REVIEW article

Front. Neurosci., 04 June 2020

Sec. Neurodegeneration

Volume 14 - 2020 | https://doi.org/10.3389/fnins.2020.00556

LRRK2-Related Parkinson’s Disease Due to Altered Endolysosomal Biology With Variable Lewy Body Pathology: A Hypothesis

  • 1. Institute of Parasitology and Biomedicine “López-Neyra”, Consejo Superior de Investigaciones Científicas (CSIC), Granada, Spain

  • 2. Life Sciences Institute, University of Michigan, Ann Arbor, MI, United States

  • 3. Department of Anesthesiology, New Jersey Medical School, Rutgers, The State University of New Jersey, Newark, NJ, United States

Abstract

Mutations in the gene encoding for leucine-rich repeat kinase 2 (LRRK2) are associated with both familial and sporadic Parkinson’s disease (PD). LRRK2 encodes a large protein comprised of a GTPase and a kinase domain. All pathogenic variants converge on enhancing LRRK2 kinase substrate phosphorylation, and distinct LRRK2 kinase inhibitors are currently in various stages of clinical trials. Although the precise pathophysiological functions of LRRK2 remain largely unknown, PD-associated mutants have been shown to alter various intracellular vesicular trafficking pathways, especially those related to endolysosomal protein degradation events. In addition, biochemical studies have identified a subset of Rab proteins, small GTPases required for all vesicular trafficking steps, as substrate proteins for the LRRK2 kinase activity in vitro and in vivo. Therefore, it is crucial to evaluate the impact of such phosphorylation on neurodegenerative mechanisms underlying LRRK2-related PD, especially with respect to deregulated Rab-mediated endolysosomal membrane trafficking and protein degradation events. Surprisingly, a significant proportion of PD patients due to LRRK2 mutations display neuronal cell loss in the substantia nigra pars compacta in the absence of any apparent α-synuclein-containing Lewy body neuropathology. These findings suggest that endolysosomal alterations mediated by pathogenic LRRK2 per se are not sufficient to cause α-synuclein aggregation. Here, we will review current knowledge about the link between pathogenic LRRK2, Rab protein phosphorylation and endolysosomal trafficking alterations, and we will propose a testable working model whereby LRRK2-related PD may present with variable LB pathology.

Introduction

Parkinson’s disease (PD) is a progressive neurodegenerative disorder affecting 1% of the population above the age of 60 years, and 5% of individuals over the age of 85 (). It is generally characterized by two major neuropathological hallmarks including neuronal loss, which mainly affects dopaminergic (DA) neurons in the substantia nigra pars compacta (SNpc), and the presence of proteinaceous inclusions known as Lewy bodies (LBs) and Lewy neurites rich in α-synuclein in many of the surviving neurons. The loss of these DA neurons results in the classical motor symptoms of PD, including shaking, rigidity, and slowness of movement ().

The molecular events leading to the loss of DA neurons are not well understood, and their identification is complicated by the fact that around 90% of PD cases are sporadic, meaning that there is no apparent underlying cause. However, the identification of monogenic forms of PD, in which autosomal-dominant, or autosomal-recessive mutations in certain genes cause the disease with variable penetrance is of great importance to PD research, as it allows for the generation of cellular and animal models carrying the mutations to study the mechanisms implicated in the disease ().

Point mutations in the leucine-rich repeat kinase 2 (LRRK2) gene are the most frequent cause of familial, autosomal-dominant Parkinson’s disease (PD; ; ; ; ), and sequence variations in LRRK2 are known to modify PD risk, indicating that it also plays a role in the most common sporadic form of the disease (; ). In addition, patients with LRRK2 variations present with late-onset disease and core clinical features indistinguishable from sporadic PD (). As a result, LRRK2 has become the subject of intense studies to understand some of the cellular processes that contribute to disease pathogenesis.

Leucine-rich repeat kinase 2 is a large protein which belongs to the ROCO protein family, characterized by the presense of a ROC (Ras-of-complex) GTPase, a COR (C-terminal of ROC), and a kinase domain. Apart from such catalytic core, it contains a series of protein-protein interaction domains including N-terminal armadillo, ankyrin and leucine-rich repeats, as well as C-terminal WD40 repeats (Figure 1). Over 100 LRRK2 variants have been described, and a small set of those have been shown to be pathogenic, including R1441C/G/H and N1437H in the ROC domain, Y1699C in the COR domain, and G2019S and I2020T in the kinase domain, respectively (; Figure 1). The G2019S mutation is the most common, and has been found in both familial and sporadic PD cases. In contrast to all other pathogenic LRRK2 mutations which are highly penetrant, the G2019S variant displays significantly reduced penetrance which increases with age (; ; ). This is consistent with the idea that PD can be attributed to a combination of genetic, environmental and age-related factors, and indicates that the G2019S LRRK2 variant serves as an ideal model system to investigate mechanisms underlying sporadic PD pathogenesis ().

