Parkin and PINK1 are key regulators of mitophagy, an autophagic pathway for selective elimination of dysfunctional mitochondria. To this date, parkin depletion has been associated with recessive early onset Parkinson’s disease (PD) caused by loss-of-function mutations in the PARK2 gene, while, in sporadic PD, the activity and abundance of this protein can be compromised by stress-related modifications. Intriguingly, research in recent years has shown that parkin depletion is not limited to PD but is also observed in other neurodegenerative diseases—especially those characterized by TDP-43 proteinopathies, such as amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD).
1. Introduction
Frontotemporal dementia (FTD) is the second most common type of early onset dementia under 65 years of age, characterized by atrophy of the frontal and temporal lobes resulting in behavioral and/or language dysfunction
[1,2,3][1][2][3]. Amyotrophic lateral sclerosis (ALS; alias motor neuron disease, MND) is an incurable neurodegenerative and neuromuscular disease, with degeneration of motor neurons in the brain and spinal cord which leads to progressive weakness of muscles, and gradual paralysis followed by respiratory failure
[4,5,6][4][5][6]. A total of 10% of ALS cases and c.a. 30–40% of frontotemporal lobar degeneration (FTLD) cases are caused by genetic mutations, the remaining cases being sporadic
[7]. Due to extensive clinical and genetic overlap, FTLD and ALS are thought to form a continuum of endophenotypes
[8], and according to current diagnostic criteria, the term amyotrophic lateral sclerosis–frontotemporal spectrum disorder (ALS–FTSD) is used
[9].
The common hallmark of sporadic and the majority of the genetic forms of ALS–FTSD is TDP-43 (transactive response DNA-binding protein 43 kDa) pathology, characterized by TDP-43 protein depletion from the nucleus and its cytoplasmic aggregation
[10,11][10][11]. This pathology is present in the brain and spinal cord of 97% of ALS cases and c.a. 45% of FTLD cases
[12,13][12][13].
Parkinson’s disease (PD) is a neurodegenerative disease with progressive death of dopaminergic neurons in the
substantia nigra, the brain region producing neurotransmitter dopamine (DA). PD is clinically characterized by bradykinesia, resting tremor, and rigidity
[14]. Around 15% of patients have genetic forms of the disease due to mutations in
PARK(1–23) genes
[15,16][15][16].
Mitochondrial dysfunction has been extensively described in sporadic and genetic forms of ALS–FTSD and PD
[6]. In particular, the research on mitochondrial dysfunction in genetic PD caused by biallelic
PARK2 mutations
[17] has gained pace since the discovery of parkin as a key mediator of mitophagy in 2008
[18] (
Figure 1). Only two years earlier, in 2006, TDP-43 has emerged as an important protein for neurodegenerative diseases, such as ALS–FTSD, in which parkinsonian symptoms are reported
[11,19][11][19] (
Figure 1). Further, the observation of decreased parkin in TDP-43 proteinopathies is a relatively recent finding
[20,21][20][21] (
Figure 1), and its significance in terms of mitochondrial functioning has been hardly investigated. However, since these initial observations (
Figure 1) a plethora of new research that has been performed is presented below.
Figure 1. Timeline of the initial key research observations leading to the formation of our working hypothesis: parkin deficit observed in PARK2-related PD and patients with TDP-43 proteinopathies can lead to similar consequences.
2. Parkin and PINK1—The Key Players in Mitophagy Initiation
Mitophagy is a highly coordinated process whereby defective or old mitochondria are eliminated as whole organelles. This process occurs through the lysosomal pathway with the aid of the ubiquitin–proteasome system which concomitantly marks mitochondrial outer membrane proteins for disposal
[22,23][22][23]. Mitophagy has been studied in numerous cellular and animal models either as a naturally occurring process or upon induction with various mitochondrial stressors
[24]. Among mitophagy inducers, there are agents leading to mitochondrial membrane depolarization, mitochondrial respiratory complex inhibition, mutagenic stress, and proteotoxicity (e.g., CCCP, oligomycin, antimycin A, valinomycin, MPP+, and rotenone)
[25,26][25][26].
The first step in mitophagy initiation is performed by mitochondrial kinase PINK1 (PTEN-induced kinase 1) “a sensor of mitochondrial damage”
[26]. Under normal conditions, PINK1 expression levels are generally low in cells, because it is cleaved by protease PARL upon translocation to the inner mitochondrial membrane
[27[27][28][29],
28,29], followed by subsequent ubiquitination and degradation by the proteasome
[30]. The other key player is represented by parkin (Parkin RBR E3 Ubiquitin Protein Ligase), which resides predominantly in the cytoplasm. Its expression levels in mitochondria are also normally low, as reviewed in Reference
[31].
