Thyroid hormone receptor [beta] mutation causes severe impairment of cerebellar development

Molecular and Cellular Neuroscience 44 (2010) 68–77

Contents lists available at ScienceDirect

Molecular and Cellular Neuroscience

j ourna l homepage: www.e lsev ie r.com/ locate /ymcne

Thyroid hormone receptor β mutation causes severe impairment ofcerebellar development

Aline Cristina Portella a, Fernando Carvalho a, Larissa Faustino b, Fredric E. Wondisford c,Tânia Maria Ortiga-Carvalho b, Flávia Carvalho Alcantara Gomes a,⁎a Instituto de Ciências Biomédicas, Universidade Federal do Rio de Janeiro, 21941-590, Rio de Janeiro, RJ, Brazilb Instituto de Biofísica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, 21941-590, Rio de Janeiro, RJ, Brazilc Department of Pediatrics, Division of Metabolism, Johns Hopkins University Medical School, Baltimore, Maryland 21287, USA

⁎ Corresponding author. Instituto de Ciências BioméRio de Janeiro, Centro de Ciências da Saúde, Bloco F, IlhaJaneiro, RJ, Brazil.

E-mail address: [email protected] (F.C.A. Gomes

1044-7431/$ – see front matter © 2010 Elsevier Inc. Adoi:10.1016/j.mcn.2010.02.004

a b s t r a c t

Article history:Received 25 September 2009Revised 28 December 2009Accepted 19 February 2010Available online 1 March 2010

Keywords:CerebellumThyroid hormoneTRβNuclear receptorsFoliationBergmann glia

Cerebellar development on the postnatal period is mainly characterized by cellular proliferation in theexternal granular layer (EGL) followed by migration of granular cells in the molecular layer through theBergmann glia (BG) fibers in order to form the granular layer in the adult. All these events are drasticallyaffected by thyroid hormones (TH), which actions are mainly mediated by alpha (TRα) and beta (TRβ)nuclear receptor isoforms. Here, we analyzed the effects of a natural human mutation (337T) in the TRβlocus, which impairs T3 binding to its receptor, on the mouse cerebellum ontogenesis. We report that targetinactivation of TRβ-TH binding leads to a smaller cerebellum area characterized by impaired lamination andfoliation. Further, TRβ mutant mice presented severe deficits in proliferation of granular precursors,arborization of Purkinje cells and organization of BG fibers. Together, our data suggest that the action of THvia TRβ regulates important events of cerebellar ontogenesis contributing to a better understanding of someneuroendocrine disorders. Further, our data correlate TRβ with cerebellar foliation, and provide, for the firsttime, evidence of a receptor-mediated mechanism underlying TH actions on this event.

dicas, Universidade Federal dodo Fundão, 21941-590, Rio de

).

ll rights reserved.

© 2010 Elsevier Inc. All rights reserved.

Introduction

The cerebellum is an excellent model to study mechanisms thatcontrol central nervous system (CNS)morphogenesis. Besides few celltypes, the cerebellar cells undergo sequential steps of development inspatially well-defined regions, leading to a relatively simple and well-known laminar organization (Altman and Bayer, 1997; Corrales et al.,2006; Sillitoe and Joyner, 2007).

The cerebellum is located on the back of the brain stem at themidbrain–hindbrain junction, traditionally associated to coordinatingproprioceptive–motor functions. Emerging evidences from experi-mental animal models and cerebellar disorders in humans haverecently implicated this structure in higher activities such ascognition, emotion, memory and language processing (Baillieuxet al., 2008; Callu et al., 2007; Tavano et al., 2007).

The mammalian cerebellum consists of a central vermis and twolateral hemispheres, each with its own set of fissures and folia. Thenoticeable morphological feature of the mammalian cerebellum ischaracterized by a coordinated three-dimensional complexity ofmedial–lateral and antero-posterior domains generated by a multi-

factorial phenomenon known as foliation (Sillitoe and Joyner, 2007).In mice, cerebellar foliation is characterized by transition from thesmoothed cerebellar surface to the X lobule cerebellum, a patternmostly achieved 2–3 weeks after birth (Altman and Bayer, 1997).Mapping and physiological studies implied a correlation betweenspecific folia and sensory-motor tasks (Sillitoe and Joyner, 2007;Sotelo, 2004). The conservation of the foliation pattern suggests thatthis event might be tightly genetically regulated, although it remainsunknown how position of fissures, folia size and complexity aredetermined.

Cerebellar granule precursors arise from the neuroepithelium atthe rhombic lip that forms the posterior boundary to the cerebellarprimordium. In rodents, granular cell precursors (GCPs) migraterostrally over the surface of the cerebellum during the second half ofembryogenesis, forming a second and transient germinal zone, theexternal granular layer (EGL). During the two postnatal weeks, theEGL precursors proliferate extensively and generate postmitoticgranular cells, which migrate inward through Bergmann glia (BG)fibers, bypass the Purkinje cell layer (PCL), and finally generate theinternal granular layer (IGL) (Miale and Sidman, 1961; Sotelo, 2004).

Cerebellar ontogenesis undergoes dramatic modulation by thyroidhormones (THs) (Clos and Legrand, 1973; Clos et al., 1980; Gomeset al., 2001; Martinez and Gomes, 2002, 2005). Hypothyroidism isassociated with several abnormalities in the cerebellar cortex such aspersistence of the EGL, increased neuronal death in the IGL, impaired

69A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

migration of granular cells, and deficient arborization of Purkinje celldendritic trees, followed by an altered foliation pattern (Nicholsonand Altman, 1972; Lauder et al., 1974; Lauder, 1977, 1978; Morteet al., 2002, Heuer and Mason, 2003). The molecular mechanismsunderlying such TH modulated-cerebellar foliation are completelyunknown.

TH receptors (TRs) are members of the nuclear receptor (NRs)superfamily of ligand-modulated transcriptional factors (Lazar, 2003),which includes receptors for steroid hormones, retinoic acids (RARs),1,25-dihydroxyvitamin D3, and others (Heikkinen et al., 2007). Ligandbinding to the NRs leads to receptor conformational change thatrecruits co-activators to the DNA-bound receptor. For some non-steroid NRs, TRs and RARs, the unliganded receptor interacts withcorepressors, including NCoR (nuclear receptor corepressor) andSMRT (silencing mediator of retinoid and thyroid hormone receptors)(Tagami et al., 1998), to mediate transcriptional repression, charac-terizing a ligand-independent function.

