Rhes Is Involved in Striatal Function

The development and the function of central nervous system dependon thyroid hormones. In humans, the lack of thyroid hormonescauses cretinism, a syndrome of severe mental deficiency. Itis assumed that thyroid hormones affect the normal developmentand function of the brain by activating or suppressing targetgene expression because several genes expressed in the brainhave been shown to be under thyroid hormone control. Among these,the Rhes gene, encoding a small GTP-binding protein, is predominantlyexpressed in the striatal region of the brain. To clarify therole of Rhes in vivo, we disrupted the Rhes gene by homologousrecombination in embryonic stem cells and generated mice homozygousfor the Rhes null mutation (Rhes^–/–). Rhes^–/–mice were viable but weighed less than wild-type mice. Furthermore,they showed behavioral abnormalities, displaying a gender-dependentincrease in anxiety levels and a clear motor coordination deficitbut no learning or memory impairment. These results suggestthat Rhes disruption affects selected behavioral competencies.

INTRODUCTION

Top

Introduction

The thyroid hormones thyroxine (T₄) and triiodothyronine (T₃)have many physiological effects. They exert their actions inall tissues examined and affect many metabolic pathways. Someof the most prominent effects of thyroid hormones occur duringfetal development and in early childhood. In humans, the lackof adequate levels of thyroid hormones in the first trimesterof life, such as in iodine deficiency (endemic cretinism) (8,9), or in developmental abnormalities of the thyroid gland (congenitalhypothyroidism) (22, 28, 55) results in cretinism, a syndromeof severe mental deficiency, which may be accompanied by retardedgrowth and/or neurological deficits, such as spastic diplegia.Many of these developmental effects are not reversed by latertreatment with hormone, indicating that thyroid hormone actsin a specific developmental window. Therefore, adequate levelsof thyroid hormone are required for normal central nervous systemdevelopment.

To date, several specific central nervous system genes whoseexpression is controlled by thyroid hormone have been identified.The expression of these genes may be decreased (2, 5) or increased(1, 18) in hypothyroidism. Furthermore, the total or partialabsence of thyroid hormones may also affect either mRNA stability(43, 54) or the mRNA translational process (43, 57, 60). Theidentification of thyroid hormone target genes in the centralnervous system and the understanding of their function in centralnervous system development are important to understanding thepathogenesis of neurological cretinism at the molecular level.

In order to understand the molecular basis of neurological cretinism,we studied the Rhes (Ras homolog enriched in striatum) gene(24). Rhes is predominantly expressed in the striatum, and itsexpression is controlled by thyroid hormones (59). Interestingly,several lines of evidence indicate that in neurological cretinism,there is damage of striatum, which determines a striatopallidalsyndrome with poor motor coordination and spastic diplegia (8,9, 39).

Rhes, composed of 266 amino acids, belongs to the RASD subfamilyof the Ras-related GTP-binding protein superfamily. Rhes has95% identity with TEM2 (58) and 62% identity with Dexras1 (37),which are other members of the RASD subfamily. Ras family proteinsare molecular switches that respond to extracellular signalsand regulate intracellular signal pathways controlling cellgrowth (40, 41), gene transcription (20, 61), mRNA stabilityand translation (7, 15, 52), cytoskeleton organization (33,38), peptide trafficking (23, 46, 50), and secretion (3, 45).In the central nervous system, Ras protein controls pathwaysthat are involved in synaptic plasticity, learning, and memory(10).

To assess the role of Rhes in mature striatum and in the pathogenesisof neurological cretinism, we generated mice carrying null mutationsin the Rhes gene by a gene-targeting method. In this paper,we show that mice homozygous for the Rhes mutation are viableand fertile but smaller than wild-type mice. Furthermore, theyshow a gender-dependent increase in anxiety levels and a motorcoordination deficit but no learning or memory impairment.

