LRRK2-Associated Parkinsonism With and Without In Vivo Evidence of Alpha-Synuclein Aggregates

Background: Among LRRK2-associated parkinsonism cases with nigral degeneration, over two-thirds demonstrate evidence of pathologic alpha-synuclein, but many do not. Understanding the clinical phenotype and underlying biology in such individuals is critical for therapeutic development. Our objective was to compare clinical and biomarker features, and rate of progression over 4 years follow-up, among LRRK2-associated parkinsonism cases with and without in vivo evidence of alpha-synuclein aggregates. Methods: Data were from the Parkinson’s Progression Markers Initiative, a multicenter prospective cohort study. The sample included individuals diagnosed with Parkinson disease with pathogenic variants in LRRK2. Presence of CSF alpha-synuclein aggregation was assessed with seed amplification assay. A range of clinician- and patient- reported outcome assessments were administered. Biomarkers included dopamine transporter SPECT scan, CSF amyloid-beta1–42, total tau, phospho-tau181, urine bis(monoacylglycerol)phosphate levels, and serum neurofilament light chain. Linear mixed effects models examined differences in trajectory in CSF negative and positive groups. Results: 148 LRRK2-parkinsonism cases (86% with G2019S variant), 46 negative and 102 positive for CSF alpha-synuclein seed amplification assay were included. At baseline, the negative group were older than the positive group (median [interquartile range] 69.1 [65.2–72.3] vs 61.5 [55.6–66.9] years, p<0.001) and a greater proportion were female (28 (61%) vs 43 (42%), p=0.035). Despite being older, the negative group had similar duration since diagnosis, and similar motor rating scale (16 [11–23] vs 16 [10–22], p=0.480) though lower levodopa equivalents. Only 13 (29%) of the negative group were hyposmic, compared to 75 (77%) of the positive group. Lowest putamen dopamine transporter binding expected for age and sex was greater in the negative vs positive groups (0.36 [0.29–0.45] vs 0.26 [0.22–0.37], p<0.001). Serum neurofilament light chain was higher in the negative group compared to the positive group (17.10 [13.60–22.10] vs 10.50 [8.43–14.70]; age-adjusted p-value=0.013). In terms of longitudinal change, the negative group remained stable in functional rating scale score in contrast to the positive group who had a significant increase (worsening) of 0.729 per year (p=0.037), but no other differences in trajectory were found. Conclusion: Among individuals diagnosed with Parkinson disease with pathogenic variants in the LRRK2 gene, we found clinical and biomarker differences in cases without versus with in vivo evidence of CSF alpha-synuclein aggregates. LRRK2 parkinsonism cases without evidence of alpha-synuclein aggregates as a group exhibit less severe motor manifestations and decline may have more significant cognitive dysfunction. The underlying biology in LRRK2-parkinsonism cases without evidence of alpha-synuclein aggregates requires further investigation.


Introduction
Individuals with LRRK2-associated parkinsonism uniformly demonstrate neuronal degeneration in the substantia nigra and locus coeruleus [1][2][3] , but the underlying proteinopathy is variable.A majority (60-80%) of cases demonstrate evidence of neuronal-predominant misfolded and aggregated alpha-synuclein (asyn), whether in vivo based on cerebrospinal fluid (CSF) testing or on post-mortem neuropathological examination 1,4,5 .However, over one-third may not have evidence of asyn aggregates.Understanding the clinical phenotype and underlying biology in such individuals is critical for molecularly-targeted therapeutic development 6 .Other pathologies present in some individuals with LRRK2-associated parkinsonism who do not demonstrate evidence of asyn aggregates include tauopathy, with Alzheimer's disease (AD) type tau (3R and 4R) predominating, but some demonstrate hyperphosphorylated tau resembling progressive supranuclear palsy (PSP), and less commonly TAR DNA-binding protein 43 (TDP43) 3,7,8 .
Studies to date indicate that individuals with LRRK2-associated parkinsonism with and without evidence of asyn aggregates are largely clinically indistinguishable, with a few noted differences.Asyn positive LRRK2-parkinsonism cases have been reported to have more nonmotor symptoms including hyposmia 4 , cognitive impairment, anxiety and orthostatic hypotension compared to asyn negative cases 2 .However, prior data are limited by small sample sizes and a lack of extensive clinical and biomarker characterization of cases.The Parkinson's Progression Markers Initiative (PPMI) offers the unique opportunity to address key gaps in knowledge regarding clinical, biomarker, and genetic differences in LRRK2associated parkinsonism with and without evidence of asyn aggregates, given that the cohort has had in vivo assessment of asyn aggregates in CSF as well as extensive longitudinal phenotyping in a relatively large number of cases.Indeed, findings have emerged from PPMI 4 demonstrating that among individuals with LRRK2-associated parkinsonism, absence of detectable asyn aggregates is most prevalent among those who are normosmic, especially among females.
We undertook this study with the objectives of comparing among LRRK2-associated parkinsonism cases with and without evidence of asyn aggregates whether there are (1)   differences in clinical features cross-sectionally and longitudinally (2) distinguishing features in available biofluid or imaging markers cross-sectionally and longitudinally (3) differences in prevalence of PD genetic risk.While acknowledging that LRRK2-associated parkinsonism without evidence of asyn aggregates are a biologically heterogeneous group, we hypothesized that LRRK2-associated parkinsonism without evidence of asyn aggregates would generally follow a more benign motor course.

