Genetic Evidence for Endolysosomal Dysfunction in Parkinson’s Disease: A Critical Overview

Parkinson’s disease (PD) is the second most common neurodegenerative disorder in the aging population, and no disease-modifying therapy has been approved to date. The pathogenesis of PD has been related to many dysfunctional cellular mechanisms, however, most of its monogenic forms are caused by pathogenic variants in genes involved in endolysosomal function (LRRK2, VPS35, VPS13C, and ATP13A2) and synaptic vesicle trafficking (SNCA, RAB39B, SYNJ1, and DNAJC6). Moreover, an extensive search for PD risk variants revealed strong risk variants in several lysosomal genes (e.g., GBA1, SMPD1, TMEM175, and SCARB2) highlighting the key role of lysosomal dysfunction in PD pathogenesis. Furthermore, large genetic studies revealed that PD status is associated with the overall “lysosomal genetic burden”, namely the cumulative effect of strong and weak risk variants affecting lysosomal genes. In this context, understanding the complex mechanisms of impaired vesicular trafficking and dysfunctional endolysosomes in dopaminergic neurons of PD patients is a fundamental step to identifying precise therapeutic targets and developing effective drugs to modify the neurodegenerative process in PD.


Introduction
Parkinson's disease (PD) is the second most common neurodegenerative disorder in the aging population [1][2][3][4]. It is clinically defined by the presence of bradykinesia in combination with either rest tremor and/or rigidity, and a clear beneficial response to dopaminergic therapy [5]. Neuropathologically, it is characterized by the loss of dopaminergic neurons in the substantia nigra (SN) and the presence of α-synuclein positive inclusions (Lewy bodies, LB) in surviving neurons [6][7][8]. At present, there are no approved treatments capable of slowing neurodegeneration in PD. Therefore, it is of paramount importance to shed light on the molecular mechanism causing PD neurodegeneration, because this knowledge is the indispensable prerequisite to identifying therapeutic compounds that can address the dysfunctional cellular machinery specific to this neurodegenerative disorder [9,10]. In the past two decades, PD etiopathogenesis has been linked with several deranged cellular mechanisms, ranging from mitochondrial impairment (PRKN, PINK1, PARK7) and ubiquitination defects (FBXO7) to dysfunction of the endolysosomal pathway (LRRK2, VPS35, VPS13C, ATP13A2) and synaptic vesicle trafficking (SNCA, RAB39B, SYNJ1, DNAJC6). In addition, significant parts of the risk genes associated with PD encode for endolysosomal and synaptic vesicle proteins, confirming a particular susceptibility of PD-related brain structures to the impairment of these pathways (Figure 1) [11][12][13][14][15][16][17][18]. Graphic representation of a dopaminergic neuron illustrating the relevant intracellular organelles, the involved monogenic PD genes (dark blue), and PD risk genes (light blue). In the represented model, a genetically determined impairment of the endolysosomal function in association with additional genetic and environmental pathogenic hits causes a defect in the degradation of misfolded proteins (including α-synuclein), which induces a compensatory autophagic response to address the disposal of the pathological proteins. However, a defective endolysosomal pathway implies an ineffective autophagic response, thus triggering a self-feeding vicious cycle of abnormal protein accumulation by dysfunctional engulfed lysosomes. As a consequence, the accumulation of Figure 1. Graphic representation of a dopaminergic neuron illustrating the relevant intracellular organelles, the involved monogenic PD genes (dark blue), and PD risk genes (light blue). In the represented model, a genetically determined impairment of the endolysosomal function in association with additional genetic and environmental pathogenic hits causes a defect in the degradation of misfolded proteins (including α-synuclein), which induces a compensatory autophagic response to address the disposal of the pathological proteins. However, a defective endolysosomal pathway implies an ineffective autophagic response, thus triggering a self-feeding vicious cycle of abnormal protein accumulation by dysfunctional engulfed lysosomes. As a consequence, the accumulation of misfolded aggregated proteins and engulfed lysosomes-autophagosomes leads to the progressive formation of pathological inclusions (Lewy bodies) and ultimately to neuronal distress and death. Created with BioRender.com.
Lysosomes are membrane-bound cytoplasmic organelles equipped with acid hydrolases whose major function consists in preserving cellular homeostasis by breaking down cellular (organized in autophagosomes) and extracellular (imported with endosomes) macromolecules and organelles into their fundamental components [19,20]. Lysosomal storage diseases (LSDs) are inborn metabolism defects often caused by loss-of-function (LOF) mutations in genes encoding lysosomal hydrolases. The consequent aberrant buildup of waste material causes lysosomal impairment, which can ultimately result in cellular dysfunction. Neurodegeneration is observed in association with several LSDs, clearly indicating that lysosomal impairment (or the toxic effect of mutant hydrolases) plays a relevant role in neuronal dysfunction and death [18,21]. Neurons display unique biological characteristics that may explain their increased susceptibility to the dysfunction of lysosomes and endosomal trafficking, such as their arborization and long projections, their post-mitotic state, and their highly complex interactions with other cells [22][23][24][25].
The scope of this review is to provide a critical overview of the genetic evidence supporting the central role of the endolysosomal pathway, that PD researchers can adopt as a solid starting point for future studies. We pursue this goal by reviewing monogenic PD forms and "strong" genetic PD risk variants associated with dysfunction of the endolysosomal pathway, then explore the rapidly evolving field of polygenic "weak" genetic risk variants affecting lysosomal genes, and finally present a unifying perspective linking this genetic evidence with the pathogenesis of PD [11][12][13][14][15][16][17].

