MeCP2 R168X male and female mutant mice exhibit Rett-like behavioral deficits

Authors


Abstract

Rett syndrome (RTT) is a regressive developmental disorder characterized by motor and breathing abnormalities, anxiety, cognitive dysfunction and seizures. Approximately 95% of RTT cases are caused by more than 200 different mutations in the X-linked gene encoding methyl-CpG-binding protein 2 (MeCP2). While numerous transgenic mice have been created modeling common mutations in MeCP2, the behavioral phenotype of many of these male and, especially, female mutant mice has not been well characterized. Thorough phenotyping of additional RTT mouse models will provide valuable insight into the effects of Mecp2 mutations on behavior and aid in the selection of appropriate models, ages, sexes and outcome measures for preclinical trials. In this study, we characterize the phenotype of male and female mice containing the early truncating MeCP2 R168X nonsense point mutation, one of the most common in RTT individuals, and compare the phenotypes to Mecp2 null mutants. Mecp2R168X mutants mirror many clinical features of RTT. Mecp2R168X/y males exhibit impaired motor and cognitive function and reduced anxiety. The behavioral phenotype is less severe and with later onset in Mecp2R168X/+ females. Seizures were noted in 3.7% of Mecp2R168X mutant females. The phenotype in Mecp2R168X/y mutant males is remarkably similar to our previous characterizations of Mecp2 null males, whereas Mecp2R168X/+ females exhibit a number of phenotypic differences from females heterozygous for a null Mecp2 mutation. This study describes a number of highly robust behavioral paradigms that can be used in preclinical drug trials and underscores the importance of including Mecp2 mutant females in preclinical studies.

Rett syndrome (RTT; OMIM #312750) is a regressive neurodevelopmental disorder that currently has no cure. It affects 1:10 000 females. Girls with RTT undergo a period of apparently normal development for 6–18 months followed by somatic and neurological regression characterized by altered growth, loss of motor and language skills, cognitive dysfunction and autistic-like social withdrawal. Hand stereotypies, breathing irregularities, anxiety and seizures develop during a subsequent stabilization period (Kaufmann et al. 2012; Neul et al. 2010). While the combination of clinical features in RTT is unique to this disorder, large phenotypic variability exists among RTT individuals. Approximately 95% of RTT cases are caused by mutations in the X-linked gene encoding methyl-CpG-binding protein 2 (MeCP2). More than 200 different mutations in MeCP2 have been identified (Bienvenu & Chelly 2006; Van Den Veyver & Zoghbi 2001), which likely represents the primarily source of phenotypic variability among cases (Neul et al. 2008; Samaco et al. 2013).

Numerous mouse models of RTT containing null mutations and common human nonsense (R168X, R255X, R270X, G273X and 308 truncation) and missense (A140V and T158A) point mutations have been created to facilitate an understanding of the underlying molecular mechanisms of RTT and provide avenues for preclinical testing (Baker et al. 2013; Katz et al. 2012 for review). Despite the existence of multiple mouse models, critical gaps in our knowledge of the behavioral phenotypes still remain (Katz et al. 2012). Specifically, how does behavior vary across age, sex and mutation type? The most extensively characterized RTT models are those containing large deletions (Jaenisch: Mecp2Jae and Bird: Mecp2Bird) and a C-terminal truncation (Mecp2308). Male and female Mecp2Jae and Mecp2Bird mutants and male Mecp2308 mutants phenocopy most RTT features including growth, motor, respiratory and cognitive abnormalities, although onset in Mecp2308 males occurs later. Characterization of the behavioral phenotype, including the course of symptom progression, is limited in hemizygous male and absent in heterozygous female mice containing common human point mutations (reviewed in Katz et al. 2012). Given the phenotypic variability among RTT females with different MeCP2 alleles (Bebbington et al. 2012) and the greater frequency of point mutations than large deletions in the RTT population (estimated at 11–12% per point mutation vs. 5%, respectively; Bebbington et al. 2010), more extensive characterization of RTT mouse models containing common point mutations is critically important. Characterizations of these models can provide both in vivo evidence about the effects of MeCP2 mutation type on behavioral phenotype and support preclinical trial design by aiding in the choice of appropriate model, age, sex and outcome measures for use with specific drug targets.

