Address correspondence and reprint requests to David C. Henshall PhD, Department of Physiology and Medical Physics, Royal College of Surgeons in Ireland, 123 St. Stephen's Green, Dublin 2, Ireland. E-mail: email@example.com
14-3-3 proteins are a family of signaling molecules involved in diverse cellular functions, which can mediate anti-apoptotic effects. Seizure-induced neuronal death may involve programmed (apoptotic) cell death pathways and is associated with a decline in brain 14-3-3 levels. Presently, we investigated the subcellular localization and effects of seizures on isoforms of 14-3-3 in rat hippocampus, and contrasted these to findings in human temporal lobe epilepsy (TLE). All brain isoforms of 14-3-3 were detected in the cytoplasmic compartment of rat hippocampus, while 14-3-3γ and -ζ were also present in mitochondrial and microsome-enriched fractions. Focally evoked seizures in rats significantly reduced 14-3-3γ levels within the microsome-enriched compartment at 4 h, with similar responses for 14-3-3ζ, while cytoplasm-localized 14-3-3β, -ε and -η remained unchanged. Analysis of human autopsy control hippocampus revealed similar 14-3-3 isoform expression profiles. In TLE samples, the microsome-enriched fraction also showed differences, but here 14-3-3ε and -ζ levels were higher than controls. TLE sample 14-3-3 isoform abundance within the cytoplasmic fraction was not different to controls. This study defines the subcellular localization of 14-3-3 isoforms in rat and human hippocampus and identifies the microsome-enriched fraction as the main site of altered 14-3-3 levels in response to acute prolonged and chronic recurrent seizures.
The 14-3-3 proteins are a family of 28–33 kDa acidic signaling and molecular chaperone proteins involved in diverse functions including protein trafficking, ion channel and synaptic function, cell cycle control and apoptosis (Aitken et al. 1992). Expression of 14-3-3 is highest in brain and largely neuronal, where five isoforms are expressed, namely β, ε, γ, η, and ζ. The 14-3-3σ and 14-3-3τ/θ isoforms are non-neuronally expressed (Hermeking 2003). Each isoform is encoded on a separate gene and the tissue and intracellular distribution patterns are diverse and subtype specific in many cases (Hermeking 2003). Over 100 interacting partners have been identified to date and 14-3-3 proteins are associated with all major cell compartments including the cytoplasm, endoplasmic reticulum (ER) and Golgi apparatus (microsomes), nucleus and plasma membrane (Martin et al. 1994; van Hemert et al. 2001; Nufer and Hauri 2003).
14-3-3 proteins play important roles in the regulation of apoptosis, where they mainly mediate anti-apoptotic functions, sequestering pro-apoptotic molecules including the Bcl-2 family protein Bad (Zha et al. 1996), forkhead transcription factors (Brunet et al. 1999) and apoptosis signal regulating kinase 1 (Ask1; Zhang et al. 1999). Cell death signaling proceeds after either dissociation or degradation of 14-3-3. Evidence is accumulating for critical roles of 14-3-3 proteins in the setting of neuronal death and neurodegenerative diseases (Berg et al. 2003). Seizures are known to trigger neuronal death that features molecular aspects of apoptosis and these pathways are also implicated in the pathogenesis of hippocampal sclerosis in temporal lobe epilepsy (TLE; Henshall and Simon 2005). We have previously detected dissociation of 14-3-3 from pro-apoptotic proteins including Bad (Henshall et al. 2002) and Ask1 (Shinoda et al. 2003) after seizures, and interaction of 14-3-3 with novel apoptosis-regulating proteins (Henshall et al. 2003). Damaging seizures rapidly induce a decline in brain levels of total 14-3-3 (Schindler et al. 2004), which studies incorporating later time points suggest is the result of proteolysis (Henshall et al. 2002). However, the distribution and responses of specific 14-3-3 isoforms during seizure-induced neuronal death is unknown. Presently, we examined the expression of the main brain isoforms of 14-3-3 in rat hippocampus and their response during seizure-induced neuronal death, and contrasted this to findings in human control and TLE hippocampus.
