Impact of Type II LRRK2 inhibitors on signaling and mitophagy

Much effort has been devoted to the development of selective inhibitors of the LRRK2 as a potential treatment for LRRK2 driven Parkinson's disease. In this study, we first compare the properties of Type I (GSK3357679A and MLi-2) and Type II (GZD-824, Rebastinib and Ponatinib) kinase inhibitors that bind to the closed or open conformations of the LRRK2 kinase domain, respectively. We show that Type I and Type II inhibitors suppress phosphorylation of Rab10 and Rab12, key physiological substrates of LRRK2 and also promote mitophagy, a process suppressed by LRRK2. Type II inhibitors also display higher potency towards wild-type LRRK2 compared with pathogenic mutants. Unexpectedly, we find that Type II inhibitors, in contrast with Type I compounds, fail to induce dephosphorylation of a set of well-studied LRRK2 biomarker phosphorylation sites at the N-terminal region of LRRK2, including Ser935. These findings emphasize that the biomarker phosphorylation sites on LRRK2 are likely reporting on the open vs closed conformation of LRRK2 kinase and that only inhibitors which stabilize the closed conformation induce dephosphorylation of these biomarker sites. Finally, we demonstrate that the LRRK2[A2016T] mutant which is resistant to MLi-2 Type 1 inhibitor, also induces resistance to GZD-824 and Rebastinib suggesting this mutation could be exploited to distinguish off target effects of Type II inhibitors. Our observations provide a framework of knowledge to aid with the development of more selective Type II LRRK2 inhibitors.


