Breathing abnormalities in a female mouse model of Rett syndrome

Abstract

Rett syndrome (RTT) is a female neurodevelopmental disease with breathing abnormalities. To understand whether breathing defects occur in the early lives of a group of female Mecp2 ^+/− mice, a mouse model of RTT, and what percentage of mice shows RTT-like breathing abnormality, breathing activity was measured by plethysmography in conscious mice. Breathing frequency variation and central apnea in a group of Mecp2 ^+/− females displayed a distribution pattern similar to Mecp2 ^−/Y males, while the rest resembled the wild-type mice. Similar results were obtained using the k-mean clustering statistics analysis. With two independent methods, about 20 % of female Mecp2 ^+/− mice showed RTT-like breathing abnormalities that began as early as 3 weeks of age in the Mecp2 ^+/− mice, and were suppressed with 3 % CO₂. The finding that only a small proportion of Mecp2 ^+/− mice develops RTT-like breathing abnormalities suggests incomplete allele inactivation in the RTT-model Mecp2 ^+/− mice.

Introduction

Rett syndrome (RTT) is a neurodevelopmental disease caused by disruptions of the X-linked gene encoding methyl-CpG-binding protein 2 (MeCP2) which affects approximately 1 in 10,000 live female births [1–3]. Males with this defect are usually unable to survive after birth. In addition to autistic symptoms, people with RTT show breathing disturbances, such as irregular breathing, apnea, hyperpnea, apneusis, Valsalva breathing, air swallowing, etc. [4, 5]. The breathing disorders contribute to the high rate (26 %) of sudden and unexpected death, and developmental abnormalities in the brain [6].

Many of these breathing abnormalities exist in male mice with the Mecp2 gene deletion (Mecp2 ^−/Y), which are typically used as mouse models of RTT. Our previous studies have shown that the Mecp2 ^−/Y mice lose their sensitivity to moderate hypercapnia, while their sensitivity to severe hypercapnia is normal [7]. The defect seems to be, at least partially, due to the abnormal expression of pH-sensitive K⁺ channels in the central nervous system [7]. As a result, CO₂ levels are detected only when hypercapnia becomes severe in the mice, leading to a transition from hypoventilation to hyperventilation. After the excessive CO₂ is removed from the body, hypoventilation is resumed. The defective response to moderate PCO₂ thus can lead to periodic hyper- and hypoventilation as seen in RTT.

Nearly all RTT patients are female, whereas the majority of previous studies on the RTT breathing phenotype were performed on male mouse models with disruption of the Mecp2 gene. The rationale behind such an approach is that the RTT-like symptoms can be reliably produced in these Mecp2 ^−/Y males, whereas they may be variable in female mice heterozygous with the Mecp2 gene deletion (Mecp2 ^+/−) owing to random X-chromosome inactivation. However, the female animal models of RTT resemble more closely human RTT patients with respect to their genotypes. Therefore, information generated in Mecp2 ^−/Y male mice needs to be validated in Mecp2 ^+/− females.

Indeed, breathing defects have been found in female Mecp2 ^+/− mice older than 2 months, which is an age equivalent to full grown humans [8, 9]. However, it is unclear what percentage of Mecp2 ^+/− females develop RTT-like breathing abnormalities, whether the breathing abnormalities occur in the early lives, and whether they deteriorate with growth. It is still challenging to address these questions, because the random X-chromosome inactivation causes only a certain number of mice to develop breathing disorders with variable symptoms, which requires subtle experimental approaches to reveal. In our previous studies, we have developed several methods for the analysis of the breathing irregularities in Mecp2-null mice, including population analysis of breathing variation and CO₂ sensitivity [7]. Therefore, we performed these studies to address the above questions in heterozygous Mecp2 ^+/− female mice.

Methods

Animals

Mecp2-null and Mecp2 heterozygous mice of strain B6.129P2 (C)-Mecp2tm1.1Bird, developed by Dr. Adrian Bird, were used as mouse models for Rett syndrome in this study [10]. To produce Mecp2 ^−/Y mice, Mecp2 ^+/− purchased from the Jackson Laboratory (Bar Harbor, ME) were crossed with Swiss Webster Mecp2 ^+/Y males. Genotypes of females and males were identified by following protocols of the Jackson Laboratory. Female and male offspring were used in the following studies. All animal experimental procedures were conducted in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the regulation of the Institutional Animal Care and Use Committee of Georgia State University.

