Transgenic Overexpression of 14-3-3 Zeta Protects Hippocampus against Endoplasmic Reticulum Stress and Status Epilepticus In Vivo

14-3-3 proteins are ubiquitous molecular chaperones that are abundantly expressed in the brain where they regulate cell functions including metabolism, the cell cycle and apoptosis. Brain levels of several 14-3-3 isoforms are altered in diseases of the nervous system, including epilepsy. The 14-3-3 zeta (ζ) isoform has been linked to endoplasmic reticulum (ER) function in neurons, with reduced levels provoking ER stress and increasing vulnerability to excitotoxic injury. Here we report that transgenic overexpression of 14-3-3ζ in mice results in selective changes to the unfolded protein response pathway in the hippocampus, including down-regulation of glucose-regulated proteins 78 and 94, activating transcription factors 4 and 6, and Xbp1 splicing. No differences were found between wild-type mice and transgenic mice for levels of other 14-3-3 isoforms or various other 14-3-3 binding proteins. 14-3-3ζ overexpressing mice were potently protected against cell death caused by intracerebroventricular injection of the ER stressor tunicamycin. 14-3-3ζ overexpressing mice were also potently protected against neuronal death caused by prolonged seizures. These studies demonstrate that increased 14-3-3ζ levels protect against ER stress and seizure-damage despite down-regulation of the unfolded protein response. Delivery of 14-3-3ζ may protect against pathologic changes resulting from prolonged or repeated seizures or where injuries provoke ER stress.


Introduction
14-3-3 proteins are a ubiquitous family of molecular chaperones of which seven isoforms are known in mammals (b, e, c, g, f, h, and s). 14-3-3 proteins regulate cell proliferation, differentiation, metabolism and apoptosis [1]. 14-3-3 proteins are present within synapses and are important for the function and localization of ion channels [2]. 14-3-3 proteins can also function as sweepers of misfolded proteins [3] and promote protein trafficking from the endoplasmic reticulum (ER) [4]. Genetic deletion studies have demonstrated non-redundant roles for certain isoforms in brain development and function [5,6] while aberrant expression of 14-3-3 proteins has been implicated in several diseases of the nervous system [2].
Prolonged seizures (status epilepticus) or repeated seizures over time (pharmacoresistant epilepsy) can damage the brain [7]. Excitotoxicity is a key mechanism, whereby prolonged overactivation of glutamate receptors results in loss of intracellular calcium homeostasis, oxidative stress, damage to intracellular organelles, and necrosis [8,9]. Seizures also trigger release of apoptogenic proteins from mitochondria and downstream caspasedependent and -independent neuronal death [10,11]. 14-3-3 proteins may be important upstream regulators of apoptosisassociated signalling after seizures. 14-3-3 proteins dissociate from pro-apoptotic proteins such as Bad and apoptosis signal-regulating kinase 1 (ASK-1) after experimental status epilepticus, promoting neuronal death [12,13,14]. While some 14-3-3 isoforms are downregulated after seizures [12,15], levels of the zeta (f) isoform are increased [15,16,17]. This may be neuroprotective since depleting 14-3-3f levels in vitro exacerbates kainic acid excitotoxicity [17]. Reduced 14-3-3f expression was reported during epilepsy development [18] and 14-3-3f is also involved in the function of tuberin, mutations in which result in neurological phenotypes including seizures [19]. Although 14-3-3 proteins are mainly cytosolic they are also found in the ER-containing microsomal fraction [15,20]. ER functions include regulating protein folding and trafficking and intracellular calcium storage [21]. Cell stress can result in the three-branched unfolded protein response (UPR) [22]. Ire1, an endoribonuclease, cleaves the X-box binding protein 1 (Xbp1) transcript resulting in upregulation of molecular chaperones such as Bip (glucose-regulated protein 78/GRP78); cleavage of activating transcription factor 6 (ATF6) leads to increased ATF4 levels and modulators of ER stress; activation of protein kinase RNA (PKR)-like ER kinase (PERK) which phosphorylates eukaryotic initiation factor 2a (eIF2a) leading to a shut-down in protein translation. If ER stress persists, the ATF6 and PERK branches trigger apoptosis through up-regulation of CHOP and activation of caspases [23]. ER stress may be an important pathophysiological component in experimental and human temporal lobe epilepsy (TLE) [24,25,26,27,28]. Moreover, inhibition of ER stress can protect against seizure-induced neuronal death [25,29]. Whether 14-3-3f can protect the hippocampus against either ER stress or seizure-induced neuronal death in vivo is unknown. To test this idea we studied ER stress and the response to seizures in transgenic mice over-expressing 14-3-3f.