Sample
Data were from the PPMI, a multicenter prospective cohort study.PPMI methods have been described elsewhere in detail 9 .Briefly, PPMI recruited individuals diagnosed with PD based on clinical features who were sporadic (without known pathogenic variants associated with PD) and a group with parkinsonism and known pathogenic variants in LRRK2.Inclusion criteria for the sporadic PD group were abnormal dopamine transporter (DAT) SPECT imaging by visual inspection, 2 years or less since diagnosis, not receiving dopaminergic treatment and not expected to require it within 6 months of enrollment.The LRRK2associated parkinsonism group was enrolled irrespective of treatment and if disease duration was 7 or less years.Exclusion criteria for all enrolled groups included dementia and medical conditions that preclude study activities.
The sample for this analysis is comprised of individuals with LRRK2-associated parkinsonism (LRRK2-parkinsonism) and a sporadic PD (sPD) group frequency matched to the LRRK2-parkinsonism group for age and time since diagnosis at enrollment.Inclusion criteria for this analysis were: (1) availability of asyn seed amplification assay (SAA) result (see methods below) (2) positive asyn SAA (CSFasynSAA+) result for the matched sPD group.Exclusion criteria were lowest putamen DAT specific binding ratio ≥65% of expected for age and sex in individuals who had a negative asyn SAA (CSFasynSAA-) result, presence of known pathogenic GBA1 variant (as presence of pathogenic glucocerebrosidase (GBA1) variants in individuals with LRRK2 can potentially modify the phenotype), and inconclusive or multiple system atrophy-like SAA results.
Baseline visit (time zero) for this analysis was the baseline study assessment for participants in the LRRK2 parkinsonism group and for the sPD group it was the first visit at which they were frequency matched for age and time since diagnosis.

Assessments of Motor and Non-Motor Function
Motor and non-motor assessment of signs, symptoms and function in PPMI that are assessed at baseline and at each annual visit are as follows: -Demographics: age, sex at birth, years of education, self-reported race and ethnicity -Clinical history: age at parkinsonism symptom onset, duration since PD clinical diagnosis at baseline visit, levodopa equivalent daily dose (LEDD) -Movement Disorders Society Modified Unified Parkinson's Disease Rating Scale (MDS-UPDRS) parts 1, 2, and 3.An ambulatory capacity score was calculated as the sum of MDS-UPDRS items 2.12, 2.13, 3.10, 3.11, 3.12

Genotyping
Genotyping methods in PPMI are described in detail at ppmi-info.org.Briefly, each PPMI participant receives a determination of presence or absence of pathogenic variants in the LRRK2 gene (or other genes) as well as APOE genotype.Population genetic structure was inferred with principal component analysis as described 14 .
In addition, we procured genome sequencing data from the Accelerating Medicines Partnership Parkinson's Disease (AMP-PD) project.The data processing methodology is detailed in a public GitHub repository 15 , follows the methods outlined by Nalls et al 16 , utilizing 90 risk-associated SNPs.However, for this study, we omitted two SNPs located in the LRRK2 region.We thus generated a modified polygenic risk score (mPRS), the cumulative risk weighted by the effect estimates of associated genetic variants, consisting of 88 SNPs.

Biomarker assessments
Presence of aggregated alpha-synuclein in CSF obtained at the baseline visit was assessed using the alpha-synuclein (asyn) seed amplification assay (SAA) as described 4,17 .The Fmax (highest raw fluorescence from each well), T50 (time to reach 50% of the Fmax), and TTT (time to reach a target RFU threshold) were used to define positive (CSFasynSAA+), inconclusive, negative (CSFasynSAA-), and multiple system atrophy-like (MSA-like) assays as described 4,17 .
. Dopamine transporter binding (DAT) was assessed with DATscan and SPECT as previously described 8 .Percent of expected lowest putamen specific binding ratio (SBR) for age and sex was determined using normative data from healthy controls in PPMI.

