Properties of the epileptiform activity in the cingulate cortex of a mouse model of LIS1 dysfunction

Experiments were done in brain slices prepared from male Lis1/sLis1 mice (Cahana et al. 2001). Mice (14–16 postnatal days) were maintained, managed, and sacrificed following the international laws and policies (Spanish Directive “RealDecreto1201/2005”; European Union Directive 2010/63/UE). All protocols were approved by the Ethical Committee for Experimental Research of the Universidad Miguel Hernández (code: 2018/VSC/PEA/0035). Mice were genotyped before performing the experiments by PCR. A ≈ 750 bp DNA fragment was amplified from mutant mice tissue sample lysate with the following primers: 3’GGTGGCAGTGTTGAGATG CCTAGCC5’ and 5´GCATTCCTGTAATCCAGTACCTGG 3´. Amplification was performed for 35 cycles with single initial denaturalization step of 94° for 5 min. Each cycle was composed by a denaturalization step of 94° for 40 s, hybridization step of 60° for 45 s and a polymerization step of 72° for 10 s.

Animals were killed by cervical dislocation, decapitated, and their brains were quickly removed and sliced (400 µm, coronal plane) with a vibratome (LeicaVT1000) submerged in ice-cold cutting solution (composition in mM: NaCl, 124; KCl, 2.5; PO4H2Na, 1.25; MgCl₂, 2.5; CaCl₂, 0.5; NaCO₃H, 26; glucose, 10; pH 7.4 when saturated with 95% O₂ and 5% CO₂). The slices were transferred to a modified artificial cerebro-spinal fluid (mACSF) with high extracellular potassium for enhanced excitability (composition in mM: NaCl, 124; KCl, 5; PO4H2Na, 1.25; MgCl₂,1; CaCl₂, 1.2; NaCO₃H, 26; glucose, 10; pH 7.4 when saturated with 95% O₂ and 5% CO₂); slices were incubated in this solution at 37 ºC during 30 min. and then remained at room temperature until recordings; recordings were obtained at least 60 min after preparing the slices. Anterior cingulate cortex (ACC) and retrosplenial cortex (RSC) slices were selected by their position along the rostro-caudal axis according to the mouse brain atlas by Paxinos and Franklin (2001). We considered ACC slices those placed between 1.10 and 0.14 mm from the bregma, and RSC slices those placed between − 1.06 and − 2.06 mm from the bregma.

For recording, one slice was placed in the recording chamber where it was perfused with mACSF at ≈3 ml/min (33–34 °C). For extracellular recordings, bicuculline (10 µM) was added to the ACSF (except in the experiments of bicuculline dose–response relationship shown in Fig. 2). Extracellular recordings were done with glass pipettes pulled of borosilicate glass (1.5 mm o.d., 0.86 mm i.d., with inner filament) and filled with mACSF (3–6 MOhm). In each slice a single electrode was successively placed in different recording positions within layer 2/3 to obtain electrophysiological recordings, as shown in the Results section. Local field potentials were recorded with a MultiClamp 700B amplifier (Axon Instruments, Molecular Devices, USA), filtered at 5 kHz and digitized at 20 kHz (Digidata 1200B or 1440A Axon Instruments, Molecular Devices, USA). Bicuculline, GYKI53655, cyclothiazide, and CNQX were obtained from Sigma–Aldrich (USA). Extracellularly recorded responses in layer 2/3 were evoked by applying electrical stimuli to layer 1 using bipolar concentric electrodes. Responses were evoked by square current pulses (0.1 ms of duration) whose amplitude was adjusted to 2 × the threshold intensity. The latency of the evoked responses was calculated from the onset of the stimulus artifact to the time at which the response was 10% of its peak amplitude. The frequency of the oscillatory post-discharges was measured from the autocorrelogram computed during an interval of 250–500 ms of regular oscillatory activity; the frequency was measured from the interval between the 1st (at 0 ms, with the highest autocorrelation) and the 2nd peaks of the autocorrelogram (see below, Fig. 3).

