Introduction

Thyroid hormone (T3) is involved in the development and function of the central nervous system.^1,2,3 The absence of thyroid hormone during critical periods of development results in severe and irreversible mental deficiency, whereas in adult individuals thyroid hormone deficiency or excess may lead to neurological and psychiatric manifestations.⁴ The actions of T3 are mediated through the control of gene expression after interaction with receptors of the nuclear superfamily of ligand-modulated transcription factors.⁵ Several T3 receptor proteins (TRs) are encoded by two distinct genes, TR and TR. The TR gene encodes three proteins, TR1, TRv2, and TRv3, which differ in their carboxyl terminus. TR1 binds T3 and activates or represses target genes, whereas TRv2 and TRv3 do not bind T3 and may have a weak antagonizing effect on TR action.^6,7,8,9,10 In addition to these protein isoforms, there are also truncated products of the gene known as 1 and 2, of unknown function.¹¹ Several amino terminal protein variants are produced from the TR gene: the classical receptors TR1 and TR2, and two newly identified T3-binding proteins, TR3 and TR3.¹²

Such a diversity of T3 receptor gene products has stimulated the search for isoform-specific functions, and mouse strains carrying targeted mutations of T3 receptor genes have been generated.^13,14 TR is involved in pituitary and liver function, and TR-deficient mice are a model for the recessive form of the human syndrome of resistance to thyroid hormone.¹⁵ Some discrete specific functions have also been assigned to TR1 and TR2. TR1 is involved in the development of cochlear function,¹⁶ whereas TR2 is essential for the differentiation of a specific subset of retinal cone photoreceptor cells.¹⁷ TR1 modulates cardiac function and body temperature,¹⁸ whereas the expression of the truncated TR products impairs intestinal development.¹⁹ Another indication of specific functions for each receptor is that TR1 and TR have opposite effects on female sex behavior.²⁰

In the rat brain about 70% of the total T3 receptor protein corresponds to the TR1 isoform.²¹ However, TR1-deficient mice do not show overt signs of central nervous system impairment. This is a major paradox in the thyroid hormone receptor field and may be at least partially explained by assuming that the severe neurological manifestations of profound hypothyroidism are a consequence of the repressor activity of unliganded receptors.²² Thus, neonatal hypothyroidism does not affect cerebellar development in TR1-deficient mice as when hypothyroidism is induced in wild-type animals.²³ Also, TR1 deletion can rescue the severe developmental defects observed in congenitally hypothyroid mice,²⁴ and the expression of dominant negative forms of TRs interferes with brain development and function.^25,26 In view of these findings, it is important to define specific roles for TR1 in the brain. In the present paper we describe behavioral alterations of hippocampal origin in adult TR1^-/- mice, which correlate with a reduced number of GABAergic perisomatic terminals on the CA1 pyramidal neurons. The data suggest a specific involvement of TR1 in the specification of inhibitory hippocampal circuitry and in the modulation of emotional reactivity and hippocampal-dependent memories.

Materials and methods

Animals

Adult male Balb/c wild-type and TR1-deficient mice¹⁸ 3–4 months old were used in all experiments. The wild-type mice were generated from crosses of the heterozygotes to have the same genetic background as the mutant mice. Animal care procedures were conducted in accordance with the guidelines set by the European Community Council Directives (86/609/EEC). Animals were under temperature-(222°C) and light-(12:12 light–dark cycle; lights on at 07.00) controlled conditions and had free access to food and water in the colony room. Mice for behavioral tests were individually housed. They were weighed and handled daily for around 4 days before the experiment for habituation to experimental manipulations. The experiments were always conducted between 08.00 and 14.30. Mice for histological analysis were anesthetized by intraperitoneal injection of a mixture of ketamine (4 mg/100 g body weight) and metedomidine (15 mg/100 g body weight) and the mice were perfused transcardially with 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.4. The brains were removed, post-fixed overnight in the same solution and dissected across the mid-line. The left hemibrain was processed for in situ hybridization–immunohistochemistry and the right for cytoarchitectonic and immunohistochemistry studies. The entire analysis was done in the dorsal hippocampus and the somatosensory cortex. Thyroid hormone concentrations in whole brain extracts were determined as described.²⁷ For these assays the brains of animals subjected to the cued fear conditioning test were used.

Behavioral tests

Each test was carried out in independent groups of mice. In the open field test, a total of five wild-type and four TR1-deficient mice were used. A circular open field of 140 cm diameter and 32 cm height was used. Each mouse was transferred to the center of the field, and its behavior and locomotor activity were monitored for 6 min using a video camera and computerized system (Ethovision 1.90, Noldus IT, Wageningen, The Netherlands).

In the fear conditioning tests, training and tests took place in a Pavlovian conditioning apparatus. This is a rodent observation cage (303725 cm) that was positioned inside a sound-attenuating chamber. The cage floor consisted of steel rods through which scrambled shocks from a shock generator (Model LI100-26 Shocker, LETICA IC, Spain) could be delivered. Each observation cage was cleaned with a 1% acetic acid solution before and after each session. The sound-attenuating chambers were illuminated by a 20 W white light bulb. Ventilation fans provided background noise at 68 dB.

