1. Extracellular Vesicles in Amyotrophic Lateral Sclerosis
Extracellular vesicles (EVs) are secreted from almost all cells and circulate freely around the body. They are able to cross the blood–brain barrier and therefore are a means of intercellular signalling to and from the central nervous system (CNS)
[1]. EVs contribute to many physiological processes in the CNS, including: neural growth and development; CNS inflammation; the neuroprotective response to oxidative stress; and maintaining brain vascular integrity and post synaptic retrograde signalling
[2][3][4][5][6]. EVs are also implicated in the pathological processes of neurodegeneration, and large numbers of studies have in recent years evaluated the role of EVs in neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS), Parkinson’s Disease (PD), Alzheimer’s Disease (AD), Huntington’s Disease and prion diseases
[7].
1.1. Amyotrophic Lateral Sclerosis Associated Genes Involved in Vesicular Pathways
There are now over 50 potentially causal or disease modifying genes identified for ALS
[8]. The most commonly identified gene mutation in European and North American patients is the hexanucleotide repeat expansion in
C9orf72, accounting for up to a third of all identified pathogenic mutations
[9]. Other common causal mutations are in the genes
SOD1,
TARDBP,
NEK1,
FIG4 and
TBK1 [9][10][11]. There is wide geographical variation in the genetic causes of ALS. Familial ALS is much less frequent in China (1.2–2.7% of total cases), where mutations in
SOD1 are the most frequent genetic cause (25.3% of fALS), followed by
TARDBP and
FUS [12]. This is similar to Japan, where
SOD1 mutations are the most common cause of fALS (29.8%), followed by
FUS and
TARDBP, whereas
C9orf72 expansions are rare, identified in only 1.3% of fALS cases
[13]. In Brazil, the most frequent genetic cause is
VAPB (30% of all fALS) followed by
C9orf72 (22%)
[14].
The multiple genes involved in ALS have a range of effects on multiple cellular processes including RNA processing, protein homeostasis, cytoskeletal dynamics, mitochondrial function, endosomal trafficking, autophagy and, important for this research, the formation of EVs
[8][15]. Multiple ALS associated genes are involved in vesicular trafficking and EV regulation as shown in
Table 1. The CHMP2B protein is an essential component of the ESCRT III complex, which is part of the machinery for MVB and EV formation, as discussed above
[16].
C9orf72,
VAPB,
FIG4,
SPG11 and
ALS2 encode proteins which are involved in intracellular trafficking of vesicles
[17][18][19][20][21][22]. Optineurin and SQSTM1/p62 proteins are both autophagy receptors and are activated through phosphorylation by TBK1
[23]. Autophagy and EV function are closely linked, with the level of autophagy in cells closely related to the secretion and transport of EVs
[24][25].
Table 1 lists some of the proposed pathogenic mechanisms for mutations of genes having functions related to vesicular pathways (also illustrated in
Figure 1).
Figure 1. The endosomal and vesicular pathway involvement of proteins encoded by ALS related genes. Alsin activates Rab5 to promote endosomal fusion and subsequent endosomal trafficking. C9orf72 forms a complex with SMCR8 (Smith-Magenis Syndrome Chromosome Region, Candidate 8) and WDR41 (WD Repeat domain 41) proteins. This complex interacts with Rab GTPases including Rab5 in endosomal formation and trafficking. The C9orf72 complex also regulates various steps in autophagy including MVB-autophagosome and autophagosome-lysosome fusion as well as regulating several aspects of lysosomal biogenesis, pH and reformation. VAPB encodes Vesicle-associated membrane protein-associated protein B/C found in the endoplasmic reticulum (ER). This anchors complexes involved in lipid transfer from the ER to golgi and also the recycling of phosphatidylinositol-4-phosphate (PtdIns4P). VAPB mutations disrupt ER-golgi tethering and leads to PtdIns4P accumulation with subsequent accumulation of endosomes and dysfunctional lysosomes. FIG4 is required for the homeostasis of a signalling lipid phosphatidylinositol 3,5-bisphosphate (PI(3,5)P2), which is required for endosomal and lysosomal maturation. FIG4 mutations result in enlarged endosomes and lysosomes with impaired lysosomal function. CHMP2B is responsible for the formation of intraluminal vesicles within the MVBs and may participate in the proper fusion of MVB with the lysosomes and the autophagosomes. CHMP2Bintron5 mutation results in accumulation of large endosomes and autophagosomes. VCP is involved in the initiation of autophagy and autophagosome maturation. TBK1 phosphorylates both optineurin (OPTN) and Sequestosome-1/p62, increasing their ability to bind to ubiquitinated cargo, initiating autophagy and delivery to autophagosomes. UBQLN2 interacts with LC3, a marker for starvation induced autophagy, to deliver ubiquitinated cargo to autophagosomes and is also recruited to OPTN containing vesicles. Spatacsin interacts with Rab5 for endosomal trafficking and maturation and contributes to lipid clearance from late endosomes and lysosomes. SPG11 mutations result in loss of spatacsin function, which leads to accumulation of lipids in lysosomes. Figure created using Biorender.com.
