Journal of Neurophysiology

Selective impact of MeCP2 and associated histone deacetylases on the dynamics of evoked excitatory neurotransmission

Erika D. Nelson, Manjot Bal, Ege T. Kavalali, Lisa M. Monteggia

Abstract

An imbalance between the strengths of excitatory and inhibitory synaptic inputs has been proposed as the cellular basis of autism and related neurodevelopmental disorders. Previous studies examining spontaneous levels of excitatory and inhibitory neurotransmission in the forebrain regions of methyl-CpG-binding protein 2 (Mecp2) mutant mice, models of the autism spectrum disorder Rett syndrome, have identified a decrease in excitatory drive, in some cases coupled with an increase in inhibitory synaptic strength, as a major source of this imbalance. Here, we reevaluated this question by examining the short-term dynamics of evoked neurotransmission between hippocampal neurons cultured from MeCP2 knockout mice and found a marked increase in evoked excitatory neurotransmission that is consistent with an increase in presynaptic release probability. This increase in evoked excitatory drive was not matched with alterations in evoked inhibitory neurotransmission. Moreover, we observed similar excitatory drive specific changes after the loss of key histone deacetylases (histone deacetylase 1 and 2) that form a complex with MeCP2 and mediate transcriptional regulation. These findings suggest a distinct role for MeCP2 and its cofactors in the regulation of evoked excitatory neurotransmission compared with their essential role in basal synaptic activity.

  • Rett syndrome
  • transcriptional repression
  • synaptic neurotransmission
  • knockout
  • methyl-CpG-binding protein 2

an imbalance between the strengths of excitatory versus inhibitory inputs at the neuronal level in the brain has become one of the primary hypotheses for the cellular basis of many neurodevelopmental disorders. Recent research points to an increase in inhibitory drive associated with the neurodevelopmental disorder Down syndrome (Chakrabarti et al. 2010). Fragile X and related autism phenotypes have also been proposed to result from alterations in neocortical excitatory/inhibitory balance (Benvenuto et al. 2009; Gibson et al. 2008). It has been suggested that autism spectrum disorders may be caused, at least partially, by an increase in the ratio of excitation to inhibition (Rubenstein 2010; Rubenstein and Merzenich 2003), since ∼30% of autistic individuals develop seizures, suggesting a hyperexcitability of cortical networks (Tuchman 2006).

Rett syndrome (RTT), an autism-related disorder seen primarily in females, is characterized by normal development up to the age of 6–18 mo, at which time patients fail to acquire new skills and enter a period of motor skill regression (Hagberg et al. 1983). RTT is caused by mutations in the X-linked gene encoding methyl-CpG-binding protein 2 (MeCP2) (Amir et al. 1999). Individuals with RTT show a wide range of neurological defects, including mental retardation, disturbances of sleep, problems with gait, decelerated head growth, and stereotypical hand movements, along with a loss of social and cognitive abilities. Seizures occur in the majority of RTT patients (Jian et al. 2007; Nissenkorn et al. 2010), but the frequency and severity of seizures tend to decrease into adolescence and adulthood, suggesting that possible deficits in excitatory/inhibitory imbalance may be limited to early developmental stages (Steffenburg et al. 2001). This epileptic activity is suggestive of cortical hyperexcitability (Glaze 2005), and, taken together with the observed decreases in the threshold for action potential firing, higher glutamate levels, and increased N-methy-d-aspartate (NMDA) receptor density in different areas of the cortex (Blue et al. 1999; Eyre et al. 1990; Horska et al. 2009), these observations imply that the brains of RTT patients favor excitation over inhibition.

Contrary to this expectation, early research on Mecp2-deficient mice, widely used animal models of RTT, revealed a decrease in excitatory synaptic transmission between cortical as well as hippocampal neurons, with either no change or an increase in inhibition (Chao et al. 2007; Dani et al. 2005; Nelson et al. 2006; Tropea et al. 2009). Additional imaging experiments have demonstrated a reduction in the excitation of cortical neurons in response to glutamate stimulation, with overall network activity being only mildly affected (Wood et al. 2009; Wood and Shepherd 2010). This decrease in excitation may be one explanation for the impaired long-term potentiation, a measure of augmented synaptic efficacy in response to repetitive high-frequency stimulation, observed in both hippocampal and cortical slices from Mecp2 mutant mice (Asaka et al. 2006; Moretti et al. 2006). However, both the cortex and hippocampus of Mecp2 knockout (KO) mice have been shown to be prone to hyperexcitability (Calfa et al. 2011; D'Cruz et al. 2010; Shahbazian et al. 2002; Zhang et al. 2008). Furthermore, decreases in paired pulse ratios (PPRs), a measure of short-term synaptic plasticity, were seen in cortical and hippocampal neurons in situ, suggesting an increase in excitatory neurotransmitter release (Asaka et al. 2006; Moretti et al. 2006; Nelson et al. 2006). Since short-term synaptic plasticity has been implicated in regulating the strength of synaptic transmission based on the timing of inputs and action potential firing (Abbott and Regehr 2004), it seems possible that increased excitation during short-term stimulation could result in overall hyperexcitability of neuronal networks.