Single neuron electrophysiological parameters were studied using somatic whole-cell intracellular recordings from pyramidal neurons in layer 2/3 of the dorsal agranular RSC (aRSC). Pyramidal neurons were selected by the shape of the soma and the prominent apical dendrite as seen under the microscope with DIC optics. Intracellular recordings were obtained with patch pipettes made of borosilicate glass that had 3–5 MOhms when filled with the intracellular solution (composition in mM: 130 potassium gluconate, 5 KCl, 5 NaCl, 5 EGTA, 10 HEPES, 2 Mg-ATP, 0.2 Na-GTP; pH 7.2 adjusted with KOH; 285–295 mOsm). Resting membrane potential was measured after achieve whole-cell configuration. Series resistance was compensated in current-clamp mode with the bridge balance tool of the Multiclamp. Passive responses and action potential firing were obtained in current-clamp configuration by applying a protocol of consecutive current square pulses (− 150 pA to 200 pA; 1 s of duration, 25 pA of increment). Membrane input resistance was calculated from the voltage deflection caused by small (− 25 to − 50 pA) hyperpolarizing current pulses. Action potential (AP) threshold was estimated as the potential level at which there was a clear abrupt increase in the slope. AP amplitude was measured from action potential threshold. AP duration was measured at 50% from the AP peak. Action potential after-hyperpolarization (AHP) peak amplitude was measured from AP threshold. AHP duration was measured from the peak to the 50% of the peak amplitude. Spontaneous postsynaptic excitatory currents (sEPSCs) were recorded in voltage-clamp configuration, holding the membrane potential of the cell at − 68 mV (equilibrium potential for Cl–- calculated from the extra and intracellular concentrations). Spontaneous synaptic currents detection and analysis were performed using WinEDR 3.9.1 software (Dr. J. Dempster, University of Strathclyde, Glasgow UK). All drugs used in the experiments were added to the mACSF at the specified concentrations from concentrated stocks prepared in water (bicuculline, GYKI53655, and CNQX; stocks at 1000 × the final maximum concentration) or in DMSO (cyclothiazide; stock 50 mM).

Postnatal P4 mice were anesthetized and injected with a lentivirus that expresses GFP under the promotor of CaM kinase IIa (Lv-CkIIa-GFP; SignaGene; SL100307). At P14, the injected and non-injected mice were perfused with paraformaldehyde (PFA) 4% in PBS, and the brains were maintained in PFA 4% overnight at 4 ºC. The brains were embedded in agarose at 4% and cut in slices using a vibratome (100 µm of thickness). The slices were incubated in PBST with 1% of BSA and 10% of lysine and incubated with the primary antibodies overnight at 4ºC: chicken anti GFP (1:200, AVES; GFP-1020), mouse anti GluA2 (1:500, Fisher; 32-030-0), mouse anti-GluA1 (1:500, Rockland, 200-301-D61), rabbit anti-GluK5 (1:500, Abcam, ab32672), and rabbit anti GluK1 (1:500, Abcam, ab67316). The slices were washed with PBST and incubated with the secondary antibodies, anti-chicken-FITC (Sigma; F8888) and anti-mouse-Alexa 594 (Molecular Probes; A11032), anti-mouse-Alexa 488 (Molecular probes; A11019), and anti-rabbit-Alexa 488 (Molecular Probes; A21206), and DAPI (nuclear marker). The slices were mount with mowiol-NPG and the pictures were taken with the confocal microscope Leica SPEII.

The proteins were extracted from fresh brain tissues (ACC and RSC) from postnatal 16 mice after decapitation. The extraction was made with RIPA buffer (Millipore; 20–188) in azote and sonicated later. The total amount of proteins was detected with Pierce BCA protein assay kit (Thermo Scientific; 23225). After denaturalization (at 92 ºC, 10 min with β-mercaptoethanol) 20 µg of proteins of each sample were loaded on SDS–polyacrylamide gel (12%) and electrophoresed followed by wet transfer in nitrocellulose blotting membrane (Amersham; 10600007). The membranes were incubated with primary antibodies overnight at 4 ºC (with TBST and BSA 5%): rabbit GluA2 (1:4000, Fisher; AB1768IMI), mouse anti-GluA1 (1:1000, Rockland, 200-301-D61), rabbit anti-GluK5 (1:1000, Abcam, ab32672), and mouse GAPDH (1:10,000, GeneTex; GTX627408). The membranes were washed with TBST and incubated with the secondary antibodies conjugated with HRP: anti-rabbit (1:4000, Vector; BA-1000) and anti-mouse (1:4000, Vector; BA-9200). The membranes were treated with Immobilon Forte (Millipore; WBLU0100) and exposed in dark room. Images were taken with the Amersham Bioimager 680 and to analyze the ratio of gray values between samples Image J was used. The protein level was expressed in arbitrary units (A.U.) calculated as the quotient between the area of the GluA2 band over the area of the GAPDH band.

