Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Genome-Wide Search Reveals the Existence of a Limited Number of Thyroid Hormone Receptor Alpha Target Genes in Cerebellar Neurons

  • Fabrice Chatonnet,

    Affiliation Institut de Génomique Fonctionnelle de Lyon, Université de Lyon, Université Lyon 1, CNRS, INRA, École Normale Supérieure de Lyon, Lyon, France

  • Romain Guyot,

    Affiliation Institut de Génomique Fonctionnelle de Lyon, Université de Lyon, Université Lyon 1, CNRS, INRA, École Normale Supérieure de Lyon, Lyon, France

  • Frédéric Picou,

    Affiliation Institut de Génomique Fonctionnelle de Lyon, Université de Lyon, Université Lyon 1, CNRS, INRA, École Normale Supérieure de Lyon, Lyon, France

  • Maria Bondesson,

    Current address: Department of Biology and Biochemistry, Center for Nuclear Receptors and Cell Signaling, University of Houston, Houston, Texas, United States of America

    Affiliation Karolinska Institute Novum, Stockholm, Sweden

  • Frederic Flamant

    Frederic.Flamant@ens-lyon.fr

    Affiliation Institut de Génomique Fonctionnelle de Lyon, Université de Lyon, Université Lyon 1, CNRS, INRA, École Normale Supérieure de Lyon, Lyon, France

Abstract

Thyroid hormone (T3) has a major influence on cerebellum post-natal development. The major phenotypic landmark of exposure to low levels of T3 during development (hypothyroidism) in the cerebellum is the retarded inward migration of the most numerous cell type, granular neurons. In order to identify the direct genetic regulation exerted by T3 on cerebellar neurons and their precursors, we used microarray RNA hybridization to perform a time course analysis of T3 induced gene expression in primary cultures of cerebellar neuronal cell. These experiments suggest that we identified a small set of genes which are directly regulated, both in vivo and in vitro, during cerebellum post-natal development. These modest changes suggest that T3 does not acts directly on granular neurons and mainly indirectly influences the cellular interactions taking place during development.

Introduction

3,5,3′ tri-iodothyronine (T3), the active form of thyroid hormone, acts by binding to its cognate receptors, thyroid hormone receptors (TRs), which are transcription factors from the nuclear receptor family [1]. TRs exist as three isoforms, TRα1, TRβ1 and TRβ2, encoded by the Thra and Thrb genes respectively. They can bind DNA both in the presence and absence of T3, either as homodimers or as heterodimers with RXR (Retinoid X Receptor). DNA binding occurs on so-called T3 response elements (TREs) which usually associate two consensus half-sites (5′AGGTCA3′) often organized either as direct repeats separated by four nucleotides (DR4). T3-liganded TRs have the ability to induce either transcriptional activation or repression depending on the locus and the local environment. For positive regulation, unliganded TRs change conformation upon T3 binding, releasing transcription corepressors to recruit coactivators and leaving epigenetic marks on neighboring histone tails [2]. Negative regulation mechanisms are poorly understood, and still a matter of controversy.

The cerebellum is a very suitable structure to study T3 action during neurodevelopment because it is particularly sensitive to hypothyroidism [3]. Moreover, the cerebellum has a relatively simple architecture compared to other brain regions, with only a limited number of cell types. The cerebellar cortex is organized in layers lying over the white matter and the deep cerebellar nuclei. The cerebellar cortical layers comprise of the molecular layer containing GABAergic interneurons, the Purkinje cell layer that also contains the cell bodies of Bergmann glia cells and the internal granular cell layer (IGL), containing mature granular cells, astrocytes and Golgi GABAergic interneurons. During development, granular cell precursors form a germinative layer over the surface of the cerebellar primordium called the external granular layer (EGL), where precursors divide actively and start their differentiation after exiting the cell-cycle (for review see [4]). Post-mitotic granular neurons extend their axon in the molecular layer and migrate their cell body through the Purkinje cell layer to populate the IGL. At post-developmental stages, granular neurons represent more than ninety percent of cerebellar cells.

T3 deprivation during cerebellar post-natal development affects all these cell types (reviewed in [5]). Granular cell migration is blocked and granular cell precursors are trapped in the EGL after the third post-natal week, when migration towards the IGL is normally finished [6]. Purkinje cell layering is defective and the size of their dendritic tree is reduced [7]. Bergmann glia fibers morphology is abnormal, and astrocytes [8], [9] show an over-proliferation and delayed differentiation. Oligodendrocytes [10], [11] and GABAergic interneurons [12] differentiation is delayed. Mouse genetics provided clear evidences that most if not all of these defects are due to the presence of unliganded TRα1 in all cerebellar cell types. First, Thra knock-out reverses the deleterious effect of hypothyroidism [13]. Second, dominant negative mutations of TRα1, which affect either its ability to bind T3 or to recruit coactivators, lead to a cerebellar phenotype resembling hypothyroidism [14], [15], [16]. Unlike Thra, Thrb is expressed in only a few cell types and at low level in the developing cerebellum. However, Thrb point mutation has also been found to induce important cerebellar defects [17], raising the possibility that T3 can initially act on few cell types, and indirectly exert a global and indirect influence on a network of cellular interactions.

Several attempts have been performed to elucidate the molecular events underlying the observed cellular alterations in hypothyroid cerebellum and identify the TRs target genes. Various animal models have been used to identify genes which expression levels are sensitive to T3 status [18][22] or which promoters are directly bound by TR [23]. Although a number of variations in mRNA levels were observed, most of these experiments failed to establish a direct link between cells behavior and TR regulated transcription, as changes in gene expression might be explained by variations in neurotrophic factors levels, which are secondary to T3 deficiency [21]. Cellular heterogeneity of the cerebellum, which can mask variations in gene expression restricted to one cell type, is also likely to impair the identification of direct TR target genes in these experiments. Primary neuronal cell cultures in part circumvent these problems as they allow focusing on early events triggered by T3, limiting cellular interactions and favoring neurons or glial cells by choosing adequate culture conditions and thus reducing cellular heterogeneity. Up to now, and to our knowledge, RNA analysis of cultured cells has demonstrated the direct upregulation by TRα1 (bound to identified TREs) of only four genes, which are deregulated in hypothyroid cerebellum: Hairless (Hr), synaptotagmin-related gene 1 (Srg1), [24] Krüppel-like factor 9 (Klf9) [25], [26] and A kinase anchor protein 1 (Akap1) [22]. To reach a better understanding of the genetic regulation underlying the neuronal defects observed in hypothyroid brain, we performed a genome-wide transcriptome analysis of T3 response in primary cerebellar neurons culture. This identified 15 genes, which are either up- or down-regulated by TRα1 in response to T3. We further studied whether these genes are directly regulated by TRα1 in vitro and in vivo in different cell types of the cerebellum and in a cerebellum cell line. We chose to work on the C17.2 cell line [27] since it is derived from EGL cells and thus was supposed to be closely related to our primary cultures. We used this cell line to have a more convenient access to chromatin binding since it is easier to realize chromatin precipitation in cells rather than in tissues, particularly in our case where there is no “ChIP grade” antibody for TRα1.