In this study, we set out to make a comprehensive assessment of short-term synaptic plasticity after the loss of MeCP2, with an attempt to relate any observed changes to the proposed imbalances in excitatory and inhibitory synaptic activity. As an extension of these experiments, we also measured excitatory and inhibitory short-term synaptic plasticity in hippocampal neurons lacking the MeCP2-associated transcriptional repressor proteins histone deacetylase (HDAC)1 and HDAC2 (Nan et al. 1998). Histone deacetylation has previously been implicated in the regulation of long-term synaptic plasticity, the cellular basis for learning and memory (Alarcon et al. 2004; Guan et al. 2002; Levenson et al. 2004; Vecsey et al. 2007). However, studies on the effects of specific HDACs on synapse function are relatively few (Akhtar et al. 2009; Guan et al. 2009), and further insight is needed into how these proteins might play a role in short-term synaptic plasticity.

MATERIALS AND METHODS

Cell culture.

Dissociated hippocampal cultures were prepared from the brains of male Mecp2tm1.1Bird-null KO mice (Jackson Laboratories), floxed Hdac1 mice, or floxed Hdac2 mice according to previously published protocols (Kavalali et al. 1999). Briefly, whole hippocampi were dissected from floxed mice on postnatal days 0–3 or Mecp2 KO mice on postnatal day 0. Tissue was trypsinized for 10 min at 37°C, mechanically dissociated using siliconized glass pipettes, and then plated onto Matrigel-coated coverslips. At day 1 in vitro, 4 μM cytosine arabinoside (Sigma) was added to control for overgrowth of glial cells. All animal experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee of University of Texas Southwestern Medical Center.

Lentivirus production.

Human embryonic kidney-293 cells were transfected using the Fugene 6 transfection system (Roche Molecular Biochemicals) with the expression plasmids pFUGW or pFUGW-Cre and two helper plasmids, Δ8.9 and vesicular stomatitis virus G protein, at 3 μg of each DNA per 75-cm2 flask (Dittgen et al. 2004). After 48 h of transfection, the lentivirus-containing culture medium was harvested, filtered at a 0.45-μm pore size, aliquoted, frozen in liquid nitrogen, and stored at −80°C. Floxed Hdac1 and floxed Hdac2 hippocampal cultures were infected at 7 days in vitro by adding 300 μl of viral suspension to each well. The titer was determined by counting the number of infected [green fluorescent protein (GFP)-positive] neurons per coverslip, with a high titer being >80% infection.

Electrophysiology.

Synaptic activity was recorded from hippocampal pyramidal cells using whole cell voltage-clamp techniques. Data were acquired using an Axopatch 200B amplifier and Clampex 9.0 software (Axon Instruments). Recordings were filtered at 2 kHz and sampled at 200 μs. Modified Tyrode solution containing (in mM) 150 NaCl, 4 KCl, 2 MgCl2, 2 CaCl2, 10 glucose, and 10 HEPES (pH 7.4) was used as the external bath solution with 50 μM picrotoxin to isolate α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor-mediated excitatory postsynaptic currents (EPSCs) or 10 μM 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione and 50 μM 2-amino-5-phosphonopentanoic acid to isolate inhibitory postsynaptic currents (IPSCs). For NMDA receptor-mediated EPSC recordings, the external bath solution also contained (in μM) 0 MgCl2, 10 6-cyano-7-nitroquinoxaline-2,3-dione, and 50 picrotoxin. For all recordings, the pipette internal solution contained (in mM) 115 Cs-MeSO3, 10 CsCl, 5 NaCl, 10 HEPES, 0.6 EGTA, 20 TEA-Cl, 4 Mg-ATP, 0.3 Na3GTP, and 10 QX-314 (pH 7.35; 300 mosM). Field stimulation was applied through parallel platinum electrodes immersed in the perfusion chamber delivering 20- to 30-mA amplitude (1-ms duration) pulses to generate maximum outputs. All experiments were done on cultures of 14–22 days in vitro. Control experiments were performed on hippocampal cultures from wild-type littermates of Mecp2 KO mice, lenti-GFP-infected floxed Hdac1 neurons, or lenti-GFP-infected floxed Hdac2 neurons.

