Defects in brainstem neurons associated with breathing and motor function in the Mecp2^R168X/Y mouse model of Rett syndrome

Abstract

Rett Syndrome (RTT) is an X-linked neurodevelopmental disorder caused mostly by disruption of the MECP2 gene. Among several RTT-like mouse models, one of them is a strain of mice that carries an R168X point mutation in Mecp2 and resembles one of the most common RTT-causing mutations in humans. Although several behavioral defects have previously been found in the Mecp2^R168X/Y mice, alterations in nerve cells remain unknown. Here we compare several behavioral and cellular outcomes between this Mecp2^R168X/Y model and a widely used Mecp2^Bird/Y mouse model. With lower body weight and shorter lifespan than their wild-type littermates, the Mecp2^R168X/Y mice showed impairments of breathing and motor function. Thus we studied brainstem CO₂-chemosensitive neurons and propriosensory cells that are associated with these two functions, respectively. Neurons in the locus coeruleus (LC) of both mutant strains showed defects in their intrinsic membrane properties, including changes in action potential morphology and excessive firing activity. Neurons in the mesencephalic trigeminal nucleus (Me5) of both strains displayed a higher firing response to depolarization than their wild-type littermates, likely attributable to a lower firing threshold. Because the increased excitability in LC and Me5 neurons tends to impact the excitation-inhibition balances in brainstem neuronal networks as well as their associated functions, it is likely that the defects in the intrinsic membrane properties of these brainstem neurons contribute to the breathing abnormalities and motor dysfunction. Furthermore, our results showing comparable phenotypical outcomes of Mecp2^R168X/Y mice with Mecp2^Bird/Y mice suggest that both strains are valid animal models for RTT research.

rett syndrome (rtt) is an X-linked neurodevelopmental disorder with a morbidity rate of about 0.01% in female live births. In addition to autistic features, patients with RTT typically exhibit symptoms of breathing dysfunction and motor impairment. Approximately 90% of cases of RTT are caused by disruptions in the gene encoding the methyl-CpG-binding protein 2 (MeCP2) protein, a transcriptional regulator.

Since the discovery of the defective MECP2 gene in the development of RTT, several mouse models have been created, which recapitulate many RTT features (10–12, 41). One of the most commonly used mouse models in RTT research is the strain of Mecp2^tm1.1Bird/J mice with Mecp2 knockout. Also known as Mecp2^Bird mice, they carry a large deletion of exons 3 and 4, rendering the protein nonfunctional (12). The present understanding of the effects of the Mecp2 disruption on RTT-like symptoms such as breathing abnormality, motor dysfunction, cognitive impairment, and early death is mainly derived from this mouse model. Although these findings made in Mecp2^Bird mice are important, they need to be validated in other animal models. This is particularly significant when the ultimate goal of RTT research is to formulate therapeutic modalities.

Unlike engineered gene disruptions, several naturally occurring Mecp2 mutations may cause Mecp2 disruptions to different degrees. One of them is the R168X point mutation (38). The R168X mutation, found in 8–12% of patients with RTT as one of the most common RTT-causing mutations, results in a premature truncation of the MeCP2 protein that presumably might retain some functionality (25, 44). Therefore, it is necessary to study cellular and molecular mechanisms for RTT in animal models carrying such a mutation. Indeed, a mouse model with the R168X mutation in the Mecp2 gene (Mecp2^R168X mice) has been developed recently (19, 30). Studies in this mouse model have revealed several behavioral defects including motor function, respiration, cognition, and anxiety (2, 5, 30). These findings are encouraging, as they warrant further studies on the cellular consequences of the R168X mutation.

Several previous studies have shown that Mecp2 disruption alters the excitation-inhibition balance of neuronal networks and meanwhile reduces the availability of certain neurotransmitters in the CNS (9, 15). In locus coeruleus (LC) neurons, the main providers of norepinephrine (NE) in the CNS, we have found that the Mecp2^Bird mice show cellular dysfunction manifested as defects in intrinsic properties, synaptic inputs, NE-biosynthesis enzyme expression, and CO₂ chemosensitivity (16, 24, 28, 33, 47–49). These as well as the fact that the LC neurons play a role in central CO₂ chemoreception and breathing regulation suggest that validating some of these findings in another mouse model with a natural Mecp2 disruption such as the R168X mutation may benefit the understanding of the neuronal basis for autonomic dysfunction in RTT.

Because the R168X mutation may potentially result in a partially functional protein product as opposed to the nonfunctional MeCP2 protein in Bird mice, it is possible that the phenotype of the R168X mouse model is less severe than the Bird mouse model. To elucidate how the R168X mutation affects neuronal intrinsic properties and the severity of RTT-like symptoms, we compared breathing abnormalities and motor dysfunction, two of the most consistent and reproducible phenotypes, between Mecp2^R168X and Mecp2^Bird mice and studied the cellular mechanisms associated with these phenotypes. Our results indicated that Mecp2^R168X and Mecp2^Bird mice share most RTT-like defects at behavioral and cellular levels, suggesting that the partially truncated MeCP2 protein with the R168X mutation may not be functional, at least in two types of brainstem neurons and their associated functions.

MATERIALS AND METHODS

Animals.

Experiments were performed in male Mecp2^R168X/Y and Mecp2^Bird/Y mice because the male mice with Mecp2 disruption offer a reliable Mecp2-defective condition that is not always available in heterozygous females because of the uncontrolled X-chromosome inactivation (6, 43). We also used symptomatic and asymptomatic females to validate our findings from male mice. To breed the Mecp2^R168X mice, female heterozygous Mecp2^R168X mice were purchased from the Jackson Laboratory (B6J;129S6.MeCP2^R168X, stock no. 024990; Bar Harbor, ME) and crossbred with WT males with the same genetic background. Offspring were genotyped using the PCR protocol from Jackson Laboratory. Female heterozygous Mecp2^tm1.1Bird mice [strain name: B6.129P2(C)-Mecp2^tm1.1Bird/J, stock no: 003890, Jackson Laboratory] were crossbred with wild-type (WT) C57BL/6 males to produce Mecp2^Bird/Y male offspring. Male WT littermates of the Mecp2^R168X/Y and Mecp2^Bird/Y mice served as controls. All experimental procedures in the animals were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Georgia State University Institutional Animal Care and Use Committee.

Plethysmography.

Breathing activity was recorded from conscious mice without anesthesia, as we did previously (50). In the experiment, a mouse was placed in an ∼50-ml plethysmograph chamber. The mouse was given normal room air at a flow rate of 20 ml/min and allowed to acclimate to the chamber for 10 min before the recording. The mouse was monitored using a video camera during a 15-min recording. Frequency variation was calculated using at least 200 breaths by analyzing three or four stretches of at least 50 consecutive breaths. These stretches were randomly sampled with approximately equal length. Apneas were counted during the entire duration of the recording. A breath was considered an apnea if the breath lasted more than twice the duration of the previous breath. Movement and potential sleep were checked in the video records. Periods of movement and sleep were excluded from analysis. Data were stored and analyzed offline using Clampfit 10.3 software (Molecular Devices, Sunnyvale, CA).

