même.

Expression of a Human Mutant Thyroid Hormone Receptor β1 in Purkinje cells is Sufficient to Delay Development of the Cerebellum

Christelle Cayrou1, Julie Martel1, Louise Cossette1, Dominique Grouselle1, and Jack Puymirat1

1Unit of Human Genetics, CHU Laval Research Center, University Laval, 2705 Blvd Laurier, Ste-Foy, Quebec, G1V 4G2, Canada; 2U549 INSERM, Blvd Alesia, Paris 75013, France. The work was performed in Canada

Acknowledgments : We thank Dr Samuel Refetoff for kindly supply us with the hmTRβ1 cDNA and for its insightful comments on our manuscript. This work was supported by the Canadian Institutes of Health Research (CM053101). C.C. was supported by a fellowship from the Canadian Thyroid Association.

Key words: Brain; thyroid hormones; neurons; gene expression; congenital hypothyroidism ABSTRACT

The molecular mechanisms that underlie the pronounced effects of thyroid hormones on the developing brain are unknown. Their actions are mainly mediated through three major thyroid hormone receptor (TR) isoforms; TRα1, TRβ1 and TRβ2. To determine the functions of the TRβ in the developing brain, we have generated transgenic mice that express a mutant human TRβ1 under the control of the human neurofilament promoter to drive the expression in the central nervous system. We show that the transgene is specifically expressed in brain neurons and in Purkinje cells in the cerebellum. During the second week of life, the cerebellum of hmTRβ1 mutants presents all of the morphological abnormalities observed in congenital hypothyroidism, such as alterations in the maturation of Purkinje cells and in cerebellar foliation, delay in proliferation and migration of granule cells from the external granular layer and increased apoptosis in the internal granular layer as well as abnormal gene expression. Our results suggest that the expression of unliganted TRβ1 in Purkinje cells is sufficient to produce the morphological abnormalities observed in the cerebellum of congenital hypothyroid rodents.

INTRODUCTION

Thyroid hormones (TH) are essential for the normal development of the brain where the absence of hormone at birth impairs the proliferation, migration and maturation of neurons and alter synaptogenesis and myelinisation in different species (Legrand (1982); Anderson (2001)). The mode of action of TH is mediated through thyroid hormone receptors (TRs), which are ligand-dependent transactivation factors (Glass and Holloway (1990); Oppenheimer et al. (1994)). Three TRs isoforms have been identified, TRα1, TRβ1, and TRβ2, which are derived from the TRα and TRβ genes respectively (Sap et al. (1986); Weinberger et al. (1986); Murata (1998)). However, other products from these two genes do not have receptor functions (Chassande et al. (1997); Williams (2000)). Their specific functions in mediating TH response in brain development is poorly understood. Although TRβ2 plays a major role in feedback regulation of TRH (Abel et al. (2001)), it is solely expressed in the brain (Bradley et al. (1992)), TRβ1 and Trα1 are the most abundant brain isoforms. Reports suggest that TRβ1 has a major role through the effects of TH on latter developing brain. Mutations in the TRβ gene are associated with RTH, witch can severely affect mental development (Beck-Peccoz and Chatterjee (1994); Refetoff et al. (1993)). No mutation of TRα has been yet associated with human disease. Permanent transfection experiments demonstrate that a neuronal cell line (N2a) that overexpresses TRβ1 shows, in presence of T3, arrest of proliferation and induction of morphological and functional differentiation (Lebel et al. (1994)), similar to the effect of T3 on neurons during perinatal development brain. No effect was observed in response to T3 in N2a overexpressing Trα1. Furthermore, the rise in the level of TRβ1 commences on embryonic day 19 and increases 40-fold to postnatal day 10, while TRα1 reaches levels close their adult maximum. The temporal association of this surge, together with the level of T3 and the known influence of TH on the developing brain raises the possible implication of TRβ1 in inducing these effects (Strait et al. (1990); Mellstrom et al. (1991); Bradley et al. (1992); Puymirat (1992)).

Knockout mice for TRβ and transgenic mouse mutants for TRβ1 have been developed to determine the function of TRβ1, but these models present only mild phenotypes and no brain alterations (Forrest et al. (1996); Wong et al. (1997); Hayashi et al. (1998); Abel et al. (1999); Gothe et al. (1999)) except some neuropeptides defect (Calza et al. (2000)). Recently, a TRβ knock-in mutant was developed by introducing into the mouse β locus a T3 binding mutation by homologous recombination (Hashimoto et al. (2001)). These mice show delay in cerebellum development, as do hypothyroid animals. But, it is still uncertain which isoform is responsible for these effects. Indeed, all TRβ proteins present in the periphery and in all cell types in the brain are affected by this mutation. It is therefore possible that changes observed in the developing brain may be a result of the peripheral effects of TH deficiency or of TH on glial cells.

In this study, we produced transgenic mice expressing a mutant human TRβ1 (hmTRβ1) under the control of the human light neurofilament (hNF-L) promoter to target expression of hmTRβ1 to neurons (Charron et al. (1995a); Charron et al. (1995b)). We selected the natural mutant TRβ1 G345R to insure dominant negative alteration of endogenous TRβ action (Sakurai et al. (1990)). We show that these mice express hmTRβ1 specifically in brain neurons and that the expression of this hmTRβ1? induces all of the morphological alterations observed in the cerebellum of hypothyroid animals.

