Elucidating molecular mechanisms underlying limbic epileptogenesis may reveal novel targets for preventive therapy. Studies of TrkB mutant mice led us to hypothesize that signaling through a specific phospholipase (PLC), PLCγ1, promoted development of kindling.
To test this hypothesis, we examined the development of kindling in PLCγ1 heterozygous mice. We also examined the cellular and subcellular location of PLCγ1 in adult wild-type mice.
The development of kindling was impaired in PLCγ1 heterozygous mice compared to wild-type controls. PLCγ1 immunoreactivity was localized to the soma and dendrites of both excitatory and inhibitory neurons in the hippocampus of adult mice.
This study implicates PLCγ1 signaling as the dominant pathway by which TrkB activation promotes limbic epileptogenesis. Its cellular localization places PLCγ1 in a position to modify the efficacy of both excitatory and inhibitory synaptic transmission. These findings advance PLCγ1 as a novel target for therapies aimed at preventing temporal lobe epilepsy induced by status epilepticus.
Dr. He is a senior neuroscientist at Duke University Medical Center.
Temporal lobe epilepsy (TLE) is a common and often devastating form of human epilepsy that currently lacks preventive therapy. Elucidating the molecular mechanisms of epileptogenesis can provide targets for preventive therapies. A number of attractive candidates have recently emerged including mammalian target of rapamycin (mTOR), repressor element silencing transcription factor (REST), prostaglandin E2 receptor subtype EP2, and mir-134.
Work from multiple laboratories including our own has advanced the idea that enhanced activation of the brain-derived neurotrophic factor (BDNF) receptor, TrkB, promotes development of TLE in animal models and humans. Contrary to this idea, Paradiso et al. reported that viral-mediated delivery of fibroblast growth factor-2 and BDNF following status epilepticus reduced spontaneous seizures and cell death in the pilocarpine model; whether similar results ensue from viral-mediated delivery of BDNF alone is presently unclear. Using a chemical-genetic approach, Liu et al. developed proof-of-concept evidence that inhibiting TrkB kinase and its downstream signaling transiently following status epilepticus prevented TLE. This provides a rationale for developing inhibitors of TrkB signaling for preventive therapies. Crystal structure of the ATP-binding domain of TrkB reveals striking similarity to other tyrosine kinases, underscoring the challenge of developing selective inhibitors.
An alternative strategy approach to limiting TrkB signaling would be to identify the signaling mechanisms by which TrkB activation promotes epileptogenesis. Activation of TrkB induces two major signaling pathways: phosphorylation of Y515 leads to binding of the adaptor protein, SHC; phosphorylation of Y816 leads to binding and activation of the enzyme PLCγ1. Studies of TrkB mutant mice in which phenylalanine was substituted for tyrosine at residue 515 or 816 revealed impairment of development of kindling in TrkBY816F but not TrkBY515F.[10, 11] This suggested that the proepileptogenic consequences of TrkB activation are mediated by its activation of PLCγ1. That said, multiple adaptor proteins and enzymes can bind a given motif of a receptor tyrosine kinase, underscoring the importance of determining whether PLCγ1 signaling per se promotes epileptogenesis. Null mutations of PLCγ1 result in death at embryonic day 9.5. Therefore, to test the hypothesis that PLCγ1 promotes epileptogenesis, we examined the development of kindling in Plcγ1 heterozygous mice.
PLCγ1 mutant mice were generated by targeted deletion of genomic sequences encoding the X domain and both SH2 domains of PLCγ1 as described previously. Note that line generated with replacement vector TV-1 was used in these experiments. Homozygous disruption of PLCγ1 (−/−) results in embryonic lethality at approximately embryonic day 9.0–9.5. Immunoblot analysis of mouse embryonic fibroblasts (MEFs) revealed no detectable full-length PLCγ1 protein in PLCγ1−/− MEF and reduced PLCγ1 protein content in PLCγ1 heterozygous (+/−) MEF. Thirteen heterozygotes of PLCγ1 and 11 wild-type (+/+) littermates were used for the kindling experiment. The genotype of each animal was assessed twice using polymerase chain reaction (PCR) of genomic DNA isolated from tails (before and after experiments). No overt differences were detected in appearance of +/− and +/+ mice. The average body weights of +/− (22.15 ± 1.10 g) and +/+ (23.45 ± 0.73 g) were not significantly different. We crossed inbred strains of PLCγ1 mutant mice on a 129/SvJ background to C57BL/6 for three generations.