Data are given as mean ± SEM and the number of cases. Statistical analysis was made with SigmaStat3.2 (SystatSoftwareInc., USA). Comparisons were made with the Student’s t test or the Mann–Whitney Rank Sum test (when sample values did not follow a normal distribution as checked with the Kolmogorov–Smirnov test). For percentage comparison in Fig. 3C, chi-squared (χ²) test was used. Statistical significance values were depicted using the following code: (*) p value < 0.05; (**) p value < 0.01; (***) p value < 0.001.

To investigate the consequences on cortical network activity of LIS1 dysfunction caused by LisH domain deletion in heterozygosis, we analyzed the properties and the propagation along layer 2/3 of epileptiform discharges evoked by electrical stimuli in conditions of enhanced cortical excitability. To do this, electrophysiological extracellular recordings were performed in brain slices from postnatal juvenile (P14–P16) mice, either mutant (Lis1/sLis1) or wild type (WT) littermates. Recordings were obtained under conditions of enhanced excitability (bicuculline 10 µM, 5 mM extracellular K⁺, and 1.2 mM Ca⁺⁺; see “Methods”). Following our previous study on epileptiform activity propagation (Rovira and Geijo-Barrientos 2016), we recorded from two different cingulate cortical regions, the anterior cingulate cortex (ACC) and the retrosplenial cortex (RSC; Fig. 1A); these two areas were subdivided further in primary and secondary ACC (pACC and sACC) and dorsal agranular and granular RSC (aRSC and gRSC). In each slice, a stimulus electrode was placed on cortical layer 1 and recordings with an extracellular recording electrodes were performed targeting different positions along the cortical layer 2/3 (see scheme of the experimental design in Fig. 1A, right panel). Across experiments, stimulus parameters were stable, with minor adjustments of intensity (100–200 µA, 0.1 ms; see “Methods” for details on stimulus intensity adjustment). In these conditions, single-pulse electrical stimuli (100–200 µA, applied at 0.033 Hz) reliably induced stereotyped responses similar to those associated to seizure events recorded in similar conditions, which are the outcome of the coordinated firing of a local population of neurons (Rovira and Geijo-Barrientos 2016). We refer to this type of activity as epileptiform activity (EA). Figure 1B shows two representative extracellular recordings containing EA events. EA events were characterized by an initial negative peak after a fixed latency from the stimulus artifact (indicated, respectively, by an asterisk and an arrow in Fig. 1C), which was often followed by a variable number of post-discharges of smaller amplitude and shorter duration (indicated by a bracket in Fig. 1C); we refer to these events as “evoked EA events”. The post-discharges following the initial negative peak can persist for as long as one second after the stimulus. In some slices, EA events not related to electrical stimuli were also observed during the long inter-stimulus period (30 s; Fig. 1B); we refer to these events as “spontaneous EA events”.

The presence of spontaneous activity is directly related to the enhanced excitation level of the underlying neuronal network caused by our experimental conditions. Thus, we decided to quantify the frequency of these spontaneous events in the Lis1/sLis1 mouse cortex to assess whether its cortex was more excitable. The frequency of spontaneous EA events was determined as the average during prolonged continuous recordings (10 min). Interestingly, the frequency of spontaneous EA events was similar in both genotypes (Fig. 1D). To analyze the sensitivity of the cortical circuits to the generation of EA we studied the responses to increasing concentrations of bicuculline, from 0.2 to 50 µM (Fig. 2). In the pACC, the threshold for the onset of EA was about 0.5 µM in both WT and Lis1/sLis1 slices, and the maximum size of the epileptiform discharges was obtained with concentrations above 10 µM; the half-maximal effective concentration (EC₅₀) of bicuculline was 1.99 ± 0.29 µM (WT slices, n = 4) and 2.24 ± 0.62 µM (Lis1/sLis1 slices, n = 4). In the aRSC, the values of EC₅₀ were also similar in WT and Lis1/sLis1 slices (WT slices: 2.25 ± 0.46 µM, n = 2; Lis1/sLis1 slices: 2.36 ± 0.39 µM, n = 3). The EC₅₀ was calculated from the fit of a Hill equation to the data of those slices in which a complete series of concentrations of bicuculline (from 0.2 to 50 µM) were tested, and the difference between WT and Lis1/sLis1 was not statistically significant in both ACC and RSC areas. Note that, both in WT and Lis1/sLis1 slices, there were no epileptiform discharges at concentrations of bicuculline lower than 0.5 µM. This lack of differences in the dose–response relationship of the effect of bicuculline indicates that LisH domain deletion in heterozygosis has no impact on the overall cortical network excitability.

