Nuclear and degradative functions of the ESCRT-III pathway: implications for neurodegenerative disease

ABSTRACT

The ESCRT machinery plays a pivotal role in membrane-remodeling events across multiple cellular processes including nuclear envelope repair and reformation, nuclear pore complex surveillance, endolysosomal trafficking, and neuronal pruning. Alterations in ESCRT-III functionality have been associated with neurodegenerative diseases including Frontotemporal Dementia (FTD), Amyotrophic Lateral Sclerosis (ALS), and Alzheimer’s Disease (AD). In addition, mutations in specific ESCRT-III proteins have been identified in FTD/ALS. Thus, understanding how disruptions in the fundamental functions of this pathway and its individual protein components in the human central nervous system (CNS) may offer valuable insights into mechanisms underlying neurodegenerative disease pathogenesis and identification of potential therapeutic targets. In this review, we discuss ESCRT components, dynamics, and functions, with a focus on the ESCRT-III pathway. In addition, we explore the implications of altered ESCRT-III function for neurodegeneration with a primary emphasis on nuclear surveillance and endolysosomal trafficking within the CNS.

KEYWORDS:

• Amyotrophic lateral sclerosis
• endolysosomal trafficking
• ESCRT-III pathway
• frontotemporal dementia
• neurodegenerative diseases
• nuclear pore complex
• nuclear surveillance
• protein degradation

Introduction

The endosomal sorting complex required for transport (ESCRT) is a multi-functional membrane-remodeling complex present in archaea and eukaryotes [1–9]. The ESCRT pathway plays a pivotal role in multiple cellular processes including nuclear envelope repair, nuclear pore complex (NPC) quality control, plasma membrane repair, lysosome repair, intraluminal vesicle (ILV) and multivesicular body (MVB) formation, endolysosomal trafficking, neuronal pruning, and exosome biogenesis and release (Figure 1) among others as has been recently described [4,5,10–30]. Notably, alterations in cell biological processes linked to ESCRT function, specifically nuclear surveillance and endolysosomal trafficking/autophagy, have been highlighted through studies in various model systems of neurodegenerative diseases including Amyotrophic Lateral Sclerosis (ALS) and Frontotemporal Dementia (FTD) [31–39]. Collectively, this work implicates ESCRT dysfunction as a significant contributor to ALS and FTD pathogenesis.

Figure 1. Overview of a subset ESCRT-III mediated cellular processes. A subset of known cellular processes that involved ESCRT-III membrane remodeling functions depicted in neurons.

Neurodegenerative diseases such as ALS and FTD are characterized by the selective death and dysfunction of neuronal and glial cell populations within distinct CNS regions. Specifically, neurons and glial cells within the motor circuitry of the motor cortex and spinal cord affected in ALS and neuronal and glial cells within the frontal and temporal cortices impacted in FTD. Given the cell type and CNS region specificity of these neurodegenerative diseases, ALS leads to a deterioration of motor function and FTD impacts decision-making, personality, behavior, and language [40–49]. Although largely considered distinct neurodegenerative diseases, ALS and FTD share a number of genetic and pathological underpinnings. For example, mutations in C9orf72 and CHMP2B are two examples of genetic mutations causative of both ALS and FTD [50–56]. Interestingly, amongst patients with the C9orf72 mutation, a subset (~10–15%) are diagnosed with both ALS and FTD [44,57–64]. Pathologically, nuclear clearing and associated nuclear loss of function as well as subsequent cytoplasmic mislocalization and/or aggregation of the RNA binding protein TDP-43 has been observed in the majority of ALS cases and ~50% of FTD cases [65–77]. Collectively, studies in various model systems of ALS and FTD have documented impairments in multiple cellular processes including protein degradation, nucleocytoplasmic transport, and RNA processing [39,57,73,78–89]. Intriguingly, the ESCRT-III pathway and its protein constituents play a pivotal role in both the direct and indirect regulation, coordination, and execution of cellular processes disrupted in ALS and FTD, highlighting a role for the ESCRT-III pathway in disease pathogenesis.

In this review, we will discuss the function of the ESCRT-III pathway in nuclear membrane sealing and repair, nuclear pore complex surveillance, and endolysosomal trafficking and degradation, given the documented importance and relevance of these cell biological processes to neurodegeneration, namely ALS and FTD. As the fundamental biology and dynamics of the ESCRT-III machinery has been recently and thoroughly reviewed elsewhere [4,5,8,21], the central focus of our review will be on the implications of disruptions in specific ESCRT-III functions for ALS and FTD pathogenesis. Where appropriate, we will discuss studies directly implicating ESCRT-III dysfunction in models of ALS and FTD. This review will emphasize the fundamental and pathologic involvement of the ESCRT-III nuclear surveillance pathway in the central nervous system (CNS) given that a number of recent studies have documented nuclear pore complex and nuclear envelope alterations in neurodegenerative diseases [90–97]. Lastly, we will discuss the role of mutations in the ESCRT-III protein CHMP2B in FTD and ALS.

