Epilepsia, 52(12):2315–2325, 2011 doi: 10.1111/j.1528-1167.2011.03273.x
FULL-LENGTH ORIGINAL RESEARCH
Inflammatory changes during epileptogenesis and spontaneous seizures in a mouse model of mesiotemporal lobe epilepsy *1Fabien Pernot, y1Christophe Heinrich, *Laure Barbier, zAndre´ Peinnequin, *Pierre Carpentier, *Franck Dhote, *Vale´rie Baille, *Claire Beaup, xAntoine Depaulis, and *{Fre´de´ric Dorandeu *Department of Toxicology and Chemical Risks, Institut de Recherche Biome´dicale des Arme´es, Centre de recherches du service de sante´ des arme´es, La Tronche Cedex, France; yDepartment of Physiological Genomics, Institute of Physiology, Ludwig-Maximilians Universita¨t, Schillerstrasse, Mu¨nchen, Germany; zDepartment of Biological Effects of Radiation, Institut de Recherche Biome´dicale des Arme´es, Centre de recherches du service de sante´ des arme´es, La Tronche Cedex, France; xGrenoble Institute of Neurosciences, Inserm U836-UJF-CEA, Chemin Fortune´ Ferrini, Universite´ Joseph Fourier, Site Sante´, La Tronche Cedex, France; and {Val-de-Grace School of Military Medicine, 1 place Alphonse Laveran, Paris, France
SUMMARY Purpose: Neuroinflammation appears as a prominent feature of the mesiotemporal lobe epilepsy syndrome (MTLE) that is observed in human patients and animal models. However, the precise temporal relationship of its development during epileptogenesis remains to be determined. The aim of the present study was to investigate (1) the time course and spatial distribution of neuronal death associated with seizure development, (2) the time course of microglia and astrocyte activation, and (3) the kinetics of induction of mRNAs from neuroinflammatoryrelated proteins during the emergence of recurrent seizures. Methods: Experimental MTLE was induced by the unilateral intrahippocampal injection of kainate in C57BL/6 adult mice. Microglial and astrocytic changes in both ipsilateral and contralateral hippocampi were examined by respectively analyzing griffonia simplicifolia (GSA) lectin staining and glial fibrillary acidic protein (GFAP) immunoreactivity. Changes in mRNA levels of selected genes of cytokine and cytokine regulatory proteins
The mesiotemporal lobe epilepsy (MTLE) syndrome, one of the most common forms of focal epilepsies, has been suggested to result from an initial precipitating event (e.g., febrile seizures, cerebral infection, traumatic brain
Accepted August 10, 2011; Early View publication September 28, 2011. Address correspondence to Fabien Pernot, PharmD, Ph.D., Rhenovia Pharma SAS, 20c rue de Chemnitz, 68100 Mulhouse, France. E-mail: fabien.
[email protected] 1 Both authors equally contributed to this work. Wiley Periodicals, Inc. ª 2011 International League Against Epilepsy
(interleukin-1b, IL-1b; interleukin-1 receptor antagonist, IL-1Ra; suppressor of cytokine signaling 3, SOCS3) and enzymes of the eicosanoid pathway (group IVA cytosolic phospholipase A2, cPLA2-a; cycloxygenase-2, COX-2) were studied by reverse transcription-quantitative real time polymerase chain reaction. Key Findings: Our data show an immediate cell death occurring in the kainate-injected hippocampus during the initial status epilepticus (SE). A rapid increase of activated lectin-positive cells and GFAP-immunoreactivity was subsequently detected in the ipsilateral hippocampus. In the same structure, Il-1b, IL-1Ra, and COX-2 mRNA were specifically increased during SE and epileptogenesis with a different time course. Conversely, the expression of SOCS3 mRNA, a surrogate marker of interleukin signaling, was mainly increased in the contralateral hippocampus after SE. Significance: Our data show that specific neuroinflammatory pathways are activated in a time- and structuredependent manner with putative distinct roles in epileptogenesis. KEY WORDS: Microglia, Astrocytes, Cytokines, Temporal lobe epilepsy, Mouse.
injury, or stroke) (Engel, 2001; Cendes et al., 2002), followed by a ‘‘silent’’ period, preceding the occurrence of recurrent seizures, several years later. This period of latency appears to be associated with different cascades of molecular and cellular processes, leading to the development of seizure-generating neuronal circuits, a process called epileptogenesis. Among the different mechanisms that may be involved in epileptogenesis, neuroinflammation appears as one of the most critical (de Lanerolle & Lee, 2005; Vezzani & Granata, 2005), as suggested in animal models (Vezzani & Baram, 2007) as well as in clinical reports (Ravizza et al., 2008).
