Deletion of phospholipase C �4 in thalamocorticalrelay nucleus leads to absence seizuresEunji Cheonga, Yihong Zhenga,b, Kyoobin Leea, Jungryun Leea, Seongwook Kima,b, Maryam Sanatia,b, Sukyung Leec,Yeon-Soo Kimc, and Hee-Sup Shina,b,1

aCenter for Neural Science, Korea Institute of Science and Technology, Seoul 136-791, Korea; bDepartment of Neuroscience, University of Scienceand Technology, Daejeon 305-333, Korea; and cLaboratory of Retroviruses and Gene Therapy, Department of Smart Foods and Drugs and IndangInstitute of Molecular Biology, Inje University, 64 Jeodong 2nd Street, Indang Building, Jung-Ku, Seoul 100-032, Korea

Contributed by Hee-Sup Shin, October 28, 2009 (sent for review September 2, 2009)

Absence seizures are characterized by cortical spike-wave dis-charges (SWDs) on electroencephalography, often accompanied bya shift in the firing pattern of thalamocortical (TC) neurons fromtonic to burst firing driven by T-type Ca2� currents. We recentlydemonstrated that the phospholipase C �4 (PLC�4) pathway tunesthe firing mode of TC neurons via the simultaneous regulation ofT- and L-type Ca2� currents, which prompted us to investigate thecontribution of TC firing modes to absence seizures. PLC�4-deficient TC neurons were readily shifted to the oscillatory burstfiring mode after a slight hyperpolarization of membrane poten-tial. TC-limited knockdown as well as whole-animal knockout ofPLC�4 induced spontaneous SWDs with simultaneous behavioralarrests and increased the susceptibility to drug-induced SWDs,indicating that the deletion of thalamic PLC�4 leads to the genesisof absence seizures. The SWDs were effectively suppressed bythalamic infusion of a T-type, but not an L-type, Ca2� channelblocker. These results reveal a primary role of TC neurons in thegenesis of absence seizures and provide strong evidence that analteration of the firing property of TC neurons is sufficient togenerate absence seizures. Our study presents PLC�4-deficientmice as a potential animal model for absence seizures.

epilepsy � gene knockdown � knockout mice � thalamus

Absence seizures are generalized nonconvulsive seizures char-acterized by a brief and sudden impairment of consciousness,

concomitant with bilaterally synchronized spike-and-wave dis-charges (SWDs) in the electroencephalogram (EEG) over widecortical areas (1–4). Abnormal hypersynchronized oscillatory ac-tivities in the thalamocortical network, consisting of feedforwardand feedback connections between the cortex and the thalamus,have been implicated as an underlying mechanism for the gener-ation of SWDs (5–9). Some studies using rat models of absenceseizures have suggested that the cortex plays a leading role in thegeneration of SWDs (10–13). Other studies support the hypothesisthat massive thalamocortical synchronization is driven from recur-rent oscillatory activities in the network between reticular thalamicnucleus (nRT) and thalamocortical (TC) relay nucleus (3, 8, 9, 14,15). A majority of these studies proposed a leading role for nRTneurons in the genesis of absence seizures. Relatively less attentionhas been directed on the role of TC neurons in the generation ofSWDs. Thalamocortical network oscillations are often observed tobe accompanied by a shift in the firing pattern of thalamocortical(TC) neurons from tonic to burst firing (16). Low-threshold burstfiring driven by T-type Ca2� currents in TC neurons has long beenproposed to be a critical component in sustaining the oscillationsduring the SWDs (3, 8, 17), although a controversy still remains(4, 18).

Many studies have described spontaneous appearance of SWDsin the cortical EEG from rodent models for absence epilepsy(19–24). Some showed that T-type Ca2� currents were increased inthe TC neurons of mutant mice with spontaneous absence epilepsy(25–27). Our previous study showed that mice deficient for the �1GT-type Ca2� channel were resistant to the generation of SWDs in

response to activation of type B gamma-aminobutyric acid(GABAB) receptors (17). Another study showed that �1G T-typeCa2� channels play a critical role in the genesis of spontaneousabsence seizures that result from hypofunctioning P/Q-type chan-nels, but also demonstrated that augmentation of thalamic T-typeCa2� currents is not an essential step in the genesis of absenceseizures (27). In contrast, a recent report showed that transgenicmice overexpressing the Cav3.1 gene for �1G T-type calciumchannels in the whole brain exhibited spontaneous absence epi-lepsy, an observation that suggested a causal relationship betweenthe elevation of �1G T-type calcium channel activity and absenceepilepsy (28). A limitation common to all these mouse models,however, is that the alteration of T-currents was not restricted totheTC relay nucleus but general to other brain regions including thecortex. This limitation makes it not possible to establish a conclusivelink between a change in T-type Ca2� currents in TC neurons andthe occurrence of SWDs.