In the contextual fear conditioning 11 wild-type and eight TR1-deficient mice were used. Three minutes after the mice were placed in the conditioning chamber, they received three 2-s shocks (unconditioned stimuli) at 0.75 mA. The intertrial interval was 60 s, and the mice were removed from the conditioning chambers 30 s after the final shock presentation and returned to their home cages. Thus, a conditioning session lasted approximately 5.5 min. Testing for contextual fear conditioning was performed 24 h (24-h test), 7 days (1-week test), and 1 month (1-month test) after conditioning. At training, behavioral scores were carried out for the 3-min period prior to shock (pre shock period) and for the 2.5-min period starting immediately after the first shock presentation (post shock period). Scores for each of these periods were analysed separately. At testing, mice were placed back into the original training context in the absence of shock for an 8-min context test. Using a time-sampling procedure, every 2 s, each mouse was scored blind as either freezing or active. Freezing was defined as behavioral immobility except for movement needed to breathe. Behavior was evaluated on each experimental session.

In the cued fear conditioning a total of eight wild-type and seven TR1-deficient mice were used. Training consisted of exposure of the mice to a conditioning chamber (160 s), followed by a tone (80 dB sound at 2000 Hz) that was presented for 20 s before the electric foot shock (2 s, 0.75 mA, constant current). Each mouse received three 2-s shocks and the intertrial interval was 60 s. The mice were removed from the conditioning chambers 30 s after the final shock presentation, and they were returned to their home cages. Thus, a conditioning session lasted approximately 5.5 min. One day after conditioning (24-h test), the mice were placed in a novel context (3 min) and re-exposed to the tone (3 min). Subsequently, 1 week later (1-week test), the mice were placed in a novel context, different from the one previously used in the 24-h test. After a 3 min baseline period, the training tone was played for 3 min. Using a time-sampling procedure, each mouse was scored blind every 2 s as either freezing or active, during re-exposure to the tone.

For statistical analysis, the freezing data for each observational period were transformed to a percentage of total observations and the results were expressed as meanSEM. Statistical comparisons between TR1-deficient and wild-type mice were performed using the Mann–Whitney or Student's t-tests, when appropriate. Differences were considered significant when P<0.05.

Histological analysis

A total of four wild-type and four TR1-deficient mice were used. The hemibrains were cut serially at 100-m on a vibratome in the coronal plane. Some of these sections were pre-treated with a solution of 50% ethanol and 1% hydrogen peroxide (Merk, 107210) in PB to remove endogenous peroxidase activity, then pre-incubated in a stock solution (3% normal goat or horse serum (Vector Laboratories, S-1000 and S-2000, respectively) and 0.25% Triton X-100 (Merk, 86031000) in PB) for 2 h at room temperature. Sections were then incubated in parallel for 24 h at 4°C in the above solution containing one of the following primary polyclonal (rabbit) or monoclonal (mouse) antibodies: rabbit anti-GABA transporter type 1 (GAT-1, 1:500; Chemicon, Temecula, CA, USA) or mouse anti-parvalbumin (PV, 1/4000; Sigma, C-8666). The sections were subsequently washed in PB, incubated in a species-specific biotinylated antibody (Vector Laboratories, BA-1000, BA-2020) diluted 1:200 in PB for 1 h at room temperature, and processed by the avidin–biotin–peroxidase method as above. Sections were mounted on glass slides, dehydrated, cleared in xylene and coverslipped. Adjacent sections were stained with thionin.

Counts of PV-labeled somata were performed in the CA1 field of the dorsal hippocampus, directly from light microscope slides. With the aid of an eyepiece reticule, we counted the number of PV-labeled cell somata within an area of 616616 m (total surface 0.379 mm²), centered on CA1, and including all strata from stratum oriens to stratum lacunosum-moleculare. This counting frame was used on two sections from each wild-type and mutant animal. The statistical analysis was performed with the aid of the SPSS statistical package. The distribution of the data set was determined by skew test.

Some sections were double-stained with rabbit anti-GAT-1 and mouse anti-PV. Following incubation with the primary antibodies (diluted to the above specifications in stock solution), the sections were washed in PB and incubated for 1 h at room temperature in a solution containing biotinylated goat anti-rabbit IgG (Vector Laboratories), diluted 1:200. The sections were then incubated for 2 h at room temperature in a mixture of Cy5-conjugated goat anti-mouse IgG (diluted 1:200) and Cy2-conjugated streptavidin (diluted 1:1000) (Amersham Life Science, Arlington Heights, IL, USA). Afterwards, the sections were mounted and examined in a Leica TCS 4D confocal laser scanning microscope equipped with an argon/krypton mixed gas laser and a Leitz DMIRB fluorescence microscope. Excitation peaks at 489 and 649 nm were used to visualize Cy2- and Cy5-labeled profiles, respectively. Fluorescently labeled Cy2 and Cy5 profiles were recorded through separate channels. Control sections for immunocytochemistry consisted of the replacement of the primary antibody with normal serum or replacement of the appropriate secondary antibody with an inappropriate secondary antibody. No significant staining was observed under these control conditions.

