The design and development of LRRK2 inhibitors as novel therapeutics for Parkinson’s disease

ABSTRACT

Parkinson’s disease (PD) is a common neurodegenerative disease affecting nearly 10 million people worldwide and placing a heavy medical burden on both society and families. However, due to the complexity of its pathological mechanisms, current treatments for PD can only alleviate patients’ symptoms. Therefore, novel therapeutic strategies are urgently sought in clinical practice. Leucine-rich repeat kinase 2 (LRRK2) has emerged as a highly promising target for PD therapy. Missense mutations within the structural domain of LRRK2, the most common genetic risk factor for PD, lead to abnormally elevated kinase activity and increase the risk of developing PD. In this article, we provide a comprehensive overview of the structure, biological function, and pathogenic mutations of LRRK2, and examine recent advances in the development of LRRK2 inhibitors. We hope that this article will provide a reference for the design of novel LRRK2 inhibitors based on summarizing the facts and elucidating the viewpoints.

GRAPHICAL ABSTRACT

1. Introduction

PD is one of the most common neurodegenerative diseases, second only to Alzheimer’s disease (AD) in prevalence and growing at a much faster rate, making it a significant health problem affecting the world [1]. Epidemiological studies have shown that the prevalence of PD increases progressively with age, with a prevalence of approximately 1% among individuals over the age of 60 and exceeding 4% in people older than 80 years [2]. The number of patients diagnosed with PD doubled between 1990 and 2016, reaching 6.1 million worldwide [3]. Influenced by global aging, this figure is projected to double again by 2030 [4]. Despite the widespread range of effects associated with PD, there is currently no available treatment that can effectively delay its progression. Therefore, the development of novel therapeutic drugs for PD is necessary.

The clinical manifestations of PD include motor and non-motor symptoms, both of which progressively worsen over time. Typical motor symptoms, which often present in the early stages, include tremors, bradykinesia, and rigidity [5]. Non-motor symptoms include cognitive impairment, sleep disturbances, dementia, and autonomic dysfunction [6]. Pathologically, the features of PD include the loss of dopamine neurons in the substantia nigra pars compacta (SNpc), depletion of dopamine in the striatal area, and the abnormal deposition of intra-neuronal Lewy bodies (LBs) [7]. Under pathological conditions, α-synuclein misfolds, aggregates into insoluble fibrillar protein deposits, and spreads to other neuronal cells, subsequently affecting the entire brain [8]. Abnormally deposited α-synuclein is neurotoxic and can be degraded and cleared under normal conditions. However, due to the indirect effects of several mechanisms, such as damage to the protein clearance system, mitochondrial dysfunction, neuroinflammation, and oxidative stress, abnormal proteins gradually accumulate, ultimately leading to neuronal death [9–12]. It is widely accepted that PD symptoms emerge when there is a loss of 50–70% of the nigrostriatal dopaminergic neurons in the brain [13].

The pathogenesis of PD is complex, characterized by numerous risk factors and significant heterogeneity in clinical manifestations [14]. Consequently, the treatment of patients with PD should prioritize individualized features and be initiated at the earliest possible stage [15]. Current clinical treatment of PD is based on pharmacological interventions, with a primary focus on symptomatic treatment [16], aiming to improve both motor and non-motor symptoms of patients. While dopamine-based therapies usually help alleviate motor symptoms, they do not halt the progression of the disease [17]. Several drugs are currently available for clinical treatment, and most of them act by involving either direct or indirect enhancement of the dopaminergic system, thereby alleviating patients’ symptoms to a certain extent [15]. Levodopa is one of the principal therapeutic agents used in the treatment of PD, often in combination with carbidopa or benserazide, and it has the highest efficacy in the treatment of motor symptoms [16]. However, long-term administration of levodopa may lead to the emergence of motor complications [18]. In addition, these medications do not adequately address the non-motor symptoms experienced by patients, and patients require non-dopaminergic medications for relief [19].

In the context of an aging population, the demand for novel therapeutic agents for PD is growing. It is widely acknowledged that PD is a complex, multifactorial disease influenced by various factors, including age, environmental factors, genetics, and exposure to neurotoxins [13,20,21]. Even though the majority of PD cases show a sporadic onset, approximately 5–15% of PD cases show a clear familial inheritance [22]. More than 20 genetic mutations have been identified as being associated with PD [20], and genome-wide association studies (GWAS) have identified 90 independent susceptibility loci [23]. Recent years have witnessed significant advancements in research focused on the genetic basis of PD, and these mutation loci that have been identified as responsible for monogenic forms of PD are known as PARK genes [24]. Among them, mutations in PARK8 are the most common cause of both sporadic and familial PD [25]. PARK8 is located on chromosome 12, encoding leucine-rich repeat kinase 2 protein, which is also known as Dardarin, a term that means tremor [26]. Pathogenic missense mutations in LRRK2 are highly associated with familial and idiopathic PD [27]. Drugs developed targeting this kinase have the potential to slow down the progression of PD, which current therapeutic methods are unable to achieve. Consequently, the development of therapies targeting LRRK2 has emerged as a prominent area of research, and four small molecule inhibitors have entered clinical trials. The subsequent sections of this article will describe the structure, pathogenesis, and mutations of LRRK2, providing a brief overview of the current status of research on LRRK2 inhibitors.

