Heterozygous mutation or deletion of a lissencephaly gene (Lis1) in humans is associated with a severe disruption of cortical and hippocampal lamination, cognitive deficit, and severe seizures. Mice with one null allele of Lis1 (Lis1+/– mice) exhibit significant brain malformations and slowed migration of interneuron precursors. Although hyperexcitability was demonstrated in dysplastic hippocampal slices from Lis1+/– mice, little is known about synaptic function in these animals. Here we analyzed GABA-mediated synaptic inhibition. We recorded isolated whole cell inhibitory postsynaptic currents (IPSCs) on visually identified pyramidal neurons in disorganized CA1 regions of hippocampal slices prepared from Lis1+/– mice. We observed a 32% increase in spontaneous IPSC frequency in Lis1+/– mice compared with normotopic CA1 pyramidal neurons in age-matched controls. This increase was not associated with a change in spontaneous IPSC decay or miniature IPSC frequency. Mean IPSC amplitude was increased, and event histograms indicated a greater number of large (>125 pA) events. Tonic inhibition, response to paired-pulse stimulation and evoked IPSC decay kinetics were not altered. Consistent with increased synaptic inhibition, Lis1+/– interneurons also exhibited more spontaneous firing in cell-attached recordings and increased excitation as measured by voltage-clamp recording of spontaneous excitatory postsynaptic currents (EPSCs) onto interneurons. Our results reveal a significant alteration in the function of inhibitory circuits within the malformed Lis1+/– hippocampus. Given that precisely coordinated GABAergic activity is vital to generation of oscillatory activity and place field precision in hippocampus, these alterations in synaptic inhibition may contribute to seizures and altered cognitive function in type I Lissencephaly.
Lissencephaly is a brain malformation disorder characterized by loss of normal gyri and a disorganized cerebral cortex. In humans, malformations are associated with mental retardation, developmental delay, intractable forms of epilepsy, and other serious neurological symptoms (Dobyns et al. 1993; Guerrini and Filippi 2005; McManus and Golden 2005; Walsh 1999). Classical, or type I, Lissencephaly is linked to mutations in one allele of the Lis1 gene. Lis1 protein is important in the microtubule-based motor activity of cytoplasmic dynein (Dobyns and Truwit 1995; Reiner et al. 1993; Vallee and Tsai 2006). Because microtubule motility is vital to the proper migration of neuronal precursors (Tsai and Gleeson 2005), an insufficient amount of Lis1 protein appears to result in slowed migration and subsequent brain malformation (McManus et al. 2004; Wynshaw-Boris and Gambello 2001). Although a role for Lis1 protein in neurodevelopment is well established, how neurons communicate in a disorganized, Lis1-deficient brain is not understood.
Severe anatomical alterations, including cellular disorganization and multiple principal cell layers in hippocampus, were previously reported in mice with one null allele of Lis1 (Lis1+/– mice) (Fleck et al. 2000; Hirotsune et al. 1998). In other rodent models for which disruption of normal brain architecture has been described (e.g., postnatal freeze lesion, in utero exposure to radiation or methylazoxymethanol), studies have noted significant functional disruption of synaptic communication within these disorganized neuronal networks. These phenotypes include, but are not limited to, enhanced N-methyl-d-aspartate (NMDA) receptor function (Calcagnotto and Baraban 2005; DeFazio and Hablitz 2000), decreased excitatory drive of displaced interneurons (Xiang et al. 2006), increased miniature excitatory postsynaptic current frequency and bursting rate (Zsombok and Jacobs 2007), and reduced IPSC frequency (Calcagnotto et al. 2002; Trotter et al. 2006). Human brain tissue samples obtained from patients with cortical dysplasia exhibit similar functional deficits (Andre et al. 2004; Calcagnotto et al. 2005; Najm et al. 2004). Disruption of cognitive function and generation of abnormal seizure discharge—common clinical manifestations of a malformed brain (Guerrini et al. 2003)—are likely to be linked with significant alterations in synaptic transmission.