In situ hybridization

In situ hybridization was performed using [³⁵S]-UTP-labeled riboprobes and free-floating 25 m sections as described in detail elsewhere.²⁸ Four animals of each condition were analysed. The riboprobes were synthesized from templates encoding nucleotides 1161–1428 of mouse TR1 and 21–329 of mouse TR1. After in situ hybridization, the sections were subjected to immunohistochemistry as described.²⁸ Quantitation of hybridization grains corresponding to TR1 and TR1 per PV-labeled cell somata was performed in the stratum pyramidale and oriens of the CA1 field of the dorsal hippocampus and in the somatosensory cortex of wild-type mice, directly from light microscope slides. In the somatosensory cortex, the number of grains per cell was counted in all the PV-ir cell somata present in a 363 m width column, which included all cortical layers. In the CA1 hippocampal region, the number of grains were counted in all the PV-immunoreactive(ir) somata from stratum oriens and stratum pyramidale. Quantitation was performed on three sections from three wild-type mice after correction for background hybridization. For each receptor probe, at least 44 cells were counted for CA1, and 108 cells for the somatosensory cortex, in each animal.

Results

Mice lacking TR1 have behavioral abnormalities of hippocampal origin

To evaluate behavior as well as learning and memory abilities of TR1-deficient mice, different tests were performed. First, an open field task was done to evaluate the emotionality of the mutant mice. TR1-deficient mice (n=4) spent significantly less time exploring the center of the open field compared to wild-type mice (n=5) (t=8.621; P<0.001), and showed a reduced rearing behavior (t=3.16; P<0.01). Moreover, in contrast to wild-type animals, which were constantly moving and exploring the open field, mutant mice displayed a small percentage of time freezing, which significantly differed from wild-type (Mann–Whitney test; P<0.02) (Figure 1a).

Figure 1.

TR1-deficient mice are defective in contextual but not cued fear tasks. Behavior of wild-type (open boxes) and TR1-deficient mice (solid boxes) in the open field test (a), contextual (b) and tone cued (c) fear conditioning tests. In the open field (a), the mutant mice (n=4) exhibited more freezing and less rearing behavior than the wild-type mice (n=5). In the contextual fear conditioning task (b), the mutant (n=8) and the wild-type mice (n=11) presented similar freezing values during both the post shock period at training and the 24-h test. However, the freezing levels at the 1-week test were considerably higher in the mutant mice than in the wild-type mice. On the contrary, in the cued conditioning task (c), no differences in freezing levels were found between the mutant (n=7) and the wild-type (n=8) mice at any of the training and testing (24-h and 1-week) sessions.

Full figure and legend (49K)

Subsequently, we used Pavlovian contextual and cued conditioning freezing to examine learning and memory abilities of the mice. In these paradigms, mice learn to fear an otherwise innocuous stimulus, a context or a tone, by pairing it with an aversive stimulus, such as a foot shock. When re-exposed to either the context or the tone, mice exhibit a conditioned fear response, freezing, in which all movements cease, except those required to breathe.²⁹ Cued fear conditioning depends on the integrity of the amygdala³⁰ while contextual fear conditioning is sensitive to both amygdala and hippocampal lesions.^30,31,32

In the contextual fear conditioning test, both groups of mice, wild-type (n=11) and TR1-deficient (n=8), did not differ in freezing levels either during the postshock interval at training (which indicates a similar sensitivity to the shock), or at the 24 h test (t=-0.467, NS and t=0.945, NS, respectively, Figure 1b). However, compared to wild-type animals that progressively reduced their levels of conditioning when re-exposed to the conditioned context 1 week after training, a more intense freezing was observed in TR1-deficient mice in contextual fear conditioning 1 week after training (t=-3.92; P<0.003, Figure 1b). Interestingly, this increased contextual freezing was no longer evident when the mice were tested 1 month after conditioning (t=-0.931; NS). On the contrary, no differences in freezing were observed when wild-type (n = 8) and mutant (n = 7) mice were tested in the auditory fear (cued) conditioning task, either at training (t=-1,59; NS), or at 24 h (t=0,363; NS) or 1 week (t=-1,63; NS) post-training tests (Figure 1c).