Two distinct genes encode three different isoforms capable ofbinding T3 (Triiodothyronine); TRα1, TRβ1, and TRβ2, which aredifferently expressed throughout tissues (Lazar, 2003; Ortiga-Carvalho et al., 2005). The physiological and specific functions of theTR isoforms remain largely unidentified. Data generated from mutantTRs and hypothyroid mice reveal sharp discrepancies, and suggest anintricate interplay between the different isoforms of TR (Hashimotoet al., 2001; Morte et al., 2002, 2004; Heuer and Mason, 2003;Manzano et al., 2003).

Fig. 1. TRβ mutation affects cerebellar morphology. Cresyl-violet staining of cerebellar paraTRβΔ337T/Δ337T (C). Morphometric measurements of cerebellum (D), lobule IX (E), ML (F), andthe absence of the fissure that splits lobule VI into VIa and VIb in TRβΔ337T/Δ337T. Note the a***pb0.001 (n=4 per group). I–X, cerebellar folium I–X; D, dorsal; V, ventral; P, posterior;

In order to evaluate the effect of a non-binding TRβ in cerebellarmorphogenesis, we used mice carrying a natural human mutation(Δ337 T) in the TRβ locus (Hashimoto et al., 2001). We report that theTRβ mutation dramatically affects cerebellar ontogenesis, character-ized by impaired granular cell precursors (GCPs) proliferation andBergmann glia and Purkinje cell development, followed by a severedeficit of folia formation. Our data correlate THs with cerebellarfoliation and lamination, and provide, for the first time, evidence ofTRβ role on this event.

Results

TRβ Δ337T mutation impairs cerebellar laminar structure and foliation

In order to address TRβ effects on cerebellar development, weanalyzed parasagittal sections of the cerebellar vermis of P21 animals,an age in which the final cerebellar foliation pattern is achieved.TRβwt/Δ337T and TRβΔ337T/Δ337T P21 mice showed decreased cerebellarsize (18% and 30%, respectively), represented by reductions in the MLand GL areas (Figs. 1D–G). The EGL was absent in cerebella from bothwild-type and mutated P21 mice (Supplemental Fig. S1), a featurecharacteristic of non-hypothyroid mice (Lauder et al., 1974).

In P21 mice, the cerebellar foliation pattern is characterized by thepresence of ten well-formed lobules and the appearance of some sub-lobules; lobule VI is subdivided into sub-lobules VIa and VIb by anadditional fissure, as shown for TRβwt/wt and TRβwt/Δ337T mice

sagittal sections from postnatal-day 21 (P21) mice: TRβwt/wt (A), TRβwt/Δ337T (B), andGL (G) areas revealed under-developed features in P21mice. Red asterisks in C indicatebsence of the EGL in all P21 phenotypes. Data are expressed as mean±SEM. *pb0.05;A, anterior; ML, molecular layer; GL, granular layer. Scale bar: 1 mm.

70 A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

(Figs. 1A and B). However, this fissure fails to form in TRβΔ337T/Δ337T

mice, where lobule VI remains fused (Fig. 1C, asterisk).In order to document the temporal onset and progression, as well

as the cellular basis, of TRβ effects on cerebellar development, wefocused on initial period of development, whenmost of cellular eventsof cerebellar development are still taking place. Cresyl-violet stainingof parasagittal sections of the cerebellar vermis of P9 mice revealed alarge decrease in the area of the TRβ mutant mouse cerebellum(Figs. 2A, D and G), with a reduction of 27% and 33% in the area of thetotal cerebellum (Fig. 2J) and lobule IX (Fig. 2K), respectively. In

Fig. 2.Deficit in foliationpatternand laminarorganizationof TRβmutantmice.Cresyl-violet staiTRβΔ337T/Δ337T (G, H, and I) postnatal-day 9mice. Black arrows in B and E indicate thefissure betand I indicate analteration in lobule IXdevelopment. Cerebellar layerswereanalyzed from lobulmicewereobserved.Data are expressed asmean±SEM. *pb0.05; ***pb0.001. TRβwt/wt (n=8);P, posterior; A, anterior; EGL, external granular layer; ML, molecular layer; IGL, internal granul

addition, the TRβ Δ337T mutation also dramatically impaired fissureand lobule formation. At the beginning of the second postnatal week,themouse cerebellar foliation pattern is characterized by the presenceof ten lobules (I–X), separated by seven well-defined fissures(Figs. 2A–C). In contrast, TRβΔ337T/Δ337T mice fail to form the fissurebetween lobules VI–VII (Fig. 2H), which remain fused. Additionally,lobule IX is severely affected (Fig. 2I), leading to a poor morphology incomparison to wild-type cerebellum. These alterations were lessevident in TRβwt/Δ337T mice (Figs. 2D–F), although some phenotypicvariations might exist among different animals.

ningof cerebellarparasagittal sections ofTRβwt/wt (A, B, andC), TRαwt/Δ337T (D, E, and F) andween lobules VI and VII, which is absent in TRβΔ337T/Δ337T (red arrow inH). Red arrows in Fe IX.Decreases in cerebellum(J), lobule IX (K), EGL (L),ML (M), and IGL (N) areas inmutantTRβwt/Δ337T (n=7); TRβΔ337T/Δ337T (n=7). I–X, cerebellar folium I–X;D, dorsal; V, ventral;ar layer. Scale bars: 250 μm.

71A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

Here,we found an important decrease in all layers of TRβΔ337T/Δ337T

lobule IX, especially in the EGL: 38% decrease (Fig. 2 L); ML: 25%decrease (Fig. 2 M); and IGL: 33% decrease (Fig. 2N). TRβwt/Δ337T

mice were not statistically different from the wild-type mice.Because the proliferation rate of EGL precursors varies according to

the lobule region, we analyzed the EGL by dividing it into three distinctregions: region A: between lobules VIII–IX; region B: EGL from the edgeof lobule IX without contact with any lobule; region C: between lobulesIX–X(Fig. 3G). Both TRβwt/Δ337T and TRβΔ337T/Δ337T showedadecrease inthe EGL from all regions (Figs. 3A–F); however, this alterationwasmoreevident in the homozygote animal [TRβwt/Δ337T: reduction of 18%, 9%,and 12%; TRβΔ337T/Δ337T: reduction of 38%, 33%, and 36% respectively forregions A (Fig. 3H), B (Fig. 3I), and C (Fig. 3 J)].

