LRRK2 in Parkinson’s disease: in vivo models and approaches for understanding pathogenic roles


Z. Yue, Department of Neurology and Neuroscience, Mount Sinai School of Medicine, New York, NY 10029, USA
Fax: +1 212 241 3869
Tel: +1 212 241 3155


The recent discovery of the genetic causes for Parkinson’s disease (PD) is fruitful; however, the continuing revelation of PD-related genes is rapidly outpacing the functional characterization of the gene products. Although the discovery of multiple PD-related genes places PD as one of the most complex multigenetic diseases of the brain, it will undoubtedly facilitate the unfolding of a central pathogenic pathway and an understanding of the etiology of PD. Recent findings of pathogenic mutations in leucine-rich repeat kinase 2 (LRRK2) (PARK8) that are linked to the most common familial forms and some sporadic forms of PD provide a unique opportunity to gain insight into the pathogenesis of PD. Despite rapid growth in biochemical, structural and in vitro cell culture studies of LRRK2, the in vivo characterizations of LRRK2 function generally fall short and are largely limited to invertebrates. The investigation of LRRK2 or homologs of LRRK2 in nonmammalian models provides important clues with respect to the cellular functions of LRRK2, but an elucidation of the physiology and pathophysiology of LRRK2 relevant to PD would still depend on mammalian models established by multiple genetic approaches, followed by rigorous examination of the models for pathological process. This minireview summarizes previous studies of genes for ROCO and LRRK2 homologs in slime mold, nematode worms and fruit flies. It also discusses the results obtained from available mouse models of LRRK2 that begin to provide information for understanding LRRK2-mediated pathogenesis in PD.


bacterial artificial chromosome


C-terminal of ROC






leucine-rich repeat kinase 2


Parkinson’s disease


Ras of complex proteins


tyrosine hydroxylase


Clinical symptoms of patients carrying Parkinson’s disease (PD)-associated mutations of leucine-rich repeat kinase 2 (LRRK2) are indistinguishable from typical sporadic PD. The spectra of neuropathological features of PARK8 (LRRK2) patients is broad and appears to encompass those associated with other familial PD cases such as PARK1 (α-synuclein) and PARK2 (Parkin). However, the neuropathology of PARK8 is variable and is not always associated with the presence of intracellular inclusions (e.g. Lewy body, tau tangles and ubiquitin inclusions) [1,2]. Recent studies also suggest that the penetrance of LRRK2 pathogenic mutations is incomplete [3,4].

LRRK2 encodes a large complex protein consisting of 2527 amino acids (285 kDa). It belongs to the ROCO family, which is defined by the presence of a Ras of complex proteins (ROC) domain followed by a C-terminal of ROC (COR) domain of unknown function [5]. LRRK2 also contains armadillo-like repeats, LRR, kinase and WD40 domains [6]. In vitro biochemical analysis demonstrates that LRRK2 contains kinase and GTPase activities that are apparently altered by pathogenic mutations of LRRK2 [7–12]. In addition, studies using cultured cells or neurons show that enhancement of kinase activity in PD-related mutants of LRRK2 is correlated with increased neurotoxicity, thus implicating a causal role of aberrant enzymatic activity of LRRK2 in neuropathogenesis [9,13–15]. However, whether this possible gain-of-function in the kinase activity of LRRK2 contributes to the the pathological process of PD has yet to be shown in mammalian models.

Understanding the physiological function of LRRK2 under normal conditions and in the context of PD remains a daunting task, especially given the complexity of LRRK2 protein structure, which consists of multiple functional domains that are likely to be involved in numerous cellular pathways. There is clearly a need to investigate LRRK2 structure/function-related proteins (e.g. ROCO family proteins) in various model systems, including lower eukaryotes and invertebrates, in order to obtain clues for building important hypotheses. The rapidity and efficiency of in vivo studies in many nonmammal models have already provided timely information about molecular mechanisms of many disease processes and will continue to impact our understanding of disease pathogenesis. This minireview will examine the previous studies of genes for ROCO or LRRK2 homologs in slime mold Dictyostelium discoideum, nematode worms Caenorhabditis elegans and fruit flies Drosophila melanogaster. It will also discuss the available information reported in the literature (albeit limited), as well as ongoing studies in several laboratories that have created LRRK2 rodent models.