14-3-3f Mice
Generation of 14-3-3f-overexpressing mice (hereafter referred to as 14-3-3ftg) has previously been reported [30]. The SJL mice express myc-tagged mouse 14-3-3f under the control of the ubiquitous elongation factor 1a (EF1a) promoter. Transgene expression was confirmed by Western blot analysis of the myc tag in protein lysates from tail snips, as described [30]. Heterozygous males and wild-type females were bred together to obtain heterozygous 14-3-3ftg and wild-type littermate controls. Mouse body and brain weight was recorded in 6 week old animals.

Seizure Model
All mouse experiments were performed in accordance with the European Communities Council Directive (86/609/EEC) and were reviewed and approved by the Research Ethics Committee of The Royal College of Surgeons in Ireland (REC#205) under license from the Department of Health, Dublin, Ireland. Food and water was available to mice ad libitum. Induction of status epilepticus in SJL wild-type and 14-3-3ftg mice was performed as previously described [31]. Briefly, mice were anaesthetised with isoflurane and placed in a mouse adapted stereotaxic frame. After making a mid-line scalp incision, three partial craniectomies were performed for placement of skull electrodes. A fourth full craniectomy was drilled for the placement of a guide cannula (Coordinates from Bregma: anterior-posterior (AP) = 20.3 mm and laterally (L) = 20.3 mm) [32]. The cannula and electrodes were fixed in place and mice allowed to recover before being placed in a Perspex container. The EEG was recorded using a Grass Comet Digital EEG. After baseline EEG was established, an injection cannula was lowered through the guide cannula 3.75 mm below the brain surface for the injection of kainic acid (Sigma-Aldrich) into the basolateral amygdaloid nucleus. After 40 min all mice received an i.p. injection of lorazepam (6 mg/kg) to curtail seizures and reduce mortality and morbidity. For i.c.v injections, mice were fitted with a cannula as described [33] and received 1 ml injection of 50 mM tunicamycin (Sigma-Aldrich).
EEG was analysed using Grass software with additional frequency and amplitude analyses of EEG data performed by uploading the data to an automated EEG analysis programme (LabChart pro v7 software, ADInstruments Ltd) [33]. Seizures were defined as the duration of high amplitude (.26baseline) high frequency (.5 Hz) discharges and calculated between the time of KA injection and lorazepam [33].

Hippocampal Primary Neuron Culture
Hippocampal primary neurons were cultured as previously described [34]. Briefly, neurons from wt or 14-3-3ftg E18 embryos were cultured for 6 days and then exposed to 3 mM KA for 24 h followed by assessment of cell death by propidium iodide staining. Cell death was expressed as a % of total cells.

Subcellular Fractionation
Hippocampi were fractionated to obtain the cytoplasm, mitochondria, nuclear and microsome-enriched fractions, according to previous methods [15]. Briefly, samples were homogenized in a mannitol/sucrose buffer containing a protease inhibitor cocktail and then centrifuged twice at 12006g for 10 min. The post-nuclear supernatant was then centrifuged twice at 10 0006g for 15 min and the resulting mitochondrial pellet was resuspended in a sucrose buffer and purified through a percoll bilayer by centrifugation at 41 0006g for 30 min. The crude cytosolic fraction was then centrifuged at 100 0006g for 1 h to separate the microsomal and cytosolic fractions. Following fractionation, protein samples (20 mg) were boiled in gel-loading buffer and then separated by 12% SDS-PAGE. Proteins were transferred to nitrocellulose membranes and then incubated with appropriate antibodies. Membranes were then incubated with appropriate secondary antibodies (1:2000 dilution) followed by chemiluminescence detection on a FujiFilm Las4000 imager.