Consistent with the severe disruption in hippocampal organization seen in Lis1+/– mice, acute brain slices from these animals exhibit a reduced threshold for potassium-induced epileptiform bursting, and a modest disruption of excitatory field potentials at the Schaffer collateral-CA1 synapse (Fleck et al. 2000). Cortical and hippocampal interneurons, which release the inhibitory neurotransmitter GABA, were recently shown to exhibit slowed nonradial migration in embryonic slice cultures from Lis1+/– mice, although inhibitory neurotransmission was not analyzed (McManus et al. 2004). Because GABA-mediated synaptic transmission plays a critical role in regulating a variety of CNS functions (Farrant and Nusser 2005; Mohler 2006), we focused these studies on inhibitory circuits in the malformed Lis1+/– mouse hippocampus. Our approach used patch-clamp techniques, infrared-differential interference contrast (IR-DIC) microscopy and pharmacological manipulations. Here we report an increase in spontaneous IPSC frequency onto disorganized CA1 pyramidal cells and a corresponding increase in spontaneous firing and excitatory drive of interneurons; we noted no changes in tonic or evoked inhibitory currents.
We prepared 300-μm acute brain slices from P14 to P22 mice for electrophysiological recordings. Mice were anesthetized and decapitated, and the brain was rapidly removed and placed into ice-cold, oxygenated artificial cerebrospinal fluid (ACSF) containing high sucrose (in mM, 150 sucrose, 50 NaCl, 25 NaHCO3, 10 dextrose, 2.5 KCl, 1 NaH2PO4-H2O, 0.5 CaCl2, and 7 MgCl2). The brain was blocked and glued to the stage of a Vibratome (Leica VTS1000, Bannockburn, IL), and horizontal slices containing the hippocampus were prepared in oxygenated high-sucrose ACSF at 4°C. Slices were transferred to a holding chamber containing normal ACSF (in mM, 124 NaCl, 3 KCl, 1.25 NaH2PO4-H2O, 2 MgSO4-7H2O, 26 NaHCO3, 10 dextrose, and 2 CaCl2). Slices were incubated for 40 min in a 35°C water bath before being maintained at room temperature for the remainder of the day. For recording, each slice was transferred to a recording chamber where it was submerged in oxygenated, normal ACSF flowing at 2–4 ml/min.
We obtained whole cell voltage- and current-clamp recordings from visually identified pyramidal cells and interneurons using infrared differential interference contrast (IR-DIC) optics. Patch pipettes (3–7 MΩ) were pulled using a micropipette puller (Sutter Instrument, Novato, CA) from 1.5 mm OD borosilicate glass (World Precision Instruments, Sarasota, FL). All data were acquired using an Axopatch 1D amplifier, digitized at 10 kHz with a Digidata 1320A, and recorded with pClamp 8.2 software (Axon Instruments/Molecular Devices, Sunnyvale, CA). In some experiments, 0.1% biocytin was added to the internal pipette solution for subsequent morphological confirmation of the recorded cell type, using standard procedures for processing sections with diaminobenzidine tetrahydrochloride (Xiang et al. 2006).
Voltage-clamp recordings of spontaneous and miniature inhibitory postsynaptic currents (sIPSCs and mIPSCs) were obtained at room temperature, in accordance with standard recording protocols (Bessaih et al. 2006; Calcagnotto et al. 2002; Shao and Dudek 2005; Trotter et al. 2006; Zhu and Roper 2000). Internal patch pipette solution (pH 7.20–7.25, 285–295 mosM) contained, in mM, 140 CsCl, 1 MgCl2, 10 HEPES, 11 EGTA, 2 Na2ATP, 0.5 Na2GTP, and 1.25 QX-314. We recorded sIPSCs from CA1 pyramidal neurons held at −60 mV in the presence of 3 mM kynurenic acid (Tocris, Ellisville, MO) to block glutamatergic transmission; 100 μM gabazine (SR-95531; Sigma) was added at the end of some experiments (n = 25) to confirm that all synaptic events recorded in the presence of kynurenic acid were GABAergic. To isolate mIPSCs from sIPSCs, we added 1 μM TTX (Alomone Labs, Jerusalem, Israel) and tested each batch of TTX to confirm its efficacy at blocking action potentials. For tonic inhibition experiments, we followed the protocol described by Maguire et al. (2005). Briefly, we obtained stable sIPSC recordings as previously described, with 0.005 mM GABA included in the ACSF. After a stable baseline in 3 mM kynurenic acid was established (>5 min), we perfused the slice with 100 μM gabazine and continued recording until a stable baseline was re-established and all synaptic events were eliminated.
