|Home | About | Journals | Submit | Contact Us | Français|
Certain mutations within the mammalian target of rapamycin (mTOR) pathway, most notably those affecting the tuberous sclerosis complex (TSC), lead to aberrant activation of mTOR and result in a high incidence of epilepsy in humans and animal models. Although hyperactivation of mTOR has been strongly linked to the development of epilepsy and, conversely, inhibition of mTOR by rapamycin treatment is protective against seizures in several models, the downstream epileptic mechanisms have remained elusive. Autophagy, a catabolic process which plays a vital role in cellular homeostasis by mediating the turnover of cytoplasmic constituents, is negatively regulated by mTOR. Here we demonstrate that autophagy is suppressed in brain tissues of forebrain-specific conditional TSC1 and phosphatase and tensin homlog (PTEN) knockout (KO) mice, both of which display aberrant mTOR activation and seizures. In addition, we also discovered that autophagy is suppressed in the brains of human TSC patients. Moreover, conditional deletion of Atg7, an essential regulator of autophagy, in mouse forebrain neurons is sufficient to promote development of spontaneous seizures. Thus, our study suggests that impaired autophagy contributes to epileptogenesis, which may be of interest as a potential therapeutic target for epilepsy treatment and/or prevention.
Macroautophagy, henceforth referred to as autophagy, is a catabolic process which mediates the turnover of cytoplasmic constituents via lysosomal degradation (Rubinsztein et al., 2005; Yang and Klionsky, 2010). Genetic studies primarily in yeast revealed that a set of autophagy-related (Atg) genes acting similarly to the ubiquitin-conjugation system modulates autophagosome biogenesis (Suzuki et al., 2001; Kim et al., 2002; Xie and Klionsky, 2007). Autophagy has been implicated in various physiological processes (Levine and Kroemer, 2008), including the adaptive response to starvation, quality control of intracellular proteins and organelles (mitophagy) (Wang and Klionsky, 2011), embryonic development, cellular defense against bacterial invasion (Mizushima et al., 2008), antigen presentation (Deretic and Levine, 2009), tumor suppression (Chen and Debnath, 2010; Komatsu, 2011; Takamura et al., 2011), homeostasis of axons (Komatsu et al., 2007), and neurodegeneration (Komatsu et al., 2005; Hara et al., 2006; Komatsu et al., 2006; Rubinsztein, 2006). Autophagy activity is negatively regulated by the mammalian target of rapamycin (mTOR) in response to growth factors, nutrient bioavailability, and stress (Levine and Klionsky, 2004). Recent studies revealed that mTOR regulates autophagy via phosphorylation of the autophagy-initiating kinase Ulk1 (Kim et al., 2011).
mTOR itself is tightly controlled by upstream regulators, including TSC and PTEN (Laplante and Sabatini, 2012). Loss-of-function mutations in these regulators lead to hyperactivation of mTOR (European Chromosome 16 Tuberous Sclerosis, 1993; van Slegtenhorst et al., 1997; Inoki et al., 2002; Onda et al., 2002; Baybis et al., 2004) and result in a high incidence of epilepsy in both humans (Padberg et al., 1991; Thiele, 2004; Holmes and Stafstrom, 2007) and animal models (Meikle et al., 2008; Zeng et al., 2008; Zhou et al., 2009; Sunnen et al., 2011). It is generally believed that aberrant hyperactivation of mTOR is primarily responsible for the epileptogenic process (Cao et al., 2009; Wong, 2010); however, the downstream mechanisms remain to be elucidated. Recent studies found that autophagy activity is down-regulated in mouse embryonic fibroblast (MEF) cells following TSC2 deletion (Ng et al., 2011; Parkhitko et al., 2011), although the effects of disinhibition of mTOR on autophagy in neurons remain unclear. Autophagy has also recently been found to be compromised in Lafora disease, which is strongly associated with epilepsy (Knecht et al., 2010), further supporting a possible role for autophagy in epileptogenesis. In the present study, we investigated if autophagy, a downstream component of the mTOR pathway, is altered in association with loss of function of TSC1 or PTEN and, furthermore, if inactivation of autophagy is sufficient to cause seizures.