Introduction
Autosomal dominant mutations that activate the LRRK2 (leucine rich repeat protein kinase-2) signaling pathway cause late-onset Parkinson's disease [1,2]. LRRK2 is a large 2527 residue multidomain enzyme bearing an armadillo, ankyrin, leucine-rich repeats followed by a tandem Roco type GTPase consisting of a Roc and Cor domain, a Ser/Thr kinase domain and a C-terminal WD40 domain [3]. Various pathogenic mutations have been well-characterized that lie within the LRRK2 Roc (N1437H, R1441G/C/H), Cor (Y1699C), or kinase (G2019S, I2020T) domain [4]. The G2019S mutation is located within the conserved Mg 2+ subdomain VII motif of the kinase domain and represents the most frequently observed mutation [5]. LRRK2 phosphorylates a subset of Rab GTPases, including Rab8A, Rab10 and Rab12 at a conserved Ser/Thr residue (Thr72 for Rab8A, Thr73 for Rab10, Ser106 for pRab12) [6,7]. The LRRK2 phosphorylation site is located within the Rab effector-binding Switch-II motif. LRRK2 phosphorylated Rab proteins interact with new set of phospho-binding effectors such as RILPL1 and RILPL2 [6,8].
All LRRK2 pathogenic mutations increase phosphorylation of Rab substrates including Rab10 [6,7,9]. These mutations also stimulate LRRK2 autophosphorylation at Ser1292 [10], but detection of this site is challenging especially for endogenous LRRK2, due to its low stoichiometry of phosphorylation. LRRK2 is also phosphorylated at several well studied serine residues that are termed biomarker sites including Ser910, Ser935, Ser955 and Ser973, located between the ankyrin and leucine rich repeats [11,12]. Phosphorylation of Ser910 and Ser935 promotes interaction with 14-3-3 scaffolding proteins, preventing LRRK2 from assembling into inclusion like bodies in the cytosol [11,13]. These biomarker sites have received much attention as they become dephosphorylated following pharmacological inhibition of LRRK2 [13]. Dephosphorylation of Ser935 in particular, has been widely exploited to assess in vivo efficacy of LRRK2 inhibitors [14][15][16]. It is unclear whether LRRK2 Ser910/Ser935 phosphorylation is regulated by autophosphorylation or via upstream kinases. Phosphorylation of the biomarker sites do not correlate with intrinsic LRRK2 kinase activity as mutation of these residues to Ala do not impact basal LRRK2 kinase activity [11,17]. Furthermore, various pathogenic mutations including R1441C/G, Y1699C and I2020T mutations suppress the phosphorylation of Ser910 and Ser935 through unknown mechanisms [11,13].
Evidence points towards LRRK2 controlling mitochondrial [18,19] and lysosomal [20][21][22][23][24] homeostasis. Recent work has revealed that LRRK2 inhibits basal mitophagy in a variety of cells including dopaminergic neurons and that LRRK2 inhibitors promote mitophagy in LRRK2[G2019S] knock-in mice [25]. LRRK2 has also been shown to control repair of damaged endomembranes through its ability to phosphorylate Rab8A [26,27]. Treatment of cells with agents that stress lysosomes, promote Rab protein phosphorylation by LRRK2 [28,29]. Pathogenic LRRK2 also inhibits ciliogenesis in cholinergic interneurons through LRRK2 phosphorylated Rab10 forming a complex with RILPL1 [30,31]. Cilia loss in cholinergic interneurons induced by LRRK2 pathogenic mutations may decrease the ability of these cells to sense Sonic hedgehog in a neuro-protective circuit that supports dopaminergic neurons [30].
There are two widely studied classes of kinase inhibitors namely, Type I that bind to the kinase domain in a closed active conformation and Type II that maintain the kinase in an open inactive-conformation [32]. Most kinase inhibitors are Type I compounds that bind to the ATP binding pocket in a conformation in which the critical catalytic residues including the DFGψ motif (DYGI in LRRK2) are oriented in an 'in' conformation with the α-helix and the core hydrophobic spine all aligned in an fully active conformation [33,34]. Type II inhibitors, include approved kinase compounds such as imatinib [35] and sorafenib [36]. These compounds exploit an additional binding site located adjacent to the region occupied by ATP when the kinase is in the inactive conformation. In this inactive conformation, the DFGψ motif is oriented in an 'out 'conformation pointing away from the α-helix and in addition the core hydrophobic spine residues are not aligned with each other [32,37].
R1441G and I2020T pathogenic mutations as well as treatment with Type I LRRK2 kinase inhibitors promote binding of LRRK2 to microtubule filaments [34,38,39]. Tomography and cryo-EM analysis suggest that pathogenic mutations induce the LRRK2 kinase domain to adopt a closed active conformation that becomes intrinsically capable of binding microtubules in a well-ordered and periodic manner [40,41]. Furthermore, recruitment of LRRK2 to microtubules inhibits kinesin and dynein microtubule motility in biochemical studies [41]. High resolution cryoEM studies have recently described the inactive structures of full length [42] as well as the catalytic moiety [43] of LRRK2. Comparison of the active LRRK2 structure predicted from in situ Cryo-electron tomography analysis of LRRK2 associated with microtubule filaments [40] as well as AlphaFold artificial intelligence predicted structure of LRRK2 [44], indicates that the active conformation of LRRK2 is markedly distinct from that of inactive LRRK2. Thus, Type I and Type II kinase inhibitors that trap LRRK2 in the inactive vs active conformation may have distinct physiological effects.
In this study, we investigate the cellular and in vitro properties of Type II LRRK2 inhibitors (GZD-824, Rebastinib and Ponatinib) ( Figure 1). We find that Type II inhibitors similar to Type I compounds, suppress LRRK2 mediated phosphorylation of Rab proteins and stimulate basal mitophagy. Unexpectedly, and in contrast with Type I inhibitors, GZD-824, Rebastinib as well as Ponatinib, do not induce the dephosphorylation of LRRK2 at Ser935 and other nearby residues. We also report that Type II inhibitors suppress wild-type LRRK2 more potently than pathogenic forms of this enzyme. These observations provide a framework of knowledge that will help with the development and characterization of more selective Type II LRRK2 inhibitors as tools to dissect LRRK2 signaling functions. (#ATCC CRL-1573) and HeLa (#ATCC CCL-2) cells were purchased from American Tissue Culture Collection (ATCC) and cultured in DMEM containing 10% (by vol) FBS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin. All cells were grown at 37°C, 5% (by vol) CO 2 in a humidified atmosphere and regularly tested for mycoplasma contamination.