2. LRRK2 and PD

2.1. Structure of LRRK2

LRRK2, first discovered in 2004, is a large and complex protein that is highly associated with autosomal dominant familial PD [26,28,29]. LRRK2 belongs to the ROCO family of proteins, and consists of 2527 amino acids divided into seven structural domains (Figure 1) [30]. Three of these structural domains, the Ras of Complex (ROC), the C-terminal of ROC (COR), and the Kinase Domain (KIN), constitute the catalytic core of the protein. In contrast, other structural domains, located on either side, are involved in protein-protein interactions: the N-terminal Armadillo Repeat (ARM), Ankyrin-Like (ANK), Leucine-Rich Repeat (LRR), and the C-terminal Trp-Asp-40 (WD40) structural domain [31]. Notably, all LRRK2-related mutations that may cause PD are located within the core kinase structural domain and affect kinase activity [32]. Protein kinases tend to have conserved ATP-binding pockets and regions that undergo phosphorylation reactions and can switch between active (DFG-in) and inactive (DFG-out) conformations [33]. Unlike the conserved DFG motifs found in other protein kinases, the three important amino acids that make up the activation loop of LRRK2 are Asn2017, Tyr2018, and Gly2019, collectively referred to as the DYG motif (Figure 2(b)) [34].

Figure 1.

(a) Schematic diagram of LRRK2 domain. (b) The structure of the LRRK2 monomer (PDB ID: 8FO2). (c) Magnified image of KIN and ROC domains in LRRK2. ARM, ANK, LRR, ROC, COR, KIN, and WD40 domains are colored in red, blue, orange, cyan, green, gray, and yellow, respectively. Seven pathogenic amino acid variants are highlighted in magenta. The phosphorylation sites S935 and S1292 are colored in navy blue. ATP and GDP are displayed with green carbon, blue nitrogen, red oxygen, and orange phosphorus atoms.

Figure 2.

(a) LRRK2 is involved in the pathogenesis of PD. Abnormally elevated LRRK2 activity leads to lysosomal dysfunction, impaired vesicle transport, mitochondrial dysfunction, and neuroinflammation, thus leading to PD. (b) The dfg-in conformation of the KIN domain in LRRK2 (PDB ID: 8FO9). (c) The magnified image of the DYG motif in WT LRRK2 (PDB ID: 8TXZ). (d) The magnified image of the DYS motif in G2019S LRRK2 (PDB ID: 8TZC).

The structure of LRRK2 has been explored for 20 years. The first full-length cryo-electron microscopy structure of LRRK2 (PDB ID: 7LHW) was published in 2021 [35], with a significant displacement of the catalytic structural domain compared to the previously reported half-structure of LRRK2 (PDB ID: 6VNO) [36]. Since 2023, two research teams have disclosed the crystal structures of the LRRK2-Rab29 complex and the LRRK2-inhibitor complex, revealing the specific mechanism of LRRK2 activity regulation and providing a basis for the rational design of LRRK2 inhibitors [37–39].

2.2. Function of LRRK2

LRRK2 is predominantly expressed in several tissues throughout the human body, with high levels of expression observed in periphery tissues such as the lungs, kidneys, and immune cells, while exhibiting moderate expression in the brain. This differential expression provides insight into the mechanisms by which LRRK2 is involved in PD [40]. Within the brain, LRRK2 is mainly expressed in neurons, astrocytes, and microglia, and is particularly enriched in membrane-bound organelles in the cytoplasm, such as lysosomes, the endoplasmic reticulum, and the Golgi apparatus. This suggests that LRRK2 plays an important role in membrane trafficking [41].

As a GTPase, LRRK2 plays a crucial role in a variety of cellular processes, including: regulating vesicular transport, the autophagy system, and influencing lysosomal function [42]; regulating neuronal ciliogenesis [43]; participating in mitochondrial autophagy [44]; influencing microtubule dynamics [45]; and contributing to the innate immune system [46]. The processes regulated by LRRK2 May be interrelated (Figure 2(a)).

LRRK2 exists intracellularly as monomers, dimers, and tetramers [35], with dimers being significantly enriched in the cell membrane, indicating that the dimeric form is essential for the membrane localization function of LRRK2 [47]. Recent studies have shown that the kinase activity of LRRK2 is closely correlated with its aggregation state [37]. Rab29 serves as a substrate for LRRK2, which is recruited to the cell membrane, leading to the formation of dimers and asymmetrically activated tetramers from LRRK2 monomers. Activated LRRK2 phosphorylates Rab GTPase substrates, thereby exerting physiological effects [48].

2.3. Pathogenic mechanisms of LRRK2

LRRK2 mutations are strongly associated with the pathogenesis of PD. These mutations are the most common factor contributing to late-onset familial PD and idiopathic PD [27], accounting for 5–13% of familial and 1–5% of sporadic PD [49]. Patients with PD who carry LRRK2 mutations usually develop the disease between the ages of 60 and 70 and exhibit clinical symptoms similar to those of idiopathic PD [50]. The LRRK2 mutations associated with PD are primarily clustered within the ROC-COR and KIN structural domains (Figure 1(a)), and will increase the phosphorylation levels of substrates [51–53]. In addition, LRRK2 activity is significantly elevated in postmortem brain tissue from PD patients, and oxidative stress associated with PD pathogenesis may further upregulate LRRK2 activity in vivo [54]. All of these studies suggest that LRRK2 kinase activity plays an important role in the pathogenesis of PD.