Statistical Analysis
Baseline demographic and clinical features were compared in the CSFasynSAA-and CSFasynSAA+ using two-sample Wilcoxon rank sum test, chi-square test or Fisher's exact test as appropriate.To account for differences due to age, linear regression and logistic regression adjusting for age for continuous and categorical outcomes, respectively were used to model clinical outcomes and biomarkers with SAA as an explanatory variable.Log, square root, or rank transformations were applied to models with non-normally distributed residuals.
The specific transformations used were marked on the tables and detailed in the table footers.Summary statistics were examined for motor, non-motor, and biologic variables from baseline to year 4.
Only individuals with at least 1 annual follow-up visit following baseline were included in longitudinal analyses.To assess whether the longitudinal trajectory of the outcome measures differed between CSFasynSAA-and CSFasynSAA+ groups, generalized linear mixed-effects models (LMM) with random intercept and slope and unstructured working correlation structure were employed.Specifically, CSF asyn SAA status, time in years, and their interaction were included in the models.This analysis assumed a linear fit in the link function of mean responses over time from year 1 to year 4, wherever available, using the Restricted Maximum Likelihood and Residual Pseudo-Likelihood methods when appropriate.
Continuous biologic CSF outcomes were ranked at each time point and modeled to evaluate whether the longitudinal trajectory of the mean rank response differed by CSF asyn SAA groups, assuming a linear fit in the mean rank of each response over time from year 1 to year 4, when available.Similarly, models were employed to assess whether the longitudinal trajectory of log odds for categorical response variables differed based on CSF asyn SAA status from year 1 to year 4. Random intercept only models were used for outcomes with convergence issues.Wald tests were conducted to assess the statistical significance of the interaction term between CSF asyn SAA status and time.A quadratic fit model was also tested if the linear fit did not result in a significant interaction.To explore sex differences, a three-way interaction model with sex, CSF asyn SAA, and time was also tested.An identity link and logit link were chosen for continuous and categorical response variables, respectively.Time effect p-values were reported for all models, with separate time effects provided for each CSF asyn SAA status when the interaction term was significant.
All models adjusted for baseline value of the outcome, age, sex, years since diagnosis at enrollment, and genetic principal components PC1, PC2, PC3 14 .Models involving outcomes that may be affected by PD medications, such as MDS-UPDRS Part 3, and Ambulatory Capacity Score also adjusted for time-varying LEDD in the model.
All longitudinal analyses were conducted under the assumption of missing at random (MAR).
Sensitivity analyses were employed to evaluate the plausibility of the MAR assumption.
Intermittent missing values were imputed using Monte Carlo Markov Chain methods 23 .
Multiple imputation was used for outcomes displaying significant interactions.Notably, a one-dimensional tipping point analysis was utilized to assess the significance of the .interaction term by systematically shifting the mean of missing values at year 4 for these outcomes in the opposite direction of significance, identifying the point at which the interaction term becomes nonsignificant.
To determine if differences in CSFasynSAA-and CSFasynSAA+ parkinsonism cases vary according to LRRK2 status, when cross-sectional or longitudinal analysis revealed significant differences in the LRRK2 parkinsonism CSFasynSAA-and CSFasynSAA+ group for a given outcome, the outcome was then compared in the LRRK2 parkinsonism CSFasynSAA-group to the sPD CSFasynSAA+ group using the same statistical method.
For comparison of genetic risk variants in the LRRK2-associated parkinsonism cases, we compared mPRS in CSFasynSAA+ and CSFasynSAA-groups using logistic regression with CSFasynSAA-as the reference group.In addition, we conducted an examination of individual GWAS risk variants to evaluate their association with CSF asyn SAA status, adjusting for age, sex, and the first three genetic principal components.Given the exploratory nature of this study, we set the significance threshold at 0.05 (two-tailed).
All analyses were conducted in SAS Institute Inc. (SAS Institute Inc version 9.4 Cary, NC).
The final analytic sample included 148 LRRK2-associated parkinsonism cases and a comparator group of 378 sporadic PD CSFasynSAA+ (sPD) frequency matched to them by age and disease duration.Seven participants did not have follow-up beyond baseline.Up to 4 follow-up visits (after baseline) were expected for 141 LRRK2-associated parkinsonism cases; the majority completed year 4 (31 (69%) CSFasynSAA-and 78 (81%) CSFasynSAA+ .cases).Among the 32 cases who did not complete year 4, 9 contributed data at later time points, 13 withdrew from the study before year 4, and 7 were lost to follow-up.
There were no differences in other non-motor measures or tests of cognitive function in the two groups (Table 2). .

Comparison of PD genetic risk variants
The analysis comparing risk variants was confined to LRRK2-parkinsonism cases of European ancestry (n=130) of which 48 were CSFasynSAA-and 82 were CSFasynSAA+.None of the other assessed rating scales, imaging, or biofluid biomarkers changed significantly in the two groups when the outcome was modeled as linear.When a quadratic term was introduced, the interaction with the second order term was significant for SCOPA-AUT, though there was minimal overall change in SCOPA-AUT total score (Table 4).
None of the imaging, CSF, serum, or urine biomarkers changed differently in the CSFasynSAA-and CSFasynSAA+ groups over time (supplementary table 2).Data on biofluid biomarkers were missing on a substantial number at later time points of follow-up (supplementary table 2).
Testing of a 3-way interaction term between sex and SAA status did not reveal any differences in change according to SAA and sex (supplementary table 1), but sample sizes in the subgroups at later time points were small.