Monogenic Causes of PD Associated with Endolysosomal and Vesicular Dysfunction
Early epidemiological studies revealed that 10-15% of PD patients have positive family history of the disease while most cases are sporadic [26]. Through family-based linkage analysis and, more recently, next-generation sequencing approaches, thirteen genes have been definitively implicated in the etiology of PD. Mutations in the SNCA [27,28], LRRK2 [29][30][31], and VPS35 [32,33] genes cause autosomal dominant forms, whereas mutations in the PRKN [34], PARK7 [35], and PINK1 [36] genes cause autosomal recessive forms. In addition, biallelic mutations in the ATP13A2 [37], PLA2G6 [38], FBXO7 [39,40], DNAJC6 [41], SYNJ1 [42], and VPS13C [43] have been reported as rare causes of early-onset parkinsonism with atypical clinical features. Finally, RAB39B gene mutations have been associated with a form of X-linked levodopa-responsive parkinsonism in combination with various degrees of intellectual disability [44]. The possibility of relying on these established genetic PD forms represents a solid starting point to build hypotheses and generate genetic disease models for exploring the dysfunctional mechanisms underlying the PD pathology. In the following paragraphs, we focus on those Mendelian genes that support the role of impaired endolysosomal and vesicular function in the pathogenesis of PD ( Figure 1).
The G2019S variant, the most common LRRK2 pathogenic mutation in the western world, has a prevalence of 1% in all PD cases worldwide and is extremely frequent in Berber Arabs (up to 37% of PD patients) and Ashkenazi Jews (up to 23% of PD patients) [47,[67][68][69][70][71]. The R1441C variant, the second most common LRRK2 mutation, is more frequent in Italy [72]. The R1441G is a founder mutation in the Basque population (up to 46% in
The clinical features of VPS35-related PD are similar to those of typical PD, with a relatively younger age of onset (46.9 ± 8.6 years). The reported motor features are usually classical (bradykinesia, rigidity, rest tremor) with asymmetric onset. Regarding non-motor features, dysautonomia, psychosis, and hallucinations are rare, and cognitive decline is observed in 15-30% of cases. The therapeutic response to dopaminergic therapy is good, though wearing-off and levodopa-induced dyskinesias appear in around 80% of cases [138][139][140][141]. Neuropathological findings remain to be determined [138][139][140][141]. 6