A variety of potential LRRK2-mediated mechanisms have been proposed in pathophysiological studies of PD, including neuroinflammation and dysfunction of the autophagy-lysosomal pathway (Figure 2(a)) [55].

LRRK2 is significantly expressed in peripheral immune cells and microglia in the brain, making it a modulator of neuroinflammation and proinflammatory cytokines [56]. Under the influence of reactive oxygen species (ROS) induced by rotenone, endogenous LRRK2 in neuronal cells is activated, resulting in an elevated phosphorylation level of LRRK2 substrates [54]. This scenario resembles that observed in the brain tissue of patients with sporadic Parkinson’s disease (PD), suggesting that oxidative stress, a pathogenic mechanism of PD, can upregulate LRRK2 activity. Conversely, inhibition of LRRK2 can ameliorate α-synuclein-mediated neuroinflammation and dopaminergic neuronal degeneration, indicating that LRRK2 activity can potentiate neuroinflammatory responses and exacerbate neuronal degeneration [57]. These findings indicate that LRRK2 activation due to oxidative stress is likely to amplify the neuroinflammatory response, leading to damage within the nervous system.

The influence of LRRK2 on the autophagy-lysosome system merits significant attention. During activation, Rab29 facilitates the recruitment of LRRK2 to the lysosomal membrane, where it undergoes phosphorylation [58]. LRRK2 plays a crucial role in regulating the number, size, and function of lysosomes. Pathogenic mutations in LRRK2 are associated with an increase in lysosomal size and a reduction in lysosomal function. However, these lysosomal defects can be corrected through the application of LRRK2 inhibitors [32,59]. Dysfunction of lysosomes may lead to the accumulation of aberrant proteins that need to be cleared, triggering the onset of PD, as demonstrated in PD models [60]. In addition, LRRK2 is implicated in the autophagy process, contributing to the transport of autophagosomes as well as the degradation of lysosomal contents [61]. Mutant LRRK2 May cause a decrease in mitochondrial autophagy, leading to mitochondrial dysfunction [62].

2.4. Pathogenic mutations in LRRK2

Many mutants of LRRK2 have been reported to date, but only a few are regarded as truly pathogenic, including N1437H, R1441C/G/H, Y1699C, G2019S, and I2020T, all of which are located in the catalytic structural domain of the protein (Figure 1) [63]. Overexpression of these mutant proteins in primary neuronal cultures showed toxicity, resulting in neuronal synapse shortening, altered organelle function, and cell death [64].

The prevalence of various pathogenic LRRK2 mutations is genetically specific to certain populations and even families [65]. The most common LRRK2 pathogenic missense mutation is G2019S, which is present in over 85% of PD patients with LRRK2 mutations [66]. Approximately 6% of all familial PD patients and 2% of idiopathic PD are caused by G2019S mutations, which are more common in specific populations [67]. The association between G2019S-LRRK2 overexpression and neurodegeneration has been demonstrated in a variety of animal models [68,69]. The G2019S substitution is located in the activation loop of LRRK2, potentially stabilizing the active kinase conformation DYG-in, and enhancing kinase activity [70]. Compared to wide-type LRRK2 (WT-LRRK2), G2019S-LRRK2 exhibits increased kinase activity, with a reported elevation of 2–4-fold as indicated by the phosphorylation levels of the substrate Rab10 and the LRRK2 Ser1292 site [71,72]. Although the root mean square deviation (RMSD) of the kinase structural domains between WT and G2019S LRRK2 is only 0.4 Å, the kinase activities are vastly different (Figure 2). This observation suggests that the mutation does not cause a significant change in the conformation, and provides an opportunity for the design of conformationally specific kinase inhibitors [35,70].

The I2020T mutation is located within the activation loop of the KIN structural domain, adjacent to G2019S, which also generates highly active kinases by stabilizing the active conformation [73]. Pathogenic mutations located in the structural domains of ROC (N1437H, R1441G/C/H) and COR (Y1699C) increase the binding of the kinase to GTP, thereby enhancing kinase functionality [74]. It has been shown that these mutations also increase the levels of phosphorylation and autophosphorylation of LRRK2 substrates [71,72,75]. In addition, the G2385R mutation located in the WD40 structural domain is common in Asian populations and may increase the risk of PD [76], possibly due to its impact on the dimerization of the WD40 structural domain [77]. Biochemical analyses indicate that the G2385R mutation also enhances LRRK2 kinase activity [77].

Abnormally elevated LRRK2 kinase activity is implicated in various mechanisms associated with PD. Mutations leading to elevated LRRK2 kinase activity contribute to the pathogenesis of PD, thereby providing a rationale for the therapeutic use of LRRK2 inhibitors in the treatment of this condition [78]. The LRRK2 kinase inhibitor PF-06447475 has been shown to attenuate neuronal degeneration induced by α-synuclein in an animal model [79]; DNL-201 can regulate the lysosomal pathway in both in vivo and in vitro studies, completely normalizing defective lysosomal function and potentially restoring the ability to clear abnormal proteins, showing potential for treating PD [80,81].

PD induced by mutations exhibits symptomatological similarities to sporadic PD [82]. Furthermore, WT-LRRK2 is also activated in the pathogenesis of PD [54]. Consequently, LRRK2 is likely to be involved in all forms of PD, and pharmacological agents that inhibit the activity of the LRRK2 kinase may prove beneficial not only for patients with PD associated with pathogenic mutations.