To verify the specificity of PLCγ1 antibody, we generated syn-cre+PLCγ1flox/flox(syn-cre+PLCγ1f/f) mice. Crossing PLCγ1 floxed mouse to syn-cre mouse generated progeny in which expression of the floxed gene was selectively eliminated in a subset of central nervous system (CNS) neurons. The syn-cre+PLCγ1f/f mice are viable and exhibit no overt differences from their syn-cre−PLCγ1f/f littermates.
Procedures involving animals followed National Institutes of Health (NIH) guidelines for the care and use of experimental animals. All experiments conformed to local guidelines on the ethical use of animals.
Both +/− and +/+ mice were decapitated under deep pentobarbital anesthesia (250 mg/kg, i.p.). Following decapitation, the mouse head was dipped into liquid nitrogen for 4 s to rapidly cool the brain. The hippocampi, amygdala, and cortex were rapidly dissected on ice and homogenized in lysis buffer [20 mm Tris (pH 8.0), 137 mm NaCl, 1% NP40, 10% glycerol, 1 mm phenylmethylsulfonylfluoride (PMSF), and one Complete Mini protease inhibitor tablet (Mini, Roche, Mannheim, Germany)/10 ml]. The supernatant was saved following centrifugation at 16,000 g for 10 min, aliquoted, and stored at −80°C for further biochemical analysis.
Western blotting was performed to analyze PLCγ1 content using procedures as described previously.[11, 14] Antibodies to PLCγ1 (Cell Signaling, Danvers, MA, U.S.A.) and β-actin (Sigma, St Louis, MO, U.S.A.) were used in these experiments. Results from Western blotting were quantified as described previously. Briefly, the immunoreactivity of individual bands on Western blots was measured by Image J software, and β-actin was used to control for loading and transfer. Western blotting of brain homogenates prepared from 4+/+ and 4+/− mice revealed reductions of 37%, 45%, and 44% in hippocampus, amygdala, and cortex, respectively (Fig. 1), thereby confirming reduced expression of PLCγ1 protein in the PLCγ1+/− mice.
PLCγ1 immunohistochemistry was performed using a protocol described previously. Briefly, under pentobarbital anesthesia, mice were perfused with 4% paraformaldehyde in phosphate-buffered saline (PBS), and the brains were removed, postfixed for 4 h, and cryoprotected until brains sank. Forty micrometer coronal sections were cut and used for immunofluorescent staining. The following antibodies were used for immunofluorescence staining: PLCγ1 (Cell Signaling), neuronal nuclei (NeuN, Millipore, Temecula, CA, U.S.A.), glial fibrillary acidic protein (GFAP, Sigma, Louis, MO, U.S.A.), glutamate decarboxylase 67 (GAD67, Millipore), Parvalbumin (Sigma), microtubule associated protein2 (MAP2, Sigma), Tau1 (Millipore), Synapsin 1 (Synaptic Systems, Geottingen, Germany). After 1 h incubation with blocking solution (5%NGS, 0.5%NP40 in PBS buffer), the primary antibodies were applied to floating sections overnight at 4°C. Alexa Fluor 594 goat anti-rabbit and Alexa Fluor488 goat anti-mouse secondary antibodies (Invitrogen, Eugene, OR, U.S.A.) were used to visualize the immunofluorescent staining. Images were captured using a Leica True Confocal Scanner SP2 (Nussloch, Germany) system. To quantify the colocalization of PLCγ1 with GAD67, two images under 63× objective were taken from CA1 oriens and stratum lacunosum-moleculare (SLM), respectively, and the cells with PLCγ1-ir+ or GAD67-ir+ alone, or both. The results are presented as the percentage of PLCγ1-ir+ or GAD67-ir+ cells averaged from two adjacent sections in two naive mice. The specificity of PLCγ1 antibody was verified by the reductions of immunoreactivity in CA3 and dentate gyrus in syn-cre+PLCγ1flox/flox mice compared with the WT controls (Fig. 3).