PLC�4 is highly expressed in TC neurons where it functions asa downstream signaling molecule of type 1 metabotropic glutamatereceptors (mGluR1s), which mediate corticothalamic excitatoryinputs (29, 30). We recently reported that thalamic PLC�4 pathwaytunes the firing modes of TC neurons via simultaneous modulationof T- and L-type Ca2� channels. The amplitudes of both T- andL-type Ca2� currents were increased in PLC�4-null (PLC�4�/�)TC neurons, and activation of protein kinase C (PKC), a down-stream signaling molecule of PLC�4, reversed the increase in bothof these Ca2� currents (31). These observations prompted us toexamine the role of TC firing modes regulated by thalamic PLC�4in the genesis of absence seizures.

We show here that TC-limited knockdown as well as whole-animal knockout of PLC�4 in mice induced spontaneous absenceseizures and also increased the sensitivity to drug-induced absenceseizures. Our data reveal a primary role for TC neurons in thegenesis of absence seizures, providing evidence that an alteration inthe firing properties of TC neurons caused by a disruption of asingle gene is sufficient to induce absence seizures.

ResultsPLC�4�/� Mice Show Spontaneous SWDs Accompanied by BehavioralArrests. We have recently reported that thalamic PLC�4 pathwaytunes the firing modes of TC neurons by simultaneous modulationof T- and L-type Ca2� channels (31). To examine the role ofthalamic PLC�4 in the genesis of SWDs, we collected EEGrecordings from PLC�4�/� mice and their wild-type littermates

Author contributions: E.C. and H.-S.S. designed research; E.C., Y.Z., J.L., S.K., and M.S.performed research; K.L., S.L., and Y.-S.K. contributed new reagents/analytic tools; E.C.,Y.Z., and K.L. analyzed data; and E.C. and H.-S.S. wrote the paper.

The authors declare no conflict of interest.

1To whom correspondence should be addressed at: Center for Neural Science, KoreaInstitute of Science and Technology, 39-1 Hawolgok-dong, Seongbuk-gu, Seoul 136-791,Korea. E-mail: shin@kist.re.kr.

This article contains supporting information online at www.pnas.org/cgi/content/full/0912204106/DCSupplemental.

21912–21917 � PNAS � December 22, 2009 � vol. 106 � no. 51 www.pnas.org�cgi�doi�10.1073�pnas.0912204106

Dow

nloa

ded

by g

uest

on

Apr

il 15

, 202

1

simultaneously with video monitoring of behavior. We observedhigh-amplitude fluctuations, SWDs, in the EEGs of PLC�4�/�

mice while they were awake, phenomena that were never observedin wild-type littermates. The SWD activities in PLC�4�/� micewere bilaterally synchronized, and occurred simultaneously in thecortex and the thalamus (Fig. 1A and Fig. S1). Each episode ofSWD was accompanied by a behavioral arrest (Movie S1). TheSWD pattern of PLC�4�/� mice typically consisted of sharp,spike-like discharges and slow waves. On average, PLC�4�/� miceshowed the number of SWD events of 87.9 � 8.8 per hour (range,14–210; n � 37) and an average SWD duration of 1.3 � 0.1 s (Fig.1B). Averaging the power spectrums of spontaneous SWD seg-ments in PLC�4�/� mice by Fourier analysis revealed a peak SWDfrequency of 7.8 Hz (range, 4–10 Hz; Fig. 1C). Because SWDs haveharmonic components, there were local peaks at frequencies 2-fold(15.6 Hz) and 3-fold (23.5 Hz) higher. The SWDs in PLC�4�/� micedwindled substantially after an i.p. injection of ethosuximide (200mg/kg), an anti-absence epilepsy drug (P � 0.05, paired t test; n �6), whereas injection of vehicle (0.9% NaCl) had no effect (Fig. 1A and D). At the same dose, ethosuximide had no effect on theEEG pattern of wild-type mice (Fig. S1).

PLC�4�/� TC Neurons Are Prone to be Shifted to Burst Firing Mode. Toexamine the role of firing pattern of PLC�4�/� TC neurons in thegeneration of SWDs, we investigated the firing properties of TCneurons in VB region from brain slices. Here, we observed that thefiring mode of PLC�4�/� TC neurons was easily shifted to low-threshold burst firing after a series of tonic firings (15 of 21 cells),a finding that has not previously been observed in wild-type TCneurons (0 of 17 cells) (Fig. 2 A and B). We previously reported thatmutant TC neurons showed an increased tendency for burst firingafter a slight hyperpolarization (31). Here, we further investigatedthe tendency for burst firing in wild-type and PLC�4�/� TCneurons with various hyperpolarizing steps. The difference wasclear near the resting membrane potential (Fig. 2 C–E). PLC�4�/�

TC neurons (n � 21) generated low-threshold burst firing after

hyperpolarizing steps between �67 and �64 mV, whereas none ofwild-type TC neurons (n � 17) did. There was no difference in theresting membrane potential [�60.24 � 5.63 mV (n � 17) vs.�58.74 � 6.71 mV (n � 21)] or in input resistance [167.03 � 31.45M� (n � 17) vs. 158.91 � 29.61 M� (n � 21)] between wild-typeand PLC�4�/� TC neurons as described in ref. 31.

