Abnormal Prefrontal Neural Oscillations are Associated with Social Deficits in MECP2 Duplication Syndrome

Dear Editor,

Methyl-CpG-binding protein 2 (MeCP2) participates in regulating the expression of the CREB, TDP-43 and Dnmt1 genes [1, 2]. Recent studies have shown that MeCP2 is essential for maintaining normal neural functions in the mammalian brain. Overexpression of MeCP2 causes MECP2 duplication syndrome [3], a type of autism spectrum disorder that results in a core phenotype of social deficits in human patients [4], nonhuman primates [5], and rodents [6].

Overexpression of MeCP2 can cause an abnormal increase in dendritic spines in excitatory neurons [7], which causes an excitation/inhibition imbalance in neural circuits and results in autism-like behaviors. Research in MECP2 transgenic (TG) mice indicates that the medial prefrontal cortex (mPFC) plays an important role in the social deficits that characterize MECP2 duplication syndrome. While functional magnetic resonance imaging has shown abnormal overexcitation of the mPFC in MECP2 TG mice, knocking out the MECP2 transgene in the mPFC rescues the social recognition deficits in adult TG mice [8, 9]. However, it remains unknown whether MeCP2 overexpression leads to abnormal mPFC neural activity, and the relationship between neural activity and social deficits in MECP2 duplication syndrome remains to be characterized.

In this study, MECP2 duplication syndrome models were developed in adult (20-week-old) male MECP2 TG Sprague Dawley rats, in which a human MECP2 transgene had been introduced by recombination with a microinjected P1-derived artificial chromosome. A group of wild-type (WT) adult (20-week-old) male rats followed the same experimental protocol as the control group.

To record neural activity, bilateral multi-channel macroelectrodes (3 channels in each hemisphere) were implanted into the mPFC region (anterior/posterior: 4.20 mm; mediolateral: ±0.85 mm; dorsal/ventral: −3.9 mm), and local field potentials (LFPs) were recorded with the AlphaLab SnR system (AlphaOmega), which captures the oscillatory activity of neuronal populations.

To explore the abnormal neural oscillations in the mPFC in the MECP2 duplication syndrome model, the LFPs were analyzed by power spectral analysis. All LFPs were preprocessed with the EEGLAB toolbox [10]. Since the energy of the neural oscillatory activity was mainly concentrated in the low frequency band, the LFPs were first filtered with a 150-Hz low-pass filter and resampled to 300 Hz. Then, the signals were further filtered with a 2-Hz high-pass filter to remove low-frequency baseline drift. The power-line interference was removed by adaptive notch filters [11]. The power spectral density was then estimated by the Welch method with a 2-s time window and 1-s overlap and normalized by dividing by the integral of the power spectral density from 2 Hz to 90 Hz.

The social ability of the TG (n = 11) and WT control groups (n = 11) was evaluated by behavioral analysis of social interaction. After 10 min of adaptation in the home cage, each rat engaged in social interaction with 3 unfamiliar, age-matched WT males in 3 test trials; each trial lasting 5 min. During the social challenge experiment, the average duration and frequency of active social behavior (when the target rat actively contacted or sniffed the WT partner) and passive social behavior (when the WT partner contacted or sniffed the front half of target rat without avoidance), the total duration of social behavior, and the average frequency of social avoidance were recorded. The results showed that the active social time and frequency (Fig. 1A, D), passive social time and frequency (Fig. 1B, E), and total social time (Fig. 1C) of the TG group were significantly less than those of the WT control group, but there was no significant difference in social avoidance (Fig. 1F).

Fig. 1.

Social deficits and abnormal mPFC neural oscillations in rats with MECP2 duplication. A–F Boxplots in WT and TG rats of active social time (A) passive social time (B), total social time (C), active social frequency (D), passive social frequency (E), and social avoidance (F). G LFP signals after preprocessing. H Power spectra (2–90 Hz) of TG and WT rats under solo and social conditions (colored lines, mean value; shaded area, SEM) I, J Neural responses to the social interaction of WT (I) and TG (J) rats in each frequency band (mean ± SEM, lines show each sample). K, L Boxplots of LFP power of WT and TG rats in each frequency band under solo (K) and social (L) conditions (ends of whiskers represent the minimum and maximum of data points). The line within box represents the median (odd numbers of data points) or second quartile (even number of data points); the bottom and top edges of a box represent the first and third quartiles, respectively. WT n = 11, TG n = 11 in social interaction behavioral analysis in A–F, WT n = 9, TG n = 17 in LFP recording in H–L, *P <0.05, **P <0.01, ***P <0.001, two tailed independent t-test in A–F; two-tailed paired t-test in I and J; two-tailed independent t-test in K and L. au, arbitrary unit

Based on the significant difference between the TG and WT control groups in social interaction behavioral analysis, the LFPs of the mPFC were recorded and analyzed under two conditions (Fig. 1G): one in which the target rat moved freely in an individual cage (solo), and one in which an unfamiliar, age-matched male WT rat of the same age was added to the home cage as a social partner (social). Three trials were recorded under each condition, and each trial lasted 20 min.