Grip-strength and grid-walking tests.

In the grip-strength test, an individual mouse was held by its tail to allow the forelimbs to grasp a metal bar connected to a force-electricity transducer (CB Sciences, Milford, MA). The mouse was slowly pulled up vertically until the grip was released. This was repeated three times for each mouse. The maximal force reached was measured as the grip strength. The measurements were converted to grams, and all three trials were averaged.

In the grid-walking test, a mouse was placed in a rigid box (32 cm × 20 cm × 20 cm), which was elevated to 50 cm high. The device had a metal mesh floor with 11 × 11 mm openings. The mouse walking was recorded using a video camera for 5 min. Forepaw and hindpaw foot faults were counted when a limb missed the metal floor bar (0.5 mm in diameter) completely and went through the grid hole. The foot fault ratio was calculated by the overall foot faults divided by the total steps (including both forelimbs and hindlimbs).

Survival rate.

All animals counted for survival rate were kept in the same housing condition as for other mice. Their survival time was counted from birth to in-cage death or when the mice reached a humane endpoint that was determined by a staff member in the animal facility at Georgia State University.

Brain slice.

Brain slices were prepared as described previously (47). In brief, mice were anesthetized with isofluorane. The brainstem was obtained rapidly and placed in an ice-cold, sucrose-rich artificial cerebrospinal fluid (sucrose aCSF) containing the following (in mM): 220.0 sucrose, 1.9 KCl, 0.5 CaCl₂, 6.0 MgCl₂, 33.0 NaHCO₃, 1.2 NaH₂PO₄, and 10.0 d-glucose. The solution was bubbled with 95% O₂ balanced with 5% CO₂ (pH 7.40). Transverse pontine sections (180–250 μM) containing the LC area were obtained using a vibratome sectioning system (1000 Plus; Vibratome, St. Louis, MO). The slices were transferred to normal aCSF containing the following (in mM): 124.0 NaCl, 3.0 KCl, 2.0 CaCl₂, 2.0 MgCl₂, 26.0 NaHCO₃, 1.3 NaH₂PO₄, and 10.0 D-glucose; they were allowed to recover at 33°C for 1 h and then kept in room temperature before being used for recording. One slice was transferred to a recording chamber that was perfused with oxygenated aCSF at a rate of 2 ml/min and maintained at 32–35°C (TC-344B; Warner Instruments, Hamden, CT). LC neurons were identified as described previously (17).

Electrophysiology.

Whole cell current-clamp studies were performed on cells visualized using a near-infrared charge-coupled device camera. Patch pipettes were pulled with a Sutter pipette puller (model P-97; Sutter Instruments, Novato, CA) with a resistance of 3–6 MΩ. The slice was perfused with normal aCSF and superfused with 95% O₂-5% CO₂ at 34°C. The internal (pipette) solution for current clamp contained the following (in mM): 130.0 potassium gluconate, 10.0 KCl, 10.0 HEPES, 2.0 Mg-ATP, 0.3 Na-GTP, and 0.4 EGTA (pH 7.30). Recorded signals were amplified with a Multiclamp 700B amplifier (Molecular Devices), digitized at 10 kHz, filtered at 1 kHz, and collected with the Clampex software. The temperature was maintained at 33°C during recording by a dual automatic temperature control (Warner Instruments).

To select healthy LC neurons, only cells with resting membrane potentials more negative than −40 mV and action potential (AP) amplitudes >65 mV were analyzed. All AP characteristics of LC neurons were determined after averaging at least 30 APs without any stimulus currents. AP amplitude was measured from threshold to peak. AP threshold was determined using the maximum second derivative method (31). To determine the health of Me5 neurons, cells with stable resting membrane potentials more negative than −50 mV and APs with amplitudes >55 mV were analyzed. AP characteristics of Me5 neurons were analyzed from evoked APs.

Data analysis.

The electrophysiological data were analyzed with the Clampfit 10.3 software. Data are presented as means ± SE. Statistical analysis of differences was performed using the two-tailed Student's t-test. Difference was considered significant when P ≤ 0.05.

RESULTS

Observable RTT-like characteristics.

Like Mecp2^Bird/Y mice, all Mecp2^R168X/Y mice displayed stereotypical hindlimb clasping (Fig. 1A). They also had a smaller body size than WT male littermates (Fig. 1B). When the body weight was compared, Mecp2^R168X/Y mice weighed much less (8.9 ± 0.7 g, n = 13) compared with their WT littermates at 4 wk of age (17.1 ± 1.1 g, n = 7; P < 0.001), which resembled Mecp2^Bird/Y mice in the same age (20.9 ± 1.7 g, n = 9) vs. their WT littermates (27.8 ± 1.2 g, n = 8; P < 0.01) (Fig. 1C₁). Note that Mecp2^R168X/Y mice were generally smaller than Mecp2^Bird/Y mice attributable to perhaps their genetic background as previously demonstrated to influence body weight (29). To eliminate the effect of genetic background, Mecp2^R168X/Y and Mecp2^Bird/Y were normalized to their WT littermates. After normalization, Mecp2^Bird/Y showed a significantly higher body weight (0.8 ± 0.1, n = 9) than the Mecp2^R168X/Y mice (0.5 ± 0.0, n = 13; P < 0.01) (Fig. 1C₂).

Fig. 1.

General abnormalities of Mecp2^R168X/Y mice. A: photographs of a male Mecp2^R168X/Y (R168X) mouse and its wild-type (WT) littermate at 5 wk of age. The R168X mouse displayed hindlimb clasping, and the WT littermate did not. B: male R168X mouse was smaller than its WT littermate. C₁: both R168X and Mecp2^Bird/Y (Bird) mice had lower body weight compared with their WT littermates. C₂: Bird mice weighed significantly more than R168X mice after normalization to the average body weight of WT littermates. D: survival curves of R168X and WT were normalized to 100% with the 50% death rate indicated by the vertical line. E: survival curves of Bird and WT were normalized to 100% with the 50% death rate indicated by the vertical line. **P < 0.01, ***P < 0.001 (two-tail Student's t-test). Data are presented as means ± SE.

We also compared the lifespans of these two strains of mice under the same housing condition. The lifespan of 22 Mecp2^R168X/Y mice was recorded and compared with 18 of their WT littermates. The Mecp2^R168X/Y mice started dying at 33 days of age, and 50% of them died on day 52 (Fig. 1D). In contrast, none of the WT mice died in the study period (3 mo). The 50% death rate of Mecp2^Bird/Y mice was 55 days (Fig. 1E). When the ratio of survival vs. death was compared with the χ² test, we did not find any statistical difference between Mecp2^R168X/Y and Mecp2^Bird/Y mice (P > 0.05).

Breathing disturbances.

Breathing disturbances are some of the major symptoms found in human patients with RTT and mouse models. We have previously shown that breathing frequency variation and apnea rate can be reproducibly measured and compared between Mecp2^Bird/Y and WT mice (48, 50). Using these measures, we tested and compared breathing disturbances in Mecp2^R168X/Y and Mecp2^Bird/Y mice.