Overview of ESCRT proteins and dynamics

The ESCRT machinery is comprised of five core subcomplexes: ESCRT-0, ESCRT-I, ESCRT-II, ESCRT-III, and VPS4. These ESCRT complexes can physically interact but have defined functional roles within the cell. In total, over 30 individual proteins make up the mammalian (over 20 in yeast) ESCRT machinery [2–8,98–101]. Briefly, the mammalian ESCRT-0 complex is comprised of two proteins HRS and STAM and plays an essential role in multivesicular body (MVB) formation by binding multiple ubiquitin molecules and localizing ubiquitinated proteins to the endosome [13,102–104]. VPS23/TSG101, VPS28, VPS37A, VPS37B, VPS37C, VPS37D, MVBB, and UBAP1 make up the mammalian ESCRT-I complex and collectively function as a bridge between ESCRT-0 and ESCRT-II, handing off ubiquitinated proteins from ESCRT-0 to ESCRT-II. The ESCRT-II complex, made of VPS22, VPS25, and VPS36 in mammals, is essential for delivering ubiquitinated proteins to endosomes [2,4,8,105].

The ESCRT-III complex is perhaps the most versatile and critical of the ESCRT machinery as it functions in all ESCRT processes (Figure 1). It is comprised of ~12 proteins in mammals including CHMP1A, CHMP1B, CHMP2A, CHMP2B, CHMP3, CHMP4A, CHMP4B, CHMP4C, CHMP5, CHMP6, CHMP7, and IST1/CHMP8. Unlike other ESCRT complexes, the ESCRT-III machinery is only transiently activated or assembled. Assembly begins when early acting factors such as ESCRT-I/-II, the ESCRT associated protein ALIX, or the ESCRT-II/-III hybrid protein CHMP7 bind adapter proteins, ubiquitinated cargo, and/or membranes to cluster cargo and vesicles and initiate membrane bending. In turn, this facilitates ESCRT-III subunit assembly and filamentation along bound membranes [4,7,98,106]. When ESCRT-III assembly is not required, aberrant polymerization of ESCRT-III filaments is prevented by autoinhibition via ‘closed’ conformations of ESCRT-III protein monomers [107–111]. The closed-protein conformation is dependent upon the C terminus as removal of the C terminal region converts ESCRT-III proteins into an ‘open’ conformation and enables subsequent polymerization [112]. This conformational change is essential for ECSRT-III protein assembly at target membranes and in turn ESCRT-III protein function. Under basal cellular conditions, the conversion from a ‘closed’ to ‘open’ state to remove autoinhibition is facilitated by protein – protein interactions. This activation can either occur via interactions between individual ESCRT-III subunits or resident proteins within target membranes [98,99,113–117].

ESCRT-III proteins are recruited in a sequential manner and assemble into multi-protein polymers that undergo subunit exchange to facilitate membrane remodeling [118–124]. Recent evidence suggests that membrane remodeling can occur in both an inside-out (toward the cytoplasm) and outside-in (away from the cytoplasm) manner. The binding of ESCRT proteins to membranes stabilizes membrane curvature thereby allowing the formation of membrane necks. Membrane scission for repair and remodeling is then catalyzed by recruitment and function of the AAA – ATPase VPS4 [4–6,20,98,105,109,120,122,123,125–127]. In addition, VPS4 facilitates ESCRT-III polymer disassembly [118,121,122,128–133]. VPS4 recruitment is facilitated by direct interactions with a microtubule interacting and transport domain interacting motif (MIM) sequence in the C terminus of most ESCRT-III proteins [98,133,134]. However, as is the case for CHMP7 due to an amino acid charge substitution, not all ESCRT-III proteins are capable of recruiting VPS4 through this MIM domain sequence [135]. Once recruited, VPS4 catalyzes ATP hydrolysis to exchange ESCRT-III protein subunits from polymer filaments, initiating their ‘deactivation’ and recycling [118,121,122,128–133]. For a more in depth and comprehensive review of ESCRT-III protein domain structure and interactions, we refer readers to a recent publication by McCullough and colleagues [4]. However, it is important to note, especially in considering mechanisms underlying ESCRT-III dysfunction in neurodegenerative disease, that the sequence of recruitment events and requirement of individual ESCRT-III proteins for function is not known for every membrane throughout the cell. Additionally, it is unknown whether the sequence of events and protein requirements is conserved across organisms and cell types. In fact, recent work has established fundamental differences in the requirement of specific nuclear envelope proteins required for CHMP7/ESCRT-III function in nuclear surveillance in human neurons compared to yeast and non-neuronal immortalized mammalian cell lines [90,93,114–117].

