Elevated Cerebrospinal Fluid α‐Synuclein Seeding Activity Predicts Central Lewy Body Diseases

Abstract

Background

Cerebrospinal fluid (CSF) α‐synuclein seeding activity (SSA) via a seed amplification assay might predict central Lewy body diseases (LBD) in at‐risk individuals.

Objective

The aim was to assess CSF SSA in a prospective, longitudinal study.

Methods

Participants self‐reported risk factors were genetics, olfactory dysfunction, dream enactment behavior, orthostatic intolerance, or hypotension; individuals who had ≥3 confirmed risk factors underwent CSF sampling and were followed for up to 7.5 years. Participants who developed a central LBD (LBD+) were compared to those who did not. Quadruplicate SSA areas under the curve (AUC) were averaged.

Results

Of 11 subjects with average AUCs above 500,000 units, 7 (64%) developed a central LBD compared to 1 of 20 (5%), with AUCs below the cutoff value (P = 0.0011 by log‐rank test). Conversely, 7 of 8 (88%) LBD+ participants had elevated initial AUCs.

Conclusions

Increased CSF SSA predicts central LBDs. Individuals who develop a central LBD have elevated initial SSA AUCs.

By the time a person at risk for a central Lewy body disease (LBD) such as Parkinson's disease (PD) or dementia with Lewy bodies (DLB) evinces signs of motor or cognitive dysfunction, substantial neuronal damage has likely already occurred. There is great interest currently in biomarkers that can identify central LBDs early enough that neurorescue strategies might delay the onset of symptoms.

Lewy bodies contain an abundance of the protein α‐synuclein (α‐syn), ¹ and it is widely suspected that intraneuronal α‐syn deposition causes or contributes to deficiencies of neurotransmitter catecholamines in the nigrostriatal system and heart, characteristic features of LBDs. ² , ³ , ⁴

Real‐time quaking‐induced conversion (RT‐QuIC) and similar amplification assays exploit the self‐propagating potential of specific protein aggregates to act as seeds that can grow by incorporating monomers of their constituent proteins by seeded polymerization. ⁵ Seeding activities in biospecimens can be amplified by orders of magnitude in multiwell plates with fluorescence readout. Seed amplification assays for α‐syn have detected pathological α‐synuclein seeding activity (SSA) in the cerebrospinal fluid (CSF) even in prodromal synucleinopathies. ⁶ , ⁷

The recently completed prospective, longitudinal PDRisk Study (NIH Clinical Protocol 09N0010, NCT00775853) was designed to determine whether in people with multiple PD risk factors biomarkers of catecholamine deficiency in the heart or brain predict PD during long‐term follow‐up. ⁸ , ⁹ , ¹⁰ Entry into the study was based on 4 categories of risk: genetic, olfactory, dream enactment behavior (as in rapid eye movement behavior disorder), and orthostatic intolerance or orthostatic hypotension. Details of criteria used for confirming the risk factors have been published. ⁸ Analysis of CSF samples from the PDRisk study enabled a prospective evaluation of whether increased SSA predicts central LBDs in at‐risk individuals. We also examined the relationships between CSF SSA and biomarkers of central and cardiac catecholamine deficiency in the same subjects.

Patients and Methods

Data and samples from the NINDS IRB‐approved protocols 03N0004, 09N0010, 17N0076, and 18N0140 were transferred to the secondary research protocol 000490, “Mechanisms of Autonomic and Catecholamine‐Related Disorders,” and the project reported here was approved by the NIH IRB. All the participants were tested at the NIH Clinical Center after they provided written informed consent. Testing at follow‐up was per the protocols of the original studies—mainly NIH Clinical Protocol 09N0010, “Biomarkers of Risk of Parkinson Disease.” For comparison and establishment of normal cutoff values, CSF samples were also assayed from healthy volunteers.

For this report we defined central LBDs in terms of a clinical diagnosis of PD or DLB.

RT‐QuIC assays of SSA were performed at the Rocky Mountain Laboratories (RML). In black 96‐well plates with a clear bottom (Nalgene Nunc International, Rochester, NY, USA) wells were loaded with six 0.8 ‐mm diameter molecular biology–grade silica beads (OPS Diagnostics, Lebanon, NJ, USA). Added to each well was a reaction solution (85 μL) that when combined with CSF (see later) gave final concentrations of 40 mM phosphate buffer (pH 8.0), 170 mM NaCl, 10 μM thioflavin T (ThT), 0.0015% sodium dodecyl sulfate, and 0.1 mg/mL recombinant α‐syn. The recombinant α‐syn, purified at the RML as described previously, ¹¹ was filtered through a 100‐kD molecular weight cutoff filter immediately prior to use.

Fifteen microliters of quadruplicate aliquots of CSF were added to the wells. The plate was sealed with a plate sealer film (Nalgene Nunc International) and incubated at 42°C in a BMG FLUOstar Omega plate reader (BMG Labtech, Cary, NC, USA) with cycles of 1 minute shaking (400 rpm double orbital) and 1 minute rest. ThT fluorescence measurements (450 ± 10 nm excitation and 480 ± 10 nm emission, bottom read) were taken every 45 minutes.