MATERIALS AND METHODS

Top

Materials and Methods

Tissue preservation and histological analysis. Mice were killed by CO₂ asphyxiation. Brains and embryos wereisolated, fixed, and embedded in paraffin as described previously(19). Brains were cut in 10-µm sections; embryos werecut in 7-µm sections. In situ hybridization was carriedout on paraffin-embedded sections as described previously (19).The Rhes riboprobe was transcribed from a 372-bp AflII-PvuIIgenomic fragment of the 3' untranslated region located 42 bpdownstream of the Rhes stop codon. The EGFP (enhanced greenfluorescent protein) riboprobe was transcribed from a 747-bpNcoI-BamHI fragment, which contains all of the EGFP coding region.

RNA analysis. Total RNA was isolated from adult mouse tissues, staged embryosbrains, and cultured cells by the guanidine hydrochloride procedureas previously described (16); 5 µg of total RNA, previouslytreated with RNase-free DNase I (Roche), was used for reversetranscription (RT)-PCR analysis. Reverse transcription of mRNAswas carried out with the SuperScript preamplification systemfor first-strand cDNA synthesis (Life Technologies). Single-strandedcDNAs in 2 µl of a 25-µl reaction mixture were amplifiedby PCR with Taq DNA polymerase (Roche). Glyceraldehyde-3-phosphatedehydrogenase (G3PDH) mRNA was amplified as an internal controlfor the reverse transcription reaction. The oligonucleotideprimers used were Rhes (5'-ACTAGTTCAGGACAGAGCTCTGAC-3' and 5'-CAGCAGGTGTCTTTATCCAGAGTC-3')and G3PDH (5'-TCCACCACCCTGTTGCTGTA-3' and 5'-ACCACAGTCCATGCCATCAC-3').For Northern blot analysis, 15 µg of total RNA was separatedon a 1% formaldehyde-agarose gel, blotted onto a Hybond N nylonmembrane (Amersham), and hybridized with the 372-bp AflII-PvuIIgenomic fragment labeled with ³²P.

Rhes antibody preparation. The sequence encoding full-length rat Rhes, included betweenthe NdeI and EcoRV sites, was cloned in the NdeI site and filled-inBamHI site of vector pET15b; the full-length protein fused witha 6His stretch at its NH₂ terminus (6H-Rhes) was expressed inEscherichia coli BL21(DE3). The protein was solubilized in 4M urea and injected into rabbits (30). The anti-Rhes antiserumwas purified by affinity chromatography as previously described(19).

Western blot analysis. Total protein extracts from wild-type and knockout striatumand from transfected cells were prepared as previously described(34), resolved by sodium dodecyl sulfate (SDS)-4 to 15% polyacrylamidegel electrophoresis (PAGE) on a precast gel (Bio-Rad), and transferredto a polyvinylidene difluoride membrane (Millipore). As a positivecontrol, HeLa cells were transfected with the Pb-Rhes construct,encoding Rhes protein under the control of the PGK-1 promoter.The blot was probed with a 1:20,000 dilution of polyclonal anti-Rhesantibody and developed with the ECL Plus Western blotting detectionreagent (Amersham Life Science).

Generation of knockout mice. The Rhes genomic locus was isolated by PCR screening of thephage artificial chromosome library RPCI-21 (provided by theYac Screening Center, DIBIT-HSR and IGBE-CNR, Milan, Italy).The internal ribosome entry site-EGFP cassette and the PGKneocassette, flanked by loxP sequences, were flanked by two Rhesgenomic DNA fragments: a 2.9-kb fragment including the Rhestranslational start codon and a 5-kb fragment (see Fig. 3A).The 2.9-kb genomic fragment underwent a site-directed mutagenesisreaction which allowed insertion of a stop codon in each readingframe and EcoRI and XhoI restriction sites downstrea of theRhes translational start codon. The site-directed mutagenesisreaction was performed with the QuikChange site-directed mutagenesiskit (Stratagene) according to the manufacturer's instructions.

