The first mouse model of Mecp2 deficiency has been generated by the laboratory of Adrian Bird [12]. This model has been created crossing mice carrying a floxed copy of MeCP2 with a mouse that expresses CRE recombinase on the X chromosome. The resulting mouse (Mecp2^tm1.1Bird Mecp2Bird) is deleted for exons 3 and 4 of MeCP2, respectively, encoding the methyl-binding and the transcriptional repression domains. No MeCP2 protein is detected in the mice using antibodies targeting epitopes located in the N- or C-terminus [12]. This mouse model mimics more or less the development of the RTT phenotype. The development seems unaffected during early intrauterine and postnatal periods. The KO males for MeCP2 (MeCP2^−/y) are most severely affected. The first signs are visible from the third week of postnatal life and worsen during the entire life of the animal to its death occurring at 7–8 weeks. Affected mice present first a growth deficit (reduced size) and abnormal positioning of the hind limbs (hind limb clasping). At the motor level, the mice exhibit a loss of coordination associated to the reduction in the number of spontaneous movements. Breathing instability is also observed mainly characterized by hyperventilation followed by apneas [13]. It is interesting to note an irregular tooth wear, possibly due to a defect of positioning between the maxillary and mandible bones. In contrast to males, no obvious symptoms are observed in heterozygous females (Mecp2^+/−) during early postnatal life. Studies of Mecp2^+/− mice have been limited because of potential phenotypic variability due to X chromosome inactivation effects. Even if the phenotypic variability is high in heterozygous female mice, this gender nonetheless represents the real RTT model and every effort should be made to confirm the results obtained in null males in heterozygous female mice. In the Mecp2 KO-B female mice, the first phenotypic signs appear around the 12th week of age consisting of abnormal positioning of the rear legs. Later, they tend to develop obesity and respiratory irregularities (around the 9th month) can be found in heterozygous females [12]. This last point is quite intriguing since most RTT patients exhibit an extremely low body weight. Several hypotheses could be proposed. A distinctive hormonal regulation in RTT girls and Mecp2 heterozygous female mice could explain the differences in weight gain. Alternatively, permanent access to food in mice could contribute to this phenomenon. Nevertheless, heterozygous female mice are definitively less affected than null males.

Concomitantly to the Bird model, the Jaenisch model (or Mecp2^tm1.1Jae or Mecp2Jae) [14] has been created with a deletion of exon 3 (encoding the DNA methyl-binding domain) and a portion of exon 4 of MeCP2. This model is generally considered as a model completely devoid of MeCP2 but nevertheless, it is worth noting that a small portion of MeCP2 is transcribed. Accordingly, if immunohistochemistry detects no MeCP2 protein, western blot can reveal small peptides in mutant brain mutant. Because the N-terminus is always present, including for example the NLS (Nuclear localization sequence) and a part of the TRD (transcriptional repression domain), it would be wise to consider the consequences of the expression of this mutant peptides [14, 15]. In vivo, KO-J mice appear normal until the 4–5th weeks of postnatal life. They display progressive behavioral disorders (impairment of motor skills, “hind limb clasping,” irregular breathing, tremors, epileptic seizures, weight loss), which also involve various neurological functions such as anxiety, social interactions, learning and memory. KO-J female mice have no phenotype before 4 months, even if one can observe a slightly lower weight when paired with wild type females at around 5 weeks of age. From the fourth month a decrease in locomotor activity appears. Weight gain has also been described and from the sixth month the females have trouble in positioning the hind limbs as well as slight respiratory irregularities [14–16].