of 31
The VPS35 gene encodes a subunit of the heteropentameric retromer complex, located at the endosomal membrane, where it facilitates the endosome-to-Golgi and the endosome-to-plasma membrane transport [138,142,143]. Rodent-and iPSC-based studies are providing greater insight into VPS35-related neurodegeneration [90,92,138,144,145]. The D620N variant was found to be associated with impaired autophagy, possibly due to abnormal sorting of the ATG9A autophagy receptor and decreased autophagosome formation [138,144]. Furthermore, it determines a defective sorting of cathepsin D (encoded by CTSD, see below), resulting in lysosomal dysfunction, which potentially implies an impaired degradation of aggregation-prone proteins such as α-synuclein [138,145]. A third mechanism may be identified in mitochondrial degradation, since Vps35 physiologically facilitates mitophagy via mitochondrial-derived vesicles and the mutated Vps35 might enhance the activity of Dlp1, a protein involved in mitochondrial fission, thus resulting in mitochondrial fragmentation [138,146,147]. Vps35 D620N mutation was shown to enhance Lrrk2-mediated phosphorylation of Rab10 as well as autophosphorylation, suggesting that Vps35 may be an upstream regulator of Lrrk2. Overexpression of wildtype Vps35 was demonstrated in flies and murine models to rescue retromer-mediated defects, such as lysosomal enlargement, caused by Lrrk2 G2019S overexpression or Rab29 knockdown. These observations support a possible shared mechanism of Vps35, Lrrk2, and Rab29 in PD pathogenesis [138,148]. Importantly, the demonstration of a common shared altered mechanism of these proteins involving vesicular and endolysosomal trafficking will require further validation in human-based neuronal models hopefully capable of recapitulating PD pathology. In this direction, recent studies of human iPSC-derived neurons carrying the VPS35 D620N mutation showed decreased autophagic flux [149].

VPS13C (Vacuolar Protein Sorting 13 Homolog C)
Biallelic VPS13C mutations cause autosomal recessive early-onset PD (EOPD) [43,150]. The link between VPS13C mutations and PD was identified through a genome-wide association study (GWAS) in 2014 and confirmed in 2016 through homozygosity mapping and exome sequencing study of consanguineous families, when the homozygous c.8445 + 2T > G variant was identified in a Turkish patient with PD [43,151].
The gene VPS13C encodes an intermembrane lipid transfer protein localized at contact sites between the ER and late endosomes/lysosomes and on the outer mitochondrial membrane [43,158]. Vps13C regulates lysosomal homeostasis and controls mitophagy, modulating the Pink1/Parkin pathway in cellular models [43,159]. The neurodegeneration associated with the loss of VPS13C function thus seems primarily attributable to an alteration of lysosomal homeostasis and an upregulation of Pink1/Parkin-dependent mitophagy [43,153,154,159]. However, whether this impaired mechanism observed in vitro reflects the pathological process in VPS13C-PD brains is far from being clear yet.

ATP13A2 (ATPase Cation Transporting 13A2)
Biallelic ATP13A2 mutations have been associated with a multitude of phenotypes including Kufor-Rakeb disease (KRD), neuronal ceroid lipofuscinosis, hereditary spastic paraplegia, and an amyotrophic lateral sclerosis-like form [37,160,161]. KRD is a rare AR, levodopa-responsive, rigid-akinetic parkinsonism with atypical features including pyramidal signs, supranuclear gaze palsy, and cognitive decline. It was first described in 7 of 31 1994 in a Jordanian family, and was associated with ATP13A2 through linkage analysis in 2006 [37,162,163].
ATP13A2 encodes a lysosomal P-type ATPase which has been associated with the homeostasis of metal cations (e.g., Fe 3+ , Mn 2+ , Zn 2+ ), mitochondrial clearance, and, possibly, α-synuclein detoxification [171,184,185]. In cell lines, the endogenous level of Atp13a2 repeatedly appeared below immunodetection, despite using different currently available Atp13a2 antibodies. Therefore, most of the studies showing co-localization with lysosomal and late endosomal/intraluminal vesicle markers are based on overexpression of the protein. Atp13a2-deficient mice show sensorimotor deficits, and accumulation of insoluble α-synuclein in the brain, which is exacerbated by overexpression of the human wildtype α-synuclein [186]. Apoptosis dysregulation, mitochondrial dysfunction, and ER stress have emerged from animal and cellular models. Moreover, Atp13a2 LOF determines lysosomal dysfunction with defective polyamine export and autophagosome dysfunction, as effectively explored both in vivo and in vitro [176,[187][188][189][190][191].