2315
2316 F. Pernot et al. In the brain, inflammation is characterized by glial activation, edema, and synthesis of inflammatory mediators such as cytokines or enzymes of the prostaglandin pathway. Experimental evidence has demonstrated a rapid-onset inflammatory response to acute seizures in various models that largely implicates interleukin-1b (IL-1b), among others, as a prototypic inflammatory signal (Eriksson et al., 1999; Vezzani et al., 1999; De Simoni et al., 2000; PlataSalaman et al., 2000; Turrin & Rivest, 2004; Gorter et al., 2006; Ravizza et al., 2006; Dhote et al., 2007). Some reports have demonstrated that IL-1b alters neuronal excitability and may contribute to the pathophysiologic process of epilepsy (Vezzani et al., 2000; Vezzani & Baram, 2007). Unfortunately, in the brain, IL-1 receptor antagonist (IL-1Ra), which acts by limiting IL-1b-mediated actions, is generally produced much later than IL-1b, in contrast with peripheral inflammatory reactions (Dinarello, 1996; Vezzani et al., 2008). Some other regulatory mechanisms have evolved to limit the potentially harmful consequences of cytokine signaling. A major process implicates suppressors of cytokine signaling (SOCS), a family of proteins whose expression is induced by cytokines and that, in turn, negatively regulate signaling pathways used by many cytokines, thereby modulating a wide range of inflammatory processes and considered as key regulators of inflammation (Lang et al., 2003; Yoshimura et al., 2007; Croker et al., 2008). More precisely, SOCS3 is induced by a wide range of stimuli including interleukin-6 (IL-6) family and regulates the cytokine signaling via the inhibition of the Janus kinases/ signal transducers and activators of transcription (JAK/ STAT) transduction pathway in a negative feedback loop (Heinrich et al., 1998; Schmitz et al., 2000; Yasukawa et al., 2000). Although to date studies of the pattern of activation of SOCS family proteins in rodent epilepsy models remains to be specified, it was demonstrated that SOCS3 can be highly regulated by neuronal and nonneuronal cells during seizure activity, with a potential implication during the remodeling process of epileptogenesis (Rosell et al., 2003). Interestingly, it was demonstrated that some of the seizure-induced proinflammatory signals remain up-regulated during epileptogenesis (Voutsinos-Porche et al., 2004; Lee et al., 2007; Ravizza et al., 2008; Maroso et al., 2010) and are likely to play a major role in the establishment of recurrent epileptic seizures (Ravizza et al., 2011). In the present study, we aimed at further characterizing the precise and spatial temporal patterns of neuroinflammatory events that are associated with MTLE development. We thus used intrahippocampal kainate (KA) in mice, which is characterized by unilateral histologic changes reminiscent of hippocampal sclerosis often observed in human MTLE (Suzuki et al., 1995). In these mice, a single unilateral injection of KA into the hippocampus first induces status epilepticus (SE) lasting for up to 15 h (Riban et al., 2002). Then, during the next 2 weeks, recurrent spontaneous ipsilateral hippocampal discharges progressively Epilepsia, 52(12):2315–2325, 2011 doi: 10.1111/j.1528-1167.2011.03273.x
develop to become stable for the lifetime of the animals (Heinrich et al., 2011). These recurrent seizures are largely confined to the ipsilateral, sclerotic hippocampus, but can also be detected in the contralateral hippocampus (Bouilleret et al., 1999; Riban et al., 2002; Meier et al., 2007). Here, we investigated and compared in both hippocampi (1) the time course and spatial distribution of neuronal death associated with seizure development, (2) the timing and regional activation of astrocytes and microglia, and (3) the spatiotemporal pattern of expression of the mRNAs of neuroinflammation markers, such as IL-1b, IL-1Ra, and SOCS3, as well as two enzymes of the eicosanoids pathway: group IVA cytosolic phospholipase A2 form (cPLA2-a) and cycloxygenase-2 (COX-2).
