Cholinergic Nucleus 4 Degeneration and Cognitive Impairment in Isolated Rapid Eye Movement Sleep Behavior Disorder

Cholinergic nucleus 4 (Ch4) degeneration is associated with cognitive impairment in Parkinson's disease and dementia with Lewy bodies, but it is unknown if Ch4 degeneration is also present in isolated rapid eye movement sleep behavior disorder (iRBD).

Rapid eye movement (REM) sleep behavior disorder (RBD) is characterized by the loss of muscle atonia during REM sleep, which results in dream enactment behavior. Isolated RBD (iRBD) is associated with α-synuclein pathology and is recognized as a precursor to other α-synucleinopathies, especially Lewy body diseases, with 45% progressing to Parkinson's disease (PD) and 45% progressing to dementia with Lewy bodies (DLB) after 15 years of follow-up. 1 Consistent with PD and DLB, individuals with iRBD also have cognitive impairment. Compared to controls, studies report impairments in memory, visuospatial function, attention, and executive function. [2][3][4] In one of these studies, half of iRBD patients met the criteria for mild cognitive impairment (MCI), 3 and individuals with iRBD and MCI are at increased risk of developing dementia. 5 One of the major sources of cognitive impairment in PD and DLB is degeneration of the cholinergic basal forebrain, especially the nucleus basalis of Meynert, the major constituent of cholinergic nucleus 4 (Ch4). 6,7 Measurement of cholinergic basal forebrain nuclei volumes using probabilistic cytoarchitectonic maps applied to magnetic resonance imaging (MRI) has demonstrated reduced Ch4 volumes in DLB and in PD with cognitive impairment, 8,9 but it is unknown if Ch4 degeneration is also present in iRBD.
The aim of this study was to determine if there is evidence of cholinergic basal forebrain degeneration, that is, reduced Ch4 volume on MRI, in patients with iRBD compared to healthy controls (HCs) and whether Ch4 degeneration is associated with cognitive impairment in iRBD. Determining whether Ch4 degeneration is present in iRBD and associated with cognitive impairment has implications for the development of therapies for cognitive impairment in this population. We also examined the volumes of cholinergic nuclei 1, 2, and 3 (Ch123) as a comparison region to Ch4. As these cholinergic nuclei project to the hippocampi and the olfactory bulbs, we hypothesized that this region would be less likely to be associated with cognitive impairment in iRBD.

Clinical Assessments
The following assessments were performed at the baseline visit: (1) MoCA (cognitive screener), (2) Hopkins Verbal Learning Test-Revised (verbal memory), (3) letter number sequencing (LNS, working memory), (4) Semantic Fluency Test-animals (SFT-animals, executive function), (5) Symbol Digit Modalities Test (processing speed/attention), (6) Benton Judgment of Line Orientation (JLO, visuospatial skills), (7) Movement Disorders Society Unified Parkinson's Disease Rating Scale (MDS-UPDRS), and (8) REM sleep behavior disorder questionnaire (RBDSQ). In the iRBD and control cohorts an RBDSQ score ≥5 was used as the cutoff consistent with the diagnosis of RBD. In the iRBD cohort, PD-MCI level 1 criteria were applied to determine the percentage of participants who met the criteria for MCI. 10

MRI Acquisition and Analysis
anatomical space of the single-subject MNI template into the standard MNI template space by an affine translation along the yand z-axes of 4 and 5 mm, respectively. 21 Quality of the overlay was confirmed by visual inspection. Relative bilateral Ch4 and Ch123 GMD were calculated using a MATLAB script that multiplied the GMD value for each voxel by the weighting contained within the probabilistic map (Fig. S1). 8 To produce standardized GMD values, each relative GMD was divided by the total weighting contained within the probabilistic map. TIV-normalized GMD values were calculated by dividing the GMD value by TIV and multiplying by 1000.

Statistical Analysis
Statistical analysis was performed using Stata IC 14 (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP). We compared demographics, cognitive assessment scores, and GMD values between iRBD patients and HCs using twosample t tests or Mann-Whitney U tests when appropriate. We examined Q-Q plots to determine whether data were normally distributed. Sex distribution was compared between groups using χ 2 test. For comparisons of cognitive test scores between iRBD and HC groups, P-values were corrected for the false discovery rate using the Benjamini-Hochberg procedure. To adjust for the effects of scanner type, we used partial correlation analysis to evaluate the relationship between group (iRBD vs. HC) and TIV-normalized cholinergic basal forebrain GMD, Ch4 and Ch123. We did not need to adjust for the effects of age and sex in these analyses because the groups were well matched ( Table 1).
In secondary analyses we used a backward selection method to investigate the relationship between basal forebrain GMD (independent variable) and cognitive test scores (dependent variable) adjusted for group (HC vs. iRBD), scanner type, and TIV. We did not adjust for age or sex because both were well matched between the groups. Predictors at P < 0.1 were included in the final models. TIV was included in all models that retained basal forebrain regions. Final models were stratified by group. Statistical significance for the final models was set at P < 0.05. Homogeneity of variance was tested using the Breusch-Pagan test.