Statistical analysis.

All error bars represent SEs, and all data were tested for statistical significance by means of a two-tailed Student's t-test.

RESULTS

Alterations in excitatory short-term synaptic plasticity seen with the loss of MeCP2, HDAC1, and HDAC2 in hippocampal neurons.

To elucidate the effects of MeCP2 on short-term synaptic plasticity in the context of a possible imbalance in excitatory and inhibitory activity in Mecp2 KO mice, we performed whole cell patch-clamp recordings of dissociated hippocampal cultures made from newborn (postnatal day 0) Mecp2 KO mice and wild-type littermate controls. Dissociated cultures are particularly useful for the assessment of synaptic function at the level of individual neurons while, at the same time, limiting the contribution of network and noncell autonomous factors (Kavalali et al. 1999). We waited for the cultured neurons to mature for ∼2–3 wk in vitro to allow for complete synapse formation and maturation (Mozhayeva et al. 2002). We began examining the properties of evoked synaptic transmission by recording both excitatory and inhibitory responses to paired pulse stimulation over a range of different stimulus frequencies (20 Hz to 1 Hz). The PPRs of EPSCs and IPSCs were then calculated by dividing the average response size on the second pulse by the average response size on the first pulse. EPSCs recorded from hippocampal neurons typically show facilitation at higher stimulation frequencies (<100-ms interstimulus intervals) (Dobrunz et al. 1997). As expected, the PPRs of EPSCs recorded from control neurons were >1 when stimulated at both 10 and 20 Hz, with no facilitation occurring at longer interstimulus intervals (Fig. 1A). Mecp2 KO neurons showed significant reductions in the PPRs of EPSCs at high stimulation frequencies (20, 10, and 5 Hz) compared with control neurons (Fig. 1A). Alterations in PPRs are believed to be indicators of presynaptic properties of vesicle release, with the amount of facilitation being inversely related to the initial release probability (Pr) at the synapse. Our EPSC PPR results agree with previous findings in Mecp2 KO hippocampal neurons compared with wild-type controls (Asaka et al. 2006; Moretti et al. 2006; Nelson et al. 2006) and bolster the premise that the loss of MeCP2 increases Pr from excitatory presynaptic terminals. No alterations were seen in the PPRs of IPSCs in Mecp2 KO neurons (Fig. 1B), which supports previous studies that suggest that MeCP2 specifically regulates excitatory synapses at the cellular level (Asaka et al. 2006; Chao et al. 2007; Nelson et al. 2006; Tropea et al. 2009).

Fig. 1.

Changes in excitatory, but not inhibitory, paired pulse ratios (PPRs) in methyl-CpG-binding protein 2 (Mecp2), histone deacetylase 1 (Hdac1), and histone deacetylase 2 (Hdac2) knockout (KO) hippocampal neurons. A, top: representative traces of normalized evoked excitatory postsynaptic currents (EPSCs) recorded in 50 μM picotoxin at 20-Hz stimulation frequency. Bottom, Mecp2 KO neurons showed decreased excitatory PPRs at 5-, 10-, and 20-Hz stimulation frequencies compared with controls, whereas both Hdac1 and Hdac2 KO neurons had reduced excitatory PPRs at 10 and 20 Hz. Control, n = 48 neurons/10 coverslips; Mecp2 KO, n = 13 neurons/3 coverslips; Hdac1 KO, n = 20 neurons/3 coverslips; Hdac2 KO, n = 17 neurons/3 coverslips. *P < 0.05 for control neurons compared with Mecp2 KO, Hdac1 KO, and Hdac2 KO neurons; **P < 0.05 for Mecp2 KO neurons compared with control neurons. B, top: representative traces of normalized evoked inhibitory postsynaptic currents (IPSCs) recorded in 10 μM 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione at 5-Hz stimulation frequency. Bottom, inhibitory PPRs recorded at numerous stimulation frequencies were not significantly affected by the loss of MeCP2, HDAC1, or HDAC2. Control, n = 45 neurons/9 coverslips; Mecp2 KO, n = 10 neurons/2 coverslips; Hdac1 KO, n = 17 neurons/3 coverslips; Hdac2 KO, n = 18 neurons/4 coverslips.