Here, we provide a detailed characterization of the behavioral phenotypes of both male and female mice carrying a knock-in of the common early truncating R168X nonsense mutation (Christodoulou et al. 2003; Lawson-Yuen et al. 2007) and monitor symptom progression. We compare these phenotypes to our characterizations of Mecp2Jae (Chen et al. 2001) and the phenotypes of commonly studied Mecp2Bird (Guy et al. 2001) and Mecp2308 (Shahbazian et al. 2002) male and female mutants.

Materials and methods

Rett mouse model

All experiments were conducted on Mecp2R168X mice generously provided by the Coyle laboratory (Lawson-Yuen et al. 2007). Procedures were approved by the Tufts University Institutional Animal Care and Use Committee and conformed to standards set forth in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. F1 hybrid C57BL/6 × 129S6/SvEv Tac Mecp2R168X/+ females were obtained through rederivation. F1 Mecp2R168X/+ females were backcrossed two generations to C57BL/6 males (Jackson Laboratories, Bar Harbor, ME, USA). The offspring became increasingly symptomatic; therefore, we maintained the Mecp2R168X colony on a mixed background. Mice for these studies were generated by crossing these mixed C57BL/6 × 129S6/SvEv Tac Mecp2R168X/+ females and wild-type males for one to two generations. After weaning on postnatal day (PN) 23, pups were group housed in cages with up to five same sex littermates. Mice were maintained on a 12-h light/dark cycle with lights on at 0600 h and food and water provided ad libitum. All behavioral tests, with the exception of dark-cycle locomotor activity, were conducted during the light cycle between 1200 and 1700 h.

Behavioral testing

An observer blind to the mouse's genotype performed and analyzed all behavioral tests. Sixty-two male and 71 female Mecp2R168X and wild-type littermates were assessed on a variety of behavioral tasks examining general health, somatic growth, motor, anxiety and cognitive function (ages and testing cohorts shown in Table 1). To monitor symptom progression, mice were tested shortly after weaning (males: PN 26–PN 35; females PN 26–PN 40) and at an age when mutants are expected to be fully symptomatic based on our studies in Mecp2Jae mice (males: PN 40–45; females: PN 60–80).

Table 1. Time line for behavioral testing
 MaleFemale
  1. Superscript letters represent five cohorts of mice tested.

Physical and neurological health  
Body weightPN 24a, 43aPN 29d, 64e
Reaching and rightingPN 24a, 43aPN 29d, 64e
Shock sensitivity, hearingPN 35bPN 41d, 80e
Motor function  
Forelimb grip strengthPN 24a, 43aPN 29d, 64e
Dark-cycle activity levelsPN 23a, 42aPN 28d, 63e
Motor coordinationPN 29cPN 30d, 74e
Anxiety  
Zero mazePN 25bPN 26d, 60e
Learning and memory  
Novel object recognitionPN 27–29aPN 34–36d
Context and cued fear conditioningPN 35–36bPN 40–41d, 79–80e
Visible seizuresOngoingOngoing

Neurological battery

Mice were weighed and a neurological battery was performed to evaluate the general health and basic motor skills of the mice. Reaching and righting reflexes, and grip strength were measured as described previously (Stearns et al. 2007).

Dark-cycle locomotor activity

Locomotor activity was used to assess general motor function. Activity was measured across the 12-h dark cycle using a photobeam activity system (San Diego Instruments, San Diego, CA, USA). Mice were placed individually into a cage (47 × 25 × 21 cm) inside a rectangular arena equipped with a 3 × 8 array of photobeams. The average number of ambulatory (two adjacent) and fine (repeated single) beam breaks per hour over the 12 h was compared.

Accelerating rotarod

Balance and motor coordination were measured on an accelerating rotarod across three trials each spaced 10 min apart (San Diego Instruments).

Zero maze

Anxiety-related behavior was measured on the zero maze (Stearns et al. 2007). Mice were given 5 min to explore the open or closed arms on the zero maze and time in open arms, number of transitions to open arms and distance traveled were recorded using Anymaze (Stoelting, Wood Dale, IL, USA).

Novel object recognition

Novel object memory was assessed during three sessions. This task relies on the innate tendency of a mouse to explore unfamiliar objects vs. familiar objects. Testing was performed in an open-field arena (Schaevitz et al. 2012). Twenty-four hours prior to training, mice were habituated to the arena for 10 min. During training, mice were given 10 min to explore two identical lego objects (A + A). Short- and long-term object memory were assessed in two subsequent sessions (60 min or 24 h after the completion of training) during which mice were given 10 min to explore the familiar (A) or a novel (B or C) object. The duration of exploration (defined as the mouse's snout or forelimbs physically touching or approaching within 1 cm of an object) of familiar and novel objects was measured. The amount of time spent exploring the novel object over the total time exploring both novel and familiar objects in each session was used to measure object memory.

