The most common inherited monogenetic cause of intellectual disability is Fragile X syndrome (FXS). The clinical symptoms of FXS evolve with age during adulthood; however, neurophysiological data exploring this phenomenon are limited. The Fmr1 knockout (Fmr1 KO) mouse models FXS, but studies in these mice of prefrontal cortex (PFC) function are underrepresented, and aging linked data are absent. We studied synaptic physiology and activity-dependent synaptic plasticity in the medial PFC of Fmr1 KO mice from 2 to 12 months. In young adult Fmr1 KO mice, NMDA receptor (NMDAR)-mediated long-term potentiation (LTP) is intact; however, in 12-month-old mice this LTP is impaired. In parallel, there was an increase in the AMPAR/NMDAR ratio and a concomitant decrease of synaptic NMDAR currents in 12-month-old Fmr1 KO mice. We found that acute pharmacological blockade of mGlu5 receptor in 12-month-old Fmr1 KO mice restored a normal AMPAR/NMDAR ratio and LTP. Taken together, the data reveal an age-dependent deficit in LTP in Fmr1 KO mice, which may correlate to some of the complex age-related deficits in FXS.
Fragile X syndrome (FXS) is a major form of inherited intellectual disability (ID) caused by the transcriptional silencing of the gene FMR1. The disorder is characterized by a series of physical, cognitive, and emotional symptoms that include social deficits and a diminished ability to learn. Frequently found comorbid with other complex neurological disorders, including attention deficit hyperactivity disorder, epilepsy, and anxiety (Bailey et al. 2008), FXS is also the principal monogenetic cause of autism spectrum disorders.
The FMR1 gene product, fragile X mental retardation protein (FMRP), is an mRNA-binding protein, which is absent in the FXS brain throughout development. Principally, FMRP is found in neuronal spines where it functions as a translational repressor of a subpopulation of mRNAs linked to synaptic function. Upon suitable synaptic stimulation, FMRP is released and local mRNA translation proceeds. In the hippocampus, mGlu5 receptor (mGluR5) activation engages an important synaptic signal transduction pathway leading to local translation (Ronesi and Huber 2008). Loss of FMRP results in dysregulation of this pathway and is the basis of the mGluR theory of FXS (Bear et al. 2004). This is supported by the observation that, in a mouse model of FXS, mGluR5 antagonists not only restore many of the associated changes in the synaptic structure, development, and regulation, but also some of the behavioral deficits (Dölen et al. 2007; Michalon et al. 2012). mGluR5 antagonists have been a major focus for the development of therapeutics to treat some of the symptoms associated with FXS (Krueger and Bear 2011).
Prefrontal cortex (PFC) malfunctions are a common denominator in several neuropsychiatric diseases (Goto et al. 2010; Huntley and Howard 2010). Loss of executive function and inhibitory control are particularly prominent in the ID associated with FXS, and there is evidence of increasing ID with age (Lachiewicz et al. 1987; Wright-Talamante et al. 1996; Cornish et al. 2001; Utari et al. 2010; Schneider et al. 2013). The exact contribution of FMRP in PFC functions and in particular the synaptic basis of the age-dependent defects reported in FXS remains obscure. In the Fmr1 knockout (Fmr1 KO) mouse model of FXS, there is evidence of changes in PFC-mediated behaviors in young adult animals (Dölen et al. 2007; Gantois et al. 2013; Kramvis et al. 2013; Sidorov et al. 2014), and biochemical analysis of PFC synapses suggests alterations in glutamate receptors and accessory proteins in excitatory synapses (Krueger et al. 2011). During juvenile development, a hyperconnected state exists between layer V pyramidal neurons, similar to that found in other models of autism (Rinaldi et al. 2008; Testa-Silva et al. 2012). However, the increased connectivity between layer V neurons is normalized as Fmr1 KO mice reach adolescence. In early adulthood, it has been reported that both spike-timing-dependent and D1 receptor-mediated long-term potentiation (LTP) are modified in superficial layers of the Fmr1 KO PFC (Zhao et al. 2005; Meredith et al. 2007; Wang et al. 2008), but whether this is indicative of a general PFC dysfunction is unclear. There is evidence from the piriform cortex that an age-linked phenotype may be present in the Fmr1 KO model. In early adulthood, Fmr1 KO mice express stimulus-evoked LTP; however, an increasing loss in LTP is found as mice reach middle age and after (Larson et al. 2005). Whether a similar phenotype is present in the PFC and other brain regions associated with the cognitive decline found in FXS remains unexamined.