Histopathology and Immunohistochemistry
Fresh-frozen brains were processed on a cryostat in the coronal plane and 12 mm sections collected at the level of dorsal and ventral hippocampus [31]. Sections were nissl-stained using standard techniques [36]. DNA fragmentation was analyzed using terminal deoxynucleotidyl dUTP nick end labeling (TUNEL) (DeadEnd Fluorometric TUNEL system, Promega) as described [34]. Irreversible neuronal injury was assessed using Fluoro-Jade B (FJB) staining [34]. Briefly, sections were post-fixed, incubated in 0.006% potassium permanganate, rinsed and transferred to 0.001% FJB solution (Chemicon Europe Ltd., Chandlers Ford, UK). Sections were then rinsed, dried, cleared and mounted in DPX (Sigma-Aldrich). For immunohistochemistry, sections were post-fixed, permeabilized, blocked in 5% goat serum and incubated overnight with antibodies against; NeuN (1:400) and KDEL (1:500) (Millipore). Sections were washed and incubated with secondary antibodies conjugated with AlexaFluor 488 for NeuN or AlexaFluor 568 for KDEL (Biosciences Ltd). Sections were mounted with medium containing DAPI (Vector Laboratories Ltd) and examined using a Hamamatsu Orca 285 camera attached to a Nikon 2000s epifluorescent microscope. Counts were performed under 206 magnification lens with a counting window area of 725 mm6550 mm and were the average of those from two adjacent sections. For 14-3-3 immunohistochemistry, free floating coronal sections were stored in a 24 well plate in an anti-freeze solution containing ethylene glycol. Sections were treated with 1% H 2 O 2 to deactivate peroxidises and then blocked with 10% BSA/FBS solution for 90 min. The following antibodies were then applied to sections overnight at 4C; 14-3-3 f, c (1:1,000, Santa Cruz Biotechnology), ZnT3 (1:100, Synaptic Systems). Sections were then incubated with biotinylated antibodies with horse/donkey serum for 90 min before being treated with Avidin ABC peroxidise complex for 1 h. Sections were washed then incubated with DAB, washed again, mounted and coverslipped. For immunofluorescence microscopy, free-floating sections were stained with anti-myc and then either NeuN or GFAP followed by either AlexaFluor 488 or AlexaFluor 568-conjugated secondary antibodies.

Data Analysis
All data are presented as mean 6 standard error of the mean. Gel densitometry was undertaken using gel-scanning integrated optical density software (AlphaEaseFC v4.0). Two group comparisons were made using Student's t test (GraphPad Instat). Significance was accepted at p,0.05.

Distribution of myc-14-3-3f Transgene in the Mouse Brain
Heterozygous transgenic mice overexpressing myc-tagged mouse 14-3-3f under the EF-1a promoter on a SJL background were bred and genotyped as before [30]. Mice were born at expected rates and developed normally. EF-1a is constitutively expressed in the brain and we began by characterizing the distribution of the 14-3-3f transgene in different brain regions, by western blot detection of the myc tag ( Figure 1A). Myc-14-3-3f was present in all major subfields of the mouse hippocampus, at its expected molecular weight of 34 kD (representing the 28 kD 14-3-3f protein plus the 6 kD myc tag). Myc-14-3-3f was also detectable in the neocortex, cerebellum, striatum and brain stem ( Figure 1A). To support these data we stained tissue sections from wild-type and 14-3-3ftg mice using antibodies against 14-3-3f ( Figure 1B). Endogenous 14-3-3f was detected mainly in neurons in the CA pyramidal layers and in granule neurons in the dentate gyrus, in a somal distribution ( Figure 1B). Higher immunoreactivity for 14-3-3f was evident in all hippocampal subfields in 14-3-3ftg mice ( Figure 1B). We did not detect significant 14-3-3f immunoreactivity in glia. Immunoreactivity for 14-3-3c was not different between genotypes ( Figure 1B, far right panels). Double immunofluorescence staining of hippocampal sections for myc and either NeuN or glial fibrillary acidic protein (GFAP) confirmed the transgene was primarily expressed in neurons ( Figure 1C).
We next examined subcellular fractions from the hippocampus of wild-type and 14-3-3ftg mice to determine if the myc-14-3-3f distributed normally within cells. Immunoblotting using specific markers confirmed purified nuclear, mitochondrial, cytosolic and ER-containing microsome fractions ( Figure 1D). The b, e and g isoforms were present exclusively in the cytosolic fraction of the mouse hippocampus ( Figure 1E). In contrast, 14-3-3f and 14-3-3c were found in all cellular compartments ( Figure 1E). The myctagged 14-3-3f was also present in each subcellular fraction ( Figure 1E).