PLC�4�/� Mice Are More Susceptible to GABAB Receptor Agonist-Induced Generation of SWDs. Our previous study showed that micelacking �1G T-type Ca2� channels were specifically resistant toGABABR agonist-induced absence seizures (17). To investigate thesensitivity of PLC�4�/� mice to GABABR agonist, we used �-bu-tyrolactone (GBL) known to induce absence seizures primarily byacting on GABAB receptors (32, 33) and RS(�/�) baclofen, aselective GABAB receptor agonist (34).

Although PLC�4�/� mice exhibited spontaneous SWDs beforethe injection of drugs, the total duration of SWDs per totalrecording time was relatively short (Fig. 3 B and C). i.p. adminis-tration of GBL (40 and 70 mg/kg) or RS(�/�) baclofen (20 mg/kg)elicited 2–5-Hz paroxysmal SWDs in both PLC�4�/� mice andwild-type littermates (Fig. 3 A–C and Fig. S2). However, thequantitative difference in the total duration of SWDs between thetwo genotypes was striking in mice administered 40 mg/kg GBL(P � 0.001 at all time points; Fig. 3B) or RS(�/�)-baclofen (P �0.001 or P � 0.01, depending on time point; Fig. 3C). Averagedtemporal power spectra exhibited a stronger power density in the2–5-Hz frequency range corresponding to epileptic SWDs in GBL-or baclofen-treated PLC�4�/� mice compared with drug-treatedwild-type mice (Fig. 3 D and E). These data indicate that PLC�4�/�

mice are more susceptible to GABAB receptor agonist-inducedabsence seizures.

Silencing Thalamic PLC�4 Using Short Hairpin RNA (shRNA) InducesAbsence Seizures. A previous study using in situ hybridizationtechniques has reported abundant expression of PLC�4 mRNA inthe TC relay nuclei, whereas lower expression was observed in the

Fig. 1. PLC�4�/� mice exhibit spontaneous SWDsaccompanied by behavioral arrests. (A) SynchronizedSWDs were observed between the frontal cortex andthalamus in PLC�4�/� mice. (B) The number of sponta-neous SWDs per hour was counted. (C) A power anal-ysis of spontaneous SWDs in PLC�4�/� mice showed apeak frequency at 7.8 Hz. (D) i.p. injection of ethosux-imide (ETX) significantly reduced the number of SWDs(P � 0.05; paired t test), whereas injection of vehicle(saline) did not.

Cheong et al. PNAS � December 22, 2009 � vol. 106 � no. 51 � 21913

NEU

ROSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Apr

il 15

, 202

1

overall cortical area and PLC�4 mRNA was absent from layer IV(30). We analyzed PLC�4 expression in wild-type mice by immu-nostaining and confirmed that PLC�4 protein was highly expressedin the TC relay nuclei (Fig. 4A1, red signals) and PLC�4 expressionwas low in the cortex (Fig. 4A2). No expression of PLC�4 proteinwas detected in PLC�4�/� mice (Fig. 4B).

Because PLC�4 is expressed in other brain regions, albeit atrelatively low levels compared with the TC relay nuclei (Fig. 4A),our results do not conclusively demonstrate that the absence-seizure phenotype of PLC�4�/� mice results from a deficiency ofPLC�4 in TC neurons and a consequent alteration in the firingpattern of TC neurons. To test the hypothesis that a PLC�4deficiency in the TC relay nuclei is responsible for the generationof the absence-seizure in PLC�4�/� mice, we tested whetherselective knockdown of PLC�4 in the TC relay nuclei wouldgenerate the absence-seizure phenotype. Lentiviral (LV) vectorscarrying short hairpin RNA (shRNA) for PLC�4 were generatedand tested as described in ref. 35. High-titer LV vectors expressingshRNA for PLC�4 (LV-shPLC�4) were injected into the TC relaynuclei to knockdown PLC�4 expression locally; mice injected withpLKO vectors (expressing nonspecific shRNA) were used as acontrol group. PLC�4 immunostaining was substantially reduced inLV-shPLC�4-infected neuronal cells in a wide range of TC relaynuclei (Fig. 4D), whereas neuronal cells infected with control virusshowed normal PLC�4 expression (Fig. 4C). The optimal effectof shPLC�4 on PLC�4 expression was observed 4 weeks aftervirus injection. DAPI staining confirmed that the neurons wereintact (Fig. 4 C and D), indicating that sh-PLC�4 efficientlyreduced endogenous PLC�4 expression without damaging tha-lamic neurons.

EEGs of PLC�4�/� mice injected bilaterally with virus stockswere recorded before and after injecting RS(�/�)-baclofen. Post-hoc staining confirmed the site of viral-vector injection and PLC�4

expression levels. Only EEG data from mice in which thalamicPLC�4 expression was clearly knocked down were used for furtheranalysis. Spontaneous SWDs were observed in 7 of 12 mice injectedwith LV-shPLC�4 (Fig. 5A) and averaged 7.3 � 2.0 SWDs per hour(range, 3–17 SWDs/h, Fig. 5B). In contrast, none of the nine miceinjected with pLKO-control showed any evidence of spontaneousSWD-like patterns on EEGs (Fig. 5A). In addition, administrationof RS(�/�)-baclofen induced a greater duration of SWDs inPLC�4�/� mice infected with LV-shPLC�4 than in those infectedwith pLKO-control (P � 0.005 or P � 0.01, Fig. 5C). These dataindicate that the absence-seizure phenotype of PLC�4�/� mice wasprimarily due to the deletion of thalamic PLC�4.

