Evidence for abnormal early development in a mouse model of Rett syndrome

Authors


*P. Maciel, Life and Health Sciences Research Institute (ICVS), School of Health Sciences, University of Minho, Campus de Gualtar, 4710-057 Braga, Portugal. E-mail: pmaciel@ecsaude.uminho.pt

Abstract

Rett syndrome (RTT) is a neurodevelopmental disorder that affects mainly females, associated in most cases to mutations in the MECP2 gene. After an apparently normal prenatal and perinatal period, patients display an arrest in growth and in psychomotor development, with autistic behaviour, hand stereotypies and mental retardation. Despite this classical description, researchers always questioned whether RTT patients did have subtle manifestations soon after birth. This issue was recently brought to light by several studies using different approaches that revealed abnormalities in the early development of RTT patients. Our hypothesis was that, in the mouse models of RTT as in patients, early neurodevelopment might be abnormal, but in a subtle manner, given the first descriptions of these models as initially normal. To address this issue, we performed a postnatal neurodevelopmental study in the Mecp2tm1.1Bird mouse. These animals are born healthy, and overt symptoms start to establish a few weeks later, including features of neurological disorder (tremors, hind limb clasping, weight loss). Different maturational parameters and neurological reflexes were analysed in the pre-weaning period in the Mecp2-mutant mice and compared to wild-type littermate controls. We found subtle but significant sex-dependent differences between mutant and wild-type animals, namely a delay in the acquisition of the surface and postural reflexes, and impaired growth maturation. The mutant animals also show altered negative geotaxis and wire suspension behaviours, which may be early manifestations of later neurological symptoms. In the post-weaning period the juvenile mice presented hypoactivity that was probably the result of motor impairments. The early anomalies identified in this model of RTT mimic the early motor abnormalities reported in the RTT patients, making this a good model for the study of the early disease process.

Rett syndrome (RTT) is a major cause of mental retardation in females, affecting 1 per 10 000 to 1 per 22 000 females born (Percy 2002). The ‘classic’ progression of RTT has four stages (Kerr & Engerstrom 2001). Stage I is characterized by an apparently normal development with uneventful prenatal and perinatal periods; in this stage (around 6–18 months) some of the patients learn some words and some are able to walk and feed themselves. In stage II (regression) a deceleration/arrest in the psychomotor development is noticed, with loss of stage I acquired skills, establishment of autistic behaviour and signs of intellectual dysfunction; the hands’ skilful abilities are replaced by stereotypical hand movements, a hallmark of RTT. The pre-school/school years correspond to stage III (pseudo-stationary) and here some improvement can be appreciated, with recovery of previously acquired skills. This is followed by the progressively incapacitating stage IV that can last for years (Hagberg et al. 2002); at this final stage patients develop trunk and gait ataxia, dystonia, autonomic dysfunction (breathing anomalies, sleep and gastrointestinal disturbances) and many of them have a sudden unexplained death in adulthood.

In spite of the classic RTT description, some researchers have questioned whether RTT patients display subtle signs of abnormal development soon after birth (Engerstrom 1992; Kerr 1995; Naidu 1997; Nomura & Segawa 1990). Huppke and colleagues reported on a sample of RTT patients who presented a significantly reduced occipito-frontal circumference, shorter length and lower weight at birth (Huppke et al. 2003). This hypothesis has recently been confirmed by the work of Einspieler and colleagues (Einspieler et al. 2005b), who analysed video records of the first 6 months of life of 22 RTT patients and were able to notice abnormalities in several behaviours. All RTT patients presented an abnormal pattern of spontaneous movements within the first 4 weeks of life, with abnormal ‘fidgety’ movements that were considered a sign of abnormal development (Einspieler et al. 2005a,b). Such abnormal movements were ascribed to problems in the central pattern generators in the brain (Einspieler et al. 2005a; Einspieler & Prechtl 2005). In a different study, midwives and health visitors blinded for the clinical status of the children were able to identify in family videos potential anomalies in the early development of RTT patients, particularly anomalies in physical appearance and hand posture, as well as body movements and postures (Burford 2005). Segawa, in a retrospective study of patients’ clinical files (Segawa 2005), also reported altered presentation of several motor milestones.