Discussion
In this large sample of individuals with LRRK2-associated parkinsonism, we compared clinical, imaging and biofluid biomarker, and genetic characteristics among those with .evidence of CSF asyn aggregates compared to those without.Importantly, and unique to this cohort, all assessments occurred in vivo in participants who had received a clinical diagnosis of PD and had dopaminergic dysfunction as evidenced by DAT imaging.Taken together, our results indicate that while the CSF asyn CSFasynSAA-and CSFasynSAA+ groups are largely similar, there are some important differences.The CSF asyn CSFasynSAA-group had less severe motor dysfunction (and a trend toward more severe cognitive dysfunction at baseline).Concordantly, they had less advanced dopaminergic neuron dysfunction, as evidenced by DAT binding measures.By contrast, the CSFasynSAA-group had higher serum NfL, a biomarker that predicts increased risk of cognitive decline 24 .Interpretation of these results requires consideration for sex and age differences in the compared groups, as well as differences in disease duration at enrollment.Longitudinal analysis revealed that the CSFasynSAA-group, despite being older and receiving less dopaminergic therapy, did not decline in motor functional rating scale, in contrast to the CSFasynSAA+ group who had significant worsening of functional impairment over time.

High prevalence of LRRK2 parkinsonism cases without evidence for asyn aggregates
In the PPMI sample of LRRK2 parkinsonism cases included in this analysis, one-third had no evidence of asyn aggregates based on CSF asyn SAA.This is in contrast to sporadic PDindividuals with a clinical diagnosis of PD who do not have any known pathogenic variantswhere only 6.7% 4 .-9% 4,25of cases do not have evidence of asyn aggregates.
It is likely that most cases that are negative for CSF asyn SAA are negative for asyn in the brain.This is supported by several lines of evidence including measurement of asyn with a variety of methods and autopsy-CSF correlation 26 , including one of the cases included in this analysis that was examined postmortem and showed no Lewy pathology 4 .Nevertheless, it is possible that in some cases the CSF test is false negative.Indeed, some neuropathologically examined cases with confirmed Lewy body pathology have been CSF asyn SAA-; these are most often focal Lewy pathology, such as in the amygdala or brainstem 27 .On the other hand, there is a reported case of LRRK2-associated parkinsonism that did not demonstrate postmortem Lewy pathology but who demonstrated asyn aggregates on brain homogenate by .asyn SAA 28 .Regardless of detection of asyn, of course, this does not exclude the possibility that pathogenic variants in LRRK2 may impact asyn function without leading to Lewy pathology or abnormal CSF SAA 29 .
It is also likely that in some individuals neurodegeneration does occur independent of presence of misfolded asyn.Indeed, pathogenic LRRK2 variants have been associated with various proteinopathies including AD, various tauopathies including PSP, corticobasal degeneration (CBD), familial frontotemporal degeneration (FTD), TDP-43-associated neurodegeneration 29 .Neuropathological studies are skewed toward cases with clinical features of parkinsonism 3 , and given that the prevalence of LRRK2 pathogenic variants in the general population is not small, interpretation of results in cases with other clinical diagnoses who have been autopsied is difficult; in some cases the genetic variant may be incidental.
Having said that, a few studies that have screened for LRRK2 pathogenic variants in brain banks offer insights into the prevalence of LRRK2 pathogenic variants in a range of neurodegenerative disorders.In a series of 110 cases 30 , of which 66 were synucleinopathies, 29 tauopathies, and 3 non-specific nigral degeneration, the prevalence of positivity of pathogenic variants in LRRK2 gene was 1.8%.One case had PD based on clinical criteria and neuropathological examination, whereas another case had been diagnosed with PD based on clinical criteria, but neuropathological examination demonstrated nonspecific nigral degeneration without Lewy bodies.A p.R1441R variant was detected in another PD case 30 .
Taking together data from published case series, approximately 22% of LRRK2 associated parkinsonism cases demonstrate neuropathological findings of hyperphosphorylated tau, as occurs in PSP 7 .
Several possible biologic mechanisms could be implicated in LRRK2-mediated neurodegeneration, whether related to asyn aggregates or independent of it.All pathogenic variants in the LRRK2 gene are missense mutations and have been found throughout the gene 29 .The LRRK2 protein is a large, complex multidomain protein that functions as a protein kinase.Altered LRRK2 signaling has been implicated in dysfunction in a range of cellular processes and molecular pathways including vesicular trafficking, autophagy, lysosomal degradation, endolysosomal stress, microglial response, calcium dysmetabolism and resultant endoplasmic reticulum stress, neuroinflammation, mitophagy, and mitochondrial dysfunction [5][6][7] .The PPMI cohort is being characterized with extensive proteomics and transcriptomics data which will allow investigation of differences in these various biologic processes in asyn positive and negative cases in the future.