Both Mecp2^R168X/Y and Mecp2^Bird/Y mice displayed frequent hyper- and hypoventilation (Fig. 2, A and B). Quantitatively, the breathing irregularity was analyzed by averaging >200 breaths. The obtained standard deviation was then divided by the arithmetic mean as an indication of breathing frequency variation (48). The Mecp2^R168X/Y mice had significantly higher breathing frequency variation (0.33 ± 0.05, n = 7) than their WT littermates (0.11 ± 0.02, n = 5; P < 0.01), which was comparable to Mecp2^Bird/Y (0.27 ± 0.02, n = 11) vs. their WT littermates (0.11 ± 0.02, n = 11; P < 0.001) (Fig. 2C₁). After normalization to WT levels, no significant difference was found between Mecp2^R168X/Y (2.9 ± 0.5, n = 7) and Mecp2^Bird/Y mice (2.4 ± 0.2, n = 11; P > 0.05) (Fig. 2C₂).

Fig. 2.

Breathing abnormalities. A₁-B₂: traces of breathing activity from male Mecp2^R168X/Y (R168X), Mecp2^Bird (Bird), and WT mice. R168X and Bird mice displayed periods of slow and fast breathing as well as apneas, whereas the WT mice did not. C₁: breath frequency (f) variation was significantly higher in R168X mice than in their WT littermates, as was found in Bird mice. C₂: there was no significant difference in f variation between R168X and Mecp2^Bird/Y (Bird) mice after normalization to their WT littermate levels. D₁: apnea counts shown as apneas/h were significantly higher in both R168X and Bird mice compared with their WT littermates. D₂: no significant difference in apnea counts was found between R168X and Bird mice after normalization. **P < 0.01; ***P < 0.001; NS, P > 0.05 (two-tail Student's t-test).

Frequent apneas were observed in both strains of mice with Mecp2 disruption. Apnea counts were higher in Mecp2^R168X/Y (130.3 ± 31.3 apneas/h, n = 7) than in their WT littermates (1.2 ± 0.8 apneas/h, n = 5; P < 0.01). Similar breathing abnormality was seen in Mecp2^Bird/Y (123.5 ± 8.9 apneas/h, n = 11) over the WT (23.9 ± 7.3 apneas/h, n = 11; P < 0.001) (Fig. 2D₁). Because the average apnea rate in some of the WT was too low (nearly zero) to make regular normalization, we analyzed the difference in apnea counts of Mecp2^R168X/Y and Mecp2^Bird/Y mice vs. their WT littermates. In the study, apnea counts from each mutant mouse in both strains were subtracted by the population mean of apnea counts of WT mice, which were then averaged and compared between Mecp2^R168X/Y (130.3 ± 31.3 apneas/h, n = 7) and Mecp2^Bird/Y mice (123.5 ± 8.9 apneas/h, n = 11). No statistical difference was found with such a comparison (P > 0.05) (Fig. 2D₂).

Muscle strength and motor coordination.

Another major feature of RTT is motor dysfunction seen in both human patients and mouse models. The motor dysfunction manifests itself as impairment in muscle strength and motor coordination (13). Thus we assessed the muscle tone and motor coordination, using the grip-strength test and the grid-walking test. At 4–5 wk of age, the grip strength of Mecp2^R168X/Y (48.6 ± 4.4 g, n = 12) was significantly lower than their WT littermates (60.3 ± 4.5 g, n = 7; P < 0.05) (Fig. 3A₁). Similar reduction in grip strength was observed in Mecp2^Bird/Y mice (57.7 ± 3.4 g, n = 12) compared with their WT littermates (89.6 ± 3.3 g, n = 7; P < 0.001). There was no significant difference between Mecp2^R168X/Y and Mecp2^Bird/Y mice when the data were normalized to the WT values (0.8 ± 0.1, n = 12 in Mecp2^R168X/Y; 0.6 ± 0.0, n = 12 in Mecp2^Bird/Y; P > 0.05) (Fig. 3A₂).

Fig. 3.

Reduced motor function. A₁: Mecp2^R168X/Y (R168X) mice displayed lower grip strength compared with their WT littermates, as did Mecp2^Bird/Y (Bird) mice. A₂: grip strength of R168X mice was not significantly different from Bird mice after normalization to WT levels. B₁: compared with WT mice, both R168X and Bird mice displayed a higher percentage of foot faults in the grid-walking test. B₂: after normalization, foot faults between R168X and Bird mice were not significantly different from each other. *P < 0.05; **P < 0.01; ***P < 0.001; NS, P > 0.05 (two-tail Student's t-test).

These mice with Mecp2 disruption also showed a higher percentage of foot faults than their WT littermates. The foot faults averaged 8.8 ± 1.1% (n = 6) in Mecp2^R168X/Y and 4.6 ± 0.9% (n = 6) in their WT littermates (P < 0.05) (Fig. 3B₁). Similarly Mecp2^Bird/Y mice showed 5.5 ± 0.9% (n = 6) foot faults compared with the WT (2.1 ± 0.2%, n = 6; P < 0.01). After normalization to the WT average, however, there was no significant difference between Mecp2^R168X/Y (1.9 ± 0.2, n = 6) and Mecp2^Bird/Y mice (2.7 ± 0.4, n = 6; P > 0.05) (Fig. 3B₂).

Passive membrane properties of LC neurons.

Human patients with RTT and mouse models show autonomic dysfunction, especially breathing abnormalities known to involve the NE system in the CNS (15, 34, 51). In the CNS, most NE modulation is derived from the LC, where neurons show abnormal intrinsic membrane properties in the Mecp2^Bird/Y mice (47). The demonstration of significant breathing abnormalities in Mecp2^R168X/Y mice opened a question as to how the Mecp2^R168X/Y mutation affects membrane properties of LC neurons. Thus we performed whole cell recordings in brain slices and compared the intrinsic membrane properties of these cells between the mutant mice and their WT littermates.

LC neurons from Mecp2^R168X/Y and WT mice showed a typical nonlinear relationship of membrane potentials vs. current injections (I–V) with no postinhibition rebound (Fig. 4, A and B). The membrane potentials of LC neurons in Mecp2^R168X/Y averaged −43.0 ± 1.1 mV, n = 14 compared with cells from WT mice (−47.5 ± 1.8 mV, n = 14; P < 0.05). These resembled LC neurons in Mecp2^Bird/Y mice (−45.0 ± 1.1 mV, n = 7) vs. the WT (−49.4 ± 1.0 mV, n = 5; P < 0.05) (Fig. 4C₁). When normalized to WT levels, Mecp2^Bird/Y cells did not show a significant difference (0.9 ± 0.0, n = 7) from the Mecp2^R168X/Y (0.9 ± 0.0, n = 14; P > 0.05) in membrane potentials (Fig. 4C₂). The average input resistance of LC neurons was not significantly different between WT and Mecp2^R168X/Y or Mecp2^Bird/Y mice although neurons in null mice tended to have a higher input resistance (Fig. 4D₁). There was no significant difference in input resistance between Mecp2^R168X/Y and Mecp2^Bird/Y mice after normalization (Fig. 4D₂).