Electrode implantation and kindling procedures
Eleven +/+ and 13 +/− mice were included in the kindling experiment. A bipolar electrode for stimulation and recording was stereotactically implanted in the right amygdala under pentobarbital anesthesia (60 mg/kg) using the following coordinates with bregma as the reference: 1.2 mm posterior, 2.9 mm lateral, 4.6 mm below dura. After a postoperative recovery period of 1 week, the electrographic seizure threshold (EST) was determined by administering a 1 s train of 1 msec biphasic rectangular pulses at 60 Hz beginning at 40 μA. Additional stimulations increasing by 10 μA were administered at 1 min intervals until an electrographic seizure lasting at least 5 s was detected on the electroencephalography (EEG) recorded from the amygdala. The behavioral manifestations of seizures were classified according to a modification of the description of Racine (1972). Stimulations were subsequently administered twice per day at an intensity of the EST until the animals exhibited three consecutive seizures of class 4 or greater with limb clonus and/or tonus lasting at least 12 s. The surgery and kindling procedures were performed by an individual blinded to genotype of the animals.
After a stimulation-free period of 2 weeks following the third consecutive seizure of class 4 or greater, persistence of the hyperexcitability was examined. The animals were perfused transcardially under deep pentobarbital anesthesia with ice-cold 4% paraformaldehyde in 1× PBS for 5 min at a flow rate of 7.5 ml/min. The brains were removed, post fixed in the same solution overnight at 4°C, and cryoprotected in 30% sucrose in 1× PBS until they sank. Coronal sections of 40 μm were cut in a cryostat at −20°C, mounted in PBS on slides, air-dried, frozen and stored at −80°C until use. The sections were stained with methyl green pyronine-Y (MGPY) and analyzed for electrode placement by a person without knowledge of genotype. Only animals with correct electrode placement in the amygdaloid complex were included in the statistical analysis for kindling experiment.
The results from Western Blot and kindling development were analyzed by two-tailed Student's t-test. Data sets are presented as mean ± standard error of the mean (SEM) for the designated number of experiments.
Kindling development was inhibited in PLCγ1+/−
We hypothesized that the development of kindling would be inhibited in PLCγ1+/− mice for two reasons. First, the development of kindling was impaired in TrkB mutant mice in which phenylalanine was substituted for tyrosine residue 816 of TrkB; one protein bound by the motif of TrkB containing phosphorylated tyrosine 816 is PLCγ1. Second, use of a surrogate measure of PLCγ1 activation revealed its enhanced activation in both kindling and status epilepticus models. That said, because multiple adaptor proteins and enzymes can bind a given motif of a receptor tyrosine kinase, we sought to test whether signaling through PLCγ1, in particular, affected development of kindling. To address this question, we examined the development of kindling in PLCγ1+/− mutant mice. PLCγ1+/− mice exhibited a significant impairment of limbic epileptogenesis, as evident in the increased number of stimulations required to elicit behavioral seizures compared with +/+ (Fig. 2B and D). The number of stimulations required to evoke a limbic seizure termed class 1 or 2 was increased twofold in +/− (4.23 ± 0.62, n = 13) compared with +/+ (2.09 ± 0.28, n = 11, p < 0.01). Likewise, the number of stimulations required to induce the first class 4 or 5 seizure was significantly increased in +/− (7.38 ± 0.75) compared with +/+ controls (4.64 ± 0.41, p < 0.01). Similarly, the number of stimulations required to induce the third consecutive class 4 or greater (fully kindled stage) was increased in +/− (14.69 ± 0.93) compared with +/+ controls (9.45 ± 0.65, p < 0.001, Student's t-test). No significant difference was detected in the current required to evoke an initial electrographic seizure (179 ± 27 μA for +/− and 143 ± 20 μA for +/+) (Fig. 2A). Nor was a significant difference detected in the duration of electrographic seizure between +/− and +/+, during either the first evoked electrographic seizure or third consecutive class 4 or above seizure (Fig. 2C). Of interest, the cumulative duration of evoked electrographic seizure prior to the first seizure class 4 or 5 for +/+ was 49.18 ± 4.90 s in comparison to 90.92 ± 12.33 s for +/− mice, an increase of almost twofold (p = 0.006). That is, a greater duration of electrographic seizure activity is required prior to the first generalized seizure in +/− compared to +/+ mice. To determine whether the disruption of PLCγ1 influenced the persistence of the hyperexcitability after completion of kindling, +/− and +/+ mice were subjected to a single stimulation following a stimulation-free period of 2 weeks after kindling. All wild-type and all PLCγ1 heterozygotes exhibited an electrographic seizure of similar duration (33.31 ± 2.22 s for +/−; 30.73 ± 2.45 s for +/+) in response to the initial stimulation, and the proportion of each exhibiting seizures of class 4 or 5 were similar (WT 82%; PLCγ Hets 92%), thereby excluding any substantial elevation of threshold. In summary, kindling development was impaired but the persistence of kindling was unaffected in the PLCγ1+/− compared to +/+ mice.