RAB39B (RAB39B, Member RAS Oncogene Family)
LOF mutations in RAB39B were identified as a cause of X-linked recessive (XLR) early-onset parkinsonism (EOP) in 2014 through linkage analysis in three Australian brothers presenting intellectual disability (ID) with EOP, and confirmed via linkage analysis and direct RAB39B sequencing in a Wisconsin family with 13 males presenting a similar phenotype [44].
RAB39B encodes a neuronal GTPase involved in vesicular trafficking and recycling between synaptic terminals, endosomes, and the Golgi apparatus. Rab39B is thought to control many cellular functions including α-synuclein homeostasis and GluA2/GluA3 AMPAR subunit trafficking from the ER to the Golgi apparatus, as suggested by several murine, iPSC, and isogenic human embryonic stem cell models [44,[193][194][195]200,202,209,210].

SYNJ1 (Synaptojanin 1)
Biallelic SYNJ1 mutations were linked with autosomal recessive EOPD in 2013 through homozygosity mapping and exome sequencing, in two parallel studies describing an Iranian and an Italian family [42,211].

DNAJC6 (DnaJ Heat Shock Protein Family Member C6)
In 2012, biallelic DNAJC6 mutations were linked with juvenile-onset autosomal recessive parkinsonism through homozygosity mapping and exome sequencing in a Palestinian family [41].
To date, 21 cases of DNAJC6-related parkinsonism have been reported, carrying biallelic missense, nonsense, frameshift, or splicing mutations [41,[233][234][235][236][237][238][239]. DNAJC6related parkinsonism is typically characterized by disease onset in the second decade of life and rapid clinical progression until loss of walking within around 10 years from onset. Initial symptoms tend to be rest tremor and bradykinesia, followed by rigidity and postural instability. Although classic PD is a possible form of DNAJC6-related parkinsonism, there usually are atypical manifestations such as dystonia, pyramidal signs, postural tremor, dysarthria, anarthria, epilepsy, ID, and psychosis. Levodopa response is extremely variable and often burdened by dyskinesias. No consistent neuropathological features have yet been described [67,237,240].
In conclusion, monogenic evidence seems to indicate two different groups of genes within the endolysosomal pathway: those more involved in lysosomal-autophagic functions and those encoding for proteins related to the machinery of synaptic vesicles, which are considered a specialized form of recycling endosome present in neurons. The clinical phenotype and neuropathologic findings of the first group seem to recapitulate better the idiopathic form of PD (e.g., LRRK2, VPS35, VPS13C). Conversely, the second group is characterized by atypical phenotypes and, as far as is known, neuropathology (e.g., SYNJ1 and DNAJC6). Therefore, one could speculate that the best models to identify disease mechanisms and therapeutic compounds for the common idiopathic form of PD should be based on the first group of genes, to which GBA1 should also be added (see below). In this context, SNCA and RAB39B probably deserve a separate discussion: although they seem functionally more involved in synaptic function, they display widespread LBD, more in line with idiopathic forms of PD. Therefore, these genes may represent a bridging mechanism between the two groups described above (Figures 1 and 2, Table 1).
In conclusion, monogenic evidence seems to indicate two different groups of genes within the endolysosomal pathway: those more involved in lysosomal-autophagic functions and those encoding for proteins related to the machinery of synaptic vesicles, which are considered a specialized form of recycling endosome present in neurons. The clinical phenotype and neuropathologic findings of the first group seem to recapitulate better the idiopathic form of PD (e.g., LRRK2, VPS35, VPS13C). Conversely, the second group is characterized by atypical phenotypes and, as far as is known, neuropathology (e.g., SYNJ1 and DNAJC6). Therefore, one could speculate that the best models to identify disease mechanisms and therapeutic compounds for the common idiopathic form of PD should be based on the first group of genes, to which GBA1 should also be added (see below). In this context, SNCA and RAB39B probably deserve a separate discussion: although they seem functionally more involved in synaptic function, they display widespread LBD, more in line with idiopathic forms of PD. Therefore, these genes may represent a bridging mechanism between the two groups described above (Figures 1 and 2, Table 1).   ). The considered features are: age of onset (AOO), motor signs (bradykinesia, rigidity, rest tremor), cognitive problems (including neuropsychiatric symptoms and intellectual disability in the case of RAB39B), non-motor symptoms of PD including (dysautonomia, hyposmia, REM sleep behavior disorder), atypical signs (e.g., epilepsy, spasticity, dystonia, dysmorphisms), and therapy failure (poor response to levodopa treatment or early onset of LID and MF).