ROCO proteins in slime mold D. discoideum

The finding of the conserved ROC and COR domains in LRRK2 has stirred particular interest with respect to studying the functions of the known ROCO proteins. The first ROCO protein was identified in slime mold Dictyostelium [5] and, so far, at least two ROCO proteins, GbpC and Pats1, have been characterized in vivo in this species [16,17]. Similar to LRRK2, GbpC and Pats1 both have LRR and kinase domains flanking the central ROCO sequence. The sequence arrangement of these functional motifs ‘LRR-ROC-COR-kinase’ (LRCK) is also found in LRRK2 homologs of nematode worms and fruit flies. The investigation of ROCO structure of a prokaryotic protein in Chlorobium tepidum revealed mechanistic insight into protein dimerization and the regulation of ROC GTPase activity [18], which may be involved in intramolecular control of the kinase activity [9,19]. It is possible that the subgroup of ROCO proteins (including mammalian LRRK2), which contain the conserved functional motifs ‘LRCK’, may adopt a similar structural mechanism to regulate enzymatic activities and their cellular functions (Fig. 1).

Figure 1.

 Schematic illustration of domain structures and alignment for LRRK2 and LRRK2 structure-related proteins. ANK, N-terminal ankyrin repeat domain; DEP, Dishevelled, EGL-10, pleckstrin domain; GEF, guanine-exchange factor; GRAM, glucosyltransferases, Rab-like GTPase activators and myotublarins domain.

A series of in vivo studies have revealed that the ROCO proteins, GbpC and Pats1, are involved in multiple cellular processes: chemotaxis, cell division and development. Deletion of GbpC in D. discoideum was shown to cause a reduction of chemotactic reaction towards cAMP [20]. In addition, loss of GbpC was associated with a decrease in phosphorylation of myosin II and a change in subcellular localization of myosin heavy chain [21]. The chemotactic ‘rescue’ experiment showed that the kinase domain alone is insufficient to complement the chemotactic defect in the GbpC-deletion mutant. Furthermore, the study indicated that the LRR, ROC and kinase domains are all required for chemotaxis [16]. These results thus suggest that the functional integrity of GbpC protein requires all subdomains in the core ‘LRCK’ sequence.

ROCO protein Pats1 was originally found in a genetic screen of cellular defect in cytokinesis of D. discoideum [17]. Mutant cells with Pats1 deletion exhibit abnormal cell morphology and division deficits. It was shown that the WD40 domain of Pats1 interacted with myosin heavy chain, whereas the deletion of Pats1 caused an alteration in the localization of myosin heavy chain [17]. Moreover, over-expression of the kinase domain alone resulted in a similar phenotype to that of the deletion mutants, suggesting that deregulation of kinase activity underlies the mechanism of cytokinesis impairment. Taken together, the studies of mutant phenotypes for the two ROCO proteins in D. discoideum indicate their roles in regulating cytoskeleton structures. Interestingly, it was previously noted that human LRRK2 binds to cytoskeleton-related proteins [22], microtubules [23] and phosphorylates moesin, a protein that anchors the actin cytoskeleton to the plasma membrane [24]. Although cytoskeleton proteins are known as frequent ‘contaminants’ in the process of searching binding proteins, the in vivo evidence for the relationship between ROCO proteins and cytoskeletons in D. discoideum suggests a need to further investigate the possibility that cytoskeleton proteins are the physiological targets of LRRK2. It is possible that LRRK2 regulates cytoskeletal mobility, which is linked to various cellular vesicle trafficking events.