Female predominance among LRRK2 parkinsonism cases without evidence for asyn aggregates
We found a female predominance among the LRRK2 parkinsonism cases without evidence for pathologic asyn.There is extensive literature that demonstrates that in individuals diagnosed with PD, sex differences exist for clinical, biomarker, neuropathological, or genetic endpoints 31 .Sex differences in LRRK2-parkinsonism cases are particularly notable.
A meta-analysis 32 of 66 studies of LRRK2-associated parkinsonism (that were not biologically characterized) revealed a higher prevalence of LRRK2 pathogenic variants in females diagnosed with PD.In a study 33 of 530 LRRK2-associated parkinsonism, and compared to 759 sporadic PD cases, the male predominance observed in sporadic PD was not seen in the LRRK2-associated cases.
As mentioned, asyn-negative LRRK2 parkinsonism cases often exhibit AD pathology, and these results could in part be a reflection of sex differences in AD.For example, women have a greater burden of neurofibrillary tangles 34,35 , and women with AD pathology are more likely to manifest clinically with dementia but not to be diagnosed with DLB 35 .The effect of sex on tau may even be brain-region specific, and females may have network characteristics favoring spread of tau 36 .Women with AD pathology are more likely to have copathology with TDP/hippocampal sclerosis abd cerebrovascular disease.On the other hand, male sex is more likely to be associated with pure Lewy body pathology (absence of copathology 34 ).
There are several possible mechanisms that could explain sex differences in asyn pathology and in relation to LRRK2 that require investigation.Exposure to sex hormones has been postulated as one possible mechanism explaining a predominance of tau pathology in females compared to males.The higher likelihood of diagnosis of neurodegenerative disorders postmenopause has been observed.Estradiol may have a protective effect against hyperphosphorylation of tau 38 .Estrogen receptor colocalizes with neurofibrillary tangles.
Another possible mechanism may relate to the effect of estrogen on mitochondria and oxidative stress 38,39 .A relationship between sex hormones and neuroinflammation is another possible mechanism that may explain sex differences 39 , which is of particular relevance given the role of LRRK2 in the immune system.Sex differences in immune activation and microglial function may also play a role 31,39 .. Indeed, a study examining the serum profile of 23 immune-associated markers in sporadic and LRRK2-associated PD demonstrated sex differences in immune profile but without differences in the LRRK2 and sporadic group 40 .
While differential genetic risk factors for PD in men vs women have not been demonstrated, sex-specific effects of genotype may exist 31,41 .11 genomic loci have jointly been associated with PD and sex-specific traits, namely age of menarche and age at menopause.Many of the genes that mapped to loci shared between PD and age at menarche have been implicated in PD pathophysiology, including immune activation and regulation 41 .Sex-specific differences in LRRK2 brain expression in healthy controls (but not in PD) have been observed 41 .The effect of age on expression of genes that may be relevant in PD pathophysiology may also vary by sex 41 .. A greater burden of tau among women has been postulated to be mediated by ApoE status 42 , and upregulation of ApoE expression by estrogen was postulated as a possible mechanism 42 .
We did not find differences in ApoEe4 genotype in CSFasynSAA+ vs CSFasynSAA-but we had a small sample size and low prevalence of ApoEe4 in our sample.With larger sample sizes and by comparing proteomic or transcriptomic data, these hypotheses can be investigated in future studies.Gender differences in behavioral, occupational, environmental exposures may also contribute 7,43,44 and deserve investigation.

Lower Prevalence of Olfactory Dysfunction in the group without evidence for asyn aggregates, especially among females
A lower prevalence of olfactory deficit among LRRK2-associated parkinsonism has been previously identified 33,45 , but in PPMI it has been demonstrated that this finding is largely restricted to LRRK2-associated parkinsonism without evidence of asyn pathology 4 .In a study 33 of 530 LRRK2-associated parkinsonism, and compared to 759 sporadic PD cases, female LRRK2-parkinsonism individuals were less likely to have olfactory deficit 33 .
However, in that study, biological characterization was not present.The PPMI study sample now enables demonstration that asyn negative LRRK2 parkinsonism cases are much more likely to be normosmic 4 .One possible explanation for these findings is the preferential susceptibility of olfactory bulb 46 and anterior olfactory nucleus 47,48 cells to asyn pathology, as evidenced by data from animal models.Future studies of asyn pathology in nasal mucosa in LRRK2 cases may shed light on the observed differences in olfactory dysfunction we report here.

