Lgi1 null mutant mice exhibit myoclonic seizures and CA1 neuronal hyperexcitability

Generation of Lgi1-targeted embryonic stem cells

The structural gene of the mouse ortholog of human LGI1 contains eight exons, which spans 44.16 kb (Fig. 1). To reduce the probability of generating a truncated mutant Lgi1 protein after gene targeting, we decided to delete the genomic region extending from exon 3 to exon 8 using Cre/loxP-mediated chromosome engineering (24). The loxP sites were inserted into the desired endpoints using MICER-based targeting vectors (25), available through the Wellcome Trust Sanger Institute (Cambridge, UK). These vectors carry loxP sites and either the 3′- or the 5′-HPRT gene fragments as well as agouti and tyrosinase coat color markers, respectively (Fig. 1). The specific MICER clones selected, MHPP-6m1 and MHPN-127h1, carried genomic inserts which mapped between exon 1 and 2 and exon 8 downstream of the non-coding region of Lgi1, respectively (Fig. 1). To generate the deletion at the Lgi1 locus by cis recombination, AB2.2 embryonic stem (ES) cells were first targeted with MHPN-127h1 (Fig. 1). An ES cell clone with successful targeting of the loxP site using MHPN-127h1 was identified through neomycin selection. This clone was then electroporated with MHPP-6m1, a 3′HPRT targeting vector (Fig. 1), and the ES cells were then selected in puromycin. After 9 days of selection, the cells from the puromycin-resistant clones were pooled and electroporated with pOG231, which carries the Cre recombinase. ES cell clones which had undergone loxP recombination reconstitute the HPRT gene and these cells were then selected using hypoxanthine, aminopterin and thymidine (HAT) medium. Sib-selection of the HAT-resistant clones was performed with G418 and puromycin, where 55% were G418- and puromycin-sensitive, suggesting that they carried the desired deletion. Southern blot analysis of DNA derived from these clones identified the specific restriction enzyme fragment that was predicted to be generated as a result of the deletion of the Lgi1 gene (Fig. 1).

Fluorescence in situ hybridization analysis of ES clones with Lgi1 deletion

ES cell clones which were shown by Southern blot analysis to carry hemizygous deletions were further analyzed using fluorescence in situ hybridization (FISH). The deleted region spans a <50 Kb genomic interval, which precluded using BACs in the FISH analysis of the deleted region. Instead, individual genomic DNA probes based on the plasmid clones that mapped within the deleted region were obtained from the Mouse 10 Kb plasmid library prepared by Dr R.B. Weiss at the University of Utah. Plasmid DNA from four individual clones was pooled and used to generate the biotinylated probe. This probe was then applied to metaphase chromosome spreads derived from six independently derived ES cell clones. The mouse Lgi1 gene is located on the distal part of chromosome 19. To unequivocally identify chromosome 19 in the metaphase spread, we used BAC RP23-359N19, which is located on the proximal part of chromosome 19. As shown in Figure 2, the control BAC identifies both copies of chromosome 19 within the metaphase spread, but the Lgi1 locus was only present on one of these chromosomes, demonstrating the successful deletion of the homologous region in all of these cell clones.

Generation of the heterozygous null mice

The mutant ES cell clones were used to generate chimeras using standard procedures. Male mice with high chimerism as judged by coat color were then crossed onto albino C57BL/6J mice to generated mice with the constitutional heterozygous deletion. Germline transmission of the Lgi1 mutation was confirmed using a PCR strategy which first demonstrated the constitutional presence of the reconstituted HPRT gene that resulted from the Cre-mediated deletion event (Fig. 3). These heterozygous null mice were then bred to generate homozygous mutant mice, which initially could be identified using the PCR strategy described in Figure 3. In this approach, eight PCR reactions identified the genomic loci, of which two loci lay outside the predicted deletion on either side and four loci were located within the targeted deletion. Analysis of DNA derived from tail snips of 10-day-old mice demonstrated that, as expected, ∼25% of littermates carried the homozygous deletion. From an analysis of >200 mice, there was no apparent sex ratio distortion and the frequency of the Lgi1-targeted alleles showed normal Mendelian inheritance.