Intrathalamic Injection of a PKC Activator or a T-Type Ca2� ChannelBlocker, but Not an L-type Ca2� Channel Blocker, Reduces SWDs. Werecently observed that the amplitudes of both T- and L-type Ca2�

currents were increased in PLC�4�/� TC neurons, resulting in amarked change in TC neuron firing pattern (31). In that study, we

Fig. 2. Firing pattern of PLC�4�/� TC neuron is easily shifted to burst firingmode. Representative traces showing the firing pattern of (A) wild-type and (B)PLC�4�/� TC neurons from whole cell patch recordings. Tonic firing was inducedwith depolarizing currents, 400 pA. PLC�4�/� TC neurons were often shifted fromtonic to low threshold burst firing (B), whereas wild-type TC neurons have nevershown such a transition of firing mode (A). The bottom panel displays the appliedcurrent steps. Injection of prepulses which slightly hyperpolarize the membranepotentials elicited low-threshold burst firing in PLC�4�/� TC neurons (D), but notin wild-type TC neurons (C). (E) Spike numbers in a burst induced by variousprepulses hyperpolarizing the membrane potentials to between �73 and �63mV in wild-type (closed circles) and PLC�4�/� (open circles) TC neurons.

Fig. 3. The duration of SWDs induced by GABAB receptor agonists is greaterin PLC�4�/� mice than in wild-type littermate controls. (A) Epidural EEGsrecorded from the frontal cortex region (Cx) and local field potential recordedfrom the ventrobasal region (VB) of the thalamus were collected before andafter administration of GBL. The total durations of SWDs (in seconds) per totalrecording time (in minutes) induced in PLC�4�/� mice (closed circles) andPLC�4�/� mice (open circles) by GBL (B) or RS(�/�)-baclofen (C) were averagedand compared. The total duration of SWDs induced by the GABA drugs usedin this study was greater in PLC�4�/� mice than in wild-type littermate controls(P � 0.01 or P � 0.001). (D and E) Averaged temporal power spectrogramsexhibited a greater power density at low frequency (2–5 Hz) in PLC�4�/� mice.

21914 � www.pnas.org�cgi�doi�10.1073�pnas.0912204106 Cheong et al.

Dow

nloa

ded

by g

uest

on

Apr

il 15

, 202

1

showed that this phenotype was primarily due to down-regulationof PKC activity in TC neurons, without the changes in inputresistance or resting membrane potential, and could be readilyreversed by activating PKC. To examine whether alteration of thefiring pattern of PLC�4�/� TC neurons via down-regulation of thePKC pathway is responsible for the generation of spontaneousSWDs, we focally injected phorbol 12,13-didecanoate (PDD), aPKC agonist, into the TC relay nuclei. PLC�4�/� mice microin-jected with PDD (10 pmol) showed a substantial decrease in thenumber of SWDs compared with baseline recordings (P � 0.05;paired t test), whereas those injected with vehicle (0.9% NaCl)exhibited no change in SWDs (Fig. 6 A and B).

To verify the role of thalamic T- and L-type Ca2� channels in thegeneration of SWDs, we injected mibefradil, a T-type Ca2� channelblocker, or nifedipine, an L-type Ca2� channel blocker, into thesame target area in the thalamus and recorded EEGs. We recentlyshowed that nifedipine increases the tonic firing rate, but does notaffect burst firing in TC neurons (31). The injection of mibefradil(1 nmol) reduced the occurrence of SWDs (P � 0.005), whereasnifedipine (1 nmol) had no effect (Fig. 6B), indicating that T-typeCa2� currents in TC neurons contributed to the generation of theabsence-seizure phenotype in PLC�4�/� mice. These data stronglyindicate that the down-regulation of thalamic PKC and the resultingincrease in overall propensity for burst firing with enhanced T-typeCa2� currents primarily contribute to the genesis of SWDs inPLC�4�/� mice.

DiscussionIn this study, we addressed the question on the role of TC neurons,especially their firing properties, in the genesis of absence seizures.The observation that silencing PLC�4 in TC relay nuclei using

Fig. 4. Microinjected LV-shPLC�4 selectively knocks down thalamic PLC�4. (A)Immunostaining in wild-type mice showed that PLC�4 (red) is highly expressed inTC neurons (A1) and expressed at relatively low levels in cortical neurons (A2). (B)PLC�4 was not detectable in PLC�4�/� mice. (C) The expression of PLC�4 was notaffected by injection of pLKO-control into the thalamic regions. i–iii show en-larged views of the Inset in the Left. (i): immunostaining for PLC�4 (red); (ii): DAPIstaining (blue); (iii): merged image. (D) The expression of PLC�4 was greatlyreduced by injection of LV-shPLC�4 into the thalamic regions (Di and Diii); DAPIstaining shows cells remained intact (Dii). (Scale bar, 50 �m.)