Here, we evaluated LTP in pyramidal neurons from deep layers (layer 5) of the medial PFC (mPFC) neurons in adult Fmr1 KO mice, in both early and late adulthood. We report that LTP and NMDA receptor (NMDAR) function is impaired in the PFC of 12-month-old mice, but normal at a younger age. This age-specific reduction of synaptic plasticity may underlie part of the multiple cognitive and behavioral deficits described in aging FXS patients. Furthermore, specific inhibition of mGluR5 restored a normal AMPAR/NMDAR ratio and rescued LTP in 12-month-old Fmr1 KO mice, extending the mGluR theory of FXS.
Materials and Methods
All mice were group-housed with 12 h light/dark cycles in compliance with the European Communities Council Directive (2010/63/EU) for the Care and Use of Laboratory Animals. Male Fmr1 KO2 mice, null for protein, and mRNA on a C57Bl6/J genetic background were used, with wild-type (WT) littermates as a control group. Young adult mice (2 months) were 9–12 weeks and mature adult (12 months) were 12 ± 2 months of age.
Mice were anesthetized with isoflurane and decapitated according to institutional regulations. The brain was sliced (300 µm) in the coronal plane with a vibratome (Integraslice, Campden Instruments, Loughborough, UK) in a sucrose-based solution at 4 °C (in mM: 87 NaCl, 75 sucrose, 25 glucose, 5 KCl, 21 MgCl2, 0.5 CaCl2, and 1.25 NaH2PO4). Slices were allowed to recover for 60 min at 32–35 °C in artificial cerebrospinal fluid (aCSF) containing 126 mM NaCl, 2.5 mM KCl, 2.4 mM MgCl2, 1.2 mM CaCl2, 18 mM NaHCO3, 1.2 mM NaH2PO4, and 11 mM glucose, equilibrated with 95% O2/5% CO2. Slices were then maintained at 22 ± 2 °C until recording.
Whole-cell patch-clamp and extracellular field recordings were made from layer 5 pyramidal cells in the prelimbic cortex (Lafourcade et al. 2007). For recording, slices were superfused (2 mL/min) with aCSF. All experiments were done at 32–35 °C. The recording aCSF contained picrotoxin (100 µM) to block GABAA receptors. To evoke synaptic currents, 100–200 µs stimuli were delivered at 0.1 Hz through an ACSF-filled glass electrode positioned dorsal to the recording electrode in layer V. For extracellular field experiments, the recording pipette was filled with aCSF. The glutamatergic nature of the field excitatory postsynaptic potential (fEPSP) was confirmed at the end of the experiments using the ionotropic glutamate receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 20 µM) that specifically blocked the synaptic component without altering the non-synaptic component. LTP was induced by a theta burst protocol as previously described (Iafrati et al. 2014).
For whole-cell patch-clamp, pyramidal neurons in PFC layer 5 were visualized using an infrared microscope (BX-50, Olympus). Experiments were made with electrodes containing 143 mM cesium methanesulfonate (CH3O3SCs, voltage clamp) or 145 mM potassium gluconate (KGlu, current clamp and voltage clamp), with 5 mM NaCl, 1 mM MgCl2, 1 mM EGTA, 0.3 mM CaCl2, 2 mM Na2+-ATP, 0.3 mM Na+-GTP, 10 mM glucose buffered with 10 mM HEPES, pH 7.3, osmolarity 290 mOsm. Electrode resistance was 3–5 MΩ. If access resistance was >20 MΩ or changed by >20%, the experiment was rejected. Junction potentials were not correct. To perform current–voltage (I–V) curves and to test neuronal excitability, a series of hyperpolarizing and depolarizing current steps were applied immediately after breaking in the cell.