Atg7flox/flox (developed by Komatsu et al (Komatsu et al., 2005)), TSC1flox/flox (Kwiatkowski et al., 2002) and PTENflox/flox (Groszer et al., 2001) (from Jackson lab) mice were crossed with CaMKIIα-Cre mice (strain T29-1 (Tsien et al., 1996)). Animals were housed in a pathogen-free, temperature- and humidity-controlled facility with a 12 hr light cycle (lights on at 0700) and ab libitum access to food and water. All experiments were performed according to the guidelines set by the Animal Care and Use Committee as well as the National Institutes of Health guide for care and use of laboratory animals. Efforts were made to minimize suffering and unnecessary use of animals.
DNA extraction was accomplished by digestion of 2 mm mouse tail snips in 500μl of digestion buffer (50 mM KCl, 10 mM Tris-HCl, 2.5 mM MgCl2, 0.45% v/v NP40 and 0.45% v/v Tween 20) with 100μg/ml proteinase K at 56°C overnight. On the next day, the samples were heated to 95°C for 5 min. PCR was performed using the following primers: wild-type Atg7 sense 5″-tgcatgtctgtggttgcttc, antisense 5″-agaggggtacaggggcatac, floxed Atg7 sense 5″-ggacttgtgcctcaccagat, antisense 5″-ctcgtcactcatgtcccaga, TSC (detects WT and mutant) sense 5″-gtcacgaccgtaggagaagc, antisense 5″-gaatcaaccccacagagcat, PTEN (detects WT and mutant) 5″-caagcactctgcgaactgag, antisense 5″-aagtttttgaaggcaagatgc, and Cre sense 5″-gcatttctggggattgtta and antisense 5″-cccggcaaaacaggtagtta.
Mice were anesthetized by i.p. injection of pentobarbital (30 mg/kg) supplemented with inhalation of isoflurane. Mice were first perfused with 30 ml of PBS and subsequently perfused with 4% paraformaldehyde. Brains were removed and held in 4% paraformaldehyde for 24 hr at 4°C and then transferred into 30% sucrose solution at 4°C until they were completely submerged. Brain tissues were then frozen and sectioned into 40 μm slices using a cryostat and stored at 4°C in PBS with sodium azide.
Mice were sacrificed and the brains were removed and sliced into 2 mm thick sections in a stainless steel matrix. Similar anatomic regions were identified and a 2 mm diameter hole punch was used to isolate two punches from the desired regions for each animal. Isolated tissues were homogenized in lysis buffer consisting of 50 mM Tris, pH 7.4, 2 mM EDTA and a proteinase inhibitor set (Roche). The lysates were then mixed with equal volumes of 2X LDS sample buffer comprised of 20% β-mercaptoethanol, 20 mM NaF, 20 mM of Na3VO4, and proteinase inhibitor, and heated at 95°C for 5 min. Insoluble cell debris was removed by centrifugation at 10,000 g for 10 min. Protein samples were resolved in an 8% Bis-Tris gel in MES buffer (50 mM MES, 50 mM Tris-HCl, 1 mM EDTA and 0.1 % SDS) and transferred to a 0.45 μm nitrocellulose membrane in transfer buffer (25 mM Tris, 200 mM glycine and 20% methanol). For LC3 detection, proteins were resolved by 15% SDS-PAGE and transferred onto a PVDF membrane. Membranes were subsequently blocked in 5% nonfat dry milk in TBST (25 mM Tris-HCl, pH 7.4, 1.5 M NaCl, 0.05% Tween-20) for 1 hr at room temperature (RT) and incubated with rabbit anti-p62, anti-S6, anti-phospho S6(Ser235/236), anti-Ulk1, anti-phospho Ulk1(Ser757), anti-LC3, or anti-GAPDH antibodies (Cell Signaling Inc.) at 1:1000 dilution at 4°C overnight. Anti-Ulk1 for mouse and human were purchased from Sigma and Cell Signaling Inc., respectively. Membranes were washed in TBST and incubated with anti-rabbit HRP-conjugated secondary antibody (1:10,000) in 5% milk in TBST for 1 hr at RT. Membranes were washed in TBST followed by a final wash in TBS. Signals were visualized with ECL reagent (Pierce) and an LAS-4000 luminescent image analyzer (Fujifilm). ImageJ software was used to subtract background and to perform densitometry.