Most patients with classic RTT are heterozygous for mutations in the X-linked methyl-CpG binding protein gene (MECP2) (Amir et al. 1999), which encodes the methyl-CpG binding protein, MeCP2; this is known to bind symmetrically methylated CpG dinucleotides, and to recruit the co-repressors Sin3 yeast homologue A and histone deacetylase 1 and histone deacetylase 2 to repress transcription (Jones et al. 1998). When mutated, MeCP2 does not bind or binds ineffectively to its targets and, as a consequence, deregulation of transcription is thought to occur. Animal models of RTT were created in mice, mimicking several motor aspects of RTT and even the more emotional and social aspects of the syndrome (Chen et al. 2001; Guy et al. 2001; Shahbazian et al. 2002). The mutants are born normal and a few weeks later start to present a progressive motor deterioration, despite no gross abnormalities in the brain being noticed. Males carrying the mutation in hemizygosity display an earlier onset and are more severely affected than heterozygous females, probably as the result of X-chromosome inactivation that makes these females mosaics for the expression of the mutation, as is the case for the human condition.

The study presented here was performed using the Mecp2tm1.1Bird (Guy et al. 2001) mouse as a model. These mice were described as presenting no initial phenotype. Male Mecp2tm1.1Bird null animals begin to show symptoms at 3–8 weeks whereas heterozygous female animals manifest the disease at 3 months of age. The phenotype of these animals mimics many of the motor symptoms of RTT: stiff and unco-ordinated gait, reduced spontaneous movement, hind limb clasping, tremor and irregular breathing. Pathologically, no obvious histological abnormalities were detected in peripheral organs or in the brain. However, more recently, Kishi and Macklis reported that in the Mecp2-null mice the neocortical projection layers were thinner and the pyramidal neurons in layer II/ III had smaller somas and less complex dendritic trees in symptomatic animals than in wild-type mice (Kishi & Macklis 2004). Another study in this animal model suggested an essential role of MeCP2 in the mechanisms of synaptic plasticity (LTP and LTD) in the mature hippocampal neurons (Asaka et al. 2005).

The goal of this study was to determine whether the early neurodevelopmental process was altered in the absence of MeCP2 in mice. We assessed the achievement of milestones, considering different maturational and physical growth measures and neurological reflexes, two of the most well-known and most used neurobehavioural testing categories to address neurological disorder (Spear 1990), in the Mecp2tm1.1Bird mouse model of RTT (Mecp2-null males and Mecp2-heterozygous females). We identified an altered developmental progression of the mutant animals since the first postnatal week, in spite of their apparently normal phenotype. The differences seen suggest the presence of mild neurological deficits already at this age; the animals also presented significantly reduced activity, probably as a result of motor impairments early in life. The abnormal achievement of the developmental hallmarks, although transient, could reflect abnormalities that are likely to impact the development of more mature behaviours.

Materials and Methods

Animals

The strain used in this study was created by the Bird laboratory by transfecting the targeting vector in 129P2/OlaHsd E14TG2a embryonic stem cells and injecting these into C57BL/6 blastocysts (Guy et al. 2001). According to information from the Jackson Laboratory, from whom we acquired the animals, the original strain was bred to C57BL/6 mice and backcrossed to C57BL/6 at least five times. Female Mecp2tm1.1Bird mice were bred with C57BL/6 wild-type (wt) male mice, to obtain wt and Mecp2-mutant animals. Mice were kept in an animal facility in a 12-hour light: 12-hour dark cycle, with food and water available ad libitum. A daily inspection for the presence of new litters in the cages was carried out twice a day and the day a litter was first observed was scored as day 0 for that litter. After birth, animals were kept untouched in the home cage with their heterozygous mothers until postnatal day (PND) 3, and at PND4 animals were tagged in their feet or the tip of the ears. Neurodevelopmental evaluation tests were started at PND4 and performed daily through to PND21. Weaning was performed at 22/23 days of age. Males and females were separated and kept in independent cages, in groups of three to seven animals per cage. At weaning the tip of the tail of the mice was cut for DNA extraction by Puregene DNA isolation kit (Gentra, Minneapolis, MN) and genotyping was performed according to the protocol supplied for this strain by the Jackson Laboratory. At the fourth postnatal week animals were tested for spontaneous activity in the Open-field (OF) apparatus and the day after this animals were tested for anxiety-like behaviour in the Elevated plus-maze (EPM) apparatus. At the fifth postnatal week animals were tested in the rotarod apparatus. After completing the experiment animals were rapidly killed by decapitation, thus minimizing their suffering (in accordance with the European Communities Council Directive, 86/609/EEC).