1.2. Extracellular Vesicle Mediated Transfer of Misfolded Proteins and miRNAs in Amyotrophic Lateral LSclerosis
A pathological hallmark of ALS is TAR DNA-binding protein 43 (TDP 43) positive inclusions, which have been identified in brain stem and spinal cord tissue in over 97% of patients at postmortem
[33]. SOD1, fused in sarcoma (FUS), and dipeptide repeat proteins (DPRs) from the
C9orf72 intronic hexanucleotide repeat expansion, also aggregate to form protein inclusions
[34]. These misfolded proteins have been demonstrated to spread between cells in a prion-like manner and induce further protein misfolding
[35]. This mechanism has been used to explain the contiguous spread of disease that is often seen in ALS, where the disease spreads to adjacent neuroanatomical segments
[36]. EVs have been shown to contain aberrant protein aggregates in cell and animal models of ALS as well as in patients with ALS. Evidence is growing to support the hypothesis that EVs can spread pathological misfolded proteins between cells in ALS and that these EVs can exert deleterious effects on recipient cells, discussed in
Section 4 below. EVs have also been shown to mediate the transfer of RNAs, particularly miRNA, which can alter gene expression in recipient cells, potentially also of relevance to the progressive spread of neurodegeneration in ALS. Studies in other neurodegenerative diseases have also found that pathological proteins such as prion protein, amyloid-β, α-synuclein, and tau propagate via exosomes
[37][38][39]. A timeline of discoveries in cell based and animal models of these roles of EVs in ALS is shown in
Figure 2.
Figure 2. A timeline of experimental evidence of how EVs are involved in the spread of misfolded proteins, and how miRNA alter cellular phenotypes, in ALS cell-based and animal models. Abbreviations: MN-motor neuron, DPR-dipeptide repeat proteins
[40][41][42][43][44][45][46][47][48][49][50][51][52]. Figure created using Biorender.com.
1.2.1. SOD1
The first evidence that EVs could spread misfolded proteins was in 2003 when SOD1 was shown to be excreted by EVs in SK-N-BE neuroblastoma cells
[40]. Astrocytes with mutant SOD1 overexpression showed increased EV secretion compared to wild type cells; these secreted EVs were taken up by motor neurons, inducing cell death
[42]. Misfolded SOD1 aggregates spread between NSC-34 motor neuron-like cells via EVs, causing cell rupture and cell death
[41][53]. Once the SOD1 aggregates are introduced to neural cells, they result in a self-perpetuating induction of further SOD1 aggregation and transfer between cells
[54]. These results have also been replicated in animal models, with CNS derived EVs in SOD1
G93A mice containing misfolded SOD1 aggregates
[50]. However, while mutant SOD1 aggregates are secreted by EVs, other studies have reported impaired secretion of mutant SOD1 compared to the wild type protein in NSC-34 motor neuron-like cells and rat microglial cells, with a proposed underlying mechanism of dysfunctional secretory pathways as a result of golgi fragmentation and ER stress
[55][56].
1.2.2. TDP 43
Multiple studies demonstrate that TDP 43 is transferred intercellularly via EVs. Insoluble TDP 43 aggregates from ALS or FTD brain tissue resulted in intracellular accumulation of TDP 43 and cell death when added to SH-SY5Y neuroblastoma cells
[43]. The same study also showed that TDP 43 aggregates can transfer between cells via EVs. CSF EVs in patients with ALS and ALS/FTD contain TDP 43 and these EVs have been shown to cause propagation of TDP 43 aggregates when added to U251 glioblastoma cells
[45]. EVs isolated from human ALS brain tissue also caused TDP 43 accumulation and propagation in Neuro2a mouse glioblastoma cells
[46]. In the same study, EV release was inhibited by a sphingomyelinase inhibitor, which resulted in increased TDP 43 aggregates in the Neuro2a cells and also exacerbated the clinical phenotype of transgenic mice expressing human mutant TDP-43
A315T. This suggests that while EVs play a key role in the propagation of TDP 43 proteinopathy, the inhibition of EV secretion may precipitate greater intracellular accumulation of pathological aggregates
[46]. Free TDP 43 can also be taken up by cells, but it has been shown in human embryonic kidney 293 (HEK-293) cells that EVs containing TDP 43 are preferentially taken up compared to free TDP 43, and thus have potential to spread and exert greater cellular toxicity
[57].
1.2.3. FUS
There are few studies analysing mutant FUS or FUS binding partners. Analysis of HEK cells expressing mutant FUS
R521G confirmed that FUS is present in EVs and that many FUS binding partners are components of EVs
[47]. FUS and several FUS binding partners were also observed in EVs derived from skeletal muscle samples from sALS patients without FUS mutations
[52].
1.2.4. Dipeptide Repeat Proteins
DPRs are a group of 5 protein complexes formed as a result of repeat-associated non-AUG (RAN) translation of the
C9orf72 intronic hexanucleotide repeat expansion
[58]. Such DPRs have been shown to spread intercellularly via EVs in spinal motor neurons derived from induced pluripotent stem cells (iPSCs) expressing the expanded hexanucleotide repeat
[48].