MATERIALS AND METHODS

Development of Transgenic Mice . A 2.1 Kb blunt-end BamH I/ Nco I fragment containing the hmTRβ1 cDNA (hmTRβ1/pcDNA1 vector was kindly provided by Dr S. Refetoff, University of Chicago) encoding the mutant G345R hTRβ1 cDNA and the BGH poly(A) was cloned into the unique Hind III site of the pGChNF-L plasmid (kindly provided by Dr J. P. Julien, Montreal General Hospital, Canada). This plasmid contains a 4.9 Kb hNF-L fragment which includes 292 base pairs of the 5’-flanking sequences, a 4.6 Kb fragment of exonic and intronic region, a 3’UTR sequence to drive expression in the nervous system (Charron et al. (1995a; Charron et al. (1995b)) and the first polyadenylation site of the hNF-L gene. Enzymatic digestion and DNA sequencing established sense orientation of the insert.

Kpn I/ Sac II-digested hNF-L/hmTRβ1 was purified with QIAEX II gel extraction kit (Qiagen, Mississauga, ON), microinjected into fertilized eggs derived from C3B6F1 mice and transferred into the oviducts of pseudopregnant CD1 mice as described (Brinster et al. (1981)). Presence of the transgene in the founders was tested by Southern blotting of genomic DNA isolated from tail biopsies. Ten (10) μg of DNA were digested with EcoR I and Xho I and hybridized with a [32P]-labeled full-length hmTRβ1 cDNA probe. Screening offspring was performed by PCR analysis of tail DNA using primers specific for hTRβ1. ?The oligonucleotides sequences are as follows: sense, 5’ttgggacaaaccgaagcact3’ and antisense 5'tgtgcccgatggacttctgc3' which generate a 593 bp DNA fragment.

Brain cell cultures . Primary neurons were prepared from E16 fetal mouse cerebral hemispheres as previously described (Puymirat et al. (1992)). Briefly, cells were mechanically dissociated and plated at a density of 3 x 106 cells per 100-mm tissue culture dish. Cells were cultured in chemically defined medium and maintained in a humidified atmosphere of 5% CO2, 95% air at 37oC. Primary mixed glial cell cultures were prepared from newborn mouse cerebral hemispheres as previously described (Baas et al. 1997). Cells were mechanically dissociated and plated in Waymouth’s medium (Invitrogen, Burlington, ON.) supplemented with sodium bicarbonate (2g/liter), sodium pyruvate (110 mg/liter), penicillin (50 units/ml), streptomycin-sulfate (50mg/ml), and 10% stripped fetal calf serum. Cells were cultured for 2 weeks under a humidified atmosphere of 5% CO2, 95% air at 37oC.

Northern Blot Analysis . Total RNA was prepared from brain, kidney, spleen, liver, lung and heart of transgenic (Tg) and wild-type (Wt) mice by TRIzol (Invitrogen, Burlington, On). Total RNA (20 μg/lane) was separated on formaldehyde agarose gels and transferred onto a nylon membrane. Blots were hybridized with 32P-cDNA probe, washed with 0.1 x SSC, 0.5% SDS at room temperature for 20 min, then twice with 0.1 x SSC, 0.5% SDS at 67oC for 20 min before exposure to x-ray film. Probes used: hmTRβ1 cDNA fragment (+325-+545) to detect the transgene; mouse TRβ1 cDNA fragment (+112-+310) to detect the endogenous TRβ1, the full-length rat TRα cDNA to detect the TRα1- and α2- mRNA, BTEB cDNA fragment for BTEB mRNA. Blots were normalized using either a [32P]-3’ end-labelled 18S rRNA-specific oligonucleotide (Clements et al. (1988)) or rat cyclophilin cDNA. DNA probes were labeled with [32P] dCTP by random priming.

Western Blot Analysis . Cerebellum tissu (100 mg) from 15-day-old Tg and Wt mice was disrupted in liquid nitrogen and homogenized in 1 ml buffer A (250 mM sucrose, 20 mM Tris HCl pH 7.8; 1,1 mM MgCl2, 0,1% triton X100, 1mM PMSF and a mix of protein kinase inhibitors (Roche diagnostic, Laval, QC) 10 min at 4oC. Nuclei were removed by centrifugation for 5 min at 5,000 rpm, washed 3 times in the same buffer. Nuclear proteins were then extracted in lysis buffer B (Buffer A, 5 mM DTT, 20% glycerol and 400 mM KCL) 20 min at 4oC. The samples were centrifuged at 16,000 g x 5 min and the supernatant was dialyzed in buffer B with 150 mM KCL only. Samples were stored at –20oC until used. One hundred (100) μg of nuclear proteins were separated by 10% SDS-polyacrylamide gels, transferred to a PVDF membrane and probed with a monoclonal anti-hTRβ1 (J51, 1/1000, StressGen Biotechnology Corp., Victoria, BC) in TBS containing 0,05% Tween. Detection was performed with a horseradish peroxidase-coupled anti-mouse antibody (1:10000, Jackson Immunoresearch laboratories, West Grove, PEN) and ECL Western blotting detection reagents (Amersham Pharmacia biotech, Baie d'Urfe, QC). Western blots were normalized using a monoclonal antibody (2B3G8) prepared against a 55-kDa nuclear protein. This nuclear protein is constitutively expressed in various tissues including neural cells (Puymirat J, unpublished data).

Determination of Total T3 (TT3), free T4 (FT4) and Hypothalamic TRH . Whole blood was obtained from the retro-orbital sinus. FT4 and TT3 were measured using an heterogeneous competitive immunoassay (Immuno-1-system, Bayer Corporation, Tarrytown, NY). Hypothalamus were isolated from 30-day-old Tg and Wt mice, disrupted in acetic acid 1M pH2, 1mM PMSF, boiled 10 min, sonicated and centrifuged at 12000g at 4oC. The supernatant was lyophilized and stored at 4oC until used. Thyroliberin (TRH) was measured by using a specific Enzyme Immuno Assay for TRH in mouse and rat brain (Grouselle et al. (1990)).