Brain tissues were obtained from Albany Medical Center (AMC) and the National Institute of Child Health and Human Development (NICHD) brain and tissue bank at the University of Maryland at Baltimore. All TSC specimens were obtained from patients with a history of neurological manifestations including seizures. TSC was confirmed by neuropathological diagnosis, including the presence of cortical tubers, subependymal nodules and/or subependymal giant cell astrocytomas (SEGA). TSC samples were obtained from 16 patients, 9 males and 7 females ranging from 13 to 56 years old. Five were surgical specimens and 11 were postmortem tissues with an average post mortem interval of 5 hr. From the 16 patients, 12 tuber samples (5 fixed and 5 frozen from NICHD and 2 fixed from AMC) and 12 SEGA samples (5 fixed and 3 frozen from NICHD and 4 fixed from AMC) were collected. In some cases both tuber and SEGA were collected from the same person. Control cortical tissues were obtained from autopsies of 14 patients who had no history of epilepsy or TSC and died of non-neurological causes. Fifteen samples (5 fixed and 5 frozen from NICHD and 5 fixed from AMC) were acquired with an average postmortem interval of 5.5 hr. One set of fixed and frozen tissues was collected from the same person. All specimens assessed were histologically intact. Brain slices were prepared from paraffin-embedded tissues sectioned at 5 μm. All human tissues were obtained in accordance with a protocol approved by the Albany Medical College Institute Review Board and Committee on Human Research.
Slides were cleared in two rinses of xylene and then rehydrated via ethanol gradient treatment, including two rinses in 100% ethanol, a rinse in 95% ethanol and, finally, a rinse in 70% ethanol. Endogenous peroxidase activity was blocked by quenching in 3% H2O2 and slices were blocked with 10% BSA with 0.4% Triton X-100 for 1 hr at RT. Slices were incubated in primary antibody overnight at 4°C (anti-p62, American Research Products, 1:100 dilution; anti-phospho-S6, Cell Signaling, 1:25 dilution; anti-synapsin, Chemicon, 1:1000 dilution). Slices were again rinsed in PBS and incubated with biotinylated secondary antibody (1:200) for 30 min at RT, followed by another rinse in PBS and incubation in ABC reagent (Vector Vectastain Elite kit) for 30 min at RT. Slices were washed and staining was visualized using ImmPact DAB (Vector Lab). Specimens were dehydrated with a reverse ethanol gradient and mounted in DPX mountant. NIH ImageJ was used for image analysis. For quantification of p62 in human samples, 20× images were selected from 5 regions which contained the highest number of cells with a positive staining for each slice. All cells which were readily identified as individual cells via the presence of a recognizable soma were analyzed. Regions of interest were defined by manually outlining a cell, and the mean intensity was measured. Additionally, 5 areas on an image which contained no distinct staining were measured and the average intensity served as the background for each individual image. A cell was counted as positively stained only if it had an intensity that was greater than or equal to 1.5 times the background level.
Mice were transcardially perfused and sectioned as described above. Slices were adhered to coated coverslips and rinsed in H2O. For cresyl violet staining, slides were incubated in 0.1% cresyl violet acetate dissolved in H2O for 3 min. For H&E staining, slides were submerged in Gill 3 hematoxylin (Richard-Allan Scientific) for 5 min, rinsed in H2O for 5 min and differentiated in 1% acid alcohol for 30 seconds. Slides were then rinsed in H2O for 1 min, submerged in 0.2% ammonia and then again rinsed in H2O. Slides were dipped into 95% ethanol and treated with Eosin-Y (Richard-Allan Scientific) for 1 min. Finally, the specimens were dehydrated via ethanol gradient, rinsed in xylene and mounted with DPX.
Detection of apoptosis in Atg7KO mouse brains by TUNEL staining was performed using a NeuroTACS II kit (Trevigen). Brain sections of 40 μm were adhered to coated glass slides and rinsed in PBS for 10 min. NeuroPore solution was added for 25 min at RT. Slices were rinsed in PBS and quenched in 3% H2O2 for 5 min. A labeling reaction mix consisting of TdT dNTP, TdT enzyme, TdT labeling buffer, and Mn2+ cation was applied to sections for 60 min in a humidified chamber. Brain slices that were treated with nuclease served as a positive control. The reaction was terminated using stop buffer and the slides were washed in PBS. DAB was used to visualize apoptotic cells. Slides were rinsed in H2O, counterstained, dehydrated in a reverse ethanol gradient followed by xylene and mounted in DPX.