TR1-deficient mice show a selective alteration in the perisomatic GABAergic innervation of pyramidal CA1 neurons

The above data led us to hypothesize that lack of the TR1 receptor subtype might lead to specific alterations in the hippocampus that might be responsible for the alterations of behavior. Accordingly, we first examined the laminar architecture of the cerebral neocortex and the hippocampal formation using Nissl staining, but found no apparent differences between the wild-type and the mutant mice. We then examined potential alterations in inhibitory cortical circuits in the TR1-deficient mice by analysing the immunostaining patterns of PV and GAT-1. PV is a Ca²⁺-binding protein that labels a subpopulation of GABAergic interneurons,³³ which include chandelier cells and basket cells, the most powerful inhibitory neurons of the cerebral cortex (reviewed by DeFelipe³⁴ and by Freund and Buzsáki³⁵). GAT-1 is the predominant type of GABA transporter in the cerebral cortex,³⁶ which has been shown to be a good marker of GABAergic terminals, although it is also present in some astrocytic processes.^37,38,39

The immunostaining patterns for PV and GAT-1 in the neocortex were virtually identical in the wild-type and mutant mice. In both groups of animals, numerous PV (Figures 2a and b) and GAT-1-immunoreactive (ir) terminal-like puncta (Figures 3a and b) were seen in the neuropil and around the somata of unstained cells. In contrast, specific alterations were observed in the CA1 field of the hippocampus, as follows. In wild-type mice PV-ir neurons constitute a subpopulation of nonpyramidal cells whose somata are mainly localized in the strata pyramidale and oriens.³⁵ PV-ir terminal-like puncta were found mostly in the stratum pyramidale, forming dense clusters around pyramidal cells (Figure 2c). We use the term perisomatic terminals to refer to the terminal axonal plexus formed by both basket cells and chandelier cells (ie, axon terminals innervating the cell body, proximal dendrites, and axon initial segments). Immunohistochemistry for GAT-1 in wild-type mice shows numerous labeled terminal-like puncta, but no cell bodies. These puncta were seen homogeneously distributed in the neuropil throughout all layers of the CA1 subfield and around the somata of pyramidal cells, forming dense clusters (Figure 3c), which were similar to those formed by PV-ir in the stratum pyramidale. In double-labeling experiments, it was found that the great majority of GAT-1 puncta in the stratum oriens, stratum radiatum and stratum lacunosum-moleculare did not contain PV, but in the stratum pyramidale there was a high degree of colocalization of both markers, thus demonstrating that GAT-1-ir in the perisomatic clusters is contained in neuronal terminals (Figure 4).

Figure 2.

The density of PV-ir terminals is decreased in the CA1 region of TR1-deficient mice. PV-ir neurons in the somatosensory cortex (a,b) and CA1 hippocampal region (c,d) in wild-type (a,c) and TR1-deficient (b,d) mice. SO: stratum oriens; SP: stratum pyramidale; SR: stratum radiatum. The arrows point to the PV-ir terminal around pyramidal neurons. Scale bar, 13 m.

Full figure and legend (246K)

Figure 3.

The density of GAT-1-ir terminals is decreased in the CA1 region of TR1-deficient mice. GAT-1-ir terminals in the somatosensory cortex (a,b) and CA1 hippocampal region (c,d) in wild-type (a,c) and TR1-deficient (b,d) mice. SO: stratum oriens; SP: stratum pyramidale; SR; stratum radiatum. The arrows point to GAT-1-ir terminals around pyramidal neurons. Scale bar, 13 m.

Full figure and legend (275K)

Figure 4.

Coexpression of PV and GAT-1 in the perisomatic terminals of CA1 neurons in wild-type mice. Simultaneous detection of PV (red, (a)) and GAT-1 (green, (b)) immunoreactivity in the CA1 region of the dorsal hippocampus in wild-type mice. The pseudo-colored confocal images showed that there is a high degree of double-labeled terminals (yellow, (c)) specifically surrounding cell bodies in the stratum pyramidale, and not in the other strata of the CA1 hippocampal region as oriens and radiatum. This indicates that perisomatic terminals coexpressed both PV and GAT-1. An interneuron, with the cell body positive for PV and with PV-ir and GAT-1-ir terminals, is seen in the stratum pyramidale (arrows in (a) and (c)). Scale bar, 37.5 m.

Full figure and legend (239K)

Furthermore, the density of PV-ir and GAT-1-ir terminals in stratum oriens, stratum radiatum and stratum lacunosum-moleculare was similar in wild-type and mutant mice. However, in TR1-deficient mice there was a prominent and consistent reduction in the density of both PV and GAT-1 terminals forming the perisomatic clusters (Figures 2c and d and Figures 3c and d). In contrast, the number of PV-ir somata per surface area in CA1 was similar in wild-type and mutant mice. The average number (SD) of immunoreactive cells in a 0.379 mm² field (which included all strata from the stratum oriens to the stratum lacunosum-moleculare) was 7.92.4 for wild-type mice and 6.43.3 for the mutant mice. Statistical comparison revealed no significant difference between the two groups of animals (U=18.5, Z=-1.42, P=0.153).