Behavioral Tests . Three month-old Tg and wt mice were tested in behavioral experiments with open-field test to measure locomotors activities and motor posture pattern (Gerlai et al., 1993). The open-field apparatus was a large opaque cage (40 X 40 cm) whose bottom was divided in a 10 X 10 cm square grid. Mice were accustomed to the apparatus twice: 20 min on the day before testing and 20 min on the test day. Locomotion score and motor posture pattern were assessed by the number of squares crossed in 10 min and by the number of rearing carried out during the same period, respectively.

Tissue Preparation . Brain from P5 postnatal day mice was isolated and fixed in 4% paraformaldehyde/PBS overnight at 4oC and postfixed in 4% paraformaldehyde/PBS/30% sucrose for the following 24 h. Brain tissue was then snap-frozen and stored at –80oC before sectioning. Mice from P10 to P20 were anesthetized and perfused intracardially with 0.9% NaCl followed by 4% paraformaldehyde/PBS. The brain was isolated and postfixed as above. Cerebellums were sectioned into 12 μm slices.

Histology and Morphological Analysis. Sections were stained with Cresyl-Violet and analyzed with a Zeiss Axiophot light microscope in order to study the foliation and the size of the molecular layer and external granular layer. Six different sections, every ten sections, were analyzed for each animal (Wt n=3 , Tg n=3 ) at a magnification of 20X. Two determinations were performed for each lobule and all lobules per section were analyzed.

Immunohistochemistry . Sections were incubated with a monoclonal anti-calbindin D28K antibody (1:200, Sigma-Aldrich, Oakville, ON), which specifically stains Purkinje cells. Labeling was visualized using a horseradish peroxidase-coupled anti-mouse antibody (1:1000, Jackson Immunoresearch laboratories, West Grove, PEN) and 0.05% diaminobenzidine as chromogen. Morphometric analysis of Purkinje cells was performed on three different sections for each animal (Wt n=2 , Tg n=2 ). Only cells with a clear body and distinct area were counted. Areas of ramification and the body of Purkinje neurons were quantified by NIH image version 1.62. Purkinje cell densities were determined by counting every cell in 500 μm length of a lobule at magnification 100X. Two lobules from five sections (every fifteen section) were analyzed.

Transgene expression was visualized on 20 μm sections from P20 Tg-1 and Wt mice. Fresh brain was cut on cryostat and fixed with 100% methanol. Sections were incubated with M.O.M kit (according to laboratory instructions, Vector Laboratories, Burlington, ONT) before staining with a monoclonal anti-hTRβ1 (J51, 1/200, StressGen Biotechnology Corp., Victoria, BC) and visualized with a biotin-coupled anti-mouse antibody (1:500, DAKO Diagnostics Canada, Mississauga, ONT) and horseradish peroxidase-coupled streptavidin (1:500, DAKO Diagnostics Canada, Mississauga, ONT). Diaminobenzidine (DAB) was used as chromogen.

Proliferation Studies . 10-day-old mice were subcutaneously injected with brdU (5 mg/100g of body mass). Six hours after the injection, the mice were anesthetized and perfused intracardially with 0.9% NaCl followed by 4% paraformaldehyde/PBS. The brain was isolated and postfixed as described above. Cerebellar cryostat sections were washed with PBS, treated with 2N HCl for 30 min at 37oC, rinsed once with 0.1M Na2B407 and twice with PBS. Sections were incubated with trypsin for 30 min at 37oC, washed three times with PBS and incubated with an anti-brdU (1/1000; Sigma-Aldrich, Oakville, ON) in PBS/10% FBS/ 0.2% Tween20. Detection was performed with a horseradish peroxidase-coupled anti-mouse antibody (1:10000, Jackson Immunoresearch laboratories, West Grove, PEN) and 0.05% diaminobenzidine as chromogen. Density of immunoreactive cells was evaluated by the number of pixels in 10000 μm2 section. Two-fields/section (X200) on six sections, every seven section, were analyzed. Values were expressed as pixels/mm2 (Wt n=3 , Tg n=3 ). The volume of each lobule containing immunoreactive cells was determined by measuring their area with NIH image version 1.62.

Apoptosis Studies . TUNEL staining was performed using the in situ Apoptag-kit (oncor) according to the protocol previously described (Neveu and Arenas (1996)). Briefly, 15-day-old sections were incubated with dUTP-digoxigenin and terminal deoxynucleotidyl transferase for 1h at 37oC. After washing, sections were incubated with an anti-digoxigenin-fluorescein antibody, counterstained with DAPI and mounted in PBS-Glycerol. Cell death was quantified by counting the number of fluorescein-positive nuclei within a 10000 μm2 section. Two-fields/section (200X) on five different sections, every four sections, were analyzed. Values were expressed as number of positive cells/mm2. (Wt n=3 , Tg n=3 ).