Normal Hippocampal Morphology in 14-3-3ftg Mice
We next examined hippocampal morphology in 14-3-3ftg mice. Nissl-stained sections of 14-3-3ftg mice were indistinguishable from wild-type mice ( Figure 2A). Staining for the zinc transporter protein ZnT3 confirmed 14-3-3ftg mice have a normal distribution of mossy fibers ( Figure 2B). Mouse body weight was normal although brain weight was slightly lower ( Figure 2C, D). Immunostaining and western blotting for the mature neuron marker NeuN determined hippocampal neuron distribution and counts were normal in the 14-3-3ftg mice ( Figure 2E, F). No differences were observed for levels of the astrocyte marker GFAP or the microglia marker Iba1 ( Figure 2G, H).

Reduction in ER Chaperones and UPR Proteins in 14-3-3ftg Mice
RNAi-mediated down-regulation of 14-3-3f in organotypic hippocampal cultures results in an ER stress-like response that features increased levels of ER chaperones, including Lys-Asp-Glu-Leu (KDEL)-containing proteins such as Grp78 and Grp94 [17]. To investigate whether overexpression of 14-3-3f alters ER chaperones or the UPR in vivo we began by staining tissue sections from wild-type and 14-3-3ftg mice with antibodies against KDEL ( Figure 3A). KDEL immunoreactivity was detected in all CA neuronal populations and in the granule neurons of the dentate gyrus in wild-type mice. Lower KDEL immunoreactivity was apparent in hippocampal sections from 14-3-3ftg ( Figure 3A). This difference was confirmed using western blot analysis of lysates from CA1, CA3 and the dentate gyrus-enriched portions of the hippocampus (Figure 3B-D).
Basal mRNA levels of Grp78 and Grp94 were not different between wild-type and 14-3-3ftg mice ( Figure 3E and data not shown).
No differences were found between wild-type and 14-3-3ftg mice for CA3, CA1 or dentate gyrus levels of calnexin or eIF2a ( Figure 3F). In contrast, levels of phospho-eIF2a, ATF4 and ATF6 were lower in the CA3 subfield of 14-3-3ftg mice ( Figure 3F, G), although levels of phospho-eIF2a and ATF4 and ATF6 were not consistently different from wild-type in the dentate gyrus or CA1 subfield ( Figure 3F, G and data not shown). Levels of the spliced transcript of Xbp1 were not detectable in 14-3-3ftg mice whereas low levels of this were detected in wild-type animals, which were increased in mice subject to seizures ( Figure 3H).

14-3-3f Overexpression Protects Against ER Stressinduced Neuronal Death in vivo
In view of the lower levels of ER stress/UPR-related proteins, we postulated that 14-3-3ftg mice may show altered vulnerability to ER stress in vivo. To test this idea, wild-type and 14-3-3ftg mice were given an intracerebroventricular injection of the ER stressor tunicamycin. In wild-type mice this resulted in extensive death of dentate granule neurons 48 h later, as detected by Fluor-Jade B (FJB) staining ( Figure 4A) and TUNEL staining ( Figure 4C). In contrast, tunicamycin-induced neuronal death was significantly reduced in 14-3-3ftg mice ( Figure 4B, D). This was despite lower levels of many ER/UPR-related proteins including Grp94, phospho-eIF2a and phospho-Ire1 and lower levels of caspase-12 ( Figure 4E, F). ATF4 and ATF6 levels were not different after tunicamycin ( Figure 4E).