The nematode worm models

Despite the fact that nematode worms are used extensively as an in vivo model to study disease gene function, there are only a few reports on LRRK2 or LRK-1 (LRRK2 ortholog in Celegans) available to date. LRK-1 is the only ortholog of human LRRK2 found in Celegans. The product of LRK-1 shares a conserved ‘LRCK’ core sequence with LRRK2 (Fig. 1). The first study characterized the phenotype of mutant worms containing deletion of LRK-1. It provided important evidence implicating a role for LRK-1 in synaptic vesicle localization [25]. It showed that synaptic vesicles and their associated proteins are exclusively localized in the pre-synaptic regions but not in dendrites. By contrast, in mutant worms carrying truncated LRK-1, synaptic vesicle proteins are located in axons (pre-synaptic), as well as dendritic terminals (post-synaptic). The localization of the synaptic vesicle proteins in both pre- and post-synaptic regions as a result of the lack of LRK-1 apparently is not a random event because the mislocalization of the synaptic proteins in dendrites depends on AP-1 clathrin adaptor, which is known to be involved in dendritic transport, but not on Unc104 kinesin, a motor protein required for axonal transport. Therefore, this study suggests that LRK-1 protein, as a resident of Golgi apparatus, controls the directionality of synaptic vesicle proteins by restricting these proteins from going to the dendrites. This result further indicates a critical function of LRK-1 in establishing the polarity of synaptic vesicle proteins and perhaps in regulating synaptic vesicle life cycle in the axons. However, this study did not reveal any results regarding the viability of neurons, especially dopaminergic neurons, or any functional consequence of the mislocalization of synaptic vesicle proteins.

The above study also made reference to the partial defect of chemotaxis to volatile odorants in mutant worms carrying LRK-1 deletion [25]. Although it is unknown whether the chemotaxis deficiency involves a dysfunctional cytoskeletal system (as found in mutant slime mold carrying an ROCO proteins deletion), it would be interesting to investigate the underlying mechanism associated with the loss of LRK-1 that could be related to the pre-motor symptom of hyposmia in PD.

The second study investigated how ectopic expression of human LRRK2 wild-type or G2019S mutant in worms modifies cellular responses to rotenone, a mitochondrial toxin in nematode worms [26]. The results showed that over-expression of wild-type LRRK2 offers the transgenic worms a strong protection against rotenone toxicity, whereas over-expression of G2019S LRRK2 also protects, but to a lesser degree. Furthermore, reduced endogenous LRK1-1 expression potentiates rotenone toxicity. This report implicates a role for LRRK2 in cellular protection against mitochondria-related stress. This function may be partially impaired by the PD mutation of G2019S. Interestingly, over-expression of LRRK2 wild-type, but not the G2019S mutant, extended the lifespan of worms, indicating a beneficial role of LRRK2 in the ageing process. This result suggests that overproduction of LRRK2 (wild-type or G2019S mutant) in worms, unlike in mammalian cell cultures [9,19], does not cause toxicity [26].

The third report in nematode worms, however, shows that over-expression of worm green flourescent protein-tagged LRK-1 wild-type or LRK-1 G1876S (corresponding to human G2019S) leads to an early larval arrest [27]. Although this result may suggest that over-expresssion of worm LRK-1 is much more toxic than human LRRK2, it also raises the question of whether or not LRK-1 is a true ortholog of human LRRK2 [6]. Importantly, this study suggests a genetic link between Lrk-1 and Pink-1, a Celegans homolog of human PINK-1 that is associated with a recessive form of PD [28]. Mutant worms that lack Lrk-1 were shown to have enhanced sensitivity to endoplasmic reticulum stress induced by tunicamycin, a specific inhibitor for N-linked glycosylation. Interestingly, this enhanced sensitivity is suppressed in mutant worms with deletion of both Lrk-1 and Pink-1 genes. On the other hand, although Pink-1 mutant worms exhibit increased vulnerability to paraquat, defects in mitochondrial cristae and impairment of axonal guidance, a lack of Lrk-1 appeared to reverse the Pink-1 deletion-associated defects in double mutant Lrk-1 and Pink-1. This study suggests an antagonistic role of Lrk-1 and Pink-1 in stress response and neuronal activities [27].