Statistical Analysis . Statistical analysis was performed using one-way ANOVA (StatView software, version 4.51). Differences between groups were assessed using Fisher's PLSD. All values are given as mean ± SE, and statistical significance was set at P<0.05

RESULTS

Generation of Transgenic Mice

The DNA construct containing a 4.9 KB human NF-L fragment and the hmTRβ1 cDNA (Fig 1a), was microinjected into fertilized eggs derived from C3B6F1 mice. A total of five independent founders Tg mice were generated as determined by Southern blot analysis (data not shown). Four of which successfully transmitted the transgene to progeny. Two separate homozygous lines containing 10 and 125 copies of the hNF-L/hmTRβ1 were established by intercrossing heterozygous mice. The transgene has demonstrated simple Mendelian inheritance. All analysis presented in this study were performed in homozygous animals expressing 10 copies of the hNF-L/hmTRβ1 (Tg-1). We also verified whether similar morphological abnormalities were also observed in the other mice (Tg-2) produced that contained 125 copies of the transgene.

Expression of Mutant TRβ1 Transgene

Expression of the transgene in several tissues was determined by Northern blot (Fig 1b). A 4- and a 2.6-kb mRNA were detected in the brain of Tg mice. The 2.6-kb mRNA corresponds to the expected size of the transgene whereas the 4-kb transcript was about ~1.5-kb longer than expected. These latter mRNAs correspond to the preferential use of the hNF-L poly (A) site. The transgene was not detected in several tissues including heart, lung, liver, kidney and spleen consistent with prior observations showing that the hNF-L promoter specifies brain-restricted transgene expression (Charron et al. 1995a,b). No transgene was detected in Wt mice. The Tg-2 mice express 1.5-fold lower levels of the transgene than the Tg-1 mice.

To determine whether the transgene is specifically expressed in neurons, we next analyzed its expression in neuronal and in mix glial (containing astrocytes and oligodendrocytes) cultures. In neuronal cultures, 80-85% of the cells are neurons (based on positive neurofilament immunostaining) and the remaining cells are astrocytes (based on positive immunostaining with anti-GFAP) after 10 days in vitro (Puymirat et al. 1992). In mix glial cultures, cells were either GFAP-positive attesting their astrocyte phenotype or galactocerebroside-positive attesting their oligodendrocyte phenotype; no neurofilament-positive cells were observed after 15 days in vitro (Baas et al. 1997). As shown in Fig 1c, the transgene is expressed in neuronal but not in mix glial cultures indicating that its expression is neuron-specific.

The expression of the transgene was confirmed by Western blot analysis (Fig. 1d), using an antibody that recognizes an epitope in the hTRβ1 NH2-terminal region. This antibody detected a 55-kDa protein in the cerebellum of postnatal day 10 Tg-1 animals whereas no band was detected in the cerebral hemispheres of Wt animals. The size of the protein corresponds to the expected 55KDa size of hTRβ1. A higher band was detected in both normal and Tg mice and most likely corresponds to a non-specific protein. Furthermore, this expression is cell-specific since we notice transgene expression only in Purkinje cells (Fig 1e). To determine if granule cells were affected, we carried out RT-PCR on NT-3 RNA, a granule cell specific gene, which was downregulated in hypothyroid mouse (Fig 2b). These mice show a 2-fold decrease of NT-3 RNA expression in comparison to Wt mice. No significant difference in NT-3 RNA expression was observed between Tg and Wt mice.

To determine whether the hmTRβ1? interferes with the endogenous TRβ1 and TRα gene expression, we hybridized northern blots with probes specific for mouse TRβ1 and TRα isoforms. As shown in Fig 2a, expression of the hmTRβ1 did not affect expression of the mouse TRβ1, TRα1 and TRα2 mRNAs in the brain.

Study of the Thyroid Function in Transgenic Mice

Mean blood FT4 determined in P25 Tg mice was 1.5-fold higher than in P25 Wt mice (42.1 ± 2.1 pM vs. 27 ± 5.3 pM; P = 0.008). However, TT3 level was not modified in Tg-1 compared with Wt mice (0.6 ± 0.03 vs. 0.7 ± 0.04; P>0.05 ). Elevation of TH levels normally suppresses TRH production by TRH-neurons of the paraventricular hypothalamus. To determine whether an increase in levels of FT4 negatively regulates TRH production in these neurons, we measured the levels of TRH by a specific EIA. The hypothalamic TRH content was significantly increased by 1.3-fold in Tg compared to Wt mice (57 ± 2.3 vs. 44.7 ± 3.7; P = 0.02 ).

Retarded Growth of Transgenic Mice

No significant differences were observed on the lifespan and reproductive capacity between Tg and Wt mice. Growth retardation in Tg mice was manifested as impaired weight gain and this was detected in both male and female. Transgenic mice significantly showed a proportional weaker weight than Wt animals, at all time points examined ( P<0.0001). This reduction is about 17% in average. Transgenic mice had also a significant delayed eye opening (12.6 ± 0.22 versus 13.4 ± 0.3 days, P <0.001 ). To assess possible deficits in motor performance of the Tg mice, three month old Tg-1 and Wt mice were subjected to behavioral tests. Study of the motor performance revealed hyperactivity in Tg-1 mice (184.8 ± 9.5 versus 105.5 ± 15.4 cross-squared, P<0.0001 ) whereas no significative difference was observed on motor posture pattern between Tg-1 and Wt mice (61.4 ± 4.3 versus 45.8 ± 10.8, P=0.2 ).

Delay in Development of the Cerebellum in Transgenic Mice

Several morphological abnormalities were observed in the cerebellum of Tg-1 mice. There was an atrophia of the cerebellum (Fig 3a) from P15, which persists in the adult animals. Sections stained with Cresyl Violet showed an alteration in cerebellar foliation at all time points examined in Tg-1 and Tg-2 mice (Fig 3b). Although each layer seems to be organized, the external granular cell layer (EGL) in mutant mice was thicker than in control at P10 (Fig 3d: 37.9 ± 1.8 μm versus 29.1 ± 0.8 μm, P<0. 0001) and persists at P15 (Fig 3c). The molecular layer (ML) was significantly thinner in Tg-1 mice compared to Wt animals at P10 (less 20%, P<0.0001 ) and P15 (less 10%, P<0.0001 ) (Fig 3c and d).