In response to mitochondrial stress, PINK1 normal processing is hindered. As a consequence, it gets anchored to the translocase of the outer mitochondrial membrane (TOMM) complex where subsequent PINK1 autophosphorylation leads to its auto-activation
[32,33][32][33].
Further to this, PINK1 recruits parkin along with ubiquitin onto damaged mitochondria and activates them through phosphorylation on Ser65
[33]. To enhance this process, the Ser65-phosphorylated ubiquitin reinforces parkin activation and stabilization on the OMM
[26].
Activated parkin ubiquitinates several mitochondrial outer membrane proteins (e.g., MFN2—mitofusin 2; TOMM20—translocase of outer mitochondrial membrane 20; and VDAC1—voltage dependent anion channel 1), tagging them for proteasome degradation with long ubiquitin chains
[26,34][26][34]. Subsequent phosphorylation of these long ubiquitin chains by PINK1 on Ser65 makes it impossible for ubiquitin specific peptidase 30 (USP30) to detach ubiquitin residues from OMM proteins
[35]. This generalized polyubiquitination of OMM proteins recruits, in turn, autophagy receptors, such as optineurin (OPTN) and sequestosome-1 (SQSTM1, p62), in such a way as to start forming an autophagosome around the damaged/old mitochondrion, eventually leading to its engulfment and subsequent digestion upon fusion with a lysosome
[33].
3. Loss-of-Function (LOF) and Gain-of-Function (GOF) Mechanisms in TDP-43 Proteinopathies
TAR DNA-Binding Protein-43 (TDP-43) was first identified in 2001 as a protein able to bind the HIV-1 TAR binding sequence
[61][36], and in 2006, as the main component of aggregates found in the brains of patients with ALS and FTLD
[11,19][11][19] (
Figure 1). Recently, many reviews have been focused on elucidating the role of this protein in disease and normal development, and, for this reason, the reader is referred to these works for a more detailed description
[10,62,63,64,65][10][37][38][39][40]. Briefly, TDP-43 belongs to the class of heterogeneous ribonucleoproteins (hnRNPs) that have been long referred to as the “RNA histones”. Normally, these proteins bind nascent RNA molecules and affect all aspects of RNA processing within the cell, from capping/splicing/polyadenylation to transport/translation and eventually degradation. At the structural level, TDP-43 possesses a highly structured N-terminus domain (NTD)
[66,67][41][42] that controls protein dimerization/oligomerization
[66,68][41][43]. This NTD is followed by two RNA Recognition Motifs (RRMs) that are responsible for sequence-specific binding to RNA
[69,70][44][45] and then by an unstructured C-terminus region that plays a fundamental role in phase separation and aggregation
[71,72,73,74][46][47][48][49].
In pathological aggregates, TDP-43 is subject to various post-translational modifications that include ubiquitination, phosphorylation, acetylation, sumoylation and is also cleaved to generate C-terminal fragments
[62][37]. From a pathomechanistic point of view, there are two major disease pathways that have been proposed to become disrupted by TDP-43 aggregation and modifications: gain- and loss-of-function disease mechanisms
[10]. The gain-of-function mechanisms may include various factors such as direct toxicity of the aggregates
[75[50][51][52][53],
76,77,78], direct toxicity of the C-terminal fragments
[79[54][55],
80], or indirect toxicity caused by sequestration of other proteins that are normally in close contact with TDP-43 in the cellular environment
[81,82,83,84,85][56][57][58][59][60]. Intriguingly, it is also possible that aggregates might be protective at least during the early stages of the disease. This hypothesis is supported in a recent study based on random mutagenesis of the TDP-43 C-terminus where it has been observed that mutations that increase hydrophobicity and aggregation can decrease toxicity in yeast cells
[85][60]. Alternatively, the loss-of-function pathological mechanisms might eventually occur through the sequestration of soluble TDP-43 in the aggregates, thus leaving not enough TDP-43 to perform its normal functions within cells. Therefore, this will result in multiple defects that range from preventing DNA damage to all aspects of RNA processing
[86][61]. In support of this view, many recent studies agree that alterations in RNA metabolism could be a major contributor to ALS/FTLD processes in humans
[87][62]. Most importantly, it should be kept in mind that all these gain- and loss-of-function possibilities do not necessarily exclude each other, and these different mechanisms could play a role at different stages of the disease. In conclusion, the emerging picture from all of these studies is that, following aggregation of TDP-43, a combination of RNA processing alterations might represent the principal disease contributor in patients with ALS and FTLD
[88,89,90][63][64][65].
4. Interwoven Relations between TDP-43 and Parkin
Interestingly, apart from Parkinson’s disease, decreased levels of parkin have been found in several TDP-43 proteinopathies (
Table 1/
Figure 2C). To this date, in fact, there are several examples where the deregulation of parkin expression or its cellular localization has been linked to TDP-43 complex neuropathology and/or has been observed to occur following manipulations of TDP-43 expression levels
[20,21,91,92,93,94][20][21][66][67][68][69] (
Table 1).