MeCP2 is known to function as a transcriptional repressor in a complex containing additional proteins, including HDAC1 and HDAC2; therefore, we also explored short-term synaptic plasticity and the balance between excitation and inhibition in neurons lacking these proteins. Since constitutive Hdac1 and Hdac2 KO mice die during embryonic development, we used a viral-mediated approach to knock out these individual HDACs in hippocampal neurons. To investigate whether an acute loss of HDAC function would result in alterations in synaptic transmission, we made primary dissociated hippocampal cultures from newborn floxed Hdac1 or Hdac2 mice and allowed them to age 7 days in vitro before infecting them with a high-titer lentivirus expressing GFP-tagged Cre recombinase (Cre) or GFP as a control. One to two weeks later, we recorded evoked EPSCs and IPSCs from both GFP- and Cre-infected floxed Hdac1 and floxed Hdac2 neurons. Characterization of the successful knockdown of these proteins has been previously described (Akhtar et al. 2009). Similarly to what was seen in Mecp2 KO neurons, the PPRs of evoked EPSCs in both Hdac1 and Hdac2 KO neurons were reduced at high stimulation frequencies (20 and 10 Hz), whereas the PPRs of evoked IPSCs were unchanged across all stimulation frequencies (Fig. 1, A and B). Numerical PPR data before normalization are included in the Supplemental Material (Supplemental Fig. S1).1 These data suggest that MeCP2 may work in combination with both HDAC1 and HDAC2 to negatively regulate excitatory, but not inhibitory, initial presynaptic Pr in response to action potentials in mature hippocampal neurons.

Trains of postsynaptic currents reveal increased excitatory drive in both Mecp2 and Hdac2 KO neurons.

To further study how the loss of MeCP2, HDAC1, and HDAC2 might affect excitatory short-term synaptic transmission during bursts of action potential firing, we recorded postsynaptic responses to short trains of five action potentials at different stimulation frequencies. The average first response amplitudes of EPSCs were largely enhanced in both Mecp2 KO and Hdac2 KO neurons compared with controls (Mecp2 KO neurons: 1,174.5 ± 252.5 pA and control neurons: 608.6 ± 129.8 pA; and Hdac2 KO neurons: 1,242.1 ± 165.5 pA and control neurons: 609.8 ± 167.4 pA, P < 0.05 for both), suggesting an increase in excitatory drive in response to action potential firing. At 2, 5, and 10 Hz, the increase in evoked EPSC amplitudes in Mecp2 KO neurons was most robust during the first couple of responses and became less so as the synaptic responses began to depress with more action potentials (Fig. 2, A–C). In Hdac1 KO neurons, the amplitudes of evoked EPSCs were unchanged at all three stimulation frequencies compared with controls (average first response amplitudes: Hdac1 KO neurons 954.0 ± 226.4 pA and control neurons 972.0 ± 217.6 pA, P > 0.05; Fig. 2, D–F). In Hdac2 KO neurons, the amplitudes of evoked EPSCs were increased significantly at every response during the 2-, 5-, and 10-Hz trains (Fig. 2, G–I). The EPSC amplitude data are included in the Supplemental Material (Supplemental Fig. S2).

Fig. 2.

EPSC amplitudes were increased in response to trains of five action potentials in Mecp2 and Hdac2 KO neurons. A–C: EPSC amplitudes were significantly increased in Mecp2 KO neurons compared with controls at the beginning of 2- (A), 5- (B), and 10-Hz (C) stimulation frequencies. Control, n = 15 neurons/3 coverslips; Mecp2 KO, n = 11 neurons/3 coverslips. *P < 0.05. D–F: EPSC amplitudes were unchanged in Hdac1 KO neurons compared with controls at 2- (D), 5- (E), and 10-Hz (F) stimulation frequencies. Control, n = 15 neurons/3 coverslips; Hdac1 KO, n = 20 neurons/3 coverslips. G–I: EPSC amplitudes were significantly increased in Hdac2 KO neurons compared with controls at 2- (G), 5- (H), and 10-Hz (I) stimulation frequencies. Control, n = 17 neurons/4 coverslips; Hdac2 KO, n = 17 neurons/3 coverslips. *P < 0.05.

Trains of inhibitory postsynaptic currents are unchanged in Mecp2, Hdac1, and Hdac2 KO neurons.