Inhibitor treatments
Inhibitors were dissolved in dimethyl sulfoxide (DMSO) (Sigma, #D2650). The inhibitors were diluted 1000× into cell medium. The equivalent volume of DMSO was added to the minus inhibitor controls. Experiments were performed in six-well plates, 10 cm dishes or 15 cm dishes depending on the amount of cell extract required.

Transient transfections
Transient transfections were performed for 24 h using the linear polyethylenimine (PEI, Polyscience, #24765) method [52] with a final concentration of 2 mg/ml PEI and 6 μg of DNA for a 10 cm diameter dish with 10 ml of medium.

IC 50 LRRK2 Nictide assay
Reactions were undertaken with wild-type Flag-LRRK2 (#A15198) and Flag-LRRK2[G2019S] (#A15201) purchased from ThermoFisher Scientific. Peptide kinase assays, performed in triplicate, were set up in a total volume of 25.5 ml containing 13.7 ng LRRK2 kinase in 50 mM Tris-HCl, pH 7.5, 0.1 mM EGTA, 10 mM MgCl 2 , 20 mM Nictide (an optimized peptide substrate for LRRK2 [53]), 0.1 mM [γ-33 P]ATP (∼500 cpm/ pmol) and the indicated concentrations of inhibitor dissolved in DMSO. After incubation for 30 min at room temperature, reactions were terminated by harvesting the whole reaction volume onto P81 phosphocellulose filter plate and immersion in 50 mM phosphoric acid. Samples were washed extensively and the incorporation of [γ-33 P]ATP into Nictide was quantified by Cerenkov counting using a Tri-Carb 4910 TR PerkinElmer liquid scintillation analyzer. IC 50 values were calculated with GraphPad Prism using non-linear regression analysis.

Kinase profiling
Protein kinase profiling of the Type I (GSK3357679A and MLi-2) and Type II inhibitors (GZD-824 and Ponatinib) were undertaken at a concentration of 0.1 and 1 mM and carried out against the Dundee panel of 140 protein kinases at the International Centre for Protein Kinase Profiling (http://www.kinase-screen.mrc.ac. uk). The kinase profiling of each inhibitor was undertaken at a concentration of 1 and/or 10 mM. Results for each kinase are presented as the mean kinase activity ±S.D. for an assay undertaken in duplicate relative to a control kinase assay in which the inhibitor was omitted. Abbreviations and assay conditions used for each kinase are available (http://www.kinase-screen.mrc.ac.uk/services/premier-screen).

Quantitative immunoblot analysis
Cell extracts were mixed with a quarter of a volume of NuPage LDS Sample buffer 4× Invitrogen and heated at 95°C for 5 min. An amount of 10-30 μg of samples was loaded onto NuPAGE 4-12% Bis-Tris Midi Gel (Thermo Fisher Scientific, Cat# WG1403BOX) and electrophoresed with the NuPAGE MOPS SDS running buffer (Thermo Fisher Scientific, Cat# NP0001-02) at 90 V. After 30 min the voltage was increased at 120 V for 1 h 30 min. At the end of electrophoresis, proteins were electrophoretically transferred onto the nitrocellulose membrane (GE Healthcare, Amersham Protran Supported 0.45 mm NC) at 100 V for 90 min on ice in the transfer buffer (48 mM Tris-HCl and 39 mM glycine supplemented with 20% methanol). Transferred membrane was blocked with 5% (w/v) skim milk powder dissolved in TBS-T [20 mM Tris-HCl, pH 7.5, 150 mM NaCl and 0.1% (v/v) Tween 20] at room temperature for 1 h. The membrane was typically cropped into three pieces, namely the 'top piece' (from the top of the membrane to 75 kDa), the 'middle piece' (between 75 and 30 kDa) and the 'bottom piece' (from 30 kDa to the bottom of the membrane). All primary antibodies were used at 1 μg/ml final concentration and incubated in TBS-T containing 5% (by mass) bovine serum albumin with exception of α-tubulin and GAPDH antibodies that were diluted 1 : 5000 and 1 : 2000, respectively. The top piece was incubated with the indicated rabbit anti-LRRK2 phosphorylation site antibody multiplexed with mouse anti-LRRK2 C-terminus total antibody. The middle piece was incubated with mouse anti-GAPDH or anti-α-tubulin antibody. The bottom pieces were incubated with indicated total and phospho-Rab antibodies. Membranes were incubated in primary antibody overnight at 4°C. Prior to secondary antibody incubation, membranes were washed three times with TBS-T for 10 min each. The top and bottom pieces were incubated with goat anti-mouse IRDye 680LT (#926-68020) secondary antibody multiplexed with goat anti-rabbit IRDye 800CW (#926-32211) secondary antibody diluted in TBS-T (1 : 25 000 dilution) for 1 h at room temperature. The middle piece was incubated with goat anti-mouse IRDye 800CW (#926-32210) secondary antibody diluted in TBS-T (1 : 25 000 dilution) at room temperature for 1 h. Membranes were washed with TBS-T for three times with a 10 min incubation for each wash. Protein bands were acquired via near infrared fluorescent detection using the Odyssey CLx imaging system and quantified using the Image Studio software.