View larger version (25K):
[in this window]
[in a new window]
FIG. 3. (A) Homologous recombination in ES cells. Targeting vector and wild-type and mutant Rhes allele maps are shown. Restriction enzyme sites and probes (represented as boxes located upstream and downstream of the homologous arms) are shown. The fragments obtained from wild-type and mutant allele digestions are also shown. K, KpnI; X, XbaI; S, StuI; RI, EcoRI. (B) Identification of ES cell recombinant clones by Southern blot analysis. Left panel: DNA isolated from ES cell clones was digested with EcoRI and probed with a 0.8-kb KpnI fragment (located 5' to the genomic fragment for homologous recombination), yielding 9.5-kb and 6.7-kb bands for the wild-type and targeted alleles, respectively. Right panel: Homologous recombination was confirmed by digesting DNA from positive clones with StuI and probing it with a 0.55-kb KpnI-StuI fragment (located 3' to the genomic fragment for homologous recombination), yielding 6.5- and 5.9-kb bands for the wild-type and targeted alleles, respectively. +/+, wild type; +/–, Rhes^+/–.

The oligonucleotide primers used for the mutagenesis reactionwere 5'-CTTAGCAGGCACCTCGAGTGTGGAATTCCTACTGGACTAGGTCTTCATCATG-3'and 5'-CATGATGAAGACCTAGTCCAGTAGGAATTCCACACTCGAGGTGCCTGCTAAG-3'.The herpes simplex virus thymidine kinase cassette was positioneddownstream of the 3' homology arm. Transfection of the targetingvector and selection of the mutant embryonic stem (ES) cells(R1) were performed as described previously (21) except that400 µg of G418 (Gibco) per ml was used. Screening of EScell clones and genotyping of mice were carried out by Southernblot analysis. Genomic DNA samples were digested with EcoRIor StuI; the 0.8-kb KpnI and 0.55-kb KpnI-StuI fragments, locatedoutside the homology arms, were used as probes (see Fig. 4Afor probe positions and digestion product sizes). Embryo manipulationsand aggregations of ES cell clones with mouse blastocysts ofstrain CD1 were carried out as described previously (36). Chimericanimals with a high contribution of the 129/Sv genetic background,as judged from coat color, were bred with CD1 mice. Offspringheterozygous for the disrupted Rhes gene were mated to eachother to produce Rhes null mice.

View larger version (46K):
[in this window]
[in a new window]
FIG. 4. Identification of knockout mice. Southern blot analysis of offspring generated from mating of mice heterozygous for the Rhes null mutation. DNA isolated from tails was digested with EcoRI and probed with a 0.8-kb KpnI fragment, yielding 9.5- and 6.7-kb bands for the wild-type and targeted alleles, respectively. +/+, wild type; +/–, Rhes^+/–; –/–, Rhes^–/–.

Thyroid-stimulating hormone, glucose, and amylase measurements. Thyroid-stimulating hormone levels in blood collected from wild-typeand knockout mice on postnatal days 30 and 180 were analyzedwith a rat thyroid-stimulating hormone radioimmunoassay kit(Amersham). The glucose levels in blood collected from wild-typeand knockout mice on postnatal day 120 were analyzed as describedpreviously (49). Amylase levels were analyzed as described previously(29).

Animals. Ten adult (five males and five females) wild-type mice and 10adult (five males and five females) knockout mice were usedin all experiments. The wild-type mice were generated from crossesof the heterozygous mice to have the same genetic backgroundas the Rhes^–/– mice. The animals were housed inan air-conditioned room (temperature, 21 ± 1°C; relativehumidity, 60 ± 10%) with the lights on from 2000 to 0800h, in Plexiglas boxes (33 by 13 by 14 cm) with a metal top andsawdust as bedding. Pellet food (enriched standard diet, purchasedfrom Mucedola, Settimo Milanese, Milan, Italy) and tap waterwere continuously available. Before each test, mice were individuallyweighed. For the Morris water maze test, they were weighed boththe first day, before testing, and the last day, after testing.