For evoked IPSC (eIPSC) studies, we obtained voltage-clamp recordings of sIPSCs as previously described. A monopolar stimulating electrode was placed in stratum radiatum, 50–100 μm from the pyramidal cell being recorded. Stimulus parameters were controlled using pClamp software, an A310 Accupulser (World Precision Instruments), and an A360 Stimulus Isolator (World Precision Instruments). Pulses (100 μs) were delivered at low frequency (0.05 Hz). We obtained a plot of stimulus intensity versus response size for each cell and performed paired-pulse experiments at a stimulus strength that reliably elicited a response of ∼100 pA for the first pulse. Interpulse intervals tested in these experiments were 40, 100, 400, 500, and 700 ms. At the end of some experiments, 5 μM bicuculline (Sigma) was added to ensure that eIPSCs were mediated by GABAA receptors (n = 3).
For cell-attached recordings of spontaneous firing in interneurons, we used a slightly modified ACSF containing 5 mM KCl and obtained recordings at 32°C. We recorded spontaneous firing in voltage-clamp mode at −60 mV before switching to current-clamp mode and rupturing the patch to obtain a whole cell current-clamp recording so that the interneuron sub-type could be identified. Internal pipette solution (pH 7.20–7.25, 285–295 mosM) for spontaneous firing and current-clamp recordings contained, in mM, 120 K-gluconate, 10 KCl, 1 MgCl2, 0.025 CaCl2, 0.2 EGTA, 2 Na2ATP, 0.2 Na2GTP, and 10 HEPES. We injected 1,000-ms depolarizing and hyperpolarizing current steps to measure membrane and firing properties of each interneuron recorded; passive membrane properties were examined occasionally throughout the experiment to ensure the health of the cell and the integrity of the seal.
Recordings of spontaneous excitatory postsynaptic currents (sEPSCs) on interneurons were obtained at 32°C in the presence of 5 μM bicuculline, in accordance with standard recording protocols (Alkondon and Albuquerque 2002; Behr et al. 2000; Epsztein et al. 2006; McBain and Dingledine 1992; Perez et al. 2006). We used the previously described current-clamp internal patch pipette solution and recorded events at a holding potential of −60 mV. Immediately after obtaining a whole cell recording, we switched to current-clamp mode and injected current steps to identify the interneuron subtype (cells with firing properties reminiscent of CA1 pyramidal neurons were discarded) and then switched back into voltage-clamp mode to record sEPSCs. All voltage-clamp recordings were low-pass filtered at 1 kHz and band-pass filtered at 60 Hz (Hum Bug, AutoMate Scientific, Berkeley, CA). At the end of some experiments, 3 mM kynurenic acid was added to confirm that sEPSCs were successfully isolated (n = 3). Whole cell access resistance and holding currents were monitored throughout all recordings; cells for which these values changed by >25% were excluded from analysis.
sIPSCs, mIPSCs, and sEPSCs were analyzed using Mini Analysis 5.2.5 software (Synaptosoft, Decatur, GA). Each event was manually selected with the investigator blind to the genotype of the animal. Analysis commenced after allowing sufficient time for stabilization of the recording and for complete wash-in of pharmacological agents (>4 min). Decay times were measured as the time from the peak to the time at which the event amplitude was of peak value, and these values were averaged for each cell. Mini Analysis software and Microsoft Excel were used for construction of amplitude, decay, and inter-event interval histograms. Histograms were generated from data across all cells, with an equal number of events used from each cell. Amplitudes of tonic GABAergic currents were measured by hand using Clampfit 8.2 software (Axon Instruments); data are expressed as percent change in holding current after gabazine application. Amplitudes and decay time constants of evoked IPSCs were measured in Clampfit 8.2 software. Decay time constants were determined by fitting a curve to the decay phase of an averaged eIPSC. For presentation of paired-pulse modulation responses in ⇓⇓⇓⇓Fig. 5, we show the profile of response types only for a standard interpulse interval of 100 ms (Kim and Alger 2001; Kogo et al. 2004; Storozhuk et al. 2002; Xiao et al. 2006), though all interpulse intervals were analyzed and found to be similar. Spontaneous firing rates from cell-attached recordings were measured using Mini Analysis software. All parameters from current-clamp recordings of interneurons (input resistance, afterhyperpolarization amplitude, firing frequency, firing frequency adaptation, spike width) were measured as described in Butt et al. (2005). All data are presented as means ± SE. Unless otherwise noted, two-tailed, unpaired Student's t-test were used to determine statistical significance of experimental results; in some cases, χ2 tests were used. Because sIPSC amplitudes did not appear to conform to a normal distribution, we used the Kolmogorov-Smirnov test to calculate significance of these results, and we also used this test to compare inter-event interval distributions. Results were deemed statistically significant if the calculated P value was <0.05.