The same observer, who was blinded for the genotype of the animals and for the performance of the animals on the previous day, evaluated all the described tests. Tests were always performed in the same circadian period (between 1100 and 1800 h) and whenever possible at the same hour of the day. All the animals were separated from their parents at the beginning of each test session and kept with their littermates in a new cage, with towel paper and sawdust from their home cage. Once the test sessions finished for all the members of a litter, the animals were returned to their home cage. Table 1 shows attributable scores for each test. Throughout the text when Mecp2-heterozygous animals are referred to they are always females and Mecp2-null animals is always used to refer to male animals. All the controls used were littermates of the Mecp2-mutant (male and female) animals.

Table 1.  Attributable scores in milestones performance of Mecp2-mutant and wild-type animals
 Score
 0123
Ear openingclosedopen 
Eye openingboth closedone openboth open 
Surface righting reflexstays in dorsal positionfights to uprightrights itself 
Postural reflexnot presentpresent 
Negative geotaxisturns and climbs gridturns and freezesmoves but fails to turndoes not move
Wire suspensionnot presentpresent 

Pre-weaning behaviour

Maturation measures

Body weight The body weight of mice was registered every day from PND4 through to PND21 (weight ± 0.01 g).

Anogenital distance (AGD) The distance between the opening of the anus and the opening of the genitalia was registered (distance ± 0.5 mm).

Ear opening The day when an opening in the ear was visualized was registered.

Eye opening We registered the state of the eyes from the day when animals started to open the eyes until the day when every animal in the litter had both eyes opened. An eye was considered open when any visible break in the membrane was noticed.

Developmental measures

Surface righting reflex (RR) Mice were restrained on their back on a table and then released. The performance of the animal (to turn or not) was scored and the time taken to surface-right, in a maximum of 30 seconds, in three consecutive trials, was registered. To determine the score for each day, the median value was calculated for the three trials.

Postural reflex (PR) Animals were put in a small box and shaken up and down and left and right. Existence of an appropriate response (animals splaying their four feet) was scored.

Negative geotaxis (NG) Animals were put in a horizontal grid and then the grid was turned through 45° so that the animal was facing down. The behaviour of the animal was observed for 30 seconds and registered as shown in Table 1.

Wire suspension (WS) The animals were forced to grasp a 3-mm wire and hang from it on their forepaws. The ability of the animals to grasp the wire was scored and the time for which they held the wire (maximum 30 seconds) was registered.

Post-weaning behavioural tests

Open field

Animals were placed in the centre of a 43.2 × 43.2-cm arena with transparent walls (MedAssociates Inc., St Albans, VT) and their behaviour was observed for 5 min. Activity parameters were collected (total distance travelled, speed, resting time and the distance travelled and time spent in the predefined centre of the arena versus the rest of the arena). The number of rears, the time that animals spent exploring vertically and the number of bolus faecalis were also registered by observation.

Elevated plus maze

Animals were placed in an EPM apparatus consisting of two opposite open arms (50.8 × 10.2 cm) and two opposite closed arms (50.8 × 10.2 × 40.6 cm) raised 72.4 cm above the floor (ENV-560, MedAssociates Inc.) and behaviour (number of entries in each arm and the time spent in each of the arms) was registered for 5 minutes.

Rotarod

Mice were tested in a rotarod (TSE systems, Bad Hamburg, Germany) apparatus to evaluate their motor performance. The protocol consisted of 3 days of training at a constant speed (15 r.p.m.) for a maximum of 60 seconds in four trials, with a 10-min interval between each trial. At the fourth day, animals were tested for each of six different velocities (5 r.p.m., 8 r.p.m., 15 r.p.m., 20 r.p.m., 24 r.p.m. and 31 r.p.m.) for a maximum of 60 seconds in two trials, with a 10-min interval between each trial. The latency to fall off the rod was registered.