RT–PCR analysis of RNA derived from the hippocampus and cortex of null mutant mice, as expected, showed the absence of Lgi1 gene expression compared with wild-type and heterozygous littermates (Fig. 3). To examine temporal expression of Lgi1, we performed RT–PCR analysis on hippocampal mRNA from wild-type and heterozygous mutant littermates. Lgi1 expression was identified in all animals from 4 days old (P4) through 23 days old (P23) (Fig. 3). In this semi-quantitative analysis, there was evidence for a lower expression level in the heterozygotes as might be expected, although it is still clear that LGI1 expression is present at all time points during the first month of life in these animals.

Cosegregation of coat color markers at the genotypic and phenotypic levels in mutant mice

The MICER clones used to generate the targeted knockout carry either the agouti or tyrosinase coat color markers, which can potentially correlate with specific genotypes within littermates (Fig. 1A). We crossed the original heterozygous targeted mice onto an albino C57BL/6J background where the influence of these genes is often more readily observed in terms of eye and coat color in particular. After six backcrosses, the Lgi1 wild-type mice showed pink eyes and a white coat color, the heterozygous mutants had dark ruby eyes with white coats and the homozygous null mice showed dark ruby eyes with cream-colored coats (Supplementary Material, Fig. S1). These coat-color phenotypes are caused by the presence of one and two copies of the tyrosinase coat color marker in the heterozygous and homozygous mutant mice (Fig. 1A). These phenotypic markers negated the need to genotype mice in this colony in subsequent analyses.

Lgi1 null mice show early-onset seizures

Since LGI1 mutations cause predisposition to ADPEAF in humans, we monitored the mice for seizure activity from birth. Homozygous mutant null mice were viable and developed apparently normally with only a slight reduction in overall body size in some cases. After 12–20 days, however, these mice demonstrated generalized myoclonic seizure phenotypes, which involved uncontrolled jumping and running accompanied by erratic movements which were preceded and followed by periods of relative inactivity involving preening or slow walking. The series of intermittent seizure events were noted to last for ∼20 min (Supplementary Material, videos M1 and M2). Preceding the onset of seizures, these mice were noted to exhibit ‘flagpole’ tail dorsiflexion. During the later stages of the seizures, mice would fall on their sides and exhibit running/galloping motions with the fore and hind limbs. Mice that were followed further would die shortly after an isolated seizure event. These seizure phenotypes were highly consistent between individual mice. Heterozygous null mice have never been observed to exhibit this phenotype and, over a 12–18 month observation period, have not been noted to suffer the premature death experienced by the homozygous null mice.

Hyperexcitability in CA1 hippocampal neurons in Lgi1 null mice

To investigate whether the loss of Lgi1 resulted in altered synaptic function, we investigated extracellular field responses in the CA1 region of the hippocampi from Lgi1-mutant and wild-type mice. When perfused in the modified artificial cerebrospinal fluid (ACSF) (0-Mg²⁺, low-Ca²⁺ and high-K⁺-ACSF), wild-type slices generated characteristic spontaneous epileptiform discharges with burst charges 11.7 ± 1.3 per min (mean ± S.E.M., n = 7). In contrast, the spontaneous burst discharge frequency was significantly increased in the Lgi1-mutant mice (n = 7, P < 0.01; Fig. 4A and B). These results indicate that CA1 neurons in Lgi-mutant mice exhibit higher spontaneous epileptiform-discharge activity in comparison with those in control littermates. To investigate possible mechanisms of the increased discharge activity in mutant mice, we analyzed the intrinsic properties of CA1 pyramidal neurons in current clamp mode. Neurons fired in response to the depolarizing current (+200 pA), but become hyperpolarized when clamped at −200 pA (Fig. 5A). The firing frequency and spikes showed no difference between mutant and wild-type mice (n = 9, P > 0.05; Fig. 5A and B). Moreover, there was no difference in the resting membrane potential (n = 9, P > 0.05; Fig. 5C) and input resistance between the two mice (n = 9, P > 0.05; Fig. 5D). These results indicate that Lgi1 mutation did not change the intrinsic properties of pyramidal neurons in the hippocampal CA1 region. This finding also suggests that the cellular architecture of pyramidal neurons was not altered in the mutant mice, consistent with the gross morphological characterization described below.