Spontaneous inhibitory transmission in the Lis1+/− hippocampus
To analyze synaptic inhibition, we obtained voltage-clamp recordings of spontaneous IPSCs from 56 CA1 pyramidal cells in hippocampal slices prepared from wild-type (WT) and age-matched Lis1 mutant mice (WT, n = 28 cells; Lis1+/–, n = 28 cells). sIPSCs were recorded at a holding potential of −60 mV with 3 mM kynurenic acid in the ACSF to block glutamatergic transmission. Under IR-DIC, we recorded from Lis1+/– pyramidal cells located within the center of a diffuse, but recognizable, “principal” cell layer as well as cells located within clearly heterotopic locations. In all experiments, we did not observe differences between these populations of Lis1+/– pyramidal cells and have pooled them for analysis. Here we report a modest, but significant, increase in GABA-mediated synaptic transmission in Lis1+/– mice compared with controls. In WT hippocampal slices, mean sIPSC frequency was 3.89 ± 0.44 Hz, whereas in Lis1+/– mutants, sIPSC frequency was increased by 32% to 5.14 ± 0.39 Hz (P < 0.05; Fig. 1, A–C). Correspondingly, interevent intervals were decreased (P < 0.05, Kolmogorov-Smirnov test; Fig. 1E). Mean decay time was unchanged (WT: 4.87 ± 0.20 ms; Lis1+/–: 4.67 ± 0.15 ms; Fig. 1D), and the overall distribution of decay times was similar (Fig. 1F).
The mean amplitude of sIPSCs was also higher in Lis1+/– mice (43.11 ± 2.51 pA) compared with controls (37.01 ± 1.47 pA; P < 0.05, Kolmogorov-Smirnov test; Fig. 2A). An amplitude histogram of sIPSCs demonstrated that this increase could not be attributed to a rightward shift in the overall distribution of event amplitudes but instead reflected a threefold increase in the frequency of large (>125 pA) events in Lis1+/– mice (Fig. 2B, *). Only 4% of recorded wild-type sIPSCs were >125 pA, whereas 12% of Lis1+/– sIPSCs were >125 pA (Fig. 2C).
Because large events could be due to spontaneous firing of interneurons, we next recorded mIPSCs from 33 pyramidal cells (WT, n = 16 cells; Lis1+/–, n = 17 cells) in the presence of 1 μM TTX and 3 mM kynurenic acid. With action potentials blocked, these large events were eliminated (Fig. 2C). Analysis of mIPSC kinetics revealed no differences between WT and Lis1+/– mice (Table 1; Fig. 3, A and B). mIPSC frequency was unchanged (WT: 3.40 ± 0.46 Hz; Lis1+/–: 3.10 ± 0.37 Hz; Fig. 3C); in agreement with these data, an interevent interval plot also showed no differences between WT and mutant animals (Fig. 3F). Similarly, mean mIPSC decay time was unchanged (WT: 4.36 ± 0.15 ms; Lis1+/–: 4.70 ± 0.17 ms; Fig. 3D), and the distributions of decay times were similar between WT and mutant animals (Fig. 3G). With large events eliminated in the presence of TTX, mean mIPSC amplitude was also unchanged (WT: 28.64 ± 1.68 pA; Lis1+/–: 32.16 ± 1.73 pA; Fig. 3E) as was the distribution of individual mIPSC amplitudes across all cells (Fig. 3H).
To further assess inhibitory synaptic function in Lis1 mutant mice, we also recorded tonic inhibitory currents onto disorganized CA1 pyramidal cells. Tonic currents are mediated by extrasynaptic GABAA receptors that are bound by ambient levels of GABA, providing an important source of shunting inhibition to pyramidal neurons (Farrant and Nusser 2005). We hypothesized that due to an increase in overall GABA release, ambient GABA levels might be higher in mutant mice, leading to an increase in tonic inhibitory currents as compared with WT littermates. To test this hypothesis, we recorded from 25 pyramidal cells (WT, n = 10; Lis1+/–, n = 15) at −60 mV in the presence of 3 mM kynurenic acid and at the end of each recording, bath applied the competitive GABAA receptor antagonist gabazine (100 μM). We measured percent change in holding current after gabazine application, and found no difference in mean tonic current size between WT and Lis1+/– mice (WT: 10.36 ± 1.63%; Lis1+/–: 9.92 ± 1.01%; Fig. 4).