Parvalbumin interneurons of the CA1 hippocampal region express preferentially TR1 over the TR1 isoform

Both the behavioral and the immunohistochemical data described so far suggested that lack of TR1 induced alterations in hippocampal circuitry, at the level of the basket cells, which could not be compensated by the TR gene products. We therefore studied the possibility that basket cells in the hippocampus expressed preferentially TR1 over TR1. To this goal, we performed in situ hybridization for TR1 and TR1, combined with PV immunohisto- chemistry. The patterns of receptor expression throughout the brain were identical to those previously reported.^40,41 At the cellular level, these experiments revealed differential expression of receptor subtypes in cells of neocortex and hippocampus that were positive for parvalbumin. Figure 5 shows an example of the colocalization of TR1 and of TR1 (as detected by the silver grains after in situ hybridization) with PV-positive interneurons in the CA1 field (panels (a) and (b)) and in the somatosensory cortex (panels (c) and (d)). In the CA1 field, PV-positive cells express mainly TR1, whereas in the somatosensory cortex they express mainly TR1. After quantification of the data, we found that the average number (SE) of hybridization grains per PV-ir somata in the stratum oriens and stratum pyramidale of the CA1 region was 22.62 and 71.23.6 for TR1 and 9.41.3 and 22.11.4 for TR1, respectively. However, in the somatosensory cortex in a 363-m wide column (which included all cortical layers) the grains were 17.20.7 for TR1 and 25.81 for TR1. Statistical analysis revealed significant differences (P< 0.001) in the expression of TR1 compared to TR1 in the stratum pyramidale and neocortex.

Figure 5.

Expression of thyroid hormone receptor isoforms in PV-ir neurons. In situ hybridization for the receptors was combined with immunohistochemistry for PV. The figure shows expression of TR1 (a,c) and TR1 (b,d) in PV-ir neurons of the CA1 region of the dorsal hippocampus (a,b) and somatosensory cortex (c,d) in wild-type mice. Colocalization of the silver grains with immunoreactivity (arrows) reveals that the PV-ir neurons of CA1 (a) express preferentially TR1 over TR1 (b), whereas the opposite holds for the somatosensory cortex (c,d). Scale bar, 8 m.

Full figure and legend (168K)

Thyroid hormone concentrations in the brains of wild-type and TR^-/- mice

To discard the possibility that the morphological and behavioral alterations of TR^-/- mice are due to hypothyroidism, we measured thyroid hormone concentrations in whole-brain extracts from wild-type mice (n=8) and TR1^-/- mice (n=7). The results showed that T4 was similar in both types of animals (wild-type: 2.330.29 ng/g; TR1^-/-:2.630.39 ng/g, P=0.33), whereas T3 was slightly elevated in the mutant mice (wild type: 1.210.27 ng/g; TR1^-/-: 1.620.31 ng/g, P<0.05).

Discussion

Our findings provide the first data on the specific involvement of TR1 in brain structure and function. TR1-deficient mice presented behavioral alterations that correlated with structural defects in the hippocampus. In the open field test, the behavior displayed by mutant mice revealed a higher emotionality than the wild-type mice, as indicated by their reduced time devoted to exploring the center of the field, and their lower rearing and increased freezing levels. However, mutant mice showed a higher freezing response than wild-type animals when tested 1 week after training in the contextual fear conditioning task. In contrast, no differences were found at any testing time in the auditory conditioning task. The observed behavioral alterations were not due to differences in overall locomotor activity, which is normal in TR1-deficient mice.¹⁸

The different memory retrieval results displayed by TR1-deficient mice in contextual and cued fear conditioning can be due to the different neural systems implicated in these tasks.³⁰ Contextual fear conditioning depends on the function of both the amygdala and the hippocampus, with the latter involved in the formation of an integrated representation of the context. In contrast, cued fear conditioning relies mainly on the amygdala. Therefore, the result of the behavioral tests pointed to a defect at the level of the hippocampus. Indeed, TR1-deficient mice display a prominent and selective reduction of perisomatic terminals expressing the Ca²⁺-binding protein PV and the GABA transporter GAT-1 in the CA1 field of the hippocampus. This suggests a specific involvement of TR1 in the specification of inhibitory hippocampal circuitry. PV immunohistochemistry in the hippocampus selectively labels the axons of basket and chandelier cells,^35,42,43 which are involved in the regulation of efferent activity.⁴⁴ Our results therefore suggest that these inhibitory interneurons are affected in the mutant mice. Since the number of PV-ir neurons were similar in wild-type and TR1-deficient mice, it is likely that the reduction in perisomatic terminals is due to a poor development of the axonal arborizations of basket and chandelier cells rather than to a reduction of the number of cells. Despite this, one has to take into account that each single chandelier or basket cell gives rise to several thousand boutons. For example, a single basket cell may give rise to up to 12 000 boutons.⁴³ Therefore, it might be possible that the loss of just a few chandelier and basket cells, which were not apparent in our quantitative analysis, could have profound consequences on the density of perisomatic terminals, and be responsible for the observations reported here.