Cellular localization of PLCγ1 in hippocampus
Understanding how PLCγ1 signaling could promote limbic epileptogenesis would be facilitated by knowing its cellular localization within the CNS of adult mice, leading us to perform immunohistochemical analyses. Meaningful interpretation of these experiments requires establishing the specificity of the PLCγ1 antibody. Towards this end, we performed PLCγ1 immunohistochemistry in conditional PLCγ1 knockout mice (syn-cre+PLCγ1f/f) and wild-type controls, which included syn-cre−PLCγ1f/f and syn-cre−PLCγ1+/+ mice. The PLCγ1f/f mice were crossed to a transgenic mouse in which Cre recombinase is driven by a synapsin1 promoter, a line in which Cre recombinase is expressed in CA3 pyramidal cells and dentate granule cells but to a lesser extent in CA1 pyramidal cells. A reduction of PLCγ1-immunoreactivity (PLCγ1-ir) in CA3 pyramidal and dentate granule cell body layers was evident in syn-cre+PLCγ1f/f compared to control mice. Note the relative preservation of PLCγ1-ir+ in CA1 (arrow heads) compared to CA3 (arrows) or dentate gyrus in syn-cre+PLCγ1f/f (Fig. 3B) compared to control mice (Fig. 3A). The anatomic distribution of the reduced PLCγ1-ir+ in the syn-cre+PLCγ1f/f compared to control mice parallels the expression of syn-cre and supports the specificity of the PLCγ1 antibody used in this experiment.
Consistent with the pattern of its messenger RNA (mRNA) distribution,[17, 18] PLCγ1 was widely expressed in most brain regions of wild-type mice. Within the hippocampus, staining of the principal cell bodies, such as pyramidal neurons of CA1 and CA3 as well as dentate granule cells, was prominent (Fig. 3A). In addition, some PLCγ1-ir+ cells were scattered throughout dentate hilus, stratum lucidum of CA3, and stratum radiatum, oriens, and lacunosum-moleculare in both CA3 and CA1, a distribution consistent with γ-aminobutyric acid (GABA) interneurons (Figs. 3A and 4C,D). Notably, the staining intensity of these cells was stronger than that of principal neurons.