As previously noted, LRRK2 has both GTPase and kinase activities, with its kinase activity playing a significant role in the pathophysiology of PD. The development of LRRK2 kinase inhibitors is currently regarded as one of the most attractive therapeutic strategies for PD. Numerous ATP-competitive LRRK2 inhibitors have been developed in recent years, and studies have shown that the inhibition of LRRK2’s kinase activity is neuroprotective [83]. The remainder of this article will summarize the current advancements in LRRK2 kinase inhibitors, to guide researchers interested in the development of such inhibitors.

3. LRRK2 inhibitors

Four small molecule compounds have entered clinical trials, DNL-201 and DNL-151, developed by Denali Therapeutics, and NEU-411 and NEU-723, developed by Neuron23 (Table 1). DNL-201 is a highly selective and central nervous system (CNS)-permeable LRRK2 inhibitor that has shown potential efficacy in inhibiting PD during Phase I and Phase Ib clinical trials [80]. DNL-151 is currently the only LRRK2 inhibitor that has advanced to Phase III clinical trials, exhibiting excellent CNS permeability and no serious adverse effects in most subjects [81]. Due to the superior pharmacokinetic profile of DNL-151, Denali announced that it would focus exclusively on this drug. In 2023, Neuron23 initiated clinical trials for NEU-411 and NEU-723; however, for unknown reasons, the Phase I clinical trial of NEU-723 was forced to be terminated.

Table 1.

LRRK2 inhibitors in clinical trials for PD treatment.

Compound	Company	Status	Clinical trial
DNL-151	Biogen/Denali Therapeutics Inc.	Phase III	NCT05418673
DNL-201	Denali Therapeutics Inc.	Phase I	NCT03710707
NEU-411	Neuron23 Inc.	Phase I	NCT05755191
NEU-723	Neuron23 Inc.	Phase I	NCT05633745

Inhibitors targeting LRRK2 can be broadly categorized into type I and type II based on the classification of protein kinase inhibitors. Type I inhibitors bind competitively to ATP within the ATP-binding pocket of the kinase’s structural domain, while type II inhibitors bind to the inactive conformation of the DYG-out of the LRRK2 kinase [84]. Most small molecules targeting LRRK2, including those that have entered clinical trials, are conventional type I inhibitors.

The development of novel LRRK2 inhibitor molecules began in 2011. Before this, only relatively nonspecific LRRK2 inhibitors had been reported, and these compounds exhibited low selectivity for LRRK2. The natural product staurosporine (1, Figure 3) and its analog K-252a (2, Figure 3) are representative examples, with the former reaching IC₅₀ values of 2.0 and 1.8 nM against LRRK2 (WT and G2019S mutant), respectively [85]. The oncology drug sunitinib (3, Figure 3), which targets tyrosine kinases, also has effective LRRK2 inhibition, with IC₅₀ values of 15 and 19 nM for WT and G2019S mutant LRRK2, respectively. However, at a concentration of 1 μM, sunitinib is poorly selective with significant inhibition of 12 out of 85 protein kinases [86]. In the following years, researchers used high-throughput screening techniques to identify many LRRK2 inhibitors with novel chemical backbones.

Figure 3.

The structure of LRRK2 inhibitors.

3.1. Pyrimidine containing LRRK2 inhibitors

In 2011, a screening of kinase inhibitor libraries identified two selective molecules: LRRK2-IN-1 (4, Figure 3) and CZC-25146 (5, Figure 3). LRRK2-IN-1 demonstrated inhibitory effects on both WT and G2019S mutant LRRK2 with IC₅₀ values of 13 nM and 6 nM, respectively, and exhibited high selectivity in binding assays involving over 440 kinases [87]. Another inhibitor, CZC-25146, exhibited IC₅₀ values of 4.76 and 6.87 nM against LRRK2 (WT and G2019S mutant) and showed inhibitory activity against only five out of a total of 184 protein kinase binding assays [88]. Despite their nanomolar-level inhibitory activity and good kinase selectivity, these two compounds are unable to cross the blood-brain barrier, limiting their utility in animal PD models. Consequently, they are primarily utilized as tool molecules for evaluating the inhibitory effects of LRRK2 in cellular systems.

To address the challenge of low penetration across the blood-brain barrier, researchers have reported many compounds featuring an aminopyrimidine backbone, inspired by the structures of 4 and 5. TAE684 (6, Figure 3), an anaplastic lymphoma kinase (ALK) inhibitor, showed similar inhibitory activity to 4, with IC₅₀ values (WT and G2019S) of 7.8 and 6.1 nM, respectively. 6 in mouse models exhibited excellent pharmacokinetic properties and significant brain exposure, however, its selectivity for kinases was poor and it was unable to effectively inhibit phosphorylation in the mouse brain [89].

HG-10-102-01 (7, Figure 3) was optimized based on the above compounds and showed IC₅₀ values of 20.3 and 3.2 nM against WT and G2019S LRRK2, respectively. 7 results from the minimization of the 2,4-diaminopyrimidine backbone of 4-6, informed by the predicted docking conformation between the homology model and the inhibitor molecules of LRRK2. 7 has a lower relative molecular mass, high kinase selectivity, and improved pharmacokinetic properties. It is the first compound with considerable potential to exert phosphorylation inhibition in the brain as well [90,91].