Alteration of Purkinje cells maturation

The development of Purkinje cells is also severely affected in hypothyroid animals. Arborization of Purkinje cells was responsible for the ML thickness. In hypothyroidism, alteration of the Purkinje cell development coincides with depletion of this thickness. We therefore analyzed Purkinje cell development in Tg mice after their labeling with a monoclonal anti-calbindin antibody (Fig 4a). No significant difference was observed in the number of Purkinje cells between Tg and Wt mice ( P=0.13 ), but their maturation was severely affected. At P10, the area of Tg-1 cell bodies was 2.1-fold (Fig 4c; P<0.0001 ) smaller than that observed in Wt mice. Furthermore, their Purkinje dendritic tree was hypoplasic as attested by the size of the area of arborization, which was 1.8-fold (Fig 4b; P<0.0001 ) smaller compared to Wt animals.

Delay in arrest of the EGL proliferation and increase of apoptosis in IGL

To determine whether the persistence of the EGL in Tg-1 mice at P10 corresponds to the persistence of granule cell proliferation, BrdU incorporation was performed. There was a significant 1.5-fold increase in the number of BrdU-positive granular cells in the EGL of Tg-1 mice compared to Wt animals (Fig 5a; P=0.004 ). Once granule cells ceased their mitosis, they leave the EGL and migrate into the IGL where a greater number of them die in hypothyroid animals. To determine whether increased apoptosis occurs in the IGL of Tg-1 mice, like in hypothyroid animals, we next analyzed granule cell survival by TUNEL assay. In normal mice, cell death was higher at P10 and declined by 2-fold at P15 to reach undetectable levels in the adult. Based on these data, apoptosis was studied at P15. Tg-1 mice present a 3.7-fold (Fig 5b; P=0.0005 ) increase of TUNEL-positive cells in the IGL but also in the EGL and the molecular layer compared with that observed in Wt mice.

To determine whether all these phenotypes are specific, we analyzed the Tg-2 mice. As shown in Fig. 3b and 4a, these mice exhibit the same cerebellar morphological abnormalities as observed in Tg-1 mice but the effects were less pronounced, in good agreement with their lower transgene expression level.

Expression of BTEB mRNA in The Cerebellum of Transgenic Mice

We have previously identified the basic transcription factor (BTEB) as a thyroid-hormone-regulated gene in the developing rat brain (Denver et al., 1999). We have also demonstrated that this gene is regulated in neuro-2a cells that overexpress TRβ1 but not in cells overexpressing TRα1, suggesting a specific regulation via TRβ1. Furthermore, BTEB was implicated in the thyroid hormone effect on neuron ramification (Cayrou et al., 2002). To determine if its expression is altered in the developing brain of hmTRβ1 mutants, we analyzed its expression in the developing cerebellum. As shown in Fig 6a, the levels of BTEB RNA were low at P5 in Wt mice and increased significantly by 2.6-fold up to P10 ( P=0.003 ). It still increased thereafter to reach maximum levels at P15 (4.2-fold level observed at P5, P=0.006 ). In the developing brain of Tg-1 mice, BTEB RNA was expressed at similar levels as observed in Wt mice at P5 but increased thereafter more slowly (increase P10 vs. P15: P=0.15 ). At P15, the level of BTEB RNA was significantly 1.3-fold lower than that observed in controls ( P = 0.005) and remained lower at P20 ( P=0.03 ). The levels of BTEB RNA in the cerebellum of Tg-2 mice was also decreased compared to the levels of BTEB RNA in Wt mice but remained higher than that observed in Tg-1 mice (not shown).

DISCUSSION

In order to investigate the molecular basis that underlies brain abnormalities in congenital hypothyroidism, we produced transgenic mice that overexpress a mutant human TRβ1. We selected the natural mutant TRβ1 G345R that causes severe RTH in affected individuals (Sakurai et al. (1990)). To determine the consequences of a dominant negative effect occurring specifically in neurons, we used the hNFL promoter to target the expression of the transgene in neurons at high levels (Charron et al. 1995a,b). We confirmed that the transgene was expressed by neurons but not by glial cells. Our mice are phenotypically distinct from the Knock-out TRβ mice (Forrest et al. (1996)) but notably similar to TRβ (Δ337Τ) knock-ins (Hashimoto et al. (2001)). Knockout mice show severe dysfunctions of the hypothalamus-pituitary-thyroid axis but, surprisingly, no phenotypic alteration of brain development like hypo- or hyperthyroidism. HmTRβ1 mice display hypothyroid changes in cerebellum development despite weakly but significantly elevated FT4 levels. These results indicate that dominant negative inhibition by mutant TRβ1 creates these observed phenotypes. This hypothesis was bolstered by a previous study using mutant TRβ1 expressed in the liver of transgenic mice that showed a dominant negative effect of mutant RT on endogenous TRβ1 (Hayashi et al. (1996)). More recently Hashimito and colleagues also showed a dominant negative effect of mutant TRβ in Knock-in mice (Hashimoto et al. (2001)).