FIGURE 1

Whilst only the G2019S LRRK2 mutation seems to increase LRRK2 kinase activity when assayed in vitro (West et al., 2005, 2007; ; , ; ; ; ; ; ; ; ; ; ; ), all pathogenic LRRK2 mutants converge on enhancing LRRK2 kinase substrate phosphorylation when assayed in vivo (). Highly potent, selective and brain-permeable kinase inhibitors have been developed and are in various stages of clinical development (West, 2017). At the same time, extensive research efforts are under way to gain detailed knowledge of the cellular deficits mediated by pathogenic LRRK2, as this may point toward additional targets for disease-modifying therapeutics.

A multitude of studies indicate that pathogenic LRRK2 alters vesicular trafficking events which all ultimately impact upon proper endolysosomal functioning (, ; ; ; , ; ; Yakhine-Diop et al., 2014; ; ; ; ; ; ; ; ). Phosphoproteomic studies have identifed a subset of Rab proteins including Rab1, Rab3, Rab5, Rab8, Rab10, Rab12, Rab29, Rab35, and Rab43 as physiological substrates for LRRK2 (; ; ; ). Since Rab GTPases are key players of all intracellular vesicular trafficking steps (Wandinger-Ness and Zerial, 2014), this raises the possibility that pathogenic LRRK2 may cause such endolysosomal deficits by impairing the functioning of one or several of these Rab substrates. The lysosome is key for the degradation of a large variety of proteins, protein aggregates, and defunct organelles. Therefore, aberrant lysosomal functioning mediated by pathogenic LRRK2 may be expected to cause the buildup of protein aggregates including α-synuclein, yet a significant portion of LRRK2-related PD cases do not display any apparent LB pathology. Here, we will summarize current knowledge about the link between pathogenic LRRK2 activity, Rab phosphorylation and concomitant lysosomal deficits, which may contribute to cell death associated with LRRK2-related PD in either the absence or presence of α-synuclein pathology.

Protein Aggregation in Sporadic and Familial PD

The presence of proteinaceous inclusions caused by the aggregation of misfolded proteins in distinct regions of the brain is a common neuropathological feature of many neurodegenerative disorders. Tau is a microtubule-binding protein which accumulates as neurofibrillary tangles in the brains of patients with Alzheimer’s disease (AD) as well as other neurodegenerative diseases including sporadic PD (Zhang et al., 2018). Intracellular aggregates of α-synuclein are common to all α-synucleinopathies and are called LBs in PD and in Dementia with Lewy Bodies (DLB), glial cytoplasmic inclusions in multiple system atrophy (MSA), and axonal spheroids in several less well-characterized neuroaxonal dystrophies (). Alpha-synuclein pathology is observed in over 50% of autopsy-confirmed AD brains (), and both α-synuclein and AD-related pathology can be detected in up to 25% of cognitively healthy elderly subjects (). The observed pathology in healthy individuals may reflect a presymptomatic disease stage, or the existence of diverse α-synuclein strains not equally toxic to cells (). In either case, these findings indicate that neither LB nor tau pathology are specific to a given disease entity.

Current criteria for the pathological diagnosis of sporadic PD include neuronal loss in the SNpc accompanied by LB pathology. In support of this association, there are familial PD cases where LB pathology is consistently present, including α-synuclein-related, or glucocerebrosidase (GBA)-related parkinsonism. Pathogenic missense mutations in α-synuclein, as well as triplication of the α-synuclein locus, cause autosomal-dominant PD with nearly complete penetrance, whilst duplication of the α-synuclein locus causes late-onset PD with reduced (30–50%) penetrance and variable clinical presentations, even within families (). Thus, α-synuclein dosage seems to be important, but may not be the only contributor to disease manifestation. Neuropathological examinations of clinically affected patients due to either α-synuclein mutations or multiplications reveal prominent cell loss in the SNpc, associated with severe LB pathology in all cases (; Figure 2). Where analyzed, around 50% of cases also display tau pathology, which is similar to the percentage of tau pathology found in sporadic PD cases (; Zhang et al., 2018).

FIGURE 2

Homozygous mutations in the GBA gene, which encodes glucocerebrosidase (GCase), a lysosomal enzyme involved in the metabolism of glycosphingolipids, cause Gaucher disease, the most common autosomal-recessive lysosomal storage disease (LSD). Heterozygous GBA mutations are the strongest genetic risk factor for sporadic PD, and can also increase risk for other α-synucleinopathies. Reduced GBA activity per se is not a necessary cause of PD, since most patients heterozygous or homozygous for GBA mutations never develop the disease (). However, LB pathology is prominently observed in virtually all GBA-heterozygous mutation carriers manifesting PD (Figure 2), and where analyzed, is accompanied by tau pathology (). In addition to heterozygous GBA carriers, the risk of developing PD is also increased in various other LSDs due to mutations in different enzymes which metabolize lipids within lysosomes (). Importantly, in the brains of those patients who develop parkinsonism, the consequent accumulation of several distinct lipids is associated with α-synuclein pathology (Figure 2), as seen in Niemann-Pick disease type C1, Fabry’s disease, Krabbe’s disease, or Sandhoff disease, as well as in several rodent models carrying mutations in such LSD genes (; ; ; ; ).