Fig. 4.

Altered passive membrane properties of locus coeruleus (LC) neurons. A₁ and A₂: whole cell current-clamp recordings from R168X and WT LC neurons (top) with steps of hyperpolarizing current injections (bottom). B: I–V plots of the LC neurons in A₁ and A₂ that show inward rectification (IR). C₁: LC cells from R168X mice displayed a more depolarizing membrane potential (V_m) than the WT cells, which was also seen in LC cells from Bird mice. C₂: there was no significant difference in resting membrane potential between R168X and Bird mice after normalization to WT levels. D₁ and D₂: input resistance (R_m) was not significantly different in R168X and Bird mice compared with input resistance in WT mice before and after normalization to WT levels. E₁: IR ratio was measured as the input resistance at −70 mV divided by that at −100 mV, which was significantly higher in R168X and Bird mice than their WT littermates. E₂: there was no significant difference in IR ratio between the two strains of mutant mice after normalization to WT values. *P < 0.05; NS, P > 0.05 (two-tail Student's t-test).

The nonlinear I–V relationship indicated that these cells also have inward rectification (IR). We have previously shown that, in male Mecp2^Bird/Y mice, LC cells have enhanced IR attributable to increased Kir 4.1 expression, which likely contributes to the resting membrane potential and dampens neuronal hyperexcitability (47). Furthermore, enhanced IR may minimize inhibitory input to the cell to correct its hyperexcitability. The IR ratio was thus calculated by dividing input resistance at −70 mV by that at −100 mV as described previously (47). The rectification ratio of LC neurons was higher in Mecp2^R168X/Y mice (2.0 ± 0.2, n = 15) than in WT (1.5 ± 0.1, n = 14; P < 0.05), which were similar to Mecp2^Bird/Y (1.8 ± 0.1, n = 7) vs. their WT littermates (1.2 ± 0.1, n = 6; P < 0.05) (Fig. 4E₁). However, the rectification ratio of LC neurons was not significantly different between Mecp2^R168X/Y and Mecp2^Bird/Y mice after normalization (Fig. 4E₂).

LC neuron excitability.

Increased LC neuronal hyperexcitability has been found in Mecp2^Bird/Y mice, which is likely to contribute to defects in the NE system in the mice (33, 47). Therefore, we measured the spontaneous firing activity as well as other active membrane properties of LC cells in these mice. The spontaneous firing activity of LC neurons from Mecp2^R168X/Y mice (6.6 ± 0.7 Hz, n = 14) was significantly higher compared with their WT littermates (3.9 ± 0.3 Hz, n = 14; P < 0.01). These were similar to Mecp2^Bird/Y mice (6.8 ± 0.5 Hz, n = 7) vs. their WT littermates (3.4 ± 0.5 Hz, n = 5; P < 0.01) (Fig. 5, A–C₁). After normalization to the WT values, the firing rates were not significantly different between Mecp2^R168X/Y and Mecp2^Bird/Y mice (Fig. 5C₂).

Fig. 5.

Active membrane properties of LC neurons. A and B: recordings from spontaneously firing LC neurons in WT and R168X mice shown in the same time and amplification scales. C₁: spontaneous firing rate (FR) of LC cells from both strains of mutant mice was significantly higher than their WT littermates. C₂: normalized spontaneous FR was not significantly different between R168X and Bird mice. D and E: both peak frequency (F_p) and steady-state frequency (F_s) were higher in R168X and Bird mice than WT cells. F: normalized F_p and F_s were not significantly different between R168X and Bird mice. G: spike frequency adaptation (SFA) was calculated as F_p/F_s. Compared with WT cells, neither R168X nor Bird cells showed a significant difference. H₁: average amplitudes of slow afterhyperpolarization (sAHP) of R168X and Bird cells were not significantly different from WT cells. H₂: normalized sAHP amplitude was not significantly different between R168X and Bird mice. I₁: average sAHP area of R168X and Bird cells was not significantly different from WT cells. I₁: normalized sAHP area was not significantly different between R168X and Bird mice. *P < 0.05; **P < 0.01; NS, P > 0.05 (two-tail Student's t-test).

Spike frequency adaptation (SFA) is a response of neurons to continuing excitatory inputs. In the present study, the SFA was calculated as the firing frequency of the first two APs as the peak state (F_p) divided by the firing frequency of the APs at steady state (F_s). Both F_p (34.6 ± 3.9 Hz, n = 8) and F_s (34.2 ± 2.4 Hz, n = 8) were significantly higher in Mecp2^R168X/Y mice than in their WT littermates (24.1 ± 1.4 Hz, n = 8 and 22.2 ± 1.4 Hz, n = 8, respectively) (Fig. 5D). Similarly, LC neurons from Mecp2^Bird/Y mice also showed significantly higher F_p and F_s than WT cells (Fig. 5E). Neither F_p nor F_s showed a significant difference between Mecp2^Bird/Y and Mecp2^R168X/Y neurons (Fig. 5F). SFA also was not significantly different between Mecp2^R168X/Y, Mecp2^Bird/Y, and their WT littermates (Fig. 5G).

Afterhyperpolarization (AHP) following each AP contributes to setting the firing rate of neurons. Because we observed changes in spontaneous firing rate, we studied AHP characteristics. The AHPs consisted of fast (fAHP) and slow (sAHP) components. AHP amplitude was measured from the AP threshold level to the lowest hyperpolarizing potential of the AHP. We did not find any difference in the proportions of cells with fAHP and sAHP between the mutant strains and their WT littermates. There was no significant difference in sAHP amplitude between Mecp2^R168X/Y (−24.0 ± 2.2 mV, n = 8) and WT cells (−26.0 ± 2.1 mV, n = 8; P > 0.05) (Fig. 5H₁). The average sAHP amplitude between Mecp2^Bird/Y (−26.0 ± 1.4 mV, n = 8) also was not significantly different compared with WT cells (−26.0 ± 1.6 mV, n = 6; P > 0.05) (Fig. 5H₁). After normalization, Mecp2^R168X/Y (0.9 ± 0.1, n = 8) and Mecp2^Bird/Y (1.0 ± 0.1, n = 8; P > 0.05) were not significantly different (Fig. 5H₂). We also compared the sAHP area and found no significant difference between Mecp2^R168X/Y (1,731.2 ± 395.1 mV × ms, n = 8) and WT cells (1,976.9 ± 248.9 mV × ms, n = 8; P > 0.05) (Fig. 5I₁). There was also no significant difference between Mecp2^Bird/Y (2,368.6 ± 311.3 mV × ms, n = 8) and WT cells (3,179.9 ± 603.3 mV × ms, n = 6; P > 0.05) (Fig. 5I₁). Mecp2^R168X/Y (0.9 ± 0.2, n = 8) and Mecp2^Bird/Y (0.7 ± 0.1, n = 8; P > 0.05) were not significantly different after normalization (Fig. 5I₂).