We chose to study cerebellar development because it is a major TH target structure in the developing brain and hypothyroid effects were previously reported (Legrand (1982)). Tg mice show all the defects observed in the cerebellar development of hypothyroid animals: persistence of granule cell progenitors still at P15 in the EGL with a greater proliferative zone in oEGL at P10, atrophia of Purkinje dendritic trees and increase of apoptosis in IGL at P15. The different phenotype intensities observed in our two transgenic mice lines expressing hmTRβ1 RNA at different levels demonstrate specific action of the transgene. Interestingly, hmTRβ1 proteins were only expressed in nuclear Purkinje cells, in agreement with NF-L expression in the developing cerebellum (Chan et al. (1997; Gilad et al. (1989)), but we showed that both granule and Purkinje cell development appear to be affected by hmTRβ1. Purkinje cells express only TRβ isoforms whereas postmitotic granule cells express only TRα isoforms (Bradley et al. (1992)). The dominant effect of hmTRβ1 in our mice is therefore specific of endogenous TRβ. A major difference between hmTRβ1 and TRβ Knock-in mice is the mutation of all TRβ isoforms in Knock-in mice that could explain the big difference in TH levels between the two transgenic mice. A recent report on knockout TRβ2 mice shows the critical role of the TRβ2 isoform on TRH regulation and on negative feedback regulation of the hypothalamus-pituitary-thyroid axis by TH (Abel et al. (2001)). We only observe a weak TRH increase in the hypothalamus and pituitary of Tg mice (data not show) despite a significant increase of TT4 serum level. Abel and co-workers obtained similar results with pituitary-specific mutant TRβ1 expression (Abel et al. (1999)). These data suggest that mutant TRβ1 have only a weak dominant negative effect on TRβ2 isoform functions. This implies that the major dominant negative effect of hmTRβ1? in Purkinje cells is presumably specific to endogenous TRβ1 .

Some neurotrophins, like neurotrophin-3 (NT-3) or BDNF (Neveu and Arenas (1996); Lindholm et al. (1997)), play a major role in cerebellar development, survival and maturation of cells. Neurotrophin-3 promotes differentiation and survival of Purkinje cells by increasing branching of trees (Lindholm et al. (1993); Neveu and Arenas (1996)). Their receptor, TrKC, is also essential for these effects. Only Granule cells in the cerebellum express this neurotrophin, whereas TrKC is present in all neurons. Despite the shortage arborization of Purkinje cells, hmTRβ1? has no effect on the expression of NT-3 RNA, which is decreased normally in hypothyroidism. TRβ1 could possibly regulate TrKC pathway in Purkinje cells, causing a decrease of branching. This may explain results obtained by Morte on the lack of effect of GC-1 to restore Purkinje cell arborization in hypothyroid animals (Morte et al. (2002)). BDNF preferentially stimulates proliferative granule cell neurite outgrowth and migration across ML (Segal et al. (1992)). Recent results on Knock-out BDNF show its implication on granule cell migration (Borghesani et al. (2002)). Interestingly, these mice present relative same defects than those seen in our mouse model (Schwartz et al. (1998)). Like for the NT-3 receptor, TRβ1 could regulate the TrKB pathway in Purkinje cells, abolishing BDNF action by Purkinje cells on granule cell maturation.

Our results indicate the major role played by Purkinje cells and TRβ1 in the postnatal cerebellum and also in the essential repression of gene transcription by unliganted RT in the pathogenesis of hypothyroidism. Also, hmTRβ1 is specifically expressed in neurons demonstrating that the morphological changes do not result from the blockage of RT in glial cells. In summary, our study strongly suggests that TRβ1 plays a major role in the effects of TH on the developing cerebellum and that these effects are the consequence of a specific interaction of TH with TRβ1 in Purkinje cells.

  1. References

Abel ED, Ahima RS, Boers ME, Elmquist JK, Wondisford FE (2001). Critical role for thyroid hormone receptor beta2 in the regulation of paraventricular thyrotropin-releasing hormone neurons. J Clin Invest 107:1017-23.

Abel ED, Kaulbach HC, Campos-Barros A, Ahima RS, Boers ME, Hashimoto K, Forrest D, Wondisford FE (1999). Novel insight from transgenic mice into thyroid hormone resistance and the regulation of thyrotropin. J Clin Invest 103:271-9.

Anderson GW (2001). Thyroid hormones and the brain. Front Neuroendocrinol 22:1-17.

Beck-Peccoz P, Chatterjee VK (1994). The variable clinical phenotype in thyroid hormone resistance syndrome. Thyroid 4:225-32.

Borghesani PR, Peyrin JM, Klein R, Rubin J, Carter AR, Schwartz PM, Luster A, Corfas G, Segal RA (2002). BDNF stimulates migration of cerebellar granule cells. Development 129:1435-42.

Bradley DJ, Towle HC, Young WS, 3rd (1992). Spatial and temporal expression of alpha- and beta-thyroid hormone receptor mRNAs, including the beta 2-subtype, in the developing mammalian nervous system. J Neurosci 12:2288-302.

Brinster RL, Chen HY, Trumbauer M, Senear AW, Warren R, Palmiter RD (1981). Somatic expression of herpes thymidine kinase in mice following injection of a fusion gene into eggs. Cell 27:223-31.

Calza L, Forrest D, Vennstrom B, Hokfelt T (2000). Expression of peptides and other neurochemical markers in hypothalamus and olfactory bulb of mice devoid of all known thyroid hormone receptors. Neuroscience 101:1001-12.

Chan SO, Peng D, Chiu FC (1997). Heterogeneous expression of neurofilament proteins in forebrain and cerebellum during development: clinical implications for spinocerebellar ataxia. Brain Res 775:107-18.