Table 1. Effects of TDP-43 pathology, TDP-43 depletion, or overexpression on parkin protein and mRNA levels.
TDP43-Proteinopathy Model |
Cell Type/Treatment Length |
Parkin mRNA/Protein |
Accompanying Changes |
References |
Patients with sporadic ALS (n = 12) vs. control subjects |
ca. 1000 motor neurons/- |
Trend for decreased mRNA (microarray) |
- |
[95] | [70] |
Patients with sporadic ALS (n = 11) vs. control subjects (n = 3) |
Spinal cord motor neurons—only those with TDP-43 inclusions/- |
Decreased protein (IF) |
- |
[21] |
Carriers of | PGRN | mutations from families with FTLD |
Human primary skin fibroblasts with | PGRN | mutations |
Decreased mRNA/protein by ca. 60% (qRT-PCR) |
Unchanged MFN2 and VDAC1 mRNA and protein |
[94] | [69] |
Mouse TDP-43 knockdown |
Striatum injection of antisense oligonucleotides/2 weeks |
Decreased mRNA by ca. 70% (RNAseq) |
- |
[20] |
Mouse TDP-43 knockdown |
Brain and spinal-cord injection of antisense oligonucleotides/2 weeks |
Decreased mRNA by ca. 80% (qRT-PCR) |
- |
[21] |
TDP-43 knockdown in human neurons (TDP-43 expression reduction by 60–75%) |
Human neurons (iPSC-derived and HUES6 line) lentiviral shRNA constructs/na |
Decreased mRNA by ca. 25% (qRT-PCR) |
- |
[21] |
TDP-43 silencing (siRNA) in HEK293T |
Human HEK293T (DMSO vs. mitochondrial uncoupler CCCP; siTDP-43 or si CTRL)/na |
Decreased protein cytoplasmic localization (IF) |
Decreased prohibitin 2 (PHB2) |
[91] | [66] |
TDP-43 silencing (siRNA) in skin fibroblasts derived from patients with FTLD |
Human primary skin fibroblasts with | PGRN | mutations and control fibroblasts (siTDP-43 or siCTRL)/48 h |
Decreased protein by ca. 40% (WB) |
- |
[94] | [69] |
Overexpression of wild-type TDP-43-HA or mutant TDP-43-Q331K |
Primary mouse neurons/motor cortex and human HEK293T cells/48 h |
Decreased endogenous parkin mRNA and protein by c.a. 50% (qRT-PCR, WB) |
Increased PINK1 protein |
[92] | [67] |
Exogenous co-expression of wild-type TDP-43-HA and intron-free human parkin or intron-free PINK1 |
Human HEK293T cells/48 h |
Decreased intron-free parkin mRNA and protein by c.a. 50% (qRT-PCR) |
Increased cleaved PINK1 protein forms insoluble cytoplasmic aggregates |
[92] | [67] |
Transgenic | Drosophila | knock-in of wild-type human TDP-43-H |
Fly heads/na |
Decreased mRNA and protein by c.a. 45% (qRT-PCR, WB) |
- |
[92] | [67] |
Wild-type TDP-43 overexpression |
Human HEK293T (DMSO vs. mitochondrial uncoupler CCCP; wild-type pLX-TDP-43-v5 vector/na) |
Increased protein cytoplasmic localization (IF) |
Increased prohibitin 2 (PHB2) |
[91] | [66] |
Wild-type TDP-43 overexpression |
Human primary skin fibroblasts with transiently silenced PGRN (48 h) overexpressing wild-type flag-TDP-43 (24 h) |
Increased protein by c.a. 40% (WB) |
Increased PGRN protein |
[94] | [69] |
Wild-type TDP-43 overexpression |
Human skin fibroblasts with | PGRN | mutations overexpressing wild-type flag-TDP-43 (48 h) |
Decreased protein by c.a. 50% (WB) |
- |
[94] | [69] |
Transgenic mouse with heterozygous knock-in of human mutant TDP-43 (A315T) |
Whole-brain tissue |
mRNA and protein reduced by 70% compared to wild-type controls (qRT-PCR, WB) |
Abnormal neuronal mitochondrial cristae, fusion and fission defects; |
[96] | [71] |
Overexpressed wild-type TDP-43 |
Human M17 neuroblastoma cells |
Increased protein by c.a. 50% (WB) |
- |
[93] | [68] |
Transgenic mouse with knock-in of human mutant TDP-43A315T |
|
Increased mRNA by c.a. 50% (qRT-PCR) |
- |
[93] | [68] |