To examine the nature of PLCγ1-ir+ cells, we conducted three immunohistochemical experiments. For the first step, double staining was performed using antibody to PLCγ1 together with antibodies to either a neuronal marker, NeuN, or an astrocyte marker, GFAP. PLCγ1-ir+ cells (red), including principal cells and scattered cells, coincided with NeuN-ir+, but not with GFAP-ir+ (green, Fig. 4B,A, respectively), establishing the neuronal localization of PLCγ1-ir+ cells. Next, to further analyze the type of PLCγ1-ir+ neurons, glutamate decarboxylase 67 (GAD67) was used as a marker of GABA-containing neurons. The double staining revealed that more than half of scattered PLCγ1-ir+ neurons were GAD67-ir+: 62% in stratum lacunosum-moleculare (SLM) and 73% in stratum oriens of CA1. Likewise, the majority of GAD67-ir+ neurons exhibited PLCγ1 immunoreactivity: 95% in SLM and 90% in stratum oriens (Fig. 4C). To begin to examine which subsets of GABA interneurons exhibit PLC-ir+, colocalization of parvalbumin, a marker of fast-spiking basket cells with PLCγ1-ir+ was examined. Parvalbumin-ir+ cells distribute broadly within hippocampus and most of them exhibited PLCγ1-ir+ (Fig. 4D). Parvalbumin-ir was cytoplasmic and dendritic; however, PLCγ1-ir was mainly somatic (Fig. 4D). In sum, PLCγ1-ir is localized to both principal neurons and GABA interneurons in adult hippocampus, including parvalbumin subtype of interneurons.
To further explore the subcellular loci of the PLCγ1-ir within hippocampal neurons, sections from wild-type mice were colabeled with an antibody to PLCγ1 and markers of axons (Tau) or dendrites (MAP-2). PLCγ1-ir prominently localized in the soma of neurons (Fig. 5A, arrow heads) and short processes in subsets of cells in dentate hilus (Fig. 5A, arrows). The colocalization of these processes with MAP2-ir+ supports their identity as proximal dendrites (Fig. 5A). Similarly, PLCγ1-ir prominently localized in CA1 pyramidal cells. Within CA1 pyramidal cells, PLCγ1-ir is mainly evident in soma and proximal portion of apical dendrites (not shown). Axon marker, tau, nicely labeled the axons of the dentate granule cells, named mossy fiber. Surprisingly, no overlap of tau- and PLCγ1-ir was evident in mossy fiber axons within stratum lucidum of CA3 (Fig. 5B, arrows). These findings demonstrate that PLCγ1-ir is largely confined to neuron soma, some within proximal dendrites, but no detectable axonal localization within the axons of dentate granule cells.
Our prior studies led us to hypothesize that signaling mediated by PLCγ1 promotes limbic epileptogenesis. To test this hypothesis, we examined the rate of kindling development in mice heterozygous for PLCγ1. We also ascertained the cellular distribution of PLCγ1 in hippocampus of adult mice using a PLCγ1-specific antibody. The following two principal findings emerged: (1) The rate of development of kindling was impaired in PLCγ1 heterozygotes in comparison to wild-type control mice; and (2) PLCγ1-immunoreactivity was evident in neurons but not astrocytes and expressed in both principal excitatory neurons and GAD-67 interneurons. PLCγ1 immunoreactivity was detectable in soma and dendrites but not axons of these neurons. These findings implicate a causal role for PLCγ1 signaling in limbic epileptogenesis and provide anatomic locales at which to pursue the cellular mechanisms responsible for its proepileptogenic actions.
We hypothesized that the development of kindling would be inhibited in PLCγ1+/− mice for two reasons. First, the development of kindling was impaired in TrkB mutant mice in which phenylalanine was substituted for tyrosine residue 816 of TrkB, a motif that can bind PLCγ1 and promote its activation. Second, use of a surrogate measure of PLCγ1 activation revealed its enhanced activation in the kindling model. That said, because multiple adaptor proteins and enzymes can bind a given motif of a receptor tyrosine kinase, we sought to test whether signaling through PLCγ1, in particular, affected development of kindling. The lack of a drug that selectively inhibits PLCγ1 necessitated a genetic approach to test this hypothesis. The embryonic lethality of PLCγ1 null mutants led us to test this hypothesis in PLCγ1 heterozygotes. Although reductions of PLCγ1 protein approximated only 40%, clear impairments of development of kindling were evident in the PLCγ1+/− mice. The magnitude of the impairments was less than that of trkBPLC/PLC mice, the differences likely due to residual PLCγ1 available for activation in the PLCγ1+/− mice. That said, the present results support the conclusion that the dominant signaling pathway mediating the proepileptogenic consequences of TrkB activation in the kindling model involves PLCγ1.