When we repeated the same action potential train stimulation to evoke IPSCs, we found that IPSCs recorded from Mecp2 KO neurons were similar to controls at every response during the 2-, 5-, and 10-Hz trains (average first response amplitudes: Mecp2 KO neurons 1,145.9 ± 349.2 pA and control neurons 1,287.4 ± 202.0 pA, P > 0.05; Fig. 3, A–C). In Hdac1 KO neurons, the amplitudes of evoked IPSCs during each train were unchanged at all three stimulation frequencies, although there was a slight, nonsignificant trend toward decreased IPSC amplitudes (average first response amplitudes: Hdac1 KO neurons 1,066.8 ± 214.9 pA and control neurons 1,408.2 ± 239.2 pA, P > 0.05; Fig. 3, D–F). In Hdac2 KO neurons, evoked IPSC amplitudes also remained similar to controls during the 2-, 5-, and 10-Hz trains (average first response amplitudes: Hdac2 KO neurons 1,571.2 ± 258.5 pA and control neurons 1,103.6 ± 266.7 pA, P > 0.05; Fig. 3, G–I). The numerical IPSC amplitude data are included in the Supplemental Material (Supplemental Fig. S3). All together, the train stimulation data indicate that both MECP2 and HDAC2 play specific roles in controlling the magnitude of evoked excitatory synaptic activity. These data further demonstrate the involvement of MeCP2 and HDAC2 in controlling excitatory, but not inhibitory, synaptic function. Interestingly, the loss of both MeCP2 and HDAC2 resulted in increased excitatory responses, indicating an enhanced excitation in response to action potential stimulation and a greater excitatory-to-inhibitory ratio, which is in contrast to previous results suggesting decreased excitatory synaptic activity in Mecp2 KO neurons during spontaneous activity (Chao et al. 2007; Dani et al. 2005; Nelson et al. 2006).

Fig. 3.

IPSC amplitudes were unchanged in response to trains of five action potentials in Mecp2, Hdac1, and Hdac2 KO neurons. A–C: in Mecp2 KO neurons, IPSC amplitudes were similar to controls at 2- (A), 5- (B), and 10-Hz (C) stimulation frequencies. Control, n = 10 neurons/2 coverslips; Mecp2 KO, n = 10 neurons/2 coverslips. D–F: in Hdac1 KO neurons, IPSC amplitudes were not significantly different from controls at 2- (D), 5- (E), and 10-Hz (F) stimulation frequencies. Control, n = 17 neurons/3 coverslips; Hdac1 KO, n = 17 neurons/3 coverslips. G–I: in Hdac2 KO neurons, IPSC amplitudes were similar to controls at 2- (G), 5- (H), and 10-Hz (I) stimulation frequencies. Control, n = 18 neurons/4 coverslips; Hdac2 KO, n = 19 neurons/4 coverslips.

Excitatory postsynaptic response depression is faster in Mecp2 KO neurons but not in Hdac1 or Hdac2 KO neurons.

The short-train stimulation data indicated possible changes in the rate of response depression during brief bursts of activity, particularly in Mecp2 KO neurons. Therefore, we measured both excitatory and inhibitory responses to 20 s of action potential stimulation at 10-Hz frequency in Mecp2 KO neurons and compared the results with responses recorded from control, Hdac1, and Hdac2 KO neurons. After the initial facilitation in response amplitude seen during high-frequency stimulation, long trains of action potentials applied at 10 Hz typically depress neurotransmission, which is believed to be due to a rapid decrease in the number of vesicles available for release from presynaptic terminals. When the kinetics of EPSC and IPSC depression were examined in Mecp2 KO neurons, we observed a more rapid depression of EPSCs compared with controls (Fig. 4, A and B), whereas the depression of IPSCs was unchanged (Fig. 4, C and D). This observation provides further confirmation that alterations in presynaptic function occur specifically in excitatory, but not inhibitory, synapses. Interestingly, whereas in Hdac2 KO neurons there was a small but significant reduction in the degree of facilitation, in Hdac1 neurons changes seen in EPSCs did not reach significance. Under all conditions, we did not detect a significant difference in IPSC response depression in response to 10-Hz stimulation compared with controls (Fig. 4, A–D). The numerical EPSC and IPSC depression data before normalization are included in the Supplemental Material (Supplemental Fig. S4). These results suggest that while initial excitatory Pr appears to be affected by the loss of MeCP2, HDAC1, and HDAC2, more prolonged stimulation begins to differentiate Mecp2 KO neurons, perhaps due to alterations in the number of synaptic vesicles available for release over brief periods of high-frequency stimulation that lead to changes in short-term synaptic plasticity. This discrepancy may also suggest a functional redundancy between HDAC1 and HDAC2 in the regulation of evoked neurotransmitter release.