Genetic Risk Factors for PD Associated with Endolysosomal Dysfunction
Despite these great advances in the field of PD Mendelian genetics, a monogenic cause can be identified in only a minor fraction of PD cases. In general, non-monogenic PD is thought to be a multifactorial disorder influenced by genetic and environmental factors. The emergence of new technological approaches and the increasing sizes of international PD cohorts are leading to the identification of common variants with small effects contributing to PD [244,245]. Pathway analysis of the genes in which these variants are located has shown the involvement of several pathways including mitochondrial biology, inflammation/immune response, and lysosomal-autophagic functions [246,247]. Here, we focus on the risk genes involved in the endolysosomal function (Figure 1).
The GBA1 gene encodes the glucosylceramidase beta 1 (GCase), a lysosomal hydrolase that breaks glucosylceramide into glucose and acylsphingosine [257,258]. Biallelic GBA1 mutations cause Gaucher's disease (GD), a lysosomal storage disorder biologically characterized by a significant reduction in GCase activity, which leads to toxic accumulation of glucosylceramide and glucosylsphingosine [259]. Monoallelic GBA1 mutation carriers do not clinically develop GD nor present a pathological accumulation of glucosylceramide and glucosylsphingosine [250,260]. The link between monoallelic GBA1 mutations and PD emerged between 1985 and 1988 when parkinsonism was described as a possible neurological manifestation in GD patients and their relatives carrying heterozygous GBA1 mutations [260][261][262][263][264]. This link was confirmed by subsequent large-scale genetic studies, including GWAS [11].
Point mutations and complex alleles in GBA1 (resulting from rearrangement with the pseudogene GBAP1) are both reported as strong genetic risk factors for PD. Missense mutations are most represented, including the main causative mutations for GD (e.g., N370S and L444P) as well as PD risk variants that are not associated with GD (e.g., E326K and T369M) [256,[289][290][291][292][293][294]. Severity classification for GBA1 mutations considers their impact on the GD phenotype and correlates with the residual GCase activity. The degree of severity of GBA1 mutations, which can be defined as "severe" or "mild", has a differential effect on penetrance, age at onset, and clinical progression of PD [282,295]. However, it is still not completely clear whether this severity classification derived from GD patients can be automatically translated to the PD field [256,267,271,282,[295][296][297].
The disease mechanisms of GBA1-related PD are yet to be understood and may not perfectly match with those of GD, as there are PD risk variants that do not cause GD in homozygosis (e.g., E326K and T369M), and since no definitive correlation has been demonstrated to date between GCase activity and GBA1-related PD [256,290,292,293,298,299]. Nevertheless, the GCase protein was identified in 32-90% of Lewy bodies in patients with GBA1-related PD or Dementia with Lewy bodies (DLB) and less than 10% of Lewy bodies in patients with idiopathic PD, suggesting a possible