Parkinson’s disease (PD) is the second most common neuropathology after Alzheimer’s disease [1]. It has the fastest-growing prevalence and most deaths among neurological disorders, affecting nearly 6.1 million ageing people in 2016 globally, which is more than doubled from the past generation [2]. PD is a multisystem disease that affects the nervous system, especially the central nervous system (CNS), with the pathological characteristic of losing more than 50% of the dopaminergic neurons in the substantia nigra pars compacta ^[3][4][3,4]. Dopaminergic neurons are the main source of producing rich dopamine (DA). The depletion of DA manifests in the progressive degeneration of motor mobility with the clinical symptoms of bradykinesia, rigidity, rest tremor, and postural instability ^[1][5][1,5]. In addition, the neuropathology of PD also exhibits heterogeneous clinical symptoms in non-motor function (deficit in olfactory, depression and cognition, rapid eye movement sleep behaviour disorder, constipation, and central autonomic control) at all stages of the PD pathological process, even predating the motor symptoms [6].

PD is considered as multifactorial neuropathology with complex aetiology, which was found to be highly correlated with ageing, abnormal alpha-synuclein (α-Syn) accumulation in dopaminergic neurons, exposure to polluted environment, and genetic susceptibility ^[1][3][7][1,3,7]. However, the risk of PD pathogenesis was found to interplay with multiple different factors and eventually give rise to different clinical symptoms. More often, the disease mechanisms may not be completely identical in all PD patients. Hence, the complexity of PD disease increases the difficulty to reveal a clearer pathogenesis pathway.

To date, there are only approximately 10–15% of patients reported as family history of PD symptoms, while the remaining 85–90% of the PD population are classified as sporadic PD ^[1][8][1,8]. More than 20 genes were found to be involved in familial PD, in which the penetrant mutation is rare, with an occurrence of 2% in the PD population ^[7][8][9][7,8,9]. These mutations present a broad range of risks and demonstrate heterogeneous mechanisms associated with PD. Among the PD causative genes, the 51-exon Leucine-rich repeat kinase (LRRK2) gene is the most common mutation, where the variants are reported in 0.7% of all the people showing PD symptoms [8]. Hence, a missense mutation in LRRK2 G2019S received the most attention because it appears to be the most penetrant mutation in the autosomal dominant PD and sporadic PD, with an influence of 40% and 10%, respectively ^[4][5][10][4,5,10]. Moreover, the occurrence of LRRK2 variants affecting a wide span of diverse ethnicities in different geographic distributions further define its role of genetic susceptibility in the PD population [11].

There are nearly 100 mutations identified at the different domains of LRRK2, in which some are related to the pathogenesis of PD, including G2019S, R1441C/G/H, I2020T, and Y1699C. The mutation sites are highly conserved and segregate in families with PD disease ^[5][8][10][5,8,10]. These mutations lead to gain-of-function mechanisms, where kinase activity of LRRK2 will increase to phosphorylate a group of Rab proteins in different subcellular localities, thus affecting downstream pathways to drive PD pathogenesis [14]. Interestingly, in in vitro and in vivo studies, knocking out LRRK2 recapitulated aspects of PD pathogenesis such as dopaminergic neuron loss, accumulation of α-Syn, impairment of protein degradation, and dysregulation of autophagy ^{[15][16][17][18]}[15,16,17,18]. However, there is a lack of evidence of phenotypic impact, as there was no increase in the risk of PD in humans carrying loss of function variants of LRRK2 [19].

Many studies also show that the aberrant function of LRRK2 to PD is not necessary by mutation alone, instead, some upstream factors such as Rab29 and α-Syn could exacerbate the activity of LRRK2 ^[14][20][21][14,20,21]. Therefore, evidence suggests that the mutant LRRK2 is inter-related with other proteins and impacts diverse cellular biological processes, increasing the complexity and enigma of the understanding of PD pathogenesis.