Seizure Severity is not Altered by 14-3-3f Overexpression in Mice
Reduced 14-3-3f levels have been shown to increase vulnerability to kainic acid in organotypic hippocampal slice cultures [17].
To determine whether over-expression of 14-3-3f has effects on seizure-induced neuronal death in vivo we examined the response of mice to status epilepticus. Intra-amygdala microinjection of kainic acid produces prolonged seizures which spread to the hippocampus via the entorhinal cortex and perforant pathway. Mossy fibres from dentate granule neurons synapse directly on CA3 pyramidal neurons, which are particular vulnerable to seizure-induced neuronal death in this model [31]. Protein levels of kainic acid receptors appeared normal in 14-3-3ftg ( Figure 5A). Forty minute EEG recordings from skull-mounted electrodes detected no differences in baseline total power or frequency parameters in 14-3-3ftg mice ( Figure 5B, C). Next, we recorded seizures in wild-type and 14-3-3ftg mice. Seizure durations in wild-type mice were similar to those reported previously for SJL mice in this model [31]. Electrographic seizure EEG was not different in 14-3-3ftg mice ( Figure 5D, E).

14-3-3f Overexpression Protects Against Seizure-induced Neuronal Death in vivo
Hippocampal damage in SJL mice subjected to intra-amygdala kainic acid-induced status epilepticus is found extensively in the CA3 subfield, with additional neuronal death in the CA1 subfield and hilus [31]. We next examined seizure-damage 72 h after status epilepticus in wild-type and 14-3-3ftg mice. As expected, FJB staining of tissue sections from wild-type mice revealed hippocampal damage was most extensive in the ipsilateral CA3 subfield ( Figure 6A). Seizure damage was also present in the CA1 subfield and hilus ( Figure 6A). FJB-staining was reduced in 14-3-3ftg mice in all subfields ( Figure 6A, B). Supporting these results, staining for irreversible DNA fragmentation using TUNEL confirmed extensive neuronal death in the CA3, CA1 and hilus of wild-type mice, which was reduced in 14-3-3ftg mice ( Figure 6A, B). 14-3-3ftg mice also displayed higher levels of normal NeuN staining in each hippocampal subfield ( Figure 6B and data not shown). Western blot analysis of protein lysates from wt and 14-3-3ftg mice revealed significantly lower levels of several proteins related to ER stress and the UPR in the CA3 subfield, including Grp 78 and 94, phospho-eIF2a and caspase-12 (p,0.05; n = 3 per group, data not shown).
Finally, to support the in vivo seizure data we exposed primary cultures of hippocampal neurons to kainic acid, an in vitro model of excitotoxicity [34]. Kainic acid treatment produced ,70% cell death in primary hippocampal cultures from wild-type SJL mice. In contrast, cell death was significantly lower in primary hippocampal cultures from 14-3-3ftg mice ( Figure 6C).