The cutoff value for low CSF 3,4‐dihydroxyphenylacetic acid (DOPAC) was ≤1.22 pmol/mL, for low ¹⁸F‐dopamine‐derived radioactivity ≤6000 nCi‐kg/cc‐mCi, for the putamen/occipital cortex (PUT/OCC) ratio of ¹⁸F‐DOPA‐derived radioactivity ≤2.7, and for the washout percentage of putamen ¹⁸F‐DOPA‐derived radioactivity ≥20%. ⁸ , ⁹ , ¹⁰ , ¹²

Assay personnel of the RML received coded specimens and were strictly blinded until the data were tabulated. NINDS personnel were blinded to the RML SSA testing results until the data were tabulated for analysis.

The area under the curve (AUC) from ThT fluorescence plots was calculated using the trapezoidal method. Means from the quadruplicate AUCs for each CSF sample were plotted using GraphPad Prism (GraphPad Software, LLC, Boston, MA, USA). A cutoff value of 500,000 arbitrary units was chosen based on the data in the healthy volunteers. Data for AUC were compared to those for maximum ThT and lag time. ¹¹ In scatterplots, cutoff values for ThT max and lag time were placed visually based on data from the healthy volunteers.

For statistical testing and graphics, GraphPad Prism for Mac OS 10.1.0 was used. A Kaplan–Meier survival curve was constructed for the groups with increased versus normal average AUC at study entry, with curve comparison using the Mantel–Cox log‐rank test. t‐Tests for independent means were used to compare the group developing a central LBD during follow‐up (LBD+) to the group that did not (LBD–). Receiver operating characteristic (ROC) curves were constructed to compare the LBD+ and LBD– groups.

Statistical analyses were exploratory. There were no adjustments for multiple comparisons or exclusions of outlying data. A P‐value <0.05 defined statistical significance.

Results

Values for the CSF SSA were obtained for 31 at‐risk individuals (8 LBD+, 23 LBD–, Supplementary Data Workbook in Data S1) and 6 healthy volunteers. Among 11 PDRisk study participants with average AUCs above the cutoff value, 7 (64%) went on to be diagnosed with a central LBD (5 PD, 2 DLB); among 20 subjects with average AUCs below the cutoff value, 1 (5%) was diagnosed subsequently with a central LBD (P = 0.0011 using log‐rank test). Survival curve analysis showed that the relative hazard ratio was 5.95 (Fig. 1A).

FIG. 1.

Survival and receiver operating characteristic (ROC) curves for cerebrospinal fluid (CSF) α‐synuclein seeding activity (SSA). (A) Participants were stratified by initial mean SSA area under the curve (AUC) above the cutoff value of 500,000 arbitrary units (red) or below the cutoff value (gray). (B) ROC curve comparing PDRisk study participants who were diagnosed with a central Lewy body disease (LBD) during follow‐up (LBD+) versus those not diagnosed with a central LBD during follow‐up (LBD–). PDRisk study participants with elevated mean initial SSA AUC had a high risk of later development of a central LBD. For comparison of LBD+ versus LBD– groups, there was imperfect specificity due to 4 LBD– participants with elevated initial mean SSA AUCs.

Upon initial evaluation, 88% of the LBD+ group had an average AUC above the cutoff value, in contrast with 17% of the LBD– group (area under the ROC curve = 0.83, P = 0.0068; Fig. 1B).

Group mean AUC of SSA tended to be higher in the LBD+ than LBD− group (Fig. 2A). Increased AUC values were noted in 4 LBD– participants (arrows in Fig. 2A). Review of the medical records did not reveal any medical conditions or drug treatments known to evoke α‐syn oligomerization or misfolding in these participants. The LBD+ group had lower mean cardiac ¹⁸F‐dopamine‐derived radioactivity (Fig. 2B), lower mean CSF DOPAC (Fig. 2C), higher mean putamen washout percentage of ¹⁸F‐DOPA‐derived radioactivity (Fig. 2D), and lower mean PUT/OCC ratio of ¹⁸F‐DOPA‐derived radioactivity (Fig. 2E).

FIG. 2.

Individual values (with means ± SEM [standard error of the mean]) for biomarkers in at‐risk individuals who subsequently developed a central Lewy body disease (LBD+, red) and those who did not (LBD–, gray). (A) SSA mean AUC, average area under the curve for synuclein seeding activity; (B) ¹⁸F‐DA, interventricular septal myocardial ¹⁸F‐dopamine derived radioactivity (nCi‐kg/cc‐mCi); (C) CSF DOPAC, cerebrospinal fluid 3,4‐dihydroxyphenylacetic acid (pmol/mL); (D) ¹⁸F‐DOPA washout %, percentage decrease in putamen ¹⁸F‐DOPA‐derived radioactivity between the peak value and the value in the 15′ static image ending at 120′ after tracer administration; (E) ¹⁸F‐DOPA PUT/OCC, putamen/occipital cortex ratio of ¹⁸F‐DOPA‐derived radioactivity in the 15′ static image ending at 120′ after tracer administration; horizontal dashed lines show cutoff values. Numbers in italics are P‐values for 2‐tailed independent‐means t‐tests comparing the LBD+ versus LBD– groups. Arrows in the panel for SSA mean AUC highlight 4 samples with high values in the LBD– group.