Behavioral tests. The tests below were carried out according to the referencesgiven in parentheses: open-field (14), passive avoidance (13,17), elevated plus-maze (35, 53), rota-rod (11), and Morriswater maze (12, 42).

Statistical analysis. Analyses of variance were performed on body weight, duration,and frequency data in each behavioral category measured in theopen-field and elevated plus-maze tests and on learning performancedata for the Morris water maze. Post hoc comparisons were performedby Tukey's Honestly Significantly Different test. In the passiveavoidance test, Mann-Whitney analysis was applied to evaluatethe main effect of treatment and Wilcoxon analysis to evaluatethe main effect of sex. In the rota-rod test, in order to assessthe significance of the difference between wild-type and knockoutmice, the Mann-Whitney U test was used. The main effect of speedwas analyzed by the Friedman nonparametric analysis of variance,and Wilcoxon analysis was used to evaluate the main effect ofsex. When no main effect of sex and/or genotype x sex interactionswas found, the sex variable was not considered in the analysis.

RESULTS

Top

Results

Expression of Rhes during brain development and in adult tissues. We analyzed the expression pattern of Rhes in the developingmouse brain by RT-PCR. Total RNA was extracted from the brainsof CD1 staged embryos. RT-PCR analysis detected Rhes mRNA startingfrom embryonic day 13.5 (E13.5) (Fig. 1A). We analyzed the expressionpattern of Rhes in CD1 adult mouse tissues by Northern blotanalysis. Rhes was expressed at very high levels in brain andat low levels in kidney, thyroid, lung, heart, and testis (Fig.1B). No Rhes mRNA was detected in liver. A low level of Rhesexpression was also detected in the rat thyroid cell line FRTL-5and in the mouse ES cell line R1 (44) (Fig. 1B). Since R1 cellswere grown on mouse embryonic fibroblasts (MEFs), we also analyzedRhes expression in MEF cells. The absence of the hybridizationband in total RNA from MEF cells demonstrated bona fide Rhesexpression in ES cells and not in the MEF layer.

View larger version (61K):
[in this window]
[in a new window]
FIG. 1. Rhes expression profile during embryo development and in the adult mouse. (A) (Top panel) RT-PCR analysis performed with total RNAs prepared from E12.5 (a), E13.5 (b), E14.5 (c), E15.5 (d), E16.5 (e), E17.5 (f), and E18.5 (g) embryo brains. Total RNA from adult brain (h) was used as a positive control for PCR. The oligonucleotides were designed according to the sequence of the mouse Rhes gene and amplified a 403-bp fragment of the Rhes 3' untranslated region. (Middle panel) Specificity of PCR products checked by hybridization with a 207-bp PstI-NcoI fragment derived from the Rhes sequence between the PCR primers. (Bottom panel) G3PDH gene amplification performed as an internal control for RT-PCR. (B) (Upper panel) Northern blot performed with total RNAs prepared from mouse adult tissues (brain [a], liver [b], kidney [c], thyroid [d], lung [e], heart [f], and testis [g]) and TL5 (h), R1 (i), and MEF (l) cells. The probe used was the 372-bp AflII-PvuII fragment of the mouse Rhes genomic locus, located 42 bp downstream of the translational stop codon. The lower panel shows the 28S and 18S rRNAs of each sample.

To obtain further insights into the distribution of Rhes mRNAduring embryogenesis and in the adult brain, we performed insitu hybridization experiments (Fig. 2). Rhes mRNA was prominentlyexpressed in the striatum, but it was also present in the accumbensnucleus (ventral part of striatum), in the olfactory tubercle,in the piriform cortex, and in the hippocampus dentate gyrus.No signal was detected with a Rhes sense riboprobe. Surprisingly,no signal was detected by in situ hybridization in the brainof E15.5 and E17.5 embryos, suggesting that, during embryogenesis,Rhes mRNA levels are below the limit of detection of our insitu hybridization technique.