Statistical analysis

In the pre-weaning behaviour analysis, because there were problems with achieving the assumptions required for repeated measures testing, such as sphericity and homogeneity of variances, using the data obtained, we used regression methods to compare the performance between Mecp2-mutant and wt littermate control mice. To do this, variables scored 0 or 1 were analysed by logistic regression [Score = f(day, genotype sex)]. For continuous variables, a linear or a quadratic regression was applied. Interaction between the independent variables (day, genotype and sex) was also studied and reported when it was observed. The surface righting reflex and wire suspension times were analysed as survival times through the Kaplan–Meier test. The Negative Geotaxis was analysed (classification in three classes) by a χ2 test and the percentage of animals meeting the criterion (score = 0) by linear regression was found. In the post-weaning behaviour tests, data were analysed using Student’s t-test. A critical value for significance of P < 0.05 was used throughout the study.

Results

Pre-weaning behaviour analysis

In this and in all other variables under study we always analysed male and female animals separately. The number of animals used in the analysis of maturation markers and neurological reflexes in the pre-weaning period was: Mecp2-null n = 13, wt littermate males n = 11, Mecp2-heterozygotes n = 16, wt littermate females n = 9.

Physical growth and maturation

Body weight

We weighed Mecp2-mutant and wt littermate control mice everyday from PND4 to PND21 and analysed the data with a quadratic regression. As expected, the body weight was statistically different between male and female animals, with female mice being heavier than male mice (P = 0.013), and the day of analysis had a significant influence on the body weight (P < 0.001). When we analysed the influence of the Mecp2 genotype of mice in the body weight, we noticed that the body weight evolution of Mecp2-null mice was not different from that of the wt littermate controls, in the first 21 days of postnatal development (P = 0.156). Surprisingly, however, Mecp2-heterozygous mice presented a significantly reduced body weight when compared to their wt littermate controls (P < 0.001) (Fig. 1a,b). The effect of genotype was not seen from the beginning of the study, but from around PND10 onwards.

Figure 1.

Physical growth and maturation parameters of the Mecp2-heterozygous female mice and the Mecp2-null male mice during the pre-weaning period. (a,b) Body weight evolution from PND4 to PND21 of Mecp2-mutant animals and their wt littermate controls. Mecp2-heterozygous females had a significant reduction in body weight that started to be noticeable after PND10 (P < 0.001). (c,d) Anogenital distance measurement from PND4 through PND21 of Mecp2-mutant animals and their wt littermate controls. Mecp2-mutant mice presented a significant reduction in the AGD (P < 0.001). (Mecp2-heterozygous females, n = 16; wt females, n = 9; Mecp2-null males, n = 13 and wt males, n = 11. Values are mean ± SEM. AGD, anogenital distance; PND, postnatal day; ko, knock-out; wt, wild-type; *P < 0.05).

Ear and eye opening

We observed mice daily from PND4 and registered the day when at least one eye was open and the day when both eyes were open. The day an aperture was seen in the ear was also registered. No differences existed between genotypes or gender regarding the mean day of aperture of eyes and ears (supplemental table 1).

Anogenital distance

We took this measure from PND4 to PND21 in all mice and analysed the data using a linear regression method. As body weight might influence the anus–genitalia distance, previous studies (Degen et al. 2005) introduced a correction: the AGD value was divided by the weight of each animal at each postnatal day (AGD/weight). We calculated the coefficient of correlation between the AGD and the body weight of the mice (R = 0.907 for male mice and R = 0.917 for female mice) and because our findings suggested that these two variables were highly associated we decided not to use this correction.

The AGD of male mice was higher than that of female mice (P < 0.001), as expected, and the day of testing affected this distance, which was higher the later the measure was taken (P < 0.001). We found that male and female Mecp2-mutant animals presented a statistically significant reduction in the AGD throughout the pre-weaning period, when compared to their respective wt controls (P < 0.001) (Fig. 1c,d).