Discussion
In the present study we report that transgenic overexpression of 14-3-3f in mice alters basal levels of proteins involved in the UPR pathway. 14-3-3f overexpression nevertheless potently protected the hippocampus against ER stress and status epilepticus in vivo. These results suggest that restitution or delivery of 14-3-3f may protect against neurologic or neurodegenerative injuries in which excitotoxicity, ER stress or an impaired UPR response is a causal patho-mechanism.
14-3-3 proteins are increasingly recognized as crucial molecular chaperones during brain development, in neuronal function and disease [2]. There is some functional redundancy among certain 14-3-3 isoforms [39], although loss of 14-3-3f does not appear to be compensated by increased levels of other isoforms [5,6]. Likewise, 14-3-3 isoforms display selectivity in their capacity to prevent neuronal death according to the nature of the stressor, which may derive from their interacting partners and subcellular distribution [40]. The present study is the first to characterize the brain phenotype of mice overexpressing 14-3-3f. The animals were originally developed to explore tumor-promoting effects of 14-3-3f overexpression [30] and the mice do develop tumors but beyond the time when mice were used presently (E. J-M., unpublished data). The overexpressed 14-3-3f was found throughout the mouse hippocampus and detectable in the cytosol, nucleus, mitochondria and ER-containing fractions. Thus, the transgene expressed and distributed similarly to the endogenous protein. This excludes features of the phenotype of the mice being due to erroneous distribution of extra 14-3-3f to cell populations or compartments in which the protein is not ordinarily present. 14-3-3f protein was recently reported to be restricted to the CA pyramidal neurons and granule neurons of the hippocampus in adult mice [6,15,17]. Our studies confirmed these cells were immunoreactive for 14-3-3f, and expressed the myc-tagged transgene, but we detected the transgene in other brain regions, consistent with other studies showing 14-3-3f to be present outside the hippocampus in adult brain [41].
14-3-3f has been reported to regulate neurite outgrowth and is required for normal hippocampal development [6,42]. Indeed, deletion of 14-3-3f results in mis-located pyramidal neurons and changes to the mossy fiber pathway within the infra-and suprapyramidal layer and a predicted increased excitatory drive onto CA3 pyramidal neurons [6]. We did not detect any hippocampal abnormalities in 14-3-3ftg mice, indicating that overexpression of 14-3-3f does not produce an opposite phenotype to gene deletion. We also detected no differences in body weight, reported for 14-3-3f 2/2 mice [6] although a smaller brain weight was noted. The cause of this is uncertain and differences in neuron or glia populations were not observed. Indeed, consistent with normal brain development, we found no differences in baseline or seizure 14-3-3f Protects against ER Stress and Seizures PLOS ONE | www.plosone.org EEG in 14-3-3f transgenic mice, establishing an equivalent episode of status epilepticus was likely incurred and thus differences in damage can be confidently assigned to altered cell death/survival signalling rather than due to reduced seizures.
14-3-3 proteins regulate apoptosis by sequestering pro-apoptotic proteins including Bad and ASK-1, which are activated by excitotoxic insults to the brain and contribute to neuronal death in models of seizure and stroke [12,43]. Overexpression of 14-3-3f did not alter resting levels of apoptosis-associated proteins to which it is known to bind, or change levels of autophagy-related proteins. Kainic acid receptors, which are trafficked from the ER by a 14-3-3-regulated mechanism [4], were also at normal levels in 14-3-3ftg mice. Nevertheless, additional 14-3-3f may function as a pool to buffer against pro-apoptotic proteins released from endogenous 14-3-3f or other isoforms during apoptosis. The main molecular adjustment in 14-3-3ftg mice, however, was downregulation of proteins involved in the UPR pathway. Most prominent was reduced levels of KDEL-containing proteins although all branches of the UPR were affected. This was probably due to posttranslational mechanisms since transcript levels of several tested genes were normal. 14-3-3f overexpression did have effects on mRNA, however, reducing Xbp1 splicing, a consequence of the activation of the IRE1 branch of the UPR that promotes increased expression of ER chaperones [22]. Taken together, these data suggest overexpressed 14-3-3f may produce a selective adjustment of the UPR. Since this included reduced levels of proteins involved in folding, this is consistent with 14-3-3f functioning as a sweeper of mis-folded proteins [3]. ER stress is implicated as a patho-mechanism underlying neurodegeneration in several diseases, including epilepsy [44], although NMDA receptor-induced neuronal death can occur independently of ER stress in vivo [45]. Studies here demonstrated that overexpressed 14-3-3fwas capable of protecting against ER stress induced by tunicamycin. Tunicamycin causes ER stress by preventing N-glycosylation of proteins, thus resulting in a build-up of proteins in the ER and triggering the UPR and ER stressinduced apoptosis, although direct effects on neurotransmission have been reported [46]. Tunicamycin injection into the brain caused the selective death of neurons within the dentate gyrus, consistent with in vitro reports [17]. 