Charron G, Guy LG, Bazinet M, Julien JP (1995a). Multiple neuron-specific enhancers in the gene coding for the human neurofilament light chain. J Biol Chem 270:30604-10.

Charron G, Julien JP, Bibor-Hardy V (1995b). Neuron specificity of the neurofilament light promoter in transgenic mice requires the presence of DNA unwinding elements. J Biol Chem 270:25739-45.

Chassande O, Fraichard A, Gauthier K, Flamant F, Legrand C, Savatier P, Laudet V, Samarut J (1997). Identification of transcripts initiated from an internal promoter in the c-erbA alpha locus that encode inhibitors of retinoic acid receptor- alpha and triiodothyronine receptor activities. Mol Endocrinol 11:1278-90.

Clements JA, Matheson BA, Wines DR, Brady JM, MacDonald RJ, Funder JW (1988). Androgen dependence of specific kallikrein gene family members expressed in rat prostate. J Biol Chem 263:16132-7.

Forrest D, Hanebuth E, Smeyne RJ, Everds N, Stewart CL, Wehner JM, Curran T (1996). Recessive resistance to thyroid hormone in mice lacking thyroid hormone receptor beta: evidence for tissue-specific modulation of receptor function. Embo J 15:3006-15. abs.html.

Gilad GM, Gilad VH, Dahl D (1989). Expression of neurofilament immunoreactivity in developing rat cerebellum in vitro and in vivo. Neurosci Lett 96:7-12.

Glass CK, Holloway JM (1990). Regulation of gene expression by the thyroid hormone receptor. Biochim Biophys Acta 1032:157-76.

Gothe S, Wang Z, Ng L, Kindblom JM, Barros AC, Ohlsson C, Vennstrom B, Forrest D (1999). Mice devoid of all known thyroid hormone receptors are viable but exhibit disorders of the pituitary-thyroid axis, growth, and bone maturation. Genes Dev 13:1329-41.

Grouselle D, Tixier-Vidal A, Pradelles P (1990). Enzyme immunoassays for thyroliberin (TRH) and TRH-elongated peptides in mouse and rat hypothalamus. Neuropeptides 17:155-62.

Hashimoto K, Curty FH, Borges PP, Lee CE, Abel ED, Elmquist JK, Cohen RN, Wondisford FE (2001). An unliganded thyroid hormone receptor causes severe neurological dysfunction. Proc Natl Acad Sci U S A 98:3998-4003.

Hayashi Y, Mangoura D, Refetoff S (1996). A mouse model of resistance to thyroid hormone produced by somatic gene transfer of a mutant thyroid hormone receptor. Mol Endocrinol 10:100-6.

Hayashi Y, Xie J, Weiss RE, Pohlenz J, Refetoff S (1998). Selective pituitary resistance to thyroid hormone produced by expression of a mutant thyroid hormone receptor beta gene in the pituitary gland of transgenic mice. Biochem Biophys Res Commun 245:204-10.

Lebel JM, Dussault JH, Puymirat J (1994). Overexpression of the beta 1 thyroid receptor induces differentiation in neuro-2a cells. Proc Natl Acad Sci U S A 91:2644-8.

Legrand J (1982). [Thyroid hormones and maturation of the nervous system]. J Physiol 78:603-52.

Lindholm D, Castren E, Tsoulfas P, Kolbeck R, Berzaghi Mda P, Leingartner A, Heisenberg CP, Tessarollo L, Parada LF, et al. (1993). Neurotrophin-3 induced by tri-iodothyronine in cerebellar granule cells promotes Purkinje cell differentiation. J Cell Biol 122:443-50.

Lindholm D, Hamner S, Zirrgiebel U (1997). Neurotrophins and cerebellar development. Perspect Dev Neurobiol 5:83-94.

Mellstrom B, Naranjo JR, Santos A, Gonzalez AM, Bernal J (1991). Independent expression of the alpha and beta c-erbA genes in developing rat brain. Mol Endocrinol 5:1339-50.

Morte B, Manzano J, Scanlan T, Vennstrom B, Bernal J (2002). Deletion of the thyroid hormone receptor alpha 1 prevents the structural alterations of the cerebellum induced by hypothyroidism. Proc Natl Acad Sci U S A 99:3985-9.

Murata Y (1998). Multiple isoforms of thyroid hormone receptor: an analysis of their relative contribution in mediating thyroid hormone action. Nagoya J Med Sci 61:103-15.

Neveu I, Arenas E (1996). Neurotrophins promote the survival and development of neurons in the cerebellum of hypothyroid rats in vivo. J Cell Biol 133:631-46.

Oppenheimer JH, Schwartz HL, Strait KA (1994). Thyroid hormone action 1994: the plot thickens. Eur J Endocrinol 130:15-24.

Puymirat J (1992). Thyroid receptors in the rat brain. Prog Neurobiol 39:281-94.

Puymirat J, l'Hereault S, Dussault JH (1992). Expression of thyroid hormone receptors mRNAs in rat cerebral hemisphere neuronal cultures. Brain Res Dev Brain Res 69:173-7.

Refetoff S, Weiss RE, Usala SJ (1993). The syndromes of resistance to thyroid hormone. Endocr Rev 14:348-99.

Sakurai A, Miyamoto T, Refetoff S, DeGroot LJ (1990). Dominant negative transcriptional regulation by a mutant thyroid hormone receptor-beta in a family with generalized resistance to thyroid hormone. Mol Endocrinol 4:1988-94.

Sap J, Munoz A, Damm K, Goldberg Y, Ghysdael J, Leutz A, Beug H, Vennstrom B (1986). The c-erb-A protein is a high-affinity receptor for thyroid hormone. Nature 324:635-40.