Materials and Methods Intrahippocampal injection of kainate C57BL/6 male mice of 8–10 weeks of age were anesthetized with chloral hydrate (400 mg/kg, i.p.) and placed in a stereotaxic frame. A solution of 50 nl of KA (20 nM) in physiologic saline was injected in the dorsal hippocampus (see Text S1). Upon recovery from anesthesia, animals injected with KA displayed mild clonic movements of the forelimbs, rotations, and immobility that lasted for up to 15 h as described before (Riban et al., 2002). Because 100% of mice that experienced SE develop spontaneous recurrent seizures (Heinrich et al., 2006, 2011), only data from animals with a characteristic SE were included in the present study. Video–electroencephalography recordings Within 2 h after the injection of KA, six mice implanted with hippocampal and cortical electrodes were placed in acrylic glass test cages and were connected to a digital video–electroencephalography (EEG) recording device (Coherence, Deltamed, France; sampling rate = 256 Hz). Continuous EEG and video acquisition was then performed for 24 h, under red light conditions during the dark period (12/12 h, light on at 7:00 a.m.), while the animals were freely moving. Immunohistochemistry and histochemistry Mice were deeply anesthetized with an overdose of pentobarbital (80 mg/kg, i.p.) at set times after intrahippocampal microinjection (2 and 24 h, and 3, 7, and 21 days). Tissue sections of each injected brain were analyzed after hemalun-phloxin (H&P) staining (Lillie & Fullmer, 1976) using a standard protocol (Baille et al., 2005) in order to verify (1) the correct location of the injection site, (2) the right position of the bipolar electrode, and (3) the occurrence of neuronal death in the dorsal hippocampus. Detection of DNA fragmentation was performed using the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) technique described previously
2317 Neuroinflammation in Temporal Lobe Epilepsy (Gavrieli et al., 1992). Microglial cells were detected with peroxidase-labeled isolectin B4 according to a previously described method (Streit, 1990). Astrocytes were immunostained using an antibody directed against glial fibrillary acidic protein (GFAP). RNA quantification of selected neuroinflammatory markers Saline and KA-injected mice were killed by decapitation at set time points, that is, 5 h, and 1, 2, 7, and 21 days after hippocampal injection (n = 5 at each time and per condition). Because of the massive changes in the hippocampal structure of KA-injected mice (see Results), the number of animals sampled at key time points, that is, SE (5 h), epileptogenesis (7 days), and recurrence of seizures (21 days) was increased to 11 animals. Six naive animals were also sampled at 5 h (n = 6) (see Text S1 and Table S1).