Fig. 4.

EPSC and IPSC depression in response to 10-Hz stimulation in Mecp2, Hdac1, and Hdac2 KO neurons. A: representative traces of normalized evoked EPSCs during the first 2 s of 10-Hz field stimulation. B: average EPSC amplitudes recorded from Mecp2 KO, Hdac1 KO, Hdac2 KO, and control neurons measured during the first 2 s of 10-Hz field stimulation. Control, n = 44 neurons/10 coverslips; Mecp2 KO, n = 11 neurons/3 coverslips; Hdac1 KO, n = 11 neurons/3 coverslips; Hdac2 KO, n = 17 neurons/3 coverslips. While all three KO neurons showed a reduction in the second response amplitude compared with controls, the depression of Mecp2 KO EPSCs was significantly faster during the first second. **P < 0.05 for control neurons compared with Mecp2 KO and Hdac2 KO neurons; *P < 0.05 for Mecp2 KO neurons compared with control neurons. C: representative traces of normalized evoked IPSCs during the first 2 s of 10-Hz field stimulation. D: average IPSC amplitudes recorded from Mecp2 KO, Hdac1 KO, Hdac2 KO, and control neurons showed no differences in response depression during 10-Hz stimulation. Control, n = 46 neurons/9 coverslips; Mecp2 KO, n = 9 neurons/2 coverslips; Hdac1 KO, n = 17 neurons/3 coverslips; Hdac2 KO, n = 12 neurons/4 coverslips.

Rate of MK-801 block of NMDA EPSCs is faster in Mecp2, Hdac1, and Hdac2 KO hippocampal neurons.

An additional method for measuring Pr from excitatory synapses is the MK-801 blocking assay (Hessler et al. 1993; Rosenmund et al. 1993). MK-801 is a use-dependent channel blocker of NMDA receptors, and the blocking rate of NMDA EPSCs is considered proportional to Pr. To probe further into the roles of MeCP2, HDAC1, and HDAC2 in the control of excitatory Pr, we treated our cultures with MK-801 during 0.2-Hz field stimulation and recorded NMDA EPSCs. In Mecp2 KO neurons, the rate of MK-801 block was significantly faster than controls, as were the rates of MK-801 block in both Hdac1 and Hdac2 KO neurons, although the effects were much smaller (Fig. 5, A and B). The numerical decay of evoked NMDA EPSC amplitudes in the presence of MK-801 data before normalization is included in the Supplemental Material (Supplemental Fig. 5). Besides implicating enhanced presynaptic Pr from excitatory synapses, an increase in the blocking rate of NMDA EPSCs could also be due to a possible alteration in the properties of postsynaptic NMDA receptors. While we cannot entirely rule out this possibility, previous research on different brain regions from both RTT patients and Mecp2 mutant mice has indicated alterations in NMDA receptors (Asaka et al. 2006; Blue et al. 1999); this finding strongly agrees with the short-term depression data suggesting increased excitatory Pr with the loss of MeCP2, HDAC1, and HDAC2 function.

Fig. 5.

Decay of evoked N-methyl-d-aspartate (NMDA) EPSC amplitudes in the presence of MK-801 (10 μM) in control, Mecp2 KO, Hdac1 KO, and Hdac2 KO neurons. A: representative normalized traces of evoked EPSCs (1st, 10th, and 20th) before and during the perfusion of MK-801 at 0.2-Hz field stimulation. B: average NMDA EPSC amplitudes measured during 0.2-Hz field stimulation. Control, n = 3 neurons/3 coverslips; Mecp2 KO, n = 4 neurons/4 coverslips; Hdac1 KO, n = 3 neurons/3 coverslips; Hdac2 KO, n = 4 neurons/4 coverslips. Decay of the NMDA-evoked currents was fit to a quadratic polynomial for control neurons and a single exponential for Mecp2 KO, Hdac1 KO, and Hdac2 KO neurons. The KO neurons showed reductions in response amplitudes at rates significantly faster compared with controls. **P < 0.001 for control neurons compared with Mecp2 KO neurons; *P < 0.05 for control neurons compared with Hdac1 KO and Hdac2 KO neurons.

