Healthy control, de novo PD, and MCI-LB subjects were enrolled between November 2018 and April 2023 at the University of Alabama at Birmingham. The study was approved by the Institutional Review Board at UAB, and full written informed consent was obtained on each participant.

Nine subjects with MCI-LB were enrolled in this exploratory arm of the Alabama Udall study. Criteria for MCI-LB subjects were based on the DSM-5 definition of Probable Mild Neurocognitive Disorder with Lewy Bodies (16) and the McKeith et al. (13) criteria. Inclusion criteria included at least two of the following five core features (fluctuating cognition; visual hallucinations, stereotyped illusions, and/or recurring delusions; REM sleep behavior disorder, Parkinsonism; and impaired visuospatial performance). Exclusion criteria for MCI-LB subjects included clinical diagnosis of moderate or greater dementia or treatment with dopaminergic therapies for Parkinsonism.

64 subjects with newly diagnosed PD and 70 age and sex-matched healthy controls (HC) were enrolled in the Alabama Udall cohort. Inclusion and exclusion criteria for HC and PD subjects have been previously described (14). Briefly, criteria for PD participants included PD diagnosis using UK Brain Bank criteria (17) of less than 2 years at time of enrollment, no previous treatment with PD medications, Hoehn and Yahr stage I-III, and minimum age of 40. Diagnosis with autoimmune or inflammatory disorders, active treatment with immunosuppressants, and serious comorbidity potentially affecting study participation were exclusion criteria for all subjects in the study.

After consent, participants underwent an extensive battery of clinical assessment over one or two visits. Clinical assessments for all subjects included demographics, medical history, immune disorder questionnaire, vaccination history, medication use, family history, behavioral history, PD Screening Questionnaire, vital signs, neurological examination, Epworth Sleepiness Scale, Hamilton Anxiety and Depression scales, Movement Disorder Society-United Parkinson’s Disease rating scale (MDS-UPDRS), Modified Schwab and England Activities of Daily Living, Montreal Cognitive Assessment (MoCA), Parkinson’s Disease Quality of Life Questionnaire (PDQ-39), Rapid Eye Movement Behavior Disorder Questionnaire, Pittsburgh Sleep Quality Index (PSQI), Scales for Outcomes in Parkinson’s Disease – Autonomic Dysfunction (SCOPA-AUT), and University of Pennsylvania Smell Identification Test (UPSIT). Subjects also participated in comprehensive cognitive testing meeting the Level II diagnostic mild cognitive impairment (MCI) assessment criteria recommended by the MDS task force (15). A minimum of two tests were performed in each of seven cognitive domains: attention, language, verbal memory, visual memory, executive function, visuospatial ability, and processing speed. The tests performed were Wechsler Adult Intelligence Scale - Fourth Edition (WAIS-IV) Digit Span and Letter-Number Sequencing (18); Hopkins Verbal Learning Test - Revised (HVLT-R) (19); 10/36 Spatial Recall Test (20); Judgment of Line Orientation (JLO) (21); Hooper Visual Organization Test (22); Boston Naming Test (BNT) (23); Animal Naming (24); Delis-Kaplan Executive Function System (DKEFS) Color/Word Interference (25); and Trail Making Test (TMT) (26). For each cognitive domain, normally distributed z-scores were calculated using the best available normative mean and standard deviation. Normative groups were stratified by age and education whenever possible. The cognitive composite score is the average of all domain z-scores.

Blood samples from each participant was used for flow cytometry studies, complete blood count, plasma collection, and DNA isolation. Unused plasma and DNA samples were banked at the UAB Center for Clinical and Translational Science (CCTS) Sample Processing and Analysis Network (SPAN) and at the NINDS BioSEND program.

Plasma cytokine and chemokine levels from participants was determined with the V-PLEX Proinflammatory Panel 1 and the V-PLEX Chemokine Panel 1 using the MSD SECTOR Imager 2,400 at the UAB Diabetes Research Center Human Physiology Core, as previously described (14).