Results

Genome-wide Search of T3 Regulated Genes

In order to identify changes in gene expression which are mainly related to immediate and direct action of thyroid hormone on post-natal cerebellum neuronal cells, we prepared primary cultures of cerebellar cells and maintained them in conditions favoring neurons survival, at the expense of glial cell proliferation. It has been shown that in this system, granular neurons, which represent the majority of cerebellar cells, switch from proliferation to migration and finally terminal differentiation, including dendritogenesis [28], thus reproducing the normal stages of development. We verified by immunocytochemistry that in these conditions most of the cells (80%) express the Tuj1 neuronal markers, and that prolonged culture (10 days) allowed the cells to grow fine dendritic-like processes. Two days after seeding, cells were treated with T3 for various amounts of time (6, 16, 24 or 48 hours) or left untreated before RNA extraction. Total RNAs were then amplified and cDNA prepared to serve as probes for microarrays hybridization. (See dataset GSE24793 on the NCBI Gene Expression Omnibus (GEO), http://www.ncbi.nlm.nih.gov/geo/).

Only probe-sets with a sufficient level of expression (P in columns “detection” of data matrix available at the GEO depository, covering about 40% of all annotated genes) and that showed a fold change superior to 2 were considered. We detected signals for markers for each cerebellar cell-type, indicating that our culture system contained all the cell constituents of the cerebellum and we also found that its composition evolved over time. For example, at early time points we detected high levels of Pax6 and Ccnd2, two markers of granule cell precursors, which decreased over time. Conversely, we observed low levels of Gabra6, a marker of differentiated granule cells, which increased over time. We also detected Pcp2 expression, a marker of Purkinje cells, Pdgfra and Plp1, markers of oligodendrocyte precursors and of mature oligodendrocytes, respectively, and Slc1a3 (Glast) and Gfap, two markers of Bergmann glia and astrocytes.

To address the robustness of T3 regulation in primary cultures of cerebellar cells, we used Q-RT-PCR to compare the microarray results acquired with a first set of primary cultures to RNA expression of both the same and another independent primary cultures experiment. We generated primers for the most up- or down-regulated genes for each time point and for genes showing an overall constant tendency (up- or down-regulated for all time points) from the microarray results. Given the relatively low numbers of probe-sets corresponding to these criteria, we also selected genes showing a change in expression of at least 1.6 fold for one or more time points, and genes that were previously known to be regulated by T3, like Ccnd2 or Hr. We confirmed that T3 induced changes in gene expression for a number of genes and the changes in expression pattern increased over time. 6 genes were up-regulated and 9 were down-regulated in a significant and reproducible manner (Table 1). We focused later investigations on this group of 15 putative TR target genes The group of up-regulated genes included the previously reported T3 target gene Klf9. Hr expression was too low to be detected by microarray hybridization, but its expected induction by T3 was confirmed by Q-RT-PCR. This result is consistent with Q-RT-PCR being a more sensitive detection method of RNA than microarray hybridization, and indicates that our survey, although revealing expression of about 40% of the annotated genes in the cultured cells, might have missed some changes in gene expression, due to limited sensitivity. However, detection of mRNA expressed by very limited cell populations (see above, ie, Pcp2 in Purkinje cells) suggests that it was not the case.

thumbnail
Table 1. Q-RT-PCR confirmation of T3 mediated regulation in primary cultures of cerebellum neuronal cells.

https://doi.org/10.1371/journal.pone.0030703.t001

In vitro Study of Target Genes Regulation of Expression

The fact that most of the genes that we identified were distinct from those that have been identified in previous studies prompted us to address the possibility that some genes might be regulated in specific cell types, or at specific differentiation steps, whereas others would be regulated more broadly. To study the expression of the T3 regulated genes on another cellular background we used C17.2 cerebellum immortalized pluripotent cell line, which originates from EGL [27]. Although the exact nature of these cells is unclear [29], they are able to differentiate into granular-like neurons when grafted in brain [30]. We first restored C17.2 cells response to T3 by expressing a tagged version of TRα1 in a stable manner and measured the expression of Hr in response to TR and T3 (Fig. 1A). We found that both C17.2 and C17.2/TRα1 cells can differentiate within 5 days to Tuj1+ neuronal-like cells in serum-free Neurobasal medium. Interestingly, in this differentiation assay, the presence of T3 completely prevented the appearance of Tuj1+ cells in C17.2/TRα1 but not in C17.2 (data not shown). We then addressed the expression of the 15 novel TR target genes, identified in primary neuronal cultures, in C17.2/TRα1 cells after 48 hours of T3 stimulation in the presence or absence of serum. Figure 1b indicates that, soon after serum deprivation, T3 regulation was visible in these cells, with 3 exceptions (Plxna2, Spata13 and Vnn1). Interestingly, the T3-induced changes in gene expression in presence of serum were mainly limited to Hr induction. Although thought to be expressed only in the oligodendrocytes lineage, Plp1 was activated in serum starved C17.2/TRα1 cells, which did not express other typical oligodendrocyte markers, like Mbp gene, encoding the Myelin Basic Protein (data not shown).

thumbnail
Figure 1. Transfection of TRα1 into C17.2 cells restores their response to T3 treatment.