The EGL thickness also varied within the same lobule of an animal,being thicker between lobules (regions A and C) and thinner at theedge (region B). This feature was preserved in all three genotypes(Supplemental Fig. S2). Together, these findings indicate that TRβ is

Fig. 3. Decreased EGL thickness in TRβmutantmice. Micrographs of cerebellar cortex fromdiffeand TRβΔ337T/Δ337T (C and F) of postnatal-day 9 mice. Red details in (A–F) delimitate EGL thicknregions show a smaller EGL thickness inmutant mice (H–J). Data are expressed as mean±SEMexternal granular layer; ML, molecular layer; PCL, Purkinje cell layer; IGL, internal granular lay

required for the cerebellar laminar organization and foliation pattern,and suggest a key function for TRβ in EGL maintenance.

Impaired GCPs proliferation in TRβ mutant mice

At the end of the third postnatal week in mice, the EGL hascompletely disappeared; its maintenance is a balance between theproliferation rate of GCPs and their exit from the layer, events that areseverely affected by TH (Nicholson and Altman, 1972; Lauder et al.,1974). In order to assesswhether the defect in TRβ signaling impairs theexpansion of the GCPs pool, we performed in-vitro and in-vivo BrdUlabeling experiments (Fig. 4). In-vitro analysis revealed that granularprogenitors derived from TRβΔ337T/Δ337Tmice showed an 18% reductionin BrdU incorporation (Figs. 4A–D and G), suggesting a deficiency in theproliferation of these cells. This effect was not due to in vitro impairedsurvival of GCPs, because we did not detect any obvious alteration incultures of TRβ mutant granular neurons (Supplemental Fig. S3). The

rent lobule IX regions stained by cresyl-violet of TRβwt/wt (A and D), TRβwt/Δ337T (B and E),ess. Cerebellum in G illustrates the three regions where EGL thickness was measured. All. **pb0.01; ***pb0.001. TRβwt/wt (n=7); TRβwt/Δ337T (n=6); TRβΔ337T/Δ337T (n=5). EGL,er. Scale bar: 50 μm.

Fig. 4. Impaired GCPs proliferation and altered granular cell migration in TRβ mutant mice. Cerebellar cultures (n=2 per group) prepared from postnatal-day 7 TRβwt/wt (A and C)and TRβΔ337T/Δ337T (B and D) mice were incubated for 24 h in the presence of BrdU (1 μg/ml), followed by quantification of the number of BrdU-positive cells (G). For in-vivo assays,postnatal-day 9 (n=3 per group) TRβwt/wt (E) and TRβΔ337T/Δ337T (F) mice were intraperitoneally injected with BrdU (100 mg/kg), and the number of EGL/BrdU-positive cellsanalyzed after 1 h of BrdU incorporation (H). Insets in E and F depict DAPI staining of the same field. The TRβ Δ337Tmutation greatly decreased the proliferation rate of granular cellprecursors. Data are expressed as mean±SEM. **pb0.01; ***pb0.001. EGL, external granular layer; ML, molecular layer; PCL, Purkinje cell layer; IGL, internal granular layer. VIII,lobule VIII; IX, lobule IX. Scale bars: 50 μm.

72 A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

impaired proliferation of TRβ mutant GCPs precursors was furtherconfirmed by in-vivo analysis of the EGL of mutant mice. Similarly, weobserved a 14% decrease in the number of BrdU-positive cells in the EGLof TRβ mutant mice (Figs. 4E–F and H), indicating that the TRβ Δ337Tmutation impairs the proliferation of GCPs.

TRβ Δ337T mutation induces abnormal Bergmann glia and Purkinjecell development

THs modulate the maturation and the number of Bergmann gliacells in vitro and in vivo (Clos and Legrand, 1973; Messer et al., 1985).In order to assess whether TRβ Δ337T mutation impairs BGdevelopment, we first employed a BG/GCPs mix culture from thepostnatal cerebellum. Whereas most of the glial cells derived fromTRβwt/wt (Figs. 5A and D) and TRβwt/Δ337T (Figs. 5B and D) miceexhibited an elongated morphology, characteristic of BG cells, thegreat majority of those from TRβΔ337T/Δ337T (Figs. 5C and D) showed aprotoplasmic morphology (75%). Remarkably, the culture of TRβΔ337T/

Δ337T glial cells showed a 50% reduction in the glial population derivedfrom the mutated cerebellum (Fig. 5E).

We next investigated BG organization in vivo, through immunos-taining assays for the glial marker GFAP in parasagittal sections of P9mice lobule IX (Figs. 5F–K). In contrast to in-vitro data, arborization ofBG fibers was not significantly abnormal in mutant mice (Figs. 5H–K);nevertheless, supporting the in-vitro data, the number of BG fiberswas greatly decreased in both TRβwt/Δ337T (15%) and TRβΔ337T/Δ337T

(25%) (Fig. 5 L), suggesting that TRβmight be a mediator of TH actionin BG maturation.

Because the PC–BG interaction is a crucial step in cerebellar celldevelopment (Yamada et al., 2000; Sotelo, 2004), we sought to analyzePC in the early stageof cerebellardevelopment. PC andBGare intimatelyassociated in all phenotypes, as evidenced by the BGprocesses spanningthe entire ML from PCL to the pia (Figs. 5F–K). Consistent with previousresults, Calbindin-D28K protein immunodetection revealed a dramaticdecrease in the branching of PC (Figs. 5M–O) in both P9 TRβwt/Δ337T

(Figs. 5N and N') and TRβΔ337T/Δ337T (Figs. 5O and O'). Thus, our datashow that the maturation of PC and BG cells strongly depends on TRβ,and support the hypothesis that the cerebellar malformations found inTRβΔ337T/Δ337T might be associated with deficit in cellular interactions.

Discussion

We demonstrated in the present study that TRβ is required forTH-induced cerebellar ontogeny. This is supported by observationsthat TRβ mutants develop severe abnormalities, including defects incerebellar foliation and laminar organization, associated with impairedPurkinje cell arborization, a deficit in the number of Bergmann gliafibers, and a decreased rate of GCPs proliferation. Our study has madetwo main contributions. Primarily, it is the first to indicate a receptor-mediated mechanism (TRβ-dependent) by which TH elicits cerebellarfoliation; additionally, our data support the idea that cerebellardevelopment is the result of an interplay between two TH receptorssegregated in different cell types: TRα, mainly expressed by GCPs(Bradley et al., 1992), and TRβ, expressedbyPCs (Strait et al., 1991); thussupporting the role of cellular interactions in cerebellar morphogenesis.