Evoked synaptic transmission and paired pulse modulation in the Lis1+/– hippocampus
To further characterize inhibitory synaptic transmission in Lis1+/– mice, we evoked IPSCs onto 19 CA1 pyramidal cells (WT, n = 8; Lis1+/–, n = 11) using a monopolar electrode placed in s. radiatum. To isolate eIPSCs, we included 3 mM kynurenic acid in the ACSF and recorded responses at a holding potential of −60 mV. We measured peak amplitude of each eIPSC and calculated a paired-pulse ratio (PPR) for each cell, examining paired-pulse modulation at interpulse intervals of 40, 100, 400, 500, and 700 ms (Fig. 5, A and B). For all cell types, we observed examples of paired-pulse facilitation (PPF; defined as a PPR >1) and paired-pulse depression (PPD; PPR <1), with some cells showing an absence of any paired-pulse modulation (PPR = 1). The distributions of cells with PPR greater than, less than, or equal to 1 were similar between genotypes at all interpulse intervals tested; for clarity, shown in Fig. 5 is the distribution at a standard interpulse interval of 100 ms (WT, PPR >1: 37.5%; WT, PPR <1: 37.5%; WT, PPR = 1: 25.0%; Lis1+/–, PPR >1: 36.4%; Lis1+/–, PPR <1: 45.5%; Lis1+/–, PPR = 1: 18.2%; P > 0.05, χ2 test; Fig. 5C). A plot of the amplitude of the second response as a percentage of the first response [(eIPSC2/eIPSC1) × 100] reveals that the profile of paired-pulse modulation at different intervals was also similar between WT and mutant animals (for all interpulse intervals, P > 0.05, Student's t-test; Fig. 5B). We also examined elPSC decay time constants to determine whether abnormalities in GABA receptor subunit composition or GABA reuptake were functionally evident in Lis1+/– mice. We found that values were comparable between WT and mutant mice (WT: 40.79 ± 5.27 ms; Lis1+/–: 39.76 ± 4.68 ms; Fig. 5D).
Spontaneous firing and excitatory drive onto interneurons in the Lis1+/− hippocampus
Because of the observed increase in large, action-potential-dependent spontaneous IPSCs in Lis1+/– mice, we tested the hypothesis that interneurons in the Lis1+/– hippocampus exhibit increased spontaneous firing by recording from 37 CA1 interneurons (WT, n = 20; Lis1+/–, n = 17) in the cell-attached configuration (Cossart et al. 2001) (Fig. 6). Under IR-DIC, we recorded from fast-spiking (FS), regular-spiking nonpyramidal (RSNP), and burst-spiking (BST) interneurons in s. oriens, s. radiatum, and s. lacunosum-moleculare locations. The mean spontaneous firing rate of interneurons in WT mice was 0.16 ± 0.12 Hz with only 20% of recorded WT interneurons firing at rest. Nonfiring cells were assigned a rate of 0 Hz and included in the overall calculation of mean firing rate. These data confirm that spontaneous firing of interneurons is rare in slices from control animals (Cossart et al. 2001). However, firing activity for interneurons recorded in Lis1+/– slices was more common with 71% of cells exhibiting spontaneous activity (P < 0.001, χ2 test; Fig. 6C). We also assigned a firing rate of 0 Hz to nonfiring Lis1+/– interneurons and found that the mean firing rate of Lis1+/– interneurons was significantly increased, at 1.21 ± 0.46 Hz (P < 0.05; Fig. 6D). We also compared the firing rates of only those interneurons that exhibited spontaneous firing, excluding all interneurons with a rate of 0 Hz. By this analysis, we found the mean firing rate for WT interneurons to be 0.82 ± 0.50 Hz (n = 4), whereas the rate for Lis1+/– interneurons was 1.71 ± 0.60 Hz (n = 12). At the conclusion of cell-attached recording, we ruptured the patch and switched into current-clamp mode to confirm interneuron identity (Freund and Buzsaki 1996; Flames and Marin 2005) (Fig. 6, A and B, bottom).
