Leucine-rich repeat kinase 2 (LRRK2) is a complex signalling protein that is a key therapeutic target, particularly in Parkinson's disease (PD). In addition, there is now evidence showing that LRRK2 expression and phosphorylation levels have potential as markers of disease or target engagement. Indeed, reports show increases in LRRK2 protein levels in the prefrontal cortex of PD patients relative to controls, suggesting that increase in total LRRK2 protein expression is correlated with disease progression. LRRK2 phosphorylation levels are reduced in experimental systems for most disease mutants, and LRRK2 is also rapidly dephosphorylated upon LRRK2 inhibitor treatment, considered potential therapeutics. Recently, the presence of LRRK2 was confirmed in exosomes from human biofluids, including urine and cerebrospinal fluid. Moreover, phosphorylation of LRRK2 at phosphosites S910, S935, S955 and S973, as well as at the autophosphoryation site S1292, was found in urinary exosomes. In this review, we summarize knowledge on detection of LRRK2 in human biofluids and the relevance of these findings for the development of PD-related biomarkers.
Parkinson's disease (PD) is one of the most frequent neurodegenerative disorders. The last 20 years have seen a surge in our knowledge of pathological mechanisms, driven by the identification of genes linked or associated with PD. At least 20 genes have been linked to familial forms of PD, while >20 genetic risk loci have been reported from PD genome-wide association studies (GWAS) . Some of the genes that have received the most attention are those encoding α-synuclein (SNCA), glucocerebrosidase (GBA), parkin (PARK2), Pten-induced kinase 1 (PINK1), microtubule-associated protein tau (MAPT) or leucine-rich repeat kinase 2 (LRRK2). Studies revealing biological functions of PD genes in health and disease have made it possible for the research community to begin to develop and test novel therapeutic strategies. For instance, the dominant genes SNCA (encoding α-synuclein) and LRRK2 have both been found to exert cellular toxicity when overexpressed and/or when gain-of-function mutants are expressed . While several questions still remain as to the precise cellular functions of α-synuclein or LRRK2, the study of these disease proteins has led to the development of therapeutic strategies targeting these proteins that are currently in the late phases of preclinical research [3–5] or in early-stage clinical trials.
However, as the field progresses on the road to neuroprotective or neuroregenerative therapies, another aspect of PD research requires special attention, i.e. the development of biomarkers for PD. Biomarkers (‘biological markers’) are defined by the National Institutes of Health Biomarkers Definitions Working Group as ‘a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention’ . Despite advances in our understanding of PD pathophysiology, its diagnosis and the evaluation of disease progression and treatment efficacy are still largely based on clinical criteria. In addition, there is currently no reliable method available to identify PD in subjects prior to the apparition of motor symptoms. This element is of particular importance in light of our knowledge that abnormal dopaminergic cell death begins at stages prior to the development of motor symptoms and reductions in life quality. Therefore, there is a pressing need for indicators to improve the accuracy of PD diagnosis (distinguishing PD from healthy subjects and other neurodegenerative conditions with Parkinsonism) and to evaluate the prognosis/disease progression and the response to symptomatic and disease-modifier therapeutics.
In the search for disease biomarkers, one approach is to evaluate disease-associated proteins or markers of their activity. For instance, the α-synuclein protein that is encoded by the SNCA gene is currently under evaluation as a disease biomarker. α-Synuclein is an aggregation-prone protein that is abundant in Lewy bodies and Lewy neurites that correspond to proteinaceous deposits present in, respectively, cell bodies and neurites of diseased brains. Interestingly, α-synuclein is also present in human cerebrospinal fluid (CSF), and the use of levels of α-synuclein in CSF as a PD biomarker is under evaluation [7,8]. Interestingly, evidence is now revealing that the LRRK2 protein may constitute a potential biomarker for several aspects related to PD.
Interest in LRRK2 has been stimulated by the genetic evidence demonstrating the importance of LRRK2 in PD. Indeed, mutations in LRRK2 are responsible for monogenic forms of PD and GWAS showed that genomic variations at the LRRK2 locus identified increased risk of sporadic PD . LRRK2 is a large and complex protein containing multiple functional domains , including GTPase and kinase functions [11,12]. While the precise biological role of LRRK2 is still a subject of intense research, studies put forward a gain of toxic function hypothesis for LRRK2 as a disease mechanism. Indeed, in experimental systems, overexpression of LRRK2 or expression of the G2019S — mutant form of LRRK2 that activates kinase function — generally leads to moderate toxicity . More importantly, LRRK2 appears to play a role in regulating toxicity of other toxic factors, such as α-synuclein, or challenges to the immune system [14–17].