Fig. 5. TRβmutation leads to abnormal Bergmann glia and Purkinje cell development. A–E) In-vitro assays: Mixed cerebellar cultures prepared from postnatal-day 7 (n=2 per group) TRβwt/wt, TRβwt/Δ337T, and TRβΔ337T/Δ337T mice were grown for3 days and stained for the glial marker GFAP. Morphology (D) and number (E) of GFAP-positive cells were evaluated. Most of the TRβwt/wt (A) and TRβwt/Δ337T (B) GFAP-positive cells showed an elongated form, characteristic of Bergmann glia cells;whereasmost of the TRβΔ337T/Δ337T glial cellswere poorly radialized (C). Quantitative analysis revealed ahuge decrease in the number of GFAP-positive cells in TRβΔ337T/Δ337T (E). F–O) In-vivo assays: Confocalmicrographs ofGFAP (green; F–K) and PCmarker, Calbindin-D28K (red; G, I, K;M–O') stainingof cerebellar parasagittal sections of TRβwt/wt (F, G,M,M'), TRβwt/Δ337T (H, I, N, N'), and TRβΔ337T/Δ337T (J, K, O, O') P9mice (n=3per group). TRβwt/Δ337T and TRβΔ337T/Δ337Tmice showed adecrease inthenumber of BGgliafibers that reach the pial surface invivo (L), aswell as a dramatic decrease in thedendritic arborization of Purkinje cells. Data are expressed asmean±SEM. *pb0.05. EGL, external granular layer;ML,molecular layer; PCL, Purkinjecell layer; IGL, internal granular layer; VIII, lobule VIII; IX, lobule IX. Scale bars: 20 μm (C, J, K); 50 μm (O–O'). 73

A.C.Portella

etal./

Molecular

andCellular

Neuroscience

44(2010)

68–77

74 A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

The TRβmutatedmice used in this study constitute an animalmodelfor thyroid hormone resistance (TRH), a syndrome characterized byclinical hypothyroidism despite high levels of TH (Refetoff et al., 1967).The TRβ mutation generates mice with a hypothyroid state for thosetissues that express this isoform; however, because of high levels of TH,tissues with TRα expression display the hyperthyroid phenotype.Analysis of P21 mice did not detect EGL in mutant mice (SupplementalFig. S1), suggesting that TRβ mutation does not lead to a classichypothyroid cerebellum. Importantly, expression of the mutated TRβisoform does not affect expression of TRα (Hashimoto et al., 2001).Assessment of the relative contribution of TRα versus TRβ in cerebellardevelopment is a complex task, and the generation of several TRmutants, either knockout or knockin, has provided contradictory data(Hashimoto et al., 2001; Morte et al., 2002, 2004; Heuer and Mason,2003; Manzano et al., 2003). The TRβ mutant is unable to bind T3, andconstitutively binds to nuclear CoRs. Binding of this mutant isoform toanother functional nuclear receptor (such as TRα, TRβ, or RXR) inhibitsits function, i.e., has a negative dominant effect (Pazos-Moura et al.,2000). This mutation also shows ligand-independent effects, which aresupported by the greater severity, depending on the system, of thehomozygous mouse phenotype compared to mice lacking TR. This isconsistent with observations that TRβ−/− (Forrest et al., 1996) andTRα−/− animals (Morte et al., 2002, 2004) do not show obviousabnormalities in the cerebellum similar to those observed in the presentstudy. We previously demonstrated that the TRβmutation has few andvariable effects on heart function (Ortiga-Carvalho et al., 2004) andretina development (Pessôa et al., 2008) when in heterozygosis. Asimilar pattern was observed in this study: the mean values of most ofthe cerebellar parameters of heterozygotemice were always near wild-type values; however, individual values and measurements variedwidely, similarly to humans expressing the mutation in heterozygosis.To simplify, we will mainly discuss wild-type versus homozygousmutant mice.

One of the most morphological features of the cerebellum is itsthree-dimensional complexity driven by the postnatal foliationprocess. The functional roles of lobules, as well as the parasagittalmolecular domains and topographical circuitry organization of thecerebellum, are still speculative. However, increasing evidence hassuggested that the folia might serve as a platform for organizingcerebellum circuitry, because each lobule is represented by differentcombinations of termination of afferents, which supply the cerebel-lum with specific sensory and motor information (Sillitoe and Joyner,2007). A correlation between THs levels and foliation has beensuggested (Lauder et al., 1974), although a molecular mechanismunderlying this effect has not been identified. In the present study, wedemonstrated that cerebella from TRβ mutant mice show abnormalfoliation during development.

One of the mechanisms proposed for fissure and lobule formation isthe relationship between the rapidly expanding superficial sheet of theEGL, through proliferation of GCPs, and the more slowly differentiatingunderlying cortex (Doughty et al., 1998; Corrales et al., 2006). In supportof a correlation between GCPs proliferation and cerebellar folding,genetic and experimental manipulations where GCPs are partiallyeliminated through irradiation (Doughty et al., 1998) or throughdownregulation of Shh signaling, a morphogen known to increaseGCPs proliferation (Corrales et al., 2006), both produce cerebellumwithsimple foliation pattern. This is also in agreementwith thepoor foliationpattern found in hyperthyroidism,which causes premature depletion ofthe EGL, in contrast to hypothyroidism, inwhich themaintenance of theEGL is associated with increased foliation (Lauder et al., 1974).Unliganded TRβ leads to decreased size of the EGL, associated withimpairedproliferation ofGCPs. Because TRβmutants showhigh levels ofTH, the observed effects on GCPs could be mediated by TRα expressedby GCPs. Although we cannot completely rule out these effects; this isunlikely to be responsible for all the events described here, because wedid not observe increased GCPs proliferation in TRβ mutant mice, a

characteristic of the hyperthyroid cerebellum (Supplemental Fig. S4).GCPs proliferation is highly dependent on interactions with Purkinjecells (PCs), as revealed by decrease in GCPs number in several workswhere PCs were lost after cell ablation or genetic mutation (Herrup,1983; Smeyne et al., 1995; Wetts and Herrup, 1982). The fact that GCPsdo not express TRβ, together with the observation that the TRβmutation leads to a deficit in PCs differentiation (present results andthose from Hashimoto et al., 2001), suggests that the deficit in GCPsproliferation observed in the present study is not cell-autonomous, butrather is caused by a secondary effect on PCs known to provide trophicsupport for the proliferation of GCPs (Dahmane and Ruiz i Altaba, 1999).An example of TH effects on granular cells through PCs has beenpreviously provided by works of Poguet et al., which demonstrated thatGCPs number is affected by cyclin D2 levels regulated by T3 actionthrough GCPs-TRα and PCs-TRβ (Poguet et al., 2003).