Fruit fly models of LRRK2

The fruit fly homolog of LRRK2 is dLRRK, which also contains the conserved ‘LRCK’ core sequence (Fig. 1). To date, at least four studies have reported using fruit fly D. melanogaster to investigate the in vivo functions of human LRRK2 or dLRRK. The first study showed that the mutant flies lacking dLRRK exhibited impaired locomotive activity and a significant reduction of tyrosine hydroxylase (TH) immunostaining in dopaminergic neurons. Although the number of dopaminergic neurons appears unaltered, they display abnormal morphology, suggesting that they are under pathogenic stress or undergoing slow degeneration [29]. Two other studies, however, did not reproduce the behavioral and TH deficits in mutant flies carrying deletion of dLRRK. Instead, they observed unchanged numbers of TH+ neurons in these mutants, indicating that dLRRK is dispensable for the survival of dopaminergic neurons [30,31]. In addition, Wang et al. [30] showed that mutant flies containing C-terminal kinase domain truncated dLRRK are selectively sensitive to H2O2, but not to paraquat, rotenone or β-mercaptoethanol. By contrast, Imai et al. [31] showed that dLRRK null flies are relatively resistant to general oxidative stress, such as paraquat and H2O2 treatment, compared to wild-type flies [31]. Furthermore, dLRRK null flies have significant reduced levels of 4-hydroxy-2-nonenal of lipid peroxidation, an indication of oxidative damage. Although the exact role of dLRRK in oxidative stress remains unclear, all studies in fly models reported to date consistently demonstrate that dLRRK is not essential for the early development and viability of dopaminergic neurons.

The results obtained from studies of transgenic flies over-expressing dLRRK or human LRRK2 have been somewhat inconsistent between the different groups. Although Lee et al. [29] indicated that over-expression of a pathogenic mutant or wild-type dLRRK did not cause any significant defects in transgenic flies, two other independent reports demonstrated that expressing mutants of dLRRK or LRRK2 in flies causes selective degeneration of dopaminergic neurons as well as motor function deficits [31,32]. Of these two reports, however, one showed that even over-expressing wild-type human LRRK2 led to the toxicity of dopaminergic neurons and impairment of motor function (although to a lesser degree than LRRK2 G2019S) [32], whereas the other indicated that over-expressing wild-type dLRRK did not affect the number of dopaminergic neurons or motor function [31].

Interestingly, two studies have shown a relationship of LRRK2 or dLRRK to the dopamine physiology. Liu et al. [32] found that treatment of l-DOPA improved the motor impairment of transgenic flies caused by LRRK-G2019S but not the degeneration of TH+ neurons. The results obtained by Imai et al. [31] suggested that dLRRK is involved in negatively regulating homeostatic levels of dopamine. They demonstrated that the over-expression of a PD-pathogenic mutant of dLRRK (but not wild-type dLRRK) resulted in a reduction in brain dopamine levels compared to that of nontransgenic flies. Conversely, dopamine content was elevated in mutant flies with a dLRRK deletion. This increase in dopamine content is likely to be a result of dopamine release, uptake or metabolism, but not to an alteration of TH+ neuron numbers [31].

Finally, Imai et al. [31] provided evidence that both dLRRK and LRRK2 kinase can phosphorylate eukaryotic initiation factor 4E-binding protein, a negative regulator of eukaryotic initiation factor 4E-mediated protein translation and a key mediator of various stress responses. They proposed a model in which LRRK2 mediates the pathological effect in part through modulating translation initiation [31].

In summary, these studies in fruit flies have provided important in vivo information regarding the potential function of LRRK2 (Table 1). Indeed, certain observations reported in fly models may appear to be conflicting. However, it is possible that the different genetic backgrounds, the genomic locus of insertion for gene disruption, transgenic expression levels and nutrient conditions are responsible for the divergent results. These issues need to be resolved in the future in order to understand better the physiological function of LRRK2, as well as the pathogenic effect of PD mutations of LRRK2. Once validated in mammalian models, selected models could serve as a robust system for revealing the genetic pathways of LRRK2 in PD and for screening chemical compounds to intervene with the LRRK2-mediated pathogenesis.