Spontaneous seizure activities were detected by continuous video monitoring of 16 Atg7KO, 10 TSC1KO, and 10 PTENKO mice from 8PM until 8AM 5 days/week starting at 3 weeks of age and continuing to 10 weeks for TSC1KO and PTENKO and 20 weeks for Atg7KO. Videos were analyzed by trained researchers, and behavioral seizures which presented with forepaw clonus, rearing, rearing with falling or wild running were identified. Videos were examined from 3 weeks of age until a behavioral seizure was identified, at which point the animal was counted as positive for spontaneous seizures. We acknowledge that this method may fail to identify subtle seizure events or seizures without behavioral manifestation, which may result in underestimation of the actual onset of seizures. Five mice from each group with positive behavioral seizures were further monitored for electrographic seizures with electroencephalography (EEG). Mice were sedated using isoflurane inhalation and placed into a stereotaxic chamber. Three electrodes were linked to stainless steel surgical screws (J.I. Morris). Two screws were implanted at AP −1.8, ML 1.8 (left and right) serving as differential recording electrodes. A third screw was placed into the skull serving as a ground. Electrodes were held in place using dental cement (Harvard Apparatus). Mice were given 5 days to recover prior to recording. EEG recordings were performed in free-moving mice hooked to a Pinnacle 8200 EEG system (Pinnacle Technology) along with continuous video monitoring. EEG data were analyzed with the Sirenia Seizure program by Pinnacle and seizures were manually identified by characteristic high-frequency and high-amplitude firing. Additionally, most seizures were followed by brief post ictal suppression. All electrographic seizures were verified behaviorally by video recording.
For experiments determining LC3 flux, mice were sacrificed and brains were removed, quickly placed into ice-cold dissection buffer (75 mM sucrose, 87 mM NaCl, 25 mM NaHCO3, 7 mM MgCl, 1.25 mM NaH2PO4, 2.5 mM KCl, 25 mM D-glucose and 0.5 mM CaCl2) and subsequently mounted onto a block for sectioning. Using a vibratome, slices were cut in an ice-cold slurry of dissection buffer while bubbled with oxygen. After cutting, the slices were quickly transferred to separate dishes containing ACSF (125 mM NaCl, 25 mM NaHCO3, 1 mM MgCl, 1.25 mM NaH2PO4, 2.5 mM KCl, 10 mM D-glucose and 1 mM CaCl2) bubbled with oxygen with or without 100 μM chloroquine diphosphate (MP Biomedical). Slices were incubated for 2 hr at room temperature and hippocampal tissues were collected using a 2 mm hole punch. Lysates for western blot were prepared as described above. LC3 flux was calculated by taking the ratio of LC3II/LC3I in chloroquine-treated versus control slices. Because ECL signals for LC3II were very weak and for LC3I very robust, digital images for LC3 western blots were acquired by multiple exposures ranging from 5 sec to 2 min. Images with a short and long exposure were used for quantification of LC3I and LC3II, respectively.
Statistics were performed using GraphPad Prism and include t-tests for comparisons between 2 groups and one-way ANOVA with a Tukey post hoc for 3 or more groups.
We first set out to determine if autophagy is impaired in mouse brains under conditions where mTOR is hyperactivated. We utilized TSC1 and PTEN conditional KO mice by crossing TSC1flox/flox (Kwiatkowski et al., 2002) or PTENflox/flox (Groszer et al., 2001) mice with a CaMKIIα-Cre mouse line (Tsien et al., 1996), in which Cre recombinase is robustly expressed in neurons within the cortex and hippocampus, key brain structures involved in epileptogenesis (Meyer and Beck, 1955). Using video recording, we found that both TSC1KO and PTENKO mice began to display behavioral seizure activity as early as 5 weeks of age, with 80–90% of mice developing seizures by week 10 (Fig 1a). No seizures were observed in control TSCflox/flox, PTENflox/flox, or CaMKIIα-Cre mice. TSC1KO and PTENKO mice showed a significant decrease in survival rate, with a steep drop occurring between postnatal weeks 6–8 (Fig 1b), indicating a positive correlation between the onset of seizures and mortality. Video recording confirmed that all deaths which occurred during recording or during routine handling immediately followed severe seizures (accounting for 40% and 50% of total mortality for TSC1KO and PTENKO, respectively), suggesting that seizures are primarily responsible for the mortality. To verify that the behavioral seizure events observed in TSC1KO and PTENKO mice also presented with electrographic abnormalities, we selected 5 pairs of TSC1KO and PTENKO mice with behavioral seizures for cortical EEG and verified that the seizures were presenting with increased amplitude and frequency of firing (Fig 1c & 1d). Western blot analysis revealed that TSC1KO mice at postnatal day 28 started to display higher levels of mTOR activity as indicated by phosphorylation of S6 at Ser235/236 than that in age-matched controls, (Fig 1e & 1f). The level of p-S6 in TSC1KO mice is also slightly elevated at postnatal day 28, compared with TSC1KO mice at postnatal day 14 and 21, but did not reach a statistical significance. This indicates that the epileptogenic window is a 2–3 week period beginning at postnatal week 4. We also observed that the level of p-S6 at postnatal day 28 in wild-type mice was reduced compared with postnatal day 14 and 21. These data are consistent with a recent study that reported that mTOR activity is lower in the mature brain than in the immature brain (Talos et al., 2012).