2. Main Mechanisms of LRRK2-Associated PD

The main finding in pre-clinical studies of LRRK2-PD is that elevated LRRK2 kinase activity or phosphorylation of targets is linked to various aspects of pathogenesis. Firstly, LRRK2 causes α-Syn neurotoxicity by decreased clearance of α-Syn and increased propagation and aggregate formation of α-Syn in a kinase-dependent manner. In the transgenic models with G2019S, α-Syn neurotoxicity induction is reversible with LRRK2 kinase inhibition, but not reversible in non-transgenic models ^[22][28]. Intriguingly, in models of synucleinopathy, LRRK2 inhibition can restore neuronal loss and motor deficits but worsens α-Syn clearance ^[23][83]. The similarity between these two studies is that LRRK2 inhibition is ineffective in preventing α-Syn neurotoxicity in the absence of pathogenic LRRK2. This suggests that though LRRK2 hyperactivation is present in idiopathic PD, it is not a main contributor to α-Syn neurotoxicity, but rather other activities lead of LRRK2 lead to neurodegeneration. The differences in these findings could be attributed to different mouse models. The former utilized an α-Syn PFF PD mouse model, whereas the latter utilized human α-Syn overexpressing transgenic mice. While the studies in this frevieldw have attributed neurotoxicity to α-Syn inclusion, some studies suggest α-Syn aggregation is a protective mechanism ^[24][97]. Rather than being a causative factor, Lewy body pathology could be an accompaniment of neuronal death. Therefore, the reversal of neurotoxicity with LRRK2 kinase inhibition could be associated with other exacerbated activities of LRRK2. LRRK2 kinase activity modulates calcineurin-independent and calcineurin-dependent pathways ^[23][25][33,83]. LRRK2 impacts both inflammatory pathways, as it can phosphorylate RCAN1, which is an inhibitor of calcineurin and activates NFATc2, a downstream substrate of calcineurin ^[26][104]. Other inflammatory pathways influenced by kinase activities of LRRK2 include upregulation of AGE-RAGE, which activates the NF-κB signalling pathway ^[27][34]. The NF-κB signalling pathway is indirectly upregulated in microglia by LRRK2, as LRRK2 decreases NF-κB inhibitory signalling by downregulating PKA activity ^[28][38]. In addition to contributing to inflammatory pathways, glia to neuron crosstalk induces neurotoxicity through other ways. In both astrocytes and microglia, both WT and pathogenic LRRK2 had a decreased ability to clear α-Syn ^[29][30][24,31]. This accumulation of α-Syn in astrocytes can be linked to how LRRK2-G2019S compromises CMA and macroautophagy ^[31][36]. LRRK2 G2019S astrocytes had a decreased ability in responding to oxidative stress through distinct means, SERCA inactivation and the SHP (receptor)-PIAS1 (protein)-XBP1 (transcription factor) pathway ^[32][33][46,48]. In LRRK2 KO microglia, there was a decrease in antioxidant response after treatment with pre-formed fibrils of αSYN ^[34][39]. These three results suggest that diversion of typical activities of LRRK2, i.e., the hyperactivation of kinase, will affect LRRK2 redox signalling. LRRK2 also facilitates microglial activation in a kinase-dependent manner ^[35][85]. A newer study found that LRRK2 G2019S impacts microglia response to inflammation by Rab8a function in iron uptake and transport ^[36][105]. Mitochondrial dysfunction and ER stress are the main mechanisms in PD. LRRK2 causes it through two main ways: calcium dyshomeostasis and defects in the mitochondria life cycle. Calcium dyshomeostasis in the mitochondria and ER was exacerbated by LRRK2 mutants G2019S, Y1699C, and R1441C and by the knockout of LRRK2 ^[37][38][39][43,44,45]. This shows that elevated LRRK2 activity will not exclusively impact calcium homeostasis, as a loss of physiological LRRK2 activity impacts calcium homeostasis as well. Mutations of LRRK2 such as G2019S and R1441G cause defects in the mitochondrial life cycle by the inhibition of mitophagy ^[40][41][42][49,50,51]. In two of the three studies, one involving G2019S and another with R1441G, kinase inhibition was unable to reverse the defects in mitophagy, suggesting that factors aside from kinase hyperactivity reduce mitophagy ^[40][42][49,51]. The R1441G study involved LRRK2 mutant mice, whereas the G2019S studies used patient fibroblasts. The difference in response to inhibition of kinase activity despite the utility of similar models could be because of differences in monitoring of mitophagy in human primary fibroblasts. The former study focused on measuring Miro1 intensity as opposed to mitophagy induction ^[40][41][49,50]. In general, the life cycle of the mitochondria is affected by LRRK2 because expression of the enzymatic core of LRRK2 alone (kinase and COR domain) leads to reduced mitochondrial biogenesis ^[43][53]. Many downstream substrates of LRRK2 are involved in vesicle trafficking. A delicate balance of activities of LRRK2 is needed in physiological conditions for vesicular transport, particularly in microtubule-mediated vesicular transport events ^[44][54]. The increase in phosphorylation of downstream factors of LRRK2 such as SJN1, auxilin, and Synapsin 1 is linked to dysfunction in synaptic vesicle trafficking ^[45][46][47][56,57,60]. Defects in autophagy and lysosomal processes are caused by LRRK2 kinase hyperactivity, disassociation of normal binding partners of LRRK2, altered recruitment of AV proteins, and phosphorylation of Rab10. Downstream Rab GTPases of LRRK2, including Rab10, play a pertinent role in LRRK2-PD pathology. From the studies in this frevieldw, Rab5, Rab7, Rab10, Rab29, and Rab35 were linked to various aspects of PD pathogenesis in a kinase-dependent manner. Rab5 and Rab35 were linked to α-Syn neurotoxicity by the downregulation of endosomal formation ^[29][48][49][24,26,82]. The negative regulation of CME was due to Rab5, Rab7, and Rab10 ^[50][58]. Lysosomal stress and impaired autolysosome formation was linked to Rab10 and Rab29, respectively ^[51][52][64,66]. In addition, the interference of ciliogenesis was mediated by binding of RILPL1 to pRab10 ^[53][54][80,81]. A separate study found that pRab8a and pRab10 were associated with RILPL1 and pRab8a and pRab12 was linked to RILPL2-related cilia defects ^[55][106]. Other PD-related phenotypes linked to Rab GTPases include impaired mitophagy. The phosphorylation of Rab10 prevented a necessary accumulation of Rab10 at the mitochondria that was necessary to induce mitophagy ^[56][107]. Of the five Rab GTPases implicated in LRRK2-PD, Rab10 has a role in several mechanisms and was found to be relevant in other forms of familial PD, including PINK1 and GBA1 ^[56][57][107,108]. This suggests a convergence of pathways with Rab10 as a substrate. Furthermore, pRab10 elevation is detectable in clinical samples of both LRRK2-PD and iPD patients ^[58][109]. Studies have been conducted to investigate the mechanisms that influence the phosphorylation of Rab10 by LRRK2. PPM1H, a phosphatase, was found to be a negative regulator of Rab10 phosphorylation ^[59][110]. The differential distribution of PPM1H in tissues could be an explanation for the variation of pRab10 in clinical samples ^[58][109]. Lastly, a downstream target of LRRK2 is tau. Previous studies have shown the in-vitro hyperphosphorylation of tau by LRRK2 ^[60][61][62][111,112,113]. Expression of LRRK2 G2019S in Drosophila promoted the phosphorylation of tau by the recruitment of glycogen synthase kinase 3β (GSK3β) ^[63][114]. The accumulation of insoluble and phosphorylated tau was observed in a transgenic LRRK2 mouse model of tauopathy ^[62][113]. Studies in this frevieldw reveal contrasting findings to previous knowledge of LRRK2 and tau. Two studies, despite different cell types, show that tau pathology observed in LRRK2-PD is kinase-independent ^[64][65][77,79]. Rather than the LRRK2 kinase activity contributing to tau neurotoxicity, it is attributed to proteosomal impairment and actin polymerization ^[64][66][77,78]. Two separate in vivo studies found that the oligomerization of LRRK2 promoted by its mutant forms resulted in overstabilization of F-actin, which has consequences for the accumulation and spread of proteins such as tau and α-Syn ^[66][67][78,84]. Overall, despite a lack of consensus that the kinase activities of LRRK2 directly impact tau pathology, there is evidence that LRRK2 promotes the spread of tau pathology.