Contextual and cued fear conditioning

Associative learning was assessed using an automated fear conditioning system for mice (Coulbourn Instruments, Allentown, PA, USA) with the following parameters: Day 1, acquisition: 3-min habituation to context A, followed by two tones (80 dB) and 0.6 mA footshock pairings. Day 2, contextual retention: 5 min in context A with no tone or shock. Cued retention: performed at least 1 h after contextual retention consisting of 3 min in context B, followed by 3-min tone (80 dB). Freezing during contextual and cued retention was recorded by an automated system (Graphic State 3.0 software; Coulbourn Instruments). Prior to statistical analysis, all freezing in intervals greater than 2 seconds in length during acquisition, contextual and cued retention were converted to a percentage freezing value.

Statistical analysis

Data were analyzed using one-way analysis of variance (anovas) with genotype as the between-group factors and mixed-model anova with testing session as the repeated measure. Post hoc Bonferroni analyses were used to examine differences between genotype and session for significant anovas. To assess memory on the novel object recognition task, a one-sample t-test was used to determine whether performance within a group was statistically different than 50%, denoting chance. All analyses were performed using SPSS software (SPSS Inc., Chicago, IL, USA) with P < 0.05 considered significant.

Results

Somatic development is altered in Mecp2R168X male, but not female, mutant mice

Body weight is highly variable in RTT girls and in mouse models of RTT. Individuals may be significantly underweight or overweight (Renieri et al. 2009; Tarquinio et al. 2012). Additionally, muscle tone is decreased in RTT females and in some mouse models of RTT (Neul et al. 2010; Stearns et al. 2007). To determine whether Mecp2R168X mutants display a similar phenotype, we conducted a neurological battery including analysis of body weight and forelimb grip strength.

Males

Body weight was significantly reduced in Mecp2R168X/y mice at both PN 24 (F1,31 = 8.93, P = 0.005) and 43 (F1,56 = 24.35, P < 0.001) (Fig. 1a), as was forelimb grip strength (PN 25: F1,28 = 5.94, P = 0.021; PN 43: F1,28 = 36.18, P < 0.001) (Fig. 1c). Reaching and righting reflexes assessed during the neurological battery were unaffected in Mecp2R168X/y mice; all mice performed normally (data not shown). Lifespan was not directly measured; however, most Mecp2R168X/y male mutants were very sick (lethargic, ruffled fur and respiratory abnormalities) at euthanasia on PN 50.

Figure 1.

Weight and grip strength in Mecp2R168X mutant male and female mice. Weight is significantly reduced in Mecp2R168X/y males (a) at PN 25 and PN 43 compared with littermate controls. In Mecp2R168X/+ females (b) weight is unaffected at PN 29, but significantly reduced at PN 64 compared with littermate controls. Grip strength is significantly reduced in Mecp2R168X/y males (c) at PN 25 and 43, but unaffected in Mecp2R168X/+ females (d) at PN 29 and PN 64 compared with their sex-matched littermate controls. n given within bars. All values represent average ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Females

Body weight was normal at PN 29 (F1,37 = 0.13, P = 0.72), but significantly reduced in Mecp2R168X/+ females compared with wild-type controls at PN 64 (F1,28 = 5.29, P = 0.03) (Fig. 1b). Forelimb grip strength at both PN 29 and PN 64 (Fig. 1d) was not significantly different between Mecp2R168X/+ mice and wild-type mice (PN 29: F1,37 = 0.06, P = 0.82; PN 64: F1,28 = 2.42, P = 0.13). Reaching and righting reflexes were also normal in all Mecp2R168X/+ female mice. In addition, we observed tonic-clonic seizures in 3.7% (n = 5 of 136 heterozygous females) of Mecp2R168X/+ heterozygous mice, after about 4 months. Seizures occurred during cage changing and appeared to be stimulated by touch. The seizures were very severe such that affected females typically died within a month of the first observation. Apart from heterozygous females that died young following seizures (average age at death 182 ± 52 days), lifespan was greater than 1 year.

Mecp2R168X male and female mutant mice exhibit impaired motor function

Motor dysfunction is a prominent feature in girls and women with RTT (Bebbington et al. 2012) that is closely mirrored in mouse models using measurements of activity levels and motor coordination (e.g. Stearns et al. 2007).