Ulk1 is a critical component that mediates initiation of autophagy (Kim et al., 2011). A recent study revealed that mTOR phosphorylates Ulk1 at Ser757 (p-Ulk1 Ser757), resulting in inhibition of autophagy (Kim et al., 2011). We therefore examined if Ulk1 phosphorylation at Ser757 is altered in vivo following genetic disinhibition of mTOR via TSC1 or PTEN deletion. We found that both TSC1KO and PTENKO mice displayed a significant increase in Ulk1 phosphorylation at Ser757 in both cortex and hippocampus when compared with wild-type litter mates, indicative of reduced autophagy activation (Fig 2a & b & 3a & b). Phosphorylation of S6 at Ser235/236 is also elevated in TSC1KO and PTENKO mouse cortex and hippocampus, consistent with mTOR hyperactivation (Fig 2a & c & 3a & c). Neither S6 nor Ulk1 displayed any change in phosphorylation in the cerebellum (Fig 2a–c & 3a–c), where Cre expression is minimal (Tsien et al., 1996). The mTOR inhibitor rapamycin reduced the levels of phosphorylated Ulk1 concurrent with a reduction in S6 phosphorylation (Fig 2a–c & 3a–c), confirming that increased Ulk1 phosphorylation reflects activation of mTOR due to inactivation of TSC1 and PTEN.
Modification of microtubule-associated protein 1 light-chain 3 (LC3) from LC3I to LC3II is another marker of autophagy activity (Tanida et al., 2008). The basal level of autophagy in the brain was very low, as little LC3II was detected (Fig 2d & 3d). To increase the sensitivity of detection, we monitored autophagy flux by blocking degradation of LC3II, a strategy that has been used previously (Kaushik et al., 2011; Zois et al., 2011). We employed chloroquine (CQ) to attenuate degradation of LC3II by inhibiting lysosomal activity (Zois et al., 2011). We observed a significant accumulation of LC3II in hippocampal tissues prepared from control mice, whereas no accumulation occurred in TSC1KO (Fig 2d & e) or PTENKO (Fig 3d & e) mice. These data indicate that autophagy activity is down-regulated in TSC1KO and PTENKO mouse brains.
We next extended our study to brain tissues from human TSC patients. TSC is an autosomal dominant disorder caused by mutations in either the TSC1 or TSC2 gene (European Chromosome 16 Tuberous Sclerosis, 1993; van Slegtenhorst et al., 1997; Crino et al., 2006) and characterized by severe neurological manifestations, including cognitive dysfunction (Prather and de Vries, 2004) and a high incidence of early-onset, intractable epilepsy (in 60–90% of patients) (Thiele, 2004; Holmes and Stafstrom, 2007). Typical pathological changes include cortical tubers (Yamanouchi et al., 1997), subependymal nodules, and SEGA (Trombley and Mirra, 1981). We found that both tubers and SEGA displayed a significant increase in phosphorylation of Ulk1at Ser757 (Fig 4a–b) along with an increase of phosphorylated S6 when compared with control cortical tissues (Fig 4a & c). p62/SQSTM1 is an ubiquitin-binding protein and a substrate of autophagy; therefore, its levels generally reflect autophagy flux (Komatsu et al., 2006; Ichimura and Komatsu, 2010). Western blotting revealed significant accumulation of p62 in cortical tubers and SEGA (Fig 4a & d). Immunohistological staining revealed strong p62 staining in morphologically distinct cell types: enlarged, dysmorphic giant ganglion-like cells in cortical tubers, and spindle-shaped astroglial-like cells in SEGA (Fig 4e & f). In contrast, only a few weakly stained p62 positive cells were found in control cortical brain specimens. In addition, phosphorylation of S6 was also slightly increased in giant cells and astroglial-like cells (Fig 4e). Amorphic and fairly weak staining of synapsin was also observed in both tubers and SEGA, perhaps reflecting the loss of neuronal identity and defects in synapse formation (Fig. 4e).