3. LRRK2 Role in iPD and Familial PD

Studies reported in this frevieldw show that in the presence or absence of pathogenic variants, LRRK2 can contribute to the pathogenesis of PD. Comparisons were drawn from SH-SY5Y cells transfected with LRRK2 mutants, fibroblasts of LRRK2-PD patients, and post mortem brains of sporadic PD patients. Elevated protein levels associated with increased susceptibility to mitochondrial calcium dysregulation were observed for all groups, suggesting a shared mechanism between LRRK2-PD and sporadic PD ^[37][43]. Furthermore, kinase inhibition was able to reduce translational impairment in fibroblasts derived from both LRRK2 familial and sporadic PD patients ^[68][75]. It was also demonstrated that in patients with idiopathic PD, there was enhanced kinase activity of WT LRRK2 which was associated with abnormalities in mitochondria and lysosomal function ^[69][115]. Therefore, targeting LRRK2 activities may have broad therapeutic utility in idiopathic PD, not only in those who carry a LRRK2 mutation.

4. Potential Biomarkers in LRRK2-PD

The most significant biomarker in the detection of PD would be the increased kinase activity of LRRK2. LRRK2 kinase activity can be quantified by its autophosphorylation at serine1292 or the levels of downstream pRab10. In a Norwegian cohort, elevated levels of s1292 were identified in male patients that were carriers of G2019S ^[70][116]. These elevated pS1292 levels were found in brain and urinary exosomes. Future studies are needed to investigate pS1292 levels in larger, more diverse cohorts and to explore other LRRK2-PD mutations. In this frevieldw, some studies highlight the possibility of novel biomarkers in LRRK2-associated PD. In vitro and in vivo studies have identified the overexpression of Rab35 and linked it to α-Syn pathology ^[48][49][26,82]. In addition to that, elevated mtDNA damage was found in iPD and LRRK2-PD patient fibroblasts ^[71][41]. Impairment of translation was also observed in fibroblasts of iPD and LRRK2-G2019S patients ^[68][75]. Further studies with human cohorts are needed to assess if these biomarkers are detectable in patient serum.