Materials and methods
All animal procedures were performed in a facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) in accordance with protocols approved by the Institutional Animal Care and Use Committee and the principles outlined in the National Institute of Health Guide for the Care and Use of Laboratory Animals. Studies were performed according to previously described techniques (Schindler et al. 2004; Shinoda et al. 2004). Briefly, adult male Sprague–Dawley rats (280–320 g) underwent seizures induced by unilateral intra-amygdala microinjection of kainic acid (Sigma, St Louis, MO, USA) which results in neuronal death within the ipsilateral CA3 and hilar region of the hippocampus (Schindler et al. 2004; Shinoda et al. 2004). Under isoflurane anesthesia, rats were intubated, ventilated and their femoral vein was catheterized. Subsequently, rats were placed in a stereotaxic frame and affixed with skull-mounted recording electrodes (Plastics One Inc., Roanoke, VA, USA) bi-temporally over the hippocampi and a third across the frontal cortex to record EEG. A craniectomy was also performed for placement of the injection cannula. The animal was then removed from the frame and anesthesia discontinued. EEG recordings were commenced (Grass model 8–16) and then a 31-gauge internal cannula (Plastics One Inc.) was inserted into the lumen of the guide to inject kainic acid (0.1 μg in 0.5 μL phosphate buffer) into the amygdala. The EEG was continuously monitored until lorazepam (6 mg/kg intravenously) was administered to terminate seizures after 40 min. All animals experienced injury-causing polyspike paroxysmal seizures. Rats were killed 4 h following lorazepam.
Human brain samples
This study was approved by the Legacy Health System Institutional Review Board and informed consent was obtained from all patients. Clinical data for patients have previously been published (Shinoda et al. 2004). Patients with medically intractable TLE were referred for surgical resection of the left or right temporal lobe by an epileptologist following neurological assessment, video EEG recording and MRI/neuroimaging. No patients had experienced status epilepticus during the year in which their surgery was performed. All patients were taking anticonvulsant medication prior to surgery. The hippocampus was frozen in liquid nitrogen and stored at −70°C until use. A coronal slab of ∼1 mm thickness was subsequently prepared from the sample for fractionation. Control hippocampi were obtained from the Brain and Tissue Bank for Developmental Disorders at the University of Maryland, Baltimore, MD, USA. These specimens were similar fresh frozen, en bloc hippocampi from people who died of causes not related to known neurological disease. Subcellular fractionation was performed on a selection of control (C1–C3; gender ratio 2 male : 1 female; age 39 ± 12 years, range 15–53 years; post-mortem interval 8 ± 2 h) and epilepsy specimens (E2–E4; gender ratio 2 female : 1 male; age 23 ± 4 years, range 16–29 years) as previously described (Yamamoto et al. 2006).
Subcellular fractionation and western blotting
Hippocampi were fractionated to obtain the cytoplasm, mitochondria and microsome-enriched fractions, according to previous methods (Schindler et al. 2004; Yamamoto et al. 2006). Briefly, samples were homogenized in a mannitol/sucrose buffer containing a protease inhibitor cocktail and then centrifuged twice at 1200 g for 10 min. The post-nuclear supernatant was then centrifuged twice at 10 000 g for 15 min and the resulting mitochondrial pellet was re-suspended in a sucrose buffer and purified through a percoll bilayer by centrifugation at 41 000 g for 30 min. The crude cytosolic fraction was then centrifuged at 100 000 g for 1 h to separate the microsomal and cytosolic fractions. Following fractionation, protein samples (50 μg) were boiled in gel-loading buffer and then separated on 12% sodium dodecyl sulfate–polyacrylamide electrophoresis gels. Proteins were transferred to polyvinylidene difluoride membranes (Bio-Rad Laboratories, Hercules, CA, USA) and then incubated with antibodies against 14-3-3β (sc-17288), 14-3-3ε (sc-1020), 14-3-3γ (sc-731), 14-3-3η (sc-17287), 14-3-3σ (sc-7683, sc-7681), 14-3-3τ (sc-732), 14-3-3ζ (sc-1019; Santa Cruz Biotechnology, Santa Cruz, CA, USA). For verification of fraction quality the following antibodies were used: α-tubulin (Santa Cruz Biotechnology), Bim (Stressgen, Victoria, BC, Canada), cytochrome IV oxidase (CoxIV; Molecular Probes, Eugene, OR, USA), proteosome 20S (Affinity Bioreagents, Golden, CO, USA) and caspase 12 (Cell Signaling Technology, Beverly, MA, USA). Membranes were then incubated with appropriate secondary antibodies (1 : 2000 dilution) followed by chemiluminescence detection (NEN Life Science Products, Boston, MA, USA), and then exposed to Kodak X-OMAT film (Kodak, Rochester, NY, USA).