We next investigated whether direct inhibition of autophagy is sufficient to alter seizure susceptibility. Atg7 is an E1-like enzyme which mediates conversion of LC3I to LC3II in the autophagy pathway (Yang and Klionsky, 2010). To address the role of autophagy in epileptogenesis, Atg7flox/flox mice were crossbred with CaMKIIα-Cre mice. Atg7flox/flox; CaMKIIα-Cre (Atg7KO) mice were viable with normal body weight at birth when compared to their wild-type littermates. Although LC3II levels were low in control mice, we found that LC3II nearly disappeared in the cortex and hippocampus of 6–7 week-old Atg7KO mice, but remained in the cerebellum (Fig 5a & b). p62 was barely detectable in control mice, but, in stark contrast, was markedly accumulated in cortical and hippocampal tissues of Atg7KO mice (Fig 5a & c). We also monitored phosphorylation of S6 to determine if inactivation of autophagy by Atg7 deletion has any impact on mTOR activation. We did not find any significant change in p-S6 (Fig 5a & d). Immunohistological staining of p62 revealed accumulation throughout the forebrain, with a pronounced increase in cortical neurons, hippocampal CA1 and CA3 pyramid cells, and a few cells in the dentate gyrus (Fig 5e). Together, these data confirmed that ablation of Atg7 results in inhibition of autophagy in the cortex and hippocampus.
In previous studies, mice with broad deletion of either Atg7 or Atg5 in either neural progenitors (using a nestin-Cre line) or in cerebellar Purkinje neurons (using the Pcp2-Cre line) displayed neurodegeneration, but neither of them were reported to develop spontaneous seizures (Hara et al., 2006; Komatsu et al., 2006; Komatsu et al., 2007). To ascertain whether inhibition of autophagy specifically in relatively mature forebrain neurons can lead to epilepsy, we employed video monitoring to document spontaneous recurrent seizures in 6–12 week-old Atg7KO mice. We found that approximately 75% of KO mice developed spontaneous behavioral seizures (Fig 6a), which began as early as 6–7 weeks of age, but the majority of mice developed seizures around postnatal weeks 8–10. Seizures lasted an average of approximately 30 seconds and presented with forepaw clonus, rearing, rearing with falling, and periods of wild running which were occasionally followed by a tonic phase and death. We further selected 5 Atg7KO mice with positive behavioral seizures for cortical EEG and verified that the seizures were presenting with abnormal electrographic activity (Fig 6c & d). Additionally, Atg7KO mice displayed reduced survival, beginning around 6–8 weeks of age, eventually falling to approximately 75% survival by 20 weeks (Fig 6b). Of note, all mortality which was caught during video/EEG monitoring or routine handling was confirmed to be from sudden death immediately following a severe seizure episode, suggesting that most mortality was likely associated with seizures.
Previous studies have shown that inactivation of either Atg7 or Atg5 in neuronal progenitor cells leads to neurodegeneration (Komatsu et al., 2006); however, our Atg7flox/flox; CaMKIIα-Cre mice at 6 to 7 weeks old did not display any overt neuronal loss based on H&E or cresyl violet staining, nor any significant increase in TUNEL staining (Fig. 7a & 7b).
Hyperactivation of mTOR is generally recognized as the mechanism underlying epileptogenesis in rodent genetic TSC1 and PTEN models as well as in human TSC patients (Baybis et al., 2004; Meikle et al., 2008; Zeng et al., 2008; Zhou et al., 2009). The mTOR signaling pathway regulates a myriad of cascades, including protein translation, Akt activity, autophagy and others (Laplante and Sabatini, 2012), however, the exact contribution to epileptogenesis from these downstream mechanisms remains elusive. Here, we showed that autophagy is inhibited in the brains of TSC1KO and PTENKO mice as well as in human TSC patients. Furthermore, mice deficient in autophagy activity through deletion of Atg7 in the forebrain developed spontaneous seizures. Therefore, our data suggest that impaired autophagy is a contributing mechanism to epileptogenesis in response to hyperactivation of mTOR.