5. Therapeutic Strategies for LRRK2-Associated PD

5.1. Direct Inhibition of LRRK2

Direct inhibition of LRRK2 kinase activity has a potential to impact other forms of familial PD. In this frevieldw, LRRK2 G2019S and GBA1-N370S-derived patient astrocytes shared hallmarks of PD ^[72][37]. Pre-clinical studies have shown that LRRK2 kinase inhibition can restore defects in GBA1 D490V mutant astrocytes ^[73][118]. The same group found that while defects caused by loss of GBA1 function can be attenuated by LRRK2 inhibition, the activity of the enzyme coded by GBA1, β-Glucocerebrosidase (GCase), is unaffected ^[74][119]. However, in that study, instead of mutant GBA1, GBA1-heterozygous-null iPSC-derived neurons were used, whereas in DA neurons derived from LRRK2-PD patients with G2019S or R1441C mutations, a reduction in GCase activity was observed in a manner mediated by Rab10 ^[75][120]. In contrast to the previous study, GCase activity in DA neurons with LRRK2 or GBA1 mutations can be increased with LRRK2 kinase inhibition. A study in 2022 corroborated these findings to some extent. In G2019S iPSC-derived neurons, GCase protein level was reduced, but in patient-derived fibroblasts and peripheral blood mononuclear cells, a positive correlation between LRRK2 kinase and GCase activity was found ^[57][108]. These studies indicate that LRRK2 kinase plays a role in regulating GCase activity, though there are differences specific to cell type. As these studies collectively suggest an interplay between these two PD-related genes, a case for GCase activation in LRRK2-PD can be also made. While most of the studies in this frevieldw focus on the role of LRRK2 G2019S in PD, other variants of LRRK2 that are not directly linked to the kinase function of LRRK2 play a role in the pathogenesis of PD. Like G2019S, R1441G in the GTPase domain can cause mitochondrial defects through Drp1 ^[42][51]. Protein synthesis deficiency, vesicle trafficking dysfunction, and autolysosomal dysfunction are linked to R1441C/G mutations. Hence, inhibitors of the GTP binding activities of LRRK2 have been proposed as a different approach. However, at present only pre-clinical work justifies its therapeutic potential. In SH-SY5Y cells expressing LRRK2 R1441C, a GTPase domain mutation, impairments were seen in the neurite transport of mitochondria and lysosomes. Treatment with GTP-binding inhibitors were able to prevent these defects ^[76][121]. GTP-binding inhibitors could also reduce kinase activity. LRRK2 GTP-binding inhibitors 68 and Fx2149 promoted LRRK2 ubiquitination, increased ubiquitinated aggregation, and contributed to an aggresomal response. This could be linked to improved clearance of protein aggregates ^[77][122]. An alternative to reducing LRRK2 activity would be to reduce total levels of LRRK2 protein. Pre-clinical studies have shown that LRRK2 antisense oligonucleotides (ASO) can ameliorate α-Syn pathology, reduce elevated levels of ER Ca²⁺ and defects in mitophagy, and reduce locomotor deficits ^[78][79][80][123,124,125]. Furthermore, treatment of LRRK2 ASO in transgenic mice expressing LRRK2 G2019S or WT LRRK2 was able to decrease phosphorylation of Rab10 and correct autophagic defects caused by LRRK2 ^[80][125]. Currently, there is an ongoing phase I clinical trial for BIIB09, a LRRK2 ASO ^[81][126]. Studies exploring the physiological role of LRRK2 have described a role in immune response—T-cell function as well as the innate immune response ^[82][83][84][127,128,129]. LRRK2 kinase activity as a negative regulator of macroautophagy has been previously described ^[85][130]. Therefore, inhibition of LRRK2 activities or total reduction in protein could dysregulate immune or vesicle trafficking functions. Therefore, apart from direct inhibition of LRRK2 activity and protein level, potential pharmacological interventions involving inhibition or upregulation of LRRK2 targets could be considered.