LC AP characteristics.

AP width was visibly different in both Mecp2^R168X/Y and Mecp2^Bird/Y compared with WT cells (Fig. 6, A and B). AP amplitude of LC neurons from Mecp2^R168X/Y mice (78.0 ± 3.7 mV, n = 8) was not significantly different from that of WT mice (79.8 ± 4.6 mV, n = 8; P > 0.05) (Fig. 6C₁). AP amplitude of LC neurons from Mecp2^Bird/Y (90.7 ± 4.7 mV, n = 8) also was not significantly different from their WT littermates (89.2 ± 3.9 mV, n = 6; P > 0.05). When AP amplitude was normalized to WT levels, there was no significant difference between Mecp2^R168X/Y mice (1.0 ± 0.5, n = 8) and Mecp2^Bird/Y mice (1.0 ± 0.1, n = 8; P > 0.05) (Fig. 6C₂). AP threshold (−33.3 ± 0.9 mV, n = 8; P < 0.05) was significantly lower in Mecp2^R168X/Y mice than in their WT littermates (−29.1 ± 1.4 mV, n = 8) (Fig. 6D₁). The threshold of Mecp2^Bird/Y mice (−39.1 ± 0.8 mV, n = 8) was also significantly lower compared with their WT littermates (−33.5 ± 2.0 mV, n = 6; P < 0.05). After normalization, the AP threshold of Mecp2^R168X/Y mice (1.1 ± 0.0, n = 8) was not significantly different compared with Mecp2^Bird/Y mice (1.2 ± 0.0, n = 8; P > 0.05) (Fig. 6D₂).

Fig. 6.

LC neuron action potential (AP) morphology. A and B: representative action potentials from R168X, Bird, and WT littermate cells. C₁: action potential amplitudes were not significantly different between either R168X and WT cells or Bird and WT cells. C₂: normalized to WT levels, there was no significant difference between R168X and Bird cells. D₁: action potential threshold was significantly different between R168X and WT cells and between Bird and WT. D₂: action potential threshold of R168X and Bird mice was not significantly different after normalization. E and F: both the AP half-duration (APD₅₀) and rise time were significantly different when R168X and Bird cells were compared with their WT littermates. G: decay times of R168X and Bird mice were not significantly different from WT cells. H: after normalization, there were no significant differences between R168X and Bird mice in either APD₅₀ (D_1/2), rise time, or decay time. *P < 0.05; **P < 0.01; NS, P > 0.05 (two-tail Student's t-test).

AP width measured at 50% AP amplitude (APD₅₀) was significantly longer in Mecp2^R168X/Y (1.5 ± 0.1 ms, n = 8) compared with WT cells (1.1 ± 0.1 ms, n = 8; P < 0.05) (Fig. 6E). The APD₅₀ from Mecp2^Bird/Y mice (1.7 ± 0.1 ms, n = 8) was also significantly longer than that of WT cells (1.2 ± 0.1 ms, n = 6; P < 0.01). The rise times of Mecp2^R168X/Y and Mecp2^Bird/Y mice were significantly higher compared with their WT littermates (Fig. 6F). Although decay times were slightly higher in Mecp2^R168X/Y and Mecp2^Bird/Y mice compared with their WT littermates, they were not significantly different (Fig. 6G). After normalization to the WT values, there was no significant difference between Mecp2^R168X/Y and Mecp2^Bird/Y mice in APD₅₀, rise time, and decay time (Fig. 6H).

LC delayed excitation.

Delayed excitation (DE) is a characteristic property of LC neurons, which is described as a delayed occurrence of APs following hyperpolarization. This property, like IR, also affects how LC cells respond to inhibitory and excitatory input, further affecting their firing activity. DE was analyzed by fitting the time delay of the first AP with the Boltzmann equation D = 1/{1 + exp [−(V − V_½)/k]}, where D is the delay period, V is the hyperpolarizing membrane potential, V_½ is the half-inactivation, and k is the slope factor. All cells showed typical DE (Fig. 7, A and B). The V_½ of LC cells from Mecp2^R168X/Y mice was slightly more hyperpolarized than that of WT cells but was not significantly different (−70.2 ± 1.2 mV, n = 8, Mecp2^R168X/Y vs. −66.5 ± 3.6 mV, n = 8, WT; P > 0.05) (Fig. 7, C and E₁). The V_½ of Mecp2^Bird/Y cells was not significantly different from WT cells (−63.2 ± 3.1 mV, n = 8, Mecp2^Bird/Y vs. −64.9 ± 4.9 mV, n = 6, WT; P > 0.05) (Fig. 7, D and E₁). Mecp2^R168X/Y and Mecp2^Bird/Y cells were not significantly different after normalization (Fig. 7E₂). The slope factor also was not significantly different in Mecp2^R168X/Y cells (5.6 ± 0.6, n = 8, Mecp2^R168X/Y vs. 7.1 ± 0.9, n = 8, WT; P > 0.05) or Mecp2^Bird/Y cells (6.8 ± 1.3, n = 8, Mecp2^Bird/Y vs. 9.3 ± 1.2, n = 6, WT; P > 0.05) (Fig. 7F₁). Mecp2^R168X/Y and Mecp2^Bird/Y were not significantly different after normalization (Fig. 7F₂).

Fig. 7.

Delayed excitation (DE) in LC neurons. A and B: whole cell current-clamp recordings of LC neurons from R168X and WT mice, respectively. With a series of hyperpolarizing steps followed by a depolarizing pulse, the neurons showed DE. Arrows indicate the delay of the first spike that increased with increasing hyperpolarizing steps. C and D: relationship between action potential delay and membrane potential was fitted with the Boltzmann equation in representative R168X, Bird, and WT neurons. E₁: there was no difference in the V_½ between R168X, Bird, and WT cells. E₂: after normalization, there was no significant difference between R168X and Bird cells. F₁: slope factor between R168X and WT mice was not significantly different. F₂: there was no difference between R168X and Bird cells after normalization. NS, P > 0.05 (two-tail Student's t-test).

Excessive activity of Me5 neurons.

The abnormal motor function may involve the propriosensory system in addition to the pyramidal and extrapyramidal systems. Indeed, we have recently shown that the Mecp2 disruption altered membrane properties of mesencephalic trigeminal nucleus neurons (Me5) in Mecp2^Bird/Y (23). Hence, we studied how the Mecp2^R168X/Y mutation affects the membrane properties in these cells. The Me5 cells are typically silent in brain slices. To test their excitability, we measured the average instantaneous firing rate in response to depolarizing current injections. Me5 neurons from Mecp2^R168X/Y displayed a higher firing frequency in response to the same depolarizing current injections than those from WT mice (Fig. 8, A–C₁). When the maximal firing rate was compared at the same current injection, Me5 neurons from Mecp2^R168X/Y mice fired significantly faster than their WT littermates (46.2 ± 16.2 Hz, n = 10 in Mecp2^R168X/Y vs. 5.9 ± 4.0 Hz, n = 15 in WT; P < 0.05), which was similar to Me5 cells from Mecp2^Bird/Y mice (126.3 ± 18.4 Hz, n = 9) vs. the WT (66.8 ± 20.6 Hz, n = 9; P < 0.05) (Fig. 8, C₁ and C₂). When the difference in firing rates was compared between null and WT mice, no significant difference was found between Mecp2^R168X/Y and Mecp2^Bird/Y mice (Fig. 8D₁).