The importance of autophagy within neurons has long been controversial, as in vivo analysis has demonstrated very low levels of autophagy in the brain, even following prolonged nutrient starvation (Mizushima et al., 2004). Since the development of conditional Atg7 and Atg5 knockout mice, it has been repeatedly demonstrated that, despite low basal levels, autophagy plays a critical role in the central nervous system, regulating maintenance of axons, neurodegeneration, and mitophagy (Komatsu et al., 2005; Hara et al., 2006; Komatsu et al., 2006; Rubinsztein, 2006; Komatsu et al., 2007). Our results are consistent with previous findings, as we detected very low levels of LC3II in the brain (Fig 5a & b). However, we demonstrate here for the first time that inhibition of autophagy alone is sufficient to promote the development of spontaneous recurrent seizures when Atg7 is deleted specifically in relatively mature neurons (Fig 6). Of particular note, in previous studies, mice with broad deletion of either Atg7 or Atg5 in neuron progenitors displayed neurodegeneration, but were not reported to develop spontaneous seizures (Hara et al., 2006; Komatsu et al., 2006; Komatsu et al., 2007). In contrast, in the present study, deletion of Atg7 in mature neurons did not result in overt neurodegeneration detectable by TUNEL staining (Fig 7). It is conceivable that subtle neurodegeneration beyond our detection limits could have occurred. Alternatively, Atg7 deletion may have less of an impact on mature neurons than on immature neurons.
The precise mechanism by which impairment of autophagy leads to neuronal hyperexcitability and seizures remains to be elucidated. Previous studies revealed that Atg7 deletion led to axon degeneration in cerebellar purkinje cells (Komatsu et al., 2007) and interfered with axonal growth in hypothalamic neurons (Coupe et al., 2012). However, a very recent study reported that deletion of Atg7 protects against axonal degeneration induced by a neurotoxin or axotomy (Cheng et al., 2011). These findings suggest that autophagy may regulate homeostasis of axons in a context-dependent manner. Mossy fiber sprouting is a characteristic pathological change in the hippocampus in epileptic brains (Sutula et al., 1989). It involves an abnormal growth of the mossy fiber axons of dentate granule cells. Although its role in epileptogenesis remains debatable, it will be of interest to see if there is abnormal mossy fiber growth in TSC1flox/flox;CaMKIIα-Cre and Atg7flox/flox;CaMKIIα-Cre mice. Mitophagy is a cellular process by which damaged mitochondria are degraded through the autophagy pathway (Wang and Klionsky, 2011). Autophagy plays an important role in homeostasis of mitochondria (Harris and Rubinsztein, 2011). A recent study reported that embryonic deletion of TSC1 leads to a 10-fold increase in mitochondria (Goto et al., 2011). Therefore, altered mitochondrial homeostasis from hyperactivation of mTOR and impaired autophagy may trigger epileptogenesis. In addition, recent studies revealed that neurons with a TSC1 deletion became sensitive to stress (Di Nardo et al., 2009; Ng et al., 2011). As both the mTOR and autophagy pathways are critical in responding to stress, it is possible that the neurons with impaired autophagy observed in TSC may become less able to cope with stress, resulting in altered neuronal excitability. Finally, a recent study also reported that p62 associates with mTOR complex 1 (mTORC1) and modulates mTOR signaling in response to nutrient availability (Duran et al., 2011). However, we did not find any significant alteration of p-S6 in Atg7KO mice (Fig 5a & d). Thus our data suggest that autophagy impairment leads to spontaneous seizures, independent of mTOR activity secondary to autophagy inactivation in our Atg7KO model.