In some slices, we observed the emergence of prolonged oscillatory post-discharges following the initial negative component of the EA. These oscillatory post-discharges are illustrated in the examples shown in Figs. 1C and 3A, B, and they were apparent in both evoked and spontaneous EA events. The emergence and oscillation frequency of these post-discharges is determined by local recurrent excitatory glutamatergic connections established among neighbor pyramidal neurons (Castro-Alamancos and Rigas 2002) and, therefore, the analysis of these post-discharges could reveal the presence of abnormalities in the Lis1/sLis1 cortical circuits. The results of this analysis are shown in Fig. 3. Figure 3A, B shows representative examples of evoked oscillatory post-discharges recorded from WT and Lis1/sLis1 slices. Typically, post-discharges appeared as relatively regular oscillations (see the prominent peaks in the autocorrelogram shown in Fig. 3D) that followed the initial negative peak and that lasted for several hundreds of ms. First, we quantified the percentage of slices from either WT or Lis1/sLis1 cortex presenting these oscillatory post-discharges to evaluate the emergence of this activity. The percentage of slices from Lis1/sLis1 or WT mice showing oscillatory post-discharges was not significantly different (Fig. 3C; comparisons made with the chi-squared test). The presence of oscillatory post-discharges was particularly scarce in the gRSC. Then, we compared the frequency of these oscillatory post-discharges (Fig. 3E). In slices from WT animals, the oscillation frequency was similar across cortical regions and comparable to previously published results under similar experimental conditions (Castro-Alamancos and Rigas 2002; Rovira and Geijo-Barrientos 2016). In the ACC (pAAC: WT 16.91 ± 0.50 Hz n = 7 vs Lis1/sLis1 14.2 ± 1.17 Hz, n = 5; p = 0.038, Student’s t test. sAAC: WT 16.02 ± 0.49 Hz n = 5 vs Lis1/sLis1 11.36 ± 0.33 Hz, n = 5; p < 0.001, Student’s t test) and in the aRSC (WT 15.71 ± 0.88 Hz n = 5 vs Lis1/sLis1 12.95 ± 0.37 Hz, n = 7; p = 0.009, Student’s t test) the oscillation frequency was smaller in the Lis1/sLis1 cortex (Fig. 3E). In the gRSC, the oscillation frequency could not be measured because in this area the oscillatory post-discharges were scarce and very irregular. However, while the frequency of the oscillatory post-discharges was lower in the Lis1/sLis1 cortex, their total duration was similar to WT cortex; in the pACC the duration was: WT 1.34 ± 0.11 s (n = 6), Lis1/sLis1 1.64 ± 0.19 s (n = 6); in the sACC was: WT 1.65 ± 0.77 s (n = 6), Lis1/sLis1 1.75 ± 0.09 s (n = 5); and in the aRSC was: WT 1.19 ± 0.03 s (n = 5), Lis1/sLis1 0.83 ± 0.17 s (n = 6); All differences were not significant. This oscillatory component of the EA is caused by complex mechanisms involving synaptic mechanisms and intrinsic electrophysiological properties of the interconnected neurons (Avoli et al. 2002; De Curtis et al. 2012). This makes difficult to derive more concrete conclusions about the mechanisms underlying these findings of a lower frequency and a similar duration of the post-discharges.

EA evoked by electrical stimuli is not restricted to the vicinity of the stimulus electrode, but can propagate through the cortex; this propagation may also reach the contralateral hemisphere (relative to the stimulus electrode) when the callosal connectivity between both areas is intact (Walker et al. 2012; Rovira and Geijo-Barrientos 2016). EA propagation along layer 2/3 was studied and compared between WT and Lis1/sLis1 ACC or RSC using a procedure previously described (Rovira and Geijo-Barrientos 2016) Briefly, this procedure consisted in the extracellular recording of EA evoked by electrical stimuli applied to layer 1 with an extracellular electrode successively placed at 0.4 mm intervals from the stimulus electrode (Fig. 4A). Also, the recording electrode was placed at homotopic sites in the contralateral hemisphere (marked as 0.4´–2.0´ in Fig. 4A) to study the propagation to the contralateral hemisphere. Figures 4B, C shows examples of EA recorded in the ipsi- and contralateral hemispheres. Figures 4D, E shows the latencies of the EA recorded at the different recording sites in the ACC (Fig. 4D) and RSC (Fig. 4E) in WT and LIS1/sLis1. In the ipsilateral hemisphere, the longer latencies of the responses recorded at increasing distances from the stimulus represent the propagation of evoked responses from the stimulation; the velocity of this propagation was calculated in each slice from the linear fit of the latency values.