Individual values for the 3 indices of CSF SSA (AUC, lag time, ThT max) were correlated with each other (Fig. S1). For all 3 indices there were 4 LBD– participants with abnormal values (arrows in Fig. S1).

Discussion

In this prospective, longitudinal study, increased CSF α‐SSA identified preclinical central LBDs. About two thirds of participants with elevated average SSA AUCs went on to be diagnosed with a central LBD, in contrast with only 1 who had an average SSA AUC below the cutoff value. Conversely, 88% of the subjects in the LBD+ group had an average AUC above the cutoff value upon initial evaluation, in contrast with 17% of the LBD– group.

Four LBD– participants had increased CSF SSA, meaning imperfect specificity. Prior studies also noted low percentages of some CSF SSA positivity in cases with no known α‐syn pathology. ¹³ , ¹⁴ The bases and clinical significance of such findings remain unclear. Indices of catecholamine deficiency in the brain (CSF DOPAC, PUT/OCC ratio of ¹⁸F‐DOPA‐derived radioactivity, washout percentage of putamen ¹⁸F‐DOPA‐derived radioactivity) and heart (¹⁸F‐dopamine‐derived radioactivity) were strongly predictive, although no single measure had perfect sensitivity and specificity in distinguishing the LBD+ and LBD– groups.

This was a small study of highly selected participants. Generalizability to the overall population of at‐risk individuals with fewer risk factors is unknown. ¹⁸F‐Dopamine PET is available only at the NIH Clinical Center, meaning the data based on this modality cannot be replicated.

In the future, diseases such as central LBDs may be defined by combinations of pathophysiological biomarkers instead of by the clinical findings. Recent movement in this direction has come from a combined biomarker study of patients with neurogenic orthostatic hypotension (nOH). ¹⁵ LB forms of nOH were found to involve cardiac noradrenergic deficiency, olfactory dysfunction, and increased α‐syn‐tyrosine hydroxylase colocalization indexes in skin biopsies. None of these biomarkers alone separated the LB from the non‐LB groups, but combining the data for those variables completely separated LB from non‐LB forms of nOH. Moreover, independently of the clinical diagnosis, the biomarker triad identified a pathophysiologically distinct cluster of nOH patients that corresponded exactly to the LB nOH group. Because in the present study the 4 subjects in the LBD– group with high SSA AUCs had normal values for cardiac ¹⁸F‐dopamine‐derived radioactivity, the PUT/OCC ratio of ¹⁸F‐DOPA‐derived radioactivity, and putamen ¹⁸F‐DOPA washout percentage, combining CSF SSA with other pathophysiological biomarkers might provide a highly specific as well as sensitive test.

Author Roles

(1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical analysis: A. Design, B. Execution, C. Review and critique; (3) Manuscript preparation: A. Writing of the first draft, B. Review and critique

D.S.G.: 1A, 1B, 1C, 2A, 2B, 3A

P.A.: 1A, 1C, 2C, 3B

C.H.: 1C, 3B

P.S.: 1C, 3B

J.G.: 1C, 3B

A.G.H.: 1C, 2C, 3B

B.C.: 1A, 2C, 3B

Disclosures

Ethical Compliance Statement: The NIH IRB approved the study. The patients provided written informed consent before participating in any of the research reported here. We confirm that we have read the journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding Sources and Conflicts of Interest: The research reported here was supported by the Divisions of Intramural Research, NIH, NINDS, and NIAID. The authors declare that there are no conflicts of interest relevant to this work.

Financial Disclosures for the Previous 12 Months: In the previous 12 months David S. Goldstein received a royalty for a book published in 2006. The authors declare that there are no additional disclosures to report.

Supporting information

Figure S1. Scatterplots relating individual values for biomarkers in at‐risk individuals who subsequently developed a central Lewy body disease (LBD+, red) and those who did not (LBD–, gray). (A) Cerebrospinal fluid (CSF) 3,4‐dihydroxyphenylacetic acid (DOPAC) versus mean area under the curve (AUC) for α‐synuclein seeding activity (SSA); (B) interventricular septal myocardial ¹⁸F‐dopamine‐derived radioactivity versus average AUC; (C) interventricular septal myocardial ¹⁸F‐DA‐derived radioactivity versus CSF DOPAC; (D) AUC versus lag time of SSA (hours); (E) AUC versus maximum thioflavin T (ThT max); (F) ThT max versus lag time. Linear regression lines of best fit with 95% confidence intervals are shown. Numbers in italics are correlation coefficients and P‐values across all subjects. Individual values for the 3 indexes of CSF SSA were correlated with each other. Arrows indicate 4 LBD– participants with abnormal values for CSF SSA.

Data S1. Supporting information.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.