Table 1. In vivo models for LRRK2 and LRRK2 homologs.
Truncation of endogenous LRK-1WormSubtle defects in movement; partially defective in chemotaxis to volatile odorants; impairment of polarized synaptic vesicle localization[25]
Transgenic expression of human LRRK2WormOver-expression of LRRK2 wild-type protects against rotenone toxicity and extend life span; over-expression of LRRK2 G2019S also protects but to a lesser extent[26]
Transgenic expression of green flourescent protein-tagged LRK-1 wild-type or G1876S mutantWormOver-expression of either LRK-1 wild-type or G1876S (corresponding to human G2019S) leads to an early larval arrest[27]
Disruption of endogenous LRK-1WormAntagonistic action of worm Lrk-1 versus Pink-1 in stress response and neuronal activities[27]
Transgenic expression and disruption of endogenous dLRRKFlyNo obvious behavioral abnormality associated with transgenic over-expression of dLRRK wild-type or mutant; deletion mutant shows impaired locomotive activity and a significant reduction of TH immunostaining in dopaminergic neurons[29]
Disruption of endogenous dLRRKFlyNo obvious behavioral deficits; unchanged TH+ neurons; enhanced sensitivity to H2O2[30]
Disruption of endogenous dLRRKFlyRelatively resistant to general oxidative stress; reduced oxidative damage; unchanged TH+ neurons; increased dopamine content[31]
Transgenic expression of dLRRKFlyOver-expression of ‘pathogenic’ dLRRK mutant caused loss of TH+ neurons in aged mice and reduced dopamine content; over-expression of wild-type dLRRK2 or kinase-dead mutant had no effect on viability of TH+ neurons[31]
Transgenic expression of human LRRK2FlyOver-expression of LRRK2 wild-type or G2019S mutant causes loss of TH+ neurons and impairment of motor function (with worse phenotype in G2019S mutant flies); treatment of l-DOPA improves motor function but not neurodegeneration[32]
LRRK2 KOMouseNo description of characterization[34]
LRRK2 KOMouseSurvived normally; display no overt behavioral abnormality; unaltered number of dopaminergic neurons for up to 24 months[41, and unpublished results]
BAC transgenics/murine LRRK2MouseOne line expressing FLAG tagged LRRK2 wild-type (> 20-fold) shows regulated expression pattern and unaltered TH+ neuron morphology or number[11]
BAC transgenics/human LRRK2MouseHuman BAC mice show very similar expression pattern to mouse BAC transgenic lines [11]; over-expression of LRRK2 wild-type (> 20-fold), G2019S or Y1699C (seven- to 11-fold and up to 12 months) did not cause overt behavioral abnormalities[42,47]
Tetracycline-regulated transgenics/human LRRK2 G2019SMouseNo obvious neuropathologies or motor abnormalities at 12 months and older[41]
BAC transgenics/human LRRK2 R1441CMouseBAC mice expressing LRRK2 R1441C develop typical motor function abnormality related to PD; no obvious loss of midbrain TH+ cells; age-dependent and levodopa-responsive slowness of movement associated with reduced dopamine release and axonal pathology of nigrostriatal dopaminergic projection[33]

Rodent models of LRRK2 and bacterial artificial chromosome (BAC)-mediated LRRK2 transgenic mice

Although there have been several studies reporting the generation of genetically engineered LRRK2 mice (including targeted deletion and transgenic expression), no systematic investigation of these mice has been described to date (one report was published recently during the preparation of this manuscript [33]). Therefore, the physiological role of LRRK2 in the mammalian central nervous system remains largely elusive.

Analysis of LRRK2 expression in mouse brain shows that it is broadly distributed in many regions, including the cerebral cortex, hippocampus, striatum, amygdala, cerebelluam and olfactory bulb, as well as in ventral tegamental area and substantia nigra (albeit at low levels) [11,34–40]. Analysis of LRRK2 expression levels during pre- and post-natal stages reveals that the LRRK2 protein appears at embryonic day 17 (E17) and is increasingly produced over the early post-natal stage [11,34], reaching peak levels by 2 months [11].

LRRK2 knockout (KO) mice

Biskup et al. [34] were the first to report the generation of LRRK2 KO mice. Taking advantage of the lack of LRRK2 expression in these mice, they performed the comprehensive evaluation of a panel of commercial antibodies against LRRK2 for their staining specificity. However, no characterization of these mice was shown [34]. Although without showing any experimental data, a study by Wang et al. [41] indicated that LRRK2 KO mice survive normally, and that they do not develop any obvious neuropathological abnormalities or motor dysfunctions up to 12 months of age. Indeed, no loss of dopaminergic neurons or motor behavioral deficits was observed even at 24 months of age in LRRK2 KO mice (Dr H. Cai, personal communication). This result, along with the study showing the developmental expression levels of LRRK2, suggests that the role of LRRK2 in early embryonic development is negligible, but may be important for cellular function at the adult stage. Furthermore, although it is possible that the lack of LRRK2 function can be compensated for by LRRK2 function-related molecules (e.g. LRRK1), this observation in LRRK2 KO mice is consistent with the findings in nematode worms and fruit flies that the deletion of the single homolog of LRRK2 in either species has no effect on the viability of dopaminergic neurons. Therefore, we propose that LRRK2 (as well as LRRK2 homologs dLRRK and LRK-1) does not play a major role in a cellular pathway that is critical for neuronal survival. Rather, it is involved perhaps in specific neuronal functions that can only distantly modulate neuronal survival or death in an age-dependent manner.