interaction between GCase and α-synuclein [12,17,256,269,300]. Several pathogenetic hypotheses for GBA1-related PD have been proposed to date, following large studies on PD brains, Drosophila, murine models, cell models including midbrain-like organoids, and iPSC-derived dopamine neurons [256,301] The suggestions include different mechanisms that may depend on both gain-and loss-of-function mutations, including ER stress, ER-Golgi transport impairment, α-synuclein clearance impairment, lysosomal and autophagic dysfunction, lipid homeostasis disruption, mitochondrial dysfunction, and neuroinflammation [256,269,[301][302][303][304][305][306][307][308][309][310][311][312]. An intriguing hypothesis is that inactive GCase located on the surface of the lysosomal membrane may change the composition of glycolipids of the membrane affecting the lysosomal internalization of α-synuclein for degradation [97,313]. Another important study showed that lysosomal glucosylceramide accumulation due to GCase deficiency directly interacts with α-synuclein and leads to its accumulation, and vice versa, so that the pathological buildup of α-synuclein may inhibit the transport of GCase to the lysosome, inducing a toxic vicious cycle for neurons [303].
Pathogenic GBA1 variants alone are not sufficient to develop PD, because in groups of patients carrying the same mutations, some develop PD while others do not. Among GBA1 mutation carriers, the development of PD probably depends on other genetic and environmental factors. In this context, two risk-modifier variants (rs356219 in SNCA and rs1293298 in CTSB) were initially identified and relatively little impact was attributed to them, while a larger cumulative impact was reported for common variants affecting genes involved in lysosomal function [12]. Furthermore, other SNCA variants and TMEM175 M393T (see below) may influence the onset age of GBA1-PD [12]. Extending the research into risk-modifying factors to rare predicted deleterious variants in lysosomal genes, the largest contribution to the development of PD in GBA1 mutation carriers was attributed to a second deleterious variant in GBA1 or a deleterious variant in genes associated with mucopolysaccharidoses, evidencing the importance of the overall lysosomal burden in the development of PD [12,17,314].
SMPD1 variants L302P and P330fs, highly prevalent among Ashkenazi Jews, were repeatedly associated with PD in this population through case-control studies, as about 1.5% of PD patients carried these mutations, compared with 0.4% of controls [318][319][320][321][322][323]. Previous studies showed that reduced ASMase levels lead to α-synuclein accumulation [323]. It is therefore possible to assume that these mutations increase the risk of developing PD by reducing lysosomal localization of ASMase and thus causing accumulation of pathological α-synuclein [323].
Moreover, several other studies on European and Chinese populations identified additional SMPD1 variants suspected to be associated with PD [254,[324][325][326]. Nevertheless, it is important to note that only some of the SMPD1 mutations causing NPD type A/B were demonstrated to be associated with PD [323].