In the ipsilateral side of ACC and RSC the latencies were longer in Lis1/sLis1 slices (Fig. 4D, E) indicating that the velocity of propagation of EA was lower in both cortical areas of Lis1/sLis1 slices (Fig. 4F) These propagation velocities were: in the ACC WT 47.69 ± 2.16 mm/s (n = 25), Lis1/sLis1 37.34 ± 2.43 mm/s (n = 15), p = 0.004; in the RSC WT 76.83 ± 12.89 mm/s (n = 12), Lis1/sLis1 49.16 ± 3.92 mm/s (n = 23), p = 0.011. The latencies and propagation velocities found in WT slices were comparable to those previously described in a different mouse strain (Rovira and Geijo-Barrientos 2016); in WT ACC the velocity of propagation was slower than in WT RSC (p = 0.043, Mann–Whitney rank sum test). In Lis1/sLis1 cortex the velocity of propagation was also slower in the ACC than in the RSC (37.34 ± 2.43 mm/ms, n = 15, vs 48.33 ± 3.12 mm/ms, n = 21, respectively; p = 0.009, Mann–Whitney rank sum test); however, the difference between ACC and RSC was larger in WT tan in Lis1/sLis1 cortex (161% vs 129% change from ACC to RSC, respectively); this fact is probably a consequence of the different organization of the neuronal local circuits of layer 2/3 of the ACC and RSC, and suggests an heterogeneous impact of LIS1 dysfunction along cingulate cortical areas.

In the contralateral hemisphere, the latencies were longer than in the ipsilateral hemisphere; the contralateral latencies, however, represent probably the building-up of the EA in the contralateral layer 2/3 and not the interhemispheric conduction time of the callosal axons, which is shorter (about 8 ms in the aRSC, Sempere-Ferrandez et al. 2018). In the ACC there were no differences between WT and Lis1/sLis1, and in the RSC the contralateral latencies were longer in Lis1/sLis1 slices (Fig. 4D, E). The interhemispheric time interval was measured as the latency different between homotopic recording sites (Fig. 4G); This time interval was longer in Lis1/sLis1 slices, but only in the RSC, indicating again the different consequences of LIS1 dysfunction in both cortical areas.

The above data about the propagation of EA clearly indicate the presence of alterations in the structure and/or the function of the cortical circuits of layer 2/3 in Lis1/sLis1 cortex. These data show also that the consequences of LIS1 mutation differ depending on the cortical area. Several mechanisms may be implicated in these abnormalities, such as changes in the intrinsic properties of cortical neurons or changes in the structure of the neuronal circuits; The propagation of EA in the neocortex is mainly supported by excitatory synapses (Pinto et al. 2005) hence alterations in glutamatergic transmission may also be implicated.

To test whether the above results could be caused by alterations of neuronal electrophysiological properties we performed whole-cell recordings of layer 2/3 pyramidal neurons. Intracellular recordings were performed in gRSC, given that it shows a more acute phenotype than rostral ACC cortex. In layer 2/3 of this cortical area there are both regular spiking and late spiking pyramidal neurons (Kurotani et al. 2013; Robles et al. 2020); we only recorded and studied regular spiking pyramidal neurons because the generation of epileptiform activity in this cortical area is restricted to this type of pyramidal neurons (Robles et al. 2020). Figure 5A shows representative recordings of regular spiking pyramidal neurons from WT and Lis1/sLis1 gRSC. The electrophysiological properties were similar across genotypes (Fig. 5B); note that the shorter action potential duration observed in regular spiking pyramidal neurons from Lis1/sLis1 parietal cortex at postnatal day 30 (Valdés-Sánchez et al. 2007) was not apparent in our data; this could be to the earlier postnatal development stage of our experiments (P14-16 respect to P30 in Valdés-Sánchez et al. 2007) or to the presence of differences among different cortical areas (RSC in our experiments vs parietal cortex in Valdés-Sánchez et al. 2007). Overall, the firing rate in response to step-current injections, the resting membrane potential, and the threshold fort spike firing remained unaltered indicating that LIS1 dysfunction did not affect the intrinsic excitability of pyramidal neurons (Fig. 5B).