Based on the structure of 7, the introduction of ethylamino and trifluoromethyl groups at the 4 and 5 positions of the aminopyrimidine, respectively, along with the incorporation of a fluorine atom on the benzene ring, produced GNE-7915 (8, Figure 3). This compound exhibits an IC₅₀ value of 9 nM against phosphorylated LRRK2. In comparison to the other inhibitors, 8 demonstrates a desirable balance of activity, kinase selectivity, and pharmacokinetic properties both in vitro and in vivo. Furthermore, toxicity studies conducted in animal models have also contributed to the risk assessment associated with the development of similar LRRK2 inhibitors [92].

In the same year, based on 8, the substitution of the aniline moiety with an aminopyrazole was implemented to reduce the size of the compound and to eliminate the potential formation of an ortho-quinoneimine reactive metabolite. This modification resulted in the synthesis of compound 9 (Figure 3), with an intracellular potency (pLRRK2 IC₅₀) of 28 nM against LRRK2. Compared to 8, the solubility of compound 9 was improved while maintaining excellent pharmacological and pharmacokinetic properties (Figure 4(a)) [93].

Figure 4.

The modification strategy of LRRK2 inhibitors. (a) Compound 9. (b) Compounds 10 and 11.

To circumvent glucuronidation in vivo, GNE-0877 (DNL-201, 10, Figure 3) was obtained by combining aminopyrimidine and gem-disubstituted cyano pyrazole with 9 as the lead compound. Following homology modeling for binding prediction, it was suggested that the cyano group could form an electrostatic interaction with Arg1957, thereby enhancing the binding affinity. Intracellular assays revealed a significant increase in the inhibitory potency of 10 (pLRRK2 IC₅₀ = 3 nM) with no evidence of glucuronidation. 10 showed high selectivity in a selectivity analysis against 188 kinases and demonstrated a favorable pharmacokinetic profile in a rat model. However, 10 was identified as a reversible inhibitor of CYP1A2, with concomitant pan-inducibility toward other hepatocyte CYP enzymes (Figure 4(b)).

To address this property, a further optimized modification of 10 resulted in the synthesis of GNE-9605 (11, Figure 3). To reduce the inhibition of CYP1A2, a polar group was introduced further along the molecular structure while keeping the pyrazole ring unchanged. Additionally, to enhance impaired brain penetration, methylpyrazole was substituted with chloropyrazole. 11 exhibited strong inhibitory effects (pLRRK2 IC₅₀ = 18.7 nM), was not an inducer or inhibitor of any CYP isoforms, and demonstrated similar selectivity and pharmacokinetic properties to 10. Comparative studies in BAC mice and crab-eating monkey models were carried out to evaluate 10 and 11. Ultimately, 10 was chosen for further investigation due to its in vivo unbound brain IC₅₀ being 10-fold higher than that of 11, as well as its lack of significant self-induction. DNL-201 has undergone several clinical trials to date, which have amply demonstrated the potential of LRRK2 inhibitors in the treatment of PD (Figure 4(b)) [80,94].

In 2014, a pyrrolopyrimidine-based lead compound was obtained by high-throughput and virtual screening. Subsequent optimization of this compound led to the development of the inhibitor PF-06447475 (12, Figure 3), which exhibits IC₅₀ values of 3 and 11 nM against WT and G2019S LRRK2, respectively [95].

Based on the structure of 8, wherein the trifluoromethyl group forms an intramolecular hydrogen bond with the NH group, JH-II-127 (13, IC₅₀ = 6.6 and 2.2 nM for WT and G2019S LRRK2, Figure 3) featuring a thickened bicyclic ring was designed. Compared with 8, 13 demonstrated enhanced potency in inhibiting LRRK2 across various mutant forms. Docking studies revealed that the Met1947 interacts with 5-Cl, while the pyrrolopyrimidine ring forms three hydrogen bonds with Met1949 and Ala1950. These findings provide a reference for understanding the binding mode, which can inform the design of additional inhibitors [96].

A fragment-based screening revealed that purine or pyrazolo[3,4-d]pyrimidine fragments exhibit exceptionally high ligand efficiency. The study utilized checkpoint kinase (CHK1) and a 10-point mutant of CHK1 as a crystallographic surrogate for LRRK2, facilitating the binding of fragmented molecules and guiding the optimization of the compounds. This process discovered 14 and 15 (Figure 3), which exhibited high activity and kinase selectivity. In mouse brain tissue, 14 and 15 inhibited the phosphorylation of Ser935 with IC₅₀ values of 16 and 23 nM, respectively; neither inhibited other kinases in a broad-spectrum kinase analysis. Docking analyses showed that both compounds contain methyl groups occupying spatial regions that confer critical selectivity for LRRK2 (Ala2016, Leu1949), which may be pivotal for the high kinase selectivity observed in these inhibitors [97].

In 2021, computer predictions and QSAR modeling identified the potential of aminoquinazoline analogues as LRRK2 inhibitors. Structural optimization yielded 16 (Figure 3) which demonstrated potent activity, high selectivity, excellent oral bioavailability, and CNS penetration, indicating therapeutic value [98].

3.2. Arylbenzamide derivatives

Aminopyrimidines are not the only effective structural category. In 2012, an arylbenzamide-based inhibitor named GSK2578215A (17, Figure 3) was discovered. It had IC₅₀ values of 10.9 and 8.9 nM for WT and G2019S LRRK2, respectively, and showed significant selectivity in kinase assays. Unlike the first-generation LRRK2 inhibitors 4 and 5, 17 also exhibited favorable permeability across the blood-brain barrier. Unfortunately, similar to 6, 17 failed to exert phosphorylation inhibition despite considerable exposure within the brain [99].