Males

Locomotor activity levels were assessed across the 12-h dark cycle at an early (PN 24) and a late (PN 43) time point in Mecp2R168X/y mice to determine whether activity decreases over time (Fig. 2a,c). At PN 24, Mecp2R168X/y mice displayed reduced ambulatory (F1,31 = 4.42, P = 0.044), but not fine motor movements (F1,31 = 2.15, P = 0.153), across the 12-h dark cycle. When we assessed activity at PN 43, Mecp2R168X/y mice exhibited both reduced ambulatory (F1,31 = 28.48, P < 0.001) and fine motor movements (F1,31 = 15.26, P < 0.001). Comparing the number of ambulatory or fine movements across days revealed a significant effect of day and a significant day × genotype interaction for ambulatory (day: F1,28 = 83.91, P < 0.001; day × genotype: F1,28 = 16.46, P < 0.001), but not for fine movements (day: F1,28 = 19.00, P < 0.001; day × genotype: F1,28 = 0.15, P = 0.70). Post hoc analysis comparing the number of ambulatory beam breaks between PN 24 and 43 revealed that both wild-type and Mecp2R168X/y mice moved significantly more on PN 43 than PN 24 (P ≤ 0.001). Motor coordination was assessed on the rotarod at PN 29 (Fig. 2e). Latency to fall was reduced in Mecp2R168X/y mice compared with wild-type littermates, suggesting impaired motor coordination (F1,16 = 31.20, P < 0.001).

Figure 2.

Motor function is impaired in Mecp2R168X mutant male and female mice. The number of ambulatory (two consecutive beam breaks, a, b) and fine (repeated breaking of the same beam, c, d) motor movements is shown in 1-h intervals across the 12-h dark cycle. Ambulatory movements are significantly reduced in Mecp2R168X/y males (a) at both PN 25 and PN 43. In Mecp2R168X/+ females, ambulatory motor activity (b) is unaffected at PN 28, but significantly reduced at PN 63 compared with sex-matched wild-type littermates. Fine motor movements in Mecp2R168X/y males (c) are unaffected on PN 25, but significantly reduced on PN 43. In Mecp2R168X/+ females, fine movements (d) are unaffected at PN 28, but significantly reduced on PN 63. Latency to fall from the rotarod was significantly shorter in Mecp2R168X/y males (e) at PN 29 and Mecp2R168X/+ females (f) at both PN 30 and PN 74. n given within bars. All values represent average ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Females

Locomotor activity was also assessed in Mecp2R168X/+ females at an early (PN 28) and a late (PN 63) time point (Fig. 2b,d). On PN 28, there were no significant differences in either ambulatory (F1,37 = 1.53, P = 0.22) or fine motor movements (F1,37 = 0.71, P = 0.40) between Mecp2R168X/+ mice and wild-type controls. By PN 63, however, Mecp2R168X/+ mice displayed reduced ambulatory (F1,29 = 9.10, P = 0.005) and fine motor movements (F1,29 = 7.68, P = 0.01) compared with wild-type females. Motor coordination was impaired in Mecp2R168X/+ mice at both PN 30 and PN 74 (Fig. 2f) as evidenced by decreased latency to fall on the rotarod compared with wild-type females (PN 30: F1,37 = 11.50, P = 0.002; PN 74: F1,28 = 16.32, P < 0.001).

Anxiety-like behavior is altered in Mecp2R168X male, but not female, mutant mice

RTT girls display anxious behavior particularly in novel environments (Mount et al. 2002). Conversely, many mouse models of RTT do not recapitulate this phenotype and, instead, show reduced anxiety-like behavior on tasks such as the zero maze, elevated plus maze and open-field tests (e.g. Samaco et al. 2013; Stearns et al. 2007).

Males

In this study, we used the zero maze to assess anxiety at PN 25 prior to testing the mice on any other behavioral task (Table 2). Percent time spent in open arms and the number of open arm entries were significantly increased in Mecp2R168X/y mice compared with wild-type littermates, indicative of reduced anxiety (% time in open arms: F1,26 = 47.47, P < 0.001; open arm entries: F1,26 = 34.34, P < 0.001). This difference was not due to a motor impairment. Mecp2R168X/y mice traveled a significantly greater distance than wild-type littermates on the maze (F1,27 = 12.24, P < 0.001).

Table 2. Zero maze
 Male (PN 25)Female (PN 26)Female (PN 60)
 WTR168XWTR168XWTR168X
  1. WT, wild type.