A. T3-induced Hr expression is detected earlier and stronger in transfected C17.2/TRα1 cells (black bars) than in non-transfected cells (white bars), *p<0.05, Student’s t-test difference between non-transfected cells and C17.2/TRα1 cells. B. The level of change in expression induced by T3 and measured by Q-RT-PCR in C17.2/TRα1 cells is indicated for each target gene, using non-treated cultures as reference (represented as log2 of the fold change). White bars indicate T3 treatment in proliferative medium (containing serum), black bars indicated T3 treatment in serum-deprived medium allowing for differentiation. Most genes show a response only in serum-deprived medium, *p<0.05, Student’s t-test, difference between serum containing and serum deprived cultures.

https://doi.org/10.1371/journal.pone.0030703.g001

The C17.2/TRα1 cell line appears therefore as a convenient in vitro model to study the T3/TRα1 signaling pathway in a neural context. It offers the possibility to assess if TRα1 is bound to consensus binding sequences (TR response element, TRE) at proximity of the target genes transcription start site (TSS). Using previous bioinformatics data [22], putative TREs conserved between mouse and human were first identified within 25 kb of transcription start sites (TSS). When no conserved sequence was found, Nubiscan online tool (http://www.nubiscan.unibas.ch/) was also used to extend the search to near-consensus sequences in this region. This gave us a list of putative TREs for each locus (Table 2). Taking advantage of the proteinG tag present on TRα1, we performed Chromatin Affinity Purification (ChAP) to enrich fractionated chromatin into TRα1 bound complexes. This confirmed the presence of TRα1 in areas covering some of previously reported TREs, upstream to Klf9 [26], Hr [22] and Dbp [23], and one downstream to the transcription start site of Gbp3 (Table 2). The transcriptional regulation for these genes is thus likely to result directly from the activation of TRα1 bound to neighboring regulatory sequences. We also used a recombination mediated cassette exchange assay [31] to address the ability of upstream promoter sequences of these genes to drive the stable expression of renilla luciferase in C17.2/TRα1 cells and mediate T3 regulation in a physiological chromatin context. This confirmed the presence of T3 regulatory elements within 6 kb of upstream sequences for Hr but not for Gbp3 and Anxa8 (data not shown).

thumbnail
Table 2. Chromatin occupancy by TRα1 in C17.2/TRα1 cells.

https://doi.org/10.1371/journal.pone.0030703.t002

In vivo T3 Target Genes Expression and Regulation by TRα1

As Thra is expressed in all cerebellum cell types, we used transgenic mice which express a dominant negative mutation of the TRα1 receptor (TRα1L400R) to address the in vivo relevance of the gene regulation observed in cultured cells. TRα1L400R can be expressed from the so called TRαAMI allele of Thra only when a transcription stop cassette is deleted by CRE/loxP recombination. In TRαAMI/S mice, excision takes place soon after fertilization, and TRα1L400R expression is nearly ubiquitous. TRαAMI/S mice display the defects in EGL migration and Purkinje cells arborization, which define congenital hypothyroidism in post-natal cerebellum [14], [16]. Q-RT-PCR of cerebellum RNA showed that significant alterations of gene expression between wild-type and TRαAMI/S mice at various developmental stages for the identified genes, indicating that these genes can be regulated by liganded TRα1 during cerebellum development, at least at one developmental stage (Figure 2). These results confirmed the predominant in vivo function of TRα1 and the relevance of the cell culture systems that we used. Expression kinetics revealed a variety of expression patterns, corresponding to the dynamic changes that occur during cerebellar development, and suggested that, for each gene, T3 sensitivity is restricted to a temporal window, sometimes limited to one time point. For some genes, opposite effects on expression of a target gene in wild-type and TRαAMI/S were observed at different time points, suggesting that a combination of different T3-mediated regulation mechanisms takes place during development. These variations in mRNA levels can however be influenced by modifications in cellular composition of the cerebellum due to the adverse developmental effects of TRα1L400R and subsequent bias in whole cerebellum mRNA content.

thumbnail
Figure 2. Kinetics of T3 target genes expression in wild-type and TRαAMI/S mice in the cerebellum as measured by Q-RT-PCR.

Expression levels were calculated for each target gene by Q-RT-PCR in wild-type and TRαAMI/S littermates at P4, P8, P15 and P21 (minimum 3 animals of each genotype for each time point). Data are expressed as mean ± SD using wild-type P4 values (A), for genes with decreasing or stable expression levels over time or P21 (B), for genes with increasing expression levels, as a reference for each genotype. *p<0.05; **p<0.01 for comparisons between wild-type and TRαAMI/S mice for each time point (Student’s t-test).

https://doi.org/10.1371/journal.pone.0030703.g002

To address this last possibility, we examined expression patterns for the identified target genes, using adult and fetal gene expression databases (Allen Mouse Brain Atlas, http://mouse.brain-map.org/welcome.do and Gensat, http://www.ncbi.nlm.nih.gov/projects/gensat/) and cell-type specific transcriptome analysis based on tagged ribosomes immunoprecipitation [32] (Table 3). Preferential Anxa8, Plxna2 and Spata13 expression seems to occur in granular cells and in their EGL precursors. Therefore, the observed increase in Anxa8, Plxna2 and Spata13 mRNA at P21 might at least in part reflect the persistence of EGL in TRαAMI/S cerebellum at late stage, rather than a transcriptional regulation. Although granular neurons and their precursors represent the majority of cells present in cerebellum and in the primary cell cultures that we initially used, it is striking that the number of genes found to be expressed preferentially in this lineage is very limited (Anxa8, Hr, Plxna2). By contrast, we picked up specific expression of glial cells (Plp1) and GABAergic neurons (Dbp, Pfkfb3, Tgm2), although these cell types were clearly under-represented in the primary cells culture system that we initially used.

thumbnail
Table 3. Database analysis of expression patterns and gene functions.