However, and perhaps surprisingly, not all PD-related gene mutations present with α-synuclein pathology. An increasing amount of evidence from LRRK2 patients indicates that PD diagnosis can be associated with a lack of LB-containing neuropathology (). Whilst all LRRK2-PD patients display neuronal loss in the SNpc, around 50% of patients do not show any apparent LB pathology, with variable tau pathology in most cases (; ; ; ; ; ). The absence of LB pathology is observed in G2019S LRRK2-PD carriers, as well as in carriers of the highly penetrant LRRK2 mutations such as R1441C/G, Y1699C, or I2020T (). Consistent with these neuropathological findings, only around 40% of G2019S LRRK2-PD patients display increased α-synuclein aggregates in the cerebrospinal fluid (CSF), in comparison with around 90% of sporadic cases (). Thus, LRRK2-related PD can be grouped into two classes, one with and one without LB pathology (Figure 2), both accompanied by DA cell loss in the SNpc.

The underlying reason for this remains unknown, but may relate to the presence of additional cellular alteration(s) present in the LB-positive LRRK2-PD patients. Genome-wide association studies (GWAS) have established that LSD gene variants increase risk of developing PD (; ; ; ; ). Strikingly, a significant percentage of sporadic PD cases carried one or multiple putative pathogenic variants in distinct LSD genes (). Thus, the LRRK2-PD patients who do display LB pathology may carry additional pathogenic LSD gene variants, with the concomitant lipid accumulation causing α-synuclein aggregation (Figure 2). In support of this possibility, a recent study employing a small set of G2019S LRRK2 patient-derived fibroblasts and iPSC-derived DA neurons revealed a decrease in GCase activity as compared to healthy control cells (Ysselstein et al., 2019). Such reduced GCase activity may play a role as disease modifier, and it will be important to assess GCase activity from a larger set of LRRK2 PD patient-derived cells, and determine whether the decrease in GCase activity correlates with a detectable increase in the lipids which serve as substrates for the enzyme, which may then be able to trigger α-synuclein accumulation.

Glucocerebrosidase is a lysosomal enzyme involved in the metabolism of glycosphingolipids. Loss of GCase enzyme function causes cellular accumulation of glycosphingolipids such as glucosylsphingosine and glucosylceramide, as well as related metabolites and cholesterol (; Xu et al., 2010). A secondary accumulation of such glycosphingolipids is also observed in several other LSDs (Xu et al., 2010), highlighting the possibility that increased glycosphingolipid levels may trigger α-synuclein oligomerization in the brains of those LSD patients. Bis(monoacylglycero)phosphate (BMP) is a negatively charged glycerophospholipid specifically found in endosomal and lysosomal membranes and thought to play a role in glycosphingolipid degradation and cholesterol transport (). Secondary accumulation of BMP has also been reported in several LSDs including Gaucher disease, mucopolysaccharidosis, Niemann-Pick disease, Fabry disease, and gangliosidoses (; ; ; ; ). Interestingly, recent studies have reported an increase in BMP levels in the urine from LRRK2 mutation carriers as compared to non-carriers, and slightly higher BMP levels in those LRRK2 carriers with PD as compared to those without disease (). In the future, it will be important to determine whether increased BMP levels correlate with an increase in α-synuclein aggregates in the CSF of those LRRK2 mutation carriers (). In either case, these data are consistent with the notion of altered lipid levels also in LRRK2-PD patients.

In sum, and at least with respect to LRRK2-related PD, current findings indicate that LBs are neither necessary nor sufficient for cell loss and the clinical expression of parkinsonism. Similarly, tau pathology per se may not be the major driving force at least in some cases, as LRRK2-related cell death can occur in the absence of LBs and the presence of a very low burden of tau pathology in the midbrain (; ). Finally, the present data indicate that the formation of LBs occurs in a manner independent of pathogenic LRRK2 kinase activity, but possibly mediated by altered lipid metabolism.

Alpha-Synuclein and PD

Alpha-synuclein is a small, natively unfolded protein which plays an important role in regulating synaptic vesicle release (; ). It is prone to aggregation, giving rise to oligomeric, or protofibrillar structures. Elevated levels of α-synuclein oligomers are observed in the CSF of sporadic PD patients as compared to control subjects (; ), supporting their role in PD pathogenesis. Alpha-synuclein oligomers comprise a very heterogeneous population, and not all of them are toxic to cells. However, for reasons that have not been fully elucidated, certain oligomeric species can be highly toxic, and can disrupt presynaptic transmission, the functioning of several organelles as well as proper protein degradation pathways (; ). The oligomerization of α-synuclein precedes its aggregation into mature LBs, which as such may play a cytoprotective role.