Immune cell profiling by flow cytometry was performed as previously described (14). 500 μL of fresh peripheral blood was blocked with Human TruStain FcX (BioLegend, San Diego, CA) to prevent non-specific Fc receptor binding. Red blood cells (RBCs) were lysed using RBC Lysis Buffer (BioLegend San Diego, CA). After washing, cells were fixed with 2% paraformaldehyde, and cell viability was assessed using the Aqua Live/Dead Kit (ThermoFisher Scientific). Cells were labeled with specific fluorochrome-conjugated monoclonal antibodies to immune cell surface markers. Antibodies used were from BioLegend: anti-CD45 Pacific Blue (clone HI30); anti-CD3 Brilliant Violet 605 (clone OKT3); anti-CD4 PE (clone OKT4); anti-CD4 eFluor 450 (clone OKT4); anti-CD8α FITC (clone HIT8a); anti-CD14 FITC (clone HCD14); anti-CD16 APC (clone 3G8); anti-CD19 Brilliant Violet 650 (clone HIB19); anti-CD27 Brilliant Violet 421 (clone O323); anti-CD38 PerCP-Cyanine5.5 (clone HB-7); anti-CD45RA APC (clone HI100); anti-CD56 BUV395 (clone 14. G2a); anti-CD62L APC/Cyanine7 (clone DREG-56); anti-CD66b PE (clone G10F5); and anti-CCR7 PE-Cy7 (clone G043H7). Cells were incubated with antibodies (diluted at 1:100) for 20–30 min at room temperature. Flow cytometry was performed using a FACSymphony (BD Biosciences), and immune cell subset analysis was conducted with FlowJo software (Tree Star, Inc., Ashland, OR).

Study sample characteristics were described using means and standard deviations for continuous variables and frequencies (percentages) for categorical variables. Differences between groups for the entire study sample were analyzed using independent samples t tests (two groups) or one-way ANOVA (three groups) for continuous variables and Pearson chi-square tests (or Fisher’s exact tests) for categorical variables. For cytokine and chemokine analyses, samples with levels below the detectable range were imputed as a value of equal to half the limit of detection (LoD) in the analysis. Sensitivity analyses were also performed where these values were excluded, and where these values were imputed with 0 and with a value equal to the LoD. Differences in neuropsychological tests (cognitive composite and domain z-scores) and flow cytometry measurements between MCI-LB and PD subjects were examined using linear regression analysis. For flow cytometry measurements and blood counts, due to concerns about effects of sample size imbalance on estimates, we also performed sensitivity analyses using 1:1 matching between MCI-LB and HC subjects for analysis, based on an exact match for gender and nearest neighbor match on age (Supplementary Table 6), and using nonparametric tests (Supplementary Table 7). In keeping with the exploratory nature of our study, our primary analysis did not correct for multiple comparisons, but sensitivity analyses using a Hochberg step-down procedure were also performed (27). Statistical analyses were performed using version 4.2.0 Patched of the R programming environment (28). Matched samples were identified using version 4.5.5 of the MatchIt package for R (29).

Clinical data have been uploaded to the NINDS Data Management Resource. DNA, plasma, and CSF samples have been stored at the NINDS BioSEND repository. Additional data are available to investigators upon request.

28 potential subjects for the MCI-LB arm were referred for this study between September 2020 and November 2022 (Figure 1). 25 were eligible after screening. Out of the eligible candidates, 16 refused to participate or did not respond to calls for scheduling visits. Nine subjects ultimately consented and enrolled in the study. Enrollment of the main Alabama Udall cohort of HC and PD participants has been previously described (14).