Gel bands were analyzed semiquantitatively using time-matched (60-s) exposures and gel-scanning integrated optical density software (AlphaEaseFC V4.0; Alpha Innotech Corporation, San Leandro, CA, USA). Data are presented as mean ± SEM and optical density values presented either directly or expressed as fold change of ipsilateral versus contralateral (rat) or control versus TLE (human) and analyzed using a Mann–Whitney U-test. Significance was accepted at p < 0.05.
Relative abundance of 14-3-3 isoforms within subcellular fractions of rat hippocampus
The composition of subcellular fractions was validated using markers for intracellular compartments and organelles as shown in Fig. 1(a). CoxIV was detected exclusively in the mitochondrial compartment while the proteosome 20S and caspase 12 were present in microsome-enriched fractions. The Bcl-2 family protein Bim was detected in cytoplasmic and, to a lesser extent, mitochondrial fractions, while α-tubulin was mainly detected in the cytoplasm with lower but similar amounts found in mitochondrial and microsomal compartments (Fig. 1a).
Initially, we undertook an assessment of the pattern of expression of 14-3-3 isoforms in each fraction using rat hippocampus contralateral to the seizure focus (Figs 1b–d). Western blotting determined the main neuronal 14-3-3 isoforms were all present in the cytoplasmic fraction, including 14-3-3β (∼29 kDa), 14-3-3ε (∼31 kDa), 14-3-3γ (∼29 kDa), 14-3-3η (∼27 kDa) and 14-3-3ζ (∼28 kDa; Fig. 1b). 14-3-3τ (∼29 kDa), which may be present in white matter, was also detected, while immunoblotting using two different antibodies did not detect 14-3-3σ, a non-CNS isoform (Fig. 1b). Four of six 14-3-3 isoforms were present in the mitochondrial fraction (Fig. 1c). Only 14-3-3γ and 14-3-3ζ were found within the microsome-enriched fraction (Fig. 1d).
Levels of cytoplasm-only 14-3-3 isoforms are not affected by seizures
Next, we analyzed the response of each isoform to seizures by comparing levels within the seizure-damaged ipsilateral hippocampus at 4 h, to levels in the undamaged contralateral hippocampus. 14-3-3β, 14-3-3η and 14-3-3τ isoforms were only found in the cytoplasm and seizures did not significantly alter their levels in this fraction or the mitochondrial or microsome-enriched compartments (Figs 2a–d). The 14-3-3ε isoform, which was found in the mitochondrial as well as cytoplasmic fraction, was similarly unaffected by seizures (Figs 2e–g).
Levels of 14-3-3γ are reduced in the microsome-enriched fraction following seizures
We next examined the first of the two 14-3-3 isoforms normally present in the microsome-enriched fraction. 14-3-3γ was robustly expressed in each hippocampal fraction, but cytoplasmic and mitochondrial levels were not affected by seizures (Figs 3a and b). In contrast, 14-3-3γ levels were significantly lower in the microsomal fraction of the ipsilateral, seizure-damaged hippocampus compared with the undamaged contralateral hippocampus (Figs 3a and b).