View larger version (122K):
[in this window]
[in a new window]
FIG. 2. Rhes expression in adult brain. Sagittal sections of CD1 adult mouse brain: striatum (A, B, C, and D), accumbens nucleus (E, F, G, and H), olfactory tubercle and piriform cortex (I, L, M, and N), and hippocampus (O, P, Q, and R). Sections A, E, I, and O were stained with hematoxylin and eosin; sections B, C, F, G, L, M, P, and Q were hybridized to the Rhes antisense riboprobe; sections D, H, N, and R were hybridized to the Rhes sense riboprobe. Sections C, G, M, and Q was higher magnifications of sections B, F, L, and P, respectively.

Targeting of the mouse Rhes locus. In order to inactivate the Rhes locus, we constructed a targetingvector (Fig. 3A). The 5' homology arm was mutagenized to inserta stop codon in each frame immediately downstream of the Rhestranslational start codon. The targeting vector contained theinternal ribosome entry site-EGFP cassette and PGKneo cassetteflanked by loxP sequences between the homology arms (6). Furthermore,the construct was flanked by a cassette for the herpes simplexvirus thymidine kinase gene for negative selection. The targetingvector was designed to abolish the synthesis of Rhes and toplace the EGFP coding sequence downstream of the Rhes regulatorysequences (Fig. 3A). The targeting construct was transfectedinto RI ES cells, and the recombinant clones were identifiedby Southern blot analysis (Fig. 3B). Chimeric mice were generatedwith five independently targeted ES cell clones. In the progenyof one chimera, we obtained 48 ES cell-derived animals (agouti),of which 27 carried the wild-type Rhes allele and the remaindercarried the targeted Rhes allele, consistent with a Mendelianpattern of transmission.

Rhes protein is not essential for normal embryo development. To assess whether Rhes function is essential for normal development,we interbred Rhes^+/– mice. The heterozygous mice werefertile; each mating was productive, and the litter sizes wereindistinguishable from those obtained with wild-type mice. Thegenotypes of the newborns were assessed by Southern blot analysis(Fig. 4). Of 117 mice examined, 28 were wild type, 67 were heterozygous,and 22 were homozygous for the mutant allele. The ratio of thethree classes of animals was not significantly different fromthe expected values for normal transmission of the wild-typeand mutant alleles. The Rhes^–/– mice were viable,showing that the Rhes gene product does not play a vital role,at least in the CD1 background. Furthermore, mice homozygousfor the Rhes knockout allele mated and were fertile, and thelitter sizes were indistinguishable from those of wild-typeand heterozygous matings.

In situ hybridization and Western blot analysis on knockout mouse brain. To determine Rhes expression in Rhes^–/– mice, weperformed in situ hybridization experiments and Western blotanalyses. Since the main Rhes expression site was the striatum,we focused on this region of the brain. In situ hybridizationshowed that Rhes mRNA was barely detectable in the striatalregion of the knockout mouse brain (Fig. 5), while it was clearlydetected in wild-type animals. We also carried out an in situhybridization experiments with an antisense riboprobe of EGFPbecause, in the targeting vector, the EGFP gene was controlledby the Rhes promoter. Figure 5 shows that EGFP mRNA was detectedin the striatum of the Rhes^–/– mouse. No signalwas detected with the sense riboprobes for either Rhes or EGFP.