Analysis of 14-3-3 binding
This was undertaken as described previously [11]. Briefly, wild-type FLAG-LRRK2 was transiently expressed in HEK293 cells with HA-Rab29. Twenty-four hours post-transfection cells were lysed with lysis buffer and snap frozen. LRRK2 was immunoprecipitated from 4 mg of lysate using 20 μl anti-FLAG-M2 agarose beads (Sigma, #A2220) for 2 h in rotation at +4°C. Immunoprecipitates were washed once with lysis buffer supplemented with 0.5% (by vol) Nonidet P-40 (instead of Triton X-100) and 0.5 M NaCl followed by two washes. Samples were eluted in 2× LDS sample buffer (20 ml per 10 ml of resin) and heated at 70°C for 10 min. After centrifugation through 0.22 mm Spin X filter, the eluted protein was subjected to immunoblotting with the indicated antibodies. An amount of 10 μl of the final reaction mixture was needed for the detection of binding of endogenous 14-3-3 and 1 ml for detection of LRRK2.

Immunofluorescence microscopy
HeLa cells were seeded on coverslips and transfected with the plasmid reported in each image 24 h before fixation. Following the treatments with the inhibitors described in figure legends, cells were washed twice in DPBS and fixed with 3.7% Paraformaldehyde pH 7.3 (Agar scientific, #R1026), 200 mM HEPES, for 20 min. Cells were then washed twice with DMEM, 20 mM HEPES, and then incubated for 10 min with DMEM, 10 mM HEPES. After two washes in DPBS, fixed cells were permeabilized with 0.2% (by vol) NP-40 in DPBS for 10 min. Cells were blocked using 1% (by mass) BSA in DPBS, then incubated for 1 h with primary antibodies diluted in 1% (by mass) BSA in DPBS, washed three times in 0.2% (by mass) BSA in DPBS, and incubated for 1 h with secondary antibodies. The coverslips were washed three times with 0.2% (by mass) BSA in PBS. Coverslips were washed once more in double distilled water and mounted on slides (VWR, #631-0909) using ProLong™ Diamond Antifade Mountant (ThermoFisher, #P36961). For localization of the Golgi compartment, Golgin-97 was stained with rabbit anti-GM97 (CST, #D8P2K) and visualized with goat anti-rabbit conjugated to Alexa Fluor 405 (Life Technologies, #P10994). HA-Rab29 was stained with mouse anti-HA (Abcam, clone HA.C5 #ab18181) and visualized with goat anti-mouse conjugated to Alexa Fluor 594 (Life Technologies, #A-11032). The images were collected on an LSM710 laser scanning confocal microscope (Carl Zeiss) using the Å∼63 Plan-Apochromat objective (NA 1.4).

Human neutrophils isolation
Neutrophils were isolated from 20 ml of whole blood collected from two healthy donors using an immunemagnetic negative isolation described previously [54]. Five million cells from each condition were re-suspended in 1 ml of RPMI medium at and treated the indicated inhibitors or DMSO as a control for 45 min at room temperature. Cells were pelleted by centrifugation at 335×g for 5 min, the supernatants decanted, and the pellets lysed in lysis buffer freshly supplemented with 0.5 mM diisopropylfluorophosphate to inhibit high levels of intrinsic serine protease activity present in neutrophils [54]. Neutrophil lysates were incubated on ice for 30 min, centrifuged at 20 800×g for 20 min at 4°C, snap-frozen in liquid nitrogen and stored at −80°C. The human biological samples were sourced ethically, and their research use was in accord with the terms of the informed consents under an IRB/EC approved protocol.