Measurements of breathing activity by plethysmography

Breathing activity of unanesthetized mice was measured by plethysmography. The individual mouse was kept in the plethysmograph chamber (~40 ml) connected to an empty reference chamber. The animal was allowed to adapt to the chamber for at least 20 min. After this period, movements of the mouse monitored with a digital camera were reduced. A recording was taken for 30 min under normal air-ventilation conditions. Breathing activity of the mouse was analyzed offline. Records with the interference of animal movements were rejected from further analysis. A total of 52 individual Mecp2 ^+/− female mice were tested. Thirteen were tested on 2–4 occasions with at least a month before subsequent testing. One breath cycle consisted of a period of inspiration followed by a period of expiration. Times for inspiration and expiration were detected using Clampfit 10 software, and used for calculation of the frequency of breathing as described previously [7].

CO₂ ventilation

In the experiments, mice were allowed to adapt to the chamber for 20 min when the plethysmograph chamber was ventilated with room air. After a 10-min recording of the baseline breathing activity with normal air, the ventilation air was switched to 3 % CO₂ by balancing with normal air in room air for 5 min at a speed of 50 ml/min. The gas mixture was obtained from a local supplier containing 21 % O₂. The total volume of the plethysmograph chamber was ~40 ml. Since the mouse occupied most of this volume, a steady state of hypercapnia was reached within 1 min as in previous experiments [7].

Data analysis

The ventilation parameters breathing frequency (f) and apneas per hour were analyzed. To calculate variability of f, breathing activity was analyzed with the Clampfit 10 software. After an average of at least 200 successive breathing cycles, the arithmetic mean value and standard deviation (SD) were obtained. The f variability or variation index was then calculated as the division of SD by the mean value. Apneas were detected manually, and counted as events per hour based on the time frame of recording. Apneas were defined as at least one sustained breath equivalent to one regular breath cycle. The entire recording period (30 min) under normal air ventilation was used to analyze the number of apneas. Breathing f variation data is reported as mean ± SE, and apneas are presented as median ± interquartile range (IQR). Differences in f variation between two groups were determined using Student’s t-test or one-way ANOVA for multiple groups. Differences in apneas between groups were determined using a Mann–Whitney U test for two groups or Kruskal–Wallis analysis for multiple groups. Group distribution was tested with the k-means clustering statistics analysis (http://scistatcalc.blogspot.com/2014/01/k-means-clustering-calculator.html).

Results

Breathing frequency variation in female Mecp2 ^+/− mice

Because of the random expression of the X chromosome, Mecp2 ^+/− heterozygous females may have breathing phenotypes resembling either wild-type (WT) or Mecp2 ^−/Y mice. Therefore, we measured breathing activity of the females in comparison with the WT, which consisted of normal male and female mice (Mecp2 ^+/+) without Mecp2 deletion. The WT mice (n = 56) showed stable breathing activity. In contrast, a clear variation in f and apnea were seen in Mecp2-null males (Mecp2 ^−/Y) (Fig. 1b).

Fig. 1

Mecp2 ^+/+ and Mecp2 ^−/Y mice show typical normal and abnormal breathing patterns, respectively. a Mecp2 ^+/+ mice exhibit normal inspiration (downdraft) and expiration breathing patterns. b Mecp2 ^−/Y mice display stereotypical irregular breathing patterns. Arrows indicate an example of apnea. c An example trace of a Mecp2 ^+/− mouse that exhibits normal breathing patterns. d An example trace of a Mecp2 ^+/− mouse that exhibits breathing patterns similar to Mecp2 ^−/Y mice

To analyze our data quantitatively, the variation index was used to measure the breathing f variations as we described previously [7]. Consistent with our previous observations [7], the f variation was much greater in the Mecp2 ^−/Y mice, showing a significant difference from the WT mice. In contrast, certain Mecp2 ^+/− females showed breathing activity like the Mecp2 ^−/Y mice, while the breathing of others resembled the WT (Fig. 1c), suggesting that they are not a single homogenous group.