https://doi.org/10.1371/journal.pone.0030703.t003

Cell Autonomous and Indirect Regulations

As TRα1L400R expression in transgenic mice depends on the Cre/loxP recombination system, we can address whether the in vivo regulation by TRα1 on T3 sensitive gene is a cell autonomous process, a feature that is an important indicator of a direct transcriptional control, by expressing TRα1L400R in defined cell types. We crossed TRαAMI mice with transgenic mice expressing the CRE recombinase in restricted cell populations. The TRαAMI/O mice expresses a tamoxifen inducible version of the recombinase in the granular neurons of the posterior lobules of the post-natal cerebellum (Otx2CreERT2, [44]) and we therefore expected an inhibition of TRα1 function in the granular cell lineage in this model. TRαAMI/P mice express TRα1L400R during fetal life in Purkinje cells and GABAergic interneurons progenitors (Ptf1a-Cre, [45]). TRαAMI/C mice express the mutation in the oligodendrocytes and their committed precursors (Cnp-Cre [46]). Tamoxifen treatment of TRαAMI/O at P1 followed by Q-RT-PCR analysis at P8 revealed a change in gene expression only for Cdh1 and Klf9 (Table 4), in keeping with the putative expression patterns established previously. TRαAMI/C mice displayed deregulation of Igsf3, a possible indication of regulation in the oligodendrocytes lineage. However, most changes in gene expression were observed in TRαAMI/P mice, where TRα1L400R expression was restricted to GABAergic neurons. Cdh1, Dbp, Pfkfb3 and Tgm2 are thus possibly regulated by TRα1 in neuronal cell types that were under-represented in the initial culture conditions. Interestingly, some of the changes in gene expression found in TRαAMI/P mice are clearly not cell autonomous consequences of TR1L400R expression in GABAergic neurons, but affect expression either in oligodendrocytes (Plp1) or granular neurons (Plxna2). This might reflect an indirect response of this last cell type to diffusible factors secreted by GABAergic neurons under T3 stimulation.

thumbnail
Table 4. Cell autonomous effect of in vivo expression of a dominant negative TRα1 mutation.

https://doi.org/10.1371/journal.pone.0030703.t004

Discussion

By using a cell culture system aimed at maximizing T3 response of cerebellum cells, we were able to identify new T3 target genes, which are either up- or down-regulated during normal cerebellum development. Among the genes identified by our screen, only four were already reported as being regulated by T3, sometimes in other brain areas and at different stages: Hr [47], Snap25 [48], Dbp [49] and Klf9 [50]. Among the identified genes, some are regulated during early development, possess TREs in their regulatory sequences, which can be actually occupied by TRs, and are sensitive to the expression of a dominant negative TRα1 mutation in a cell-autonomous manner. These genes (Hr, Dbp, Klf9, Gbp3) are, beyond any reasonable doubt, direct TR targets. For the others the demonstration is only partial and alternative explanations for changes in genes expression are possible. In some cases (Plp1, Plxna2, Tgm2), slow kinetics could be an indication that they are secondary targets, activated by transcription factors which expression is itself regulated by T3, or by some of the neurotrophic signaling pathways that are activated by T3. Nevertheless, this study provides with a list of genes that may be important for cerebellar development and which role in this process would be worth investigating.

This study is at least the fourth attempt to identify T3 target genes in the cerebellum by broad transcriptional analysis, and there is little overlap between datasets. Although the discrepancies between the present and previous studies might in part be due to technical reasons, we believe that most differences can be explained by the fact that the repertoire of TRs target genes is highly dependent on cell types and, as indicated by our in vivo data, developmental stage. Here we favored the discovery of genes regulated by T3 in neurons, especially in the most abundant granular neurons, whereas previous attempts focused on whole cerebellum at early [18] or late [20][22] time points. Interestingly, only two of these genes (Hr, Klf9) have been identified as T3 target genes in neurons from several brain areas and at different stages in mice and even in frogs [22], [26], [50], [51], [52] and seem to possess the singular property to be inducible by T3 in various cellular contexts.

Current knowledge provides few indications that the gene regulations that we observed might significantly impact cerebellar development. However, Klf9 knock-out mice show a deficit in the rotarod test, a sign of cerebellar disorder [36]. Its function in granular neurons remains unknown. Zbtb20 encodes a transcription factor able to specify neuronal identity [53], [54]. Plp1 is a gene encoding one of the major myelin component, proteolipid protein, expressed in oligodendrocytes. Thyroid hormone is known to be essential for oligodendrocytes differentiation [10] and thus for Plp1 expression [55]. According to our data, Plp1 is however unlikely to be a direct T3 target gene, and its in vivo down-regulation might just reflect a delay in oligodendrocytes differentiation. The fact that we detect changes in its expression in microarrays might be due to a less oligodendrocyte-restricted expression that previously known [56]. According to our observation Plxna2 gene is also sensitive to T3 signaling in vivo, but its deregulation by expressing the mutant TRα1 only in GABAergic neurons although this gene is expressed in granular cells suggest that Plxna2 is not a direct TRα1 target gene. It encodes a semaphorin receptor family member, which, by binding to its ligand, SEMA6A, promotes granule cells migration [39]. For both Plp1 and Plxna2, some unknown primary event is triggered by T3, possibly in Purkinje neurons, as indicated by the TRaAMI/CRE models. Purkinje neurons cells secrete a number of neurotrophic and growth factors that coordinate cerebellum development [57]. The majority of the other newly identified TR target genes encode proteins with known functions, relevant to synaptic transmission (SNAP25), transcription factors (Zbtb20, DBP), adhesion and extracellular matrix proteins (ANXA8, CDH1, TGM2), signal transduction (Spata13) and cell survival (Tgm2). However, mouse genetics have not yet established their involvement in cerebellum post-natal development. The metabolic function of the Pfkfb3 gene product, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3, is well known. This enzyme regulates the steady-state concentration of fructose-2,6-bisphosphate, a potent activator of a key regulatory enzyme of glycolysis, phosphofructokinase. Glycolysis is sensitive to T3 in several cell types, and a key parameter for neuronal activity. Glucose consumption was found to be reduced in the brain of mice expressing a TRα1 dominant-negative mutation [58]. Overexpression of Pfkfb3 is however detrimental to neurons [59] as it favors the accumulation of toxic free radicals, a possible consequence of T3 excess in several systems [60]. This provides a possible explanation for increased neuronal cell death in hyperthyroid post-natal cerebellum [18]. As we found little glucose consumption in our primary cultures (data not shown), pyruvate being probably used as a substitute for energetic metabolism, we cannot address the metabolic consequence of Pfkfb3 stimulation in our in vitro system.