Alpha-synuclein can act as a prion protein, spreading throughout the nervous system. Indeed, different studies have found LBs in fetal grafted neurons in PD patients (; ). Alpha-synuclein oligomers can be released as free-floating proteins or via exosomes. They can then infect adjacent, healthy neurons, where they can act as a seed to induce intracellular α-synuclein accumulation, thereby further contributing to disease progression (, ). In addition, innate-immune astrocytes and microglia can take up α-synuclein oligomers, which may cause multiple proinflammatory changes which are neurotoxic as well. For example, microglial uptake of α-synuclein has been shown to induce the activation of NADPH oxidase, leading to elevated ROS production and contributing to tissue damage over time (Zhang W. et al., 2005).

According to the Braak hypothesis, PD progresses across synaptically connected neuroanatomical locations in the brain, with the appearance of LBs in distinct areas reflecting the distinct clinical manifestations (). However, only around 50% of PD patients have patterns of LB pathology consistent with the Braak staging system (; ), and some sporadic () as well as familial cases (see above) have no discernible LBs at all. It remains possible that these patients are negative for LBs, but display the presence of toxic, oligomeric α-synuclein species. To address this possibility, postmortem material from LRRK2-PD patients negative for LB could be analyzed for the presence of oligomeric forms of α-synuclein, for example employing a proximity ligation assay to detect such α-synuclein species (). Alternatively, α-synuclein aggregation may not mediate cell death, at least in LB-negative PD patients. Rather than via a propagated pathogen, cell death may be triggered by cell-autonomous or regionally autonomous mechanisms unrelated to protein aggregation ().

Mechanisms Underlying α-Synuclein Oligomerization: a Hypothesis

Whilst point mutations in α-synuclein have been shown to promote increased formation of oligomers in vitro (), the molecular mechanism(s) responsible for the formation of wildtype α-synuclein oligomers are less clear. The enhanced oxidative stress inherent to DA neurons in the SNpc may increase its propensity to oligomerize (), and various post-translational modifications such as nitration or phosphorylation have been reported to enhance oligomerization (; ; ; ).

In addition, impaired degradation may lead to the formation of oligomeric α-synuclein species. The ubiquitin-proteasome system and the autophagy-lysosomal pathway are the two major systems mediating protein degradation in eukaryotic cells. Whilst the proteasome, a multi-subunit protease complex which selectively degrades ubiquitin-tagged proteins, may contribute to the degradation of a proportion of α-synuclein species, most studies indicate that the major route for the degradation is via the autophagy-lysosomal pathway (). Different types of autophagy have been described, including chaperone-mediated autophagy (CMA), macroautophagy and microautophagy, and all have been implicated in the degradation of distinct α-synuclein species. CMA is the process by which substrate proteins are selectively targeted to lysosomes and translocated into their lumen through a coordinated action of chaperones and LAMP2A, which translocates substrates directly into the lysosomal lumen (; ). Endogenous wildtype α-synuclein has been shown to be efficiently degraded by CMA (), and inhibition of this pathway leads to the formation of α-synuclein oligomers, perhaps secondary to increasing α-synuclein levels (). In contrast, pathogenic mutant α-synuclein, as well as dopamine-modified, or oligomerized α-synuclein species are poor CMA substrates (; ). Whilst a role for macroautophagy in the clearance of α-synuclein oligomers is supported by various studies (Webb et al., 2003; ; Yu et al., 2009), others have questioned its relevance. For example, distinct compounds which inhibit lysosomal degradation were found to increase α-synuclein oligomers, but inhibition of macroautophagy was without effect (). Similarly, genetic inhibition of macroautophagy in a rodent model did not lead to detectable α-synuclein aggregates (). Thus, it is tempting to speculate that in addition to macroautophagy, at least a portion of α-synuclein oligomers may be degraded via microautophagy.

In mammalian cells, microautophagy occurs in late endosomes/multivesicular bodies (LE/MVB) rather than in lysosomes, and has been termed endosomal microautophagy (eMI; ). This process allows for bulk degradation of small protein oligomers, lipid droplets, mitochondria, or even portions of the nucleus (; ; ; ). Cargo is bound to the LE membrane, and trapped into intraluminal vesicles dependent on components of the ESCRT complex, a protein complex required for MVB formation, but independent of LAMP2A, the CMA receptor (; ; ). Some substrates can be directly degraded in LE, whilst others are degraded in the lysosome upon LE-lysosomal fusion (). In the future, it will be important to obtain direct experimental evidence for the possible relationship between eMI and PD. However, a role for eMI in α-synuclein oligomer degradation is consistent with its localization to the membrane of intraluminal vesicles of LE/MVBs in axons and cell bodies as determined by correlative light and electron microscopy in neurons overexpressing the tagged protein (). Importantly, eMI is regulated by direct alterations in the lipid composition of the LE membrane, specifically by lipid raft-like regions rich in sphingolipids and cholesterol (). Thus, alterations in the lipid composition of LE membranes mediated by GBA and several other LSDs may cause deficits in eMI-mediated degradation of oligomeric α-synuclein species.