TMEM175 (Transmembrane Protein 175)
GWAS approaches in PD cases repeatedly identified an association peak on chromosome 4 (TMEM175/GAK/DGKQ locus) representing the fourth strongest risk locus. Two coding variants in this locus localized in the gene TMEM175 (M393T and Q65P) were found to be associated with PD. Transmembrane protein 175 is an integral membrane protein involved in potassium ion transmembrane transport in endolysosomes. Interestingly, the M393T variant was shown to be associated with reduced GCase activity. Therefore, TMEM175 variants are probably responsible for the GWAS association in the TMEM175/GAK/DGKQ locus, as also supported by a Mendelian randomization study [12,13,155,[327][328][329][330][331][332][333].

SCARB2 (Scavenger Receptor Class B Member 2)
SCARB2 variants have been repeatedly identified as risk factors for PD [12,334,335]. SCARB2 encodes the scavenger receptor class B member 2, which, among other endolysosomal activities, transports the GCase protein from the Golgi apparatus to the lysosome. However, unlike the TMEM175 M393T variant, the identified SCARB2 risk variants do not seem to affect GCase activity [335][336][337][338].

Polygenic "Lysosomal Burden"
An important genetic study using a sequence kernel association test (SKAT-O) investigated deleterious variant burden among lysosomal storage disorder genes using whole exome sequencing data from vast cohorts of PD cases and control subjects. The authors identified a significant burden of rare, potentially damaging lysosomal gene variants in PD patients compared with controls. The association persisted even when the GBA1 gene was excluded from the analysis, suggesting a significant "lysosomal burden" in idiopathic forms of PD. Consistent results were obtained in two independent replication cohorts. The lysosomal genes found to drive the association were GBA1, SMPD1, CTSD, SLC17A5, and ASAH1. Remarkably, in the discovery cohort, the majority of PD cases had at least one deleterious variant in an LSD gene, and 21% carried multiple damaging alleles [254]. Additional clues linking lysosomal dysfunction and "idiopathic" PD include the identification of a deficiency in the activity of the GCase enzyme in blood, CSF, and brain structures not only in GBA1-PD but also in idiopathic forms [339,340].
A recent study investigated the role of deleterious variants affecting LSD genes in modulating the penetrance of GBA1 risk variants in PD. The analysis in the discovery cohort revealed a significantly increased burden of deleterious variants in GBA1-PD patients compared to healthy GBA1 mutation carriers. Moreover, the authors demonstrated that the two strongest modifiers of GBA1 penetrance were a second variation in GBA1 (5.6% vs. 1.4%) and variants in genes causing mucopolysaccharidoses (6.9% vs. 1%) [17]. Furthermore, the largest GWAS on PD to date, based on about 40,000 PD cases including genetic and idiopathic forms, with 1.4M controls, found that an SNP in the gene GALC (rs979812) is associated with PD [11]. GALC encodes galactosylceramidase, a lysosomal hydrolase involved in ceramide catabolism, similar to ASAH1, GBA1, and SMPD1 [13,21,341,342]. Interestingly, the GALC rs979812 variant seems to be associated with increased enzymatic activity of galactosylceramidase [343].
Enrichment analysis based on the same GWAS study revealed that pathways related to lysosomal function are among the most significantly enriched [11]. Other endolysosomal genes identified through a GWAS approach were BAG3, CTSB, GPNMB, GRN, GUSB, HIP1R, LAMP3, NEU1, RAB29, SCARB2, SH3GL2, and TMEM175 [13]. Another GWAS demonstrated significant effects of BAG3, GBA, LAMP3, SCARB2, SNCA, and TMEM175 loci on age at onset of PD [12]. However, it is worth noting that large GWAS, for which perfect matching of cases and controls is virtually impossible, can result in spurious associations because of population stratification. This is also a major concern when studying recently admixed populations, particularly regarding variants with very small effect size [344]. Moreover, the scope of GWAS is limited to identifying genetic associations between PD status and genetic variants, tagging genomic regions encompassing several candidate genes. The "true" causal genes in each identified locus and the mechanisms by which they confer increased risk of PD often remain unclear. This important issue can be addressed by complementing GWAS data with quantitative trait loci (QTL) datasets correlating risk genotypes with gene expression, methylation, and proteomic data (i.e., Mendelian randomization studies-MRS); this knowledge of the molecular mechanisms by which genetic variants localized in PD risk loci increase the disease risk is a key step to translating genetic evidence into possible therapeutic targets. Remarkably, the link between endolysosomal impairment and synaptic dysfunction within PD derives also from these approaches, which have recently confirmed the causal role of several "lysosomal" and "synaptic" genetic hits (i.e., ARSA, CTSB, GALC, IDUA, RAB29, RAB7L1, SH3GL2, SMPD1, STX1B, TMEM175, VAMP42, and ZSWIM7) [333,343,[345][346][347][348][349].