Our initial broad survey identified a small number of genes in T3 sensitive genes in granular neurons, whereas it revealed T3 response of genes expressed in Purkinje neurons, and even in oligodendrocytes. This is a striking observation, considering that this two last cell types were a small minority of the cultured cell population and that EGL persistence at P21 and granular cells defects in differentiation and migration are the major landmarks of hypothyroidism. The fact that transcriptome analysis was sensitive enough to detect variations of rare mRNA prompts us to suggest that the total number of T3 target genes in granular neurons is very small, although it includes two of the best characterized examples: Hr and Klf9. That few genes are regulated by T3 in the granular cell precursors which populate the EGL is consistent with recent results showing that Thra expression is weak in the proliferating granular cells precursors [61], [62]. However, this explanation does not hold for post-mitotic granular neurons, but granular cells were mainly in precursor state when we did our primary cultures (P1–P4). The broad survey that we performed convinces us that most of the T3 direct response does not occur in the granular lineage that represents the vast majority of the cerebellum cell population at this stage, although thyroid hormone could have a major role on granular cell terminal differentiation at later stages. At early stages, our data suggests that T3-mediated gene regulation is mainly taking place in few cell types, including Purkinje cells, that play a pivotal role in a network of cellular interactions, mediated by direct contacts and neurotrophic factors secretion. The small number of target genes suggests that the cellular alterations observed in hypothyroid cerebellum are mainly indirect consequences of minor initial imbalance in gene expression, and that T3 controls the expression of few key developmental genes in cerebellum neurons. If our interpretation is correct, congenital hypothyroidism in cerebellum might be viewed mainly as desynchronization of the cellular interactions which govern proper neurodevelopment.

Materials and Methods

Animals

All animals used in this study were housed, raised, bred and used in accordance with European directive 86/609/EEC and in compliance with national and international rules and laws on animal welfare. Heterozygous mice expressing the TRα1L400R ubiquitously were produced by using the sycp1CRE transgene expressed in spermatogonia [14]. Cell specific recombinations were obtained by crossing TRαAMI mice with OtxCRE-ERT2 [44], Ptf1aCRE [45], CnpCRE [46]. Recombination patterns were verified by PCR [14] and by crossing with reporter transgenic ROSAYFP mice and immunocytochemistry [63].

Primary Cultures of Newborn Mice Cerebellar Neurons

Newborn mice (P1–P4) were decapitated and the cerebellum was rapidly recovered in a Petri dish containing HBSS medium without Ca2+or Mg2+supplemented with 1 mM pyruvate and 10 mM Hepes. Cerebella were pooled and then washed and dissociated in 1 mL of the same medium. 2 mL of HBSS (with Ca2+and Mg2+)/1 mM pyruvate/10 mM Hepes was added and cells were centrifuged 1 min at 200 g. Supernatant was removed, cells were resuspended in Neurobasal medium containing 2% B27 supplement, 0.5 mM L-glutamine and 0.5% penicillin-streptomycin and plated in 24-wells plates coated with poly-lysine (0.05 mg/mL). All cell-culture media were purchased form Gibco-Invitrogen.

Cultures were maintained for 48 hours after seeding to favor neuronal survival at the expanse of glial cells and then treated with 0.1 µM T3 for 6, 16, 24 and 48 hours (one culture for each time point). Control cultures were submitted to the same scheme of medium change without addition of T3. Batches of cells were allowed to grow for up to 10 days and screened for expression of neuronal or glial markers by immunohistochemistry. This set of cultures and another set of primary cultures comprising the same conditions have been used for confirmation by Q-RT-PCR.

RNA Extraction

Total RNA was extracted from cell cultures or mice cerebella with the Qiagen RNeasy kit according to manufacturers recommendations. RNA qualities and quantities were verified by capillary electrophoresis (BioAnalyser, Applied Biosystems) before reverse-transcription.

Microarrays Hybridization

1 µg of each RNA was reverse transcribed and cDNA were submitted to one round of linear amplification before hybridization to whole-genome mice Affymetrix microarrays (GeneChip Mouse Genome 430 2.0 Array). One microarray was used for each condition. Data were analyzed using the Affymetrix GCOS software, data were normalized by Affymetrix MAS5.0 software, providing an estimation of expression levels for each probe-set and a p-value of the significance of the expression change. Data were pairwise compared for each time point between a T3-treated culture and a matched untreated culture prepared simultaneously from the same cell batch. This work was done by the Bio-informatics and Expression Analysis core facility at Karolinska Institutet (Sweden).

Quantitative Reverse Transcription-PCR (Q-RT-PCR)

For Q-RT-PCR on mouse cerebella, 1 µg of each RNA sample was reverse-transcribed using MLV reverse-transcriptase (Promega). Quantitative PCR were then performed in 96 wells-plates using the hypoxanthine guanine phosphoribosyl transferase gene as a reference after carefully checking that our conditions did not significantly modify its expression level (less than one Ct of difference between compared samples, see also [22]) and the following mix: 4 µL of primers (see primers list in Table S1) mix 1.2 mM, 6 µL of SYBRGreen mix (either Qiagen Quanti-tect mix or Biorad iQ supermix) and 2 µL of cDNA diluted 1 in 20. A standard curve was made for each gene and each measure was made in triplicates. Melting curves were analyzed to ascertain homogeneity of the amplified products. Expression levels were calculated using the 2−ΔΔ(Ct)method [64]. Statistical differences have been calculated using unpaired Stutent t-test assuming different group sizes and variances.

C17.2 Culture and C17.2/TRα1 Cells

C17.2 neural cells [18] were grown in DMEM medium (4.5 g/L glucose, glutamine, pyruvate) containing 10% FBS, 5% horse serum, 1% glutamine, 1% penicillin-streptomycin. Cells proliferate in this medium and were split 1/10 each week. Sera used for cell culture were deprived of T3 either by activated charcoal or resin treatment. When cultured without serum, serum removal was done simultaneously with T3 treatment. For testing target gene expression after transfection and in medium with or without serum, three independent experiments have been done.

pCEMM-GS-TRα1 was constructed by inserting the murine TRα1 reading frame amplified by PCR into the BamHI site of pCEMM-NTAP [65] create a reading frame encoding a GS-TRα1 protein, tagged at its N-terminus by a fragment of protein G. A downstream IRES-gfp cassette located on the transcription unit ensured co-expression of a green fluorescent protein. pCEMM-GS-TRα1 was co-transfected into C17.2 cells with pPGK-Puro and puromycine resistant cells were submitted to two rounds of sorting to select cells with stable expression of both gfp and GS-TRα1. This was further confirmed by Western blotting using an anti-TR antibody (Santa Cruz).

