In the current study we have investigated the impact of WT and mutant forms of ASYN on lysosomal pathways in neuronal cells, as a possible pathogenetic mechanism for their toxic effects. For this purpose we have generated inducible rat PC12 and human SH-SY5Y cell lines expressing human WT and A53T ASYN, as well as two mutant forms (ΔDQ/WT and ΔDQ/A53T) that lack the CMA-targeting motif, and thus are not targeted to this pathway, do not interact with Lamp2a, and do not interfere with the degradation of other CMA substrates. We have used a similar approach in cortical neurons, where we have overexpressed these ASYN forms using adenoviral transduction.
In proliferating PC12 and SH-SY5Y cells, expression of A53T ASYN caused CMA impairment, as indicated by the marked decrease of total lysosomal degradation, in the face of unaltered macroautophagic, i.e. 3-MA-dependent, degradation. The fact that CMA impairment is responsible for lysosomal dysfunction in this setting was confirmed by the lack of changes in lysosomal degradation in cells expressing the double mutant ΔDQ/A53T. These data show for the first time in a cellular context that targeting of A53T ASYN to CMA is responsible for the decrease of total lysosomal degradation in neuronal cells. Interestingly, in these cycling cells, such lysosomal dysfunction was not associated with cell death or with compensatory induction of macroautophagy, as indicated also by the lack of changes in the ratio of LC3-I to LC3-II.
The situation in cortical neuron cultures was somewhat different, but still some essential features of the effects of A53T ASYN on lysosomal pathways were confirmed in this primary neuron setting. A53T ASYN again caused impairment of CMA, although on this occasion this did not lead to global lysosomal dysfunction, due to the compensatory activation of macroautophagy, which was identified through two different assays, the induction of 3MA-dependent degradation and the increase of the LC3-II to LC3-I ratio. In this case therefore, the activation of macroautophagy and the accumulation of autophagosomes were “productive”, in that they led to degradation of substrate proteins within lysosomes. A53T ASYN caused death that was in excess of that conferred by WT ASYN, and this death, as well as the macroautophagy induction, was abrogated with the double mutant ΔDQ/A53T. These data raised the possibility that the compensatory induction of productive macroautophagy may have deleterious consequences. This hypothesis was confirmed through pharmacological and molecular inhibition of this process, which led to improved survival. Taken together, these data support the concept that A53T ASYN causes toxicity in primary cortical neurons in part through CMA dysfunction and resultant aberrant macroautophagy activation. Thus, they confirm in a neuronal cell context our original hypothesis that mutant ASYNs may confer toxicity via CMA blockade 
, and for the first time provide conclusive evidence that the compensatory activation of macroautophagy mediates death in this setting.
In neuronally differentiated SH-SY5Y cells, WT ASYN expression overall led to similar effects with the expression of A53T in cortical neuron cultures. As we have previously reported (Vekrellis et al., in press), cell death occurred in this setting, unlike the situation in the cycling cells. Death was associated with CMA inhibition and increase of autophagosome formation, as assessed by the conversion of LC3-I to LC3-II. However, in this case, this was not associated with an increase of macroautophagy-dependent degradation, and therefore represented “non-productive” macroautophagy. Lysosomal changes did not occur with expression of the ΔDQ mutant, and death was significantly attenuated, confirming that CMA inhibition played a role in these effects. As in cortical neuron cultures with A53T ASYN expression, survival was increased with pharmacological or molecular inhibitors of macroautophagy. Therefore, even non-productive macroautophagy can lead to neuronal death.
These data in neuronally differentiated SH-SY5Y cells are important, because they indicate that, in certain settings, WT ASYN can also act as a CMA blocker and this effect can lead to cell death. This has also been suggested by Yang et al. 
and may relate to the fact that SH-SY5Y cells are dopaminergic, and dopamine may form adducts with WT ASYN that can act as CMA blockers in in vitro
. Consistent with this idea, our results suggest that inhibition of dopamine synthesis in this cell system leads to restoration of lysosomal function and improved survival (). This is significant, given that in the vast majority of PD patients it is the WT protein that is implicated as the pathogenetic agent.
Expression of A53T ASYN in neuronally differentiated SH-SY5Y cells led to more profound lysosomal dysfunction and accelerated death compared to WT ASYN expression. As with the WT protein, LC3-II accumulated, and this accumulation did not occur when the double mutant ΔDQ/A53T was expressed. However, A53T ASYN expression led also to profound inhibition of macroautophagy-dependent degradation, and this effect was not attenuated when ΔDQ/A53T was expressed. ΔDQ/A53T expression was marginally less toxic than A53T ASYN. These results suggest that A53T in this setting may also affect lysosomal function independent of CMA, through yet unknown mechanisms.