To understand whether the Mecp2 ^+/− females consist of two different populations with respect to their f variation and what percentage of mice develop RTT-like f variation, we studied the distribution of f variation index against ages. Firstly, we analyzed the f variation distribution of WT and Mecp2 ^−/Y mice. Figure 2a shows that the f variation of the Mecp2 ^−/Y mice tends to distribute between 0.2 and 0.4, while that of the WT is mostly located around 0.1 or between 0 and 0.2. Using the mean value and standard deviation of these data, Gaussian distribution curves were made to represent the data (Fig. 2a).

Fig. 2

Separation of Mecp2 ^+/− mice based on f variation. a The f variation distribution of both WT (n = 56) and Mecp2 ^−/Y (n = 13) mice against ages showed typical Gaussian distributions. WT mice are represented as open circles, Mecp2 ^−/Y mice are represented as black triangles. b The f variation distribution of Mecp2 ^+/− mice show two peaks that fit fairly to the Gaussian distributions of WT and Mecp2 ^−/Y mice in (a). The vertical line at 0.20 indicates the intersection point that was used as a threshold to Mecp2 ^+/− mice with and without breathing f variation. Of 75 Mecp2 ^+/− mice, 52 individual Mecp2 ^+/− female mice were tested. Thirteen were tested 2 to 4 times with no less than a month between tests. c Mecp2 ^+/− mice were separated based on the k-means clustering method. RTT-like Mecp2 ^+/− mice are represented as black triangles, and WT-like Mecp2 ^+/− mice are represented as open triangles. d RTT-like Mecp2 ^+/− and WT-like Mecp2 ^+/− mice indicated by the threshold level show no significant differences from those indicated by the k-means clustering method. e With the threshold separation, 13 Mecp2 ^+/− mice (Mecp2 ^+/−[R]) showed f variation like the Mecp2 ^−/Y, which have significant difference from the WT but not the Mecp2 ^−/Y mice. The rest of the WT-like Mecp2 ^+/− mice (Mecp2 ^+/−[N], n = 62) showed no significant difference in f variation from the WT. f When the f variation was separated using the k-means clustering method, the RTT-like Mecp2 ^+/− mice (n = 13, Mecp2 ^+/−[r]) have significant difference from WT (n = 56) but not from Mecp2 ^−/Y mice. Vice versa for the WT-like Mecp2 ^+/− mice (Mecp2 ^+/−[n], n = 62). Data are presented as mean ± SE (*P < 0.05, ***P < 0.005)

We then compared the f variation index of Mecp2 ^+/− females with the Gaussian distribution curves obtained from WT and Mecp2 ^−/Y mice. Figure 2b shows that although most of the Mecp2 ^+/− females have f variations around 0.1 similar to the WT mice, several Mecp2 ^+/− mice seem to have much greater f variations that were around 0.3. To prove that the apparent separation of the f variation index within the Mecp2 ^+/− female population is statistically significant, we performed the k-means clustering analysis. Using this statistics method, the f variation index of the Mecp2 ^+/− females was divided into two populations (Fig. 2c).

A characteristic of the f variation distribution is the overlap without a clear boundary, which was seen not only among Mecp2 ^+/− females (Fig. 2b) but also between WT and Mecp2 ^−/Y mice (Fig. 2a). Therefore, we chose to use the interception of the two Gaussian distribution curves as the threshold to separate the RTT-like phenotype (Mecp2 ^−/Y[R]) from the normal (Mecp2 ^−/Y[N]), which is 0.2 in f variation as shown in Fig. 2b. Statistically, the f variation of Mecp2 ^−/Y[R] mice was not different from that of RTT-like phenotype mice isolated by the k-means clustering (Mecp2 ^−/Y[r]) (Fig. 2d). So was Mecp2 ^−/Y[N] mice in comparison with Mecp2 ^−/Y[n] mice. Therefore, the interception of the two Gaussian distribution curves seems to be a useful way to separate the Mecp2 ^+/− females with different breathing f variations.