Because there is no difference in the intrinsic properties of pyramidal neurons, we next investigated whether inputs onto pyramidal neurons were altered. First, we recorded miniature excitatory postsynaptic synaptic currents (mEPSCs), a response generated by spontaneous glutamate release. The amplitudes of mEPSCs were similar between mutant and wild-type mice, suggesting the LGI1 mutation did not alter the density and response of postsynaptic receptors (n = 12, P > 0.05; Fig. 6A and C). Remarkably, however, the frequency of mEPSCs was dramatically increased in the mutant mice (n = 12, P < 0.01; Fig. 6A and B), indicating increased release of glutamate. Consistently, AMPA receptor-mediated EPSCs (P < 0.01) and NMDA receptor-mediated EPSCs (P < 0.05) were larger in mutant CA1 pyramidal cells than wild-type controls (n = 7 for wild-type mice and n = 8 for Lgi1-mutant mice, respectively; Fig. 6G and H). We have also investigated the effect of LGI1 mutation on GABAergic transmission by measuring miniature and evoked inhibitory postsynaptic currents (mIPSCs and eIPSCs, respectively). The mIPSC frequency (n = 8, P > 0.05; Fig. 6D and E) and amplitude (P > 0.05; Fig. 6D and F) and the eIPSCs amplitude (n = 7, P > 0.05; Fig. 6I and J) showed no difference between mutant and wild-type mice. These results suggest that Lgi1 mutation had no effect on inhibitory synaptic transmission. Instead, it enhances excitatory synaptic transmission by increasing the release of the excitatory neurotransmitter glutamate. These results provide a potential mechanism of epileptiform activity in mutant mice.

Gross anatomical analysis of the null mutant mice

To determine whether the Lgi1 null mice showed any gross developmental abnormalities that might explain the seizures, six individual mice from each of the three genotypes underwent extensive routine histopathological and gross anatomy analysis (conducted by the Molecular and Comparative Pathobiology Service at Johns Hopkins University). Overall, no specific abnormalities were observed in the null mutant mice at this level of resolution, with specific attention paid to the hippocampus and cortex as well as other organs demonstrated to express Lgi1 in our previous studies (11).

The targeting vectors for generating the targeted Lgi1 mutation

The MICER clone-based targeting vectors were obtained from the Wellcome Trust Sanger Institute (25). The end sequences of the genomic insert of MHPN-127h1 were verified by sequencing using primers TTGGCCGATTCATTAATGCAG and TGAAGAAAGTTGAGGAGAGTTTC, whereas the end sequences of the genomic insert of the MICER clone MHPP-6m1 were verified by sequencing using the primers CGTCCCATTCGCCATTCAGGC and AGACAATAGCAGGCATGCTG.

Gene targeting in ES cells

Culture of the AB2.2 line of ES cell line and the method for gene targeting have been described previously (36,37). For generating the deletion at the Lgi1 locus, MHPN-127h1 and MHPP-6m1 were linearized by digestion with HpaI and SmaI, respectively, prior to transfection. The linearized targeting vectors were transfected into ES cells by electroporation, which were selected in G418 or puromycin. Positive clones were identified by Southern blot analysis with one of the following probes: a PCR product external to the genomic insert of MHPN-127h1 and a 2.3 kb PvuII fragment from the BlueScript plasmid vector backbone.

Generation of the targeted Lgi1 mutation in ES cells and mice

The pOG231 cre-expression vector (38) was electroporated into double-targeted clones. ES cell clones with recombined products were then selected in the HAT medium as described previously (24,38). Clones of ES cells that carried the desired deletion at the Lgi1 locus generated by cis recombination were identified by hybridizing Southern blots of NdeI-digested genomic DNA from HAT-resistant ES cell clones that carried the desired deletion between exon 2 and 3′ non-coding region of Lgi1 to the radiolabeled probe B, a 2.3 kb PvuII fragment that was isolated from the BlueScript plasmid vector backbone (Fig. 1). Germline-transmitting chimeras were generated from ES cell lines carrying the targeted Lgi1 allele by microinjection of blastocysts isolated from albino C57B6/J-Tyr^c-Brd females as described previously (39).