Unifying Perspective and Conclusions
The previous paragraphs support the hypothesis that genetic variants of genes involved in endolysosomal function are major determinants of PD pathogenesis. Nonetheless, the existence of PD genes and risk variants in genes not directly involved in this pathway suggests that isolated dysfunction of the endolysosomal pathway is unable to explain the full picture, and additional genetic or environmental hits are important factors at play, especially in late-onset non-monogenic forms ( Figure 1).
It is an established fact that neuropathological studies of PD brains display abnormal aggregates of misfolded proteins, among which the major actor seems to be aggregated α-synuclein. Other pathological proteins have also been described, for example, aggregated tau in Lrrk2-PD patients [83] (Figure 1). A dysfunctional endolysosomal system may determine the reduction of α-synuclein turnover, increasing its cytoplasmic concentration and ultimately promoting the formation of oligomeric and fibrillary species, which are considered damaging for dopaminergic neurons [350,351]. An additional link to endolysosomal dysfunction is the presence of membranaceous structures in Lewy bodies, the pathological hallmark of the disease [352]. Moreover, dysfunctional endomembrane trafficking may induce the exploitation of alternative α-synuclein clearance routes, involving the plasma membrane and exocytosis, and ultimately promoting the spread of disease ("prion-like spread hypothesis" of PD) [353,354]. In our view, a vicious circle started by abnormal protein accumulation and subsequent activation of dysfunctional endolysosomal and autophagic processes may be the culprit of this neurodegenerative disease ( Figure 1). However, it should be noted that it remains open to debate whether α-synuclein aggregates should be portrayed as the cause or a result of the neurodegenerative process in PD. Taking this forward, a deficiency of functional monomeric α-synuclein has been proposed as a pathogenetic mechanism underlying PD ("proteinopenia" vs. "proteinopathy" hypothesis) [355]. In any case, the strong genetic evidence supporting the role of vesicular and endolysosomal impairment in PD should be carefully considered when formulating any new hypothesis for PD pathogenesis.
The causes of the selective vulnerability of dopaminergic neurons in PD are still unclear. Some peculiar characteristics of these neurons, such as the high metabolic-energetic demands, pace-making activity for the regulated release of dopamine, the pro-oxidant metabolism of this neurotransmitter, and the presence of long and complex arborizations, may suggest possible mechanisms [356]. In this scenario, normal endolysosomal and synaptic functions seem particularly important in the continuous process of recycling synaptic vesicles and the maintenance of long and complex neuronal projections, which are known to recede in PD brains in a dying-back process [357][358][359]. Conversely, high energy needs and oxidative stress management may indicate the mitochondrion as the prime suspect. Unsurprisingly, genetic evidence supports a very significant role also of the mitochondrial pathway (e.g., PRKN and PINK1 mutations) [356].
This review highlights the great importance of research in the field of PD genetics. Starting from the genetic findings accumulated over the past few decades and solidly based on that knowledge, a great number of in vitro and in vivo functional studies of disease models have been conducted, leading to an increased understanding of PD disease mechanisms. However, it is likely that some of the abnormalities observed in these models, although reproducible, are not pathogenetically linked to PD in humans. Some of the animal models do not present the neuropathological or phenotypical features of PD, calling into question the significance of the findings. It is probable that more advanced models, such as patient-derived midbrain organoids carrying specific gene mutations, represent more suitable tools to recapitulate the disease process that occurs in PD patients.
In conclusion, multiple lines of genetic evidence support a major role of vesicular and endolysosomal dysfunction in the pathogenesis of PD. As we have shown, several genes involved in these pathways have been demonstrated to be causative in monogenic forms of PD or have been shown to be associated with increased risk of PD. The crucial role of the endolysosomal system in PD has also generated therapeutic prospects for PD treatment. Indeed, several studies have suggested that the maintenance or upregulation of lysosomal activity may protect against neurodegeneration in PD [360]. In this view, promising therapeutic endolysosomal and autophagic targets include enzymatic activity enhancement (e.g., ambroxol), substrate reduction agents (e.g., venglustat), lysosomal activation (e.g., farnesyltransferase and mTOR inhibitors), and autophagic induction (e.g., trealose and nilotinib) [360]. Currently, many active therapeutic trials in humans have targetted this pathway in patients with genetic and idiopathic forms of PD [269,[361][362][363][364][365]. Hopefully, an increased understanding of the specific mechanisms by which dysfunction of these pathways causes PD may allow for the development of targeted drugs that can modify the invariably progressive course of PD.

Conflicts of Interest:
The authors declare no conflict of interest.