Expression and detection of LRRK2
LRRK2 is expressed in multiple tissues , including brain, lung, kidney, gut, muscle, heart, skin and in biofluids such as CSF, urine  and blood, with LRRK2 found in peripheral blood mononuclear cells (PBMCs), including lymphocytes and monocytes [20,21]. In the central nervous system, LRRK2 is expressed in several brain structures important for PD, such as striatum, hippocampus, cortex and cerebellum and to a lower extent in the substantia nigra [22–24]. At the cellular level, LRRK2 is expressed in neurones, including layer V motor cortex neurones and medium spiny striata neurones. Expression in glial cells appears to be absent or low, although LRRK2 expression is induced in CD68-, Cd11b- and isolectinB4-positive cells in the brain upon challenge with lipopolysaccharide . A cell preferential expression of LRRK2 is also found in the kidney, where LRRK2 is expressed primarily in collecting duct cells . The expression of LRRK2 in the brain and kidney may also be related to the presence of LRRK2 in CSF and urine. Indeed, LRRK2 is found in exosome isolates, i.e. cell-derived vesicles of 30–100 nm in diameter, from both of these fluids (ref.  and our own unpublished results), indicating that LRRK2-expressing cells in the kidney or brain may release LRRK2-positive exosomes in urine and CSF, respectively. Interestingly, the presence of LRRK2 in urinary exosomes was first reported in a mass spectrometry-based screen of urinary exosome proteins , but similar studies in CSF have failed to pick up LRRK2 protein .
The detection of LRRK2 protein in tissue and biofluids has been facilitated by the development of rabbit and mouse monoclonal antibodies with specificity and sensitivity proven in immunohistochemistry and western blotting applications . In addition to proteins detecting unmodified epitopes on LRRK2, several antibodies are also capable of phosphosite detection. Indeed, a key property of LRRK2 is that it is a highly phosphorylated protein, displaying at least 15 phosphorylation sites that can be categorized as autophosphorylation sites and sites phosphorylated by other kinases that we will call heterologous phosphorylation sites (reviewed in ref. ). Interestingly, both categories of sites occur in clusters, i.e. a cluster of autophosphorylation sites at or proximal to the ROC GTPase domain, including the T1410, T1491 or T1503 sites, and a cluster of heterologous sites in the interdomain region between the ankyrin repeat and leucine-rich repeat domains. Some notable exceptions to these clusters are the S1292 and T2483 autophosphorylation sites (BST-2016-0334F1,Figure 2).
Tissue expression distribution of LRRK2.
LRRK2 domain structure and location of phosphorylation sites.
To analyze exosomes, biofluids are typically subjected to a centrifugation-based fractionation method. Biofluids, such as CSF or urine, are first subjected to a low speed spin (500 × g) to eliminate all cellular and large debris. This step is followed by a medium speed spin (10–20 k × g) that will sediment large membrane fragments and large vesicles such as ectosomes. Finally, the resulting supernatant is subjected to a high speed ultracentrifugation (100–200 k × g spin) capable of sedimenting the 30–100 nm large exosomes . Through western blot analysis of urinary and CSF exosomes, it has been determined that LRRK2 is present in these biofluids at high pg/ml range in urine and low pg/ml range in CSF (ref.  and our own unpublished observations). Anti-LRRK2 antibodies have also been used in immunofluorescence to confirm the presence of LRRK2 in urinary exosomes co-labeled with the exosome marker TSG101 . The low level of LRRK2 observed in CSF is in line with the lack of detection of LRRK2 in CSF through mass spectrometry screening and underlines the need to develop more sensitive techniques of LRRK2 detection.
Potential for LRRK2 as a biomarker
Several pieces of evidence point to the interest of measuring LRRK2 or LRRK2 phosphorylation as a disease marker or a marker of target engagement. First, LRRK2 is expressed in tissues and biofluids that are not only relevant to disease, but are also testable in a clinical diagnostic test . For instance, studies performed in postmortem brain tissue show increases in LRRK2 protein levels in the prefrontal cortex of PD patients relative to controls [32,33], suggesting that increase in total LRRK2 protein expression is correlated with disease in this structure and, therefore, that LRRK2 levels may constitute a disease biomarker. While it is not feasible to test brain tissue during disease progression in a biochemical test, LRRK2 found in exosome isolates from CSF may be an indicator of fluctuations of LRRK2 expression in the brain.
Besides total LRRK2 levels, additional evidence suggests that phosphorylation levels at several LRRK2 phosphosites also constitute promising biomarkers. This is supported by findings that LRRK2 phosphorylation is modulated in a majority of disease variants. For the S910–S935–S955–S973 phosphosites, levels are reduced for most mutants , whereas for phospho-S1292, levels are increased for most mutants  (reviewed in ref. ). Also, all of these five sites are rapidly dephosphorylated in cell culture, mammalian animal models and in human PBMCs upon LRRK2 inhibitor treatment [3,37], considered potential therapeutics (see Figure 2B for an alignment of sequences at the level of LRRK2 phosphorylation sites across multiple species). Therefore, both LRRK2 levels and LRRK2 phosphorylation levels are promising markers for both disease and pharmacodynamic responses.