  2. ***P < 0.001.

n141516161616
% Open arm time16.4 ± 2.647.7 ± 3.6***23.3 ± 3.128.1 ± 3.230.5 ± 3.227.9 ± 2.9
Open arm entries8.0 ± 0.824.0 ± 2.7***15.3 ± 2.120.0 ± 2.223 ± 2.620.4 ± 1.6
Distance (m)4.7 ± 0.49.7 ± 1.0***4.9 ± 0.56.2 ± 0.712.0 ± 1.110.4 ± 0.7

Females

Anxiety-like behavior in Mecp2R168X/+ mice was assessed at both an early (PN 26) and a late (PN 60) time point (Table 2). At both ages, Mecp2R168X/+ mice spent a similar percentage of time in the open arms (PN 26: F1,30 = 1.21, P = 0.28; PN 60: F1,30 = 0.26, P = 0.62) and entered the open arms a similar number of times (PN 26: F1,30 = 2.45, P = 0.13; PN 60: F1,30 = 1.07, P = 0.31) as wild-type littermates, suggesting normal anxiety-like behavior. Measures of anxiety-like behavior were not confounded by motor impairments in Mecp2R168X/+ females as distance traveled on the maze was not significantly different between groups at either age (PN 26: F1,30 = 2.06, P = 0.16; PN 60: F1,30 = 1.66, P = 0.21).

Cognition function is impaired in Mecp2R168X male and to a lesser extent in female mutant mice

Although cognitive function is difficult to assess in individuals with RTT, many are believed to have intellectual impairments (Velloso Rde et al. 2009). Similarly, the majority of RTT mouse models display cognitive deficits using both novel object recognition and associative fear conditioning paradigms (Katz et al. 2012). The novel object recognition task relies on an animal's innate tendency to spend more time with a novel object than a familiar object. In the fear conditioning task, mice learn to associate both the environment (context) and a tone (cue) with foot shocks during an acquisition phase. Freezing, an adaptive response to fear in mice, is measured 24 h later when the animals are presented with either the original context without the tone (context) or a novel context with the tone (cue).

Males

On PN 28, short-term memory (measured 1 h after object training) and long-term memory (measured 24 h after object training) are intact in wild-type mice (Fig. 3a). At both 1 and 24 h after object training, wild-type male mice spent a significantly larger percentage of time with the novel object vs. a familiar object than chance (50%) [one-sample t-test: 1 h: t(1,17) = 13.47, P < 0.001; 24 h: t(1,17) = 5.32, P < 0.001]. In Mecp2R168X/y mice, object memory is intact at 1 h [one-sample t-test: 1 h: t(1,13) = 8.44, P < 0.001], but not 24 h after object training [one-sample t-test: 24 h: t(1,13) = 0.53, P = 0.60] suggesting impaired long-term memory. Object memory across sessions was significantly impaired in Mecp2R168X/y males compared with wild-type mice (anova: F1,30 = 12.07, P = 0.002). Differences in performance between Mecp2R168X/y mice and wild-type littermates are not the result of impaired motor function as there were no significant differences between distance traveled between wild-type and Mecp2R168X/y (training: 19.7 ± 1.8 vs. 15.9 ± 1.5; 1 h: 15.4 ± 1.5 vs. 14.8 ± 1.7; 24 h: 21.7 ± 1.6 vs. 17.4 ± 1.4 m, respectively) during testing sessions (all F1,30 < 3.85, P > 0.06). Nor can altered performance on this task be attributed to differences to general interest in objects as there were no significant differences between total time spent with objects (Fig. 3a) during testing sessions (all F1,30 < 2.54, P > 0.12).

Figure 3.

Cognitive function is impaired in Mecp2R168X mutant male and to a lesser extent in mutant female mice. Cognitive function was assessed using both a novel object recognition (a, c) and an associative fear conditioning (b, d, e) task. On the novel object recognition task, percent time with the novel object vs. a familiar object is significantly greater than chance (50%) in wild-type males (a) and females (c) at both 1 and 24 h after training, whereas percent time with the novel object is only significantly greater than chance at 1 h, but not 24 h after training in Mecp2R168X mutant males (a) and females (c) indicated impaired memory at 24 h. Time spent with objects was similar in Mecp2R168X mutant males (a) and females (c) and sex-matched littermates, suggesting that impaired memory is not due to reduced object interest. On the associative fear conditioning task, percent time freezing during acquisition is significantly higher in both Mecp2R168X mutant males at PN 35 (b) and females at PN 80 (e), but normal in Mecp2R168X/+ females at PN 40 (d). Mecp2R168X/y males displayed a significantly lower percent time freezing during both the contextual and cued retention tasks 24 h later (b), suggesting impaired long-term memory. Mecp2R168X/+ mice perform no differently on the contextual and cued memory tasks than wild-type controls at either PN 40 (d) or PN 80 (e). n: males: 14–18/group and females: 8–16/group. All values represent average ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 vs. wild-type, #P < 0.05 vs. chance (dotted line).

