Dysregulation of autophagy and mitochondrial function in Parkinson’s disease

Parkinson’s disease (PD) is the second most common neurodegenerative disease. Although most of PD cases are sporadic, more than 5 % are inherited [1]. The PD related genes mainly include (1) SNCA, the gene encoding α-synuclein, which is the main component of Lewy bodies [2]; (2) LRRK2, which encodes a multi-domain large protein. Variable LRRK2 mutants contribute to over 10 % of familial and about 3 % of sporadic PD cases [3]; (3) PINK1 and Parkin, which are genes involved in mitochondrial turnover and maintenance. Their mutations are associated with early-onset PD [4]. (4) DJ-1, whose mutants contribute to 1–2 % of autosomal recessive PD [5].

Given the role of autophagy in the elimination of cellular dysfunctional proteins/organelles, the contribution of impaired autphagy to PD has attracted increasing interest. Accumulating evidence indicates autophagy defects are involved in the PD pathogenesis. PD related VPS35 D620N mutant inhibits autophagy and impairs the trafficking of the autophagy protein ATG9A [6]. In addition, in the substantia nigra (SN) of PD patients, aberrant autophagy activity is identified [7]. Inducing autophagy by manipulating Polo-like kinase 2 or activiting chaperone-mediated autophagy (CMA) can reduce α-synuclein aggregation [8, 9]. Finally, PD related neurotoxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) can induce autophagy through CDK5-mediated phosphorylation of endophilin B1 [10]. Rotenone could impair mitophagy to damage neurons which can be rescued by DJ-1 overexpression [11].

Recently, increasing reports support the role of mitochondrial dysfunction in the degeneration of DA neurons during PD initiation and progression. Mitochondria are double membrane organelles and participate in multiple cellular processes, including calcium homeostasis, energy supply, metabolic synthesis and apoptosis [12]. Mutations in mtDNA are more frequent in PD patients when compared with the age-matched population [13]. Interestingly, one recent study reports that α-synuclein overexpression augments mitochondrial Ca²⁺ transient by enhancing endoplasmic reticulum-mitochondria interactions and physiological levels of α-synuclein are essential to maintain mitochondrial function and morphology [14]. MPTP and rotenone can disturb mitochondrial function, impair respiratory chain complex and induce oxidative stress [15, 16]. Thus, excessive mitochondrial stress in response to environmental toxins or impaired capability of neurons to remove damaged mitochondria through mitophagy may result in the death of neurons.

This review discussed the recent findings in the relationship among several PD related genes, autophagy and mitochondrial dysfunction (Table 1).

Autophagy, which derived from the Greek words for “self-digesting”, refers to the cellular catabolic process to remove the cytosolic components, including mainly dysfunctional organelles, misfolded proteins, and surplus or unnecessary cytoplasmic contents to lysosome for digestion [17].

Autophagy can be divided into three major types: macroautophagy, microautophagy and chaperone-mediated autophagy (CMA). This classification is dictated by the way in which the cellular components are being delivered to lysosomes [17].

Macroautophagy (MA) is the most common and well conserved form of autophagy, and enables the bulk degradation of cellular contents [18]. It mainly involves five steps: initiation, nucleation, elongation, fusion, and degradation [19]. In this process, the pre-autophagosomal structure (PAS) elongates from membrane of Golgi, mitochondria or endoplasmic reticulum and sequesters the cullular contents or organelles to form the autophagosome. Subsequently, the autophagosome fuses with lysosomes to allow the cargo to be degraded [20]. Mitophagy describes the clearance of damaged mitochondria by macroautophagy. Increasing evidence prove that dysregulation of mitophagy play important roles in neurodegenerative diseases [21].

Microautophagy is the process in which cytoplastic contents are directly engulfed into lysosomes for degradation without an intermediate autophagosome [22]. Also, microautophagy is constitutionally active under normal conditions and maintains the turnover of cellular nutrients via selective degradation of proteins and organelles [23]. Unlike macroautophagy, nutritional deprivation or stress cannot activate microautphagy [24]. Currently, knowledge about the role and function of microautophagy, especially in neurodegenerative disease, is limited and further studies are needed.