The profile for 14-3-3ζ was broadly similar to that of 14-3-3γ in that seizures induced a trend to lower 14-3-3ζ levels in the microsome-enriched fraction, while levels remained largely unchanged in the cytoplasm and mitochondrial fractions (Figs 4a and b).
Unaltered levels of α-tubulin in the microsome-enriched fraction after seizures
As levels of the two 14-3-3 isoforms expressed in the microsome-enriched compartment were both lower in seizure-damaged brain, we wished to determine whether other proteins present in this fraction decline as a result of seizures. Accordingly, we analyzed α-tubulin levels in each fraction (Fig. 4c). Semiquantitative densitometry determined there were no significant differences in α-tubulin levels between ipsilateral and contralateral hippocampus in any subcellular fraction (Fig. 4c).
14-3-3 isoforms in human hippocampus and alterations in temporal lobe epilepsy
Having established the profiles of each 14-3-3 isoform in rat we next undertook experiments to define the profiles of the main 14-3-3 neuronal isoforms β, ε, γ, η and ζ in human hippocampus and asked whether 14-3-3 isoforms are regulated in the hippocampus of patients with intractable TLE. Fractionation of human samples was undertaken in the same manner as for rat and as previously described (Yamamoto et al. 2006), and fractionation quality verified using specific organelle markers for each sample (data not shown).
The cytoplasmic fraction of control human hippocampus contained each 14-3-3 isoform previously detected in rat (Fig. 5a). As with rat, 14-3-3ε, 14-3-3γ and 14-3-3ζ had relatively higher expression than 14-3-3β and 14-3-3η. The mitochondrial fraction of control human hippocampus contained the same four 14-3-3 isoforms detected in the rat hippocampus (Fig. 5b). Abundance of 14-3-3 isoforms in the human microsome-enriched fraction was also comparable with rat, whereby 14-3-3γ and 14-3-3ζ were predominantly present and at a similar ratio to rat. However, very low levels of 14-3-3ε could also be detected (Fig. 5c).
Semiquantitative comparison of control brain 14-3-3 levels to levels in TLE hippocampus revealed no significant differences for any of the five isoforms in the cytoplasmic fraction (Fig. 5d). Within mitochondrial fractions, the same isoforms were present in TLE samples as controls (Fig. 6a). However, levels of 14-3-3η were significantly lower in TLE samples than controls (Figs 6a and b). Finally, we detected significantly higher levels of 14-3-3ε (Figs 6b and c) and 14-3-3ζ (Figs 6b and c) in the microsome-enriched fraction of TLE samples compared with controls. TLE brain samples also showed a trend toward higher levels of 14-3-3γ than controls in this fraction, although this did not reach statistical significance (Fig. 6c and data not shown).
The present study defines the distribution and fraction abundance of each neuronal isoform of 14-3-3 in rat and human hippocampus and reveals seizures selectively alter 14-3-3 isoform levels, particularly within the microsome-enriched fraction. Our data extend previous work aimed at examining the relative abundance and subcellular localizations of the main neuronal 14-3-3 isoforms (Martin et al. 1994; Pollak et al. 2006). In line with previous reports (Watanabe et al. 1993), 14-3-3γ appeared to be particularly abundant in the hippocampus, although our approach was qualitative rather than quantitative. We also readily detected 14-3-3ε and -ζ, along with rather lower levels of 14-3-3β and 14-3-3η, profiles similar to mouse hippocampus (Pollak et al. 2006). We did not detect 14-3-3σ in rat or human hippocampus, in line with its known non-CNS expression. Our detection of low levels of 14-3-3τ, a supposedly T-cell and white-matter isoform may also reflect a signal from blood component contamination or partial cross-reactivity with 14-3-3ζ (Wiltfang et al. 1999).