The histological alterations of GAT-1 and PV-ir observed in the mutant mice suggest a marked decrease of perisomatic GABAergic inhibition. In the light of the behavioral findings, it is tempting to speculate that a decreased GABAergic inhibition could lead to the potentiation of the strength at which hippocampal-dependent fear memories are stored and, therefore, to a reduced capacity to forget the negative information no longer relevant. This is clearly not a behavioral advantage, first because excessive and prolonged hippocampal activation may lead to structural damage^45,46 and, secondly, because increased emotionality and potentiated negative memories are characteristic of psychopathological states such as depression, and cannot be considered adaptive. In agreement with this interpretation, TR2-deficient mice, which overexpress TR1 in the brain three- to four-fold higher than wild-type mice,¹⁰ are less prone than wild-type mice to show depression-like behaviors (S-O Ögren and B Vennström, unpublished observations).

The results of thyroid hormone determinations in the brain suggest that the behavioral alterations of TR1^-/- mice are not due to hypothyroidism, despite the reported 30% reduction of serum-free T4 in these mice.¹⁸ T4 and T3 concentrations in the cerebral cortex are normal in TR1^-/- mice during the neonatal period.²³ In the adult mice employed in the present work, we find that whole-brain T4 is normal, whereas T3 is slightly elevated. The data suggest that mechanisms, for example, involving deiodinases, operate to compensate for the mild hypothyroidism. Given the pattern of receptor distribution shown in Figure 5 it is unlikely that the structural defects found in our studies are a consequence of moderately elevated T3 concentrations. T4 has been shown to exert actions at the membrane level, but these can be discarded in view of normal T4 concentrations.

In conclusion, our data suggest that TR1 is involved in the regulation of hippocampal function. The phenotype displayed by the TR1-deficient mice is not a transient phenotype and the products of the TR gene could not compensate for the loss of TR1. This is most likely to be due to the fact that PV-ir interneurons express little TR1, and there was no TR1 overexpression in the TR1-deficient mice. Furthermore, mice with TR gene deletion did not show alterations in several tests, such as the open field test, the Morris water maze test, contextual fear conditioning tests and hippocampal field potentials.¹⁵ In humans, TR mutations give rise to the syndrome of resistance to thyroid hormone, which is recognized clinically because pituitary resistance leads to elevated serum levels of both TSH and thyroid hormones. Concerning TR1, human syndromes arising as a result of mutations or deletions of the gene have not been described, and no mutations have been found after scanning the TR gene in several psychiatric syndromes.⁴⁷ Despite this, we believe that our results raise the intriguing possibility that inactivating mutations or deletions of TR1 in humans may have a correlate in behavior, as in the mutant mice. Since TR1 is involved in cardiac function, it could be further speculated that the affected individuals would present psychiatric symptoms in addition to cardiac alterations.⁴⁸ Association of such syndromes with the thyroid system would not be easily recognized in the clinical setting because only slight alterations of the pituitary–thyroid axis are present in TR1 deficiency, at least in mice.¹⁸