CMA is characterized by the high selectivity in substrates targeting [25]. The substrates for CMA present a unique pentapeptide motif, which is known as KFERQ. Notably, this motif is a pattern recognition motif associated with the hydrophobicity and charge of amino acid residues, rather than 100 % compliance with specific amino acid residues [26]. Interestingly, post-translational modification could perfect some KFERQ-similar motif for CMA through changing the hydrophobicity and charge of amino acid residues [27]. Briefly, CMA can be divided into four main steps: recognition, unfolding, translocation and degradation [28]. Cytoplastic heat shock cognate protein 70 (Hsc70) binds to the substrates after recognizing the KFERQ motif and delivers it to lysosomal surface, where the substrates could be unfolded with the help of Hsc70 and other chaperons. Afer that, the unfolded substrate is translocated into lysosome via lysosome associated membrane protein type 2A (LAMP2A) for degradation under the lysosomal acid environment.

Although the mechanism underlying PD pathogenesis is still unclear, mitochondrial dysfunction has been increasingly confirmed to be a vital contributor. The main characterizations of mitochondrial dysfunction include ROS over-production, ATP depletion, mitochondrial DNA depletion, caspase release and electron transport complex (ETC.) enzyme defection [29]. Neurotoxins applied to establish PD models, including MPTP and rotenone, damage mitochondrial function by inhibiting complex I activity [30]. One recent study specifically knockouts Ndufs4 gene in mice midbrain DA neurons and reports significant compromise in complex I activity. However, Ndufs4 deficit mice show no obvious neurodegeneration, no loss of striatal innervation and no movement impairment, but with increased vulnerability to MPTP, suggesting that impaired complex I activity is involved in the pathogenesis of PD [31]. Another study shows that complex I inhibition can lead to abnormal oxidition of optic atrophy 1 (OPA1), which in turn results in mitochondrial structural abnormalities and dopaminergic neurodegeneration. All these pathogenic changes are abolished by OPA1 upregulation, hinting that OPA1 is a novel therapeutic target for PD [32].

Mitophagy means the way in which damaged mitochondria are eliminated by autophagy. Through mitophagy, cells regulate both the number and quality of mitochondria in response to the metabolic stress. One recent study using genome-wide RNAi screen identifies sterol regulatory element binding transcription factor 1 (SREBF1), F-box and WD40 domain protein 7 (FBXW7) as key regulators for PINK/Parkin pathway, which is reponsible for mitophagy activition [33]. Moreover, SREBF1 is considered as a risk locus for sporadic PD [34]. Thus, this study hints that mitophagy may act as a common mechanistic link between sporadic and autosomal recessive PD. Mutation in F-box domain-containing protein Fbxo7 is related with early-onset autosomal recessive PD. One study reports that Fbox7 is involved in mitochondrial maintainance via directly interaction with Parkin to induce mitophagy, which further confirms the importance of mitophagy in PD pathogenesis [35].

Our study shows that the myocyte enhancer factor 2D (MEF2D) is present in neuronal mitochondria and heat shock protein 70 (Hsp70) interacts with MEF2D and mediates its mitochondrial translocation. Mitochondrial MEF2D is responsible for the expression of the gene NADH dehydrogenase 6 (ND6), which encodes an essential component of the complex I. Blocking MEF2D function leads to impaired complex I activity, oxidative stress and neuronal death. More importantly, in the postmortem PD brain samples, both mitochondrial MEF2D and ND6 levels are decreased compared with age-matched controls [36]. Our further study shows that MPTP caused a significant decrease in the half-life and total level of MEF2D mRNA. Down-regulation of MEF2D mRNA alone can reduces the viability of SN4741 cells and sensitizes the cells to neurotoxin [37].