Spontaneous AMPAR- and NMDAR-mediated excitatory post-synaptic currents (EPSC) were recorded in the voltage-clamp configuration at −70 or +40 mV in the presence of 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione to block AMPAR-mediated events (NBQX, 10 µM).
Data Acquisition and Analysis
Data were recorded on a MultiClamp700B (Axon Instruments), filtered at 2 or 10 kHz (spike characterization), digitized (10 or 50 kHz, DigiData 1440A, Axon Instrument), collected using Clampex 10.2, and analyzed using Clampfit 10.2 (all from Molecular Device, Sunnyvale, CA, USA). AMPAR/NMDAR ratios were calculated from monosynaptic EPSCs by dividing the averaged EPSC (30 events) amplitude recorded at −70 mV, by the averaged EPSC (30 events) current 30 ms after the peak AMPAR-mediated event at +40 mV, a time point where rapidly gated AMPAR currents are negligible. Analysis of both area and amplitude of fEPSPs and EPSCs was performed (graphs depict amplitudes for patch-clamp experiments and areas for field recordings). The magnitude of LTP was calculated 35–40 min after theta burst stimulation (TBS) as a percentage of baseline responses. LTP data were collated and analyzed from throughout the lifetime of the project, except during experimental treatments where a subset of interleaved control experiments was analyzed. Spontaneous NMDAR- and AMPAR-mediated EPSCs were analyzed with Axograph X (Axograph). Spontaneous events were identified by passing a variable amplitude prototypical AMPAR (τrise = 0.5 ms, τdecay = 3 ms; 10 ms window; noise threshold: −3) or NMDAR (τrise = 3 ms, τdecay = 50 ms; 50 ms window; noise threshold: +5) event over the recording as previously described (Iafrati et al. 2014). Overlapping events were ignored. Tau-weighted (τw) was calculated from a double exponential fit of the NBQX-isolated NMDAR currents at +40 mV with the formula: τw = (τ1* a1) + (τ2* a2), where a1 and a2 are the relative amplitudes of the 2 exponential components. Intrinsic properties were quantified in MATLAB (MathWorks) using custom routines. Sag values were calculated from a hyperpolarizing current clamp step leading to a minimum membrane potential of at least −100 mV (typically −400 pA) and were expressed as a percent of the total change in membrane potential. Statistical analysis of data was performed with GraphPad Prism (GraphPad Software, Inc., La Jolla, CA, USA) using tests indicated in the main text after outlier subtraction.
All values are given as mean ± standard error, statistical significance was set at *P < 0.05 and **P < 0.01, and n values correspond to individual animals.
Drugs were added at the final concentration to the recording aCSF media from DMSO (Picrotoxin and CNQX) or water [2-Methyl-6-(phenylethynyl)pyridine (MPEP) and NBQX] stocks. Picrotoxin was from Sigma (St. Quentin Fallavier, France), and MPEP, CNQX, and NBQX were from the National Institutes of Mental Health's Chemical Synthesis and Drug Supply Program (Rockville, MD, USA).