View larger version (89K):
[in this window]
[in a new window]
FIG. 5. Rhes and EGFP expression in knockout mouse brain. The in situ hybridization experiments were carried out on sagittal sections of adult mouse brain from CD1 (wild type [WT]) (sections A, B, C, and D) and Rhes^–/– (knockout [KO]) (sections E, F, G, H, I, L, M, and N) mice. Sections A, E, and I were stained with hematoxylin and eosin; sections B, C, F, and G were hybridized to the Rhes antisense riboprobe; sections D and H were hybridized to the Rhes sense riboprobe; sections L and M were hybridized to the EGFP antisense riboprobe; section N was hybridized to the EGFP sense riboprobe. Sections C, G, and M are higher magnifications of sections B, F, and L, respectively.

To explore the presence of the Rhes protein, we prepared a rabbitanti-Rhes polyclonal antibody. We proved the efficacy and thespecificity of the antibody by Western blotting on HeLa cellextracts transfected with either a Rhes expression vector oran empty vector. Rhes protein was only detected in the cellstransfected with the Rhes expression vector (Fig. 6, lane c)while no cross-reacting proteins were seen in the control extracts(Fig. 6, lane d). However, Western blotting carried out on striatumprotein extracts with the same antibody revealed the presenceof protein bands that were present both in wild-type (Fig. 6,lane b) and knockout (Fig. 6, lane a) animals that we interpretedas nonspecific. Three bands were present only in the wild-typemice and were completely absent in knockout extracts. The lowerband displayed a mobility very similar to that of Rhes expressedin HeLa cells. The slight difference could be due to differentposttranslational modifications happening in the in vitro andin vivo models. The upper bands identified proteins of highermolecular mass. Interestingly, the Rhes locus shows severaltranscripts, one of which (accession number BC036988) encodesa protein containing Rhes at the C terminus and extending 48additional amino acids at the N terminus without an initiatormethionine, suggesting that the actual protein could be longer,like the one(s) that we detected in the Western blot.

View larger version (51K):
[in this window]
[in a new window]
FIG. 6. Expression of Rhes protein in wild-type and Rhes^–/– striatum. Rhes^–/– (a) and wild-type (b) striatum protein extracts (100 µg of total protein/lane) were analyzed by Western blotting with anti-Rhes polyclonal antibody; 0.4 µg of total protein from HeLa cells transfected with a Rhes construct (c) was used as the positive control; and 0.4 µg of total protein from HeLa cells transfected with the empty vector (d) was used as the negative control. Arrows indicate proteins present in wild-type and absent in knockout extracts.

Taking together the RNA and protein data on Rhes expression,we conclude that the targeted Rhes allele presented in thispaper does not produce any detectable Rhes protein. We havealso presented evidence of a novel Rhes-related protein whosepresence was also abolished in the knockout mouse that we generated.

Rhes affects body weight. We observed that knockout mice weighed less than wild-type mice,as clearly shown by a significant main effect of genotype [F_(1,16) = 9.247 and P = 0.0078; see Fig. 8]. Furthermore, a genotype-ageinteraction [F_{(2, 32)} = 3.487 and P = 0.0427] but not a genotype-sexinteraction was found. At each age analyzed (postnatal days70, 100, and 130), a marked weight difference between wild-typeand knockout mice in both males and females was revealed bypost hoc analysis (P < 0.01; Fig. 7). Since Rhes is expressedin the thyroid gland and in pancreatic ß-cells (14bis),we tested the thyroid-stimulating hormone, glucose, and amylaselevels in age-matched homozygous knockout and wild-type mice.No statistically significant differences were found (data notshown).

View larger version (15K):
[in this window]
[in a new window]
FIG. 8. Locomotor activity in the open-field test. The distance traveled by wild-type and knockout mice is shown. Data are means ± standard error of the mean. Data for males and females were pooled (n = 10). The asterisk indicates a significant difference between wild-type and knockout mice (P < 0.05).

View larger version (21K):
[in this window]
[in a new window]
FIG. 7. Body weights of wild-type and knockout mice measured on postnatal days 70, 100, and 130. Data are means ± standard error of the mean, n = 5. The double asterisk indicates a significant difference between wild-type and knockout mice (P < 0.01).