As a genetically inherited disease, it is expected that TSC function is impaired from birth, and thus TSC mutation exerts its epileptogenic effects either through modification of brain development or through its continual effects on relatively mature neurons. Using the CaMKIIα Cre line, we found that deletion of either TSC1 or Atg7 in relatively mature mouse neurons at 21–28 days of age (Tsien et al., 1996) is sufficient to promote epileptogenesis, supporting the notion that the effects of loss of TSC1 or Atg7 in mature neurons are capable of recapitulating the epileptic phenotype. However, because Cre is expressed around postnatal day 14 to 21 in CaMKIIα Cre line, we can’t necessarily rule out a contribution from a possible alteration in later brain development. In light of our findings, we propose that epileptogenesis in the TSC1 model is not restricted to a developmental time point, but may in fact be a continuous process throughout the course of life. This hypothesis is also clinically relevant. Although the vast majority of TSC patients first develop epilepsy as infants, a small fraction of patients do not show seizures until adolescence or adulthood (Chu-Shore et al., 2010). Our observation provides credence to the concept that pharmacological intervention, by targeting the TSC/mTOR/autophagy pathway, could still be effective at a later point in life in TSC patients or in those with a later onset of seizure manifestation.
TSC impairment affects multiple cell types, including various subsets of neurons and glia. Recent studies revealed that in vivo broad deletion of TSC1 in neurons results in seizures and premature death (Meikle et al., 2008). Mice with TSC1 deletion in astrocytes also developed severe seizures (Zeng et al., 2008). However, interestingly, other studies have shown that TSC1 deletion restricted to GABAergic neurons does not result in spontaneous seizures (Wang et al., 2007; Fu et al., 2011), indicating that the pathogenic effect from TSC1 inactivation may be neuron subtype-specific, i.e., affecting excitatory and inhibitory neurons differently. In the present study, we found that mice with selective deletion of TSC1 mainly in excitatory neurons (using the CaMKIIα-cre line) developed very severe seizures. Thus, our data suggest that dysregulated mTOR in excitatory neurons may play a major role in epileptogenesis. We demonstrated that deficiency of autophagy activity in excitatory neurons also results in epilepsy. Future studies will determine if there are any similar epileptogenic effects of inactivation of autophagy in GABAergic inhibitory neruons and astrocytes.
In the present study, mice with a deletion of TSC1 or PTEN developed much more severe seizures, including earlier onset and higher mortality, than did Atg7KO mice (Fig 1a & b, 6a & b). Given the multiple processes under the control of the mTOR pathway, it is likely that other epileptogenic alterations occur concomitantly following disinhibition of mTOR due to loss of TSC1 or PTEN function. Therefore, our data suggest that impairment of autophagy is likely one of several epileptogenic mechanisms downstream of mTOR. In addition, we observed that the degree of autophagy impairment in Atg7KO mice exceeds that of TSC1KO or PTENKO mice. The possible interpretation is that Atg7 is essential for autophagy whereas mTOR plays a modulatory role. This is exemplified by the extensive accumulation of p62 in Atg7KO mice (Fig 5a, c & e), which is mostly absent in TSC1KO and PTENKO mice (unpublished data). Interestingly, p62 accumulation is striking in human TSC brains. We speculate that interspecies variation may account for the differences in p62 accumulation between human TSC and TSC1KO mice. It may also reflect the time period in which autophagy has been impaired, ranging from 13–56 years in the human TSC patients in the present study versus only a few weeks following TSC1 deletion in mice.
In summary, our data suggest that impaired autophagy is a contributing factor to epileptogenesis in TSC. We believe that future studies to identify other contributing factors downstream of mTOR are a priority. We hope that, through identification of additional pathogenic mechanisms in TSC, it will become possible to devise novel therapeutic strategies to prevent or reverse the epileptogenic modifications brought on by TSC mutation, and ultimately to alleviate seizures while avoiding adverse side effects such as the immunosuppression and growth retardation typically associated with the mTOR inhibitor rapamycin (Anderl et al., 2011). In addition, mTOR activity was found to be elevated in multiple acquired epilepsy models (Buckmaster et al., 2009; Zeng et al., 2009; Huang et al., 2010; Raffo et al., 2011; Talos et al., 2012; van Vliet et al., 2012); therefore, it will be of interest to determine if autophagy is also altered in acquired epilepsy.
This work was funded by grants from NIH/NINDS R01NS062068 (Y.H.), NARSAD Young Investigator Award (Y.H.), NIH/NINDS R01NS060123 (Z.Y.), and NIH/NIDDK K08DK088950 (X. Z.). We thank Christine Sheehan and Ann Dence for assistance with brain slide preparation. Human tissues were obtained from the NICHD Brain and Tissue Bank for Developmental Disorders at the University of Maryland, Baltimore, MD.
The authors declare no competing financial interests.