Finally, the lipid composition of the membrane may directly contribute to α-synuclein oligomerization independently of effects on protein degradation. Alpha-synuclein monomers are known to bind to synthetic lipid membranes which causes them to adopt an α-helical conformation prone to forming different types of oligomers (; ; ). The membrane lipid composition seems to play an important role, with α-synuclein showing preference for binding to lipid rafts (; ; ), and recent studies show that the sphingolipid glycosylceramide can trigger α-synuclein oligomerization (Zunke et al., 2018).

In sum, we propose that α-synuclein aggregates may be a common neuropathological readout, but due to distinct underlying mechanisms (Figure 3). In familial PD cases due to α-synuclein mutations, it may be due to an enhanced propensity of the mutant protein to oligomerize, and/or impaired CMA-mediated turnover. In PD cases due to α-synuclein multiplications, oligomerization may occur as a result of increased protein levels. In PD cases due to mutations in GBA, increased spingolipids may directly trigger α-synuclein binding and oligomerisation, and/or indirectly cause an accumulation of oligomers due to impaired eMI mediated by alterations in the raft-like properties of the LE/MVB. Lipid perturbations may be the main driver also in sporadic PD cases, and due to additional genetic contributions (see above), and/or manifesting in an age-dependent manner, since decreased GBA activity and increased glycosphingolipid levels have been described in the brain with increasing age (; ). In all cases, the limited capacity of eMI to sequester increasingly abundant oligomers may cause their toxic cellular buildup over time, with oligomer-laden LE/MVB reflecting LB precursors. A LE/MVB-related mechanism for the biogenesis of LBs is consistent with recent ultrastructural studies revealing that LBs from sporadic PD patients are surrounded by a membrane, and comprised of intraluminal α-synuclein along with abundant intraluminal vesicular structures, membrane fragments and dysmorphic organelles (), a topology very reminiscent of a LE/MVB with its surrounding membrane and intraluminal vesicles.

FIGURE 3

Mechanisms Underlying LRRK2-Related PD

Independent studies in different model systems using different and complementary experimental approaches indicate that LRRK2 impairs macroautophagy (; ). Pathogenic LRRK2 overexpression results in an accumulation of autophagosomes in different cell types, including SH-SY5Y and HEK293 cells (; ). Furthermore, data from patient-derived fibroblasts indicate that mutant LRRK2 causes an increase in autophagosome numbers, with an impaired response to starvation-induced autophagy (; ). Similarly, iPSC-derived DA neurons from G2019S LRRK2-PD patients display an increase in autophagosome numbers due to a decrease in autophagosome clearance (). This is consistent with our study reporting that the increase in autophagosome numbers is paralleled by a decrease in the number of acidic lysosomes, which is expected to cause lysosomal degradative deficits (). Similarly, pathogenic LRRK2 has been consistently reported to cause impaired mitophagy in various cell types, likely due to limited lysosomal clearance of mitochondria (; ; ; ; ; Walter et al., 2019). Altogether, these data indicate that mutant LRRK2 causes a deficit in lysosomal functioning, thereby impairing the autophagic clearance of protein aggregates and dysfunctional organelles.

Indeed, multiple independent studies indicate that disease-associated LRRK2 variants alter lysosomal biology, even though not due to alterations in protein expression levels as assessed from G2019S knockin tissue, in contrast to tissue obtained from LRRK2-deficient mice (). Different lysosomal readouts have been analyzed, including altered lysosome localization, number, size or luminal pH, with only few studies directly assessing the potential effects of those alterations on lysosomal protein degradation capacity. Studies in a Drosophila model overexpressing G2019S LRRK2 indicate that pathogenic LRRK2 promotes the perinuclear clustering of lysosomes in a manner dependent on Rab7a and microtubules (). A decrease in lysosome number accompanied by an increase in lysosome size, similar to the phenotype of enhanced perinuclear lysosomal clustering observed in Drosophila, has also been reported in cultured astrocytes derived from transgenic mice overexpressing G2019S LRRK2 (). Importantly, these morphological alterations were shown to be accompanied by impaired lysosomal degradation (). Lysosomal structural alterations, namely swollen lysosomes, have also been described at the electron microscopy level in cultured neurons overexpressing G2019S LRRK2 (). In cellular systems expressing endogenous amounts of pathogenic LRRK2, lysosomal phenotypes have been reported as well. For example, increased lysosome size and perinuclear clustering is observed in fibroblasts from G2019S LRRK2 PD patients as compared to healthy controls (). Interestingly, both readouts are reversed upon inhibiting the activity of TPC2, an endolysosomal ion channel implicated in calcium signaling from acidic organelles (), consistent with our findings that pathogenic LRRK2-mediated autophagic deficits are reverted upon TPC2 inhibition (). Similarly, a recent study reported increased lysosomal pH, decreased lysosomal protein degradation capacity as well as altered lysosomal calcium dynamics in cortical neurons from BAC transgenic rats expressing R1441C LRRK2 (). Modest but significant lysosomal alterations have also been reported in cultured cortical neurons from G2019S knockin mice, with an increase in lysosome number and a decrease in lysosomal acidity, accompanied by increased insoluble α-synuclein and enhanced α-synuclein secretion (). Thus, a picture is emerging whereby pathogenic LRRK2 causes lysosomal structural alterations which are paralleled by impaired lysosomal protein degradation capacity, increased intralysosomal pH and altered lysosomal calcium signaling, all of which may contribute to PD pathogenesis.