14-3-3f strongly protected  against tunicamycin-induced neuronal death despite the lower resting levels of UPR/ER stress proteins. Even after treatment, levels of many UPR/ER stress proteins remained lower in 14-3-3ftg mice, excluding an effect secondary to normalization of levels. Higher 14-3-3f levels may therefore reduce the stress caused by tunicamycin and over-expressed 14-3-3f may be sufficient to protect in the absence of a complete complement of ER stress chaperones. 14-3-3f delivery may be a means to reduce ER stress where the normal UPR response is inadequate or otherwise disrupted, such as in certain neurologic and neurodegenerative diseases [47]. The data also compliment the findings of 14-3-3f silencing in hippocampus, which triggers up-regulation of KDELcontaining proteins and sensitizes against tunicamycin-induced cell death [48].
A major finding in the present study was that overexpression of 14-3-3f potently protected against seizure-induced neuronal death in vivo. Protection was found for both pyramidal and hilar interneurons, indicating 14-3-3f overexpression protects regardless of neuronal phenotype or location within the hippocampal circuitry. These results compliment earlier work that showed lowering 14-3-3f levels in the mouse hippocampus increased neuronal death after kainate treatment of organotypic cultures [48]. The extent of protection is similar and in several cases greater than achieved by deletion of pro-apoptotic members of the Bcl-2 family in the same model [49,50,51]. This would be consistent with 14-3-3f function either upstream or being involved in curtailing cell death via actions in more than one compartment. Over-producing an anti-apoptotic protein may also be more effective than deleting a pro-apoptotic protein. Again, the protection obtained was in spite of a lower compliment of ER stress chaperones in these mice. This agrees with in vivo evidence that NMDA-dependent neuronal death in vivo is not ER stressdependent [45]. This could mean that either 14-3-3f is more effective than these proteins at protection in this model, or that the protection derives from functions in addition to ER stress-related cell death. This is likely; we detected 14-3-3f overexpression throughout the cell and 14-3-3 is able to sequester various proapoptotic Bcl-2 family proteins which contribute to seizureinduced neuronal death in the model [12,52]. Whether or not the reduced hippocampal damage in 14-3-3ftc mice would result in a beneficial effect on the post-status epilepticus epilepsy phenotype is uncertain, although our previous studies in which hippocampal damage was reduced by genetic or other means, led to fewer spontaneous recurrent seizures [51,53].
Levels of several proteins associated with the UPR have been found to be higher in the hippocampus of patients with temporal lobe epilepsy [26,27]. Levels of these proteins were generally lower, however, in the hippocampus of 14-3-3f transgenic mice suggesting the model does not phenocopy this molecular feature of human temporal lobe epilepsy. Nevertheless, elevated 14-3-3f levels were reported in the microsome-containing fraction of hippocampus from patients with temporal lobe epilepsy [15]. Although we cannot directly compare the higher 14-3-3f levels in the transgenic mice with those in patient brain, we may speculate, on the basis of our results here, that increased 14-3-3f in intractable human epilepsy could be protective against ongoing neuron loss. We note, however, that 14-3-3f is not uniformly neuroprotective, and fails to prevent neurodegeneration in models of Parkinson's disease [40]. 14-3-3f can also promote phosphorylation of Tau and chronic over-expression might have potentially deleterious effects on the brain [54,55], which could be assessed using the present model. Any strategy to enhance 14-3-3f expression may provide neuroprotection only against the acute effects of prolonged seizures or perhaps stroke, which share common patho-mechanisms such as excitotoxicity and apoptosisassociated signalling [11].
In summary, the present study demonstrates that 14-3-3f overexpression results in a selective downregulation of UPR pathways and confers protection against ER stress-and seizureinduced neuronal death in the mouse hippocampus. Restitution or overexpression of this 14-3-3 isoform may be a potential therapeutic approach for status epilepticus but not necessarily all CNS diseases associated with impaired 14-3-3f expression [2].