The power spectra of the mPFC LFPs of the TG (n = 17) and WT control (n = 9) groups in the solo and social conditions were compared to reveal the difference in neural oscillations between these groups (Fig. 1H). In this study, the neural oscillations were investigated in the following frequency bands: theta (4–8 Hz), alpha (8–12 Hz), low beta (12–20 Hz), high beta (20–35 Hz), low gamma (35–60 Hz) and high gamma (60–90 Hz).

First, the neural response in the mPFC to social interaction was investigated in the TG and WT control groups. The power of high gamma oscillations (55–90 Hz) significantly increased from the solo condition to the social condition in the TG group (Fig. 1J). However, the power of high beta oscillations (20–35 Hz) significantly decreased in the WT control group (Fig. 1I). Synchronization in the high gamma band and desynchronization in the high beta band were found in the TG and WT control groups, respectively, in the social condition.

When the power spectra of the TG and WT control groups were compared in the solo condition, the TG group had significantly higher power than the WT control group in the theta (4–8 Hz) and alpha (8–12 Hz) bands (Fig. 1K) but significantly lower than the WT control group in the low gamma band (35–60 Hz). No significant difference was found between the two groups in the social condition (Fig. 1L). These data indicated that MeCP2 overexpression led to abnormal neural activity in the mPFC and hypersynchronized gamma oscillations in response to a social challenge.

The relationship between neural oscillations and social deficits was further investigated by analyzing the Spearman correlation coefficients between the power differences of neural oscillations from the solo condition to the social condition and the scores for social behaviors in the TG (n = 11) and WT control (n = 8) groups. The neural response to the social challenge showed a significant correlation with social behavior in the TG group. The power increase in the high gamma band (60–90 Hz) was negatively correlated with active social time (r = −0.655, P = 0.034; Fig. 2A), while a significant positive correlation was found between the gamma band (35–90 Hz) power increase and total social time in the WT control group (r = 0.786, P = 0.028; Fig. 2D). The power increases in the theta and alpha bands (4–12 Hz) were negatively correlated with passive social time (r = −0.718, P = 0.017), while the power increase in the beta band (12–35 Hz) showed a positive correlation in the TG group (r = 0.800, P = 0.005; Fig. 2C).

Fig. 2.

Correlation between neural oscillation and social behavior in rats with MECP2 duplication (TG) and WT controls. A Correlation between power difference in high gamma (60–90 Hz) and active social time. B Correlation between low beta (12–20 Hz) power and active social time. C Correlation between power difference in theta-alpha (4–12 Hz), beta (12–35 Hz), and passive social time. D Correlation between power difference in gamma (35–90 Hz) and total social time. E Correlation between gamma (35–90 Hz) power and total social time. n = 8 for WT and n = 11 for TG in correlation analysis. The regression lines represent P <0.05 in A–E Spearman correlations. au, arbitrary unit

Moreover, the Spearman correlations between the power of neural oscillations in social conditions and social behavior scores were also analyzed in both groups. The TG group showed a negative correlation between gamma band (35–90 Hz) power and total social time (r = −0.652, P = 0.030; Fig. 2E). In addition, the power of the low beta band (12–20 Hz) was negatively correlated with active social time (r = −0.655, P = 0.034; Fig. 2B). No correlation was found in the WT control group.

In this study, the MECP2 TG rat model showed significant social deficits in the behavioral analysis of social interaction, replicating MECP2 duplication syndrome. Meanwhile, transgenic MECP2 rats also had abnormal neural oscillations in the mPFC, reflected by increased theta and alpha power and decreased low gamma power. The gamma neural oscillations made a significant contribution to the social deficits. The gamma oscillations under the social condition were negatively correlated with the total social time, and the high gamma oscillations increased in the social challenge condition in the TG group. The increase in high gamma oscillations was negatively correlated with active social time in the TG group, while the increase in gamma oscillations was positively correlated with total social time in the WT control group.

The abnormality of mPFC gamma oscillations and the essential role of this abnormality in social deficits in MECP2 duplication syndrome are similar to phenomena reported in other subtypes of autism spectrum disorder. The Neuroligin 3 R451C knock-in mouse model of autism also shows a decrease in low gamma power in the mPFC during social behavior, and the dysfunction of gamma oscillations contributes to a deficit in social novelty preference [12]. Moreover, clinical research has shown that abnormal gamma oscillations are an important disease phenotype in autism patients [13]. These results indicate the importance of gamma oscillations in identifying disease characteristics and revealing abnormal neuronal population activity associated with deficits not only in MECP2 duplication syndrome but also in other types of autism spectrum disorder.

Moreover, alpha and beta oscillations were found to be involved in passive social responses. The results indicated that different neural oscillations are associated with active and passive social behaviors.

A recent study of MECP2 duplication syndrome indicated that MeCP2 overexpression causes an abnormal increase in dendritic spines in excitatory neurons, which results in an excitation/inhibition imbalance as a characteristic abnormality of neural microcircuit activity in this disease [14]. However, the characteristics of oscillatory neuronal population activity in MECP2 duplication syndrome and their correlation with deficits in social behavior are still unknown. Upon exploring this topic, the present study, we found abnormal characteristics of neural oscillations in the mPFC in MECP2 duplication syndrome and established a correlation between neural oscillations and social behavior. These results helped clarify the core phenotype of an autism model by exploring its mechanism in terms of neural oscillations.

Conflict of interest

The authors declare no competing interests