Targeting LRRK2 Mutations in Parkinson's Disease

Keywords: :

• computational chemistry and molecular modeling
• drug design
• drug discovery
• inflammation therapeutics
• neurological therapeutics
• pharmacokinetics/pharmacodynamics
• rare diseases

Parkinson's disease (PD), originally named ‘shaking palsy’ [1], is a chronic neurodegenerative disorder manifesting hallmark symptoms of resting tremor and bradykinesia [2]. Since the first clinical characterization of PD in 1817, our understanding of this disease has expanded to recognize both motor and nonmotor symptoms. Modern treatments for PD remain largely symptomatic, meaning that they relieve symptoms but apparently do not target the underlying cause(s) of the disease, and are not without severe side effects in many cases. With an increasing global prevalence of PD (∼10 million), an urgent unmet medical need remains: to slow, prevent and perhaps even reverse the neuronal injury that underlies the debilitating symptoms of PD [3].

PD is characterized neuropathologically by the degeneration of dopaminergic neurons in the substantia nigra compacta (SNpc) and by the appearance of intraneuronal Lewy bodies, which are formed in part by aggregated pathologic forms of α-synuclein, a presynaptic protein. Although originally thought to have minimal genetic influence, research over the last two decades has established numerous genetic causes and genetic risk loci for PD [4,5]. Of the genetic variants that have been associated with PD, by far the most common genetic cause of familial and sporadic PD are mutations in LRRK2 (PARK8) [6,7]. Transcription and translation of this gene yield a large (286 kDa) multidomain protein, LRRK2, a member of the ROCO superfamily of proteins. LRRK2 is composed of a tandem Ras of complex (Roc) G-domain linked to a kinase domain through a carboxy-terminal of Roc (COR) sequence. Outside of the characteristic ROCO family motifs, LRRK2 possesses protein–protein interaction (PPI)-associated domains, namely WD40, armadillo (ARM), ankyrin (ANK) and leucine-rich repeats (LRR). Although the precise physiological function of LRRK2 remains elusive, evidence suggests its primary functions are largely performed through its two enzymatic domains, kinase and ROC-GTPase, which catalyze phosphorylation and GTP-GDP hydrolysis, respectively. The roles of the multiple PPI domains of LRRK2 are also the subject of intensive research and the proposed cellular functions of LRRK2 are numerous. Indeed, LRRK2 has been implicated in neurite outgrowth, vesicle trafficking, cytoskeletal maintenance and autophagy, to name a few. While elucidation of the exact mechanisms of LRRK2 in the PD context is ongoing, one fact remains clear: modulating at least one of the critical functions of LRRK2 may provide a valid therapeutic target for medicinal chemistry efforts.

To date, of the nearly 100 known mutations of LRRK2 [8], only a few are established as dominantly inherited causes of PD. Some notable examples are the G2019S, I2020T, R1441C/G/H, N1437H and Y1699C mutations occurring in catalytic domains of LRRK2; however, all increase the kinase activity of LRRK2 by varying amounts [9]. The G2019S mutation is the most prevalent in PD populations, accounting for around 6% of familial and 2% of sporadic PD, and has a high frequency in specific geographic populations. The G2019S mutation (and I2020T variant) occurs in the kinase domain of LRRK2, specifically in subdomain VII ‘DFG’ motif (DYG in LRRK2) and generates a hyperactive kinase, increasing K_cat but not V_max (wildtype [WT]) LRRK2 kinase [9]. This hyperactive kinase activity thus is strongly supported by genetic evidence as a significant contributor to the pathogenesis of PD in G2019S LRRK2 mutation carriers. It is perhaps unsurprising then that modulation of the kinase domain of LRRK2 as a means to treat PD has motivated multiple medicinal chemistry campaigns over the last decade-plus, producing a wide variety of extremely efficacious LRRK2 kinase inhibitors, for example, MLi-2, GNE-7915 and PF-360 developed by Merck [10], Genentech [11] and Pfizer [12], respectively. These potent, brain-penetrant and highly kinome-selective LRRK2 inhibitors can be considered type I kinase inhibitors, binding to the kinase hinge region of the active site and competing with ATP, while the kinase is in the DYG-in configuration. To the best of our knowledge, published brain-penetrant LRRK2 kinase inhibitors are nonspecific toward pathogenic mutant LRRK2 variants over WT LRRK2. Interestingly, despite initial safety concerns raised with kinome-selective LRRK2 kinase inhibitors GNE-0877 and GNE-7915 [13], clinical trials are proceeding with likely structurally similar ATP-competitive inhibitors [14].