The nine MCI-LB subjects were compared to the 70 HC and 64 de novo PD subjects enrolled in the broader Alabama Udall cohort. Baseline demographics (Table 1), medical history (Supplementary Table 1), and family history (Supplementary Table 2) were generally comparable between groups. The MCI-LB group showed a higher percentage of disabled employment status (Table 1). This group also showed a higher rate of renal disease and COVID-19 infection (Supplementary Table 1). The higher rate of COVID-19 infection is likely related to the fact that these subjects were enrolled after the onset of the COVID-19 pandemic, while a large percentage of the HC and PD participants were enrolled prior to the spring of 2020, when the COVID-19 pandemic started in the United States. With regard to family history, the MCI-LB group also showed a higher rate of family history of PD and schizophrenia (Supplementary Table 2). Vaccination history at baseline also differed, with higher rates in the MCI-LB group (Supplementary Table 3), likely related to enrollment of these participants after the COVID-19 vaccination was available.

As expected, there were significant differences between HC, PD, and MCI-LB participants with regard to clinical assessment scales, including the UPSIT, Schwab and England Activities of Daily Living, Hamilton Anxiety and Depression, SCOPA-AUT, and MDS-UPDRS scores (Table 2). Both PD and MCI-LB groups showed lower UPSIT scores and Schwab and England ADL scores and higher SCOPA-AUT and UPDRS scores compared to HC.

We next evaluated the cognitive profile of HC, PD, and MCI-LB participants. MCI-LB subjects had the lowest MoCA scores (Table 2). Differences in cognitive composite scores and in individual cognitive subdomains were observed among the three groups (Table 3). Linear regression analysis revealed that MCI-LB participants did more poorly in attention, verbal memory, general memory, language, processing speed, and executive subdomains compared to PD participants (Table 3). These differences suggest that the criteria used to recruit MCI-LB subjects were able to distinguish these MCI-LB subjects from de novo PD subjects, who have less cognitive impairment compared to subjects with early DLB.

We examined the percentages of innate immune cells monocytes (CD56⁻CD66b⁻CD3⁻CD19⁻ cells), NK cells (CD3⁻CD19⁻ cells), and neutrophils (CD66b⁺ cells) and adaptive immune cells (CD4⁺ T-cells, CD8⁺ T-cells, and CD19⁺ B-cells) in the MCI-LB subjects and HCs to determine whether immune changes are observed in early stages of DLB. We initially compared all 70 HC to the MCI-LB cohort. Additionally, given the mismatch in the number of participants in the two groups, we repeated our analysis in which we compared the MCI-LB cohort to a similar number of age- and sex-matched HC participants. With regard to complete blood counts and differential, MCI-LB subjects showed higher total neutrophil count compared to all HC in unadjusted analyses, although this difference was not statistically significant after controlling for multiple comparisons or in the matched HC comparison (Supplementary Table 4). Of note, one MCI-LB blood sample was not able to be processed for CBC. Plasma cytokines and chemokines did not differ between MCI-LB and HC groups (Supplementary Table 5).

We did observe several differences in immune cell populations as determined by flow cytometry. The most dramatic change in MCI-LB participants was a ~ 60% reduction in mature neutrophils when compared to all HC subjects (Table 4) or matched HC subjects (Supplementary Table 6; Figure 2). MCI-LB participants also showed a significant increase in CD56⁻CD16⁺ NK cell subset 1 when compared to all HC with correction for multiple comparisons (Table 4). This difference was also observed when MCI-LB participants were compared to matched HC participants, but was no longer statistically significant after adjusting for multiple comparisons in the matched analysis (Supplementary Table 6; Figure 2).

Sensitivity analysis examined the immune cell profiles controlling for covariates including COVID infection, COVID vaccination, and renal disease as factors that were significantly different between the HC and PD groups. We found that mature neutrophils remained significantly different between the two groups (Supplementary Table 8). Interestingly, in this analysis, we also observed an increase in regulatory B-cells (Bregs) in the MCI-LB group compared to HC when including these covariates in the analysis (Supplementary Table 8).