The LRRK2-RAB Axis Underlying Endolysosomal Deficits

To dissect the mechanism(s) underlying the endolysosomal alterations caused by pathogenic LRRK2, we employed the well-established epidermal growth factor receptor (EGFR) trafficking and degradation assays. Under high concentrations of ligand, the receptor undergoes a purely degradative trafficking route, whilst low ligand concentrations favor receptor recycling, allowing us to probe for effects of pathogenic LRRK2 on both trafficking routes (Figure 4). Pathogenic LRRK2 expression was found to cause a deficit in EGFR degradation which was reversed upon addition of distinct LRRK2 kinase inhibitors or when overexpressing active Rab7a (). The deficit in EGFR trafficking and degradation correlated with the appearance of Rab7a-positive tubular structures, which could be observed in cells expressing pathogenic LRRK2 as well as in fibroblasts from PD patients carrying the G2019S LRRK2 mutation as compared to healthy controls. In both cases, the trafficking deficit and formation of Rab7a-positive tubules were associated with a significant decrease in the levels of active Rab7a as measured using biochemical pulldown assays ().

FIGURE 4

Whilst Rab7a is not required for the trafficking to LE/MVB nor the formation of intraluminal vesicles at the LE/MVB, it is required for transfer of cargo from LE/MVB to the lysosome (), and known to regulate endolysosomal positioning, retromer-mediated trafficking, lysosome reformation, and autophagosome maturation (; ; ; ; ; Yu et al., 2010). Thus, pathogenic LRRK2 may interfere with these various endolysosomal trafficking steps via decreasing Rab7a activity. Even though ubiquitously expressed, missense mutations in Rab7 are known to cause Charcot-Marie-Tooth type 2B disease, a peripheral neuropathy (; ; ; ). In addition, in a Drosophila model, loss of Rab7 in neurons causes adult-onset degeneration beginning with a loss of synaptic function (), warranting further in vivo studies to address the role of Rab7 in neurodegeneration and specifically in LRRK2-related PD pathogenesis.

The mechanism by which pathogenic LRRK2 causes a decrease in Rab7a activity remains unknown, and various scenarios are possible. LRRK2 phosphorylates a subset of Rab proteins on a conserved residue within the switch II domain, which is implicated in GDP/GTP exchange as well as in interactions with various regulatory and effector proteins (, ). A phosphomimetic Rab8a mutant displays impaired binding to GDP dissociation inhibitor 1/2 (GDI1/2), which is essential to target and extract the Rab protein from the membrane, and also to Rabin8, the guanine nucleotide exchange factor (GEF) required to activate the protein (; ). Whilst such binding differences are less pronounced with endogenously phosphorylated Rab8a (), these biochemical studies nevertheless are consistent with the idea that LRRK2-mediated phosphorylation of Rab8a causes its inactivation (). By analogy, LRRK2-mediated phosphorylation of Rab7a may contribute to its inactivation, even though current evidence suggests that Rab7a is not a LRRK2 kinase substrate either in vitro or in vivo (, ; ), in contrast to its reported phosphorylation by the related kinase LRRK1 (). Alternatively, LRRK2 may regulate Rab7a protein levels, even though the absence of detectable differences in protein levels from G2019S LRRK2 PD patients as compared to healthy controls makes this scenario unlikely (). A third scenario includes LRRK2-mediated alterations in the functioning of at least some of the Rab proteins which are LRRK2 kinase substrates, thereby causing downstream deficits in Rab7a activity.

The most prominent LRRK2 kinase substrates include Rab8a and Rab10 (, ). Rab8a is localized to the Golgi and a tubular early recycling compartment, and is known to regulate post-Golgi exocytic membrane trafficking, and endocytic recycling steps (; ; ; ). Since Rab8a has also been described to modulate endolysosomal trafficking events (), we probed for a possible link between alterations in Rab8a and the endolysosomal degradative trafficking steps of the EGFR which are impaired by G2019S LRRK2 (). Expression of active Rab8a, or upregulation of the Rab11-Rabin8 cascade known to activate Rab8a, rescued the G2019S LRRK2-mediated trafficking deficits. Conversely, knockdown of Rab8a impaired endolysosomal trafficking and degradation of the EGFR, which could be rescued upon expression of active Rab7a. In sum, either expression of G2019S LRRK2 or knockdown of Rab8a cause a decrease in the activity of endogenous Rab7a, accompanied by identical endolysosomal trafficking deficits (Figure 5; ; ).

