Spatiotemporal progression of ubiquitin-proteasome system inhibition after status epilepticus suggests protective adaptation against hippocampal injury

The cell and molecular mechanisms by which injury to the brain leads to the later emergence of chronic, spontaneous seizures, a process termed epileptogenesis, remains incompletely understood [1]. Epileptogenesis is associated with large-scale changes to gene expression which is thought to impact on neuronal survival and to drive remodelling of neuronal networks, aberrant neurotransmitter receptor expression and function, gliosis, neuroinflammation and other characteristic changes [2]. It is increasingly recognized that targeting single genes is unlikely to be sufficient to disrupt epileptogenesis and ways to target larger signalling networks are required [1]. This has driven a strong focus on transcription factors and epigenetic mediators [3], systems genetics approaches [4] and post-transcriptional gene silencing by non-coding microRNAs [5]. This is apt because it is mRNA transcription that provides the greatest contribution to the abundance of proteins during stress [6, 7]. In contrast, protein turnover has been reported to contribute less than 10% to overall protein abundance [7]. Nevertheless, protein aggregates and dysfunction of the molecular machinery responsible for regulating protein turnover is evident in numerous neurodegenerative and neurologic disorders [8‐10]. Thus, protein turnover may be especially important for the pathogenesis of diseases associated with the brain. Whether altered protein turnover is important in epileptogenesis is unknown.

The ubiquitin proteasome system (UPS) is the major intracellular pathway leading to the degradation of aberrant and/or misfolded soluble proteins [11]. The UPS is a highly conserved and tightly controlled ATP-consuming proteolytic system. Ubiquitin (Ub) is first activated by an E1 enzyme, and then transferred to an ubiquitin-conjugating E2 enzyme [11]. Substrates to be degraded are selectively recognized by E3 Ub ligases [12]. E3 ligases mediate the subsequent transfer of activated Ubs from E2 to the substrate leading to an increasing polyubiquitination chain [11]. Once Ub moieties reach four or more per substrate chain, the substrate is usually recognized by the 26S proteasome [13]. The 26S proteasome is composed of two 19S regulatory protein complexes important for substrate recognition and deubiquitination and a barrel-like shaped 20S protein complex with a catalytic core in charge of substrate cleavage into smaller peptides and amino acids [14].

The UPS is present and functions in multiple sub-cellular compartments including cytoplasm, nucleus, mitochondria, cell membranes and pre- and post-synaptic compartments in neurons [15‐17]. In the nervous system, the UPS has been shown to regulate turnover of proteins critical for neuronal survival and neurotransmission [16, 17] by controlling the presence of functional glutamate [18] and γ-amino butyric acid (GABA) receptor levels [19], spinogenesis [20], dendrite growth and arborisation [21] and the formation of new synapses [22]. Proteasome inhibition has been proposed to play a causative role in numerous acute and chronic diseases of the brain including ischemia [10] and neurodegenerative diseases such as Alzheimer’s, Huntington’s and Parkinson’s [9, 23]. Consequently, UPS-targeting drugs have been proposed as potential therapeutic treatment strategies for various brain diseases [9].

Emerging evidence suggests the function and dysregulation of the UPS is important in epilepsy. Glutamate-induced excitotoxicity leads to an inhibition of the proteasome [24]. Genes linked to ubiquitin metabolism are prominent among transcripts regulated by seizures in animal models [25]. The deubiquitinating enzyme USP9X regulates excitability and seizures in animals and humans [26]. Genetic or pharmacologic targeting of the E3 ubiquitin ligase CRL4A (CRBN) increases seizure susceptibility [27]. The immunoproteasome, a particular form of the proteasome believed to play a key role during inflammation, is upregulated in brain tissue from patients with temporal lobe epilepsy (TLE) [28] and in the hippocampus of rats with chronic epilepsy [29]. Lafora disease, a progressive myoclonic epilepsy characterized by the presence of glycogen-like intracellular inclusions named Lafora bodies, results from mutations in the ubiquitin E3 ligase malin [30]. In addition, 20S proteasome subunits have been shown to be sequestered in Lafora bodies [31]. Mutations in the E3 ligase Ube3a have been shown to cause Angelman syndrome, a rare neurological disorder commonly associated with intractable seizures [32] and the E3 ligase MDM2 controls levels of the transcription factor p53 which regulates cell death and seizure thresholds in experimental and human epilepsy [33, 34].

However, up to date there have been no studies carried out to characterize where exactly UPS inhibition occurs after seizures in vivo, which cell-types are involved and whether proteasome inhibition impacts on seizure-induced brain damage. Here we used a UPS inhibition green fluorescent protein (GFP)-reporter mouse [35] to map UPS inhibition in vivo following status epilepticus. We also tested a proteasome inhibitor for effects on seizure-induced neuronal death.