Another knockout mouse (Mecp2^tm1Tam) has been created carrying a deletion of the entire exon 3 and part of exon 4, replacing the coding sequence for the MBD with a floxed Pgk-Neo cassette [17]. In this model, the region encoding the TRD and the C-terminal part Mecp2 is not transcribed [17]. This construction leads to a complete lack of Mecp2 transcript and protein in these mice. The authors generated Mecp2^tm1Tam line on a pure 129 or mixed 129 × C57BL/6 background, which could partially affect some of the behavioral tests. Mecp2^tm1Tam male mice exhibit a phenotype close to that observed in Mecp2 KO-B and Mecp2 KO-J. For instance, the locomotion of the mice was assessed by their activity on freewheels and Mecp2^tm1Tam male mice exhibit a significantly lower level of activity. Mutant male mice remain for a shorter time on the accelerated rotarod. In addition, Mecp2tm1Tam male mice present a reduction of fear and anxiety on the elevated plus maze apparatus. Finally, Mecp2tm1Tam male mice show impaired fear conditioning and contextual association. In addition to these classical test authors mentioned that mutant mice frequently present an abnormal gait, a folding of rear legs, tousled hair, abnormal breathing, tremors and seizures (however, no statistical study was reported for these parameters). From 5 weeks, the condition of male Mecp2 Tam begins to worsen. No difference in body weight was observed between wild-type females and heterozygote female mice. At 6 months females present front limb stereotypies and motor deficits [17, 18].

In order to study the role of MeCP2 in the regulation of the autonomic functions, the group led by Jeffrey Neul in Houston [19] crossed the floxed MeCP2 mouse with Hoxb1-Cre mice [20] that expresses the Cre recombinase in the distal part of the hindbrain: the caudal pons (including the locus coeruleus), the medulla oblongata and the spinal cord. This area contains the respiratory centers such as the retrotrapezoid core and the pre-Bötzinger complex but also the noradrenergic cell groups A5, A1/C1, A2/C2 which represent strong respiratory modulators [21]. In addition to the brainstem, these mice exhibit also a lack of Mecp2 in many peripheral autonomic tissues. The analysis of the phenotype of Hoxb1-cKO mouse reveals a disruption of motor coordination with the conservation of behaviors depending of a neural substrate preferably located in the forebrain. The cortical activity recorded by EEG is normal, suggesting that the abnormal activity observed in MeCP2^−/y mice is the consequence of a lack of MeCP2 in areas other than those expressing Hoxb1. It is important to note that these data were obtained using a 129S6 genetic background while most other studies with models of Mecp2 dysfunction were studied in a C57BL/6 background. Genetic background differences need to be taken into account to interpret the results of any study using mice [22]. When they grow older these mice tend to be obese, develop heart arrhythmia, and have abnormal baseline breathing. They have a reduced life span [19].

After the initial publication of the KO-J mouse model (see above), the same group inactivated Mecp2 in the postmitotic forebrain neurons expressing CaMKII to assess their contribution to the Mecp2-null phenotype. In this model the mutation was induced after the second postnatal day. The loss of Mecp2 preferentially in excitatory neurons of the forebrain causes a less severe phenotype than the constitutive knockout mice [14]. A more detailed analysis of the behavior of these mice at postnatal stages has shown that conditional inactivation of Mecp2 in the forebrain is sufficient to induce motor deficits (coordination), anxiety, and impaired social interactions [23].

To study the impact of the lack of MeCP2 in catecholaminergic or serotonergic neurons, the group of Huda Zoghbi invalidated Mecp2, respectively, in the neurons expressing tyrosine hydroxylase (Th), the rate-limiting enzyme in the catecholamine synthesis [24] or in neurons expressing PC12-ets factor 1 (Pet1) a factor involved in the acquisition of the neurochemical identity of serotonergic neurons [25] using the Cre-loxP system. In the Th-cKO mice, the deletion is occurring in Th-positive neurons including dopaminergic nuclei (substantia nigra pars compacta, ventral tegmental area) and noradrenergic (locus coeruleus, medulla) central and peripheral tissues such as the catecholaminergic medulla. At the neurochemical level, this conditional KO causes a large reduction of dopamine and norepinephrine levels. From a behavioral point of view, animals exhibit impaired motor function but retain normal memory and motor skills learning and show no signs of anxiety or impairment of social interactions. It is interesting to note that the respiratory phenotype does not seem to be exacerbated by the lack of MeCP2 in Th-expressing neurons. Pet1-cKO animals are more aggressive when exposed to conspecific partner mice. Surprisingly, grooming or stereotyped movements, which are known to be modulated by serotonin, are not modified. Similar to the Th-cKO mice, the deletion in the Pet1 neurons does not aggravate the respiratory phenotype. Overall, the two conditional knockout lines (cKO and Th-Pet1-cKO) survive up to 18 months of age while constitutional MeCP2^−/y mice die around 69 days after birth. More recently, Lang and colleagues [26] created a new mouse model with a selective preservation of MeCP2 in catecholaminergic cells. This catecholaminergic rescue is sufficient to improve the behavioral phenotype of male and female Mecp2-deficient mice. Motor coordination is improved and epileptic manifestations are reduced, and the lifespan of the animals is significantly increased.