With such a pooling, 10 of 75 female Mecp2 ^+/− mice tested (13 %) were separated from the rest. These Mecp2 ^−/Y[R] mice had an f variation index of 0.24 that was highly significantly different from the WT (f variation index = 0.11). The f variation of the Mecp2 ^−/Y[R] had no significant difference from the Mecp2 ^−/Y males (f variation index = 0.27) (Fig. 2e). The f variation index of the other group of Mecp2 ^+/− females, Mecp2 ^+/− [N] (0.10), resembled more the WT (Fig. 2e). Similar results were obtained by pooling Mecp2 ^+/− mice using the k-means clustering, in which 13 (17 %) showed significantly high f variation (Fig. 2f). Statistical analysis of these two results showed no significant difference (P > 0.05, χ ² test) although the k-means clustering seems more sensitive.

Based on these results obtained with two independent methods, it is likely that the Mecp2 ^+/− females are not made of a homogenous group regarding breathing irregularity, and approximately 15 % of Mecp2 ^+/− females show the RTT-like phenotype.

Apnea

A prolonged period of breathing cessation known as apnea is a hallmark of breathing dysfunction in RTT patients. Although apneas are seen in the WT mice as well, the frequency of the apneas was extremely low. In contrast, Mecp2 ^−/Y mice displayed high frequencies of apneas. Therefore, the apnea count was measured as number of events per hour (apneas/h), and used as another indication of the severity of breathing abnormality. Apnea was defined as at least one breath cycle missing as described above. The entire recording period (30 min) under normal air ventilation was used to analyze the number of apneas.

Similar to f variations, apnea counts showed normal (Gaussian) distributions in the WT and Mecp2 ^−/Y mice, in which most WT and Mecp2 ^−/Y mice were separate from each other, although overlaps were also found (Fig. 3a). At the intercept of two Gaussian distribution curves, a threshold value (38 apneas/h) was obtained. Using the threshold value, the Mecp2 ^+/− females were divided into two groups (Fig. 3b). Of 75 female Mecp2 ^+/− mice tested, 17 (23 %) were isolated as Mecp2 ^+/−[R] with RTT-like phenotype from the rest (Mecp2 ^+/−[N]). A similar number (17, 23 %) was isolated using the k-means clustering method (Fig. 3c). Statistical analysis of these two results showed no significant difference (P > 0.05, χ ² test). Consistently, apnea counts of the Mecp2 ^+/−[R] were not significantly different from Mecp2 ^+/−[r] (P > 0.05, Fig. 3d).

Fig. 3

Separation of Mecp2 ^+/− mice based on number of apneas. a Both WT (n = 56, open circles) and Mecp2 ^−/Y (n = 13, black triangles) mice show typical Gaussian distributions of the number of apneas. b Mecp2 ^+/− mice are represented as black squares. The vertical line represents the separation level at 38 apneas/h. c Mecp2 ^+/− mice were separated based on the k-means clustering method. RTT-like Mecp2 ^+/− mice (n = 17) are represented as black triangles, and WT-like Mecp2 ^+/− (n = 58) mice are represented as open triangles. d RTT-like Mecp2 ^+/− and WT-like Mecp2 ^+/− mice indicated by the threshold level show no significant differences from those indicated by the k-means clustering method. e When separated using the determined threshold level, Mecp2 ^−/Y mice and the RTT-like Mecp2 ^+/− mice have no significant difference and WT (n = 56) and the WT-like Mecp2 ^+/− mice (n = 58) have no significant difference. Data is presented as median ± IQ (*P < 0.05, ***P < 0.005). f When separated based on the number of apneas using the k-means clustering method, Mecp2 ^−/Y (n = 13) mice and the RTT-like Mecp2 ^+/− mice (n = 17) have no significant difference and WT (n = 56) and the WT-like Mecp2 ^+/− mice (n = 58) have no significant difference. Data is presented as median ± IQ (*P < 0.05, ***P < 0.005)

After the median and IQR were calculated, a Kruskal–Wallis test showed that the Mecp2 ^+/−[R] females had no significant difference from Mecp2 ^−/Y mice, but were significantly different from the WT in their medians (Fig. 3e); and vice versa for the rest of the WT-like Mecp2 ^+/− females, i.e., Mecp2 ^+/−[N]. Separation of apneas/h with the k-means clustering analysis was identical to the threshold results (Fig. 3f).