In the present study we show an increased accumulation of polyubiquitinated proteins in the hippocampus after prolonged seizures and during epilepsy suggesting a seizure-induced inhibition of the UPS. Most strikingly and unexpectedly, seizure-induced UPS impairment was mainly evident in brain areas resistant to seizure-induced cell death including neurons and astrocytes. In line with a neuroprotective role of proteasome inhibition during seizures, mice treated with the specific proteasome inhibitor epoxomicin displayed less neurodegeneration. These findings support proteasome inhibition as an endogenous protective mechanism against seizure-induced cell death.

Despite differences in disease etiology (e.g. β-amyloid, polyglutamine expansion, neurodevelopmental abnormalities and other) several chronic brain disorders share common clinical symptoms (e.g. cognitive deficits, seizures and psychological problems such as depression and anxiety) implying the emerging concept of existing shared pathological pathway activation among different brain diseases [48‐50]. Proteasome inhibition and the accumulation of polyubiquitinated aggregates are a common characteristic of chronic neurodegenerative diseases such as Alzheimer’s and Huntington’s [9, 51] and an impairment of the UPS has also been reported after acute insults to the brain such as stroke and traumatic brain injury [10, 52]. Likewise, emerging evidence suggests a dysfunction of the UPS in neurological diseases such as epilepsy [29]. However, previous studies lacked evidence showing a direct link between seizures and proteasome inhibition and whether seizures lead to an inhibition of the proteasome in vivo has not been fully established. Studies were limited to determining the expression levels of specific subunits of the immunoproteasome [28, 29] or to identify the impact certain mutations of genes involved in UPS functions have on pathology (e.g. different E3 ligases, deubiquitinating proteins) [26, 31]. Recent in vitro experiments have suggested that glutamate-induced excitotoxicity may lead to UPS inhibition mediated by NMDA receptor activation [24]. We extend these studies here by showing that prolonged seizures lead to an impairment of the UPS with the subsequent accumulation of polyubiquitinated substrates in the brain. We do not know what causes an inhibition of the proteasome after seizures. In line with in vitro results [24], we have observed a down-regulation of chymotrypsin-like activity in the hippocampal subfield CA1 after KA-induced seizures. This may contribute to the accumulation of polyubiquitinated proteins in the hippocampus after seizures. However, the hippocampal subfield showing the strongest increase of Ub^G76V-GFP reporter protein levels and polyubiquitinated conjugates after status epilepticus in our model was the DG subfield which showed only a marginal inhibition of the catalytic activity of the 20S proteasome. Previous studies have shown a dose-dependent increase in Ub^G76V-GFP reporter in hippocampal neurons treated with epoxomicin [35], therefore, our results suggest that other mechanisms may contribute to proteasome inhibition. Moreover, no differences in 20S proteasome activity was observed at the time of lorazepam administration, further suggesting only a small contribution of the catalytic down-regulation of the 20S proteasome to UPS impairment. Other factors may contribute to the accumulation of polyubiquitinated proteins after prolonged seizures. Status epilepticus leads to a depletion of intracellular ATP [53], possibly impeding E1-mediated Ub activation and compromising a correct functioning of the ATP-consuming 20S proteasome [11]. Increased intracellular Ca²⁺ during excitotoxicity might also lead to a dysfunction of the proteasome [24]; however, other reports have suggested that increased intracellular Ca²⁺ levels lead to increased proteasome activity [54]. Furthermore, status epilepticus leads to an increase in ER-stress [55] and an impairment of autophagy [56], a second major intracellular protein-degrading mechanism, thereby potentially leading to an overload of substrates which exceeds the capacity of the proteasome. Regardless of the mechanism, our results show that seizures lead to an impairment of the UPS with the subsequent accumulation of poly-ubiquitinated substrates in the hippocampus. Interestingly, seizures are a common co-morbidity of many neurodegenerative diseases such as Alzheimer’s and Huntington’s [57, 58] possibly contributing to the accumulation of polyubiquitinated aggregates during these diseases.

Our second major finding was that proteasome inhibition was most evident in hippocampal subfields resistant to seizure-induced cell death. This was unexpected since proteasome inhibition is typically associated with a failure of homeostasis and a driver of neurodegeneration [8]. We do not know why UPS impairment was most evident in these brain regions. However, comparing GFP increases between both brain hemispheres and the different hippocampal subfields, it is tempting to speculate that the level of proteasome inhibition reflects seizure progression throughout the ipsilateral and contralateral hippocampus [40] with the DG being the major site of excitatory input into the hippocampus [59]. While proteasome inhibition was most apparent in neurons at early stages after status epilepticus, astrocytes were the main cell population affected later by UPS impairment. This might be due to the fact that neurons are most likely the cell population activated during the early stages of seizure-induced pathology, while glial responses occur at later stages [60]. Interestingly, proteasome inhibition has been shown to decrease astrogliosis in vitro and in vivo [61]. Furthermore, recent studies demonstrated that astrogliosis alone, in the absence of any other pathology, is sufficient to cause epilepsy in mice [62, 63]. This may suggest a potential anti-epileptic effect provided by proteasome inhibition in glia. UPS impairment in astrocytes is not unique to our seizure model and has been described for other diseases including Huntington’s disease [64]; amyotrophic lateral sclerosis [65] and prion disease [39]. Moreover, astrocytes seem to be well adapted to support long-lasting UPS impairment probably due to the efficient induction of heat shock responses [66].