The five cardinal lobes seen in all mammals are formed during theembryonic period and determine the principal fissures and lobes.Postnatally, additional fissures successively divide the cardinal lobesinto the 10final lobes of rodents. By 21 posnatal days, thefinal foliationpattern of mice cerebellum is already established. TRβ mutant micepresent a fusion of lobules VI and VII at 21 postnatal days; however theoverall place of primary fissures was conserved suggesting a deficit inthe onset of foliation rather than a complete absence of signals thatmodulate fissure formation. This is also supported by the observationthat EGL thickness in the crownof the foliawas smaller than in the baseof the fissure either in wild-type as well as mutant mice, acharacteristic reported to be essential to fissure formation (Corraleset al., 2004). In addition, TRβ-target inactivation althoughdramaticallyimpaired GCPs proliferation did not completely abolish it either invitro or in vivo. Taken together, our data suggest that mutation in theTRβ gene leads to a decrease in the signals that modulate GCPsproliferation and cerebellar foliation. Although the nature of thesesignals is not completely elucidated, good candidates are themorphogen sonic hedgehog (Shh) (Dahmane and Ruiz i Altaba,1999; Wechsler-Reya and Scott, 1999; Corrales et al., 2006; Chanet al., 2009), β1-class integrins (Graus-Porta et al., 2001), and theextracellular matrix protein, laminin (Blaess et al., 2004), for which adisruption of the signaling pathways closely correlates with impairedGCPs proliferation and cerebellar foliation.

Several genes have been reported to be altered by THs, eitherdirectly or indirectly, in the developing cerebellum (Takahashi et al.,2008), including neurotrophins such as NT-3 and BDNF (Neveu andArenas, 1996; Koibuchi et al., 1999; Manzano et al., 2003), and growthfactors such as EGF (Alvarez-Dolado et al., 1998; Gomes et al., 1999;Martinez and Gomes, 2002, 2005). Our data shed light on theimportance of TH receptors in the regulation (either directly orindirectly) of these genes in vivo, and point, for the first time, to amolecular mechanism, receptor-mediated, by which THs modulatecerebellar development and foliation.

TRβ mutated mice show abnormal numbers and a morphologicaldeficit of BG fibers, suggesting that unliganded TRβ plays animportant role in BG differentiation. These data are consistent withthe finding that the TRα−/− cerebellum shows reduced GFAP andnestin immunostaining, a feature restored by hypothyroidism,suggesting that liganded TRβ is detrimental to astroglial celldifferentiation in the absence of TRα1 (Morte et al., 2004). Wepostulate that even in the presence of TRα, astroglial differentiationmight be the result of an interplay between the two TR isoforms.Alternatively, a deficit in BG fibers might be a secondary effect on PCs,rather than a primary effect on BG, since it is known that BGtransformation proceeds in concert with dendritic differentiation ofPCs (Yamada et al., 2000). Whether TRβ mutation affects neuronalmigration remains to be established, but this would be consistent withthe modulation of migratory molecules by THs such as reelin,tenascin, and laminin (Alvarez-Dolado et al., 1998, 1999; Farwelland Dubord-Tomasetti, 1999; Martinez and Gomes, 2002).

Fig. 6. Molecular model for TRβ-mediated effect in cerebellar development. TRβ mutation leads to deficit in cerebellar foliation, decreased GCPs pool proliferation and impaireddevelopment of Purkinje cells and Bergmann glia differentiation. EGL, external granular layer; ML, molecular layer; PCL, Purkinje cell Layer; IGL, internal granular layer; P9, postnatal-day 9 mice; P21: postnatal-day 21 mice; I–X: cerebellar lobules from I to X.

75A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

TH are essential for several events of brain development in allvertebrates, including neuronal migration and differentiation, synap-togenesis and myelination, although TH receptors involved in theseeffects are still controversial (Bernal and Nunes, 1995; Gomes et al.,2001; Porterfield and Hendrich, 1993; de Escobar et al., 2004). Wesuggest that cerebellar development is dependent on TRβ signaling.Non-binding THβ leads to dramatic changes on cerebellar layersformation, mainly characterized by decrease of GCP proliferation andpoor Purkinje cell arborization and Bergmann glia elongation (Fig. 6).Our data, together with the emerging evidence of the involvement ofthe cerebellum in high-order processing, might provide new insightsinto the mechanisms underlying neurological disorders associatedwith thyroid hormone deficiency.

Experimental methods

Animal

TRβΔ337 Tmicewere generated aspreviously described (Hashimotoet al., 2001). All mice were propagated in a mixed 129/C57/BL6background strain, and direct comparisons were made with littermatecontrols. Animals were euthanized in accordance with the guidelinesestablished by the National Institutes of Health and approved by theInternational Animal Care and Use Committee of the Federal Universityof Rio de Janeiro.

Tissue preparation, histological analysis, and immunohistochemical assays

Animals of different genotypes (wild-type: TRβwt/wt; heterozy-gote: TRβwt/Δ337T; and homozygote: TRβΔ337T/Δ337T) at postnatal days9 (P9) and 21 (P21) were transcardially perfused with 0.9% salinesolution and 4% paraformaldehyde in PBS. Cerebella were dissected,cryoprotected in 10%, 20%, and 30% sucrose at 4 °C, and embedded inO.C.T. (Tissue-Tek) for frozen sectioning in a cryostat. Cerebellarparasagittal sections were cut at a thickness of 12 μm through thevermis and placed on slides for staining with cresyl-violet forhistological analysis, or permeabilized with 0.2% Triton X-100 in PBSfor immunohistochemical staining. Tissueswerewashedwith PBS andincubated with 5% normal goat serum (NGS, Invitrogen, Carlsbad,California) in 0.05% Triton X-100 in PBS (blocking solution) for 1 h,and subsequently immunoreacted with the following antibodies:anti-Calbindin-D28K (Chemicon, California) and anti-glial fibrillaryacidic protein (GFAP; DAKO Cytomation, Glostrup).