Theta Burst-Induced LTP Is Selectively Impaired in the PFC of 12-Month-Old Fmr1 KO Mice
In the mouse mPFC, moderate TBS induces a robust NMDAR-mediated LTP of deep layer pyramidal excitatory synapses throughout life (Iafrati et al. 2014). Previous reports suggest that late onset deficits in LTP may be present in a subset of cortical structures in Fmr1 KO mice (Larson et al. 2005; Gocel and Larson 2012); however, whether a similar late onset deficit in LTP is present in the mPFC is unknown. We asked if a parallel deficit in LTP might be present in the Fmr1 KO mouse. Taking a cross-sectional approach, we challenged mPFC brain slices with a short TBS protocol at 2 and 12 months. As expected, TBS induced a robust LTP in WT mice in young adulthood, and in contrast to reported deficits in layer II/III PFC LTP (Zhao et al. 2005; Meredith et al. 2007), we were also able to induce LTP in Fmr1 KO littermates (Fig. 1A). The cumulative distribution plot of individual experiments from both groups showed a similar potentiation in fEPSP post TBS (Fig. 1B; P = 0.136, Kolmogorov–Smirnov test); furthermore, the mean percent LTP was the same in both groups (Fig. 1C; percent LTP: WT 23.97 ± 7.20%; 12-month-old Fmr1 KO 21.20 ± 4.95%; P = 0.874, Mann–Whitney U test). In line with the idea that the PFC remains plastic throughout normal aging, in 12-month-old WT mice TBS induced an increase in fEPSPs identical to younger animals. However, when we challenged 12-month-old Fmr1 KO animals with TBS-LTP, we observed a significant impairment (Fig. 1D). Although post-tetanus potentiation was observed in both genotypes (compare the first 3 min post TBS, Fig. 1D), the potentiation of the fEPSP was much reduced in 12-month-old Fmr1 KO animals for the duration of the recording. Cumulative probability distribution plots of the normalized individual experiments further showed that the potentiation in fEPSPs was left-shifted and smaller in 12-month-old Fmr1 KO mice (Fig. 1E; P = 0.003, Kolmogorov–Smirnov test). Furthermore, the mean percent LTP (Fig. 1F) was reduced by approximately 50% (percent LTP: WT 20.81 ± 3.96%; 12-month-old Fmr1 KO 10.88 ± 2.69%; P = 0.024, Mann–Whitney U test). These data reveal that the ability to express LTP is impaired in 12-month-old Fmr1 KO mice, but unaffected in young adult Fmr1 KO mice.
Age-Specific Augmentation of the AMPAR/NMDAR Ratio in 12-Month-Old PFC Neurons
In the valproate rat model of autism, LTP and the relative synaptic weight of AMPAR to NMDAR covariate in an age-dependent manner (Rinaldi et al. 2007; Walcott et al. 2011; Martin and Manzoni 2014). We recently reported that the reduction of LTP in adult, but not juvenile valproate-treated, rats was accompanied by a large increase in the ratio between AMPAR current and NMDAR synaptic current, the AMPAR/NMDAR ratio (Martin and Manzoni 2014). Consequently, we searched for changes in the relative weight of AMPAR and NMDAR at mPFC synapses in young and 12-month-old Fmr1 KO mice.
Using AMPAR currents at −70 mV as a synaptic control, we recorded evoked NMDAR currents at +40 mV (Fig. 2A). In young adult mice, the AMPAR/NMDAR ratio was similar in both groups of animals (WT 2.44 ± 0.34, n = 10; Fmr1 KO 2.32 ± 0.26, n = 6; P = 0.946, Mann–Whitney U test). However, in 12-month-old mice, we found a significant increase in the AMPAR/NMDAR ratio in the Fmr1 KO compared with age-matched controls (Fig. 2B; WT 2.34 ± 0.30, n = 12, Fmr1 KO 3.87 ± 0.46, n = 12; P = 0.021, Mann–Whitney U test). The relative magnitude of the response at +40 mV indicated a probable reduction in NMDAR currents (Fig. 2A, right). The AMPAR/NMDAR ratio in 12-month-old Fmr1 KO mice was also increased relative to young Fmr1 KO animals, whereas in WT mice no age-linked change in the AMPAR/NMDAR ratio was observed.