Fig. 5. Deletion of thalamic PLC�4 leads to the genesis of absence seizures. (A)LV vectors containing control shRNA or shPLC�4 construct were injected bilater-ally into wild-type mice and EEGs were recorded from frontal and parietal lobes.Lower: mice injected with LV-shPLC�4 showed sporadic SWDs; Upper: miceinjected with pLKO-control never showed such a high-amplitude paroxysmal EEGpattern. (B) Seven of 12 mice injected with LV-shPLC�4 showed spontaneousSWDs. The number of SWDs varied from 3–17 per hour. (C) The total duration ofSWDs per minute induced by 20 mg/kg RS(�/�)-baclofen was greater in miceinjected with LV-shPLC�4 than in mice injected with pLKO-control.

Fig. 6. Intrathalamic injection of a PKC agonist or a T-type Ca2� channelblocker, but not an L-type Ca2� channel blocker, reduces SWDs. (A) The EEG ofPLC�4�/� mice was recorded before and after bilateral injection of drugs usinga cannula system. (B) The number of SWDs in PLC�4�/� mice was significantlyreduced after the injection of 1 pmol PDD (P � 0.05) or 1 nmol mibefradil (P �0.005), whereas the injection of vehicle or 1 nmol nifedipine did not affect thefrequency of SWDs. All drugs were injected in a total volume of 0.2 �L.

Cheong et al. PNAS � December 22, 2009 � vol. 106 � no. 51 � 21915

NEU

ROSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Apr

il 15

, 202

1

shRNA induced absence-seizure phenotypes strongly supports thehypothesis that the absence-seizure of PLC�4�/� mice is primarilydue to the lack of PLC�4 in TC neurons. SWDs are generalizedover wide cortical areas, requiring a massive synchronization of thethalamo-cortico-thalamic loop. Some studies proposed a leadingrole of the cortex in the generation of SWDs (10–12). In one suchstudy using GAERS, a genetic rat model of absence seizures,infusion of ethosuximide into the cortex produced an immediatecessation of SWDs (13). Others suggested that thalamic synchrony,between nRT and TC neurons, reinforce larger thalamocorticalnetworks into synchronized oscillations (3, 8, 9, 14, 15). Most ofthese studies advocated a central role of nRT neurons in generatingSWDs. The role of TC neurons in the generation of SWDs hasreceived less attention. Here, we showed that deleting PLC�4 in TCrelay nuclei induces an absence-seizure phenotype that includes theappearance of spontaneous SWDs and concomitant behavioralarrests. The frequency of spontaneous SWDs was approximately 8Hz, a value similar to the frequencies of spontaneous absenceseizures reported in other genetic models (21–24, 27). Our resultsthus establish an animal model for absence seizures driven specif-ically by an anomaly in the TC neurons.

Our results strongly support that T-type Ca2� channels in TCneurons play a critical role in the genesis of SWDs in the mice.Low-threshold burst firing in TC neurons driven by T-type Ca2�

currents has long been thought to be critical in the thalamocorticalnetwork oscillations (3). Previous studies suggested that the in-creased T-type Ca2� currents were related to the generation ofabsence seizures based on the observation on mutant animalsexhibiting spontaneous absence epilepsy (25, 26). Our previousstudies showed that �1G T-type Ca2� channel-deficient mice wereresistant to GABAB receptor agonist-induced SWDs (17) and thatan �1G-null mutation abolished the SWDs of �1A�/� and �1Atg/tg

mice and markedly reduced SWDs in �4lh/lh and �2stg/stg mice (27),indicating that �1G T-type Ca2� channels play a critical role in thegenesis of absence seizures. Furthermore, a recent study reportedthat augmentation of T-type Ca2� currents by transgenic overex-pression of �1G T-type Ca2� channels in the whole brain wassufficient to generate spontaneous SWDs (28). However, thealteration of T-currents in these mouse models was not restricted tothe TC relay nucleus but general to other brain regions. Thislimitation bred a controversy over the role of T-type Ca2� channelsin TC neurons in the genesis of SWDs (4). In fact, low-thresholdburst firings in TC neurons have been rarely observed during SWDsin recordings in vivo from rat and cat absence seizure models (18,20, 36). Here, we showed that infusion of mibefradil, a T-type Ca2�

channel blocker, into the TC relay nuclei substantially suppressedthe SWDs in PLC�4�/� mice, indicating that T-type Ca2� channelsin TC neurons play a role in the genesis of SWDs in these mice.Because mibefradil has been reported to block other voltage-gatedCa2� channels including L- and R-type Ca2� channels and sodiumchannels at high concentrations (37), we could not completelyexclude the possibility that the other channels were partly involvedin the mibefradil effect. However, the nifedipine infusion experi-ments supported that the mibefradil effect was at least not byblocking L-type Ca2� channels at the concentration used in thisstudy. Here, we also observed that the SWDs were reduced byinfusion of a PKC agonist into the TC region, which is consistentwith our previous finding that a PKC agonist, but not an inactivestructural analog, reversibly decreases T-type and L-type Ca2�

currents (31). PKC regulation and interaction site on T-type Ca2�

channels have been reported, although it is still controversial if PKCup-regulates or down-regulates T-type Ca2� channels (38). Theseobservations support that T-type Ca2� channels in TC relay nucleiplay a critical role in the generation of absence seizure in mutantmice.