FIGURE 5

In the future, it will be important to determine the link between other kinase substrates such as Rab10 and the LRRK2-mediated endolysosomal trafficking deficits. In addition, further studies are warranted to determine how downregulation of Rab8a, either by RNAi, or by LRRK2-mediated phosphorylation, causes a decrease in Rab7a activity. Crosstalk between these two Rab proteins may be mediated by a competition for shared GEFs, even though the currently identified GEFs for Rab8a and Rab7a are distinct (; ; ; Yoshimura et al., 2010; Yasuda et al., 2016; ). Alternatively, the two Rab proteins may compete for GTPase-activating proteins (GAPs), which display a certain degree of promiscuity and little reported correlation between their Rab binding and Rab-GAP activities (; ). TBC1D15 has been described as a GAP for Rab7a (Zhang X. M. et al., 2005; ). TBC1D15 expression causes the same endolysosomal trafficking deficits as pathogenic LRRK2, which are rescued by GTP-locked Rab7a expression and correlate with a decrease in endogenous Rab7a activity (). Interestingly, TBC1D15 has been described to preferentially interact with phospho-deficient as compared to phospho-mimetic Rab8a (). Thus, the LRRK2-mediated phosphorylation of Rab8a may cause a decreased interaction with TBC1D15, allowing it to act on Rab7a, similar to what has been recently described for the retromer-mediated regulation of Rab7a activity via TBC1D5 ().

Another attractive hypothesis relates to a possible regulation of Rab7a activity by distinct lipids. Interestingly, several studies indicate that Rab7a activity is regulated by cholesterol (, ; ; ). In many LSDs, the endolysosomal accumulation of glycolipids is associated with the secondary accumulation of cholesterol (), and overexpression of Rab7a has been reported to correct such accumulation (). In addition, increased endolysosomal cholesterol impairs the GDI-mediated extraction of Rab7a from the membrane, with possible consequences for the balance between active and inactive Rab7a (). Conversely, a role for Rab8a in cholesterol trafficking has been described as well. The endolysosomal cholesterol and sphingolipid deposition in LSD fibroblasts from Niemann-Pick type C disease is rescued by overexpression of Rab8a, whilst depletion of Rab8a from wildtype fibroblasts causes an accumulation of endolysosomal cholesterol (). Finally, expression of pathogenic G2019S LRRK2 or loss of Rab8a causes an accumulation of the EGFR in a Rab4-positive endocytic compartment (), and a cholesterol-mediated impairment of the Rab4-dependent recycling pathway has been previously described (). Thus, further studies are warranted to probe for a link between LRRK2, Rab8a and altered intracellular cholesterol levels, which may underlie the various reported trafficking deficits along the endosomal/endolysosomal system via direct inactivation of Rab7a (Figure 5).

Irrespective of the precise mechanism(s) underlying Rab7a inactivation mediated by pathogenic LRRK2, the consequences may involve a loss of lysosomal identity and function, further exacerbated in the context of lysosomal stress. Our EGFR trafficking studies were performed in the presence of serum starvation which favors high flux through the endocytic system, which is accompanied by the stimulation of lysosome reformation. This process is regulated by Rab7a, and involves the generation of LAMP1-positive tubules which pinch off from the lysosomal compartment (Yu et al., 2010). Impaired lysosome reformation, as evidenced by excess lysosomal tubulation, will lead to an eventual loss of lysosome identity and function, since the tubule-derived vesicles initially contain lysosomal membrane components but are devoid of acidic pH, or intraluminal proteases (Yu et al., 2010). Interestingly, upon induction of lysosomal tubule formation, these structures were reported not to be preserved upon fixation, with staining for lysosomal membrane markers instead revealing clustered lysosomes, measured as an increase in size and concomitant decrease in numbers as compared to control conditions (), identical to the structural lysosomal changes observed upon pathogenic LRRK2 expression. Thus, all reported alterations in lysosome morphology and function are consistent with a possible pathogenic LRRK2-mediated impairment of lysosome reformation in a manner mediated by Rab7a.

Additional roles for LRRK2 under conditions of lysosomal stress triggered by distinct lysosomotropic agents have been described. Whilst LRRK2 is largely not localized to lysosomes under normal conditions (), these agents are able to potently trigger its lysosomal recruitment. For example, prolonged treatment of cells with chloroquine, which becomes protonated and trapped in acidic compartments including LE/MVB and lysosomes, causes endolysosomal enlargement, translocation of endogenous wildtype LRRK2, accumulation of Rab8a and Rab10, and the promotion of lysosome secretion (). Whilst these data suggest that wildtype LRRK2 plays a role in lysosomal homeostasis, the effect of pathogenic LRRK2 variants on this readout remains to be determined. On the other hand, recent studies using LLOME, a membranolytic polymer causing lysosomal damage, was described to cause the endolysosomal recruitment of wildtype LRRK2, accumulation and phosphorylation of Rab10 and Rab35, and subsequent recruitment of JIP4 (). Interestingly, this was associated with the formation of JIP4-positive, LAMP1-negative tubules, and was more pronounced when expressing G2019S LRRK2 as compared to wildtype LRRK2 (). Thus, and under conditions of lysosomal stress, the pathogenic LRRK2 kinase-mediated processes may further compromise lysosomal identity and function, even though the pathophysiologically relevant lysosomal stressor(s) remain to be determined.