Through the execution of structure-activity relationship (SAR) studies on 17, a structurally similar inhibitor, 18, was identified (Figure 3). This was the first compound in the series to demonstrate a high oral bioavailability (F = 91.6%). However, both 17 and 18 showed high tissue binding in the mouse brain (99.7% and 99.6%), and 18 displayed a low solubility (11 μM), which may account for its inability to produce an inhibitory effect within the brain [100].

To improve the physicochemical properties of the compounds, various substitutions on the structure of 18 were reevaluated, leading to the synthesis of compound 19 (Figure 3). The results of the SAR study indicated that the best choices for the N-aryl group were 3-pyridyl and 4-pyridazinyl, while the presence of a saturated ring resulted in decreased activity. The best substitutions for the phenyl ring were hydrogen or fluorine at the para-position. Additionally, the substituent at the 5-position of the benzamide can be replaced with different alkyl or heterocyclic groups, among which the N-connected groups can improve solubility, and the O-linked substituents will enhance human liver microsomal clearance. The morpholine analog 19 showed a good balance of activity, physicochemical properties, and metabolic stability, with an improved free unbound fraction in the brain (4.6%) and significant inhibition of LRRK2 phosphorylation in rat brains [101].

Based on the triazopyridine and diaminopyrimidine obtained by high-throughput screening, a series of arylbenzamide-based inhibitors containing a benzothiazole moiety were designed. Pharmacophore fusion generated the lead compound 20 (Figure 3), which showed IC₅₀ values of 0.7 and 0.37 μM against WT and G2019S LRRK2, respectively. SAR analyses discovered compound 21 (Figure 3), which demonstrated slightly greater potency than 20. Docking studies showed that the nitrogen and oxygen atoms of the morpholino group in the benzothiazole are important for activity and that the halogen atom at the 6-position performs a certain function. The inhibition of LRRK2 by 20 and 21 promoted the proliferation of neural stem cells and neural progenitor cells, exhibiting neurogenesis. This series of experiments suggests that LRRK2 inhibitors have the potential to serve as multifunctional drugs (Figure 5) [102].

Figure 5.

Compounds 20 and 21 resulted from pharmacophores fusion and SAR study.

In 2021, a high-throughput screening identified a thiazole-4-carboxamide derivative with potential. Due to its poor kinase selectivity, the thiazole was replaced with a pyridine ring, which was structurally optimized to develop 22 (Figure 3). Compound 22 had picomolar potency and improved selectivity in rats, exhibiting low clearance and high mean residence time (MRT) in both rats and rhesus monkeys. Unfortunately, 22 failed multiple times in the in vitro rat P-glycoprotein (P-gp) efflux transporter assay [103].

3.3. Indazole derivatives

In 2017, Merck identified 3,5-disubstituted indazole lead compounds through high-throughput screening utilizing WT LRRK2. SAR studies of the substituents at 3- and 5-positions of this series of compounds led to the discovery of MLi-2 (23, Figure 3), an inhibitor with high potency (LRRK2 IC₅₀ = 0.76 nM), high brain permeability, and kinase selectivity which acted in a dose-dependent manner. In a rat model, 23 exhibited good bioavailability (F = 39 ± 18%), moderate plasma exposure (Cmax = 0.48 ± 0.36 μM), and moderate half-life (MRT = 4.0 ± 1.0 h). However, 23 induced enlargement of type II pneumocytes in some mice [104,105].

LifeArc’s high-throughput screening discovered a lead compound based on 5-azaindazole. SAR studies of substituents at positions 3 and 4 resulted in 24 and 25 (Figure 3). Docking studies showed that the azaindazole rings in this series of compounds form two hydrogen-bonded interactions with Glu1948 and Ala1950, which may contribute to the activity of the homologous series of inhibitors [106,107].

In 2020, the G2019S-LRRK2 selective inhibitor EB-42486 (26, Figure 3), which contains an indazole structure, was obtained through structural modification of a lead compound identified from high-throughput screening. The gem-dimethyl substitution on the dihydropyrimidine significantly enhanced the inhibitor’s selectivity for the G2019S mutant. 26 exhibited IC₅₀ values of 6.6 and <0.2 nM against WT and G2019S-LRRK2, respectively, with a 33-fold selectivity for G2019S compared to WT LRRK2. Furthermore, the selectivity for intracellular inhibition of Ser935 phosphorylation exceeded 300-fold, offering an opportunity for the development of selective inhibitors and inspiring further research ideas [108].

Compound 27 (Figure 3), which also contains an indazole structure, stood out in the development of selective inhibitors. The compound was optimized using structural guidance for the docking of mutant CHK-1 with the inhibitor. Its indole core forms hydrogen bonds with Glu1948 and Ala1950, and the benzonitrile moiety disrupts the hydrogen bonding network established by Ser2019, thereby enhancing selectivity for the mutant. 27 has excellent ADME (Absorption, Distribution, Metabolism, and Excretion) properties and pharmacokinetic characteristics, with IC₅₀ values of 14 and <1 nM against WT and G2019S-LRRK2, respectively, demonstrating up to 2500-fold selectivity in cellular assays. Studies related to 27 have provided ideas for the development of selective inhibitors targeting G2019S LRRK2 [109].