Behavioral analysis of Rhes^–/– mice. (i) Passive avoidance and Morris water maze tests. In both the passive avoidance and Morris water maze tests, nosignificant difference between wild-type and knockout mice wasfound (data not shown).

(ii) Open-field test. In the first 15 min of the open-field test test, no main effectof genotype was found in distance moved, and the genotype x5-min block interaction missed statistical significance [F_(2,36) = 2.704 and P = 0.0805; Fig. 8]. However, post hoc comparisonrevealed that knockout mice had an altered locomotor profile,moving less than wild-type mice during the first 5 min (posthoc, P < 0.05; Fig. 8).

(iii) Elevated plus-maze test. In agreement with previous studies (26, 51), wild-type malesdisplayed higher anxiety levels than wild-type females in theelevated plus-maze test. This profile was found inverted inknockout mice, females being more anxious than males, as illustratedby a significant genotype x sex interaction observed in thefollowing measures (Fig. 9): percent of open arm entries [F_(1,16) = 7.011 and P = 0.0175], number of open arm entries [F_(1,16) = 4.438 and P = 0.0500], percent of time spent in the openarms [F_{(1, 16)} = 5.059 and P = 0.0389], and percent of timespent in the closed arms [F_{(1, 16)} = 4.878 and P = 0.0421].In the last measure, a clear difference between wild-type andknockout females was evident (post hoc, P < 0.05), indicatingabnormally high anxiety levels in knockout females comparedto wild-type females. Other measures confirmed increased anxietylevels in knockout mice, especially in knockout females (Fig.9). A significant genotype-sex interaction was observed in percentof head-dipping [F_{(1, 16)} = 6.347 and P = 0.0228] and percentof stretch attend postures [F_{(1, 16)} = 5.998 and P = 0.0262],both performed in the protected area. Moreover, knockout miceshowed less locomotion than wild-type mice, spending a longtime in immobility [F_{(1, 16)} = 4.336 and P = 0.0500]. It isworth noting that four knockout mice (two males and two females)out of 10 fell from the plus-maze during the test, while noneof the wild-type mice did (P < 0.05; Fisher test).

View larger version (18K):
[in this window]
[in a new window]
FIG. 9. Behavioral endpoints in the plus-maze test. The number of total and closed arm entries, the percentage of open arm entries, the percentage of closed arm time, the percentage of protected head dipping and stretch attend postures, and immobility shown by wild-type and knockout mice are indicated. Data are means ± standard error of the mean (n = 5). For immobility, data for males and females were pooled (n = 10). The asterisk indicates a significant difference between wild-type and knockout mice (P < 0.05).

(iv) Rota-rod test. In the rota-rod test, though both genotypes displayed a similarover-trial profile, decreasing their tendencies to fall fromthe mast with the increase in speed (Friedman ${chi}$ ² in the overallgroup = 17.73, P = 0.0014), knockout mice always had worse performancesthan wild-type mice (main effect of genotype; Mann-Whitney U= 81.5, P = 0.0172) (Fig. 10). Specifically, knockout mice fellwithin significantly shorter times at the two fastest speeds(post hoc, P < 0.01).

View larger version (13K):
[in this window]
[in a new window]
FIG. 10. Balance and motor coordination in the rota-rod test. The time before falling at each speed level shown by wild-type and knockout mice is indicated. Data are means ± standard error of the mean. Data for males and females were pooled (n = 10). The double asterisk indicates a significant difference between wild-type and knockout mice (P < 0.01).