Concluding Remarks

An increasing amount of evidence indicates that lysosomal deficits are central to PD pathogenesis. In the case of PD due to mutations in LRRK2, structural and functional lysosomal alterations may arise due to impaired lysosome reformation in a manner dependent on Rab7a, with lysosomal identity and function further compromised in the context of additional lysosomal stress. Such LRRK2-mediated lysosomal dysfunction may lead to progressive cellular demise and manifestation of PD symptoms in the absence of LB formation (Figure 6). Lysosomal deficits mediated by pathogenic LRRK2 may not be sufficient to cause α-synuclein aggregation, but may manifest only in the context of additional lysosomal stress. Alternatively, and as proposed here, the mechanisms leading to LB formation in a subset of LRRK2-PD patients may be independent of the LRRK2-mediated lysosomal deficits, and involve α-synuclein oligomer formation triggered by altered lipid composition due to the existence of additional genetic LSD variants in those patients, or due to additional environmental and/or age-related changes (Figure 6). Conversely, patients with mutations in GBA display reduced GCase enzyme activity, but the majority never develop PD. Disease manifestation in GBA carriers may also require additional environmental, genetic or age-related triggers, resulting in lipid alterations, and concomitant α-synuclein aggregation (Figure 6).

FIGURE 6

The pathogenic LRRK2-mediated lysosomal alterations are evidenced by an increase in lysosomal pH, expected to decrease the activity of various lysosomal proteases. Interestingly, LRRK2 kinase inhibitors were recently shown to cause reacidification of lysosomal pH and a partial correction of cathepsin B activity in GBA-mutant astrocytes (). In addition, decreased GCase activity has been described in G2010S LRRK2-PD fibroblasts and iPSC-derived neurons, which was corrected upon LRRK2 kinase inhibitor application or Rab10 expression, and was mimicked upon knockdown of Rab10 (Ysselstein et al., 2019). However, LRRK2 kinase inhibitor treatment or Rab10 expression resulted in a similar increase in GCase activity also in healthy control cells and in cells heterozygous for GBA mutations (Ysselstein et al., 2019), suggesting a mechanism independent of mutation status or disease state, perhaps related to altered endolysosomal membrane trafficking events. Indeed, impaired Golgi-lysosomal transport of GBA enzyme causes a decrease in its activity (), such that endolysosomal trafficking deficits may explain the reported decrease in GCase activity in pathogenic LRRK2-expressing cells (Ysselstein et al., 2019; Figure 6). Whilst further work is required to dissect the precise crosstalk between LRRK2 and GBA, these data highlight the importance of the LRRK2-Rab axis in regulating lysosomal homeostasis, possibly relevant beyond LRRK2-related PD. In addition, they indicate that studies addressing the role of Rab function in cell culture and rodent models will be of paramount importance toward understanding PD-related cellular mechanisms, and that modulators of Rab activity may have potential as disease-modifying agents.

Statements

Author contributions

SH conceived and wrote the manuscript. PR-R generated all artwork. PR-R, MR-L, RF, and YN provided intellectual input for the contents and edited the manuscript.

Funding

This work was supported by the Michael J. Fox Foundation and by intramural support from Rutgers University.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling Editor declares a past co-authorship with one of the authors SH.

References

Summary

Keywords

LRRK2, Parkinson’s disease, Rab protein, lysosome, phosphorylation, α-synuclein, glucocerebrosidase, lysosomal storage disorder

Citation

Rivero-Ríos P, Romo-Lozano M, Fasiczka R, Naaldijk Y and Hilfiker S (2020) LRRK2-Related Parkinson’s Disease Due to Altered Endolysosomal Biology With Variable Lewy Body Pathology: A Hypothesis. Front. Neurosci. 14:556. doi: 10.3389/fnins.2020.00556

Received

30 March 2020

Accepted

06 May 2020

Published

04 June 2020

Volume

14 - 2020

Edited by

Evy Lobbestael, KU Leuven, Belgium

Reviewed by

Takeshi Iwatsubo, The University of Tokyo, Japan; Rina Bandopadhyay, University College London, United Kingdom; Dan Berwick, St George’s, University of London, United Kingdom

Updates

Copyright

*Correspondence: Sabine Hilfiker, ;

This article was submitted to Neurodegeneration, a section of the journal Frontiers in Neuroscience

Disclaimer

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Outline

Figures

Cite article

Copy to clipboard


Export citation file


Share article

Article metrics