It is worth noting that WT and mutant ASYNs are toxic to differentiated but not proliferating neuroblastoma cells. The factors accounting for this difference could include subtle differences in the generated ASYN species, differences in clearance mechanisms, or involvement of cell cycle molecules or other proteins differentially expressed in the two states. It is possible that the lysosome plays a greater role in the degradation of proteins in the differentiated compared to the proliferating state. In support of this, total long lived lysosomal degradation (inhibited by Baf) in the cycling cells was detectable only after removal of serum from the medium (data not shown), suggesting that in their normal state these cells rely very little on lysosomes for protein turnover. Furthermore, the degree of lysosomal dysfunction caused by both WT and A53T ASYN is more prominent in the differentiated cells, and it is only in these cells that we detected LC3-II accumulation. Therefore, specific effects on lysosomes may mediate the preferential toxicity of aberrant ASYN in the neuronally differentiated state.
From our results, it appears that, depending on the exact context, CMA inhibition conferred by aberrant ASYN may, or may not, lead to induction of the process of macroautophagy. Massey et al. 
were the first to report that specific CMA inhibition may lead to activation of macroautophagy, and we have also confirmed it 
. In our current experiments, this effect, as assessed by LC3-II accumulation, occurred in cortical neurons and in differentiated SH-SY5Y cells, but not cycling cells. The reasons for these differences and the mechanisms through which such compensatory activation of macroautophagy occurs are unclear. It is worth noting however that in every case in which we observe such compensatory activation of macroautophagy, defined again as an increase of LC3-II to –I ratio, there is cell death, and this death is attenuated by macroautophagy inhibition.
This raises the issue of the exact nature of the lysosomal effects of aberrant ASYN that are linked to toxicity. In contrast to macroautophagy induction which, as mentioned above, correlates with toxicity, lysosomal dysfunction (as defined by impairment of lysosomal-dependent long-lived protein degradation) does not. For example, no generalized lysosomal dysfunction occurs in cortical neurons, and yet there is lysosome-dependent death, and the converse is true in cycling cells. However, in cases where profound generalized lsysomal dysfunction occurs (the case of A53T ASYN expression in differentiated SH-SY5Y cells), this does appear to influence cell viability. It would appear therefore that autophagosome production/formation, and not general lysosomal dysfunction, is mainly responsible for the toxicity observed following CMA inhibition by aberrant ASYN (). As mentioned, autophagosome formation can exert toxicity regardless whether it leads to “productive” macroautophagy or not. Therefore, the toxic event appears to be the formation and accumulation of autophagosomes, and not their fusion with lysosomes or the excess degradation of substrate proteins. Such toxic effects could be related to progressive damage and destabilization of the membranes of the accumulating autophagic vacuoles, leading to the cytoplasmic release of vacuolar hydrolases and cell death 
, or to impaired vesicular transport 
Schematic diagram of the lysosomal effects and relevant cell death pathways induced by aberrant ASYN.
The role of macroautophagy in cell homeostasis remains controversial since macroautophagy contributes to cell survival under stress such as starvation, but can also contribute to cell death 
. In particular as regards to neurodegeneration, this can apparently be induced both by lack and by excess of macroautophagy 
Regarding alpha-synucleinopathies, data, including our own 
and data presented here in supplementary Figure S2
, suggest that various forms of ASYN may be degraded by this process, and that, therefore, a strategy of macroautophagy induction may be beneficial in terms of removing such aberrant species and thus preventing their toxic effects 
. The data presented here though would argue that this represents an especially risky strategy, as, once ASYN levels accumulate and begin to exert toxic effects, they may do so in part via macroautophagy activation, and therefore further pharmacological macroautophagy activation may potentiate these effects.
It is worth noting in this context that toxicity induced by another genetic aberration leading to PD, that of mutant leucine rich repeat kinase 2 (LRRK2), has also been linked to macroautophagy induction. Inhibition of macroautophagy in differentiated SH-SY5Y cells reversed the detrimental effects of mutant G2019S LRRK2 on neuronal process length 
. Decrease on neuronal process length represents a prominent feature of the degenerative phenotype associated with PD-associated LRRK2 mutations. Therefore, activation of macroautophagy may represent a more general mechanism through which aberrant forms of proteins linked to PD cause neurodegeneration.
We do not wish to imply that the lysosomal alterations presented here represent the only mechanisms through which ASYN exerts its toxicity. It is clear that this is part of the picture and that other processes play a role. Interestingly, monomeric ASYN is sufficient to exert CMA blockade 
, and therefore the effects observed here cannot be accounted for by “fibrils” or “oligomers”. Other cellular effects mediated by such species may also be important for ASYN toxicity, as we and others have shown 
(Vekrellis et al., in press). Despite this, it is obvious that targeting the CMA pathway may provide some therapeutic benefits in PD. Improving CMA function may not only serve to accelerate ASYN degradation, but also to mitigate potential deleterious consequences of aberrant ASYN on this system.