Caspase-1 clipping causes complications for α-synuclein

The study of human neurodegeneration has increasingly focused on a clinically diverse group of diseases that share the common pathologic feature of insoluble aggregates of α-synuclein (αSyn) in various populations of neurons and glia. Scientists have begun identifying factors that may initiate or accelerate αSyn aggregation and how this process relates to progressive impairment of neuronal structure and function. In PNAS, companion papers by Wang et al. (1) and Bassil et al. (2) establish a compelling link between the accumulation of C-terminally truncated forms of αSyn created by caspase-1 and cellular dysfunction in culture and in vivo, all potentially initiated by inflammatory insults.

It has become apparent that, under physiological conditions, αSyn exists in several conformational and multimeric states, but their dynamic interrelationship and how some of them are altered into abnormal species that effect injury have not been clearly defined. Some answers are arising from αSyn missense mutations that cause familial Parkinson’s disease (PD). These occur in the N-terminal third of the protein and may induce structural changes that decrease the intramolecular interaction between the N and C termini (3) and/or decrease the intermolecular propensity to form α-helical tetramers and related multimers (4). Interfering with the latter process can lead to an excess of unfolded monomers prone to abnormal oligomerization (5). In rare families that inherit αSyn gene multiplication, a causal link between the lifelong increase in monomer levels, neurotoxicity, and PD is likely (6).

There is also mounting evidence that PD-linked missense mutations can enhance the production of C-terminally truncated αSyn species and such truncation of αSyn may also occur in healthy brains (7). The increased brain levels of these fragments observed in PD and dementia with Lewy bodies (DLB) (7, 8) and the vulnerability of dopaminergic neurons in αSyn transgenic (tg) rodents to enhanced C-terminal truncation [e.g., due to overexpression (9) or PD-relevant toxins (8)] all argue for a critical threshold of truncated αSyn. Notably, such truncated species can accelerate the aggregation of the full-length αSyn protein in vitro (10) and in vivo (11), and thus may contribute to or even initiate the aggregation process. Numerous laboratories have identified truncated αSyn species in LBs of PD and DLB (7, 12) as well as in the oligodendroglial inclusions (GCIs) of multiple system atrophy (MSA) (13). Most cases of these diseases present in a “sporadic” (i.e., not overtly genetic) fashion, suggesting a role for environmental triggers. In this context, αSyn aggregation and neurotoxicity can follow long-term, low-dose exposure to pesticides or the immune response to viral infection (14).

Is there a common pathway that leads to αSyn truncation in vivo and how can we protect the C terminus? In the last decade, some studies have focused on the susceptibility of human αSyn to PD-relevant environmental toxins. A common finding in both familial PD αSyn tg mice and wild-type animals exposed to either PD-linked pesticides or inflammogens is an increase in neuronal levels of αSyn and/or its shift to insoluble, aggregated forms (8, 15, 16). Previous work by one of us identified a significant increase in C-terminally truncated αSyn in mice exposed to low doses of paraquat or 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (8, 17). Paraquat exposure typically results in activation of calpain (18) (Fig. 1), a calcium-dependent protease implicated in PD (19). This protease is known to cleave αSyn at residue 122 into a C-terminally truncated, aggregation-prone form, similar to that seen in affected regions of PD and DLB brain (8, 10, 12). There is a dual adverse effect of activating calpain-1: it cleaves αSyn and thus induces neurotoxicity on the one hand, and it inhibits autophagy and thereby full proteolytic degradation of αSyn on the other (8). In a related work, coexpression of the natural calpain inhibitor, calpastatin, decreased αSyn truncation and aggregation in vivo (20). These studies support the occurrence of calpain-dependent αSyn truncation in the aggregation process of synucleinopathies. It should be noted that other proteases can potentially cleave αSyn, including the extracellular matrix metalloproteinases (21), and these proteases are also involved in brain inflammation.

Fig. 1.

Hypothetical interactions of caspase-1 and calpain-1 in protein degradation, αSyn truncation, and neurodegeneration. Inflammatory signals can activate caspase-1 and subsequent cleavage of the calpain-1 inhibitor, calpastatin. Cleavage of certain cytoskeletal proteins may also contribute to pathologic membrane changes and increased calpain-1 activation. Environmental toxins can further contribute to aberrant protease activity by elevating membrane permeability and thereby intracellular Ca²⁺ levels, thus inducing pathological calpain-1 activation. Calpain-1 and caspase-1 each degrade essential cytoskeletal and cytosolic substrates, including αSyn. This cascade of adverse conditions could alter cellular physiology and shift normally multimeric αSyn toward an excess of the more protease-sensitive monomers, enabling their subsequent truncation into aggregation-prone forms and ultimately resulting in neuronal dysfunction and loss. Calcium channel blockers and selective protease inhibitors against caspase-1 or calpain-1 or else increased αSyn C-terminal interactors could each act as protectants against aberrant cleavage of αSyn into aggregation-prone species.