Nuclear membrane sealing and repair

Maintenance of nuclear envelope integrity is essential for maintaining cellular compartmentalization and separation of genetic material from the cytosol of eukaryotic cells. Impaired nuclear envelope integrity can result in the uncontrolled exchange of cellular materials, DNA damage, and ultimately impaired cellular function [136]. Pathologic alterations to nuclear envelope integrity or mutations in resident nuclear envelope proteins have been linked to multiple diseases including tauopathies, laminopathies, ataxia, and muscular dystrophy [96,137–143]. In addition, during cell migration or in the case of cancer cell metastasis, cells navigate through confined spaces leading to nuclear constriction and rupture [22]. Thus, mechanical stress to the nucleus or via forces exerted through the Linker of the Nucleoskeleton and Cytoskeleton (LINC) complex via actin, microtubules, and intermediate filaments within the cytoplasm can compromise nuclear integrity [144–148].

In order to prevent catastrophic loss of nuclear integrity and uncontrolled nuclear ruptures, the ESCRT-III pathway safeguards the nuclear periphery by sealing holes within nuclear membranes in coordination with resident nuclear envelope proteins [22,25,26,29,119,149–151] (Figure 2). The ESCRT-II/ESCRT-III hybrid protein CHMP7 plays a fundamental role in initiating ESCRT-III function in nuclear envelope sealing (Figure 2). CHMP7 passively diffuses into the nucleus where it is then appropriately positioned to interact with the inner nuclear membrane protein LEM2/LEMD2 [30,114–117,152]. In yeast, direct binding to membrane lipids is required for CHMP7’s nuclear localization and interaction with LEMD2 [153]. Once localized to nuclear ruptures, LEMD2 phase separation promotes ‘activation’ of CHMP7 and recruitment of ESCRT-III factors to facilitate nuclear envelope sealing [116]. However, the requirement and contribution of additional ESCRT-III subunits and the cell-type specific requirements of ESCRT-III proteins in nuclear envelope resealing remains unknown. As much research to date has focused on the role of the CHMP7 in nuclear envelope sealing during ‘regulated’ formation of membrane holes during cell division, this is particularly important when considering the role of the ESCRT-III pathway in maintaining nuclear integrity in non-dividing cells such as neurons.

Figure 2. ESCRT-III functions at the nuclear envelope and NPC. (a) Nuclear envelope sealing: upon nuclear envelope rupture, barrier to autoregulation factor (BAF) is recruited to ‘plug’ larger holes prior to sealing. For smaller ruptures, CHMP7 localizes to the nuclear periphery where it is exposed to the inner nuclear membrane protein LEMD2. Protein – protein interactions between CHMP7 and LEMD2 facilitate activation of CHMP7 and subsequent recruitment and polymerization of ESCRT-III subunits to seal nuclear envelope holes. VPS4 recruitment mediates ESCRT-III polymer disassembly and subunit reuse. (b) NPC surveillance and quality control: upon NPC misassembly within the nuclear envelope, CHMP7 localizes to the nuclear periphery where it is positioned to interact with LEMD2. CHMP7 is activated through physical association with LEMD2. ESCRT-III subunits are recruited to generate and stabilize membrane buds. Polymerization of ESCRT-III subunits and recruitment of VPS4 facilitates scission of membrane buds containing NPCs for degradation. VPS4 mediates ESCRT-III polymer disassembly and subunit exchange for reuse.

Substantial evidence suggests that CHMP7/LEMD2 are involved in sealing small (e.g. <100 nm) holes within the nuclear envelope [30,114–117,152,154]. However, repair of larger ruptures requires barrier to autoregulation factor (BAF/BANF1). Upon nuclear rupture, BAF rapidly translocates and binds to DNA, thereby accumulating at the rupture site and effectively ‘plugging’ the nuclear membrane hole. BAF then recruits LEM domain proteins to sites of nuclear rupture to facilitate repair [149]. Interestingly, a recent study suggests that assembly of nuclear envelopes is dependent on CHMP7 in the absence of BAF mediated hole closure [150]. This is consistent with a report demonstrating the formation of ‘grommets’ comprised of ESCRT-III proteins in order to maintain nucleocytoplasmic compartmentalization prior to the sealing of larger nuclear envelope holes during spindle body extrusion during mitosis in yeast [119].