For BrdU (5-bromo-2′-deoxyuridine) immunodetection and de-termination of EGL proliferation, P9 mice were intraperitoneallyinjected with BrdU (10 mg/ml, 100 mg/kg body weight; SigmaChemical Co., St. Louis, MO) 1 h prior to being euthanized. Thecerebellar sections were washed in distilled water and incubated in2 N HCl at 50 °C for 10 min, followed by washing with 0.1 M boratebuffer and incubationwith rat anti-BrdU antibody (Accurate Chemical& Scientific Corporation, NY).

76 A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

After the primary antibody reactions, sections were incubated withfluorescence-conjugated secondary antibodies (Alexa Fluor goat anti-rabbit, anti-mouse, and anti-rat immunoglobulin G; Molecular Probes,Oregon). The sections were then labeled with DAPI (4′,6-diamidino-2-phenyindole, dilactate; Sigma Chemical Co., St. Louis, MO) to reveal thenuclei, and were mounted on glass coverslips (Knittel Gläser, Bielefeld)in N-propyl-gallate (VETEC, Rio de Janeiro). Tissues were analyzedunder a ZEISS AxioPlan microscope equipped with a ZEISS AxioCamdigital color camera or a ZEISS confocal microscope.

Morphometric analysis

All measurements were made using cresyl-violet-stained midlinesagittal sections, by using a digital camera (ZEISS AxioCam) attachedto a ZEISS AxioPlan microscope. The areas of the cerebellum andlobules IX, EGL, ML, IGL, GL, and the thickness of the EGL in wild-typeand mutant mice were measured through the vermis. The measure-ments were made using Axiovision AC software. Quantitative datawere obtained from 4 sections per mouse, and at least 3 mice pergenotype and age group were evaluated.

Cerebellar cell cultures

BG-enriched cultures were obtained from postnatal-day 7 (P7)mice. Briefly, cerebella from P7 mice were stripped off the meninges,cut into small pieces, and then exposed to buffer A (0.003 g/mL BSA,0.014 M glucose, 0.075 MMgSO4, and 0.25 mg/mL trypsin) for 15 minat 37°C. Next, trypsin activity was inhibited with FCS and the tissueswere mechanically dissociated into single cells with a Pasteur pipettein buffer B (0.003 g/mL BSA, 0.014 M glucose, and 0.081 M MgSO4).The cell suspension was then centrifuged for 10 min at 1000 rpmthrough buffer C (containing 0.043 g/mL BSA, 0.014 M glucose, and0.075 M MgSO4), and the pellet was resuspended in Neurobasalmedium supplemented with 2% B27 (Invitrogen, Carlsbad, California).An amount of 25,000 cells was plated onto 96-well culture flasks,previously coated with polyornithine (1.5 μg/ml), and maintained inthe same medium described above, for 24 h for GCPs, or 72 h for BGanalysis. This protocol resulted in a mixed cell culture consistingmostly of GFAP- and β-tubulin III-positive cells in a 3:7 ratio. Nearly allGFAP-positive cells showed the typical radial glia cell morphology,and immunoreacted with an antibody against the astrocyte-specificglutamate transporter (GLAST), as expected for BG cells (Yamadaet al., 2000).

Immunocytochemistry assays

Cultures were fixed with 4% paraformaldehyde for 20 min andfurther permeabilized with 0.2% Triton X-100 for 5 min at roomtemperature. Then, cells were washed with PBS and incubated with10% bovine serum albumin (BSA, Sigma Chemical Co., St. Louis, MO) inPBS (blocking solution) for 1 h, and subsequently immunoreactedwith anti-GFAP (DAKO Cytomation, Glostrup) antibody.

BrdU-immunodetection and cell-proliferation assays were per-formed as previously described (Gomes et al., 1999). Briefly, granularneuron cultures were incubated for 2 h in the presence of 1 mg/ml ofBrdU (Sigma Chemical Co., St. Louis, MO) prior to fixation with 4%paraformaldehyde. Cultures were washed with distilled water andthen incubated in 2 N HCl at 50 °C for 10 min, followed by washingwith 0.1 M borate buffer for 10 min at room temperature.

After the primary antibody reactions, cells were incubated withfluorescence-conjugated secondary antibodies (Alexa Fluor; MolecularProbes, Oregon). Nuclei were counterstained with DAPI (SigmaChemical Co., St. Louis, MO), and the cells were mounted on glasscoverslips (Knittel Gläser, Bielefeld) in N-propyl-gallate (VETEC, Rio deJaneiro), and then analyzed under a ZEISS AxioPlan microscope.

Statistical analysis

Data were analyzed using Student's t test (preceded by the F test forvariances), or one-wayANOVA followedby Tukey's test for comparisonsbetween groups. A confidence interval of 95% and a p-value of less than0.05 were considered statistically significant. Data depicted in thegraphs represent the mean±SEM of results.

Acknowledgments

We thank Ismael Gomes and Marcelo Meloni for technicalassistance and the Program for Technological Development in Toolsfor Health-PDTIS-FIOCRUZ for the use of its facilities. This study wassupported by grants from the Fundação Carlos Chagas Filho deAmparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ), ConselhoNacional de Desenvolvimento Científico e Tecnológico (CNPq),Coordenação de Aperfeiçoamento de Pessoal de Nível Superior(CAPES), and Sociedade Brasileira de Endocrinologia e Metabologia(SBEM).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.mcn.2010.02.004.

References

Altman, J., Bayer, S.A., 1997. The Development of the Cerebellar System: In Relation toits Evolution, Structure, and Functions. CRC Press.

Alvarez-Dolado, M., González-Sancho, J.M., Bernal, J., Muñoz, A., 1998. Developmentalexpression of Tenascin-C is altered by hypothyroidism in the rat brain.Neuroscience 84, 309–322.

Alvarez-Dolado, M., Ruiz, M., Del Rio, J.A., Alcántara, S., Burgaya, F., Sheldon, M.,Nakajima, K., Bernal, J., Howell, B.W., Curran, T., Soriano, E., Muñoz, A., 1999.Thyroid hormone regulates reelin and dab1 expression during brain development.J. Neurosci. 19, 6979–6993.

Baillieux, H., de Smet, H.J., Paquier, P.F., de Deyn, P.P., Mariën, P., 2008. Cerebellarneurocognition: insights into the bottom of the brain. Clin. Neurol. Neurosurg. 110,763–773.

Bernal, J., Nunes, J., 1995. Thyroid hormones and brain development. Eur. J. Endocrinol.133 (4), 390–398.