Neurological reflexes

Surface righting reflex

No differences between sexes were found in the acquisition of this reflex (P = 0.668), and the animals’ ability to regain an upright position improved with age (P < 0.009), as expected. Mecp2-mutant animals did not present differences in the age of acquisition of this reflex (P = 0.534 and P = 0.161 for Mecp2-null and Mecp2-heterozygous mice, respectively) (supplemental Fig. 1a,b). When we considered the time these animals took to surface-right, Mecp2-heterozygous mice presented statistically significant differences, with mutant females taking longer than wt littermates to regain an upright position (P = 0.031) (Fig. 2a,b). Nevertheless, when Mecp2-null mice and wt controls were compared no differences were found. There were no differences, in this last parameter, between sexes (P = 0.216).

Figure 2.

Abnormalities in milestone achievement in the Mecp2-heterozygous and the Mecp2-null mice during the pre-weaning period. (a,b) Time taken to surface-right in the surface righting reflex test. Female Mecp2-heterozygous mice took longer to regain an upright position than their wt littermates (P < 0.05). (c,d) Percentage of animals presenting the postural reflex between PND9 and PND17. A delay in the acquisition of this parameter was observed in both the Mecp2-null animals (P < 0.001) and the Mecp2-heteroygous females (P = 0.006). (e,f) Percentage of animals presenting the negative geotaxis reflex. Female Mecp2-heterozygous animals (P = 0.002) and Mecp2-null males (P < 0.001) showed a worse performance than wt littermates. (g,h) Time that animals held the wire in the wire suspension reflex (in a 30-second test). Mecp2-null male animals held the wire for longer (P = 0.010), although differences in Mecp2-heterozygous females did not reach significance. (Mecp2-heterozygous females, n = 16; wt females, n = 9; Mecp2-null males, n = 13 and wt males, n = 11. Values are mean ± SEM. PND, postnatal day; ko, knock-out; wt, wild-type, *P < 0.05).

Postural reflex

There were no differences between genders in the ontogeny of this reflex (P = 0.118) and, as expected, the day affected its establishment (P < 0.001). The pattern of acquisition of the PR was statistically different between Mecp2-null (P < 0.001) and Mecp2-heterozygous (P = 0.006) mice, when compared to their respective wt controls, with a worse outcome for mutant animals. Both Mecp2-null and Mecp2-heterozygous mice showed a delay in the acquisition of the PR reflex (Fig. 2c,d). The acquisition of the PR by wt animals started at PND9 for females and PND10 for males and at PND16 all wt animals presented the PR. In the mutant mice the reflex first appeared on PND11 for females and PND12 for males and only at PND17 did all mutant animals present the PR. The Mecp2-mutant animals showed a delay of 2 days in relation to the day of first appearance of PR in the wt animals.

Negative geotaxis

In respect to mouse behaviour, this reflex was scored from 0 to 3 (see Table 1). Scores 2 and 3 were not frequent and so, to simplify the analysis of the data, we decided to recode the behaviours for the analysis. Score 0 and score 1 were maintained and score 2 was changed to include the previous scores 2 and 3. In this task, both male and female Mecp2-mutant mice had a worse performance than their respective wt littermate controls (Fig. 2e,f). The percentage of animals meeting the criterion for a score of 0 was dependent on the day (P < 0.01) and genotype (P < 0.01), whereas sex was not significant (P = 0.07). Moreover, differences were found in the acquisition of the NG reflex between genotypes in both sexes (in both cases P < 0.01), resulting from a difference in the performance of the animals in classes 0 and 2. When we tested the animals in a weaker version of this test (at 30° inclination), Mecp2-null animals still performed worse than wt controls in this task whereas heterozygous females did not differ significantly from wt animals (data not shown).

Wire suspension

There were no differences in the establishment of this reflex between male and female mice (P = 0.176) and the day affected the establishment of the reflex (P < 0.001), as expected. The performance of Mecp2-null and Mecp2-heterozygous mice and their respective wt controls in the acquisition of the reflex (animals grasp the wire or do not grasp) was similar, with no statistical differences when compared among each other (P = 0.605 for males and P = 0.214 for females). This reflex was acquired between PND11 and PND18 for both Mecp2-mutant and wt mice of both genders. Another parameter that was taken from this analysis was the wire suspension time. As body weight might influence the time animals hold on to the wire, the curves of the wire suspension holding time were corrected taking into account the body weight. We analysed this parameter from PND15 onwards because from this day more than 50% of the animals held on to the wire for more than 1 second. The wt females held the wire for a significantly longer time than wt male mice (P = 0.046), but there were no differences between mutant male and female mice (P = 0.730). Surprisingly, Mecp2-null and Mecp2-heterozygous mice stayed on the wire longer than their respective wt littermate controls and the differences were statistically significant between Mecp2-null and wt littermate controls (P < 0.001) (Fig. 2g,h). Even when we analysed the data relative to all days (PND11–PND21), the same conclusions were reached (P = 0.010).