Electrophysiological recordings in slices

Brain slices were prepared using standard procedures (40). Briefly, transverse hippocampal slices (0.40 mm) were prepared from P14–P18 Lgi1-mutant mice and wild-type littermates using a Vibroslice (Leica VT 1000S, Leica Instruments, Nussloch, Germany) in an ice-cold solution containing (in mm): 64 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 10 MgSO₄, 0.5 CaCl₂, 26 NaHCO₃, 10 glucose and 120 sucrose. Slices were allowed to recover for at least 2 h (1 h at 34°C followed by 1 h at 22°C) in ACSF containing (in mm): 126 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 2 MgSO₄, 2 CaCl₂, 26 NaHCO₃ and 10 glucose. Slices were transferred to the recording chamber and superfused with ACSF (2 ml/min) at 34°C that was saturated with 95%O₂/5%CO₂. CA1 pyramidal neurons were visualized using an infrared-differential interference contrast microscope with a ×40 water-immersion lens (Axioskop2 Fsplus, Carl Zeiss, Germany) and an infrared-sensitive CCD camera. Whole-cell patch-clamp recordings were performed using an Axon Multiclamp 700B amplifier (Molecular Devices, Sunnyvale, CA, USA) and digitized by the pClamp 9.2 software (Molecular Devices). Both mEPSCs and mIPSCs were recorded at a holding potential of −65 mV. For mEPSCs recording, glass pipettes were filled with the solution containing (mm): 120 K-gluconate, 20 KCl, 10 HEPES, 0.1 EGTA, 2 MgCl₂, 10 sodium phosphocreatine, 0.2 leupeptin, 4 Mg-ATP and 0.3 Na-GTP (pH 7.3, 280 mOsm with sucrose). K-gluconate in the pipette solution was substituted with 140 mm KCl (final concentration) for mIPSC recording. For eEPSC and eIPSC recording, K-gluconate was substituted with 120 mm CsCH₃SO₃ and 5 mm lidocaine N-ethylchloride (QX-314) was added in the pipette. The resistance of pipettes was 3–5 MΩ. AMPA and NMDA receptor-mediated EPSCs were recorded at a holding potential of −65 and +40 mV, respectively; eIPSCs were recorded at a holding potential of −65 mV. Tetrodotoxin (1 µm) was included in the perfusion solution for mEPSC and mIPSC recording and 20 µm bicuculline was used to block GABA_A receptor-mediated currents to record mEPSCs and eEPSCs. For mIPSC and eIPSC recording, dl-2-amino-5-phosphonovaleric acid (50 µm) and 6-cyano-7-nitroquinoxaline-2,3-dione (20 µm) were supplemented to block NMDA and AMPA receptors, respectively. eEPSCs and eIPSCs were generated with a two-concentric bipolar stimulating electrode (25 mm pole separation) positioned ∼100 µm from the neuron under recording. Single pulses of 0.2 ms were delivered at 0.1 Hz and synchronized using a Mater-8 stimulator (A.M.P.I., Israel). Basic biophysical and firing properties were recorded in current-clamp whole-cell configuration using the K-gluconate-based pipette solution coupled with regular perfusion ACSF. De- and hyperpolarizing current steps (+200 and −200 pA, respectively, 200 ms) were applied for characterization of intrinsic properties. Data were filtered at 2 kHz, and sampled at 10 kHz. Neurons with a resting potential of at least –60 mV and resistance that fluctuated within 15% of initial values (<20 MΩ) were analyzed. Mini events were analyzed with MiniAnalysis software (Synaptosoft). The amplitude histograms were binned at 1 pA. Differences between cumulative amplitude histograms were evaluated using the Kolmogorov–Smirnov test. Spontaneous epileptiform burst discharge was recorded extracellularly in the stratum pyramidal of CA1 in a modified ACSF (0 mm MgSO₄, 5 mm KCl, 1.6 mm CaCl₂) by using pipettes (2−3 MΩ) filled with modified ACSF.

