Accurate MS-based Rab10 phosphorylation stoichiometry determination as readout for LRRK2 activity in Parkinson’s disease

Pathogenic mutations in the Leucine-rich repeat kinase 2 (LRRK2) are the predominant genetic cause of Parkinson’s disease (PD). They increase its activity, resulting in augmented Rab10-Thr73 phosphorylation and conversely, LRRK2 inhibition decreases pRab10 levels. However, there is no assay to quantify pRab10 levels for drug target engagement or patient stratification. We developed an ultra-sensitive targeted mass spectrometry (MS)-based assay for determining Rab10-Thr73 phosphorylation stoichiometry in human samples. It uses synthetic stable isotope-labeled (SIL) analogues for both phosphorylated and non-phosphorylated tryptic peptides surrounding Rab10-Thr73 to directly derive the percentage of Rab10 phosphorylation from attomole amounts of the endogenous phosphopeptide. We test the reproducibility of our assay by determining Rab10-Thr73 phosphorylation stoichiometry in human neutrophils before and after LRRK2 inhibition. Compared to healthy controls, neutrophils of LRRK2 G2019S and VPS35 D620N carriers robustly display 1.4-fold and 3.7-fold increased pRab10 levels, respectively. Our generic MS-based assay further establishes the relevance of pRab10 as a prognostic PD marker and is a powerful tool for determining LRRK2 inhibitor efficacy and for stratifying PD patients for LRRK2 inhibitor treatment.


INTRODUCTION
Parkinson's disease (PD) is the second most common neurodegenerative disease, and no disease-modifying therapies exist to date 1,2 .
Although most PD cases are idiopathic, mutations in several genes have been linked to familial forms of the disease 3 . Among those, mutations in the Leucine-rich repeat kinase 2 (LRRK2) comprise the predominant genetic cause of PD and account for 1% of sporadic and 4% of familial cases worldwide, and much higher in some populations 4 . At least six pathogenic missense mutations in LRRK2, including the most frequent G2019S substitution, have been identified 5 and several studies confirmed that these mutations increase its kinase activity [5][6][7][8] . LRRK2-associated PD is clinically largely indistinguishable from idiopathic PD, suggesting that LRRK2 inhibition may be useful as disease-modifying therapy for a larger group of patients 4 . Clinical trials with selective LRRK2 kinase inhibitors are ongoing and have already passed phase 1.
LRRK2 directly phosphorylates Rab10 at Thr73 and all previously tested pathogenic forms of LRRK2 enhance this phosphorylation 8,9 . Intriguingly, the PD-associated D620N mutation of the retromer complex protein VPS35 also activates LRRK2 kinase activity, which in turn results in augmented Rab10 phosphorylation 15 . Thus, multiple PD-associated factors are interconnected and dysregulation of a common LRRK2-Rab signaling pathway may be an underlying cause of PD.
The LRRK2 autophosphorylation site Ser1292 has been widely used for assessing LRRK2 kinase activity [16][17][18][19] . However, its phosphorylation levels appear to be extremely low and there is no sensitive phospho-specific antibody available to reliably detect and quantify phosphorylation at this site 6,20,21 . Instead, we recently developed several highaffinity antibodies for detecting pRab proteins in cells and in tissues 14 . Among those, a highly specific clone detects pThr73 levels in human peripheral blood cells of (mutant) LRRK2 G2019S and VPS35-D620N carriers 11,15 . While pRab10 levels were markedly increased in VPS35 D620N carriers, no statistically significant differences in pRab10 levels were detected by immunoblotting analysis when comparing controls to LRRK2-G2019S carriers in peripheral blood mononuclear cells (PBMCs) and neutrophils 11,15 . The reproducible quantification of immunoblots, particularly the detection of small (<2) foldchanges is challenging. In contrast, mass spectrometry-based quantification has become the gold standard and has several advantages over traditional biochemical methods, as it is more specific and more precise. Importantly, it allows simultaneous detection of both the phosphorylated peptide and the total protein pool and hence calculation of the absolute fraction of the phosphorylated protein, also known as phosphorylation stoichiometry or occupancy 22 .
There are several strategies differing in their accuracy, throughput, and applicability for measuring phosphorylation stoichiometry by MS [22][23][24][25][26][27] . One way is to compare the intensities of modified and unmodified peptides by label-free proteomics 27  Alternatively, a SIL recombinant protein that is chemically or enzymatically phosphorylated can also be used as the spike-in standard 28,29 . MS strategies determining stoichiometry that are combined with tailored targeted methods are especially suited for accurate quantification of low levels of a given phosphorylated analyte 30 . In particular, instruments with an Orbitrap as a mass analyzer can be operated in targeted scan modes such as high resolution selected ion monitoring (SIM) or parallel reaction monitoring (PRM). In both methods, precursor ions are isolated with a narrow m/z range by a quadrupole mass filter before introduction into the Orbitrap analyzer, thereby providing increased signal to noise ratio for the ion of interest and allowing attomole-level limits of detection 31,32 . As no fragmentation is involved in SIM, quantification of selected ions relies on the high resolution in the Orbitrap mass analyzer. In contrast, in PRM, fragment ions are used for quantification of the peptide. While this approach is more specific, the overall sensitivity of PRM can be lower, as the signal of a given precursor ion is distributed across multiple fragments, although the signal to noise ratio can be higher 32 .
Here we describe an accurate and ultra-sensitive

Rab10-pThr73 serves as a readout for LRRK2 activity in human peripheral blood neutrophils
To explore which Rab GTPases were expressed in human peripheral blood neutrophils and which of them could serve as a readout for LRRK2 activity 5 in a quantitative MS-based assay, we first isolated neutrophils from whole blood using a negative selection approach. We assessed the recovery and the purity of the enriched cells by flow cytometry ( Supplementary Fig. 1). Proteomic analysis resulted in 5,488 quantified proteins, for which we estimated the copy numbers per cell using the proteomic ruler approach 33 (Fig. 1C). In the same lysates, we enriched phosphorylated Rab proteins with the previously described pThr-specific Rab antibody and subjected the eluates to LC-MS/MS analysis 9 .
We detected 19 Rab GTPases in total, of which 7 were previously shown to be LRRK2 targets 9 ( Fig.   1D and Supplementary Fig. 3     Next, to determine whether our assay was sufficiently sensitive for quantifying the percentage of Rab10-pThr73 in human peripheral blood, we isolated neutrophils from five healthy volunteers. Upon treatment with either 30 nM or 100 nM of MLi-2, we found a decrease in Rab10-pThr73, as judged by immunoblotting (Fig. 3E).
When we determined the phosphorylation stoichiometry by MS, we found very low levels of  Table 2).