Uncontrolled ESCRT-III function and regulation itself can compromise nuclear integrity and initiate DNA damage [154] highlighting an essential role and delicate balance for ESCRT-III function in facilitating nuclear envelope repair and maintaining nuclear integrity. In addition to the characteristic autoinhibitory conformations of ESCRT-III proteins [107–112], regulatory factors such as CC2D1B and CDK1 prevent unnecessary CHMP7 – LEMD2 interactions and ESCRT-III function at the nuclear envelope [135,155]. Specifically, CC2D1B can bind to CHMP7 and impact the timing of ESCRT-III subunit recruitment and the quality of ESCRT-III polymers during nuclear envelope sealing in cell division [155]. On the other hand, CDK1 can phosphorylate CHMP7 thereby inhibiting interactions with LEMD2 and preventing ESCRT-III polymer assembly during the M phase and exit of cell division [135]. The interaction between CHMP7 and LEMD2 can additionally be influenced by membrane binding, lipid synthesis, and alternative splicing of LEMD2 [152,153,156].

Nuclear membrane invaginations containing NPCs, mRNA, and other proteins have been documented in model systems of FTD and postmortem human FTD tissues [92,95,96,157–159] and Profilin 1 (PFN) centric models of ALS [160]. In FTD, these invaginations appear to localize to areas containing perinuclear aggregates of the microtubule associated protein Tau (MAPT) and reduction of Tau aggregation/accumulation diminishes nuclear abnormalities [95]. While the mechanisms underlying Tau or PFN1 mediated nuclear invagination remain unknown, it is plausible that destabilization of the microtubule or actin cytoskeleton respectively either due to mutations in cytoskeletal proteins or mechanical stress induced by protein aggregation, compromise the integrity and stability of the of the LINC complex leading to nuclear envelope disruption. While it is unclear whether these invaginations contain nuclear perforations, impaired nucleocytoplasmic compartmentalization can also be observed [95,158]. Interestingly, a recent study demonstrated that reduction of BAF promotes the pathologic accumulation of Tau and overexpression of LEMD2 or CHMP7 exerts protection against tau aggregation [159]. However, the impact of BAF, LEMD2, and CHMP7 manipulation on nuclear invaginations and whether their impact on Tau aggregation is mechanistically linked to nuclear envelope integrity directly remains unknown at this time. Nonetheless, collectively, these studies suggest that impaired or altered function of CHMP7 and the ESCRT-III pathway in maintenance of nuclear envelope integrity may contribute to FTD pathogenesis.

Nuclear pore complex quality control

In addition to its role in maintaining nuclear envelope integrity, CHMP7 and the ESCRT-III pathway plays a crucial role in maintaining NPC homeostasis and managing NPC quality control in yeast and human cells [30,93,161–163] (Figure 2). NPCs are comprised of multiple copies of ~30 individual nucleoporin proteins totaling >1000 individual protein molecules per NPC in mammals. These large multi-protein structures function to regulate and coordinate essential cellular processes including nucleocytoplasmic transport and genome organization [164–177]. NPCs assemble in a coordinated manner dependent on the sequential recruitment and insertion of nucleoporin proteins and nuclear pore subcomplexes into the nuclear envelope and partially assembled NPC [169,178–188]. The proper assembly and maintenance of NPC integrity over the lifetime of a cell is essential for continued function and ultimately cellular health.

Recent studies support a role for CHMP7 and the ESCRT-III pathway in ensuring the proper assembly and insertion of NPCs in yeast [30,117]. During NPC surveillance, NPC intermediates, but not mature and properly assembled NPCs, are bound by the LEM family of inner nuclear membrane proteins. This positions them in close proximity to ESCRT-III subunits which are recruited to sites of misassembled or improperly inserted NPCs by LEM family inner nuclear membrane proteins. NPC intermediates are then cleared in a VPS4-dependent manner. Reduction of ESCRT-III subunits compromises NPC surveillance leading to the accumulation of NPC intermediates within storage of improperly assembled nuclear pore complexes (SINC) compartments which are retained in mother cells to safeguard nucleocytoplasmic compartmentalization in daughter cells [30]. Interestingly, these SINCs resemble perinuclear vesicles generated during the nuclear egress of viruses and megaRNP particles [189–191], the process of which has also been linked to ESCRT-III pathway function [2–4,8,12]. Although the precise mechanism by which ESCRT-III clears aberrant NPC intermediates is not fully understood, NPC and nucleoporin degradation have been linked to both the lysosomal (autophagy) and proteasomal degradation pathways [30,115,117,192,193]. ESCRT-III subunits and/or VPS4 may facilitate budding of NPC intermediates and/or directly facilitate stabilization of nuclear envelope invaginations during NPC assembly. However, ESCRT-III surveillance of NPC insertion and assembly was subsequently shown to be coupled to nuclear envelope sealing and linked to the function of the yeast orthologue of CHMP7, Chm7 [117]. Although the mechanisms that lead to CHMP7 influx and ESCRT-III activation upon NPC misassembly remain unknown, one might speculate that improperly assembled NPCs compromise nucleocytoplasmic compartmentalization thereby culminating in CHMP7 nuclear influx and facilitating interactions with LEMD2. Indeed, a recent study has demonstrated that impaired NPC permeability barrier integrity promotes excessive nuclear influx of CHMP7 in ALS neurons [90].