Post-weaning behaviour analysis

Exploratory activity

At the fourth week of age, animals were tested in the OF apparatus, to evaluate their spontaneous activity, for a period of 5 min (Mecp2-null n = 14, wt littermate males n = 16, Mecp2-heterozygous n = 12, wt littermate females n = 10). Globally, no differences were found between Mecp2-mutant and wt animals in the time they spent and distance they travelled in the centre of the arena in relation to the total area of the arena, in the time animals spent exploring vertically or in the number of rears (Supplemental table 2). We found that Mecp2-null animals travelled a smaller total distance (P = 0.049) at a lower speed (P = 0.000) than wt controls (Fig. 3a–c). Null animals produced a significantly higher number of bolus faecalis (P = 0.031) (Supplemental table 2), which could be a consequence of their neuroautonomic disorder.

Figure 3.

Mecp2-mutant female and male mice present reduced spontaneous activity without altered exploratory capacity at 4 weeks of age, in the open-field paradigm. (a) Mecp2-null male mice travelled a smaller total distance (P = 0.049), (b) at a lower speed (P = 0.000) than their respective wt littermate controls. Female heterozygous animals did not present differences in any of the parameters analysed. (Mecp2-heterozygous females, n = 12; wt females, n = 10; Mecp2-null males, n = 14 and wt males, n = 16. Values are mean ± SEM. PND, postnatal day; ko, knock-out; wt, wild-type, *P < 0.05).

Anxiety-like behaviour

The day after OF testing, animals were tested in the EPM apparatus, in a 5-min session (Mecp2-null n = 13, wt littermate males n = 13, Mecp2-heterozygous n = 11, wt littermate females n = 8). There were no differences between Mecp2-mutant animals and wt controls in the percentage of time animals spent in the open arms nor in the percentage of entries in the open arms in relation to total arms entries, but Mecp2-null animals presented a smaller number of closed arms entries (P = 0.014) (Fig. 4a–c).

Figure 4.

Mecp2-mutant female and male mice do not present anxiety-like behaviour at 4 weeks of age in the elevated plus-maze paradigm. Neither Mecp2-null male nor Mecp2-heterozygous female mice presented differences in (a) the percentage of open arms time and (b) the percentage of open arms entries, which are measures of the state of anxiety that the animals exhibit in a new environment. (c) Mecp2-null animals had fewer entries into the closed arms than their wt littermate controls (P = 0.014) suggesting the existence of a locomotor impairment. (Mecp2-heterozygous females, n = 11; wt females, n = 8; Mecp2-null males, n = 13 and wt males, n = 13. Values are mean ± SEM. PND, postnatal day; ko, knock-out; wt, wild = type, *P < 0.05).

Motor co-ordination

At 5 weeks of age, Mecp2-mutant animals were tested in the rotarod to evaluate their motor co-ordination (Mecp2-null n = 11, wt littermate males n = 11, Mecp2-heterozygote n = 16, wt littermate females n = 9). After 3 days of training, mice were tested at different speeds. Mecp2-null and Mecp2-heterozygous mice, when compared to wt control mice, presented a reduced latency to fall off the rod. This reduction was statistically significant at 15 r.p.m. for male (P = 0.046) and at 20 r.p.m. for female (P = 0.023) mice (Fig. 5a,b).

Figure 5.

Mecp2-mutant mice present motor problems at 5 weeks of age. The latency to fall off the rod was lower for the Mecp2-null mice at 15 r.p.m. (a) and for Mecp2-heterozygous females at 20 r.p.m. (b) than the latency exhibited by their respective wt controls. (Mecp2-heterozygous females, n = 16; wt females, n = 9; Mecp2-null males, n = 11 and wt males, n = 11. Values are mean ± SEM. PND, postnatal day; ko, knock-out; wt, wild-type, *P < 0.05).