It is also not surprising that the deletion of the LRRK2 gene does not lead to degeneration of dopaminergic neurons in mice, given that disruption of all known PD-related genes, such as α-synuclein, Parkin, DJ-1 and PINK-1, has not been associated with any obvious loss of dopaminergic neurons in mice. It is intriguing to note that none of these PD-related genes are essential for neural development and differentiation, which is also in support of the hypothesis that dysfunction of these genes only leads to disruption of neuronal functions mostly at the adult stage in PD.

BAC transgenic mice of LRRK2

To date, three laboratories have reported the availability of LRRK2 transgenic mice without providing details of the characterization of these mice. Two laboratories, including ours, generated BAC-transgenic mice expressing murine FLAG-tagged LRRK2 [11] and human LRRK2 [42], whereas the third indicated the usage of a tetracyline-regulated system for the transgenic expression of human G2019S LRRK2 [41]. The application of BAC-transgenic mice was initially described in 1997 [43] and has grown significantly over the past decade because of its usefulness in studying gene function in vivo, particularly in the central nervous system [44]. Growing evidence demonstrates the power of this transgenic approach in conferring correct transgene expression under endogenous promoter control with little concern about positional effect [45]. The BAC transgenic approach has been successfully used in establishing mouse models for neurodegenerative diseases [46] and is expected to contribute to an understanding of the disease mechanisms in vivo.

The application of BAC transgenics is especially advantageous over conventional transgenics for studying LRRK2. The main reasons are: (1) generation of LRRK2 BAC transgenic mice does not involve the synthesis of full-length LRRK2 cDNA, which is a > 7 kb nucleotide and technically difficult to manipulate as a result of the large size; (2) the entire genomic sequence of mouse or human LRRK2 is approximately 180 kb, which is the average length of BAC clones that are readily available in public domains; and (3) LRRK2 BAC transgenes with introduced PD mutations are suitable for modeling the LRRK2-mediated pathological process as a result of the dominant disease transmission for LRRK2 mutations. Our laboratory has previously generated numerous BAC transgenic lines expressing FLAG-tagged LRRK2 wild-type. Examination of the transgene expression in the brain shows a similar distribution pattern in all transgenic lines. We identified one BAC line that produces FLAG-LRRK2 wild-type protein at a level twenty-fold greater than the endogenous LRRK2 protein [11]. Unexpectedly, the transgenic mice overloaded with the exogenous LRRK2 did not show obvious neurotoxicity or motor function abnormalities over 20 months (X. Li and Z. Yue, unpublished results), despite FLAG-LRRK2 purified from transgenic brain displaying robust kinase and GTPase activity [11].

Melrose et al. [42] previously reported the generation of BAC transgenic mice producing human LRRK2 wild-type or mutants. Although no information was given about the viability of TH+ neurons, it was indicated that BAC transgenic mice over-expressing LRRK2 wild-type [20-fold for up to 24 months), mutant G2019S or Y1699C (seven- to 11-fold for up to 12 months) did not show an overt behavioral phenotype [47]. Consistent with these observations, tetracycline-regulated transgenic mice producing LRRK2 G2019S were also reported to be spared of obvious neuropathologies or motor abnormalities at 12 months and older [41].