5.2. Indirect Inhibition of LRRK2

Inhibition of signalling pathways linked to neurodegeneration could be considered as a potential therapeutic intervention. There are signalling pathways activated or upregulated by LRRK2 linked to neurodegeneration. ASK1, an upstream regulator of P38 and JNK, when directly phosphorylated by LRRK2, results in neuronal apoptosis ^[86][68]. Therefore, ASK1 inhibitors which are currently undergoing clinical trials, some in phase III, could be used in the context of PD or broadly in neurodegenerative diseases ^[87][131]. Another inhibitor that is proposed are JNK inhibitors. These inhibitors, mainly in phase I trials, are being developed as anti-tumorigenic agents ^[88][132]. JNK inhibitors may be used in combination with other PD drugs. A study found that pathogenic LRRK2 results in JNK activation which led to motor impairment and neuronal death in a Drosophila model ^[89][69]. JNK inhibition alone only partially attenuated neurotoxicity ^[89][69]. Hence, combination therapy of LRRK2 inhibitors and JNK inhibitors would be more effective. Another signalling pathway that can be targeted is the AGE-RAGE intracellular signalling. G2019 enhances the interaction between AGE and RAGE, leading to inflammation and oxidative stress ^[27][34]. In a pre-clinical rat study, an anti-RAGE antibody was found to be able to block systemic inflammatory responses ^[90][133]. However, as it was unable to cross the blood–brain barrier, it may have limited applicability in the context of PD. Upstream to LRRK2, there are genes causing its aberrant expression. Some in vitro studies have shown that the knockdown of these genes can attenuate neurotoxicity in LRRK2-associated PD ^[91][92][88,89]. Two long non-coding RNA, HOTAIR and MALAT1, increase the expression of LRRK2, thereby increasing its activity as well. Since gene knockdown is not feasible in humans, alternatives need to be explored. Both MALAT1 and HOTAIR have been proposed as therapeutic targets in cancer. Anti-lncRNAs were able to suppress HOTAIR activity in targeting solid tumours in vivo ^[93][134]. To suppress MALAT1 activity, siRNAs were able to elicit a degradation of MALAT1 ^[94][135]. A similar approach could be used to inhibit these lncRNA in PD models to assess its applicability. TXNIP is synergistic with LRRK2 and causes ER stress by preventing the release of antioxidants and is linked to α-Syn pathology. In human 3D midbrain organoids, inhibition of TXNIP was found to reduce LRRK2-induced phenotypes ^[95][102]. Another strategy is targeting downstream mechanisms such as mitochondrial dysfunctions. Firstly, pre-clinical work has shown that the reduction in Miro levels can attenuate neurotoxicity, as that allows damaged mitochondria to be degraded ^[96][136]. Secondly, the suppression of PERK can prevent degradation of ER–mitochondria contacts, which is necessary for mitochondrial health. In a mouse model for frontotemporal dementia, PERK inhibition was able to prevent further neuronal loss and lower levels of phosphorylated tau ^[97][137]. Thirdly, SERCA is a potential target as well. LRRK2 deactivates SERCA by direct association, which leads to ER stress and mitochondrial dysfunction ^[32][46]. As kinase inhibition of LRRK2 does not activate SERCA, another approach must be conducted, such as the pharmacological activation of SERCA through an allosteric activator. Current pre-clinical work utilizes SERCA activators for diseases such as type-2 diabetes and Duchenne Muscular Dystrophy ^[98][99][138,139]. Lastly, Drp1 inhibition could be a strategy as well to prevent neuronal death linked to mitochondrial dysfunction. A pre-clinical study found that Drp1 inhibition protected against neurotoxicity ^[100][140]. Antisense oligonucleotides (ASOs) are a potential strategy to reduce the total protein levels of LRRK2 and other interacting proteins. ASOs for α-Syn are being studied in pre-clinical models and have been found to prevent neurodegeneration associated with LRRK2 ^[101][141]. Another protein target implicated in LRRK2-PD, tau, which is abundant in carries of LRRK2 mutations, could also be addressed with a similar approach ^[102][23]. The use of tau ASOs could be necessary to address tau pathology in LRRK2-PD, as some studies have shown that it is LRRK2 kinase-independent ^[64][65][77,79]. While many of these targets are synergistic with LRRK2, there are some that are antagonistic targets. In upregulating the activity or levels of these targets, it may reverse or reduce the damage caused by overactivated LRRK2. There were two targets found to be upstream of LRRK2: Prx2 and Fbxl18. Prx2 was found to be an upstream inhibitor of LRRK2, as it could reduce pRab10 levels, a downstream effector of LRRK2 ^[103][101]. It is effectively an inhibitor of the kinase activities of LRRK2. Fbxl18 is able to target phosphorylated LRRK2 for degradation, which may be beneficial as self-phosphorylation of LRRK2 is also reflective of its kinase activities ^[104][70]. There were also factors that were not upstream of LRRK2 but could mediate the downstream effects of LRRK2. In pre-clinical studies, the upregulation of NCLX was able to help with Ca2⁺ influx caused by LRRK2 ^[38][44]. SP1 is a factor associated with α-Syn that has opposite effects on neuronal health compared to mutant LRRK2 ^[105][103]. Upregulating the activity of SP1 could be considered as a strategy in pre-clinical work. Lastly, the restoration of auxilin or DNAJC6 was able to prevent synaptic vesicle endocytosis dysfunction caused by LRRK2 ^[46][57]. Clathrin-mediated endocytosis (CME) is a key process mediated by auxilin; this could explain the functional antagonism between LRRK2 and auxilin ^[106][142].