Discussion

Delayed somatic physical growth and maturation of Mecp2-mutant mice

Among the physical growth and maturation parameters assessed in this study, differences were seen in body weight and in AGD. The body weight was significantly reduced in the Mecp2-heterozygotes, but, unexpectedly, this difference in body weight was not seen between Mecp2-null and wt control male mice in spite of their earlier disease onset. However, the curves of Mecp2-null and wt males start diverging at PND20 and would probably follow this trend at later ages. In fact, it is already known from the original publication on this model that Mecp2-null mice present a smaller body weight than wt littermate controls at 4 weeks of age (Guy et al. 2001). The same authors suggested that, given the differences observed between mice with different genetic backgrounds, the effects of MeCP2 in body weight could be mediated by one or more modifier genes. One of these modifier genes could be sex-linked and thus provide a possible explanation for the results we obtained. Also, the AGD is reduced in both male and female Mecp2-mutant mice suggesting that these animals present a slower sexual maturation. In the case of Mecp2-null mice it has been reported that their testes are always internal and they do not mate because they are too debilitated or die before adulthood. However, adult Mecp2-heterozygous mice are fertile and, as far as we know, they do not present reduced fertility and they raise normal litters (Guy et al. 2001). Taken together these results also support the evidence that MeCP2 has an effect in somatic growth markers and not only in neuronal cells (Huppke et al. 2003; Nagai et al. 2005).

Pre-weaning behaviour in the Mecp2-mutant animals suggests early neurological dysfunction

In the present study, a delay in the achievement of the postural reflex and of the surface righting reflex (only in females) was evident between Mecp2-mutant and wt animals. Both reflexes depend on the development of dynamic postural adjustments and imply the integrity of muscular and motor function (Altman & Sudarshan 1975; Dierssen et al. 2002). Acquisition of the negative geotaxis reflex, a dynamic test that reflects sensorimotor function and depends on colliculus maturation (Dierssen et al. 2002) was also disturbed. Despite those impairments, in another neurological reflex – the static wire suspension test, which is highly compensated by information from the visual and proprioceptive systems –Mecp2-mutant animals did not perform worse than wt controls. Mecp2-null animals held the wire for a longer time, even though there were no differences between Mecp2-null and wt controls as to when the mice started to grasp the wire. Thus, the fine motor skills of the forepaws did not appear to be affected in the mutant mice. The longer time holding the wire could, however, reflect the incapacity of the mutant mice to initiate a voluntary movement, which could constitute a possible sign of dyspraxia, as observed in RTT patients (Kerr & Engerstrom 2001).

All the above-mentioned reflexes are sensitive to the function of the vestibular system, of which the role is to provide information on the position and movement of body and head in space, and so they depend largely on brainstem (medullary) structures (Altman & Sudarshan 1975). The positional information is transmitted from the inner ear to the central vestibular system located in the hindbrain and integrated with information from other neural systems [for a review see (Smith et al. 2005)]. The data we obtained on neurological reflexes is particularly interesting in the light of the studies in human RTT patients that suggest dysfunction of the brainstem, where the vestibular system is located, as responsible for the early pathogenesis in RTT (Einspieler et al. 2005b; Segawa 2005). Interestingly, MeCP2 binds directly to the brain-derived neurotrophic factor (BDNF) promoter region (Chen et al. 2003; Martinowich et al. 2003) and regulates its transcription in an activity-dependent manner. BDNF appears to have an important role in the maturation and maintenance of the vestibular system, as mice deficient for BDNF and its receptor TrkB demonstrate neuronal loss in the vestibular sensory ganglia (Huang & Reichardt 2001). It is, thus, possible to speculate that the levels of this neurotrophin in the vestibular pathways could be deregulated in the Mecp2-mutant mice and in this way could also contribute to possible dysfunction in the vestibular system.