During acquisition of the fear conditioning task on PN 35, Mecp2R168X/y mice froze for a significantly larger percentage of time than their wild-type littermates (F1,27 = 6.54, P = 0.016) (Fig. 3b), likely coinciding with the motor impairments displayed by Mecp2R168X/y male mice at this age. Despite increased baseline freezing, Mecp2R168X/y male mice froze for a significantly smaller percentage of time on both the context (F1,27 = 10.04, P = 0.004) and cued (F1,27 = 22.56, P < 0.001) memory tasks 24 h later (Fig. 3b), suggesting impaired memory.

Females

On the novel object recognition task at PN 34, short-term (1 h) and long-term (24 h) memory for a familiar object were intact in wild-type females (Fig. 3c). At both 1 and 24 h after object training, wild-type female mice spent significantly more time with the novel object than chance (50%) [one-sample t-test: 1 h: t(1,7) = 10.77, P < 0.001; 24 h: t(1,7) = 3.58, P < 0.009]. Similar to Mecp2R168X/y males, Mecp2R168X/+ females displayed intact object memory at 1 h [one-sample t-test: 1 h: t(1,10) = 11.34, P < 0.001], but not 24 h after training [one-sample t-test: 24 h: t(1,10) = 0.46, P = 0.66], suggesting impaired long-term object memory. Object memory across sessions was significantly impaired in Mecp2R168X/+ females compared with wild-type females (anova: F1,17 = 5.77, P = 0.028). Impaired cognitive performance in Mecp2R168X/+ mice was not the result of motor dysfunction as distance traveled during the sessions was similar between Mecp2R168X/+ and wild-type littermates (training: 22.8 ± 1.0 vs. 22.1 ± 1.6; 1 h: 16.4 ± 1.3 vs. 17.9 ± 1.9; 24 h: 18.4 ± 1.4 vs. 20.4 ± 1.9 m, respectively; all F1,17 < 0.70, P > 0.42). Nor can cognitive dysfunction be attributed to differences to general interest in objects as the total time spent with objects (Fig. 3c) was similar between Mecp2R168X/+ and wild-type littermates during all sessions (all F1,17 < 0.51, P > 0.48).

On the associative fear conditioning task, female mice were assessed at both an early (PN 40) and a late (PN 80) time point. At PN 40, the percent time freezing during acquisition (F1,23 = 0.09, P = 0.77) and during the context (F1,23 = 0.45, P = 0.51) and cued (F1,23 = 0.04, P = 0.84) memory tasks 24 h after acquisition were similar between Mecp2R168X/+ females and wild-type littermates (Fig. 3d), suggesting normal memory in Mecp2R168X/+ mutants at this age. At PN 80, Mecp2R168X/+ females spent a significantly larger percent time freezing during the acquisition phase than did their wild-type littermates (F1,28 = 9.78, P = 0.004) (Fig. 3e). Mecp2R168X/+ female mice froze for a similar percentage of time on both the context (F1,28 = 1.79, P = 0.19) and cued (F1,28 = 0.007, P = 0.94) memory tasks 24 h later.

Discussion

The behavioral phenotype of Mecp2R168X male and female mutant mice, in this study, mirrors many of the clinical symptoms in RTT individuals and other RTT mouse models (Table 3). Body weight is dysregulated in Mecp2R168X mutants as in RTT girls. Here, both male and female mutants are underweight as juveniles and young adults, whereas previously Mecp2R168X/y male mice exhibited both underweight and overweight phenotypes, depending on age (Lawson-Yuen et al. 2007). Weight differences between studies are likely owing to the influence of background strain (mixed C57BL/6 and 129S6/SvEv Tac vs. pure 129S6/SvEv Tac) (Samaco et al. 2013).