In these recording conditions, spontaneous excitatory synaptic currents (sEPSC) were easily detected (Fig. 5C). The quantification of sEPSC may give information about the excitatory input received by pyramidal neurons. In pyramidal neurons from Lis1/sLis1 animals the frequency of sEPSCs was lower, and their peak amplitude was larger than in WT neurons (Fig. 5C, D). The sEPSCs were detected and measured in periods of 30 s selected just before the firing of a large burst of synaptic activity caused by an epileptiform discharge. The different sEPSC frequency may be caused by a prolonged firing of some neurons after the epileptiform discharges; in the neurons recorded intracellularly we did not detect this type of tonic firing, but we cannot exclude the presence of a few neurons that maintain a continuous firing in the inter EA periods in the Lis1/sLis1 cortex; in the WT cortex the frequency of sEPSCs was lower, but again, we cannot determine whether this difference was due to lower number of firing neurons, to a lower firing frequency or both. On the other side, the different amplitude of the sEPSCs could be related to changes in the postsynaptic receptors (see below the data from experiments using AMPA receptor blockers) or to the transmitter release mechanisms.

The experiments of sEPSC recording were based on continuous prolonged recording of the membrane current at a holding potential of − 70 mV. In these recordings, it was apparent the presence of large barrages of excitatory currents caused by epileptiform activity (Fig. 5E); the frequency of these bursts of excitatory currents was similar in WT and Lis1/sLis1 animals (0.018 ± 0.0057, n = 8 and 0.022 ± 0.007, n = 13, respectively), which is consistent with our finding of similar frequency of spontaneous epileptiform activity events recorded with extracellular electrodes (Fig. 1D).

Ipsilateral and contralateral propagation of EA depends on excitatory glutamatergic transmission between pyramidal neurons highly reliant on AMPA type receptors (Alefeld et al. 1998; Pinto et al. 2005). This suggest that, among other possible mechanisms, the slower propagation velocity of EA observed in Lis1/sLis1 slices could be caused by abnormalities in AMPA type glutamate receptors. To check for alterations of AMPA receptors in Lis1/sLis1 animals we studied the effect on EA propagation of several drugs acting on AMPA receptors: a selective AMPA receptor blocker (GYKI53655), an AMPA/kainate receptor blocker (CNQX), and a positive modulator of AMPA receptors (cyclothiazide). To do these experiments we recorded extracellularly in the ipsilateral hemisphere of the pAAC and the aRSC, at approximately 0.8 mm from the stimulus electrode. In the pACC and the aRSC from WT slices, bath application of GYKI53655 resulted in a reversible, dose-dependent increase in the latency of epileptiform discharges (Fig. 6). In the pACC this latency increment was significantly larger in Lis1/sLis1 slices than in WT slices, but in the aRSC the latency increment was similar in both genotypes (Fig. 6C). On the other hand, the application of CNQX also induced an increase in latency in both pACC and aRSC, but in both cortical areas the latency increase was larger in the Lis1/sLis1 cortex than in the WT cortex (Fig. 6D). In fact, in the aRSC area the increase of latency induced by CNQX was clearly larger and apparent at lower concentrations than the increase induced in pACC. Since the specific blocking action of GYKI53655 on AMPA receptors has been used to disclose effects caused by kainate receptors (Paternain et al. 1995), these results point to the possibility that abnormalities in different types of receptors could be implicated in the differences observed in the propagation of EA in both cortical areas of Lis1/sLis1 animals (see “Discussion”). We next tested the effect of cyclothiazide (CTZ, applied at 100 µM), a positive modulator of AMPA receptors; neither in the pACC nor in the aRSC CTZ produced a change in latency, either in WT slices or in Lis1/sLis1 slices (Fig. 7).

There are multiple possible abnormalities in the AMPA/kainate receptors of Lis1/sLis1 cortex, and we checked the possibility of an altered number of AMPA receptors or an altered subunit composition of AMPA receptors by immunohistochemistry or western blot using an antibody against the GluA1 and GluA2 subunits of AMPA receptors and against the GluK1 and GluK5 subunits of kainate receptors. We explored the GluA2 subunit, given the critical role of this subunit in determining the calcium permeability and the mature properties of the AMPA receptors (Yuan and Bellone 2013). Both immunohistochemistry (to detect the distribution of these subunits) and western blot experiments (to detect the protein level) with a primary antibody against these subunits showed no significant differences between WT and Lis1/sLis1 in pACC and aRSC cortical areas (Fig. 8). These results discard a change in the level of receptors containing this subunit as a possible cause of the electrophysiological alterations observed in the Lis1/sLis1 cortex, but they leave a large number of possibilities open, among them the implication of other subunits or functional alterations in glutamate receptors.