While ESCRT-III surveillance of NPC assembly and insertion has been described in yeast, it is unclear whether analogous mechanisms take place in dividing mammalian cells. Nonetheless, two studies have implicated the ESCRT-III pathway in the maintenance of NPCs throughout the lifetime of non-dividing mammalian cells [93,161,163]. A subset of nucleoporins represent some of the longest lived proteins in the mammalian CNS [194]. In order to sustain nucleocytoplasmic compartmentalization and cellular health, NPCs must be maintained throughout the lifetime of cells, in particular when considering infrequent protein turnover. In non-dividing muscle myoblasts, knockdown of CHMP2A and CHMP3 significantly increased the retention of ‘old’ Nup93 molecules within NPCs [163]. In addition, increased nuclear localization of CHMP7 appears to lead to the abnormal reduction of specific nucleoporins from the NPC in induced pluripotent stem cell (iPSC) derived neurons (iPSNs) [93]. Collectively, these studies implicate at least a subset of ESCRT-III proteins in the piecemeal turnover and degradation of individual nucleoporin proteins from existing NPCs. Although the molecular mechanisms that facilitate and regulate ESCRT-III mediated nucleoporin turnover in fully assembled NPCs remain largely unknown, this currently a topic of active investigation given the potential implications for age-related neurodegenerative diseases where ensuring properly executed nucleoporin molecule turnover is thought to be essential for maintaining NPC functionality.

As has been recently reviewed, nucleoporin and nucleocytoplasmic transport alterations are prevalent in neurodegenerative diseases such as ALS and FTD [195–198]. Multiple studies have now documented diminished nuclear localization or expression of specific nucleoporins in sporadic and genetic forms of ALS and FTD [91,93–95,160]. Recent work has established a role for CHMP7 and the ESCRT-III nuclear surveillance pathway in the pathologic initiation of NPC injury [90,93,161,199] (Figure 3). Using an iPSN model of sporadic ALS (sALS) and C9orf72 ALS/FTD, we demonstrated that abnormal and excessive nuclear localization of CHMP7 precedes the pathologic reduction of specific nucleoporins from ALS nuclei and NPCs [93,94]. Importantly, in iPSNs, CHMP7 relocalization is not initiated by altered NPC composition resulting from nucleoporin reduction [93]. However, it is currently unknown whether protein coding variants in nucleoporins, such as those recently identified in Nup50 [200], can subtly alter the structure or function of the NPC and contribute to nuclear localization and/or function of CHMP7 and ESCRT-III nuclear surveillance. Interestingly, our group recently established that alterations in NPC permeability barrier integrity are dependent on the LINC complex protein SUN1 and facilitate increased nuclear influx of CHMP7, thereby increasing its nuclear localization in sALS iPSNs [90]. Although the potential contribution of impaired nuclear export of CHMP7 to increased nuclear localization remains unknown, this study suggests that subtle alterations to NPC or nuclear envelope structure and integrity may compromise the integrity of the passive diffusion permeability barrier of the NPC central channel in the pathologic disruption of CHMP7/ESCRT-III nuclear surveillance. Notably, these studies suggest that the mechanisms underlying ESCRT-III surveillance of NPC assembly during cell division and maintenance throughout the lifetime of a cell may at least be in part distinct. In support of this, two publications now provide evidence that LEMD2 does not facilitate CHMP7/ESCRT-III NPC maintenance in iPSNs [90,93]. In contrast, as is typical for overall ESCRT functionality in multiple cellular processes [2–4,8,105,109,201], pathological nucleoporin reduction requires the recruitment of VPS4 in a manner dependent on CHMP7 in iPSNs [161]. A subsequent study confirmed these findings by demonstrating that increased expression of VPS4 leads to excessive degradation of nucleoporins in a Drosophila model of C9orf72 ALS/FTD [199]. Interestingly, expression of a dominant negative VPS4 led to the intranuclear accumulation of POM121 following its reduction from NPCs in iPSNs [161] suggesting that the ESCRT-III pathway may facilitate the nuclear internalization/budding of nucleoporins for physiologic and/or pathologic degradation. Together, these studies suggest that NPC injury observed in ALS and FTD may at least be in part due to pathologic overactivation of the ESCRT-III nuclear surveillance pathway. Future and current investigations into the molecular mechanisms underlying ESCRT-III mediated disruptions in NPC homeostasis is of high priority for our group and is essential for our understanding of pathomechanisms of neurodegenerative disease.