Schwartz PM, Levy RL, Borghesani PR, Segal RA (1998). Cerebellar pathology in BDNF -/- mice: the classic view of neurotrophins is changing. Mol Psychiatry 3:116-20.

Segal RA, Takahashi H, McKay RD (1992). Changes in neurotrophin responsiveness during the development of cerebellar granule neurons. Neuron 9:1041-52.

Strait KA, Schwartz HL, Perez-Castillo A, Oppenheimer JH (1990). Relationship of c-erbA mRNA content to tissue triiodothyronine nuclear binding capacity and function in developing and adult rats. J Biol Chem 265:10514-21.

Weinberger C, Thompson CC, Ong ES, Lebo R, Gruol DJ, Evans RM (1986). The c-erb-A gene encodes a thyroid hormone receptor. Nature 324:641-6.

Williams GR (2000). Cloning and characterization of two novel thyroid hormone receptor beta isoforms. Mol Cell Biol 20:8329-42.

Wong R, Vasilyev VV, Ting YT, Kutler DI, Willingham MC, Weintraub BD, Cheng S (1997). Transgenic mice bearing a human mutant thyroid hormone beta 1 receptor manifest thyroid function anomalies, weight reduction, and hyperactivity. Mol Med 3:303-14.

Figure legends

Figure 1. a ) Schematic representation of the transgene construct. The black box represents the cDNA of hmTRβ1 and the white box represents the 5'- 292 pb human NF-L promoter The mutation in the hormone binding domain of the hTRβ1 is a shift of glycine (G) 345 to arginine (R). Squared boxes represent the four exons of the hNF-L gene and the inter-boxes correspond to hNF-L introns. Grey boxes indicate poly(A) site of BGH (after the hmTRβ1 cDNA) and of hNF-L (after the last intron). ( b ) Northern blot analysis of transgene expression in different tissues of P15 Tg-1 and -2 mice and Wt mice. 20 μg of total RNA from kidney, spleen, liver, lung, heart and brain was analyzed with a probe corresponding to the +325-+545 fragment of hmTRβ1. The bottom panel corresponds to the hybridization of the same gel with a 18S rRNA-specific oligonucleotide. ( c ) Northern blot prepared with 20 μg total RNA derived from 10-day old neuronal cultures (1) or 15-day old mix glial cultures (2) and probed with hmTRβ1 fragment. ( d ) Western blot analysis: One hundred (100) μg of nuclear protein prepared from P10 cerebellum of Tg and Wt mice was separated on a 10% SDS-polyacrylamide gel and revealed with monoclonal antibodies directed against either a NH2-terminal hTRβ1 epitope or an unknown nuclear protein unregulated by TH (2B3). ( e ) Immunohistochemical staining of P20 Tg and Wt cerebellum with monoclonal antibodies directed against hTRβ1. The weak staining in the Wt mouse slide is due to cross reaction with endogenous TRβ1.

Figure 2. ( a ) Northern blot analysis of endogenous TRβ1, TRα1 and TRα2 expression. Twenty (20) μg of total RNA from P15 Tg and Wt cerebellum mice were hybridized with a rat TRβ1 or TRα specific probe. ( b ) RT-PCR of neurotrophin-3 from P5 Tg , Wt and Wt treated with PTU mice. Numbers behind the ethidium bromide-stained gel image represents mean of fold-increase of each quantification (NT-3/cyclophilin) from three experiments. ***, P<00001.

Figure 3. Cerebellar development in transgenic mice. ( a ) atrophia of cerebellum in P15 Tg mice. ( b ) Cresyl violet-stained cerebellum sections at P5, P10, P15 and P20. ( c ) Cerebellar cortex from P10 and P15 Tg and Wt mice. * represents absence of EGL in P15 Wt mice. ( d ). Quantification of the size of molecular layer (ML) at P10 and P15 and external granular layer (EGL) at P10 of Tg-1 mice ***, P<00001

Figure 4. Immunohistochemical staining of Purkinje cells at P5, P10 and P15 in Tg-1 and -2 mice and Wt mice. ( a ) Cerebellum sections were stained with monoclonal anti-calbindin D28K. Areas of Purkinje cell arborization ( b ) and cell bodies ( c ) ( n=12 for Tg-1 mice and n=10 for Wt mice) were measured from two different animals per condition at P10. ***, P<00001

The number of Purkinje cells per 500 μm was counted in Tg-1 and Wt mice

Figure 5. ( a ) Analysis of granule cell proliferation in the External Granular Layer (EGL) by BrdU incorporation. Ten-day old Tg-1 mice and Wt mice were injected with BrdU 6 hours before brain isolation. Proliferation was expressed by the number of pixels in the EGL by mm2. ( b ) Apoptosis analysis of granule cells in 15-day old Tg-1 mice and Wt mice by TUNEL staining. Apoptotic cells were counted in all layers of the cerebellum cortex (EGL, ML and IGL) **, P<001 and ***, P<00001

Figure 6. ( a ) Expression of BTEB mRNAs in the developing cerebellum. Total RNA was prepared at different time periods (P5 to P20). Twenty (20) μg of total RNA was analyzed by Northern blot using a BTEB cDNA probe. All blots were normalized with a cyclophilin cDNA probe. At right, BTEB expression at P15 Tg-1, Tg-2 and Wt mice ( b ). Quantification of BTEB mRNA in the developing cerebellum of Tg-1 mice and Wt mice **, P<001 and *, P<005 cerebellum of Tg-1 mice and Wt mice **, P<001 and *, P<005