Since Mecp2 ^+/− mice with significant f variations were not completely identical to those with severe apneas, Mecp2 ^+/− mice with either alone could have a mild breathing abnormality. Therefore, we further examined these mice. Using k-means clustering analysis, 9 of 75 Mecp2 ^+/− mice (Mecp2 ^+/− [v]) showed significant f variation and apneas (Fig. 4a, b). With respect to f variation and apnea, the Mecp2 ^+/− [v] mice were not significantly different from the Mecp2 ^+/−[R] mice, but significantly different from the Mecp2 ^+/−[N] females (Fig. 4c, d).

Fig. 4

RTT-like Mecp2 ^+/− mice with both significant f variation and apneas. a, b Mecp2 ^+/− mice with significant f variations and apneas (Mecp2 ^+/−[v], large diamond) indicated by the k-means clustering analysis were compared with Mecp2 ^+/− mice showing significant f variation only (Mecp2 ^+/−[r], gray triangles). Regarding severity of f variations, Mecp2 ^+/−[v] mice were not significantly different from Mecp2 ^+/−[r]. c, d Severity of apnea was not significantly different between Mecp2 ^+/−[v] and Mecp2 ^+/−[r], either. Data are presented as mean ± SE (***P < 0.005) d Data is presented as median ± IQ (***P < 0.005)

Similar results were obtained based on the threshold levels determined, in which 7 of 75 Mecp2 ^+/− mice showed both significantly high f variation and apneas. The Mecp2 ^+/− [v] mice isolated with the threshold did not show more severe f variation and apnea than the Mecp2 ^+/−[R] mice, either (not shown). These results thus suggest that either f variation or apnea may be used to isolate Mecp2 ^+/− female mice with RTT-like breathing abnormalities from the rest of the females.

Age contribution

The severity of breathing disorders seen in Mecp2 ^+/− mice may change with the progression of age. Based on the threshold levels identified, age differences between Mecp2 ^+/− mice at ages of 3 weeks to 6 months were analyzed between RTT-like and normal Mecp2 ^+/− mice with respect to f variation and apnea (Fig. 5a, b). Both breathing abnormalities were seen in mice of 3 weeks of age, and the occurrence ratio of breathing abnormalities did not increase with aging (from 3 weeks to 6 months). Approximately 13 % of Mecp2 ^+/− females displayed RTT-like breathing f variations, and approximately 28 % of Mecp2 ^+/− mice showed RTT-like levels of apnea in this age range in these groups of mice, which were very close to our observations in the general mouse population above. The f variation and apnea counts were compared in the three age groups between RTT-like and normal Mecp2 ^+/− mice. The RTT-like Mecp2 ^+/− mice showed significantly higher levels of f variation and apnea counts than the normal Mecp2 ^+/− mice (Fig. 5c, d). Out of the 13 Mecp2 ^+/− mice that were retested, 5 initially displayed apnea counts above the determined threshold of 38 apneas/h, but upon subsequent testing did not display apnea numbers above the threshold. One Mecp2 ^+/− mouse did not display apneas above the threshold, but upon subsequent testing showed a number of apneas above the threshold. The other 7 Mecp2 ^+/−never displayed a number of apneas above the threshold (data not shown).