Type II inhibitors target wild-type LRRK2 with greater potency than pathogenic LRRK2[G2019S] in vitro
We first assessed the in vitro potency of GZD-824, Rebastinib and Ponatinib towards full length wild type and G2019S pathogenic LRRK2 mutant in a biochemical assay using the synthetic peptide substrate termed Nictide [53] at an ATP concentration of 100 μM. This revealed that GZD-824 inhibited wild-type LRRK2 and LRRK2[G2019S] with an IC 50 Table 1). The two Type I LRRK2 kinase inhibitors used in this study, namely GSK3357679A and MLi-2, inhibited LRRK2 with sub-nanomolar potency with MLi-2 inhibiting LRRK2[G2019S] ∼2-fold more strongly than wild-type LRRK2. GSK3357679A displayed similar potency towards wild-type LRRK2 and LRRK2[G2019S] ( Table 1).
We also assessed the selectivity of GZD-824 (Supplemental Figure S1A), Rebastinib (Supplemental Figure S1B) and Ponatinib (Supplemental Figure S1C) towards a panel of 140 protein kinases. This revealed that all Type II inhibitors were not selective for LRRK2 and inhibited other serine/threonine as well as tyrosine kinases. At 0.1 and 1 μM, GZD-824 suppressed the activity of 22 and 43 of the 140 kinases assayed by over 80%, respectively (Supplemental Figure S1A). An amount of 1 μM Rebastinib suppressed 18 out of 140 kinases in the panel by more than 80%. An amount of 0.1 and 1 μM Ponatinib suppressed 14 and 20 kinases respectively by 80%. In contrast, GSK3357679A (Supplemental Figure S1D) and MLi-2 [16] are highly selective kinase inhibitors. Type II LRRK2 inhibitors suppress LRRK2 mediated Rab10 phosphorylation without suppressing Ser935 phosphorylation We next assessed the ability of GZD-824 (Figure 2A), and Ponatinib ( Figure 2B) to inhibit LRRK2 mediated phosphorylation of Rab10 at Thr73 in LRRK2 wild-type mouse embryonic fibroblast cells (MEFs). Cells were pre-incubated for 2 h with the indicated doses of GZD-824 or DMSO as a control and phosphorylation of Rab10 (Thr73), Rab12 (Ser105) and LRRK2 at Ser935 monitored by quantitative phospho-immunoblotting. GZD-824 induced a dose dependent inhibition of Rab10 and Rab12 phosphorylation with an IC 50 of ∼300 nM (Figure 2A). An amount of 1 μM GZD-824 reduced Rab10 and Rab12 phosphorylation to almost background levels like that observed with the MLi-2 Type I inhibitor [16]. Strikingly, GZD-824 did not induce dephosphorylation of LRRK2 at Ser935, under conditions which MLi-2 induced complete dephosphorylation of this site (Figure 2A). Increasing GZD-824 concentrations to 10 μM did not reduce LRRK2 phosphorylation at Ser935 significantly (Supplemental Figure S2A). We observed that concentrations of 10 μM or higher GZD-824 impacted cell morphology and attachment, so we employed GZD-824 at 5 mM or less in subsequent studies. Ponatinib was a weaker inhibitor, and only partially reduced Rab10 phosphorylation ∼50% inhibition at 10 μM, also without affecting Ser935 phosphorylation ( Figure 2B). Concentrations of Ponatinib above 10 μM were also observed to impact cell morphology and attachment. GZD-824 suppressed Rab10 and Rab12 phosphorylation in a time dependent manner, with a 50% reduction observed within ∼10 min ( Figure 2C). Rab10 and Rab12 phosphorylation remained suppressed and Ser935 phosphorylation unaffected over a 24 h time course of 1 μM GZD-824 treatment ( Figure 2D). GZD-824 also inhibited Rab10 phosphorylation in a dose dependent manner in primary human neutrophils isolated from two donors with maximal inhibition observed at 300 nM GZD-824 ( Figure 3). Phosphorylation of Ser935 as well as phosphorylation of other biomarker sites Ser910, Ser955 and Ser973 was unaffected by concentration of GZD-824 up to 3 μM (Figure 3).