Table 3. Comparison of the behavioral phenotype of mouse models of RTTThumbnail image of

Impairments in muscle tone, fine and gross motor skills, as well as the presence of involuntary movements and stereotypies comprise a large number of the criteria used to assess clinical severity in RTT cases (Kerr et al. 2001; Pelka et al. 2006). Consistent with a severe motor phenotype in humans with the R168X mutation (Neul et al. 2008), both Mecp2R168X male and female mutant mice display deficits on multiple motor tasks. Impaired motor coordination is evident in both male and female Mecp2R168X mutants at the earliest ages examined (PN 29 and 30). Hypoactivity is present in Mecp2R168X/y males at PN 23, but not in Mecp2R168X/+ females at PN 28. Hypoactivity is evident in females at PN 63. Early deficits in complex motor coordination with the later onset of hypoactivity may reflect progressive motor dysfunction in RTT individuals (Neul et al. 2010). In Mecp2R168X/y male mice at both PN 24 and 42, motor deficits are associated with decreased muscle tone, measured by forelimb grip strength. In Mecp2R168X/+ females, muscle tone is unaffected at either PN 28 or 64 suggesting a less severe motor phenotype in female than male mutants.

The motor phenotype, evident at weaning, is similar between Mecp2R168X/y males and our studies utilizing Mecp2Jae/y mice (Schaevitz et al. 2012). Because we first observed locomotor decreases in Mecp2Jae/+ females on PN 32 and decreases in Mecp2R168X/+ on PN 63, it appears that motor dysfunction progresses more slowly in Mecp2R168X/+ females. The high penetrance of motor impairments across sex, age and mutation type indicates that assays of motor function are robust and reliable outcome measures for preclinical trials. Furthermore, the dissociation between onset of motor coordination deficits and hypoactivity in Mecp2R168X/+ may provide a unique opportunity to evaluate therapeutic efficacy during an early symptomatic stage, a stage not apparent in Mecp2R168X/y male or Mecp2Jae/+ male or female mutants.

Performance on anxiety and cognitive behavioral tasks is more variable. Anxiety-like behavior is reduced (less anxious) in Mecp2R168X/y males similar to our previous reports in Mecp2Jae/y (Stearns et al. 2007) and in males with other Mecp2 mutations (Kerr et al. 2012). Interestingly, anxiety is unaffected in Mecp2R168X/+ females at either PN 26 or 56, which is in contrast to decreased anxiety displayed in Mecp2Jae/+ (Stearns et al. 2007) and Mecp2Bird/+ females (Samaco et al. 2013). Reduced anxiety-like behaviors prevalent across RTT mouse models, with the exception of normal anxiety in Mecp2R168X/+ females and enhanced anxiety in Mecp2308/y male mice (De Filippis et al. 2010), are dissimilar to heightened anxiety typically displayed by RTT females in novel situations (Mount et al. 2002). Thus, Mecp2R168X/y mutant males provide no additional benefit over other RTT mouse models for studying anxiety. Testing on additional anxiety tasks is required to determine whether altered anxiety can be detected in Mecp2R168X/+ females.

While many RTT individuals are reported to have intellectual impairments (Velloso Rde et al. 2009), the exact nature of cognitive dysfunction has been difficult to assess in humans (Berger-Sweeney 2011). Here, cognitive function was assessed utilizing two tasks that rely on subtly different brain circuits (Albasser et al. 2013; Maren & Quirk 2004). Early cognitive dysfunction in Mecp2R168X/y males on both the novel object recognition (PN 29) and associative fear conditioning tasks (PN 35) is consistent with impaired performance of Mecp2Jae/y (Schaevitz et al. 2010; Stearns et al. 2007) and Mecp2Bird/y males (Mcgraw et al. 2011) and suggests widespread brain abnormalities. In young Mecp2R168X/+ females, cognitive function is impaired on the object recognition (PN 35), but not on the fear conditioning task (PN 40). The dissociation between cognitive performance on these two tasks is consistent with our studies in Mecp2Jae/+ females (Nag & Berger-Sweeney 2007; Stearns et al. 2007). In contrast, at 8 weeks Mecp2Bird/+ females display contextual fear conditioning deficits (Samaco et al. 2013). To determine whether cognitive function decreases with age, we examined fear conditioning memory at PN 80. Freezing is normal during contextual and cued memory tasks; however, freezing during acquisition is increased in Mecp2R168X/+ females at PN 80. Therefore, it is difficult to determine whether normal freezing during retention tests was the result of normal memory or simply motor dysfunction.