Decreased NMDAR Synaptic Currents in 12-Month-Old Fmr1 KO Adult PFC Neurons
We first assessed synaptic function by recording evoked fEPSP. The input/output relationship was similar in 12-month-old WT mice and Fmr1 KO littermates, showing that the overall strength of mPFC glutamatergic synapses was unaltered (Supplementary Fig. 1). Visually identifying pyramidal neurons in layer 5, we measured spontaneous activity. Consistent with the absence of a change in input/output properties, spontaneous activity was similar in both WT and Fmr1 KO 12-month-old mice (Fig. 3A). The mean amplitude and distribution of events of spontaneous EPSCs (sEPSC) was almost identical between Fmr1 KO animals and WT littermates (Fig 3B; P = 0.966, Kolmogorov–Smirnov test). Furthermore, the mean interval between sEPSCs was unchanged between 12-month-old Fmr1 KO and WT littermates and the distribution of events was similar, indicating that spontaneous events occur at the same frequency (Fig 3C; P = 0.890, Kolmogorov–Smirnov test). At resting membrane potential, sEPSCs are predominantly due to AMPAR channel openings. Therefore, to measure NMDAR-mediated spontaneous events, we recorded sEPSCs at +40 mV, while blocking AMPAR-mediated currents. Surprisingly, in contrast to AMPAR events, we found that NMDAR spontaneous events from pyramidal neurons from Fmr1 KO animals were less frequent and of a smaller amplitude than WT littermates (Fig. 3D). Analyzing spontaneous NMDAR events by comparing their cumulative distributions, the amplitude of events in Fmr1 KO animals was left-shifted, indicating a reduced number of large events compared with WT controls (Fig. 3E; P = 0.002, Kolmogorov–Smirnov test). There was also a reduction in the mean spontaneous amplitude, which was 12% lower in Fmr1 KO slices (P = 0.049; Mann–Whitney U test). Furthermore, the recordings indicated that the relative interval between events was particularly affected in 12-month-old Fmr1 KO animals with a strong right-shift in the cumulative plot (Fig 3F; P < 0.001, Kolmogorov–Smirnov test). This resulted in a highly significant increase of the interval between NMDAR-mediated events in 12-month-old Fmr1 KO animals (P = 0.005, Mann–Whitney U test). When we recorded spontaneous NMDAR-mediated events in young adult PFC, there was no significant difference in the amplitude (WT: 41.95 ± 1.55 pA, n = 5; Fmr1 KO: 38.86 ± 0.75 pA, n = 5; P = 0.135, Mann–Whitney U test) nor in the interevent interval (WT: 0.74 ± 0.11 s, n = 5; Fmr1 KO: 0.64 ± 0.05 s, n = 5; P= 0.500, Mann–Whitney U test) between Fmr1 KO and WT littermates.
Thus, in contrast to the piriform cortex, where a stable reduction in NMDAR synaptic currents is seen throughout adulthood in the Fmr1 KO mouse (Gocel and Larson 2012), our data demonstrate that, in the PFC, the alteration of synaptic NMDAR currents is restricted to 12-month-old Fmr1 KO mice. Furthermore, we found that the decay kinetics of NMDAR-mediated EPSCs in these 12-month-old Fmr1 KO mice and WT littermates were similar (WT: τw = 88.20 ± 2.97, n = 10; Fmr1 KO: τw = 96.70 ± 2.25, n = 8; P = 0.121 Mann–Whitney U test). This suggests that the reduction in LTP and the increase in the AMPAR/NMDAR ratio, rather than be due to a change in NMDAR kinetics or subunit composition, derive from a general reduction of NMDAR-mediated synaptic currents in these older animals.