Figure 3. ESCRT-III mediated physiologic maintenance and pathologic disruption of NPCs in human neurons. (a) In normal human neurons, turnover of individual nucleoporin proteins to maintain NPC integrity and function is initiated by passive diffusion of CHMP7 to the nuclear space. CHMP7 is activated via currently undefined but LEMD2 independent protein – protein interactions. Activation of CHMP7 likely facilitates the recruitment and polymerization of ESCRT-III subunits. Individual nucleoporin proteins are removed from the NPC and degraded in a VPS4 dependent manner. VPS4 likely facilitates ESCRT-III polymer disassembly and subunit exchange and reuse. Exportin-1 (XPO1) actively exports CHMP7 from the nucleus, resulting in its ‘inactivation’ and maintaining low nuclear levels of CHMP7 under basal cellular conditions. Nucleoporins are synthesized and reinserted into the NPC. (b) In ALS neurons, the integrity of the passive diffusion permeability barrier is disrupted in a SUN1 dependent manner leading to a pathologic increase in the nuclear influx of CHMP7. CHMP7 is activated via currently undefined by LEMD2 independent protein – protein interactions. Activation of CHMP7 likely facilitates the recruitment and polymerization of ESCRT-III subunits. Individual nucleoporin proteins are removed from the NPC and degraded in a VPS4 dependent manner. VPS4 likely facilitates ESCRT-III polymer disassembly and subunit exchange and reuse. We hypothesize impaired interactions between CHMP7 and XPO1 or other Exportins abrogate its active nuclear export and facilitate nuclear accumulation of CHMP7 observed in ALS/FTD. In addition, we hypothesize that pathologic removal and degradation of nucleoporins is sustained via 1. sustained excessive nuclear influx of CHMP7 and/or 2. sustained LEMD2 independent activation of CHMP7 (e/g/persistent nuclear protein – protein interactions and/or 3. sustained VPS4 mediated nucleoporin removal from NPCs and/or 4. impaired nucleoporin reincorporation perhaps as a result of altered nuclear transport receptor and active nuclear import function. Collectively, sustained ‘overactivation’ of ESCRT-III mediated nucleoporin turnover gives rise to NPC disruptions in ALS/FTD human neurons.

Endolysosomal trafficking and protein degradation

Endolysosomal trafficking proceeds through a series of vesicular intermediates ultimately culminating in the degradation of cargo molecules within the lysosome [202,203]. Thus, this pathway plays a pivotal role in maintaining protein homeostasis within cells. A number of studies have identified a role for ESCRT proteins in endosomal sorting, MVB biogenesis, and endosomal and lysosomal membrane repair [13,15,28,37,99–101,204–206]. Briefly, endosomal sorting and MVB biogenesis proceeds first by the binding of ESCRT-0 complexes to endosomal membranes resulting in the clustering of ubiquitinated cargoes [2,3,13,105,125,201,207]. ESCRT-I and ESCRT-II complexes then bind ubiquitinated cargoes and deform MVB membranes to form a cargo containing bud [2,3,13,37,105,125,201,207]. Studies also suggest that ESCRT-III proteins may play a role in the induction of membrane curvature when localized to endosomal and MVB membranes at high concentrations [2,3,13,105,125,208,209]. Nonetheless, recruitment of ESCRT-III proteins is essential for formation of vesicles via scission of the neck of membrane buds [8,125,210]. Moreover, in addition to their role in recruitment of VPS4 for polymer disassembly as discussed above, ESCRT-III subunits recruit deubiquitinating enzymes for protein deubiquitination during sorting of cargoes within endosomes into MVBs [2,3,13,105,125,211].

Ultimately, MVBs fuse with either autophagosomes or lysosomes to promote protein degradation through the autophagy pathway [8,212,213]. Consistent with a role for ESCRT-III proteins in facilitating endolysosomal trafficking and degradation, reduction in CHMP4B or CHMP3 expression leads to accumulation of autophagosomes in rodent cortical neurons [214] and HeLa cells [215] respectively and knockdown of CHMP3 or CHMP5 impairs receptor degradation [216,217]. As will be discussed in more detail below, mutations in CHMP2B can also impact endolysosomal pathway function [52,218–220]. Although the molecular mechanisms by which ESCRT-III proteins regulate autophagy and lysosomal fusion remain unclear, ESCRT-III proteins also play a critical role in repair of lysosomal membranes upon damage, in turn, preventing cell death [28,221–223].

Neurodegenerative diseases are often characterized by the presence of protein aggregates at end-stage disease [85,224–230]. Thus, this pathology is suggestive of impaired protein homeostasis and degradation. A number of studies have implicated alterations in autophagy and lysosomal degradation as a contributing factor to neurodegenerative disease pathologies [34,36,38,39,231–235]. Interestingly, a recent study has demonstrated that upregulation of ESCRT pathway proteins promotes the degradation of multiple proteins associated with neurodegeneration including Tau and Huntingtin [236]. In addition, reduction in CHMP2A, CHMP2B, or CHMP6 can compromise endolysosomal function and promote Tau propagation and aggregation [237] and Tau accumulation can subsequently impact the expression of Ist1/CHMP8 therefore impeding ESCRT-III function [238]. These studies suggest that feedback loops between endolysosomal dysfunction and protein aggregation may perpetuate deficiencies in protein homeostasis in neurodegenerative disease.