In 2022, a series of structural optimizations targeting 1-heteroaryl-1 h-indazole generated 28 (Figure 3). It exhibits an IC₅₀ of 0.9 nM against LRRK2, along with good kinase selectivity, brain penetration, and pharmacokinetic properties. Unfortunately, certain observed issues led to the discontinuation of the study before 28 could enter the preclinical candidate study [110].

3.4. Indolinone derivatives

Using the broad-spectrum inhibitor 3 as a starting compound, modifications were made to improve kinase selectivity, resulting in the indolinone-centered compounds 29 and 30 (Figure 3). Compared to 3, both 29 and 30 demonstrated improved activity (LRRK2 IC₅₀ = 9 and 4 nM) and greatly improved kinase selectivity. Additionally, 29 and 30 showed good blood and brain exposure in mice, along with high oral bioavailability. This series of compounds serves as a reference for the development of highly selective LRRK2 inhibitors derived from broad-spectrum inhibitors [111].

Compound 31 (Figure 3), which is derived from the indolinone structure as its core, also showed high activity, with IC₅₀ values of 15 and 10 nM against WT and G2019S LRRK2, respectively. Docking studies revealed that the indolinone core forms multiple hydrogen bonds with the LRRK2 backbone, while the phenol and bromine atoms interact hydrophobically with the protein, and the chlorine is located in a solvent-exposed region. SAR studies suggest that phenol plays a key role in the inhibitory activity of this series of compounds [112].

Furthermore, compounds developed from indirubin derivatives, inspired by traditional Chinese medicine, have also demonstrated inhibitory activity, underscoring the potential of the indolinone scaffold in LRRK2 inhibitors [113].

3.5. Quinoline and isoquinoline derivatives

In 2013, a screening study focused on kinase inhibitors identified quinoline amide structures as LRRK2 inhibitors for the first time. To enhance CNS permeability and reduce P-gp mediated efflux, various modifications were made, resulting in the synthesis of 32 (Figure 3), which has an IC₅₀ of 7 nM [114].

In 2023, the amide-isoquinoline MK-1468 (33, Figure 3) was designed and synthesized based on the structures of the aminoquinazoline lead compound 16 and the 1-heteroaryl-1 h-indazole lead compound 28. 33 exhibited high activity (IC₅₀ = 0.4 nM) and significant kinase selectivity against LRRK2, along with favorable pharmacokinetic properties. Furthermore, 33 demonstrated efficacy in both rats and rhesus monkeys, successfully passing a series of pharmacological and safety evaluations, leading to its status as a preclinical candidate [115].

3.6. Pyrazole or triazole derivatives

High-throughput screening has yielded a series of triazole-pyridazine inhibitors, among which 34 (Figure 3) demonstrated high levels of inhibition in biochemical assays, with IC₅₀ values of 32 and 6 nM against WT and G2019S LRRK2, respectively. However, due to poor metabolic stability, further development of this series of compounds has been suspended [116]. Compound 35 (Figure 3), which shares the same backbone, exhibits IC₅₀ values of 64 and 32 nM against WT and G2019S LRRK2, respectively. However, 35 is unable to inhibit phosphorylation in the brain, significantly limiting its potency and therapeutic potential [117].

In 2022, compound 36 (Figure 3), designed based on 27 with 1 h-pyrazole as the core backbone, showed excellent activity with IC₅₀ values of 17.1 and 2.4 nM (WT and G2019S LRRK2), respectively, and about 7.1-fold inhibitory activity against the mutant. However, it showed insufficient brain penetrance [118]. In the same year, 37 (Figure 3), optimized on the base of 36, also displayed selectivity for G2019S-LRRK2, with IC₅₀ values of 49 and 4.6 nM, respectively. 37 had significantly improved ADME and physicochemical properties, enabling it to enter the rodent brain [119]. To date, no G2019S-LRRK2 inhibitor molecules have entered clinical trials.

4. Conclusions

PD is one of the most common neurodegenerative diseases, and there is an urgent clinical need for completely new drugs that can delay the onset of the disease, rather than current treatments that focus on symptom management [120]. Since 2004, when it was reported that mutations in LRRK2 are linked to PD, the understanding of the relationship between LRRK2 and the pathological mechanisms underlying PD has deepened. Evidence suggests that elevated LRRK2 kinase activity is closely associated with the progression of PD and may lead to neuronal cell death through various mechanisms, contributing to PD pathogenesis.

For this reason, several drugs have been developed to inhibit LRRK2 kinase activity, with most of them classified as type I inhibitors. However, due to limitations such as kinase selectivity, brain penetration, and unbound concentration in the brain, only a limited number of these drugs have successfully entered clinical trials. Among them, DNL-201 and DNL-151 successfully reduced the levels of lysosomal biomarkers in clinical subjects, which can be elevated due to lysosomal dysfunction caused by LRRK2 mutations [80,81]. The results of this clinical trial support the hypothesis that LRRK2 inhibitors can correct lysosomal dysfunction in PD patients, thereby addressing a key mechanism in PD pathology.

LRRK2 is expressed in various tissues, including the brain, kidney, and lung, and plays a crucial role in normal physiological functions. LRRK2 knockout rats displayed abnormal kidney, liver, and lung phenotypes [121], raising concerns about the potential toxicity of LRRK2 inhibitors. Accumulation of vacuolar-type II lung cells was also observed in pharmacological studies involving 8 and 10 [122]. Although this accumulation did not appear to significantly impact the health of the experimental animals, the prevalence of respiratory problems in PD patients suggests that the pulmonary toxicity of the drug could exacerbate the patient’s condition [123]. However, a study has shown that the effects produced by the inhibitor on lung cells are reversible and do not cause permanent loss of function [124]. Therefore, when designing LRRK2 inhibitors, it is essential to assess safety in animal models.