To explore the possibility that alterations in intrinsic firing properties of interneurons could underlie their increased activity, we obtained current-clamp recordings from 111 visually identified CA1 interneurons in slices from WT (n = 62 cells) and Lis1+/– (n = 49 cells) mice. We injected 1,000-ms current steps to measure resting membrane potential, input resistance, spike width, amplitude of the fast afterhyperpolarization, firing frequency at twice the firing threshold, and firing frequency accommodation (Table 2). In these recordings, we encountered FS, RSNP, and BST interneurons. We found that intrinsic properties were similar for all subclasses of interneurons from WT and Lis1+/– mice, indicating that a change in synaptic inputs to interneurons is likely to underlie the increased firing activity we observed in cell-attached recordings of Lis1+/– interneurons.
Finally, we hypothesized that an increase in excitatory drive onto GABAergic interneurons could underlie enhanced activity of inhibitory systems in the Lis1+/– hippocampus. We recorded isolated spontaneous EPSCs onto 41 CA1 interneurons (WT n = 23, Lis1+/– n = 18) in the presence of a GABAA receptor antagonist, 5 μM bicuculline. The mean sEPSC frequency in mutant mice was significantly increased compared with cells from age-matched WT littermates (WT: 4.19 ± 1.19 Hz; Lis1+/–: 8.56 ± 1.82 Hz; P < 0.05; Fig. 7, A–C). Mean sEPSC amplitude and decay time values were unaltered (Fig. 7, D and E). Interneuron identification was confirmed in current-clamp mode for all cells (Fig. 7, A and B, bottom).
Lis1 haploinsufficiency results in type-I Lissencephaly in humans (Dobyns et al. 1993) and similar neuronal migration disorders in mice (Gambello et al. 2003; Hirotsune et al. 1998). In humans, Lis1 deficit contributes to profound mental retardation and intractable epilepsy (Dobyns et al. 1993); in mice, Lis1 deficit is associated with impaired spatial learning and hyperexcitability of hippocampal circuits (Fleck et al. 2000; Paylor et al. 1999). To study mechanisms that may contribute to these neurological phenotypes, we analyzed inhibitory systems in Lis1+/– mice. We anticipated that GABAergic transmission would be reduced, consistent with the circuit hyperexcitability previously observed in these mice (Fleck et al. 2000) and also consistent with other studies indicating that inhibitory transmission is decreased in cortical malformation disorders (Calcagnotto et al. 2002; Trotter et al. 2006). Surprisingly, we identified a modest increase in GABA-mediated synaptic inhibition, in that the frequency of large, action-potential-dependent sIPSCs onto disorganized CA1 pyramidal neurons was enhanced compared with age-matched controls. The inhibitory current increase was associated with increased spontaneous firing rates of interneurons, as well as enhanced glutamatergic excitation of these cells. Our findings, therefore, provide evidence that Lis1 deficit is associated with a significant disruption of inhibitory circuitry in the malformed hippocampus.
In addition to known effects of Lis1 deficit on neuronal migration (McManus et al. 2004), a reduction of Lis1 protein in neurons could also contribute to functional defects via a direct effect on synaptic vesicle trafficking. Research in C. elegans indicates that Lis1 mutations may impair the proper trafficking of GABA-containing synaptic vesicles (Williams et al. 2004). Using nematodes expressing a nonsense allele of lis-1 (i.e., pnm-1) Williams et al. noted an uneven distribution of GABA vesicles, some of which were irregularly sized, in pnm-1 mutants. Similarly, in a Drosophila Lis1 mutant, large axonal swellings, possibly containing an abundance of synaptic vesicle-associated proteins, were found to “clog” axonal transport (Liu et al. 2000). In a more complex mammalian system, however, we find that Lis1 deficiency does not alter the frequency, amplitude, or decay times of mIPSCs. Similarly, decay kinetics and paired-pulse modulation of evoked IPSCs also appear unaltered in Lis1 mutant mice as compared with age-matched controls, suggesting that pre- and postsynaptic GABAergic systems are grossly normal. It remains possible, however, that our IPSC recordings at room temperature may not have uncovered minor differences in basal inhibitory transmission or subtle alterations in the phosphorylation state of postsynaptic receptors that might only be apparent at more physiological temperatures.