Importantly, mutations in genes that function in the endolysosomal or autophagy pathway (e.g. C9orf72, VAPB, VCP, TBK1, UBQLN2, CHMP2B) are causative of neurodegenerative disease [50,51,53,56,239–242] suggesting that impairments in endolysosomal function may be a primary contributor to disease pathophysiology. We refer readers to a recent review [39] for a comprehensive discussion of implications of these genetic mutations on endolysosomal trafficking and protein degradation in ALS and FTD.

From a nuclear perspective, multiple studies have now demonstrated that NPCs and nucleoporins can be degraded via autophagy pathways [192,193,243–246]. The data suggest that autophagy impairments may compromise nucleoporin turnover and may contribute to NPC disruptions in neurodegenerative disease. Although the mechanisms that connect ESCRT-III mediated NPC surveillance to autophagy and lysosomal degradation remain unknown, as discussed above, there are demonstrations that altered ESCRT-III nuclear surveillance may initiate aberrant nucleoporin degradation in ALS/FTD [90,93,161,199]. However, other reports suggest that NPCs and nucleoporins are degraded via the proteasome [30,199]. Thus, the mechanisms and pathways implicated in physiologic and pathologic degradation of nucleoporins and NPCs remain controversial. In the future, it will be necessary to determine whether distinct pathways facilitate cell type, organismal, and even nucleoporin-specific degradation.

Mutations in ESCRT-III proteins in neurodegenerative disease

Autosomal dominant mutations in the ESCRT-III protein CHMP2B have been implicated in rare genetic forms of FTD, ALS, and other motor neuron diseases [53–56,247]. The first identified mutation lies within the splice acceptor site for the final exon and results in two novel transcripts referred to as CHMP2B^Intron5 (CHMP2B^M178V) and CHMP2B^Δ10 [55,216]. A second study identified an autosomal dominant point mutation at amino acid 165 whereby glutamine is replaced by a stop codon (CHMP2B^Q165X) [56]. Both mutations, and all three resulting transcripts result in the production of truncated CHMP2B proteins lacking a portion of the C terminus [55,56,216]. Subsequent reports have identified a number of disease causative point mutations in patients with FTD (CHMP2B^D148Y) and motor neuron disease (CHMP2B^I29V, CHMP2B^T104N, CHMP2B^Q206H) [53,55,56,247]. Thus, the genetic link between CHMP2B and FTD/ALS suggests that ESCRT-III disruption may be a primary contributor to neurodegenerative disease pathogenesis.

CHMP2B FTD/ALS cases are unique in that they lack the characteristic cytoplasmic aggregation TDP-43 despite the presence ubiquitin and p62 positive inclusions at end-stage disease [248]. In contrast, a recent study demonstrated that overexpression of wildtype CHMP2B or the CHMP2B^Intron5 mutation resulted in accumulation of phosphorylated TDP-43 in N2a cells [249]. Whether this discrepancy from human pathology is cell type specific or related to artifacts of overexpression remains unclear at this time. However, endocytosis and the ESCRT machinery have both been linked to TDP-43 turnover and clearance [215,250] albeit largely in model systems based on TDP-43 overexpression. We also note that the lack of cytoplasmic TDP-43 aggregation and phosphorylation may not accurately reflect and indicate the status of TDP-43 function, specifically as it relates to loss of transcription and splicing activity in the nucleus. In fact, even in ALS, cells that harbor TDP-43 aggregated at end-stage disease are rare [65,251,252]. A substantially higher percentage of cells display a gradient-like distribution of non-aggregated TDP-43 throughout the nucleus and cytoplasm and a substantial proportion of cells even display ‘normal’ nuclear localization and expression of TDP-43 [253]. Thus, it is possible that TDP-43 nuclear function may be disrupted and/or TDP-43 May mislocalize to the cytoplasm independent of cytoplasmic aggregation, a phenomenon that may not be detected by histological analyses. As a result, future studies are necessary to determine whether TDP-43 function is disrupted in models of CHMP2B FTD/ALS and whether this contributes to disease pathogenesis as has been recently observed in other forms of ALS/FTD [63,64,70,76,254–256].