We previously identified reductions in total 14-3-3 levels during seizure-induced neuronal death in rat hippocampus (Schindler et al. 2004), and hypothesized this may be a pro-apoptotic mechanism, in line with work in other cell death systems (Masters and Fu 2001; Won et al. 2003). The present study reveals the microsome-enriched fraction as the site(s) of this change and identifies the isoform as 14-3-3γ, with perhaps further contribution by reductions of 14-3-3ζ. While these were the only two isoforms present in this fraction in rat, this change was not the result of a general decline in protein levels in this fraction as levels of an unrelated protein did not change (see Fig. 4c). Both 14-3-3γ and 14-3-3ζ are known to be released when neurons are damaged, although our data differ somewhat from a necrotic profile which would be predicted to include 14-3-3β and 14-3-3η changes (Siman et al. 2004). Thus, our data identify an unusually selective change to 14-3-3 isoform(s) within rat hippocampus during an early stage of the cell death process. Accordingly, loss/degradation or re-distribution of 14-3-3 may be an apical event in the cell death process. Indeed, the decline in 14-3-3 coincides with activation of pro-apoptotic Bad and caspase 3 in this model (Henshall et al. 2000, 2002).
The specificity of the observed changes in terms of isoform and intracellular site was unexpected, based on previous work in this model where a drop in cytoplasmic 14-3-3 had also been detected (Schindler et al. 2004). Because cytoplasmic changes were not seen in rat (or human), we would offer that the changes to overall cytoplasmic 14-3-3 may have been the result of quite small changes occurring across several isoforms. Our data support other work showing seizures regulating 14-3-3 isoform levels in a selective manner. For example, cDNA profiling by Lukasiuk et al. (2003) detected mRNA down-regulation of two isoforms following prolonged seizures in rats. Remarkably, these were also the 14-3-3γ and 14-3-3ζ isoforms, suggesting these isoforms are also regulated by seizures at a transcriptional level.
The mechanism underlying reduced microsomal 14-3-3 levels after seizures is likely to be proteolysis dependent because we did not detect increases in other fractions that might indicate translocation. 14-3-3 is known to be proteolyzed during apoptosis by caspase 3 (Won et al. 2003) and caspase 3 is active at this time during seizures in this model (Henshall et al. 2000). Moreover, we have previously detected 14-3-3 cleavage fragments in this model, albeit at later times (Henshall et al. 2002). However, other shuttling sites remain, including the nucleus, and we have observed apparent 14-3-3 nuclear accumulation in the late stages of neuronal death after seizures (Henshall et al. 2002).
The ER and Golgi apparatus are major components of the microsomes and responsible for several cellular activities including protein trafficking, regulating intracellular calcium levels and, under conditions of stress, initiating an apoptosis program (Xu et al. 2005). Several functions have been identified for microsome-localized 14-3-3s including 14-3-3ζ and 14-3-3γ in ER and Golgi protein trafficking and assembly (Dorner et al. 1999; Nufer and Hauri 2003; Yuan et al. 2003; Preisinger et al. 2004). A role for the ER has emerged in the setting of seizure-induced neuronal death (Kitao et al. 2001) and ER stress and ER-localized apoptosis signaling have been detected in hippocampus from patients with intractable TLE (Yamamoto et al. 2006). ER-localized 14-3-3-binding proteins particularly relevant in the setting of cell death may include Ask1, which is involved in ER stress-induced apoptosis (Nishitoh et al. 2002) and sequestered by 14-3-3ζ (Zhang et al. 1999). Ask1 is released from 14-3-3 during seizure-induced neuronal death in our rat model (Shinoda et al. 2003). Additionally, our group has detected binding of 14-3-3 to death-associated protein kinase (Henshall et al. 2003), which in human TLE hippocampus is abundant in the microsome-enriched fraction (Henshall et al. 2004). Thus, in the setting of seizure-induced neuronal death, the observed changes in microsomal levels of 14-3-3 may be relevant to the ER's functions, in particular regulating stress responses and apoptosis. Studies to investigate neuroprotective effects of 14-3-3 over-expression during seizures, particularly if targeted to this subcellular compartment, might now be warranted.