DISCUSSION

Animal models of RTT have consistently demonstrated alterations in the balance between excitation and inhibition in the brain (Chao et al. 2007; Dani et al. 2005; Nelson et al. 2006; Tropea et al. 2009). Previous studies have suggested that a decrease in tonic excitation may be responsible for the underlying behavioral phenotypes observed in Mecp2 KO mice; however, seizures are often observed in RTT patients. Here, we showed an increase in the ratio between excitation and inhibition in response to short-term, high-frequency stimulation as well as evidence for enhanced excitatory Pr in Mecp2 KO hippocampal neurons. Our data, taken together with the previous findings, indicate that basal, spontaneous activity in Mecp2 KO neurons displays a decrease in excitation compared with inhibition while evoked synaptic transmission shows an increase in excitatory drive.

The disparate findings between excitation and inhibition balance in spontaneous and evoked transmission is surprising but not without precedent. Spontaneous miniature and evoked synaptic currents have been shown to play different roles in the brain and have even been suggested to occur via distinct presynaptic and postsynaptic mechanisms (Atasoy et al. 2008; Fredj and Burrone 2009; Rothwell 2010; Sara et al. 2005; Sutton and Schuman 2009). Therefore, it is plausible to expect potentially opposing effects of regulatory mechanisms on these two forms of synaptic transmission (Ramirez and Kavalali 2011). Recent studies using glutamate uncaging on Mecp2 KO cortical synapses revealed a reduction in excitatory synaptic strength (Wood et al. 2009; Wood and Shepherd 2010), and a study of autaptic hippocampal neurons found a reduction in evoked EPSC magnitude attributed to a reduction in excitatory synapse number (Chao et al. 2007). Although these recent observations disagree with the basic premise of our findings (i.e., increased evoked excitatory drive), we propose a presynaptic origin for our observations as we detected alterations in short-term synaptic plasticity consistent with an increase in Pr of the excitatory neurotransmitter glutamate. Moreover, a recent study has shown that neurons from Mecp2 mutant mice were hyperexcitable in the CA1 and CA3 regions of hippocampal slices (Calfa et al. 2011). While these authors attributed this hyperexcitability to a higher level of spontaneous firing of CA3 pyramidal neurons, they also observed similar changes in presynaptic properties that indicated enhanced excitatory Pr.

The perturbations in short-term synaptic plasticity seen in Mecp2 KO hippocampal neurons could have a significant impact on the firing of action potentials and the timing of synaptic inputs within a circuit (Abbott and Regehr 2004). Indeed, a recent study discovered slower hippocampal intrinsic rhythmic activity in Mecp2-deficient mice, and while there were only subtle changes seen in the basal oscillatory activity of neurons, the hippocampal network was prone to hyperexcitability, particularly when stimulated with an excitatory stimulus (Zhang et al. 2008). Our data, taken together with the previous findings from other laboratories, highlight the presence of decreased spontaneous activity along with increased excitability upon stimulation in hippocampal neurons of Mecp2 KO mice. The high prevalence of epilepsy in RTT patients (Jian et al. 2007) and the presence of tremors and seizure activity in Mecp2-deficient mice (Chen et al. 2001; Shahbazian et al. 2002) give a strong indication for hyperexcitability in cortical and subcortical brain regions. These observations are in accordance with a mechanism in which an increase in the ratio between excitatory and inhibitory action potential-driven activity is sufficient to cause this hyperexcitable phenotype.

The function of MeCP2 as a transcriptional repressor, in part, relies on its association with the proteins HDAC1 and HDAC2 (Nan et al. 1998). Interestingly, histone deacetylation has also been implicated in the control of excitation and inhibition in the brain. For instance, the HDAC inhibitor valproic acid (VPA) is a commonly used treatment for epilepsy (Phiel et al. 2001). A recent study of rat cortical primary cultures showed that VPA treatment upregulates genes involved in neuronal excitability and downregulates genes involved in neuronal inhibition, suggesting that VPA could mediate an imbalance between excitation and inhibition, in favor of excitation, via its activity as a HDAC inhibitor (Fukuchi et al. 2009). The question remains, however, as to which specific HDACs are involved in mediating these changes. A previous study examining the roles of HDAC1 and HDAC2 in synapse maturation and function revealed a developmental switch in the imbalance between excitation and inhibition after the loss of these proteins (Akhtar et al. 2009). Since this study only looked at spontaneous miniature synaptic transmission in hippocampal neurons, we found it pertinent to assess how the ratio of excitation to inhibition during short-term synaptic transmission might be affected by the loss of HDAC1 or HDAC2.