Blaess, S., Graus-Porta, D., Belvindrah, R., Radakovits, R., Pons, S., Littlewood-Evans, A.,Senften,M.,Guo,H., Li, Y.,Miner, J.H.,Reichardt, L.F.,Müller,U., 2004.Beta1-integrins arecritical for cerebellar granule cell precursor proliferation. J. Neurosci. 24, 3402–3412.

Bradley, D.J., Towle, H.C., Young, W.S., 1992. Spatial and temporal expression of alpha-and beta-thyroid hormone receptor mRNAs, including the beta-2 subtype in themammalian nervous system. J. Neurosci. 12, 2288–2302.

Callu,D., Puget, S., Faure,A., Guegan,M., ElMassioui,N., 2007.Habit learningdissociation inrats with lesions to the vermis and the interpositus of the cerebellum. Neurobiol. Dis.27, 228–237.

Chan, J.A., Balasubramanian, S., Witt, R.M., Nazemi, K.J., Choi, Y., Pazyra-Murphy, M.F.,Walsh, C.O., Thompson, M., Segal, R.A., 2009. Proteoglycan interactions with SonicHedgehog specify mitogenic responses. Nat. Neurosci. 12, 409–417.

Clos, J., Legrand, C., Legrand, J., 1980. Effects of thyroid state on the formation and earlymorphological development of Bergmann glia in the developing rat cerebellum.Dev. Neurosci. 3, 199–208.

Clos, J., Legrand, J., 1973. Effects of thyroid deficiency on the different cell populations ofthe cerebellum in the young rat. Brain Res. 63, 450–455.

Corrales, J.D., Blaess, S., Mahoney, E.M., Joyner, A.L., 2006. The level of sonic hedgehogsignaling regulates the complexity of cerebellar foliation.Development 133, 1811–1821.

Corrales, J.D., Rocco, G.L., Blaess, S., Guo, Q., Joyner, A.L., 2004. Spatial pattern of sonichedgehog signaling throughGli genes during cerebellumdevelopment. Development131 (22), 5581–5590.

Dahmane, N., Ruiz i Altaba, A., 1999. Sonic hedgehog regulates the growth andpatterning of the cerebellum. Development 126, 3089–3100.

De Escobar, G.M., Obregón, M.J., del Rey, F.E., 2004. Maternal thyroid hormones early inpregnancy and fetal brain development. Best. Pract. Res. Clin. Endocrinol. Metab. 18(2), 225–248.

Doughty, M.L., Delhaye-Bouchaud, N., Mariani, J., 1998. Quantitative analysis ofcerebellar lobulation in normal and agranular rats. J. Comp. Neurol. 399, 306–320.

Farwell, A.P., Dubord-Tomasetti, S.A., 1999. Thyroid hormone regulates the expressionof laminin in the developing rat cerebellum. Endocrinology 140, 4221–4227.

Forrest, D., Hanebuth, E., Smeyne, R.J., Everds, N., Stewart, C.L., Wehner, J.M., Curran, T.,1996. Recessive resistance to thyroid hormone in mice lacking thyroid hormonereceptor beta: evidence for tissue-specific modulation of receptor function. EMBO J.15, 3006–3015.

Gomes, F.C.A., Lima, F.R.S., Trentin, A.G., Moura Neto, V., 2001. Thyroid hormone role innervous system morphogenesis. Prog. Brain. Res. 132, 41–50.

77A.C. Portella et al. / Molecular and Cellular Neuroscience 44 (2010) 68–77

Gomes, F.C.A., Maia, C.G., Menezes, J.R.L., Moura Neto, V., 1999. Cerebellar astrocytestreated by thyroid hormone modulate neuronal proliferation. Glia 25, 247–255.

Graus-Porta, D., Blaess, S., Senften, M., Littlewood-Evans, A., Damsky, C., Huang, Z., Orban,P., Klein, R., Schittny, J.C.,Müller, U., 2001.β1-class integrins regulate thedevelopmentof laminae and folia in the cerebral and cerebellar cortex. Neuron 31, 367–379.

Hashimoto, K., Curty, F.H., Borges, P.P., Lee, C.E., Abel, E.D., Elmquist, J.K., Cohen, R.N.,Wondisford, F.E., 2001. An unliganded thyroid hormone receptor causes severeneurological dysfunction. Proc. Natl. Acad. Sci. U. S. A. 98, 3998–4003.

Heikkinen, S., Auwerx, J., Argmann, C.A., 2007. PPARgamma in human and mousephysiology. Biochim. Biophys. Acta. 1771 (8), 999–1013.

Herrup, K., 1983. Role of staggerer gene in determining cell number in cerebellar cortex.I. Granule cell death is an indirect consequence of staggerer gene action. Brain Res.313 (2), 267–274.

Heuer, H., Mason, C.A., 2003. Thyroid hormone induces cerebellar Purkinje celldendritic development via the thyroid hormone receptor α1. J. Neurosci. 23,10604–10612.

Koibuchi, N., Fukuda, H., Chin, W.W., 1999. Promoter-specific regulation of the brain-derived neurotropic factor gene by thyroid hormone in the developing ratcerebellum. Endocrinology 140, 3955–3961.

Lauder, J.M., 1977. The effects of early hypo- and hyperthyroidism on the developmentof rat cerebellar cortex. III. Kinetics of cell proliferation in the external granularlayer. Brain Res. 126, 31–51.

Lauder, J.M., 1978. Effects of early hypo- and hyperthyroidism on development of ratcerebellar cortex. IV. The parallel fibers. Brain Res. 142, 25–39.

Lauder, J.M., Altman, J., Krebs, H., 1974. Somemechanisms of cerebellar foliation: effectsof early hypo- and hyperthyroidism. Brain Res. 76, 33–40.

Lazar,M.A., 2003. Thyroidhormone action: a binding contract. J. Clin. Invest. 112, 497–499.Manzano, J., Morte, B., Scanlan, T.S., Bernal, J., 2003. Differential effects of triiodothy-

ronine and the thyroid hormone receptor beta-specific agonist GC-1 on thyroidhormone target genes in the brain. Endocrinology 144, 5480–5487.

Martinez, R., Gomes, F.C.A., 2002. Neuritogenesis induced by thyroid hormone-treatedastrocytes is mediated by epidermal growth factor/mitogen-activated proteinkinase-phosphatidylinositol 3-kinase pathways and involves modulation ofextracellular matrix proteins. J. Biol. Chem. 277, 49311–49318.