FISH analysis

FISH was carried out basically as described by Chernova and Cowell (41). DNA probes were derived from BAC clone RP23-359A49 as well as clones 2M0204H23, 1M0418N07, 1M0559N07 and 1M0343M12 where the plasmids were isolated using standard alkaline lysis procedures and were labeled with digoxigenin-11-dUTP using the DIG-Nick Translation Mix (Roche) or with Biotin-16-dUTP using the Biotin-Nick Translation Mix (Roche). The hybridization mixture contained (in a final volume of 20 µl): 2 µg of the labeled probe, a 4-fold excess of human cot-1 DNA (Roche), a 17-fold excess of salmon sperm DNA (Invitrogen) in 50% formamide, 2X SSC and 10% dextran sulfate. The slide was denatured at 70°C in 70% formamide/2X SSC for 1 min, 30 s followed by incubation in ice-cold 70, 80 and 100% ethanol for 2 min each prior to annealing with probe. The probe mixture was denatured at 80°C for 5 min, incubated at 37°C for 45 min, applied to the surface of the slide, overlaid with a coverslip, sealed and then incubated for 20 h at 37°C. The slide was washed three times at 45°C in 50% formamide/2X SSC, two times in 1X SSC and once in 4X SSC/0.1% Tween. The probes hybridized to nuclei were detected using anti-digoxigenin-fluorescein Fab fragments (Roche) or Avidin-Rhodamine (Roche). Slide preparations were viewed using a fluorescent microscope (Nikon) and the images were captured and adjusted for signal intensity with Easyfish software (ASI).

PCR analysis

Total RNA was isolated from cortex and hippocampus of mice from postnatal days 2 (P2), 6 (P6), 11 (P11), 14 (P14) and 26 (P26) using Trizol Reagent RNA isolation system (Invitrogen Corporation, Carlsbad, CA, USA) following standard protocols. First-strand cDNA was carried out using 1–3 mg total RNA and the standard protocol for the Invitrogen SuperScript II Reverse Transcriptase Preamplification System (Invitrogen Corporation). RT–PCR reactions were carried out using specific primers, forward—GAT CCA TTC CAC GCA CCG TTC CTC, reverse—TCT TCT CTA CGT GGT CCC ATT CCA, spanning exons 1–7 to amplify unique fragments homologous to regions of mouse Lgi1 (NCBI: NM_020278) and Gapdh (NCBI: 008084), forward—CAT GTT TGT GAT GGG TGT GAA CCA CGA G, reverse—GAC AAC CTG GTC CTC AGT GTA GCC CAA G. Reverse transcription was performed using cDNA mixture supplemented with 0.3 µl of Taq DNA polymerase, 10 nm of each primer, 2.0 ml of 10X Taq Reaction Buffer, 1.25 mm MgCl₂, 125 µm of each DNTP, 100 ng/ml DNA template, 5% DMSO and PCR water to bring the final volume to 20 ml. The reaction mixtures were incubated using a PTC-100 SingleBlock System temperature cycler. PCR reactions were programmed to repeat the following cycles: one cycle with a 5 min denaturation at 95°C, 30 cycles annealing of 30 s at 94°C, 30 s at 61°C and 30 s at 72°C, followed by 5 min extension at 72°C, then bring to and hold at 4°C. The PCR products were electrophoresed on 0.8% agarose DNA grade (Thermo Fisher Scientific, Waltham, MA, USA) gels in 1X TAE. Genotyping was carried out using primers that amplify unique fragments within the mouse Lgi1 locus (Table 1). PCR reactions were performed using genomic DNA mixture supplemented with 0.3 µl of Taq DNA polymerase, 10 nm of each primer, 2.0 ml of 10X Taq Reaction Buffer, 1.25 mm MgCl₂, 125 µm of each DNTP, 100 ng/ml genomic DNA template and PCR water to bring the final volume to 20 ml. The reaction mixtures were incubated using a PTC-100 SingleBlock System temperature cycler. PCR reactions were programmed to repeat the following cycles: one cycle with a 5 min denaturation at 95°C, 35 cycles annealing of 30 s at 94°C, 30 s at 55°C and 30 s at 72°C, followed by 5 min extension at 72°C. The PCR products were electrophoresed on 0.8% agarose DNA grade (Thermo Fisher Scientific) gels in 1X TAE.