Fig 2: (A) The concentration of EB‐42168 that inhibited pSer935 LRRK2 by 90% in homozygous (HOM) G2019S LRRK2 PBMCs (172 nM) resulted in significantly less inhibition of pSer935 leucine‐rich repeat kinase 2 (LRRK2) in heterozygous (HET) G2019S LRRK2 PBMCs (36 ± 7.5%, P < 0.0001) and noncarriers (NC) peripheral blood mononuclear cells (PBMCs) (5 ± 3%, P < 0.0001). (B) The concentration of EB‐42168 that inhibited pThr73 Rab10 by 90% in HOM (582 nM) resulted in significantly less inhibition of pThr73 Rab10 in HET G2019S LRRK2 (44 ± 15%, P < 0.0001) and NC PBMCs (9 ± 12%, P < 0.0001). (C) The IC90 concentration of MLi‐2 for pSer935 LRRK2 inhibition in HOM G2019S LRRK2 PBMCs (10 nM) also inhibited the other genotypes by >90% (HET G2019S LRRK2 = 95 ± 2.7%, P = 0.047; NC = 92 ± 2.2%). (D) The MLi‐2 IC90 concentration for pT73 Rab10 (33 nM) resulted in a similarly high level of inhibition in HET G2019S LRRK2 (90 ± 3.6%) and NC (96 ± 2.9%). Bar height represents the average signal quantitated from immunoblots for pSer935 LRRK2 normalized to total LRRK2 relative to DMSO control (A and C) or pThr73 Rab10 normalized to total Rab10 relative to DMSO control (B and D). Error bars represent standard deviation (SD) of n = 3 HOM, 5 HET, and 5 NC of the G2019S LRRK2 variant. **P < 0.005, ***P < 0.0001.

Fig 3: Two weeks MLi-2 treatment does not alter ciliogenesis in R1441C LRRK2 striatal interneurons or astrocytes.Mice (8 weeks old) were fed MLi-2 LRRK2 inhibitor-containing chow or control chow for 2 consecutive weeks prior to perfusion and staining. (A) Confocal images of sections of the dorsal striatum from 8 week R1441C LRRK2 KI mice; ChAT (green, white outline); Adenylate cyclase 3 (AC3) (magenta), and DAPI (blue). (B) Quantitation of the percentage of ChAT+ and ChAT- neurons containing a cilium. (C) Percentage of ChAT+ neurons containing a cilium± MLI-2. (D) Quantitation of ChAT+ neuron ciliary length. (E) Confocal images of astrocytes identified by antibodies to GFAP (magenta, white outline), Arl13B (green), and DAPI (blue), ± MLi-2. (F, G) Quantitation of the percentage of total astrocytes (F) or GFAP/S100B+ astrocytes (G) containing a cilium. Scale bars, 10 µm. Significance was determined by t-test. B, *, p = 0.0417.

Fig 4: Schematic summary showing how VPS35 might regulate LRRK2 and LRRK2 protein kinase activity and hence PD.The mechanism by which LRRK2 is activated in vivo has not been elucidated. It is also not known what maximal activation of LRRK2 kinase activity is achievable. It is thought that the LRRK2 pathogenic mutations cause only a partial activation of LRRK2. Mutations in the kinase domain, such as G2019S, induce a moderate ~2-fold activation of LRRK2, perhaps by stabilising the active confirmation of the kinase fold. Pathogenic mutations in the ROC/COR domain of LRRK2, such as R1441G/C, increase LRRK2 kinase activity ~3–4-fold. Three recent papers [10–12] indicate that the ROC/COR mutations, promote recruitment of LRRK2 to the Rab29 protein that is located at the Golgi, which leads LRRK2 activation through an as yet unknown mechanism. VPS35 and Rab29 could operate in the same pathway with LRRK2 [32]. Our work shows that the VPS35[D620N] mutation through an unknown pathway is inducing a higher 5–6-fold activation of LRRK2 than achieved with the known pathogenic mutations. Indeed, this is supported by age-of-onset data shown in Figure 9 that suggest that Parkinson's patients with VPS35[D620N] develop the disease at a younger age than those with LRRK2 mutations. In future work, it would therefore be vital to define the mechanism by which VPS35[D620N] is activating LRRK2 and explain why VPS35[D620N] mutations induce a greater activation of LRRK2 than observed with the LRRK2 pathogenic mutants. We propose that by understanding this mechanism it might be possible to develop compounds that target the retromer complex that suppress activation of LRRK2.

Fig 5: Quantitative analysis of phosphorylation and expression of selected PD and CD-associated LRRK2 variants assessed in primary screens.FLAG-tagged LRRK2 wildtype, kinase dead (KD = D2017A), and the indicated variants were transiently expressed in HEK293 cells. Twenty-four hours post-transfection, cells were lysed and analysed by quantitative immunoblotting (as in Supplementary Figure S2A). Quantified immunoblotting data are presented as ratios of phospho-LRRK2 Ser1292/total LRRK2 (A), phospho-LRRK2 Ser935 (B), and total LRRK2/Tubulin (C), normalized to the average of LRRK2 wildtype values for each replicate (mean ± SD). Combined immunoblotting data from six independent biological replicates are shown. Dashed lines segment the graphs into corresponding regions of LRRK2 as listed in the domain schematic.