Less Severe Motor Dysfunction and Functional Impairment in the group without evidence for asyn aggregates
Despite being older, having similar disease duration, and lower LEDD at baseline assessment, the CSFasynSAA-had similar scores on MDS-UPDRS including part III ON score to the CSFasynSAA+ group.These results may indicate less severe motor involvement in the CSFasynSAA-group.Concordant with this, the CSFasynSAA-group remained stable in the MDS-UPDRS II, a multidomain, motor-predominant functional rating scale, whereas the CSFasynSAA+_group had a significant increase (declining function) over time.One possibility to explain these findings is that the underlying pathology in these cases leads to less severe affectation of dopaminergic pathways and other pathways implicated in parkinsonian motor abnormalities.The less severe DAT loss in this group supports this hypothesis.Indeed, while dopaminergic neuronal loss occurs in a range of neurodegenerative disorders, there may be disease-specific susceptibility.
In light of the differences in MDS-UPDRS II in the LRRK2-parkinsonim CSFasynSAA-vs CSFasynSAA+ cases, we next examined differences in LRRK2 CSFasynSAA-vs sPD CSFasynSAA+, as this analysis can provide insights as to whether the differences are unique to LRRK2 parkinsonism or are rather more a reflection of asyn aggregates status.Some differences in CSFasynSAA-and CSFasynSAA+ cases persisted, indicating that the differences may not be unique to LRRK2, though LRRK2 may still mediate some of these differences.
Prior studies have suggested that individuals with LRRK2-parkinsonism may be less likely to demonstrate motor complications compared to sporadic PD cases, especially among females 33 .However, in those studies biologic characterization was not available 33 .Our findings indicate that the more benign phenotype in LRRK2-associated parkinsonism may be driven by asyn-negative cases, A comparison of LRRK2-parkinsonism cases with asyn aggregates to sporadic cases with asyn aggregates is needed to determine the influence of the pathogenic variant itself on phenotype among those with asyn aggregates, and this analysis is underway in the PPMI cohort.

Differences in non-motor features in those with vs without evidence for asyn aggregates
In the few available studies that compared clinical features in LRRK2-associated cases according to asyn status, a few clinical differences have been described 2 .Kalia et al demonstrated that among cases of LRRK2-associated parkinsonism, some non-motor symptoms associated with typical sporadic PD such as anxiety, orthostasis, and cognitive changes are more likely in those with evidence of asyn aggregates 2 .
In contrast, we found that the CSFasynSAA-group had greater global cognitive dysfunction, as assessed with MoCA, at baseline.These results should be interpreted with caution given the age, sex, and education differences in the two groups; indeed, results were no longer significant after adjusting for these possible confounders.Similarly, MoCA score was lower in the CSFasynSAA-LRRK2 group compared to the CSFasynSAA+ sPD group, but given the differences in age, sex, education, and disease duration despite frequency-matching, the significance of these results is unclear.Nevertheless, it remains possible that CSFasynSAA-LRRK2 parkinsonism cases are at risk for greater cognitive dysfunction.Given that such cases may be more likely to have tauopathy-mediated neurodegeneration, and it is possible that tau-based neurodegeneration affects cortical structures preferentially leading to greater cognitive impairment.Possibly supporting this hypothesis is the finding that total tau and phospho tau levels were higher in the CSFasynSAA-group, though this finding did not remain significant when adjusting for age.Further, the CSFasynSAA-and CSFasynSAA+ groups progressed similarly in terms of cognitive decline.
There was some indication that the rate of change in autonomic symptoms differed in the CSFasynSAA+ and CSFasynSAA-groups; differences in dysautonomia according to presence of asyn has also been reported by Kalia et al 2 .However, the clinical relevance of the findings in our study is not clear; the overall burden of autonomic symptoms was similar in the two groups and mean group scores did not change substantially over time.

Biomarker differences: higher DAT binding and higher serum nFL
When examining DAT binding quantitatively, the CSFasynSAA-group had higher putamen DAT binding compared to the CSFasynSAA+ group.While the explanation for this is unclear, it may suggest that the neurodegenerative processes in CSFasynSAA-vs CSFasynSAA+ cases differentially affect dopaminergic neurons.
The CSFasynSAA-group had higher serum NfL.Serum NfL is a nonspecific marker of neuro-axonal injury and degeneration that may be abnormal in a range of neurologic disorders including FTD, multiple system atrophy (MSA), amyotrophic lateral sclerosis, stroke, multiple sclerosis, and traumatic brain injury among others 49 .It is higher in individuals diagnosed with the atypical parkinsonian disorders such as MSA and PSP compared to PD 50 .Across diseases, including in individuals diagnosed with PD, DLB, and AD, higher serum NfL is associated with greater cognitive dysfunction and predicts cognitive decline 19,51 .Consistent with this, in our study, the CSFasynSAA-group had lower MoCA at baseline, even after adjusting for age.However, the CSF asyn CSFasynSAA-group did not progress more on cognitive measures over time compared to the CSFasynSAA+ group.It is possible our study was underpowered to detect differences in longitudinal change over just a 4-year follow-up period.Alternatively, distinct biological mechanisms may subserve the progression on cognitive function.