Chromatin Affinity Purification (ChAP)

This method differs form chromatin immune-precipitation (ChIP) in that it does not involve an antibody directed against the DNA-binding protein of interest but relies on affinity binding between the tagged protein and other molecules bound to beads (in our case the protein G tag and IgG beads, see [65]). T3-treated or non-treated C17.2 cells were rinsed with PBS and cross-linked with 1% paraformaldehyde in PBS for 10 minutes at 37°C. Cells were then harvested, and stored at −80°C as pellets. Cell pellets were resuspended in 300 µL lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris pH 8) and sonicated 8 min (30s ON/30s OFF, high intensity) with a bath sonicator (Bioruptor, Diagenode). A small aliquot of cell lysate was analysed by gel electrophoresis and size of DNA fragments verified to be between 200 and 800 bp. 100 µL of the cell lysate was kept as input and the rest was diluted 1 in 10 in dilution buffer (1% Triton X-100, 1 mM EDTA, 150 mM Na Cl, 20 mM Tris pH 8) and incubated with IgG-coated magnetic beads (Invitrogen) at 4°C overnight. Negative control was obtained with non-transfected C17.2 cells. Beads were then rinsed twice in TSE1 buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris pH 8), three times in TSE2 buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 500 mM NaCl, 20 mM Tris pH8), twice in LiCl wash buffer (0.25 M LiCl, 1% NP-40, 1% Sodium deoxycholate, 1mM EDTA, 1 mM Tris pH 8), twice with TE (10 mM Tris-HCl, 1 mM EDTA pH 8) and DNA fragments were eluted in 0.1 M NaHCO3, 1% SDS. Cross-link was reversed by overnight incubation at 65°C in 200 mM NaCl. DNA was purified using the Qiagen minElute kit, and quantified by Q-PCR as described above. Calibration curves were made with dilutions of input DNA and purification rates were calculated using the 2−ΔΔ (Ct) method as input percentages. For each gene, a distal promoter DNA sequence was used as a control for non-specific background binding (see Table S2 for position of amplified sequence for both putative TREs and negative control). Enrichment was calculated as the ratio between input percentages of specifically bound quantity and non-specific binding. Preliminary experiments, using non-transfected cells, allow to set a threshold of 2-fold enrichment, as a minimum to ascertain genuine receptor binding (see Table S2 for primer sequences). Non-transfected cells gave no enrichment and no increase in purification in the precipitated fraction for any putative TRE (data not shown). Each putative has been tested on three independent ChAP experiments and results were well reproducible when enrichment was observed.

Supporting Information

Table S1.

Q-RT-PCR primers used for gene expression experiments.

https://doi.org/10.1371/journal.pone.0030703.s001

(DOC)

Table S2.

Q-RT-PCR primers used for ChAP analysis.

https://doi.org/10.1371/journal.pone.0030703.s002

(DOC)

Acknowledgments

We thank Nadine Aguilera for mouse breeding and the PBES UMS3444 facility for animal care. We also thank the personal at the Bio-informatics and Expression Analysis core facility at Karolinska Institutet for their help with microarrays hybridization and raw data analysis.

Author Contributions

Conceived and designed the experiments: FC MB FF. Performed the experiments: FC RG FP. Analyzed the data: FC MB FF. Contributed reagents/materials/analysis tools: FC FP MB FF. Wrote the paper: FC FF.