In general, cognitive tasks that are less aversive or less well entrained may produce a weaker memory trace and allow us to detect more subtle memory deficits in Mecp2 mutant females. Mecp2Jae/+ and Mecp2R168X/+ females are impaired on a non-aversive novel object recognition task that requires memory of object identity (Stearns et al. 2007) and in Mecp2Jae/+ females on a water maze task when given half the normal number of trials to learn the spatial position of a hidden platform (Lonetti et al. 2010). This study highlights the importance of assessing cognitive function prior to onset of motor deficits and suggests that additional cognitive tests will shed light on the exact nature of dysfunction in female Mecp2 mutants. Currently, it appears that novel object recognition, but not the fear conditioning protocol described herein, is an appropriate behavioral paradigm for use with Mecp2R168X/+ females in preclinical drug studies. The novel object recognition task may translate particularly well between RTT patients and mouse models of RTT. A study utilizing eye-tracking technology to investigate novel image recognition in RTT girls revealed that RTT girls recognize novel pictures, but that they spend significantly less time looking at novel pictures than typically developing children (Rose et al. 2013).

Overall, the phenotype of female Mecp2R168X/+ mice was less severe than that of Mecp2Jae/+ females and Mecp2R168X/y males with the exception of seizures. Tonic-clonic seizures were noted in a small percentage of Mecp2R168X/+ females that typically resulted in death. While approximately 60% of RTT girls have seizures (Glaze et al. 2010), to these authors' knowledge no severe spontaneous seizures have been reported in RTT mouse models (in 10+ years working with the Mecp2Jae mice, seizures were never observed; L.S. and J.B.S., unpublished observations). In the future, careful assessment of seizure activity in the Mecp2R168X mutants is warranted. Given the higher rate of susceptibility of Mecp2R168X/+ females to severe seizures than Mecp2R168X/y males or other Mecp2 mouse models, Mecp2R168X/+ female mice may provide a particularly useful model for understanding the underlying molecular mechanisms that influence seizure susceptibility in RTT. Mecp2R168X/+ females may also be useful for preclinical testing of drugs in a seizure-prone population.

Differences between Mecp2R168X/+ and Mecp2Jae/+ females in motor symptom onset, anxiety and seizure susceptibility suggest that genotype–phenotype correlations in mouse models of RTT have the potential to inform us about subtle functional differences between MeCP2 mutant proteins that may affect drug responses in clinical trials. Differences in the behavioral phenotype of female Mecp2R168X/+ and Mecp2Jae/+ mice suggest that the R168X truncated protein may be partially functional (Lawson-Yuen et al. 2007). In in vitro studies, the R168X truncated protein can both bind methylated DNA and maintain functional interactions with other proteins (Nan et al. 2007), but it is unable to recruit protein complexes involved in transcriptional repression (Stancheva et al. 2003). Alternatively, given that truncated R168X protein contains no nuclear localization signal, it may act as a gain-of-function mutation though anomalous protein–protein interactions in the cytoplasm instead of the nucleus. The mechanisms through which truncated R168X vs. a complete loss of MeCP2 protein may affect brain neurochemistry and behavior differently are yet to be discovered. Given that the phenotypes of males with these two mutations are virtually indistinguishable, this study suggests that female mutants may provide a better model in which to detect phenotypic differences between mutation types. These data emphasize the importance of conducting preclinical trials in female Mecp2 mice in addition to more commonly used males.

Currently, no effective therapeutic strategies exist for RTT. Numerous mouse models of RTT available provide excellent tools for evaluating potential treatments prior to clinical trials. While a number of therapies show promise in mice (Abdala et al. 2010; Schmid et al. 2012), translational success has been limited. Characterization of behavioral phenotypes in additional models of RTT should improve translational success by providing the opportunity to choose the model, sex, age and outcome measure most relevant to the drug under investigation. In this study, both male and female Mecp2R168X mutant mice display a number of phenotypic abnormalities reminiscent of the clinical features of RTT, which can serve as outcome measures for preclinical trials. Furthermore, phenotypic differences between Mecp2R168X/+ and Mecp2Jae/+ mutant females and their male counterparts, such as later age of symptom onset, normal anxiety-like behavior and seizure activity, suggest that Mecp2R168X/+ mutant females provide a unique phenotype in which to assess potential therapeutics.

Acknowledgments

We would like to thank H. Furgang for help with manuscript preparation and W. Rosky for his excellent care of the animals.

Ancillary