References

1. Legrand J. Effects of thyroid hormones on central nervous system. In: Yanai J (ed). Neurobehavioral Teratology. Elsevier Science Publishers: Amsterdam, 1984, pp 331–363.
2. Porterfield SP, Hendrich CE. The role of thyroid hormones in prenatal and neonatal neurological development—current perspectives. Endocr Rev 1993; 14: 94–106. | PubMed |
3. Bernal J, Guadaño-Ferraz A. Thyroid hormone and the development of the brain. Curr Opin Endocrinol Diab 1998; 5: 296–302.
4. Joffe RT, Sokolov STH. Thyroid hormones, the brain, and affective disorders. Crit Rev Neurobiol 1994; 8: 45–63. | PubMed | ChemPort |
5. Ribeiro RCJ, Appriletti JW, Wagner RL, West BL, Feng W, Huber R et al. Mechanisms of thyroid hormone action: insights from X-ray crystallographic and functional studies. Rec Prog Horm Res 1998; 53: 351–394.
6. Macchia PE, Takeuchi Y, Kawai T, Cua K, Gauthier K, Chassande O et al. Increased sensitivity to thyroid hormone in mice with complete deficiency of thyroid hormone receptor . Proc Natl Acad Sci USA 2001; 98: 349–354. | Article | PubMed | ChemPort |
7. Izumo S, Mahdavi V. Thyroid hormone receptor alpha isoforms generated by alternative splicing differentially activate myosin HC gene transcription. Nature (Lond) 1988; 334: 539–542.
8. Koening RJ, Lazar MA, Holdin RA, Brent GA, Larsen PR, Chin WW et al. Inhibition of thyroid hormone receptor action by a non-hormone binding c-erbA protein generated by alternative mRNA splicing. Nature (Lond) 1989; 337: 659–661.
9. Liu RT, Suzuki S, Miyamoto T, Takeda T, Ozata M, DeGroot LJ. The dominant negative effect of thyroid hormone receptor splicing variant 2 does not require binding to a thyroid response element. Mol Endocrinol 1995; 9: 86–95.
10. Salto C, Kindblom JM, Johansson C, Wang Z, Gullberg H, Nordstrom K et al. Ablation of TR2 and a concomitant overexpression of 1 yields a mixed hypo- and hyperthyroid phenotype in mice. Mol Endocrinol 2001; 15: 2115–2128. | Article | PubMed | ISI |
11. Chassande O, Fraichard A, Gauthier K, Flamant F, Legrand C, Savatier P et al. Identification of transcripts initiated from an internal promoter in the c-erbA locus that encode inhibitors of retinoic acid receptor- and triiodothyronine receptor activities. Mol Endocrinol 1997; 11: 1278–1290. | Article | PubMed | ISI | ChemPort |
12. Williams GR. Cloning and characterization of two novel thyroid hormone receptor isoforms. Mol Cell Biol 2000; 20: 8329–8342. | Article | PubMed | ISI | ChemPort |
13. Hsu J-H, Brent GA. Thyroid hormone receptor gene knockouts. Trends Endocrinol Metabol 1998; 9: 103–112.
14. Forrest D, Vennström B. Functions of thyroid hormone receptors in mice. Thyroid 2000; 10: 41–52. | PubMed | ISI | ChemPort |
15. Forrest D, Hanebuth E, Smeyne RJ, Everds N, Stewart CL, Wehner JM, Curran T. Recessive resistance to thyroid hormone in mice lacking thyroid hormone receptor : evidence for tissue- specific modulation of receptor function. EMBO J 1996; 15: 3006–3015. | PubMed | ISI | ChemPort |
16. Forrest D, Erway LC, Ng L, Altschuler R, Curran T. Thyroid hormone receptor is essential for development of auditory function. Nat Genet 1996; 13: 354–357.  | Article | PubMed | ISI | ChemPort |
17. Ng L, Hurley JB, Dierks B, Srinivas M, Salto C, Vennstrom B et al. A thyroid hormone receptor that is required for the development of green cone photoreceptors. Nat Genet 2001; 27: 94–98.  | Article | PubMed | ISI | ChemPort |
18. Wikström L, Johansson C, Salto C, Barlow C, Campos Barros A, Baas F et al. Abnormal heart rate and body temperature in mice lacking thyroid hormone receptor 1. EMBO J 1998; 17: 455–461.  | Article |
19. Gauthier K, Chassande O, Plateroti M, Roux JP, Legrand C, Pain B et al. Different functions for the thyroid hormone receptors TR and TR in the control of thyroid hormone production and post-natal development. EMBO J 1999; 18: 623–631.  | Article | PubMed | ISI | ChemPort |
20. Dellovade TL, Chan J, Vennstrom B, Forrest D, Pfaff DW. The two thyroid hormone receptor genes have opposite effects on estrogen-stimulated sex behaviors. Nat Neurosci 2000; 3: 472–475. | Article | PubMed |
21. Schwartz HL, Strait KA, Ling NC, Oppenheimer JH. Quantitation of rat tissue thyroid hormone binding receptor isoforms by immunoprecipitation of nuclear triiodothyronine binding capacity. J Biol Chem 1992; 267: 11 794–11 799.
22. Forrest D, Reh TA, Rusch A. Neurodevelopmental control by thyroid hormone receptors. Curr Opin Neurobiol 2002; 12: 49–56.
23. Morte B, Manzano J, Scanlan T, Vennström B, Bernal J. Deletion of the thyroid hormone receptor 1 prevents the structural alterations of the cerebellum induced by hypothyroidism. Proc Natl Acad Sci USA 2002; 99: 3985–3989.  | Article | PubMed | ChemPort |
24. Flamant F, Poguet AL, Plateroti M, Chassande O, Gauthier K, Streichenberger N et al. Congenital hypothyroid Pax8(-/-) mutant mice can be rescued by inactivating the TR gene. Mol Endocrinol 2002; 16: 24–32.  | Article | PubMed | ISI | ChemPort |