Why does the ipsilateral CA3 subfield of the hippocampus display a delayed inhibition of the UPS? The CA3 subfield is particularly vulnerable in our model [36] possibly due to the high density of kainate receptors [67]. Although a NMDA-mediated down-regulation of the UPS has been observed after excitotoxicity [24], this has been restricted to the nuclear compartment [24]. On the other hand, increased intracellular Ca²⁺ concentrations have been shown to activate the proteasome [54] which may delay the accumulation of polyubiquitinated conjugates at early time-points after status epilepticus. Another possible explanation for the delayed accumulation of polyubiquitinated proteins in CA3 may be increased inhibition of translation after seizures [68] leading to fewer substrates to be degraded by the proteasome.

Although the spatio-temporal profile of UPS inhibition has not been reported previously after status epilepticus, there have been numerous reports of UPS inhibition after ischemic brain injury [27, 69, 70]. Here, proteasome inhibition was most evident in the hippocampal subfield CA1, a brain region vulnerable to ischemia-induced cell death, whereas only a transient polyubiquitination could be observed in the remaining cell-death resistant hippocampal subfields CA3 and DG [27]. This is different to status epilepticus, where, in contrast to ischemia, cell death-resistant brain regions showed the highest increase in polyubiquitination levels, suggesting a unique pathology according to brain insult. Interestingly, during ischemia the hippocampal subfield CA1 showed the strongest reduction in the catalytic activity of the proteasome [71], similar to findings in our mouse model of status epilepticus, suggesting that some similarities may exist between different conditions. We do not know why the hippocampal subfield CA1 is more prone to a reduction in proteasome activity after seizures and why this downregulation is not accompanied by a bigger increase in polyubiquitinated proteins after status epilepticus when compared to the other hippocampal subfields. Differences in glutamate receptor recruitment during status epilepticus (e.g. NMDA subtype vs. kainic acid or AMPA receptor) with the resulting differences in intracellular calcium levels, may affect proteasome activity [24, 54]. Differences in the subfield-specific clearance of damaged proteins and/or ER-stress during status epilepticus might also contribute to differences in the accumulation of polyubiquitinated proteins.

Why does UPS impairment protect against seizure-induced cell death? The protection is presumably a result of an accumulation of proteins that would otherwise be degraded. However, we do not know their identity. The most likely candidates probably include proteins involved in inflammatory processes such as NFκB which has been shown to mediate proteasome inhibitor-mediated neuroprotection after ischemia [72, 73]. Other proteins such as heat shock proteins [74] or synaptic proteins may also be involved. This could be assessed in the future using Tandem mass spectrometry (MS-MS)-type approaches. Another option may be that accumulated proteins interrupt signalling pathways that would normally be able to progress through to kill neurons or, alternatively, the inhibition of the ATP-consuming UPS may help protect against excessive depletion of energy reserves during seizures.

Emerging evidence suggests that the UPS controls neurotransmission by regulating crucial processes such as synaptic plasticity, spinogenesis and levels of functional neurotransmitter receptors [17]. We have not observed any changes in seizure severity during status epilepticus after epoxomicin treatment. However, a potential impact of the proteasome on seizure generation or seizure severity might be more important under chronic conditions. Epileptic mice in our study showed increased polyubiquitination levels in the hippocampus suggesting a chronic impairment of the UPS which may contribute to hyperexcitability. In line with this, recent data has shown a more severe epileptic phenotype in rats treated with the proteasome inhibitor MG-132 [75], which, in contrast to epoxomicin, also inhibits calpain activity [76]. While acute/transient UPS inhibition may be protective, long-lasting inhibition may be neurotoxic. Indeed, inhibition of the UPS has been reported to elicit neurodegeneration directly [8, 45]. Therefore, it is unlikely proteasome inhibitors will be used as neuroprotective or anti-epileptic drugs. However, they provide important insight into molecular pathogenesis and, if we can identify the affected pathways/proteins involved, this may lead to novel approaches for the treatment of status epilepticus and epilepsy.

Our data show that prolonged seizures lead to an impairment of the UPS which is most prominent in brain areas resistant to seizure-induced cell death suggesting proteasome inhibition as an endogenous protective mechanism.