GZD-824 and Rebastinib inhibit pathogenic LRRK2[R1441C] and LRRK2 [G2019S] with lower potency than wild-type LRRK2
We next compared the potency at which GZD-824 suppressed Rab10 phosphorylation in wild type and LRRK2 [R1441C] ( Figure 4A) as well as LRRK2[G2019S] knock-in MEFs ( Figure 4B). In these experiments, we observed that higher doses of GZD-824 were required to reduce phosphorylation of Rab10 in the pathogenic mutant cells. We estimated that the IC 50 Figure 4C,D). As reported previously [11], the G2019S mutation did not reduce Ser935 phosphorylation and treatment of these cells with GZD-824 did not impact Ser935 phosphorylation ( Figure 4B). We also compared the ability of Rebastinib to suppress the activity of Rab10 phosphorylation in wild type and LRRK[G2019S] ( Figure 5A) and LRRK2[R1441C] MEFs ( Figure 5B). Like GZD-824, Rebastinib inhibited wild-type LRRK2 with ∼2-fold higher potency than pathogenic mutations. We also observed that Rebastinib like other Type II inhibitors tested did not induce dephosphorylation of Ser935 despite efficiently inhibiting Rab10 phosphorylation ( Figure 5A,B).

The MLi-2 inhibitor resistant LRRK2 [A2016T] mutation suppresses sensitivity to GZD-824 and Rebastinib
In previous work we have characterized a mutation within the kinase domain of LRRK2[A2016T], that does not impact on catalytic activity but reduced ∼10-fold sensitivity towards certain Type I inhibitors tested including MLi-2 [7,11,13]. This mutation has been exploited to confirm physiological responses induced by MLi-2 such as Rab protein phosphorylation are indeed mediated by LRRK2 rather than an off-target effect of the   [7]. We therefore tested whether the A2016T mutation would impact sensitivity towards GZD-824 ( Figure 6A,B) and Rebastinib ( Figure 6C,D) using previously described wild type and LRRK2[A2016T] knock-in MEFs [7]. These studies revealed that the A2016T mutation markedly blocked the ability of both GZD-824 and Rebastinib to inhibit Rab10 phosphorylation, even at 3 mM, the highest concentration studied ( Figure 6). As expected, the LRRK2[A2016T] mutation also markedly suppressed inhibition of Rab10 phosphorylation induced by the Type I MLi-2 inhibitor.
Rab29 is localized at the Golgi and co-expression of LRRK2, and Rab29 in HeLa cells strikingly results in the bulk of LRRK2 being recruited to the Golgi apparatus ( Figure 7B, Supplemental Figures S3, S4) [56]. Treatment of HeLa cells co-expressing LRRK2 and Rab29 with 5 μM GZD-824 for 4 h did not significantly impact Rab29 mediated relocalization of LRRK2 to the Golgi apparatus ( Figure 7B). In parallel experiments, treatment with 1 μM GSK3357679A for 4 h also had no impact on recruitment of LRRK2 to the Golgi by Rab29 ( Figure 7B). Control immunoblotting study for this experiment is shown in Figure 7C. Consistent with previous work [56], the LRRK2 mediated phosphorylation of endogenous Rab10 at Thr73 and autophosphorylation at Ser1292 is enhanced by overexpression of Rab29 ( Figure 7C). Both Type I and Type II inhibitor suppress LRRK2 autophosphorylation at Ser1292, as well as phosphorylation of Thr73 of endogenous Rab10. Type II inhibitors do not impact LRRK2 Ser935 phosphorylation. We also show that LRRK2 mediated phosphorylation of Rab29 at Thr71, monitored using a recently released commercial phospho-specific antibody, is blocked by both Type I and Type II inhibitors ( Figure 7C).