Genotype-phenotype correlations
While the majority of our sample carried the p.G2019S pathogenic variant, 14% had other variants, and there was a predominance of p.R1441G in the CSFasynSAA-group.These . results are consistent with findings from the literature, mainly from neuropathologically examined case series [1][2][3]5,29,52 . Among 2 G2019S cases and 27 cases with other LRRK2 variants, the majority of G2019S carriers, 70-80%, have Lewy bodies, whereas only 40-45% of other LRRK2 variants do 1 .In the original family in which the LRRK2 locus was identified as being associated with parkinsonism 53 , and in the few subsequently examined cases now known to the I2020T variant, the pathology demonstrated pure nigral degeneration in the absence of Lewy bodies or neurofibrillary tangles in about 50% of cases.Tau pathology also varies according to genotype; 90% of neuropathologically examined G2019S LRRK2parkinsonism cases have tau pathology compared to 38% of cases with other variants 5 .
Importantly, among individuals carrying the same variant, even within a family, clinical and neuropathologic phenotypic variation exists 54 .For example, in a kindred of 4 cases with R1441C mutation, parkinsonism and nigral cell loss with depigmentation and gliosis of the substantia nigra pars compacta, 2 had Lewy bodies and 1 did not have asyn pathology but had neurofibrillary tangles.One hypothesis is that the specific genetic changes alter LRRK2 protein function differently.The p.G2019S and p.I2020T variants are in the kinase domain, and are known to increase LRRK2 kinase activity 5 .On the other hand, the p.R1441G/C/H/S pathogenic variant are in the ROC domain, and increase LRRK2 kinase activity by affecting GTPase function 5 .Pathogenic variants in parts of the gene that encode any of the 3 core catalytic domains of the LRRK2 protein, namely the Roc, COR, or kinase, can be associated with nigral degeneration without asyn pathology.However, available data indicate that p.R1441C/G/H, p.Y1699C and PI2020T are more likely than G2019S to be asyn negative 2,7,55 .In rare cases, pathological findings are consistent with MSA 7 .

Genetic modifiers: comparison of PD risk variants in those with vs without evidence of asyn aggregates
To identify possible genetic underpinnings associated with CSF asyn SAA status, we compared PD risk variants in the groups.Previous research has estimated the heritability of PD at 22%, with PRS explaining approximately a quarter of this heritability within the European population 16 .Furthermore, PRS has been linked to an elevated risk of PD in carriers of the LRRK2 p.G2019S mutation, particularly noting a stronger association in cases of early-onset LRRK2-parkinsonism 44 .Interestingly, variants in MAPT 5 have been reported to increase risk of PD in LRRK2 variant carriers.Other genetic variants that may modify .manifestations of LRRK2 are in the DNM3 and VAMP4 genes 5,56 .Our investigation aimed to ascertain whether a correlation exists between mPRS and CSF asyn SAA status among LRRK2 parkinsonism cases.We did not find differences in mPRS between the groups, nor in the aforementioned genes.However, in analysis of individual risk variants, 3 were identified as possibly associated with CSF asyn SAA status, rs11557080, rs12951632, and rs6808178.
Although they would not withstand correction for multiple testing, the variant rs11557080, located in the 3' UTR of RAB29, is of particular interest due to previous studies suggesting an interaction between RAB29 and LRRK2 activity 57 .
Many studies that have investigated genetic modifiers in LRRK2-associated parkinsonism did not account for underlying pathology.In future studies, stratification of manifest cases according to evidence of asyn aggregates may yield new insights.

Study Limitations
Strengths of this study include the large sample size, in vivo assessment of asyn with a robustly validated assay, and extensive clinical and biomarker characterization of the sample longitudinally.We limited our analysis to 4-years follow-up and are not able to draw conclusions on longer-term differences in the two groups.The biomarkers compared in this analysis reflect currently available analytes in the PPMI study.They provide limited insight into potential pathogenic mechanisms that may or may not diverge in LRRK2 parkinsonism CSFasynSAA-vs SAA+ cases.However, PPMI has a comprehensive biofluids repository that will further allow exploration of other biomarkers as they are validated.Due to the small sample size of participants with non-G2019S variants, we cannot draw conclusions regarding genotype-phenotype differences.