Most of the studies cited above focused on the positive correla-tion between the enhanced peak amplitude of T-type currents andoccurrence of SWDs. In contrast, one of our previous studies

demonstrated that there was no quantitative difference in theseverity of SWDs between absence seizure mouse models withdifferent numbers of the �1G alleles (�1G�/� vs. �1G�/�) althoughtheir TC neurons exhibited 75% vs. 150% of T-currents comparedwith wild-type TC neurons (27). This study suggested that �1GT-type Ca2� currents are critical, but the augmentation of them isnot essential, in the genesis of absence seizures. In the same study,the kinetics of T-type currents, recorded in one of the epilepticanimals, was not substantially different from those recorded inwild-type animals (3). The afore-mentioned two transgenic miceoverexpressing �1G T-type Ca2� channels also showed no changein the kinetics of T-current whereas the peak current densities werenearly doubled (28). In contrast, PLC�4�/� mice showed thealtered kinetics of T-type currents and the peak current density wasincreased by approximately 50% compared with wild-type mice(31). PLC�4�/� TC neurons exhibit an increased propensity forburst firing especially near the resting membrane potential thatwould have resulted from the decreased steady-state inactivation ofT-type channels. Such alterations in the kinetics of T-type currentswould enable mutant TC neurons to easily shift their firing modeto burst firing in response to small excitatory as well inhibitoryinputs, which normally would not cause low-threshold burst firing(Fig. 2). Interestingly, a previous study reported that synaptosomal-associated protein (SNAP-25)-deficient mouse mutant, Coloboma,developed severe absence seizures with elevated thalamic T-typeCa2� currents (26). In this study, they also observed a reducedsteady-state inactivation of T-type Ca2� currents in TC neurons, afinding consistent with our observation in the PLC�4 mutant. It isalso interesting to note that the deletion in the Coloboma mutantmouse includes the genes for PLC�1 and PLC�4. These observa-tions would suggest that the increased propensity for burst firing inTC neurons may contribute critically to the generation of SWDs.Therefore, our current study provides insight into the role of T-typechannels in the genesis of absence seizures.

In this study, we could not clearly distinguish between the effectsof easy generation of burst firing and of the increased number ofspikes in a burst on the generation of SWDs because blockingT-type Ca2� channels would eliminate the burst firing of TCneurons. Burst firing in TC neurons has been implicated in sus-taining physiological and pathological oscillations including sleepspindles and SWDs (3, 8). Therefore, increased number of spike inPLC�4�/� TC neurons may contribute to sustaining the oscillationsin the thalamocortical circuit during SWDs. One could argue thatup-regulation of T-currents is due to an increase in T-type channelexpression via compensatory actions. Our recent study showed thata PKC agonist, but not an inactive structural analog, reversiblydecreases T-type and L-type Ca2� currents within minutes ofapplication (31). The present results also demonstrate that theSWDs were significantly reduced by acute infusion of a PKC agonistinto the TC region. Together, these findings support the notion thatthe up-regulation of T-current was not caused primarily by achanged expression level of T-type channels due to developmentalor compensatory changes, because the expression level could not bereversed within minutes. On the other hand, it could not becompletely excluded that a general reduction of excitation fromcortical inputs due to the lack of PLC�4 in TC neurons mightcontribute to the phenotype observed in the PLC�4�/� mice. Weobserved that an acute infusion of a T-type channel blocker into theTC region reduced SWDs in addition to the distinct change intransition between tonic to burst firing mode in mutant TC neuronsfollowing the given stimuli. These observations laid more weight onthat the changes in TC firing modes would primarily contribute tothe absence seizure phenotype in PLC�4�/� mice.

PLC�4�/� mice showed an increased susceptibility to GABABRagonist-induced SWDs in addition to showing spontaneous SWDs.GABAB receptors in TC neurons induce hyperpolarization ofmembrane potential for a long enough time to de-inactivate T-typeCa2� channels, which would induce low-threshold burst firing.

21916 � www.pnas.org�cgi�doi�10.1073�pnas.0912204106 Cheong et al.

Dow

nloa

ded

by g

uest

on

Apr

il 15

, 202

1

Therefore, longer lasting effect of these drugs in the PLC�4�/�

mice might be due to the possibility that the mutant TC neurons caneasily go to the oscillatory burst firing mode after a slight hyper-polarization of membrane potential induced by a low concentrationof these drugs.

Absence seizures that accompany pathological oscillations in thethalamocortical network interrupt the normal vigilance state. Ourprevious study showed that the thalamic PLC�4 pathway plays apivotal role in controlling the tendency of TC neurons toward eithertonic or burst-firing mode. It is well known that PLC�4 in TCneurons is functionally linked to mGluR1, which receives thecorticothalamic inputs (29). Our current data indicate that down-regulation of the thalamic mGluR1-PLC�4 pathway may lead to anabnormally easy transition from tonic to burst-firing modes in TCneurons, switching the thalamocortical network into pathologicaloscillatory activities expressed as SWDs. Our study highlights theimportance of corticothalamic inputs to TC neurons via themGluR1/PLC�4 pathway in maintenance of the normal vigilancestate.