Results Electroencephalographic recordings of the status epilepticus following intrahippocampal kainate injection It was previously shown that a single unilateral injection of KA into the hippocampus of adult mice induces an initial SE, which is followed by a latent period of 2–3 weeks before the occurrence of spontaneous discharges largely confined to the injected hippocampus (Riban et al., 2002; Meier et al., 2007; Heinrich et al., 2011). To further determine the relation between EEG paroxysmal activities and the neuroinflammatory changes, we first analyzed the EEG activities occurring in both hippocampi during the initial SE. To this aim, EEG activity was continuously recorded in KA-injected mice (n = 6) in both injected and contralateral hippocampi, as well as in the ipsilateral and contralateral frontoparietal cortices, starting upon awakening from
A
anesthesia (90–110 min after injection) and lasting for 24 h. In addition, animals were recorded at 21 days post-KA injection during the chronic phase of the disease when recurrent seizures can be observed (Riban et al., 2002; Heinrich et al., 2006). For the entire duration of the recording, lower amplitude background activity was observed in the injected hippocampus as compared to the contralateral one. This was evidenced by a peak-to-peak amplitude analysis of the signal, representing on the ipsilateral side 35.8 € 10.9% of the contralateral amplitude, throughout the 24 h recording (n = 6; Fig. 1A). As soon as the EEG recording was started, isolated spikes and bursts of spikes were observed regularly in both hippocampi and often in cortices at a frequency between 0.5 and 4 Hz. These spikes were observed 15–18 h post-KA and had higher amplitude in the contralateral hippocampus. Discharges of spikes and polyspikes were later observed (i.e., 3–5 h postinjection) and for up to 18 h, with a variable recurrence between animals (1–5 per hour; n = 6; Fig. 1A). They first occurred concomitantly in both hippocampi with higher amplitude in the contralateral side, and then they generally spread to the cortices (Fig. 1A). These discharges lasted about 20–40 s, during which time the animals remained immobile. In contrast, a few discharges of spikes, polyspikes, and spike-and-waves were observed in four animals, occurring concomitantly in both hippocampi and cortices, preceded by several sharp waves and lasting for up to 60 s (Fig. 1B). These discharges were associated with asymmetric clonic seizures of the forelimbs and, in some cases, rearing of the animals. Fully generalized tonic– clonic seizures were rarely observed. When recorded 21 days later, all mice displayed recurrent discharges in the injected hippocampus, as described previously (Riban et al., 2002; Heinrich et al., 2011). In agreement with previous studies (Riban et al., 2002; Meier et al., 2007), these data showed that early SE after KA injection occurred in both the ipsilateral
B
Figure 1. Electroencephalography (EEG) recordings during nonconvulsive status epilepticus (SE) after intrahippocampal kainate (KA) injection in mice. Panels A and B depict typical bilateral EEG recordings in the ipsilateral and contralateral hippocampi and cortices during nonconvulsive SE at 6 h (A, n = 6) and 12 h (B, n = 6) following the unilateral intrahippocampal injection of KA. (A) Example of a nonconvulsive discharge. (B) Example of a discharge associated with asymmetrical clonus of the forelimbs. HIP, hippocampus; CX, cortex. Calibration: 2 s, 500 lV. Epilepsia ILAE Epilepsia, 52(12):2315–2325, 2011 doi: 10.1111/j.1528-1167.2011.03273.x
2318 F. Pernot et al.
B
A
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
Figure 2. Neuronal damage after intrahippocampal kainate (KA) injection in mice. Representative photomicrographs of ipsilateral (A) and contralateral (B) hippocampal sections of mice EEG-recorded (Fig. 1) that were sacrificed at 21 days post-KA injection and stained with H&P. A significant cell loss in the CA1, CA3, and hilus areas was observed in the injected hippocampus (A, n = 6), In contrast, no ultrastructural changes were detected in the contralateral hippocampus (B, n = 6). Panels C–R depict representative photomicrographs showing neuronal damage at various time points following intrahippocampal KA injection. Neuronal damage was assessed by TUNEL technique (C–F) and H&P staining (G–R). At 2 h post KA, TUNEL positivity was observed in the injected hippocampus in the CA1 and hilus areas (C, n = 3). Higher power image resolution revealed that TUNEL-positive cells were localized among pyramidal neurons in CA1 (D). No neuronal damage was observed in the contralateral hippocampus (E) and CA1 area, even at a higher magnification (F). Note the increase in acidophilic cells in the ipsilateral CA1 area during epileptogenesis at 3 days (G, H; n = 5) and 7 days (K, L; n = 5). In ipsilateral hippocampus and during the chronic phase (21 days post-KA, n = 5 O, P), condensed cells were absent, suggesting cell loss. Note the massive dispersion of granule cell layer at this time point. Contralateral hippocampus displayed a normal histology of the hippocampal formation whatever the time point (I, M, Q), even at a higher magnification (J, N, R). Scale bars: 100 lm. Epilepsia ILAE Epilepsia, 52(12):2315–2325, 2011 doi: 10.1111/j.1528-1167.2011.03273.x
2319 Neuroinflammation in Temporal Lobe Epilepsy and contralateral hippocampi (