Compared with age-matched AD and control brain samples, CMA markers LAMP2A and Hsc70 are significantly decreased in SN and amygdala of PD brains, hinting the role of autophagy in PD pathogenesis [91]. Furthermore, in peripheral blood mononuclear cells from PD patients, Hsc70 protein level is decreased in all PD groups, while glucocerebrosidase protein level is reduced only in the genetic PD groups [92]. This study suggests that Hsc70 and glucocerebrosidase may serve as a screening tool for PD diagnosis. In addition, lysosomal glucocerebrosidase protein level and enzyme activity are selectively reduced in the region with increased α-synuclein inside the PD brain of early stage. The loss of lysosomal glucocerebrosidase is directly related to reduced lysosomal CMA activity, increased α-synuclein and decreased ceramide, which suggests that compromise in lysosomal function contributes to PD pathology [93]. One study also shows that the fibroblasts from PD patients with the mutation G2019S have higher level of autophagy compared with wild type fibroblasts, which contributes to the more vulnerability to MPP⁺ [76]. However, another study using the primary cultured fibroblasts from sporadic PD patients detects no changes in autophagy [94]. Further, one study uses iPSC-derived DA neurons from PD patients with GBA1 mutation and finds that these neurons has autophagic and lysosomal defects as well as increase in glucosylceramide and α-synuclein levels [95]. Together, these findings provide evidence for a link between autophagic/lysosomal system and PD pathogenesis.

Currently, there is no effective treatment to halt the progression of neuronal loss in PD. Thus, developing effective therapeutic strategy will be of great importance and will likely result from a better understanding the relationship between macroautophagy, CMA, mitochondrial homeostasis and PD pathogenesis.

First, since PD is characterized by the presence of abnormal protein aggregation, enhancing macroautophagy to remove the redundant proteins and/or dysfunctional organelles is being considered as a potential therapeutic strategy. In vivo, enhancing autophagic activity through TFEB or Beclin-1 overexpression protects nigral neurons against α-synuclein-induced toxicity [96]. Coexpression of beclin-1 activated autophagy, reduced accumulation of α-synulein, and ameliorated associated neuritic alterations. The above data support that beclin-1 could be a promising molecular target for PD treatment. In addition, some autophagy-inducing chemical agents have been shown to decrease α-synuclein levels in cell and animal PD models [97, 98].

Furthermore, modulating CMA activity could provide benefits for PD treatment. In SH-SY5Y cells, generic overexpression of LAMP2A induces CMA activity and protectes DA neurons against α-synuclein neurotoxicity [9]. Moreover, in rat SN, overexpression of LAMP2A via lentivirus can reduce total levels of α-synuclein and ameliorate α-synuclein induced neuronal degeneration and stratal terminal loss [9]. Also, our data demonstrated restoration of CMA activity is essential to maintain the normal function of neuronal survival factor MEF2D, thereby regulating ND6 transcription and promoting mitochondrial complex I activity, which protect DA neurons from MPTP induced neurotoxicity [36, 37]. Furthermore, retinoic acid derivatives can activate CMA to protect cells against proteotoxicity and oxidative stress [99].

Finally, considering the contribution of mitochondrial dysfunction to the pathogenesis of PD, restoring it might be a way to prevent or treat PD. The first attempt is to enhance mitophagy. In drosophila, overexpression of Parkin increases mitochondrial activity, reduces proteotoxicity and extends lifespan [100]. Also, another possible way is modulate mitochondrial biogenesis, which would rejuvenate mitochondrial pool. The peroxisome proliferator-activated receptor gamma, coactivator 1 alpha (PGC-1α) is a potential target. PGC-1α positively regulates the expression of genes required for mitochondrial biogenesis and the cellular antioxidant responses. Also, expression of PGC-1α-regulated genes is low in SN neurons in early PD and overexpression of PGC-1α could suppress ROS and neurodegeneration [101]. Thus upregulation of PGC-1α is a candidate neuroprotective strategy in PD. However, targeted up-regulation of PGC-1α via adeno-associated virus (AAV) in SN region of mice showed detrimental effect [102, 103]. Thus, further studies need to clarify this discrepancy.