Fig. 5

Severity and occurrence rate of breathing irregularities in Mecp2 ^+/− mice. a The ratio of RTT-like Mecp2 ^+/− mice to WT-like Mecp2 ^+/− mice based on f variation separation during the different age groups. b The ratio of RTT-like Mecp2 ^+/− mice to WT-like Mecp2 ^+/− mice based on the number of apneas/h separation during the different age groups. c The severities of f variation of the Mecp2 ^+/− during the different age groups. Data is presented as mean ± SE (***P < 0.005). d The numbers of apneas/h of the Mecp2 ^+/− during the different age groups. Data is presented as median ± IQ (***P < 0.005)

Response to CO₂

To determine whether exposing RTT-like Mecp2 ^+/− mice to an elevated level of CO₂ reduces the severity of their breathing abnormalities, a concentration of 3 % CO₂ was delivered to the animals in a plethysmograph chamber. Apneas and the f variation were analyzed in the same way as described above. RTT-like Mecp2 ^+/− mice displayed a significant reduction in f variation when exposed to 3 % CO₂ (Fig. 6a). Their irregular breathing was resumed when the high CO₂ ventilation stopped. In normal air, the f variation was 0.24 in RTT-like Mecp2 ^+/− mice. These breathing variations were suppressed with 3 % CO₂, leading to an f variation of 0.14 (Fig. 6b), which was not significantly different from the baseline f variation (0.12) of WT mice. The RTT-like Mecp2 ^+/− mice also displayed a significant reduction in apneas from 56 counts/h to 18 counts/h when exposed to 3 % CO₂ (Fig. 6c, d). Frequent apneas returned when the plethysmograph chamber was ventilated with room air.

Fig. 6

Mecp2 ^+/− mice show reduced severities of breathing irregularities in response to 3 % CO₂. a Representative traces of a Mecp2 ^+/− mouse during normal air ventilation, 3 % CO₂, and normal air washout. b, c Statistical analysis was performed in Mecp2 ^+/− mice and WT mice. Significant suppression of f variation (b, c) and apnea (d, e) was found with 3 % CO₂. Normalized f variation (c) and apnea (e). Data is presented as mean ± SE (*P < 0.05) and median ± IQ (*P < 0.05, ***P < 0.005), respectively. BL baseline, WO washout, WT wild-type

Discussion

It is known that RTT-like breathing disorders start to manifest themselves in Mecp2 ^−/Y males at 3 weeks of age [10, 11], while a fraction of Mecp2 ^+/− mice also begin to display the same symptoms at this age as shown in the present study. Since most Mecp2 ^+/− mice do not develop breathing irregularities due to random X-chromosome inactivation, the demonstration of the breathing disorders in the present study should have an impact on the understanding of RTT with female animal models.

Like human patients with RTT, Mecp2 ^−/Y male mice show hypoventilation, apneas followed by brief hyperventilation and increased variability in breathing frequency [12]. Several different studies have shown the alleviation of these irregularities by increasing the availability of neurotransmitters or their precursors such as norepinephrine, GABA and serotonin [8, 13, 14]. Our previous studies have indicated that breathing patterns of the Mecp2 ^−/Y mice can improve in response to elevated CO₂. We have found that these mice lose their sensitivity to moderate hypercapnia, while their sensitivity to severe hypercapnia appears normal. The Mecp2 ^−/Y mice do not respond to moderate hypercapnia, and display hypoventilation and breathing irregularities as under normocapnic conditions. The breathing patterns may not allow an adequate clearance of CO₂, leading to a buildup of systemic CO₂ or the development of severe systemic hypercapnia. Because the mice have a decent sensitivity to severe hypercapnia, they then hyperventilate so that the excess CO₂ is removed from the body. These events seem to occur periodically in the Mecp2 ^−/Y mice as seen in patients with RTT [7].

In one study, the percentage of Mecp2 ^+/− mice that showed apnea levels significantly greater than WT mice were studied from 8 to 12 weeks of age. The percentage of these Mecp2 ^+/− increased from 20 % at 8 weeks to 50 % at 12 weeks of age. [15]. The occurrence of apneas unaffected by age were detected in female Mecp2 ^+/− mice 4–14 months old and periodic breathing defects have been demonstrated in 9-month-old female Mecp2 ^+/− mice [8]. Concerning the lifespan of mice, it is unclear whether the defects in the old female mice might involve biological processes that are rather different from the early development of RTT patients. A recent study has demonstrated a great number of apneas in Mecp2 ^+/− female mice when compared to WT [9], in which breathing irregularities were found in Mecp2 ^+/− mice approximately 2 months old. At this age, mice are fully mature, and Mecp2 ^−/Y male mice have a low survivability rate. Apparently, the age does not correlate well with the early onset of RTT in humans. In another study, breathing abnormalities were observed in Mecp2 ^+/− mice 6–17 months, which were reduced by the selective 5HT-1a agonist, F15599 [14].