Normal Intrinsic Properties of PFC Pyramidal Neurons in Aged Fmr1 KO Mice
Given our observed age-dependent LTP deficits in the mPFC of 12-month-old Fmr1 KO mice, we wondered if there might be wider functional consequences to deep layer principal neurons due to this aberrant activity in 12-month-old mice. In the hippocampus, pyramidal cells from Fmr1 KO mice have increased potassium channel function (Gross et al. 2011; Lee et al. 2011; Brager et al. 2012), and thus we first assessed the intrinsic properties of layer 5 mPFC neurons. Irrespective of their genotype, all recorded pyramidal neurons showed similar membrane profiles in response to a series of somatic current steps (Fig. 4A). Furthermore, the membrane capacitance (Cm: WT 113.0 ± 10.95 pF, n = 14; Fmr1 KO 100.6 ± 6.463 pF, n = 23) and resting membrane potential (Vm: WT −71.71 ± 1.35 mV, n = 14; Fmr1−/− −70.45 ± 0.75 mV, n = 22) were unchanged between genotypes. The prominent sag in response to hyperpolarizing steps found in layer 5 pyramidal neurons was also similar in WT and Fmr1 KO 12-month-old animals (Fig. 4B), indicative of normal Ih function. Spike profiles from individual layer 5 pyramidal neurons were induced by a standard current step injection (Fig. 4C). Spike threshold, spiking frequency, and spike adaption were comparable between the 2 groups. Thus, Fmr1 KO in 12-month-old mice does not impact the gross intrinsic properties of PFC pyramidal.
Inhibition of mGluR5 Restores Normal LTP and AMPAR/NMDAR Ratio in 12-Month-Old Fmr1 KO Mice
One of the most promising therapeutic routes for the treatment of FXS is the reduction in mGluR5 function (Michalon et al. 2012). This strategy was developed from the now well-described hyperactivity of the receptor and its downstream partners in the hippocampus of Fmr1 KO animals (Ronesi and Huber 2008). In some cases, LTP impairments in Fmr1 KO mice can be rescued by blocking group I mGluR (Xu et al. 2012; Chen et al. 2014); but see Suvrathan et al. (2010). Therefore, we wondered if reducing mGluR5 function might allow subsequent induction of LTP in the PFC of 12-month-old Fmr1 KO. Using MPEP to block mGluR5 activity, we preincubated Fmr1 KO mPFC slices for at least 1 h in the drug before inducing TBS-LTP. With this treatment, a recovery of LTP was apparent compared with matched control Fmr1 KO mice (Fig 5A). Expressing the LTP as a percent change in fEPSP 35–40 min post TBS, a significant increase in LTP is present in MPEP-treated slices (Fig 5A, insert; Fmr1 KO: 9.96 ± 3.75%, Fmr1 KO MPEP: 20.54 ± 3.85%, P = 0.040, Mann–Whitney U test). Importantly, the LTP-induced increase in fEPSP has a magnitude similar to 12-month-old WT animals and is not significantly different from WT MPEP exposed slices (Fig. 5B). Using the same MPEP treatment regime, we also recorded AMPAR/NMDAR ratios from layer 5 pyramidal neurons from 12-month-old WT and Fmr1 KO mice. Similar to TBS-LTP, we saw an MPEP-mediated effect on the AMPAR/NMDAR ratio (Fig. 5C). Notably, the significant increase in the AMPAR/NMDAR ratio found between WT and Fmr1 KO was not present with MPEP treatment. Thus, by modulating mGluR5 function in 12-month-old Fmr1 KO animals, it is possible to recover normal LTP in the PFC, most likely due to a normalization of NMDAR function.
Here, we delineate changes in PFC synaptic plasticity in a mouse model of FXS between early and late adulthood. We found that, during early adulthood, Fmr1 KO mice express normal LTP, but that during middle to late adulthood Fmr1 KO mice show significantly reduced potentiation in response to LTP stimulation patterns. Thus, in the absence of FMRP, 12-month-old mice are defective for PFC LTP, while young adults are spared. Concurrent with the weakening in LTP, we found a reduction in NMDAR activity in 12-month-old Fmr1 KO mice, demonstrated by both a relative increase in AMPAR/NMDAR ratio and a reduction in spontaneous NMDAR currents. Finally, acute treatment with the mGluR5 antagonist MPEP reduced the AMPAR/NMDAR in 12-month-old Fmr1 KO mice and restored LTP.