Martinez, R., Gomes, F.C.A., 2005. Proliferation of cerebellar neurons induced byastrocytes treated with thyroid hormone is mediated by a cooperation between cellcontact and soluble factors and involves the epidermal growth factor-proteinkinase a pathway. J. Neurosci. Res. 80, 341–349.

Messer, A., Snodgrass, G.L., Maskin, P., 1985. Timecourse of effects of triiodothyronineonmouse cerebellar cells cultured by two different methods. Int. J. Dev. Neurosci. 3,291–299.

Miale, I., Sidman, R.L., 1961. An autoradiographic analysis of histogenesis in the mousecerebellum. Exp. Neurol. 4, 277–296.

Morte, B., Manzano, J., Scanlan, T., Vennström, B., Bernal, J., 2002. Deletion of the thyroidhormone receptorα1 prevents the structural alterations of the cerebellum inducedby hypothyroidism. Proc. Natl. Acad. Sci. U. S. A. 99, 3985–3989.

Morte, B., Manzano, J., Scanlan, T.S., Vennström, B., Bernal, J., 2004. Aberrantmaturation ofastrocytes in thyroid hormone receptor alpha 1 knockout mice reveals an interplaybetween thyroid hormone receptor isoforms. Endocrinology 145, 1386–1391.

Neveu, I., Arenas, E., 1996. Neurotrophins promote the survival and development ofneurons in the cerebellum of hypothyroid rats in vivo. J. Cell Biol. 133, 631–646.

Nicholson, J.L., Altman, J., 1972. The effects of early hypo- and hyperthyroidism on thedevelopment of rat cerebellar cortex. I. Cell proliferation and differentiation. BrainRes. 44, 13–23.

Ortiga-Carvalho, T.M., Hashimoto, K., Pazos-Moura, C.C., Geenen, D., Cohen, R., Lang, R.M., Wondisford, F.E., 2004. Thyroid hormone resistance in the heart: role of thethyroid hormone receptor beta isoform. Endocrinology 145, 1625–1633.

Ortiga-Carvalho, T.M., Shibusawa, N., Nikrodhanond, A., Oliveira, K.J., Machado, D.S.,Liao, X.H., Cohen, R.N., Refetoff, S., Wondisford, F.E., 2005. Negative regulation bythyroid hormone receptor requires an intact coactivator-binding surface. J. Clin.Invest. 115, 2517–2523.

Pazos-Moura, C., Abel, E.D., Boers, M.E., Moura, E., Hampton, T.G., Wang, J., Morgan, J.P.,Wondisford, F.E., 2000. Cardiac dysfunction caused by myocardium-specificexpression of a mutant thyroid hormone receptor. Circ. Res. 86, 700–706.

Pessôa, C.N., Santiago, L.A., Santiago, D.A., Machado, D.S., Rocha, F.A., Ventura, D.F.,Hokoç, J.N., Pazos-Moura, C.C., Wondisford, F.E., Gardino, P.F., Ortiga-Carvalho, T.M.,2008. Thyroid hormone action is required for normal cone opsin expression duringmouse retinal development. IOVS 49, 2039–2045.

Poguet, A.L., Legrand, C., Feng, X., Yen, P.M., Meltzer, P., Samarut, J., Flamant, F., 2003.Microarray analysis of knockout mice identifies cyclin D2 as a possible mediator forthe action of thyroid hormone during the postnatal development of the cerebellum.Dev Biol. 254, 188–199.

Porterfield, S.P., Hendrich, C.E., 1993. The role of thyroid hormones in prenatal andneonatal neurological development — current perspectives. Endocr. Rev. 14,94–106.

Refetoff, S., Dewind, L.T., Degroot, L.J., 1967. Familial syndrome combining deaf-mutism,stippled epiphyses, goiter and abnormally high PBI: possible target organrefractoriness to thyroid hormone. J. Clin. Endocrinol. Metab. 27, 279–294.

Sillitoe, R.V., Joyner, A.I., 2007. Morphology, molecular codes, and circuitry produce thethree-dimensional complexity of the cerebellum. Annu. Rev. Cell. Dev. Biol. 23,549–577.

Smeyne, R.J., Chu, T., Lewin, A., Bian, F., Sanlioglu, S., Kunsch, C., Lira, S.A., Oberdick, J.,1995. Local control of granule cell generation by cerebellar Purkinje cells. Mol. Cell.Neurosci. 6 (3), 230–251.

Sotelo, C., 2004. Cellular and genetic regulation of the development of the cerebellarsystem. Prog. Neurobiol. 72, 295–339.

Strait, K.A., Schwartz, H.L., Seybold, V.S., Ling, N.C., Oppenheimer, J.H., 1991.Immunofluorescence localization of thyroid hormone receptor protein beta 1 andvariant alpha 2 in selected tissues: cerebellar Purkinje cells as a model for beta 1receptor-mediated developmental effects of thyroid hormone in brain. Proc. Natl.Acad. Sci. U. S. A. 88, 3887–3891.

Tagami, T., Gu, W.X., Peairs, P.T., West, B.L., Jameson, J.L., 1998. A novel naturalmutation in the thyroid hormone receptor defines a dual function domain thatexchanges nuclear receptor corepressors and coactivators. Mol. Endocrinol. 12,1888–1902.

Takahashi, M., Negishi, T., Tashiro, T., 2008. Identification of genes mediating thyroidhormone action in the developing mouse cerebellum. J. Neurochem. 104, 640–652.

Tavano, A., Grasso, R., Gagliardi, C., Triulzi, F., Bresolin, N., Fabbro, F., Borgatti, R., 2007.Disorders of cognitive and affective development in cerebellar malformations.Brain 130, 2646–2660.

Wechsler-Reya, R.J., Scott, M.P., 1999. Control of neuronal precursor proliferation in thecerebellum by Sonic Hedgehog. Neuron 22, 103–114.

Wetts, R., Herrup, K., 1982. Interaction of granule. Purkinje and inferior olivary neuronsin lurcher chimeric mice. II. Granule cell death. Brain Res. 250 (2), 358–362.

Yamada, K., Fukaya, M., Shibata, T., Kurihara, H., Tanaka, K., Inoue, Y., Watanabe, M., 2000.Dynamic transformation of Bergmann glial fibers proceeds in correlation withdendritic outgrowth and synapse formation of cerebellar Purkinje cells. J. Comp.Neurol. 418, 106–120.