Fig. 8.

Altered membrane properties of mesencephalic trigeminal nucleus (Me5) neurons. A and B: depolarizing current injections elicited action potentials in Me5 cells from R168X and WT mice, respectively. C₁: average instantaneous FR of action potentials at increasing step current injections was compared between R168X and WT cells. C₂: FRs at the same current injections were higher in R168X and Bird cells than in cells from their WT littermates. D₁: changes in FR were compared by subtracting the WT mean value in each strain. The FR changes were not significantly different between R168X and Bird cells although Bird cells tended to fire more action potentials overall. D₂: rheobases of R168X and Bird cells were not significantly different compared with WT cells. E: there was no difference between R168X and Bird cells after normalization. F: the ratio of cells that fired few actional potentials vs. multiple actional potentials was significantly different compared with WT cells in R168X and Bird cells. G₁ and G₂: average resting membrane potentials (V_m) of R168X and Bird cells were not significantly different compared with WT cells or from each other. H₁: input resistances (R_m) of R168X and Bird cells were not significantly different from WT cells. H₂: after normalization to WT cells, the input resistances of R168X and Bird cells were not significantly different from each other. *P < 0.05; NS, P > 0.05 (two-tail Student's t-test).

The minimum current needed to elicit APs (rheobase) was not significantly different in both Mecp2^R168X/Y (475.0 ± 91.1 pA, n = 10, Mecp2^R168X/Y vs. 378.7 ± 57.3 pA, n = 15, WT; P > 0.05) and Mecp2^Bird/Y (274.3 ± 82.6 pA, n = 7, Mecp2^Bird/Y vs. 233.8 ± 31.1 pA, n = 8, WT; P > 0.05) (Fig. 8D₂). Mecp2^R168X/Y mice (1.0 ± 0.5, n = 8) and Mecp2^Bird/Y mice (1.0 ± 0.1, n = 8; P > 0.05) were not significantly different after normalization (Fig. 8E). Me5 neurons tended to fire either three or fewer APs or multiple APs sustained throughout the current injection as previously reported (40, 42). Both Mecp2^R168X/Y (5:9, Mecp2^R168X/Y; 1:17, WT; P < 0.05) and Mecp2^Bird/Y mice (5:4, Mecp2^Bird/Y; 1:8, WT; P < 0.05) had higher ratios of cells that fired multiple APs to fewer than three multiple APs than cells from WT littermates (Fig. 8F).

There were no significant differences in resting membrane potentials in either Mecp2^R168X/Y or Mecp2^Bird/Y strains compared with their WT littermates (Fig. 8, G₁ and G₂). Also neither strain showed a significant difference in membrane capacitance or input resistance measured from the linear portion of the I/V plot compared with their WT littermates (Fig. 8, H₁ and H₂).

Me5 AP characteristics.

We compared AP characteristics as we did with LC neurons (Fig. 9, A and B). The average AP threshold from Mecp2^R168X/Y mice (−38.0 ± 2.0 mV, n = 6) was significantly hyperpolarized compared with their WT littermates (−29.8 ± 2.7 mV, n = 8; P < 0.05) (Fig. 9C₁). The threshold of Mecp2^Bird/Y mice (−37.0 ± 1.2 mV, n = 6) was also significantly more hyperpolarized compared with their WT littermates (−32.7 ± 1.3 mV, n = 8; P < 0.05). After normalization, AP threshold from Mecp2^R168X/Y mice (1.3 ± 0.1, n = 6) was not different than that from Mecp2^Bird/Y mice (1.1 ± 0.0, n = 6; P > 0.05) (Fig. 9C₂). The average AP amplitude of Me5 neurons from Mecp2^R168X/Y mice (59.0 ± 1.6 mV, n = 6) was not significantly different from that of WT mice (61.1 ± 2.2 mV, n = 8; P > 0.05) (Fig. 9D₁). Mecp2^Bird/Y mice (66.9 ± 2.3 mV, n = 6) and their WT littermates (66.1 ± 1.7 mV, n = 8; P > 0.05) were also not significantly different (Fig. 9D₁). There was no significant difference in AP amplitude between Mecp2^R168X/Y mice (1.0 ± 0.0, n = 6) and Mecp2^Bird/Y mice (1.0 ± 0.0, n = 7; P > 0.05) after normalization to WT levels (Fig. 9D₂). In addition, there were no significant differences in AP overshoot of either mutant strain compared with their WT littermates.

Fig. 9.

Me5 neuron action potential morphology. A and B: representative action potentials from R168X, Bird, and WT littermate cells. C₁: action potential thresholds were significantly different between R168X and Bird cells compared with their WT littermates. C₂: R168X cells were not significantly higher than Bird cells after normalization. D₁: neither R168X nor Bird action potential amplitudes were different than cells from WT littermates. D₂: R168X and Bird mice were not significantly different after normalization. E–G: APD₅₀, rise times, and decay times not were significantly different when R168X and Bird cells were compared with their WT littermates. H: after normalization, there were no significant differences between R168X and Bird mice in APD₅₀ (D_1/2), rise time, or decay time. *P < 0.05; NS, P > 0.05 (two-tail Student's t-test).

The APD₅₀ was not significantly different in either Mecp2^R168X/Y (0.30 ± 0.0 ms, n = 6, Mecp2^R168X/Y vs. 0.38 ± 0.0 ms, n = 8, WT; P > 0.05) or Mecp2^Bird/Y mice (0.35 ± 0.0 ms, n = 6, Mecp2^Bird/Y vs. 0.38 ± 0.0 ms, n = 8, WT; P > 0.05) (Fig. 8E). Neither rise time nor decay time was significantly different in Mecp2^R168X/Y and Mecp2^Bird/Y mice than their WT littermates (Fig. 8, F and G). After normalization, APD₅₀, rise time, and decay time from Mecp2^R168X/Y cells were not significantly different than cells from Mecp2^Bird/Y mice (Fig. 8H).

Neuronal defects in female mice.