MethodsAnimals. All experiments used PLC�4�/� and their wild-type littermates in the F1hybrid generated by mating heterozygote mice (PLC�4�/�) from two geneticbackgrounds: 129/sv and C57BL/6J. Animal care and all experiments were con-ducted in accordance with the ethical guidelines of the Institutional Animal Careand Use Committee of the Korea Institute of Science and Technology. For de-tailed information, see SI Methods.

Drugs. GBL, RS(�/�)-baclofen, ethosuximide, mibefradil, nifedipine, and PDDwere purchased from Sigma (SI Methods).

In Vivo LV Vector shRNA Injection. For in vivo tests, high-titer, concentrated LVvectors expressing shPLC�4 or nonspecific shRNA (pLKO-control) were prepared

and bilaterally injected into the TC relay nuclei of approximately 8-week-oldPLC�4�/� mice (SI Methods).

In Vivo Intrathalamic Drug Injection. PLC�4�/� mice were implanted with a26-gauge guide cannula (Plastics One) for microinjection of drug stereotaxicallyplaced into TC relay nuclei (SI Methods).

EEG Electrode Implantation and EEG Recording. An epidural electrode wasimplanted in either the frontal lobe or parietal lobe, and a grounding electrodewas implanted in the occipital region. For thalamic local field recordings, aparylene-coated tungsten electrode was positioned in TC relay nuclei. Postmor-tem histological analyses confirmed the recording region (SI Methods).

Preparation of Brain Slices and Whole-Cell Patch Clamping. Firing patterns of TCneurons were recorded in acutely isolated brain slices from approximately4-week-old mice as described in ref. 31. Signals were amplified with a Multiclamp700-A amplifier (Axon Instruments) and analyzed using pCLAMP 9.2 and Mini-Analysis software (Synaptosoft). For detailed information, see SI Methods.

Tissue Processing and Immunohistochemistry. For histological analysis, brainswere frozen, cut into serial 30-�m-thick coronal sections on a freezing mic-rotome, and collected as described in ref. 39. Brain tissues were incubated with aprimary antibody against PLC�4 (Chemicon) and and then with a Cy-3-conjugated secondary antibody (Amersham). High-resolution images of braintissue were processed by confocal microscopy (SI Methods).

Data Analysis. Student’s t tests were applied to determine statistically significantdifferences between two groups. A P value �0.05 was considered statisticallysignificant. SWDswerequantifiedbycountingthenumberanddurationofSWDsusing Matlab. The average spectrograms were generated by averaging theindividual spectrograms, normalized to their respective baseline standard devi-ation. For detailed information, see SI Methods.

ACKNOWLEDGMENTS. We thank Dr. J. Choi for her valuable discussion andhelpful advice on data analysis and S. Han for technical assistance. This work wassupported by the 21C Frontier Proteomics Program of the Ministry of Education,Science and Technology, Korea, National Honor Scientist Program of Korea, andCenter of Excellence Program of the Korea Institute of Science and Technology.

1. McCormick DA, Contreras D (2001) On the cellular and network bases of epilepticseizures. Ann Rev Physiol 63:815–846.

2. Snead OC, 3rd (1995) Basic mechanisms of generalized absence seizures. Ann Neurol37:146–157.

3. Beenhakker MP, Huguenard JR (2009) Neurons that fire together also conspire to-gether: Is normal sleep circuitry hijacked to generate epilepsy? Neuron 62:612–632.

4. Crunelli V, Leresche N (2002) Childhood absence epilepsy: Genes, channels, neuronsand networks. Nat Rev 3:371–382.

5. Williams D (1953) A study of thalamic and cortical rhythms in petit mal. Brain 76:50–69.6. Danober L, Deransart C, Depaulis A, Vergnes M, Marescaux C (1998) Pathophysiological

mechanisms of genetic absence epilepsy in the rat. Prog Neurobiol 55:27–57.7. Timofeev I, Steriade M (2004) Neocortical seizures: Initiation, development and ces-

sation. Neuroscience 123:299–336.8. Huguenard JR, McCormick DA (2007) Thalamic synchrony and dynamic regulation of

global forebrain oscillations. Trends Neurosci 30:350–356.9. Huntsman MM, Porcello DM, Homanics GE, DeLorey TM, Huguenard JR (1999) Recip-

rocal inhibitory connections and network synchrony in the mammalian thalamus.Science 283:541–543.

10. Meeren HK, Pijn JP, Van Luijtelaar EL, Coenen AM, Lopes da Silva FH (2002) Corticalfocus drives widespread corticothalamic networks during spontaneous absence sei-zures in rats. J Neurosci 22:1480–1495.

11. Tan HO, et al. (2007) Reduced cortical inhibition in a mouse model of familial childhoodabsence epilepsy. Proc Natl Acad Sci USA 104:17536–17541.

12. Polack PO, et al. (2007) Deep layer somatosensory cortical neurons initiate spike-and-wave discharges in a genetic model of absence seizures. J Neurosci 27:6590–6599.

13. Manning JP, Richards DA, Leresche N, Crunelli V, Bowery NG (2004) Cortical-areaspecific block of genetically determined absence seizures by ethosuximide. Neuro-science 123:5–9.