As noted above, enhanced activation of PLCγ1 has been identified in the kindling model as revealed by enhanced immunoreactivity of the surrogate marker, p-PLCγ1 (pY783); it is important to note that no differences in expression of PLCγ1 protein were detected. What cellular and molecular consequences of PLCγ1 activation might promote the development of kindling? The cellular and subcellular locale of PLCγ1 provides a context within which to consider this question. The present findings confirm and extend analyses of PLCγ1 immunoreactivity in rat brain, analyses revealing predominant neuronal localization including dentate granule cells and pyramidal cells within both CA3 and CA1 of hippocampus. The present studies revealed PLCγ1 immunoreactivity in CNS neurons, including both principal excitatory neurons and inhibitory interneurons, a pattern similar to that of TrkB itself. No PLCγ1 immunoreactivity colocalized with an astrocyte-specific marker, GFAP. Within neurons PLCγ1 immunoreactivity was most prominent within soma and to a lesser extent within proximal dendrites; whether an antibody with increased sensitivity would detect PLCγ1 immunoreactivity within axon terminals awaits further study.
What molecular and cellular consequences of PLCγ1 activation might promote epileptogenesis? The addition of BDNF to acutely isolated hippocampal slices produces a striking reduction of the K-Cl cotransporter, KCC2, the predicted consequence of which is accumulation of [Cl−]i and a shift of EGABA in a depolarizing direction, an effect predicted to reduce synaptic inhibition. It is notable that the TrkB-mediated reduction of KCC2 expression presumably involves PLCγ1 because this effect required Y816 of TrkB. Findings from humans[22, 23] and diverse in vivo and in vitro models[21, 24-28] advance reduced KCC2 expression as one mechanism underlying hyperexcitability of epileptic tissue. With respect to the kindling model in particular, striking reductions of KCC2 mRNA and protein have been identified in the dentate gyrus and CA3 and CA1 regions of hippocampus at multiple times following kindling. The localization of PLCγ1 to dendrites positions it to mediate such an effect. PLCγ1 activation might also promote the efficacy of excitatory synaptic transmission. That is, impairments of long-term potentiation of both mossy fiber-CA3 and CA3-CA1 synapses[11, 29] have been identified in hippocampal slices isolated from trkBPLC/PLC mice. In sum, PLCγ1 activation may contribute to enhanced excitability in epileptic tissue by both enhancing excitatory and limiting inhibitory synaptic efficacy.
The present findings in the kindling model raise the question as whether PLCγ1 activation might promote limbic epileptogenesis in other models, specifically those induced by a brief episode of status epilepticus. Indeed, status epilepticus induced by pilocarpine results in a time-dependent increase of PLCγ1 activation without change in its expression. Transient inhibition of TrkB signaling was recently shown to prevent epilepsy induced by status epilepticus in a different model, one induced by microinfusion of kainic acid into the amygdala. Whether the proepileptogenic consequences of TrkB activation in the kainic acid status epilepticus model are mediated by activation of PLCγ1 is presently unclear. Activation of PLCγ1 results in hydrolysis of phosphatidylinositol 4, 5-bisphosphate (PIP2), producing diacylglycerol (DAG) and inositol 1, 4, 5-trisphosphate (IP3). IP3 binds to IP3 receptors in the endoplasmic reticulum, resulting in a transient increase of intracellular free Ca2+.[30, 31] Of interest Pal et al. implicated calcium release from endoplasmic reticulum as a causal event in epileptogenesis in the cultured hippocampal neuron model; it seems plausible that hydrolysis of PIP2 in this model may have been mediated by PLCγ1.
In sum, the present findings establish PLCγ1 as a molecular mechanism underlying development of kindling and implicate PLCγ1 signaling as the dominant signaling pathway mediating the proepileptogenic consequences of TrkB activation in this model. These findings advance PLCγ1 as a novel target for development of preventive agents for limbic epileptogenesis. Although PLCγ1 inhibitors might induce some impairment of associative learning, the brief period of exposure following status epilepticus likely required for prevention would minimize such unwanted effects.
We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. None of the authors has any conflict of interest to disclose.
This work was supported by NINDS grants NS56217 and NS060728 (J.O.M.).