DISCUSSION

Top

Discussion

The mechanism responsible for impairment of brain function inthyroid hormone deficiencies is not well understood. It is knownthat several genes in the brain are under thyroid hormone control,but it is not clear which gene(s), when deregulated, is responsiblefor the phenotype observed in thyroid hormone deficiencies orhow it does so. In this study we focused on the role of Rhes,a gene under thyroid hormone control that is expressed in thestriatum. Rhes is a member of a new subfamily of Ras-relatedsmall GTP-binding proteins recently identified. To this end,we generated mice homozygous for the Rhes null mutation. Wechecked the absence of Rhes protein in the knockout mouse striatumby Western blot analysis. This analysis also showed the absenceof another protein which probably originates from the Rhes locusby alternative transcription initiation and splicing.

The knockout mice were viable, and their general condition didnot reveal gross abnormalities with the exception of a reducedbody weight. The macroscopic analysis of adult tissues whichexpress Rhes did not show any gross abnormalities between knockoutand wild-type mice. The knockout animals mated and were fertile,and the litter sizes were indistinguishable from those of wild-typematings. No alteration of thyroid and pancreatic gland functionswas observed, even though Rhes is expressed in the wild-typeglands. Given that the absence of thyroid hormone determinessevere damage to the developing striatum in neurological cretinism(8, 9, 39), we analyzed in detail the behavioral features ofRhes knockout mice. These mice showed a significant decreasein locomotor activity compared to wild-type mice. Interestingly,Rhes deletion influenced the anxiety response in the plus-mazetest in a gender-specific manner. The impact of gender on theanxiety test has been widely studied in mice, and in the plus-mazetest, females generally show lower anxiety levels than males(35, 48). In the present study, wild-type mice behaved as expected,while an opposite trend has been found in Rhes^–/–mice, females showing higher anxiety levels than males in mostof the endpoints considered.

The main behavioral effect of Rhes deletion was a marked impairmentin motor coordination. In particular, knockout mice showed aclear impairment in the rota-rod test. This task has been provento be very sensitive to striatum integrity (11, 25, 56) andhas also been used to detect the progressive decline of striatalfunction in R6/2 Huntington gene transgenic mice (27, 31). Thus,the motor coordination impairment shown by Rhes knockout micein the rota-rod test, confirmed by the number of falls in theplus-maze, is strongly concordant with the main striatal localizationof the Rhes protein (24, 59).

The striatum is reportedly involved in cognitive abilities,from motor planning to reward seeking and procedural learning(4, 32). Consistent with the role played by the striatum inmotor activity and learning processes, these abilities are dramaticallyimpaired in advanced Parkinson's disease (47). Rhes knockoutmice showed no learning or memory impairment in the water mazeand passive avoidance tests, suggesting that Rhes protein maybe involved only in selected striatal processes not influencinglearning and memory. This finding indicates that mental retardationlinked to hypothyroidism may be independent of alteration inRhes levels or function.

In conclusion, the modest behavioral deficits of Rhes knockoutmice indicate that Rhes is involved in selected striatal competencies,mainly locomotor activity and motor coordination, suggestingthat its downregulation in hypothyroidism could be responsibleonly for a subset of symptoms, such as the striatopallidal syndrome(8, 9, 39).

ACKNOWLEDGMENTS

We thank Mario De Felice for advice in generation of knockoutmice and Giovanni Maraviglia for technical assistance. We alsothank Tommaso Russo for glucose and amylase determination andAndreas Nagy for providing the R1 ES cell line.

This work was supported in part by Telethon grant GP0208Y01,by a grant from the Associazione Italiana per la Ricerca sulCancro (to R.D.L.), by Ministero dell'Università e dellaRicerca Scientifica e Tecnologica grant "I geni dell'uomo" cluster01, and by Italian Ministry of Health project ALZ1 (to E.A.).A.R., M.T.P., P.M., and A.A. were supported by Biogem s.c.a.r.l.,Italy.

FOOTNOTES

* Corresponding author. Mailing address: Stazione Zoologica Anton Dohrn, Villa Comunale, 80121 Naples, Italy. Phone: 39 081 5833278. Fax: 39 081 5833258. E-mail: rdilauro{at}unina.it.

REFERENCES

Top