5. Future perspective

In summary, LRRK2 is strongly associated with PD, and elevated kinase activity may contribute to PD pathogenesis. After nearly two decades of research, significant progress has been made in the development of LRRK2 inhibitors, with several promising candidates advancing into clinical trials. All of these molecules are orally administered small-molecule inhibitors, featuring high brain permeability and kinase selectivity, enabling them to precisely target LRRK2. In contrast to current PD therapeutic options, LRRK2 inhibitors have exhibited neuroprotective properties and alleviated lysosomal deficiencies in patients, demonstrating their potential to decelerate the progression of early-stage PD.

However, preclinical investigations, including LRRK2 inhibitors such as DNL-201, have revealed that LRRK2 inhibitors can induce lung abnormalities in animals [121]. Given that PD treatments are primarily administered to an aging population with compromised organ function who require long-term medication, the safety standards for such drugs are extremely high. Consequently, enhancing the brain permeability of these molecules during development is crucial to minimizing potential peripheral adverse effects. Furthermore, LRRK2 inhibitors should be under thorough scrutiny to ensure their safety during long-term administration. Since the involvement of LRRK2 May be limited to the early stage of PD, rigorous testing of LRRK2 inhibitors is necessary to determine the precise timing for treatment with these molecules to achieve the best therapeutic effects.

Throughout the development of existing LRRK2 inhibitors, insufficient CNS penetration has been a common challenge encountered in the development of many inhibitors. Many molecules with excellent activity and high selectivity were discontinued precisely due to inadequate brain permeability. LRRK2, a large protein with intricate biological properties, had its detailed structure reported only in 2021 [35]. , making the development of efficient, selective, and permeable inhibitory molecules difficult in the past. However, with the disclosure of LRRK2’s crystal structure, the optimization of LRRK2 inhibitor molecules can be accelerated by using computer simulation to find sites that increase brain permeability without affecting activity.

The elucidation of the crystal structures of LRRK2 bound to inhibitors provides key clues for rational structure-based drug design [39]. Analysis of the binding modes of type I inhibitors reveals the key hydrogen bonds responsible for inhibitor and kinase binding, offering ideas for rational design and optimization of inhibitors targeting LRRK2. Additionally, since current research on LRRK2 inhibitors has focused on type I inhibitors that competitively bind to ATP, approaches aimed at reducing LRRK2 function through other mechanisms have the potential to serve as successful alternatives, which may even circumvent the lung side effects associated with kinase inhibition [125]. It has been reported that the utilization of antisense oligonucleotides (ASOs) to decrease LRRK2 expression can effectively inhibit the accumulation of α-synuclein in mouse neurons [126]. The application of type II LRRK2 allosteric inhibitors promotes the LRRK2 ubiquitination and the LB-like inclusion formation, implementing a protective strategy to sequester and degrade toxic proteins [127]. Experimental results demonstrate that these novel inhibitory strategies also possess considerable therapeutic potential. Given that LRRK2 is involved in many pathological processes of Parkinson’s disease (PD), multi-target inhibition may also be highly effective.

This paper summarizes the discovery process of inhibitor molecules with different backbone structures, aiming to aid researchers in designing LRRK2-targeted molecules for PD treatment. In the future, we anticipate an increased possibility of developing inhibitors with superior selectivity, enhanced brain penetration, and greater therapeutic efficacy, leveraging insights gained from current inhibitor designs. With dedicated efforts, it is anticipated that novel therapeutic agents for PD will soon become available.

Funding Statement

This study was financially supported by the National Natural Science Foundation of China (grant Nos. 82473781 and 82173652) and the Natural Science Foundation of Jiangsu Province (grant Nos. BK20221522). Support from Jiangsu “333 high-Level Talents Cultivation” Leading Talents (2022-3-16-203), the Qing Lan project is also appreciated.

Article highlights

Introduction

• Parkinson’s disease is a common neurodegenerative disease worldwide, but current treatments can only relieve symptoms.

• LRRK2 has been identified as a common genetic cause of PD, making it a promising therapeutic target.

Association of PD with LRRK2

• Describing the structure and physiological functions of LRRK2.

• Highlighting the role of LRRK2 in the pathogenesis of PD, including some pathogenic mutations that cause increased LRRK2 activity.

Discovery and optimization of LRRK2 inhibitors

• Briefly describing the discovery of small molecule inhibitors targeting LRRK2, and the design strategies for optimizing their selectivity and brain penetration.

• Updating the research progress on LRRK2 inhibitors against PD and introducing LRRK2 inhibitors with different scaffolds.

Conclusion and prospects

• Emphasizing the importance of evaluating the safety of LRRK2 inhibitors during drug development.

• Discussing the potential of LRRK2 inhibitors as therapeutic drugs for PD and possible future directions.

Author contributions

X Bai summarized the literature, drafted the manuscript, and drew the figures; J Zhu was responsible for proofreading the words and figures, and revising the manuscript; H Sun and Y Chen were responsible for conceiving, revising, and approving the review to be submitted to the journal

Writing disclosure

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

Disclosure statement

The authors have no competing interests or relevant affiliations with any organization or entity 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.