To distinguish symptomatic female mice from asymptomatic female mice, we used plethysmography as described in a previous report (18). We then compared the firing rates of LC cells from these mice. The average spontaneous firing rate of LC cells from symptomatic Mecp2^R168X/− mice (sMecp2^R168X/−) (5.1 ± 0.6 Hz, n = 14) was significantly higher than that of WT cells (3.3 ± 0.3 Hz, n = 13; P < 0.01) (Fig. 10, A₁-B₁). Cells from symptomatic Mecp2^Bird/− mice (sMecp2^Bird/−) (6.3 ± 0.8 Hz, n = 20) were also significantly higher compared with their WT littermates (Fig. 10B₁). There was no significant difference in neuronal FR between Mecp2^R168X/Y and Mecp2^Bird/Y mice after normalization (Fig. 10B₂). We also compared AP threshold as in the male mice and found that AP threshold (−35.9 ± 1.3 mV, n = 7; P < 0.05) was significantly lower in sMecp2^R168X/− mice than in aMecp2^R168X/− mice (−27.4 ± 3.4 mV, n = 7; P < 0.05; Fig. 10C₁). The threshold of sMecp2^Bird/− mice (−35.6 ± 2.4 mV, n = 7) was also significantly lower compared with aMecp2^Bird/− mice (−26.0 ± 3.4 mV, n = 7; P < 0.05). Normalized AP threshold of sMecp2^R168X/− mice (1.3 ± 0.0, n = 7) was not significantly different from sMecp2^Bird/− mice (1.4 ± 0.1, n = 7; P > 0.05) after dividing AP threshold of the symptomatic values by the mean asymptomatic values in both strains (Fig. 10C₂).

Fig. 10.

Firing properties of cells from female Rett mouse models. A_1-B₁: spontaneous firing activities of LC cells from asymptomatic Mecp2^R168X/+ (aR168X), symptomatic Mecp2^R168X/+ (sR168X), asymptomatic Mecp2^Bird/+ (aBird), and symptomatic Mecp2^Bird/+ (sBird) mice were compared. Both sR168X and sBird cells fired at a significantly higher rate than their asymptomatic littermates. B₂: there was no significant difference after normalization between sR168X and sBird cells. C₁: LC action potential thresholds were significantly lower in sR168X and sBird cells compared with aR168X and aBird cells. C₂: LC action potential thresholds of sR168X and sBird mice were not significantly different after normalization. D₁: as in males, the FRs at the same current injection from Me5 neurons were compared in female mice. The evoked FRs in both sR168X and sBird mice were significantly higher than their asymptomatic littermates. D₂: after normalization, there was no significant difference between sR168X and sBird mice. E₁: Me5 action potential thresholds were significantly lower in both sR168X and sBird cells compared with aR168X and aBird cells. E₂: sR168X cells were not significantly different from sBird cells after normalization. *P < 0.05; **P < 0.01; NS, P > 0.05 (two-tail Student's t-test).

When the firing rates of evoked APs were compared in the Me5 cells, we saw a similar trend as in male mice, i.e., the ceiling-level firing rate in sMecp2^R168X/− (80.4 ± 23.1 Hz, n = 11) was significantly higher than in aMecp2^R168X/− cells (23.1 ± 15.1 Hz, n = 12; P < 0.05). The average evoked FR was also significantly higher in sMecp2^Bird/− (139.5 ± 36.6 Hz, n = 13) compared with aMecp2^Bird/− mice (63.3 ± 31.0 Hz, n = 11; P < 0.01) (Fig. 10D₁). After normalization, there was no significant difference between sMecp2^R168X/− (57.3 ± 23.1 Hz, n = 11) and sMecp2^Bird/− (76.1 ± 36.6 Hz, n = 13; P > 0.05) (Fig. 10D₂). Also similar to male mice, the AP threshold from sMecp2^R168X/− mice (−35.7 ± 0.8 mV, n = 8) was significantly hyperpolarized compared with aMecp2^R168X/− (−32.8 ± 0.7 mV, n = 9; P < 0.05) (Fig. 10E₁). The threshold of sMecp2^Bird/− mice (−29.1 ± 2.0 mV, n = 7) was also significantly more hyperpolarized compared with aMecp2^Bird/− mice (−24.2 ± 0.9 mV, n = 8; P < 0.05). After normalization, AP threshold from sMecp2^R168X/− mice (1.1 ± 0.0, n = 8) was not significantly different from aMecp2^Bird/− mice (1.2 ± 0.1, n = 7; P > 0.05) (Fig. 10E₂).

DISCUSSION

This is the first demonstration of cellular properties in the Mecp2^R168X/Y mouse model. Our results have shown that breathing and motor function are clearly impaired in the Mecp2^R168X/Y mice. Our further studies of the intrinsic membrane properties of brainstem neurons indicate defects in LC neurons and Me5 neurons, which are known to play a role in breathing regulation and motor controls, respectively.

The R168X mutation.

In the R168X mutant, a point mutation causes a replacement of an arginine to a stop codon at residue 168, leading to premature termination of the MeCP2 peptide chain. This mutation occurs in ∼8–12% of patients with RTT, making it one of the most common causes for MeCP2 disruptions, alongside the T158A/M mutations (25). With the Mecp2^R168X/Y mutation, the truncated MeCP2 protein loses the transcription repressor domain although its methyl-binding domain seems to remain (44). As a result, the MeCP2 protein can no longer function as a transcriptional factor although it still can bind to DNA.

Several studies indicate that mutations closer to the NH² terminus are associated with higher severities of symptoms than those closer to the COOH terminus (4, 21). Human patients with the Mecp2^R168X/Y mutation demonstrated a high severity of motor dysfunction close to those with a large deletion, similar to the Mecp2^Bird/Y mouse model (3, 21).

Recently another Mecp2^R168X/Y mouse model has been created (7, 39). The authors reported that their males showed similar phenotypes to the original Mecp2^R168X/Y mouse model. Similarly, by comparing the results of our study that used the original Mecp2^R168X/Y mouse model with the published results from the new model, we also did not find any obvious phenotypic differences between these two Mecp2^R168X/Y mouse models. The female mice from the original model, however, showed a reduced body weight, whereas the mice from the new model did not (30). In addition, no breathing abnormalities were observed in the females of the new Mecp2^R168X model, whereas they were found in some females of the original Mecp2^R168X model. Because similar minor phenotypical variations have been reported by different research groups in the Mecp2^Bird/Y model, the phenotypic differences between the original and new Mecp2^R168X models appear modest.

Behavioral and functional findings from R168X mice.

In our present study, we have compared the behavioral and cellular phenotypes between Mecp2^R168X/Y and Mecp2^Bird/Y mice. Both strains show lower body weight compared with their WT littermates. Although this is consistent with most previous studies, subsets of male Mecp2^R168X/Y and Mecp2^Bird/Y mice showed body weight similar to and larger than the WT in some studies (12, 19). Interestingly, the Mecp2^R168X/Y mice are smaller than the Mecp2^Bird/Y mice under the same housing condition, as shown in this study. This seems to be due to the difference in their genetic background as previously reported (29). The Mecp2^R168X/Y mice were originally generated on a mixed genetic background of C57BL/6J × 129S6/SvEvTac, whereas Mecp2^Bird/Y mice were created on a pure C57BL/6 background. The Mecp2^R168X/Y mice also showed reduced lifespan and stereotypical hindlimb clasping, both of which are similar to Mecp2^Bird/Y mice (12, 19).