Given that WT LRRK2 is proposed to be involved in many crucial cellular processes, it is expected that unwanted side effects may arise from the administration of inhibitors that do not discriminate between the WT and mutant LRRK2 variants. In this way, the development of kinase inhibitors that are selective toward G2019S LRRK2 while allowing the WT to maintain crucial cellular activities would be advantageous to heterozygous LRRK2 mutation carriers. Although a direct association between LRRK2 kinase inhibitor-associated side effects and lack of selectivity for G2019S LRRK2 has not been established, a more precise approach may be an attractive therapy for certain PD patients and may finally help elucidate the molecular and cellular mechanisms of LRRK2 kinase mutations in PD pathogenesis. The design of mutant-selective LRRK2 kinase inhibitors, however, poses a significant challenge. In the context of G2019S LRRK2, the presence of a single amino acid substitution resulting in a largely identical kinase active site and a lack of x-ray structural data has resulted in a heavy reliance on molecular docking and homology modeling in LRRK2 kinase medicinal chemistry programs. Indeed, recent single-particle cryoelectron microscopy (Cryo-EM) structures of G2019S and WT LRRK2 revealed both can adopt an almost identical kinase-inactive (DYG-out) conformation, leading the authors to suggest that G2019S LRRK2 hyperkinase activity may be a kinetic effect rather than a result of structural difference in the kinase active site [15,16]. This would suggest the generation of selective G2019S LRRK2 inhibitors is a fruitless endeavor. However, these findings are contrary to previous kinetic and computational studies involving type II kinase inhibitors (preferentially binding to and inhibiting kinases in their DYG-out conformation) showing the G2019S mutation (and also I2020T) may stabilize an active (DYG-in) conformation in solution compared with WT LRRK2 [15,16]. This stabilizing effect may be achieved through the contribution of Ser2019 to a hydrogen-bonding network with adjacent catalytic residues. WT LRRK2 was shown to readily access the active (DYG-in) and inactive (DYG-out) conformations, as many kinases do, thus a preference of type II kinase inhibitors for WT LRRK2 was observed over G2019S and I2020T LRRK2 variants. If true, unique binding sites may be present and thus exploited in the G2019S kinase domain by appropriately designed small molecules. Furthermore, an x-ray structure of various LRRK2 kinase inhibitors bound to a G2019S LRRK2 surrogate (mutated CHK1) revealed the kinase to be in an active (DYG-in) conformation, suggesting the presence of type I inhibitors may be required to stabilize this conformation [17,18]. Although such x-ray structural data using LRRK2 surrogates have yielded potent inhibitors, studies [19,20] have indeed shown that significant selectivity toward the G2019S LRRK2 mutant over WT can be obtained through systematic probing of the kinase active site using a combination of iterative synthesis of molecules, follow-up screening and molecular docking. Through such a process, it is possible to obtain a data-driven model of the G2019S LRRK2 active site, visualize potential structural differences between WT and G2019S LRRK2 kinase and design more selective inhibitors. From these recent medicinal chemistry efforts [19,20], it is clear that, as suggested in the aforementioned studies with type II inhibitors [15,16], the G2019S kinase may possess a unique binding site compared with WT. This structural difference provides opportunities for G2019S LRRK2-selective inhibitor design. Indeed, specific inhibitor modifications have resulted in significant selectivity toward the GS mutant, most likely through unique Van der Waals interactions with the highly mobile glycine-rich loop of the G2019S LRRK2 kinase active site. Such interactions may not be readily accessible to other LRRK2 kinase inhibitors.

Although only differing by a single amino acid, the G2019S mutation clearly exerts a strong, ultimately pathogenic effect on the cellular functions of LRRK2. This may be through a structural effect on the kinase domain, which in turn affects global enzyme kinetics, architecture and substrate profile. In the absence of explicit x-ray structural data, selective inhibitor generation must rely on a data-driven approach to drive selective inhibitor design through the medium of computational modeling. Indeed, studies have shown that selective and efficacious mutant LRRK2 kinase inhibition is not only possible but also effective in animal models of mutant LRRK2-associated PD [19]. Alongside kinase inhibition of pathogenic LRRK2 variants, WT LRRK2 has itself also been suggested to play a role in PD pathogenesis, enabling LRRK2 inhibitors (nonselective for genetic variants) to progress through late-stage clinical trials. Time will tell if such compounds possess the ability to address the fundamental unmet need for effective but also safe PD treatments. While the advancement of these potential groundbreaking treatments for PD is highly encouraging, the progression of mutant-selective LRRK2 inhibitors provides an alternate and precise approach that may yield safer medicines. This approach would perhaps only benefit G2019S LRRK2 carriers in the first instance. However, a clinically useful mutant-selective LRRK2 inhibitor would provide unique insights into the molecular biology of LRRK2-associated PD, informing future drug discovery efforts for precision medicines for people with PD.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.