GZD-824 stimulates basal mitophagy similarly to GSK3357679A
Previous work revealed that pathogenic LRRK2[G2019S] mutation significantly reduces basal mitophagy, but not general autophagy, in primary MEFs and various mouse tissues in a manner that is rescued by inhibiting LRRK2 with structurally distinct and selective Type I inhibitors, GSK3357679A and MLi-2 [25]. We therefore monitored the impact that 10-300 nM GZD-824 had on basal mitophagy in previously generated wild type, LRRK2[G2019S] knock-in and LRRK2 knock-out MEFs [7] stably expressing the mito-QC mitophagy reporter [25]. The mito-QC reporter is localized to mitochondria via an outer mitochondrial membrane targeting sequence derived from the protein FIS1 and bears a tandem GFP and mCherry tag. When a mitochondrion is delivered to lysosomes, the low lysosomal luminal pH quenches the GFP but not the mCherry signal. The degree of mitophagy is determined by the appearance of mCherry-only puncta, which represent mitochondria engulfed into autolysosomes [57][58][59]. Treatment of wild type or LRRK2[G2019S] cells for 24 h with 30-300 nM GZD-824 stimulated basal mitophagy ∼2-fold, to the same extent as GSK3357679A ( Figure 8A,B). This effect was specific for inhibition of LRRK2, as neither GZD-824 or GSK3357679A impacted basal mitophagy in LRRK2 knock-out cells. Consistent with previous work [25], basal mitophagy was lower in LRRK2 [G2019S] MEFs compared with wild-type MEFs and treatment with Type I or Type II inhibitors increased basal mitophagy to levels similar to those observed in LRRK2 knock-out cells ( Figure 8A,B).  (A) Primary neutrophils from two healthy donors were isolated as described previously [54], and treated with or without 1 μM GZD-824 or 300 nM MLi-2 for 45 min at ambient temperature for the indicated times. Cells were lysed and 25 mg of extract was subjected to quantitative immunoblot analysis with the indicated antibodies (all at 1 mg/ml). The membranes were developed using the Odyssey CLx Western Blot imaging system. Immunoblots were quantified using the Image Studio software. Data are presented relative to the phosphorylation ratio observed in cells treated with DMSO (no inhibitor).