To date, most studies detailing the cellular consequences of CHMP2B mutations have focused on endolysosomal trafficking and autophagy functions of the ESCRT-III pathway. Histologically, multiple groups have observed pathologic accumulation of enlarged endosomes and autophagic structures in in vitro and in vivo model systems of CHMP2B FTD/ALS [52,55,214,216,257]. In contrast to the normally diffuse and predominantly cytoplasmic distribution of wildtype CHMP2B, CHMP2B^Intron5 and CHMP2B^Δ10 proteins form cytoplasmic puncta that colocalize with the endosomal and lysosomal marker CD63 when overexpressed in PC12 cells [55]. In addition, when overexpressed in primary neurons, cytoplasmic CHMP2B^Intron5 puncta colocalize with markers of recycling, late, and early endosomes [258] and CHMP2B^Intron5 expression can lead to CHMP4B sequestration within Rab7 positive endosomal structures [214]. Together, these observations are suggestive of widespread disruptions to endolysosomal trafficking and function. This is supported by observations of large endosome accumulations in patient fibroblasts, postmortem cortical tissue and primary rat neurons overexpressing CHMP2B^Intron5 [214,216,258] as well as autofluorescent accumulations in transgenic mice [52]. When expressed in HEK293 cells, CHMP2B^Intron5 more strongly associates with CHMP4B compared to its wildtype counterpart [214]. This is reminiscent of enhanced ESCRT-III subunit association upon expression of a dominant negative VPS4 [214] suggesting that ESCRT-III polymer disassembly may be impaired in the context of CHMP2B mutations.

Given that a number of disease associated mutations in CHMP2B result in C-terminal protein truncation [55,56,216], it is possible that autoinhibition is prevented leading to constitutive CHMP2B polymerization. This hypothesis is supported by the punctate distribution of CHMP2B mutants as well as the failure to dissociate from CHMP4B [55,214]. Whether CHMP2B and its potential constitutive activation compromise nuclear surveillance functions of the ESCRT-III pathway in disease are currently unclear but under investigation at least within our group. However, a recent study demonstrated that CHMP2B^Intron5 exhibits stronger binding to the microtubule-severing enzyme Spastin than its wildtype counterpart [259]. In addition, Spastin is colocalized with CHMP2B^Intron5 accumulations presumably impacting its solubility [259]. Interestingly, Spastin is recruited by IST1/CHMP8 for mitotic spindle disassembly and nuclear envelope sealing [29]. Thus, these data support a potential link between CHMP2B mutations in dysregulation of nuclear envelope and nuclear pore complex maintenance and surveillance functions of the ESCRT-III pathway.

Conclusions and perspectives

The ESCRT-III pathway is a multifaceted cellular pathway with critical roles in membrane dynamics, protein degradation, and maintenance of cellular integrity. Therefore, disruptions to ESCRT-III protein subunits and overall pathway function are likely to have widespread impacts on cellular health and survival. Recent studies have highlighted a role for ESCRT-III pathway alterations, particular NPC and nuclear envelope surveillance and endolysosomal trafficking as important contributors to neurodegenerative disease pathogenesis. However, the molecular mechanisms by which ESCRT-III dysfunction gives rise to pathophysiologic events in disease remain largely unknown. Specifically, from the perspective of NPC surveillance, little is known regarding the maintenance of NPCs throughout the lifetime of non-dividing neurons. Thus, as it relates to neurodegenerative disease pathogenesis, future studies are necessary to understand both the physiologic function and pathologic consequences of impaired ESCRT-III nuclear surveillance function in a cell type and organism-specific manner. Ultimately, understanding the mechanisms by which ESCRT-III dysfunction contributes to neurodegenerative disease, may yield novel therapeutic targets and strategies for disease. Modulation of specific ESCRT-III functions, proteins, or subunit interactions could hold promise for future therapies. In support of this idea, a recent study demonstrated that antisense oligonucleotide (ASO) mediated knockdown of CHMP7 was sufficient to repair NPCs, alleviate alterations to TDP-43 function and localization, and improve neuronal survival in genetic and sporadic ALS iPSNs [93]. Thus, continuing to unravel the intricacies underlying ESCRT-III function and dysfunction specifically in human neurons has the potential to have a significant impact on our understanding of neurodegenerative disease and novel treatment strategies.

Author contributions

ANC wrote and edited the manuscript. OK carried out all manuscript revisions and assisted with final editing. ANC conceptualized and generated figures. Conceptualization and oversight were carried out by ANC. All authors reviewed and approved the final manuscript.

Acknowledgments

We thank America Chandia Cristi and Ssu-Ying Chen for feedback on this manuscript.

Disclosure statement

ANC has submitted patents on methods and drugs to modulate various ESCRT-III proteins in neurodegeneration.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Additional information

Funding

ANC is supported by funding from NIH NINDS/NIA (R00NS123242), NIH NINDS (R01NS132836), The Robert Packard Center for ALS Research, BrightFocus Foundation, Muscular Dystrophy Association, and Target ALS (IL-2023-C6-L4 and IL-2023-C5-L3).