Abnormal acquisition of the NG reflex could reflect abnormalities in the maturation of the colliculi and the abnormal performance in the surface RR could reflect abnormalities in the labyrinthine function. Anomalies in the auditory canal cannot be the source of this dysfunction because mice with anomalies in this area present stereotypical behaviours (Khan et al. 2004) that are not exhibited by the Mecp2-mutants. Data on the pathology in this area of the mouse brain, as far as we know, is not yet available in the Mecp2tm1.1Bird mouse and future research is necessary to explore neuropathological correlates of the abnormal functional outcome in the first days of postnatal life of Mecp2-mutant mice.

The subtle but significant perturbations observed in the achievement of milestones are a first sign of early neurological pathology in the Mecp2tm1.1Bird mice. The motor problems that these mice experience later in life correlate with the developmental abnormalities and may even be a consequence of impaired neurodevelopment of pathways within the brainstem area.

Mecp2-mutant mice present reduced spontaneous activity as a results of motor impairments before the onset of overt symptoms

Adult Mecp2tm1.1Bird mice were initially described as presenting serious motor problems after a period of normal development (Guy et al. 2001). In fact, in their home cage at 4 weeks of age, juvenile Mecp2-mutant mice are, other than their reduced body weight, almost indistinguishable from their wt littermates. However, in the OF apparatus the Mecp2-null mice exhibit hypoactivity (Guy et al. 2001) despite a normal exploratory capacity. We were not able to notice any differences between Mecp2-heterozygous and wt control females in the OF, at 4 weeks of age, even though they were previously described to exhibit reduced spontaneous activity at later ages, when symptomatic (Guy et al. 2001). In the OF and EPM we did not identify an anxiety-like behaviour in either male or female Mecp2tm1.1Bird animals at 4 weeks of age. In accordance, performance of older symptomatic Mecp2-heterozygous animals in the OF also suggested that these mice do not present heightened anxiety (Guy et al. 2001). Anxiety is, however, described in other models of the RTT disorder (Gemelli et al. 2005; Moretti et al. 2005; Shahbazian et al. 2002).

In this study, at 5 weeks of age the Mecp2-null and Mecp2-heterozygous mice demonstrated motor co-ordination impairment. This is, to the best of our knowledge, the first study to identify the effect of Mecp2 mutation on sensorimotor co-ordination in the rotarod test in 5-week-old mice. Although differences in the locomotor profile of Mecp2-heterozygous mice when compared to wt controls were not identified in the OF and in the EPM apparatus, in the more sensitive and specific rotarod test, mutant females already showed motor problems at the age of 5 weeks. Motor co-ordination problems had already been previously reported in the other models of RTT, but not at such an early age: the Mecp2308/y animals are not impaired up to 10 weeks of age (Moretti et al. 2005), but are impaired at later ages (Shahbazian et al. 2002). Our findings suggest that MeCP2 is important for the acquisition of motor co-ordination abilities and that deregulation of its levels causes slight motor problems that appear early in development and become increasingly evident as development proceeds. The deficits in the rotarod are not likely to be the result of muscle weakness because the mutant animals held on longer in the WS test than the wt animals. Co-ordination is necessary for a good performance both in the dynamic reflexes and in the rotarod test. Hence, and regarding the data obtained in this study, a lack of limb co-ordination is apparently present in the Mecp2-mutant mice; given that both the NG reflex and the rotarod test are affected, we suggest that hind limbs are more severely involved. Rearing also presupposes hind limb strength (Altman & Sudarshan 1975) and as this parameter is not affected in these animals, the problem must reside in co-ordination of the hind limbs rather than in their strength.

The identification of early and subtle neurodevelopmental differences in the RTT mouse model provides an interesting analogy to the recent findings of minor neurological signs during the first months of life of RTT patients. Further analysis of neurodevelopment in these Mecp2-mutant mice, which mimic well the motor profile of RTT patients, should provide an insight into the underlying mechanisms of pathogenesis in this disease and contribute to a precocious RTT diagnosis that might be beneficial in terms of therapeutic approaches since the first months of life.

Acknowledgments

Mónica Santos is supported by Fundação para a Ciência e Tecnologia (FCT, Portugal) with the PhD fellowship SFRH/BD/9111/2002. Research in Rett syndrome is supported by FSE/FEDER and FCT, grant POCTI 41416/2001.

Ancillary