A corollary of considering LRRK2 as a biomarker is that LRRK2's substrates, such as the ezrin–radixin–moesin family of proteins , microtubule affinity-regulating kinase 1 (MARK1) , endophilin A  or Rab proteins , are also potential biomarkers. In particular, Rab proteins have demonstrated specific potential as biomarkers of LRRK2 kinase activity. Indeed, phosphorylation of Rab10 is reduced in LRRK2 knockout cells and increased in cells expressing the LRRK2 clinical and kinase-activating mutant G2019S . Rab10 phosphorylation is also reduced after treatment of cells with LRRK2 kinase inhibitors [41,42]. In these first studies exploring the dependency of Rab10 phosphorylation in LRRK2 activity, the loss of Rab10 phosphorylation appeared to be more sensitive than the loss of LRRK2 pS935 phosphorylation . Further research is now required to characterize LRRK2 substrates such as Rab10 and others as potential biomarkers of PD. The hypothesis that phospho-LRRK2 is a disease biomarker has begun to be evaluated in different studies. For instance, LRRK2 S910 and S935 phosphorylation has been tested in PBMCs of idiopathic PD patients compared with controls, even though no significant changes have been observed . Thus far, no assessment has been made of autophosphorylation sites, such as S1292, in PBMCs. In urine, testing of S1292 phosphorylation levels has yielded significant findings. Notably, elevated S1292 phosphorylation is observed in subjects carrying the G2019S mutation , a prevalent LRRK2 variant that was found to display elevated S1292 phosphorylation in cell culture and transgenic mice. In addition, S1292 is elevated in G2019S carriers with PD symptoms compared with those with none. In another study by the same group, S1292 phosphorylation is found to be significantly increased in idiopathic PD compared with matched controls . Interestingly, in this last study, elevated S1292 phosphorylation correlated with severity of cognitive impairment. It should be noted that while significant differences are observed in these studies, the S1292 phosphorylation levels in urinary exosomes are not absolute predictors of PD patients or of G2019S carriers as partial overlap is seen in the distribution of these measures in control and mutant/disease groups. Nevertheless, the findings support the hypothesis that LRRK2 is altered in disease, warranting further work evaluating LRRK2 as a biomarker.
At present, no studies have assessed whether LRRK2 or phospho-LRRK2 is altered in CSF of genetic or sporadic forms of PD. Also, thus far, studies have been performed using western blot-based measures that demonstrate sufficient sensitivity; however, this method is not precise enough and is too time consuming for use in routine clinical sample testing. Additional more quantitative methods of detecting the relatively low LRRK2 and phospho-LRRK2 levels in PBMCs, urine or CSF will be necessary to further characterize these initial findings.
Another important potential of LRRK2 is as a biomarker of target engagement for LRRK2 kinase inhibitors that are currently in development as potential therapeutics in PD. Indeed, in experimental systems such as cell lines and model organisms, treatment with LRRK2 kinase inhibitors leads to a rapid reduction in S935 and S1292 phosphorylation levels. Therefore, measure of LRRK2 phosphorylation in urine and CSF has the potential of confirming engagement of the LRRK2 target by LRRK2 inhibitors peripherally and centrally, respectively. While inhibitor-induced dephosphorylation of LRRK2 is confirmed in rodent tissues, this remains to be confirmed in biofluids.
In conclusion, in addition to being a confirmed therapeutic target for PD, emerging evidence is pointing to LRRK2 as a PD biomarker. This development makes for a welcome addition in the list of potential PD biomarkers under evaluation. However, significant steps remain to be taken to evaluate the true value of LRRK2 as a biomarker. Developing quantitative assays for total and phospho-LRRK2 is a necessary step to test larger groups of PD samples. Besides biochemical assays, another potentially valuable tool would be the development of a tracer molecule for use in non-invasive imaging. The recent emergence of highly potent, selective and brain penetrant LRRK2 inhibitors suggests that tracer development is possible. Also, the evaluation of LRRK2 as a biomarker should be placed in the larger effort to develop PD biomarkers. To date, a single biomarker with a strong specific predictive value is lacking. Rather, a more plausible scenario is that individual biomarkers each with some predictive value and/or specificity for PD may be combined in a test panel. In that context, LRRK2 may be integrated in a panel of biomarkers that together may be expected to reach significantly high levels as specific predictors of PD.
The Authors declare that there are no competing interests associated with the manuscript.