Discussion
In this study, we compared the in vitro and cellular properties of Type I and 3 distinct Type II LRRK2 inhibitors. All Type II inhibitors studied, suppressed LRRK2 protein kinase activity in cells as judged by their ability to inhibit LRRK2 mediated Rab protein phosphorylation and stimulate basal mitophagy similar to Type I compounds. The most striking difference between Type I and Type II compounds is that Type II inhibitors do not induce dephosphorylation of Ser935 or other LRRK2 biomarker phosphorylation sites. This suggests that Type II inhibitors preserve biomarker site phosphorylation on LRRK2 by maintaining the LRRK2 kinase domain in the open inactive conformation. The mechanism underlying this phenomenon is not known, but we hypothesize that in the Type II-open kinase domain conformation, the N-terminal biomarker sites on LRRK2 are exposed for phosphorylation by the upstream kinase(s) and/or rendered inaccessible to protein phosphatase(s). Vice versa, in the Type I-closed conformation, access of the N-terminal biomarker sites to upstream kinase(s)   could be restricted/or N-terminal biomarker sites could be better exposed to protein phosphatase(s). In future analysis, it would be important to compare how Type I and Type II inhibitors influence the conformation and accessibility of the biomarker sites to upstream kinases and phosphatases.
The finding that prolonged treatment of cells with GZD-824 does not impact on Ser935 phosphorylation under conditions where Rab10 phosphorylation remains suppressed, argues against the biomarker sites located in the N-terminal region of LRRK2 being controlled by autophosphorylation. Although distinct kinases have been proposed to phosphorylate these sites [60][61][62], further work is required to validate which kinase(s) target these residues in vivo. A protein phosphatase-1 (PP1) complex has been reported to dephosphorylate these  sites in vivo [63]. Future work should also explore whether interaction of LRRK2 with microtubule is involved in controlling biomarker phosphorylation. Targeting of LRRK2 to microtubules in the closed kinase conformation, could bring LRRK2 in association with a microtubule associated phosphatase and/or restrict access to an upstream protein kinase. Our conclusions are also consistent with the finding that the majority of pathogenic mutations that promote the active closed conformation of LRRK2 (R1441G/C, Y1699C, I2020T) suppress biomarker site phosphorylation [11]. Consistent with this, treatment of LRRK2[R1441C] knock-in MEFs with GZD-824 ( Figure 4) or Rebastinib ( Figure 5), would be predicted to induce the open conformation and thus rescue Ser935 phosphorylation and this is what we observe. Our results emphasize that the steady state phosphorylation of the biomarker sites on LRRK2 correlates with the open vs closed conformation of the LRRK2 kinase domain rather than the intrinsic kinase activity towards Rab substrates. The finding that treatment of cells with Type II inhibitors does not further increase biomarker site phosphorylation suggests that the bulk of the cellular LRRK2 is in the open inactive conformation.
Neither GZD-824 (Supplemental Figure S1A), Rebastinib (Supplemental Figure S1B) or Ponatinib (Supplemental Figure S1C) are selective kinase inhibitors. Our finding that LRRK2[A2016T] mutation suppresses inhibition of LRRK2 by GZD-824 as well as Rebastinib (Figure 6), could be exploited in future studies to better distinguish whether physiological effects of these inhibitors are mediated by LRRK2, or other cellular kinases inhibited by these compounds. When testing physiological effects of the Type II compounds used in this study, we would also recommend verifying that the cellular effects are observed with diverse Type II LRRK2 inhibitors as off targets of these inhibitors differ to some extent (Supplemental Figure S1). Ponatinib is clinically approved for treating chronic myeloid leukemia and Philadelphia chromosome-positive acute lymphoblastic leukemia [64]. Rebastinib is currently being explored in Phase-1 clinical trials as an TIE2 kinase inhibitor for advanced or metastatic solid tumors (https://clinicaltrials.gov/ct2/show/NCT03717415). GZD-824, Rebastinib and Ponatinib inhibit wild-type LRRK2 more potently than pathogenic LRRK2 in vitro (Table 1) and in cells (Figures 4, 5), a feature that will likely be similar for other Type II inhibitors. As pathogenic mutations favor the closed conformation of LRRK2 kinase domain, it would be expected that a higher concentration of a Type II inhibitor would be required to inhibit pathogenic LRRK2. This difference in potency between wild type and pathogenic LRRK2 for Type II inhibitors would need to be taken into consideration when evaluating the optimal therapeutic doses of a potential future Type II inhibitor compound for treating LRRK2 driven Parkinson's disease.
To our knowledge all of the well-studied LRRK2 inhibitors induce dephosphorylation of the N-terminal biomarker sites on LRRK2 suggesting that these inhibitors will be Type I compounds. Previous work in LRRK2-deficient rodents as well as nonhuman primates treated with LRRK2 inhibitors highlighted preclinical liability concerns in lung and kidney, resulting from loss of LRRK2 kinase activity that might be associated with altered autophagy and lysosome biology [65][66][67]. In future work, it would be important to explore the physiological impact that selective Type II LRRK2 inhibitors have on animal models in comparison with Type I compounds. It is possible that Type II compounds that do not induce microtubule association or lead to dephosphorylation of biomarker sites may have a different physiological impact. We would argue that the poor selectivity of GZD-824, Rebastinib and Ponatinib precludes their use as advanced tool compounds to probe and compare the detailed physiological impacts of Type I and Type II LRRK2 inhibitors. In future work we would advocate that it would therefore be important to develop more selective Type II LRRK2 tool inhibitors to enable their properties to be better studied.

Data Availability
All of the primary data that is presented in this study can be requested in electronic form by contacting Dario Alessi (d.r.alessi@dundee.ac.uk) has been deposited on the Zenodo data repository (D.O.I 10.5281/zenodo.5501303). All plasmids and antibodies (and associated datasheets) generated at the MRC Protein Phosphorylation and Ubiquitylation Unit at the University of Dundee can be requested through our reagent's website https://mrcppureagents.dundee.ac.uk/. Requests for provision of GSK3357679A should be directed to Alastair Reith (alastair.d.reith@gsk.com).