References

  1. 1. Yen PM, Ando S, Feng X, Liu Y, Maruvada P, et al. (2006) Thyroid hormone action at the cellular, genomic and target gene levels. Mol Cell Endocrinol 246: 121–127.
  2. 2. Perissi V, Rosenfeld MG (2005) Controlling nuclear receptors: the circular logic of cofactor cycles. Nat Rev Mol Cell Biol 6: 542–554.
  3. 3. Koibuchi N (2009) Animal models to study thyroid hormone action in cerebellum. Cerebellum 8: 89–97.
  4. 4. Chédotal A (2010) Should I stay or should I go? Becoming a granule cell. Trends Neurosci 33: 163–172.
  5. 5. Chatonnet F, Picou F, Fauquier T, Flamant F (2011) Thyroid hormone action in cerebellum and cerebral cortex development. J Thyroid Res. 2011: 145762.
  6. 6. Nicholson JL, Altman J (1972) The effects of early hypo- and hyperthyroidism on the development of rat cerebellar cortex. I. Cell proliferation and differentiation. Brain Res. 44: 13–23.
  7. 7. Heuer H, Mason CA (2003) Thyroid hormone induces cerebellar Purkinje cell dendritic development via the thyroid hormone receptor alpha1. J Neurosci. 23: 10604–10612.
  8. 8. Manzano J, Bernal J, Morte B (2007) Influence of thyroid hormones on maturation of rat cerebellar astrocytes. Int J Dev Neurosci. 25. pp. 171–179.
  9. 9. Trentin AG, Gomes FC, Lima FR, Neto VM (1998) Thyroid hormone acting on astrocytes in culture. In Vitro Cell Dev Biol Anim. 34: 280–282.
  10. 10. Durand B, Raff M (2000) A cell-intrinsic timer that operates during oligodendrocyte development. Bioessays. 22: 64–71.
  11. 11. Barres BA, Lazar MA, Raff MC (1994) A novel role for thyroid hormone, glucocorticoids and retinoic acid in timing oligodendrocyte development. Development. 120: 1097–1108.
  12. 12. Manzano J, Cuadrado M, Morte B, Bernal J (2007) Influence of thyroid hormone and thyroid hormone receptors in the generation of cerebellar gamma-aminobutyric acid-ergic interneurons from precursor cells. Endocrinology. 148: 5746–5751.
  13. 13. 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: 3985–3989.
  14. 14. Quignodon L, Vincent S, Winter H, Samarut J, Flamant F (2007) A point mutation in the activation function 2 domain of thyroid hormone receptor alpha1 expressed after CRE-mediated recombination partially recapitulates hypothyroidism. Mol Endocrinol 21: 2350–2360.
  15. 15. Venero C, Guadano-Ferraz A, Herrero AI, Nordstrom K, Manzano J, et al. (2005) Anxiety, memory impairment, and locomotor dysfunction caused by a mutant thyroid hormone receptor {alpha}1 can be ameliorated by T3 treatment. Genes Dev 19: 2152–2163.
  16. 16. Fauquier T, Romero E, Picou F, Chatonnet F, Nguyen XN, et al. (2011) Severe impairment of cerebellum development in mice expressing a dominant-negative mutation inactivating thyroid hormone receptor alpha1 isoform. Dev Biol. 356: 350–358.
  17. 17. Hashimoto K, Curty FH, Borges PP, Lee CE, Abel ED, et al. (2001) An unliganded thyroid hormone receptor causes severe neurological dysfunction. Proc Natl Acad Sci U S A 98: 3998–4003.
  18. 18. Takahashi M, Negishi T, Tashiro T (2008) Identification of genes mediating thyroid hormone action in the developing mouse cerebellum. J Neurochem 104: 640–652.
  19. 19. Kobayashi K, Akune H, Sumida K, Saito K, Yoshioka T, et al. (2009) Perinatal exposure to PTU decreases expression of Arc, Homer1, Egr1 and Kcna1 in the rat cerebral cortex and hippocampus. Brain Res 1264: 24–32.
  20. 20. Dong H, Wade M, Williams A, Lee A, Douglas GR, et al. (2005) Molecular insight into the effects of hypothyroidism on the developing cerebellum. Biochem Biophys Res Commun 330: 1182–1193.
  21. 21. Poguet AL, Legrand C, Feng X, Yen PM, Meltzer P, et al. (2003) Microarray analysis of knockout mice identifies cyclin D2 as a possible mediator for the action of thyroid hormone during the postnatal development of the cerebellum. Dev Biol 254: 188–199.
  22. 22. Quignodon L, Grijota-Martinez C, Compe E, Guyot R, Allioli N, et al. (2007) A combined approach identifies a limited number of new thyroid hormone target genes in post-natal mouse cerebellum. J Mol Endocrinol 39: 17–28.
  23. 23. Dong H, Yauk CL, Rowan-Carroll A, You SH, Zoeller RT, et al. (2009) Identification of thyroid hormone receptor binding sites and target genes using ChIP-on-chip in developing mouse cerebellum. PLoS ONE 4: e4610.
  24. 24. Potter GB, Facchinetti F, Beaudoin GM, 3rd , Thompson CC (2001) Neuronal expression of synaptotagmin-related gene 1 is regulated by thyroid hormone during cerebellar development. J Neurosci 21: 4373–4380.
  25. 25. Martel J, Cayrou C, Puymirat J (2002) Identification of new thyroid hormone-regulated genes in rat brain neuronal cultures. Neuroreport 13: 1849–1851.
  26. 26. Denver RJ, Williamson KE (2009) Identification of a thyroid hormone response element in the mouse Kruppel-like factor 9 gene to explain its postnatal expression in the brain. Endocrinology 150: 3935–3943.
  27. 27. Snyder EY, Deitcher DL, Walsh C, Arnold-Aldea S, Hartwieg EA, et al. (1992) Multipotent neural cell lines can engraft and participate in development of mouse cerebellum. Cell 68: 33–51.
  28. 28. Komuro H, Yacubova E (2003) Recent advances in cerebellar granule cell migration. Cell Mol Life Sci 60: 1084–1098.
  29. 29. Mi R, Luo Y, Cai J, Limke TL, Rao MS, et al. (2005) Immortalized neural stem cells differ from nonimmortalized cortical neurospheres and cerebellar granule cell progenitors. Exp Neurol 194: 301–319.
  30. 30. Riess P, Zhang C, Saatman KE, Laurer HL, Longhi LG, et al. (2002) Transplanted neural stem cells survive, differentiate, and improve neurological motor function after experimental traumatic brain injury. Neurosurgery 51: 1043–1052; discussion 1052–1044:
  31. 31. Prosser HM, Rzadzinska AK, Steel KP, Bradley A (2008) Mosaic complementation demonstrates a regulatory role for myosin VIIa in actin dynamics of stereocilia. Mol Cell Biol 28: 1702–1712.
  32. 32. Doyle JP, Dougherty JD, Heiman M, Schmidt EF, Stevens TR, et al. (2008) Application of a translational profiling approach for the comparative analysis of CNS cell types. Cell 135: 749–762.
  33. 33. Larue L, Ohsugi M, Hirchenhain J, Kemler R (1994) E-cadherin null mutant embryos fail to form a trophectoderm epithelium. Proc Natl Acad Sci U S A. 91: 8263–8267.