14. McCormick DA, Bal T (1997) Sleep and arousal: Thalamocortical mechanisms. Annu RevNeurosci 20:185–215.

15. Avanzini G, Panzica F, de Curtis M (2000) The role of the thalamus in vigilance andepileptogenic mechanisms. Clin Neurophysiol 111:S19–S26.

16. Llinas RR, Steriade M (2006) Bursting of thalamic neurons and states of vigilance.J Neurophysiol 95:3297–3308.

17. Kim D, et al. (2001) Lack of the burst firing of thalamocortical relay neurons andresistance to absence seizures in mice lacking alpha(1G) T-type Ca(2�) channels.Neuron 31:35–45.

18. Steriade M, Contreras D (1995) Relations between cortical and thalamic cellular eventsduring transition from sleep patterns to paroxysmal activity. J Neurosci 15:623–642.

19. Noebels JL, Sidman RL (1979) Inherited epilepsy: Spike-wave and focal motor seizuresin the mutant mouse tottering. Science 204:1334–1336.

20. Pinault D, et al. (1998) Intracellular recordings in thalamic neurones during spon-taneous spike and wave discharges in rats with absence epilepsy. J Physiol 509:449 –456.

21. Frankel WN (1999) Detecting genes in new and old mouse models for epilepsy: Aprospectus through the magnifying glass. Epilepsy Res 36:97–110.

22. Burgess DL, Jones JM, Meisler MH, Noebels JL (1997) Mutation of the Ca2� channel betasubunit gene Cchb4 is associated with ataxia and seizures in the lethargic (lh) mouse.Cell 88:385–392.

23. Barclay J, Rees M (1999) Mouse models of spike-wave epilepsy. Epilepsia 40:17–22.24. Noebels JL, Qiao X, Bronson RT, Spencer C, Davisson MT (1990) Stargazer: A new

neurological mutant on chromosome 15 in the mouse with prolonged cortical seizures.Epilepsy Res 7:129–135.

25. Zhang Y, Mori M, Burgess DL, Noebels JL (2002) Mutations in high-voltage-activatedcalcium channel genes stimulate low-voltage-activated currents in mouse thalamicrelay neurons. J Neurosci 22:6362–6371.

26. Zhang Y, Vilaythong AP, Yoshor D, Noebels JL (2004) Elevated thalamic low-voltage-activated currents precede the onset of absence epilepsy in the SNAP25-deficientmouse mutant coloboma. J Neurosci 24:5239–5248.

27. Song I, et al. (2004) Role of the alpha1G T-type calcium channel in spontaneous absenceseizures in mutant mice. J Neurosci 24:5249–5257.

28. Ernst WL, Zhang Y, Yoo JW, Ernst SJ, Noebels JL (2009) Genetic enhancement ofthalamocortical network activity by elevating alpha1g-mediated low-voltage-activated calcium current induces pure absence epilepsy. J Neurosci 29:1615–1625.

29. Miyata M, et al. (2003) Role of thalamic phospholipase C�4 mediated by metabotropicglutamate receptor type 1 in inflammatory pain. J Neurosci 23:8098–8108.

30. Watanabe M, et al. (1998) Patterns of expression for the mRNA corresponding to thefour isoforms of phospholipase Cbeta in mouse brain. Eur J Neurosci 10:2016–2025.

31. Cheong E, et al. (2008) Tuning thalamic firing modes via simultaneous modulation ofT- and L-type Ca2� channels controls pain sensory gating in the thalamus. J Neurosci28:13331–13340.

32. Hu RQ, Banerjee PK, Snead OC, 3rd (2000) Regulation of gamma-aminobutyric acid(GABA) release in cerebral cortex in the gamma-hydroxybutyric acid (GHB) model ofabsence seizures in rat. Neuropharmacology 39:427–439.

33. Wong CG, Gibson KM, Snead OC, 3rd (2004) From the street to the brain: Neurobiologyof the recreational drug gamma-hydroxybutyric acid. Trends Pharmacol Sci 25:29–34.

34. Bowery NG (2006) GABAB receptor: A site of therapeutic benefit. Curr Opin Pharmacol6:37–43.

35. Shin J, et al. (2009) Phospholipase C �4 in the medial septum controls cholinergic thetaoscillations and anxiety behaviors. J Neurosci In Press

36. Charpier S, et al. (1999) On the putative contribution of GABA(B) receptors to theelectrical events occurring during spontaneous spike and wave discharges. Neurophar-macology 38:1699–1706.

37. Strege PR, Bernard CE, Ou Y, Gibbons SJ, Farrugia G (2005) Effect of mibefradil onsodium and calcium currents. Am J Physiol Gastrointest Liver Physiol 289:G249–G253.

38. Iftinca MC, Zamponi GW (2009) Regulation of neuronal T-type calcium channels.Trends Pharmacol Sci 30:32–40.

39. Kang SJ, et al. (2008) Expression of Kir2.1 channels in astrocytes under pathophysio-logical conditions. Mol Cells 25:124–130.

Cheong et al. PNAS � December 22, 2009 � vol. 106 � no. 51 � 21917

NEU

ROSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Apr

il 15

, 202

1