Our recordings of spontaneous firing of hippocampal interneurons support the hypothesis that the increase in large inhibitory events (i.e., sIPSCs >125 pA in amplitude) in Lis1+/– mice is associated with increased firing of hippocampal interneurons in this disorganized circuit. Similar results were reported in other rodent studies featuring hyperexcitable circuits. For example, in kainate- and pilocarpine-treated rats, commonly used animal models of temporal lobe epilepsy, Cossart et al. (2001) observed an increase in cell-attached spontaneous firing rates of interneurons. In the case of Lis1+/– mice, the increased firing of interneurons is most likely attributable to enhanced excitatory inputs arriving from pyramidal cells, as we noted an increase in sEPSC frequency onto these interneurons. Similarly, in kainate- and pilocarpine-treated rats, the increased spontaneous firing rate of interneurons was associated with an increase in excitatory inputs to interneurons (Cossart et al. 2001). Also consistent with our findings, an increase in sEPSC frequency onto CA1 interneurons was observed in slices bathed in elevated potassium, another commonly studied manipulation eliciting epileptic discharge (McBain 1994).
An increase in GABA-mediated inhibition, as reported here for Lis1+/– mice, was also recently reported in a mouse model of autosomal dominant nocturnal frontal lobe epilepsy (Klaassen et al. 2006). Increased inhibition appears counter-intuitive in an epileptogenic process but may also contribute to the hyperexcitable phenotype observed. For example, interneurons are known to be critical for the generation and maintenance of oscillatory activity in the CNS (Bartos et al. 2007). Hippocampal parvalbumin-positive basket-type interneurons, in particular, are vital to the generation of gamma oscillations in entorhinal cortex (Cunningham et al. 2006), and GABA-mediated inhibition drives β/γ oscillatory activity in hippocampus (Trevino et al. 2007). GABA-mediated oscillations are a physiologically relevant mechanism that modulates the firing activity of large populations of principal glutamatergic neurons. Recent analysis of human electroencephalographic data suggests that high-frequency oscillations, likely controlled by the firing of inhibitory interneurons, may predict the onset of seizure activity (Medvedev et al. 2000; Worrell et al. 2005). Therefore a disruption of the inhibitory system could lead to abnormal synchronization of excitatory circuits and, perhaps, generation of abnormal electrical discharges, i.e., seizures.
Besides temporally linking neuronal firing activity, GABA-mediated oscillations play a critical role in memory consolidation (Bragin et al. 1999; Buzsaki 1989; Grenier et al. 2001). For example, interneuron-generated theta rhythms are thought to gate memory information processing in the hippocampus (Sun et al. 2001). At a cellular level, the neural correlate of learning (spatial learning, in particular) is thought to be encoded by the firing of hippocampal pyramidal (place) cells. Place cell firing patterns are fairly precise, and interneuron-mediated feedback inhibition is hypothesized to maintain this precision (Bose and Recce 2001; Frank et al. 2006; Touretzky et al. 2005). In Lis1 mutant mice, the enhanced activity of at least some interneurons, and the occasional large-amplitude inhibitory events that result, could disrupt the finely tuned patterning of place cell firing and result in a place field map with limited precision. The known relationship between place cell firing patterns and spatial memory (Hollup et al. 2001; Lenck-Santini et al. 2001) is consistent with the impaired performance of Lis1+/– mice in spatial memory tasks. For example, Lis1+/– mice failed to search in the correct quadrant of a pool during a conventional test of spatial learning performance (Paylor et al. 1999). Most importantly, the observed alteration of GABA-mediated synaptic transmission in hippocampal slices from Lis1 mutant mice could underlie some of the cognitive deficits associated with type-1 Lissencephaly.
In conclusion, we provide the first systematic assessment of inhibitory synaptic circuitry in the malformed hippocampus of a Lis1-deficient mammal. We find functional disruption of GABA-mediated synaptic transmission and increased firing rates and excitatory inputs onto interneurons in this circuit. The cumulative result of these inhibitory changes could be enhancement of the epileptogenic process and/or alteration of cognitive function, two of the primary neurological deficits observed in patients with Type-I Lissencephaly.
This work was supported by National Institutes of Neurological Disorders and Stroke Grant R01 NS-40272-05.
We thank E. Harrington and M. Dinday for mouse genotyping, A. Wynshaw-Boris for generous donation of lissencephaly mutant mice, and Y. Molina for providing inspiration during the gathering of experimental data. We also thank J. Greenwood for assisting in the analysis and interpretation of electrophysiological data and J. Greenwood and D. Jones-Davis for comments on previous versions of this manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by the American Physiological Society