  34. 34. Lopez-Molina L, Conquet F, Dubois-Dauphin M, Schibler U (1997) The DBP gene is expressed according to a circadian rhythm in the suprachiasmatic nucleus and influences circadian behavior. EMBO J. 16: 6762–6771.
  35. 35. Zarach JM, Beaudoin GM , Coulombe PA, Thompson CC (2004) The co-repressor hairless has a role in epithelial cell differentiation in the skin. Development. 131: 4189–4200.
  36. 36. Morita M, Kobayashi A, Yamashita T, Shimanuki T, Nakajima O, et al. (2003) Functional analysis of basic transcription element binding protein by gene targeting technology. Mol Cell Biol 23: 2489–2500.
  37. 37. Chesney J, Telang S, Yalcin A, Clem A, Wallis N, et al. (2005) Targeted disruption of inducible 6-phosphofructo-2-kinase results in embryonic lethality. Biochem Biophys Res Commun. 331: 139–146.
  38. 38. Klugmann M, Schwab MH, Pühlhofer A, Schneider A, Zimmermann F, et al. (1997) Assembly of CNS myelin in the absence of proteolipid protein. Neuron. 18: 59–70.
  39. 39. Renaud J, Kerjan G, Sumita I, Zagar Y, Georget V, et al. (2008) Plexin-A2 and its ligand, Sema6A, control nucleus-centrosome coupling in migrating granule cells. Nat Neurosci 11: 440–449.
  40. 40. Molnár Z, López-Bendito G, Small J, Partridge LD, Blakemore C, et al. (2002) Normal development of embryonic thalamocortical connectivity in the absence of evoked synaptic activity. J Neurosci. 22: 10313–10323.
  41. 41. Bailey CD, Tucholski J, Johnson GV (2005) Transglutaminases in neurodegenerative disorders. Prog Exp Tumor Res. 38: 139–157.
  42. 42. Pitari G, Malergue F, Martin F, Philippe JM, Massucci MT, et al. (2000) Pantetheinase activity of membrane-bound Vanin-1: lack of free cysteamine in tissues of Vanin-1 deficient mice. FEBS Lett. 483: 149–154.
  43. 43. Sutherland AP, Zhang H, Zhang Y, Michaud M, Xie Z, et al. (2009) Zinc finger protein Zbtb20 is essential for postnatal survival and glucose homeostasis. Mol Cell Biol. 29: 2804–2815.
  44. 44. Fossat N, Chatelain G, Brun G, Lamonerie T (2006) Temporal and spatial delineation of mouse Otx2 functions by conditional self-knockout. EMBO Rep 7: 824–830.
  45. 45. Pascual M, Abasolo I, Mingorance-Le Meur A, Martinez A, Del Rio JA, et al. (2007) Cerebellar GABAergic progenitors adopt an external granule cell-like phenotype in the absence of Ptf1a transcription factor expression. Proc Natl Acad Sci U S A 104: 5193–5198.
  46. 46. Lappe-Siefke C, Goebbels S, Gravel M, Nicksch E, Lee J, et al. (2003) Disruption of Cnp1 uncouples oligodendroglial functions in axonal support and myelination. Nat Genet 33: 366–374.
  47. 47. Thompson CC, Potter GB (2000) Thyroid hormone action in neural development. Cereb Cortex 10: 939–945.
  48. 48. Zhang HM, Su Q, Luo M (2008) Thyroid hormone regulates the expression of SNAP-25 during rat brain development. Mol Cell Biochem 307: 169–175.
  49. 49. Diez D, Grijota-Martinez C, Agretti P, De Marco G, Tonacchera M, et al. (2008) Thyroid hormone action in the adult brain: Gene expression profiling of the effects of single and multiple doses of Triiodo-L-thyronine in the rat striatum. Endocrinology.
  50. 50. Denver RJ, Ouellet L, Furling D, Kobayashi A, Fujii-Kuriyama Y, et al. (1999) Basic transcription element-binding protein (BTEB) is a thyroid hormone-regulated gene in the developing central nervous system. Evidence for a role in neurite outgrowth. J Biol Chem 274: 23128–23134.
  51. 51. Morte B, Diez D, Auso E, Belinchon MM, Gil-Ibanez P, et al. (2011) Thyroid hormone regulation of gene expression in the developing rat fetal cerebral cortex: prominent role of the Ca2+/calmodulin-dependent protein kinase IV pathway. Endocrinology 151: 810–820.
  52. 52. Furlow JD, Kanamori A (2002) The transcription factor basic transcription element-binding protein 1 is a direct thyroid hormone response gene in the frog Xenopus laevis. Endocrinology 143: 3295–3305.
  53. 53. Xie Z, Ma X, Ji W, Zhou G, Lu Y, et al. (2010) Zbtb20 is essential for the specification of CA1 field identity in the developing hippocampus. Proc Natl Acad Sci U S A 107: 6510–6515.
  54. 54. Nielsen JV, Blom JB, Noraberg J, Jensen NA (2010) Zbtb20 is essential for the specification of CA1 Zbtb20-induced CA1 pyramidal neuron development and area enlargement in the cerebral midline cortex of mice. Cereb Cortex 20: 1904–1914.
  55. 55. Dugas J, YC T, Speed TP, Ngai J, Barres B (2006) Functional genomic analysis of oligodendrocyte differentiation. JNeurosci 26: 10967–10983.
  56. 56. Michalski JP, Anderson C, Beauvais A, De Repentigny Y, Kothary R (2011) The proteolipid protein promoter drives expression outside of the oligodendrocyte lineage during embryonic and early postnatal development. PLoS One. 6: e19772.
  57. 57. Sotelo C, Dusart I (2009) Intrinsic versus extrinsic determinants during the development of Purkinje cell dendrites. Neuroscience 162: 589–600.
  58. 58. Itoh Y, Esaki T, Kaneshige M, Suzuki H, Cook M, et al. (2001) Brain glucose utilization in mice with a targeted mutation in the thyroid hormone alpha or beta receptor gene. Proc Natl Acad Sci U S A 98: 9913–9918.
  59. 59. Herrero-Mendez A, Almeida A, Fernandez E, Maestre C, Moncada S, et al. (2009) The bioenergetic and antioxidant status of neurons is controlled by continuous degradation of a key glycolytic enzyme by APC/C-Cdh1. Nat Cell Biol 11: 747–752.
  60. 60. Harper ME, Seifert EL (2008) Thyroid hormone effects on mitochondrial energetics. Thyroid 18: 145–156.
  61. 61. Wallis K, Dudazy S, van Hogerlinden M, Nordstrom K, Mittag J, et al. (2010) The Thyroid Hormone Receptor {alpha}1 Protein Is Expressed in Embryonic Postmitotic Neurons and Persists in Most Adult Neurons. Mol Endocrinol 24: 1904–1916.
  62. 62. 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: 2288–2302.
  63. 63. Srinivas S, Watanabe T, Lin CS, William CM, Tanabe Y, et al. (2001) Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev Biol 1: 4.
  64. 64. Livak JL, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2-DDCT method. Methods 25: 402–408.
  65. 65. Burckstummer T, Bennett KL, Preradovic A, Schutze G, Hantschel O, et al. (2006) An efficient tandem affinity purification procedure for interaction proteomics in mammalian cells. Nat Methods 3: 1013–1019.