Now, Wang et al. (1) make the intriguing observation that treating αSyn-overexpressing neuroblastoma cells with a range of inflammogens results in C-terminally truncated and insoluble αSyn and redistributes αSyn into punctate inclusions. All inflammasome activators tested induced truncation, and one (menadione) was shown to do so dose dependently. Of note, the low level of truncation seen (<20% of total αSyn) was visualized with only one αSyn antibody, so the complete spectrum of truncated species may be more complex. Because activation of the inflammasome often activates caspase-1, the authors asked whether the latter was responsible for the αSyn truncation they observed. Using specific inhibitors (VX-765 on cells and ML132 in vitro) as well as gene silencing of caspase-1, they were able to diminish αSyn truncation and increase cell viability. Next, MALDI-TOF mass spectrometry identified residues 121/122 as the specific cleavage site, and this was elegantly validated when a D121E point mutation abolished caspase-1–mediated αSyn truncation in vitro.

Are the caspase-1–generated fragments of αSyn implicated in its aggregation? Wang et al. (1) subjected αSyn cleaved by caspase-1 as well as a recombinant αSyn 1–121 protein to Thioflavin-T fluorescence aggregation assays and observed more rapid aggregation than that of holo-αSyn alone. Intriguingly, they detected colocalization of caspase-1 and αSyn in LBs of PD brain tissue.

These studies were conducted in cell culture, and their translation into brain inflammatory processes in vivo and the potential interplay between glia and neurons remained open. Here, the companion paper by Bassil et al. (2) furthers our understanding of how caspase-1–mediated truncation, oligomerization, and aggregation of αSyn in glia may mediate neurotoxicity in vivo. Bassil et al. (2) studied PLP-SYN tg mice, in which the promoter of the myelin-associated proteolipid protein (PLP) drives αSyn overexpression in oligodendrocytes and thereby models key aspects of MSA, including GCIs, neuroinflammation, dopaminergic cell loss, and motor deficits (22). The authors subsequently performed biochemical analyses on the striatum and the motor cortex, both known to have a high burden of GCIs and neuronal loss in MSA. Beside monomeric (full-length) αSyn, the authors quantified truncated and oligomeric αSyn species, but using just one αSyn antibody. In light of the caspase-1–truncated αSyn generated by inflammasome activation in Wang et al. (1), Bassil et al. (2) asked whether the inhibitor VX-765 can prevent αSyn truncation, aggregation, and dopaminergic cell loss in vivo. Probing both plain and proteinase K (PK)-treated histological sections, Bassil et al. demonstrate a reduction of PK-resistant αSyn aggregates (GCIs) in striatum (but, surprisingly, not in cerebral cortex), associated with reduced αSyn levels and improved locomotion on the beam walk test. VX-765 may also

In PNAS, companion papers by Wang et al. and Bassil et al. establish a compelling link between the accumulation of C-terminally truncated forms of αSyn created by caspase-1 and cellular dysfunction in culture and in vivo, all potentially initiated by inflammatory insults.

have lessened apparent neuroinflammation (i.e., IL-1β elevation), and dopaminergic neuron loss was ameliorated. Importantly, VX-765 treatment cancelled the positive correlation between truncated αSyn and abnormal oligomers, suggesting a direct effect of the caspase-1 truncation in the cytopathological aggregation process of this MSA model.

Together, these multiple findings firmly support a role for C-terminally truncated αSyn in the neurotoxic aggregation of synucleinopathies. However, what normally keeps the C terminus of αSyn from being cleaved by proteases? The C terminus is less conserved than the N terminus. The interaction between these termini is hypothesized to help stabilize αSyn conformations and protect against truncation and aggregation. Interestingly, rodent αSyn differs from human αSyn at amino acids 121 and 122 (mouse: D121G, N122S; rat: D121S, N122S), potentially precluding aberrant proteolytic cleavage in rodents. However, these rodent residues are uncharged, so their role in interactions with the positively charged N terminus is less clear. Similarly, changing the negatively charged aspartic acid at position 121 to a negatively charged glutamic acid abolished caspase-1 cleavage in vitro but induced toxicity in cells (1). In addition, C-terminal αSyn interaction partners or posttranslational modifications could act as steric protectants (Fig. 1). Indeed, our previous work (23) suggested that the C-terminal interactor, synphilin-1, reduces αSyn truncation and aggregation in tg mice. Similarly, C-terminal modifications may protect cleavage sites and stabilize αSyn against a proteolytically accessible conformation. In this context, unlike the caspases, which exhibit sequence specificity for substrate cleavage, the substrate specificity of calpains is defined by both primary and higher-order structures.

Finally, human use of calcium channel blockers and antiinflammatory drugs may reduce the risk of developing PD (24), perhaps by inhibiting proteolytic cleavages of αSyn (Fig. 1). However, the potential αSyn-cleaving proteases discussed here have a complex and interacting range of substrates; for example, caspases are common substrates of calpains, and caspases cleave the calpain inhibitor, calpastatin. Together with the prior work reviewed above, the two new reports raise the question of whether there is a common αSyn proteolytic pathway and whether the resulting truncated αSyn species are the major neurotoxic entities. Clearly, more experiments, including in humanized models, are needed to clarify these questions, but we are one step closer to the goal of explaining the initiation of αSyn aggregation in the synucleinopathies.