Interestingly, a more recent study by Li et al. [33] suggests that BAC transgenic mice expressing the human LRRK2 R1441C mutant develop typical motor function deficit related to PD. These mice are associated with the degeneration of TH+ axons and tauopathy, as well as TH+ cell atrophy, despite lacking obvious loss of midbrain TH+ cells. Furthermore, these BAC models develop an age-dependent and levodopa-responsive slowness of movement associated with diminished dopamine release and axonal pathology of nigrostriatal dopaminergic projection [33]. Although this transgenic line provides a promising model for further dissection of LRRK2-associated PD pathogenesis, future experiments will be needed to resolve the difference in behavioral as well as possible pathological phenotypes observed among different BAC transgenic models. Although it remains unclear at present, the different PD-related mutations examined and the distinct genetic background of the host mice, as well as the varied transgene expression levels, may be responsible for the differential phenotype of these BAC models.

It is mysterious that none of the reported LRRK2 transgenic mice show the loss of dopaminergic neurons or the accumulation of α-synuclein at substantia nigra, the hallmarks of PD pathology. Alhough the physiological function of LRRK2 has yet to be formally demonstrated in these transgenic models, current evidence suggests that the pathological consequence of over-expressing only LRRK2 wild-type or PD-related mutations in rodent model is mostly neuron dysfunction, rather than degeneration. One of the intriguing findings reported by Li et al. [47] is that the LRRK2-R1441C BAC mice show reduced dopamine release, which is consistent with previous studies conducted in fruit flies showing the connection of dLRRK to dopamine and movement control [31,32]. In addition, we found that BAC transgenic mice expressing LRRK2-G2019S also displayed a decrease in dopamine release and striatal dopamine levels in the absence of obvious neuropathology (X. Li, J. C. Patel and coworkers, unpublished results). These results suggest a pathogenic role of LRRK2 mutants in the deregulation of the striatal dopamine system. Whether other PD-related mutants of LRRK2 also have the same effect, and whether the normal function of LRRK2 is involved in striatal dopamine transmission, remains to be shown.

The above observation, therefore, is in line with the previous evidence indicating that single genetic alteration of PD-related genes, such as the over-expression of a dominant gene α-synuclein or the deletion of a recessive gene (DJ-1, Parkin or PINK1), is unlikely to recapitulate the full spectrum of PD. The lack of manifestation of the most important PD hallmarks in LRRK2 transgenic mice (e.g. dopaminergic neuron loss and deposits of α-synuclein in Lewy body) is also not surprising considering that LRRK2 PD-mutations are not fully penetrant and that LRRK2 patients display a broad range of clinical phenotypes [1–4]. Therefore, the current challenges facing us are not only to ‘tease out’ LRRK2-associated neuronal functions that are perturbed as a result of PD-related mutations, but also to identify the cellular pathways or factors that cross-talk with and thus can significantly modify LRRK2-mediated phenotypic expressions.

Concluding remarks

The invertebrate models including nematode worms and fruit flies have begun to unveil the functions of the orthologs of LRRK2 in vivo (Table 1). Although rapid analysis of these models will undoubtedly facilitate an understanding of the function of LRRK2 function in PD, the lack of a sophisticated structural and functional equivalent of the human central nervous system in these organisms limits their application when understanding the in vivo function of LRRK2 in humans. The ultimate comprehension of LRRK2 physiology and pathophysiology in PD will still depend on the establishment and detailed characterization of mammalian models of LRRK2. The collective data obtained from both KO and transgenic mouse models (albeit in preliminary form) suggest that LRRK2 is not essential for neural development and differentiation, and that it does not play a primary role in cell death pathways. However, these LRRK2 rodent models should provide valuable tools for dissecting the specific neuronal functions of LRRK2 (e.g. dopamine transmission) and likely pre-symptomatic (or early) events of the disease process. They should also be useful in testing the ‘two-hit’ or ‘multiple-hit’ hypothesis proposing that LRRK2 and other genetic or environmental factors are required to work together and facilitate the pathological process of PD.


I wish to thank Drs Chenjian Li, Huaibin Cai, Xianting Li and Sarah Funderburk for their critical comments, and Dr Huaibin Cai for sharing unpublished results. I am also grateful to Dr Nina Pan for assisting in the preparation of Fig. 1 and Table 1. This work was supported by grants to Z.Y. from the US NIH/NINDS NS061152, NS060809, RNS055683A, the Michael J. Fox Foundation, and the Bachmann-Strauss Dystonia & Parkinson Foundation.