Conclusion
In this study, we have demonstrated several characteristics that are different in LRRK2 associated parkinsonism cases with vs without evidence of asyn aggregates in the CSF.
LRRK2-associated parkinsonism cases without asyn aggregates are more likely to be female, .
normosmic, to have relatively milder motor manifestations, and to exhibit less functional decline.They may also exhibit greater cognitive impairment.We demonstrate important biomarker differences including less loss of DAT binding and higher serum NfL in the CSF asyn negative group.The PPMI cohort is being characterized with extensive proteomics and transcriptomics data.This will allow investigation of differences in various biologic processes in LRRK2-associated parkinsonism asyn positive and negative cases in the future.60 [38-90]  47 [28-80]  0.216 0.472 # DAT SBR lowest putamen=dopamine transporter specific binding ratio, percent expected for age and sex, lowest of the right or left putamen values Missing data: DAT SBR lowest putamen, n=8 (5.4%); CSF abeta, n=9 (6.1%); CSF tau and ptau, n=7 (4.7%); Serum NfL: n=27 (18.2%);Urine BMP, n=14 (9.5%).
!Scores were imputed with their upper and lower limits of detection.The Wilcoxon Rank sum test and Chi-squared test (or Fisher's exact test when at least one expected cell count is below 5) were used to compare SAA-vs SAA+ groups for continuous and categorical variables, respect Linear Regression and Logistic regression models using a-syn SAA as predictor of outcome and adjusting for age were used for continuous and categorical variables, respectively.#Model results based on the log transformation of the outcome.& Model results based on the ranking of the outcome.10) 90 ( 9) 89 ( 11) 89 ( 12) 88 ( 12) SAA-92 ( 7) 91 ( 8) 87 ( 18) 90 ( 11) 88 (13)  Epworth Sleepiness Scale : Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease-Rating Scale RBD: Rapid Eye Movement Behavior Disorder SAA: Seed amplification assay SBR: Specific binding ratio SCOPA-AUT: Scales for Outcomes in Parkinson's-Autonomic SNP: Single-nucleotide polymorphism sPD: Sporadic Parkinson's Disease .SPECT: Single-photon emission computed tomography STAI: State and Trait Anxiety Scale TDP-43: TAR DNA-binding protein 43 UPSIT: University of Pennsylvania Smell Identification Test.
. Medication OFF part 3 scores were missing on a substantial portion of participants and only medication ON state scores are included in this analysis -Modified Schwab and England -Cognitive assessment: the Montreal Cognitive Assessment (MoCA) and the following neuropsychological tests were administered to assess the respective specified domains: Hopkins Verbal Learning Test-Revised (HVLT-R) 10 ; visuospatial function: Benton Judgment of Line Orientation 15-item (split-half) version 11 , and executive function along with working memory: Letter-Number Sequencing and semantic (animal) fluency 12 .Published norms were applied, as referenced.-Psychiatric assessments: Geriatric Depression Scale-15 item (GDS-15), State and Trait anxiety scale (STAI), Epworth Sleepiness Scale (ESS), Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease-Rating Scale (QUIP-RS) -Other non-motor: REM sleep behavior disorder questionnaire; possible RBD defined as RBDSQ≥6, Scales for Outcomes in Parkinson's-Autonomic (SCOPA-AUT) -Olfactory function is assessed with the 38 item University of Pennsylvania Smell Identification Test (UPSIT).Hyposmia is defined as UPSIT score in the ≤15 percentile expected for age and sex 13 .
licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)The copyright holder for this preprint this version posted July 22, 2024.; https://doi.org/10.1101/2024.07.22

Table 2 shows
baseline motor and non-motor measures.Despite being older, having similar duration since clinical diagnosis at baseline assessment, and having significantly lower LEDD (median [IQR] 205 [100-385] vs 500 [300-765], p<0.001), the LRRK2 CSFasynSAA-group had similar scores to the LRRK2 CSFasynSAA+ group in MDS-UPDRS total score and subscores, including part III ON score (median [IQR] 16 [IQR]lowest putamen DAT SBR expected for age and sex in the LRRK2

in Motor and Non-Motor Features and Biomarkers .
Raw mean values at each follow-up time point in the CSFasynSAA+ and CSFasynSAAgroups are shown in Table4.Results of the LMMs are shown in supplementary table 1.The MDS-UPDRS II score did not significantly change over time in the CSFasynSAA-group Longitudinal Change

Table 2 :
Comparison of motor and non-motor features in LRRK2 parkinsonism CSF asyn SAA-and SAA+ cases The Wilcoxon Rank sum test and Chi-squared test (or Fisher's exact test when at least one expected cell count is below 5) were used to compare SAA-vs SAA+ groups for continuous and categorical variables, respectively.Linear Regression and Logistic regression models using a-syn SAA as predictor of outcome and adjusting for age were used for continuous and categorical variables, respectively.* Model results based on the square root transformation of the outcome.# Model results based on the log transformation of the outcome.& Model results based on the ranking of the outcome.

Table 3 :
Comparison of imaging and biofluid biomarkers at baseline in LRRK2 parkinsonism CSF asyn SAA-and SAA+ cases

Table 4 :
Longitudinal assessment of motor and non-motor features It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.