Mecp2 ^+/− mice display a range of symptoms, while the severity of RTT-like abnormalities is typically more difficult to characterize compared to Mecp2 ^−/Y males that have a clean knockout of the Mecp2 gene with more obvious symptoms. To identify Mecp2 ^+/− females with breathing abnormalities we have studied the distribution patterns using two independent methods. Our data suggest that two populations of Mecp2 ^+/− mice exist, one with a phenotype resembling WT and another displaying a significant RTT-like breathing phenotype. These two populations can be separated based on objective and quantifiable measures of breathing parameters. With these measures, a substantial number of Mecp2 ^+/− mice display breathing abnormalities as severe as those seen in Mecp2 ^−/Y males. These breathing abnormalities, though, appear to be less severe in the general population of Mecp2 ^+/− mice by comparison to Mecp2-null males due to random X-inactivation and varying MeCP2 protein levels.

Our results have shown that severe breathing defects take place as early as 3 weeks postnatal age in female Mecp2 ^+/− mice. These abnormalities were not studied at an earlier age because of technical limitations in handling premature pups of both sexes. Despite this, our results suggest that the occurrence age for the breathing abnormalities seems comparable to that of Mecp2 ^−/Y males. About 13 % of Mecp2 ^+/− mice develop irregular breathing, less than 25 % of mice show frequent apneas, and ~10 % of females have both defects. Since none of these figures are close to the prediction of random X-inactivation (~50 %), it is possible that the Mecp2 gene is not completely inactivated in 50 % of female heterozygous mice.

Our previous studies have shown that Mecp2 ^−/Y mice have impaired central chemosensitivity. The mice do not respond to moderate CO₂ levels, but their sensitivity to high PCO₂ is normal. This defect leads to periodic hyper- and hypoventilation as seen in RTT patients [7]. Hypoventilation, frequent apneas and irregular/ineffective breathing tend to produce systemic hypoxia, while hypoxia is known to be a major risk factor for the maldevelopment of the central nervous system that occurs in RTT patients. Hypoxia has also been shown to cause vasoconstriction by the suppression of Kv2.1 channels [16]. Similar to Mecp2 ^−/Y mice, the Mecp2 ^+/− female mice with severe RTT-like symptoms respond well to the hypercapnic challenge. Their breathing becomes regular with a significant reduction in apnea events when their breathing air contains 3 % CO₂. Therefore, the application of high CO₂ may alleviate the deleterious consequences of breathing abnormalities in RTT patients. Interestingly, female Mecp2 ^+/− mice between 4 months and 19 months are found to have a higher hypercapnia-response threshold, and the CO₂ chemosensitivity is improved in Mecp2-null male mice by increasing the availability of serotonin [17]. Together, these findings suggest the feasibility of correcting the breathing disorders with a CO₂ intervention, which seems useful in the therapeutic design for RTT, as the consequence of the breathing disorders may not be limited to systemic hypercapnia.

Conclusions

In conclusion, Mecp2 ^+/− female mice can be separated into two groups resembling WT and Mecp2 ^−/Y males, respectively. In mice with RTT-like disorders, breathing irregularities begin to manifest themselves during the early lives of the Mecp2 ^+/− mice before reaching adulthood. In the other group, the breathing phenotype remains similar to WT mice. We have shown that breathing abnormalities in Mecp2 ^+/− mice occur earlier than previous studies have shown and the occurrence rate is less than 20 %.The variable severities and occurrence rate of breathing irregularities in the general population of Mecp2 ^+/− mice may be due to random X-inactivation, resulting in variable amounts of MeCP2 protein. These breathing irregularities can be largely reduced under hypercapnic ventilation. Therefore, these findings indicate that the severity of breathing abnormalities can be reliably distinguished from non-disease breathing phenotypes and CO₂ intervention can reduce the severity of these breathing defects.