1.2.5. RNA Transport by EVs
While multiple types of RNA have been found in EVs, most is known about miRNAs. These are short non-coding RNAs which have gained great attention due to their ability to modify gene expression in recipient cells
[59]. MiRNAs are stable in the circulation and are found in a range of body fluids including serum, plasma, urine and CSF
[60]. There is also evidence of their transport in EVs between cells
[61]. EVs contain miRNA profiles that are vastly distinct from their host cell from which they originate
[62]. Multiple miRNAs have been identified as dysregulated in vitro in ALS models involving several different cell types. EVs have been shown to transmit miR-124-3p from neurons to astrocytes, which regulates the glutamate uptake of astrocytes
[63]. A study on astrocytes derived from iPSCs from C9orf72 ALS patients compared with iPSCs from healthy controls found that the EV miRNA content was dysregulated in ALS, with 64 dysregulated miRNAs, and downregulation of miR-494-3p as the most significant change
[51].
There are fewer in vitro studies of messenger RNA (mRNA) expression in EVs in ALS. Similar to miRNA, mRNA expression in EVs from iPSC derived motor neurons is markedly different from that of the iPSC cells themselves, being enriched for genes regulating cellular metabolism and protein homeostasis
[64].
2. Therapeutic Application of Extracellular Vesicles in Amyotrophic Lateral Sclerosis
There has been great interest in the potential use of EVs as a non-invasive method to deliver therapeutics including proteins, genetic material and drugs, in neurodegenerative diseases
[65]. The investigation of EVs as therapeutic vectors is growing with at least 20 phase 1/2 clinical trials registered in patients with cancer, SARS-CoV-2, and Alzheimer’s disease, with treatments including stem cell derived EVs, autologous EVs or drug loaded EVs
[66]. There are several properties of EVs which suggest that they could be useful in therapeutics in neurodegenerative diseases such as their ability to cross the blood–brain barrier, their low tendency to evoke an immune response and the potential for manipulation of cell surface markers to limit off target effects
[67][68]. However, there are also challenges in development of EVs for this use, such as batch-to-batch variation in their synthesis, the current lack of cost-effective large scale production protocols or of robust, reproducible methods for drug loading
[66]. EVs have been used in murine models of PD and AD to deliver small interfering RNAs (siRNA) to reduce pathological protein accumulation. EVs with α-Synuclein siRNA were peripherally injected into α-Synuclein
S129D transgenic mice. This decreased the level of α-Synuclein aggregates in brain regions pathologically affected in PD
[69]. EVs containing BACE1 siRNAs have also been used in C57BL/6 mice resulting in an over 60% reduction in BACE1 mRNA and a 55% decrease in β-amyloid 1-42 levels, a key component of plaques in AD pathology
[70].
Studies in ALS have used EVs taken from mesenchymal stem cells (MSCs). The capacity of MSCs to replicate, differentiate, secrete neuroprotective factors and produce new cells to replace damaged cells, has led to multiple phase 1/2 clinical studies investigating their potential use in ALS
[71]. However, challenges to the use of MSCs include dosing issues, variation in the differentiation state of the cells and the route of administration
[72]. As it is now recognised that MSCs exert much of their action through secretion of EVs, investigation of the MSC secretome and EVs, as a cell-free therapeutic approach, is now being explored
[73].
EVs from adipose derived stem cells (ADSCs) have been investigated in ALS cell models. Healthy human ADSCs were added to murine SOD1
G93A NSC cells resulting in slower SOD1 aggregation and improved mitochondrial function
[74]. EVs from Murine ADSCs have also been added to NSC-34 cells overexpressing SOD1
G93A, SOD1
G37R and SOD1
A4V, which were challenged to oxidative damage with hydrogen peroxide. Treatment with EVs reduced oxidative damage, increased cell viability and improved mitochondrial function
[75][76]. A subsequent study by the same group confirmed their previous findings and reported proteomic analysis on the ADSC EVs, identifying 189 proteins. Gene ontology analysis showed the most significant biological processes of the identified proteins to be cell adhesion, negative regulation of apoptosis and positive regulation of cell proliferation
[77]. The group also tested intravenous and intranasal administration of ADSC EVs to SOD1
G93A mice. The EVs were labelled with ultra-small superparamagnetic iron oxide nanoparticles and MRI confirmed the EVs did penetrate the blood–brain barrier and were deposited in the brain. The mice showed improved motor performance compared to placebo and had greater preservation of lumbar motor neurons, neuromuscular junctions and muscle fibres
[78].
EVs isolated from human bone marrow-derived endothelial progenitor cells have also been shown to reduce damage in a murine brain endothelial cell ALS model
[79]. Giunti et al. modified EVs from bone marrow derived MSCs by first treating the MSCs with interferon-γ. Treating the MSCs with interferon-γ resulted in upregulation of multiple microRNAs in EVs including miR-466q and miR-467f, which can reduce microglial activation through inhibition of the p38 mitogen-activated protein kinase pathway. The EVs reduced the levels of mRNA for Tumour Necrosis Factor and Interleukin 1b in SOD1
G93A microglial cells
[80]. This research showed that EV content can in principle be altered to exert the desired effect on target cells.