25. Itoh Y, Esaki T, Kaneshige M, Suzuki H, Cook M, Sokoloff L et al. Brain glucose utilization in mice with a targeted mutation in the thyroid hormone or receptor gene. Proc Natl Acad Sci USA 2001; 98: 9913–9918.  | Article | PubMed | ChemPort |
26. Hashimoto K, Curty FH, Borges PP, Lee CE, Abel ED, Elmquist JK et al. An unliganded thyroid hormone receptor causes severe neurological dysfunction. Proc Natl Acad Sci USA 2001; 98: 3998–4003.  | Article | PubMed | ChemPort |
27. Escobar-Morreale HF, Escobar del Rey FE, Obregon MJ, Morreale de Escobar G. Only the combined treatment with thyroxine and triiodothyronine ensures euthyroidism in all tissues of the thyroidectomized rat. Endocrinology 1996; 137: 2490–2502.  | Article | PubMed | ChemPort |
28. Guadaño-Ferraz A, Escámez MJ, Morte B, Vargiu P, Bernal J. Transcriptional induction of RC3/neurogranin by thyroid hormone: differential neuronal sensitivity is not correlated with thyroid hormone receptor distribution in the brain. Mol Brain Res 1997; 49: 37–44.
29. Fanselow MS, Bolles RC. Naloxone and shock-elicited freezing in the rat. J Comp Physiol Psychol 1979; 93: 736–744.  | Article | PubMed | ChemPort |
30. Phillips RG, LeDoux JE. Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning. Behav Neurosci 1992; 106: 274–285.  | Article | PubMed | ISI | ChemPort |
31. Kim JJ, Fanselow MS. Modality-specific retrograde amnesia of fear. Science 1992; 256: 675–677.  | Article | PubMed | ISI | ChemPort |
32. Maren S, Holt W. The hippocampus and contextual memory retrieval in Pavlovian conditioning. Behav Brain Res 2000; 110: 97–108.  | Article | PubMed | ISI | ChemPort |
33. Celio MR. Parvalbumin in most gamma-aminobutyric acid-containing neurons of the rat cerebral cortex. Science 1986; 231: 995–997.  | PubMed | ChemPort |
34. DeFelipe J. Chandelier cells and epilepsy. Brain 1999; 122: 1807–1822.  | Article | PubMed | ISI |
35. Freund TF, Buzsáki G. Interneurons of the hippocampus. Hippocampus 1996; 6: 347–470.  | Article | PubMed | ISI | ChemPort |
36. Guastella J, Nelson N, Nelson H, Czyzyk L, Keynan S, Miedel MC et al. Cloning and expression of a rat brain GABA transporter. Science 1990; 249: 1303–1306.  | Article | PubMed | ISI | ChemPort |
37. Minelli A, Brecha NC, Karschin C, DeBiasi S, Conti F. GAT-1, a high-affinity GABA plasma membrane transporter, is localized to neurons and astroglia in the cerebral cortex. J Neurosci 1995; 15: 7734–7746.  | PubMed | ISI | ChemPort |
38. Ribak CE, Tong WM, Brecha NC. GABA plasma membrane transporters, GAT-1 and GAT-3, display different distributions in the rat hippocampus. J Comp Neurol 1996; 367: 595–606.  | Article | PubMed | ISI | ChemPort |
39. DeFelipe J, Gonzalez-Albo MC. Chandelier cell axons are immunoreactive for GAT-1 in the human neocortex. Neuroreport 1998; 9: 467–470.
40. Mellström B, Naranjo JR, Santos A, González AM, Bernal J. Independent expression of the and c-erbA genes in developing rat brain. Mol Endocrinol 1991; 5: 1339–1350.  | PubMed | ISI |
41. Bradley DJ, Towle HC, Young WS. Spatial and temporal expression of - and -thyroid hormone receptor mRNAs, including the ₂-subtype, in the developing mammalian nervous system. J Neurosci 1992; 12: 2288–2302.  | PubMed | ISI | ChemPort |
42. Katsumaru H, Kosaka T, Heizmann CW, Hama K. Immunocytochemical study of GABAergic neurons containing the calcium-binding protein parvalbumin in the rat hippocampus. Exp Brain Res 1988; 72: 347–362.  | PubMed | ISI | ChemPort |
43. Sik A, Penttonen M, Ylinen A, Buzsaki G. Hippocampal CA1 interneurons: an in vivo intracellular labeling study. J Neurosci 1995; 15: 6651–6665.  | PubMed | ISI | ChemPort |
44. Miles R, Toth K, Gulyas AI, Hajos N, Freund TF. Differences between somatic and dendritic inhibition in the hippocampus. Neuron 1996; 16: 815–823.  | Article | PubMed | ISI | ChemPort |
45. Choi DW. Excitotoxic cell death. J Neurobiol 1992; 23: 1261–1276.  | Article | PubMed | ISI | ChemPort |
46. Sacaan AI, Schoepp DD. Activation of hippocampal metabotropic excitatory amino acid receptors leads to seizures and neuronal damage. Neurosci Lett 1992; 139: 77–82.  | Article | PubMed | ISI | ChemPort |
47. Feng J, Yan J, Michaud S, Craddock N, Jones IR, Cook Jr EH et al. Scanning of estrogen receptor (ERalpha) and thyroid hormone receptor (TR) genes in patients with psychiatric diseases: four missense mutations identified in ER gene. Am J Med Genet 2001; 105: 369–374.
48. Johansson C, Vennstrom B, Thoren P. Evidence that decreased heart rate in thyroid hormone receptor-1-deficient mice is an intrinsic defect. Am J Physiol 1998; 275: R640–R646. | PubMed | ISI | ChemPort |

Acknowledgements

This work was supported by grants PM98-0011, PM99-0027 and PM99-0105 from the DGICYT. BV was supported by the Swedish Cancer Society. AG-F is the recipient of a contract from the Ramon y Cajal Program of the Ministry of Science and Technology of Spain. CV and RB-P are supported by fellowships from the Community of Madrid (0177 and 01/0782/2000). We thank Prof. Gabriella Morreale de Escobar for T4 and T3 determinations, Javier Pérez for the art work, and Fernando Núñez, Pablo Señor and Miguel Marsa for the care of animals.