Deficits in synaptic plasticity have been widely reported in Fmr1 KO mice. However, the observed changes in LTP are variable, linked either to stimulation protocol or specific brain region (Sidorov et al. 2013). In contrast to our findings in layer 5 mPFC, other groups have found deficits in LTP during early adulthood in layer II/III neurons from Fmr1 KO mice (Zhao et al. 2005; Meredith et al. 2007). This layer-specific loss in LTP appears to be linked to threshold changes in calcium signaling in the postsynaptic cell and may be restored through environmental enrichment (Meredith et al. 2007) or modulation of the dopaminergic system (Wang et al. 2008). In contrast to 12-month-old layer 5 neurons, where we found reduced NMDAR function, no changes in NMDAR-mediated currents are found in these young adult neurons (Zhao et al. 2005). Instead, the age-linked loss of LTP in 12-month-old Fmr1 KO mice seen here appears to share greater similarities with deficits found in the dentate gyrus (DG) and piriform cortex. High-frequency stimulation elicits a robust LTP in CA1 pyramidal neurons, but not the DG in Fmr1 KO mice (Eadie et al. 2012; Franklin et al. 2014). In common with the biochemical reduction in NMDAR and accessory proteins found in the PFC (Krueger et al. 2011), DG neurons have lower expression of NMDAR subunits and reduced NMDAR-mediated currents (Yun and Trommer 2011; Bostrom et al. 2015). Likewise, layer Ib neurons of the lateral piriform cortex show a reduction of TBS-LTP at 12 months in Fmr1 KO mice, but not in younger animals (Larson et al. 2005). In parallel to our findings, piriform cortex neurons also show a decline in NMDAR function (Gocel and Larson 2012), correlating the reduction in LTP with the loss of NMDAR-mediated currents.
In juvenile Fmr1 KO mice, downregulation of AMPAR-mediated responses relative to NMDAR is found in both the cortex and the hippocampus (Hu et al. 2008; Pilpel et al. 2009; Till et al. 2012), affecting both somatosensory critical period patterning and synaptic plasticity (Hu et al. 2008; Pilpel et al. 2009; Harlow et al. 2010). However, in early adulthood, cortical synaptic AMPAR and NMDAR expression and function appear to normalize in the cortex (Zhao et al. 2005; Schütt et al. 2009). In agreement, we found that PFC spontaneous NMDAR synaptic currents are similar between Fmr1 KO and WT littermates in early adulthood. However, in 12-month-old mice, we found a significant reduction in both spontaneous NMDAR current amplitude and frequency. The reduction in spontaneous NMDAR frequency is likely due to the drop in event amplitude to below our detection threshold, since we found no evidence of reduced spontaneous AMPAR-mediated events in the same mice. This is consistent with our observed increase in the AMPAR/NMDAR ratio and measurement of NMDAR EPSC decay kinetics suggesting that the changes found in the 12-month-old Fmr1 KO PFC are due to a general reduction in synaptic NMDARs.
FMRP regulation of synaptic mRNAs is highly promiscuous. Over a thousand different mRNA targets have thus far been identified, including a third of the postsynaptic proteasome and a quarter of the presynaptic proteasome (Darnell et al. 2011; Ascano et al. 2012). Important targets for FMRP regulation include potassium channel family members, which in the hippocampus have been linked to deficits in LTP in Fmr1 KO mice (Lee et al. 2011). However, our recordings of the intrinsic properties from 12-month-old layer V pyramidal neurons did not identify any changes in hyperpolarization-gated currents, as reported in CA1 pyramidal neurons (Brager et al. 2012). Neither were we able to identify any changes in basic firing properties of these cells, arguing against a large change in potassium channels (Gross et al. 2011; Lee et al. 2011; Deng et al. 2013). Instead, our findings of reduced NMDAR, but not AMPAR-mediated currents, are more consistent with a direct change in synaptic glutamate receptor expression in 12-month-old Fmr1 KO mice. Supporting this idea is the finding that in young adult Fmr1 KO mice (2 months) NMDAR expression in synaptosomes is either normal or enhanced compared with WT mice in the neocortex (Schütt et al. 2009); however, in mice at an age intermediary to those used in this study (4 months), NMDAR expression is reduced by 15% (Krueger et al. 2011). Therefore, it is likely that the reduction in LTP found in 12-month-old Fmr1 KO mice is directly linked to the reduction in NMDAR function.