In the present study, we used a virus-mediated approach to knockdown the expression of either HDAC1 or HDAC2 in primary hippocampal cultures. Both HDAC1 and HDAC2 knockdown appeared to affect initial excitatory Pr, seen as a reduction in PPRs similar to that seen in Mecp2 KO neurons. Both KO neurons also mimicked the increased blocking rate of NMDA EPSCs by MK-801 seen in Mecp2 KO neurons. However, only the loss of HDAC2, but not HDAC1, resulted in an increase in excitatory evoked response amplitude, and neither Hdac1 nor Hdac2 KO neurons displayed the same increase in response depression kinetics as those seen with deletion of MeCP2. These results suggest that with respect to short-term synaptic plasticity, a reduction in HDAC2 activity can increase excitatory drive in hippocampal neurons, indicating the possibility of HDAC2 and MeCP2 working together to control excitatory synaptic strength. With respect to synaptic vesicle Pr, our results suggest that both HDAC1 and HDAC2 can act together with MeCP2 to regulate excitatory Pr during brief bursts of activity; however, loss of function of these HDACs failed to mimic the Mecp2 KO phenotype during sustained high-frequency stimulation, suggesting a potential additional regulatory role of MeCP2 in setting presynaptic efficacy independent of HDACs. Here, it is important to note that the alterations in synaptic transmission seen after the loss of MeCP2, HDAC1, or HDAC2 are all specific to excitatory synapses with little or no effect on dynamics or basal levels of evoked inhibitory neurotransmission.

Animal models with disruption of the Mecp2 gene produce behavioral symptoms that closely mimic human symptoms of RTT. Previous studies examining spontaneous synaptic activity have reported selective disruption of excitatory neurotransmission, thereby leading to an imbalance in the ratio of excitatory versus inhibitory neurotransmission. Altered excitatory-inhibitory transmission based on spontaneous transmission measurements has also been reported in other animal models of autism, such as fragile X mental retardation I KO mice for fragile X syndrome and neuroligin-3 mutant mice modeling a rare familial form of autism (Gibson et al. 2008; Tabuchi et al. 2007). Collectively, these data have led to the hypothesis that restoration of this excitatory-inhibitory balance, by either increasing glutamatergic activity or decreasing GABAergic transmission, may be a potential treatment for impairments seen in these autism models. GABAA receptor antagonists have previously been shown to reverse cognitive impairments seen in a mouse model of the related neurodevelopmental disorder Down syndrome (Fernandez et al. 2007; Rueda et al. 2008). One caveat of this approach for RTT patients is concern over whether a GABAergic antagonist or glutamatergic agonist would be a suitable treatment for a population with increased risk for seizures. The increase in the ratio between excitation and inhibition in response to short-term, high-frequency stimulation in Mecp2 KO hippocampal neurons we observed may be translated into a model in which one may expect that increasing GABAergic activity and decreasing glutamatergic activity may in fact be more beneficial to alleviate some of the symptomology associated with these disorders, including seizures. Behavioral characterization of Mecp2 mutant mice with appropriate pharmacological agents to test this hypothesis will be an important next step in elucidating how the disruption of excitation/inhibition may impact the phenotype of RTT and related disorders.

In conclusion, the results of this study give additional support to the proposed imbalance between excitation and inhibition in the brains of Mecp2 KO mice, revealing an increase in the ratio of excitatory to inhibitory activity in response to action potential firing as well as an enhancement in excitatory presynaptic Pr. None of these changes were seen in inhibitory synaptic activity after the loss of MeCP2. The knockdown of HDAC1 and HDAC2 in hippocampal neurons was able to replicate some of these findings, particularly after the loss of HDAC2. Given that HDAC2 appears to have a higher neuron-specific expression pattern than HDAC1 in the rodent hippocampus (Broide et al. 2007; MacDonald and Roskams 2008), future research into the effects of MeCP2-dependent transcriptional repression on excitatory versus inhibitory synaptic function in hippocampal neurons may choose to focus on MeCP2's association with HDAC2.

GRANTS

This work was supported by National Institute for Mental Health Grants MH-081060 (to L. M. Monteggia) and MH-066198 (to E. T. Kavalali) as well as the Division of Basic Sciences Training Program of University of Texas Southwestern Medical Center (to E. D. Nelson). E. T. Kavalali is an Established Investigator of the American Heart Association.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

ACKNOWLEDGMENTS

The authors thank Eric Olson (University of Texas Southwestern) for generously providing the floxed Hdac1 and Hdac2 mice and the members of the Monteggia laboratory for advice and discussion.

Footnotes

  • 1 Supplemental Material for this article is available at the Journal of Neurophysiology website.

REFERENCES

View Abstract