Motor dysfunction is an RTT-like phenotype. We have compared the grip strength between these two strains in this study and found that grip strength is significantly lower than that in their WT littermates. There is no significant difference between the two mutant strains in grip strength, which indicates that muscle tone is similarly impaired in both mutant strains. Motor coordination is defective in Mecp2^R168X/Y mice, as shown in the grid-walking test. When the motor coordination was compared between the two mutant strains, both showed impairments to a similar degree. These observations are thus consistent with previous studies and suggest impairment in motor coordination in both mutant strains.

Breathing dysfunction is a characteristic of RTT seen in human patients and mouse models (22, 48). Among the potential causes for the increased incidence of breathing irregularities and apneas are the defective CO₂-sensing mechanisms in patients with RTT and obstruction of the upper airways (14, 36, 48). Our study presents, for the first time, quantitative analysis of breathing abnormalities in male Mecp2^R168X/Y mice, including breathing frequency variation and apneas. These are consistent with breathing abnormalities of Mecp2^Bird/Y shown in this study as well as previous reports (1, 37, 48). No significant differences were found in breathing frequency variation and apnea counts between Mecp2^Bird/Y and Mecp2^R168X/Y mice, suggesting that both types of Mecp2 disruptions cause defects in these breathing parameters to a similar degree.

Overall, the behavioral phenotypes between Mecp2^R168X/Y and Mecp2^Bird/Y mice were indistinguishable in our study. In one previous study characterizing this mouse model, the authors compared their results to Mecp2^Jae/Y mice, another RTT mouse model that has an exon 3 deletion (30). The authors compared the onset of symptoms in the Mecp2^R168X/Y model to that of Mecp2^Jae/Y mice and concluded that there was no difference in males. Female Mecp2^R168X/X mice did show a subtle difference in increased anxiety at a later age.

Defects in brainstem neuronal activity.

The LC is the major provider of NE in the CNS affecting many systems and functions. In these RTT mouse models, LC neurons show increased spontaneous firing activity as well as other alterations in their membrane properties. We have previously found that the hyperexcitability of LC neurons in Mecp2^Bird/Y mice can be attributed to changes in intrinsic membrane properties and alterations of different ion channel expression (47). In addition to these changes in intrinsic membrane properties, the inhibitory GABA input to LC cells is reduced, further enhancing LC neuronal excitability (16). In our present study, we saw similar changes in LC properties, including hyperexcitability in Mecp2^R168X/Y mice as in Mecp2^Bird/Y mice. Because behavioral phenotypes are very similar between the two strains, the finding that their neuronal activities are similarly affected is not surprising.

As the major provider of NE in the CNS, proper LC cell function is crucial for maintaining normal function and behaviors. In human patients with RTT and RTT mouse models, the amount of available NE in the CNS is reduced (34, 51). This has been attributed to impaired NE synthesis (49). Raising the amount of available NE at synapses with the NE reuptake blocker desipramine improves symptoms such as breathing dysfunction and increases the lifespan of mouse models, further supporting its role in the development of symptoms in this disease (27, 45). Alternatively, increasing the firing rate of LC cells could also lead to more NE release from LC neuronal synaptic terminals, which might alleviate RTT-like symptoms as well.

However, our recent study using optogenetic interventions indicates that the increased firing rate of LC cells does not improve NE release from Mecp2-null neurons at all (46). Therefore, LC neuronal hyperexcitability does not seem beneficial to the NE modulation and may actually impair NE synthesis attributable to the increased energy consumption and intracellular Ca²⁺ overload (46). In fact, several studies have shown that reducing neuronal excitability by enhancing inhibitory signaling in RTT mouse models benefits RTT-like symptom relief (1, 35, 50).

Motor dysfunction is a prominent symptom of patients with RTT. Among other motor functions, patients with RTT exhibit dysfunction in facial movements such as chewing and teeth grinding (26). Consistent with findings in humans, we have observed misalignment of the jaw and teeth overgrowth in Mecp2^R168X/Y mice, which has also been reported in Mecp2^Bird/Y mice as well (12). The Me5 contains propriosensory neurons that act as a feedback mechanism for jaw and facial muscles. Because of their ideal anatomical position, the location of Me5 neurons provides an opportunity to study the propriosensory component of the motor system. We have previously found that Me5 neurons in Mecp2^Bird/Y mice display excessive firing activity as LC cells (23). The increased firing activity of Me5 neurons seems to be attributed to reorganization of I_Na and I_h (23). In our present study of Mecp2^R168X/Y mice, we have also found hyperexcitability of Me5 neurons that was comparable to the hyperexcitability found in Me5 neurons from Mecp2^Bird/Y mice. This hyperexcitation of propriosensory neurons may cause abnormal feedback to motor neurons, causing more excitation of jaw muscles and thus exacerbating the defects in the motor system.

In conclusion, we have compared several behavioral and cellular outcomes between two strains of RTT mouse models. LC neurons in both mutant strains show defects in their intrinsic membrane properties, leading to excessive firing activity. Consistent with these defects are breathing abnormalities seen in both mutant strains. The Me5 neurons in both strains of RTT models display higher firing responses to depolarization than their WT littermates. These neurons are propriosensory, regulating motor neurons via certain servo feedback circuits. Their dysfunction thus is consistent with the motor defects in muscle strength and coordination. Both LC and Me5 neurons with hyperexcitability should have an impact on the excitation-inhibition balances in brainstem neuronal networks as well as their associated functions. Interventions targeted toward reducing hyperexcitability in the CNS should help correct this excitation-inhibition imbalance. An MeCP2 protein product was previously detected in the original Mecp2^R168X/Y mouse model (19) and in several in vitro assays in HeLA cells, Xenopus, and yeast cells, indicating that a stable truncated protein can be expressed (8, 20, 32). In contrast, no protein product was detected in the new Mecp2^R168X/Y mouse model and in an in vitro assay (7, 44). Although this study did not test for protein expression, on the basis of our results, the truncation does not seem to have evident effects on the behavioral and cellular manifestations in the mice, suggesting that the transcriptional repressor domain of the protein is indeed necessary for proper functioning of the CNS in patients with RTT, whereas the methyl-binding domain alone has little effects. In addition, the comparability in phenotypical outcomes of the Mecp2^R168X/Y model with the Mecp2^Bird/Y model suggests that both strains of mice are valid animal models for RTT research.

GRANTS

This work was supported by National Institutes of Health National Institute of Neurological Disorders and Stroke Grant NS-073875.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.M.J. and C.J. conceived and designed the research; C.M.J., W.Z., N.C., Y.W., H.X., and S.Z. performed experiments; C.M.J., W.Z., N.C., and C.J. analyzed data; C.M.J., W.Z., N.C., Y.W., H.X., S.Z., and C.J. interpreted results of experiments; C.M.J., N.C., and C.J. prepared figures; C.M.J. and C.J. drafted manuscript; C.M.J., W.Z., and C.J. edited and revised manuscript; C.M.J., W.Z., N.C., Y.W., H.X., S.Z., and C.J. approved final version of manuscript.