To better understand the role of the hypothalamus in RTT, the group of Huda Zoghbi selectively inactivated Mecp2 in single minded-1 (Sim1)-positive hypothalamic neurons [27]. These Mecp2-Sim1 mice present no obvious deficits in motor coordination, tremors, learning, or memory. However, they have an enhanced physiologic response to stress associated to an increase of corticosterone secretion compared to the wild-type mice. This enhanced physiologic response to stress has been previously found in RTT patients who present an increase of urinary cortisol [28]. In the open field, Mecp2 Sim1 male mice explored the center significantly less than wild type mice, a finding consistent with enhanced anxiety-like behavior. However, light/dark box assay, one of the most used tests for studying anxiety-like behavior, did not reveal differences. The authors also found a role of Mecp2 in the regulation of social and feeding behaviors that largely rely on the hypothalamus. These mice are more aggressive and they engage in significantly more tail rattles and aggressive attacks to a new intruder than their control littermates. Finally, Mecp2 Sim1 male mice present a normal activity and normal basal metabolic rates but are hyperphagic and frequently obese possibly due to the increase in leptin and neuropeptide Y levels [27]. This last point is quite intriguing since in KO models B or J only heterozygous female mice become obese while adult male are usually skinny. Perhaps, the lack of Mecp2 in the whole body or in the hypothalamus alone will dramatically change the feeding behavior. Another possibility could be that the pure FVB background used in this study could affect the feeding behavior differently compared to B or J knockouts maintained on pure C57BL/6 background.

RTT is characterized by motor abnormalities, seizures, autistic features, and stereotyped behaviors. Most of these functions are regulated by GABA signaling, the most important inhibitory neurotransmitter in the brain. Chao and colleagues [29] produced a conditional Mecp2 knockout in neurons expressing the vesicular GABA transporter (vesicular amino acid transport, Viaat). Viaat-Mecp2^−/y mice are indistinguishable from controls until 5 weeks of age, when they exhibit hind limb clasping, forelimb stereotypies, and a significant increase of grooming behavior associated to self-injury. At 9 weeks, these mice present impaired motor coordination, learning, memory, and respiratory irregularities. In addition, EEG recordings reveal hyper excitability without LTP modifications. Social interactions were increased whereas no signs of anxiety were detected. This Mecp2-Viaat model seems to recapitulate most of the deficits observed in Mecp2 null mice.

Patients affected by RTT are normal at birth and progressively display clinical features during the first years of life. The progressive appearance of Mecp2-deficiency symptoms suggests that the function of Mecp2 in the maturing nervous system is critical for establishing normal adult neurological function. Two different groups developed an adult onset model of RTT by crossing mice harboring a floxed Mecp2 allele and a tamoxifen-inducible CreER allele to delete Mecp2 when animals are fully mature [30, 31]. By 10 weeks, McGraw and colleagues found that AKO mice are less active, have an impaired motor coordination (accelerating rotarod) and abnormal gait, and develop hind limb clasping similar to 10–11-week-old KO mice. AKO mice also develop motor abnormalities and impaired nesting ability together with impaired learning and memory. Using accelerating rotarod, Cheval and colleagues found that the latency to fall is significantly shorter for AKO mice compared to controls. Moreover, performances across the 3 days of trials suggest that AKO mice are unable to learn the motor task, but it is unclear whether this is due to a learning deficit or altered motor coordination. Most interestingly, both studies show that a removal of MeCP2 in older mice rapidly leads to their death. These results indicate that Mecp2 has an important role in neuromaintenance and associated behaviors, even if the brain is fully mature.