Enhanced postsynaptic mGluR5-mediated LTD is a hallmark of FXS (Bear et al. 2004), and genetic reduction or treatment with mGluR5 antagonists rescues many of the behavioral deficits found in the Fmr1 KO mouse (Dölen et al. 2007; Krueger and Bear 2011; Michalon et al. 2012; Busquets-Garcia et al. 2013). Furthermore, FXS individuals are reported to have enhanced expression of mGluR5 in the PFC (Lohith et al. 2013). This prompted us to ask if treatment with an mGluR5 antagonist might rescue the age-linked deficits in LTP found in the Fmr1 KO mouse PFC. Surprisingly, we found that acute treatment with mGluR5-negative allosteric modulator MPEP for 1 h ex vivo restored TBS-LTP in 12-month-old Fmr1 KO mice to WT levels. Furthermore, analysis of the AMPAR/NMDAR ratio in these MPEP-treated neurons suggested a normalization of NMDAR function. A similar acute treatment restored LTP in the hippocampus and anterior cingulate cortex in Fmr1 KO mice, arguing for the validity of this approach (Gross et al. 2011; Xu et al. 2012; Chen et al. 2014). FXS shares molecular similarities to Alzheimer's disease (Sokol et al. 2011), notably an amyloid-β peptide-dependent enhancement of mGluR5 signaling and a reduction in NMDAR-mediated transmission (Renner et al. 2010; Malinow 2012; Hu et al. 2014). Similar to our findings in 12-month-old Fmr1 KO, acute MPEP treatments are able to normalize postsynaptic mGluR5 signaling and restore NMDAR-dependent LTP in mouse models of Alzheimer's disease (Wang et al. 2004; Rammes et al. 2011; Um et al. 2013). Precisely, how mGluR5 downregulation rapidly couples to an enhancement in NMDAR function in the Fmr1 KO PFC is unclear; however, it is worth noting that a single in vivo acute MPEP injection is able to rapidly modulate both firing properties and NMDAR function in deep layer pyramidal neurons of the PFC (Homayoun et al. 2004; Homayoun and Moghaddam 2006). Regardless, the restoration of LTP in 12-month-old Fmr1 KO mice with mGluR5 antagonism supports the idea that aberrant mGluR activity is central to many of the synaptic dysfunctions found in FXS.
It is tempting to connect our age-linked Fmr1 KO observations, directly to age-dependent changes in cognitive function found in FXS individuals; however, how synaptic plasticity reflects PFC-dependent behavior is currently unclear. Nonetheless, the finding that deficits in PFC function vary temporally in the Fmr1 KO mouse extends current treatment strategies for FXS. Specifically, tailoring treatments to subject age and/or combinations of drugs might show future promise in FXS.
H.M. and O.J.M. designed research; H.M. and O.L. performed research; H.M. and O.L. analyzed data; J.T.B. contributed analysis tools and methods; H.M. and O.J.M. wrote the paper; O.J.M. supervised the entire project.
This work was supported by Institut National de la Santé et de la Recherche Médicale (INSERM); L'Agence nationale de la Recherche (ANR) Blanc France-Taiwan RescueMemo (to O.M.); ANR CYFIP-Aut; the FRAXA Research Foundation (to O.M.); NARSAD 2010 Independent Investigator Grant (to O.M.); and Foundation Jêrome Lejeune (to H.M. and O.M.).
The authors thank the FRAXA research foundation (Dr D. Nelson, Baylor College of Medicine) for providing the Fmr1 KO2 mice. O.M. and H.M. are grateful to Dr P. Chavis, Dr Po-Wu Gean, and the Chavis and Manzoni laboratories for helpful discussion. Conflict of Interest: None declared.