The expression and subcellular distribution of 14-3-3 isoforms in control human hippocampus was similar in many aspects to the hippocampus contralateral to the seizure focus in rat. This may support a conserved expression and possibly, function, profile for these proteins in this brain region. As with our rat studies, TLE hippocampus did not differ from control in levels of any 14-3-3 isoform within the cytoplasm, but did for two isoforms within the microsome-enriched fraction and one within the mitochondrial compartment. However, in contrast to the reduced 14-3-3 levels in the microsome-enriched fraction in the setting of extensive post-seizure neuronal death, TLE hippocampus showed higher microsomal levels of 14-3-3ζ, 14-3-3ε and to a lesser extent 14-3-3γ. If loss of 14-3-3 occurs in the setting of extensive seizure-induced neuronal death, raised levels of 14-3-3 might potentially be an adaptive or even protective response to recurrent, less damaging seizures. Previous studies support such a hypothesis: Forkhead box O-class 1/3a, transcriptional activator(s) of pro-apoptotic Bim, which are themselves 14-3-3 binding proteins, are activated during acute neuronal death in our rat seizure model, but suppressed in TLE brain samples and rat brain subject to repeated non-damaging seizures (Shinoda et al. 2004). Thus, specific 14-3-3 isoforms, like certain death pathway mediators, may be differentially regulated between the acute neuronal loss that follows prolonged seizures and the setting of chronic recurrent seizures that are less harmful to brain. This hypothesis may also be supported by work in the setting of ischemic preconditioning, a neuroprotective paradigm, which raises brain 14-3-3 levels (Stenzel-Poore et al. 2003). However, 14-3-3ε levels may also be elevated as a consequence of astrogliosis in TLE samples (Satoh et al. 2004) and the gene(s) that causes tuberous sclerosis, a disorder that also causes seizures, up-regulate 14-3-3ε, -γ, and -ζ levels (Hengstschlager et al. 2003). Over-expression of 14-3-3ζ has also been implicated in neurodegenerative changes in Alzheimer's disease (Hernandez et al. 2004; Umahara et al. 2004). Accordingly, raised 14-3-3 levels in TLE brain may be a beneficial response in brain subject to frequent seizures, but maladaptive or non-neuronal-related changes should not be excluded. Profiling 14-3-3 levels in other experimental paradigms such as chronically epileptic rats may provide additional insight into the differential expression patterns seen in TLE material. Studies might also explore whether 14-3-3 levels in TLE samples vary as a function of the degree of sclerosis.
Some limitations could be considered in the interpretation of the present findings. Drawing comparisons between our experimental model of prolonged seizures and clinical material from patients with recurrent epileptic seizures must be considered cautiously, particularly with small cohorts such as used presently. While enriched for ER, the microsomal fraction might contain synaptosomes and plasma membrane which are also enriched in 14-3-3 isoforms (Martin et al. 1994). Some cross-reactivity between antibodies for the various 14-3-3 isoforms is also expected, although this would not detract from the main finding that seizures alter microsomal levels of 14-3-3. In earlier studies, 14-3-3ζ and -β antibodies have cross-reacted because of N-terminal similarities (Martin et al. 1994), but anti-14-3-3ζ and -β used here were raised against internal sequences. Last, 14-3-3 isoforms are encoded on different genes and alternative splicing and other gene variation may give rise to further subspecies of isoforms that were not resolved presently but may yet perform important functions.
In conclusion, our study identifies the subcellular distribution of the main 14-3-3 isoforms in rat and human hippocampus. The regulation of a subset of 14-3-3 isoforms within specific fractions may highlight participation of intracellular organelle-related signaling pathways which influence neuronal (dys)function in disorders such as epilepsy.
The authors would like to thank Sachiko Shinoda for technical assistance and Roger Simon and Jochen Prehn for constructive criticism of the manuscript. We also thank Norman So and the Oregon Comprehensive Epilepsy Program and Drs Rosenban and Abtin for surgical collection of the specimens, and the University of Maryland Brain and Tissue Bank for autopsy specimens. This research was supported by Science Foundation Ireland (B466), Marie Curie Actions and NIH/NINDS grants NS39016 and NS41935.