Results
Of 47 iRBD participants enrolled in the PPMI, only 35 had sagittal T1 MRI sequences at baseline. There were no significant differences between the 35 iRBD participants included in this study and the 12 iRBD participants without MRI scans in regard to age, sex, RBDSQ score, MoCA score, and MDS-UPDRS total score (all P > 0.05, Table TABLE S1). For this group we identified 35 age-and sex-matched HCs. Clinical characteristics for each group are presented in Table 1. Among the iRBD participants, 33 MRIs were performed using a Siemens scanner and 2 were performed using a Philips scanner. Of the HCs, 25 MRIs were performed using a Siemens scanner, five were performed using a Philips scanner, and five were performed using a GE scanner. As expected, individuals in the iRBD cohort had higher mean scores on the RBDSQ compared to HCs. This difference remained statistically significant when the question regarding the presence of nervous system disease was excluded. MDS-UPDRS total score was significantly higher in the iRBD group compared to HCs. Similarly, individual MDS-UPDRS Part I, Part II, and Part III scores were significantly higher in the iRBD cohort compared to HCs (all P-value ≤0.0005).

Discussion
We found that Ch4 GMD is reduced in iRBD participants compared to HCs, whereas Ch123 GMD was not significantly different between the iRBD and HC groups. Our findings are consistent with prior cholinesterase PET studies. In one previous [ 11 C]-donepezil positron emission tomography (PET) study, iRBD participants had reduced neocortical binding compared to controls, 22 consistent with neocortical denervation from Ch4 cholinergic neurons. Another cholinesterase PET study in PD found that RBD was associated with cortical, limbic, and thalamic cholinergic denervation but not dopaminergic or serotonergic denervation compared to participants with PD without RBD. 23 It is possible that reduced Ch4 GMD in iRBD is part of greater regional atrophy as a previous VBM (voxel-based morphometry) study found reduced volumes of the anterior cerebellum, tegmental pons, and left parahippocampal gyrus in iRBD 24 ; however, we did not find a difference in Ch123 GMD between the iRBD group and HCs. This supports the presence of preferential or at least earlier Ch4 degeneration in iRBD. This would be consistent with a prior longitudinal MRI study in PD that showed that Ch4 atrophy preceded Ch1/Ch2 atrophy. 25 In an exploratory analysis, a recent VBM study found reduced left Ch4 GMD in iRBD with MCI compared to controls but did not detect a difference for the right Ch4 or for the iRBD group without MCI. 26 Our study suggests that bilateral reduction in Ch4 GMD is a feature of iRBD regardless of cognitive status, which is consistent with the previous PET studies reviewed earlier.
We found that participants with iRBD performed worse on all cognitive assessments compared to ageand sex-matched HCs except for semantic fluency. Previous studies also reported cognitive impairment in iRBD. 2,27 One study reported worse attention, executive function, and verbal memory in iRBD patients even when adjusted for education and age. 27 A longitudinal prospective study found that patients with iRBD performed worse on cognitive assessments involving memory and executive function at baseline, and their performance worsened over the subsequent 2 years. 2 Ch4 degeneration among our iRBD patients was correlated with impairment in tasks of executive function (SFT-animals) and attention and working memory (LNS). When adjusted for age, sex, scanner type, and TIV in a linear regression model, Ch4 GMD was found to be a significant predictor of performance only on LNS, a task assessing working memory. Prior studies have shown that cholinergic pathways projecting to the cortex play an important role in cognitive function, especially in working memory. 28,29 Moreover, Ch4 GMD has previously been shown to be correlated with performance on LNS task in PD. 9 This study examined bilateral cholinergic basal forebrain GMD in iRBD and is the first to find a relationship between Ch4 GMD and working memory in iRBD. The lack of an association between Ch4 GMD and the other cognitive test scores in iRBD may be because not all iRBD participants had cognitive impairment or this study was underpowered to detect these relationships. Limitations could arise from the use and selection of PPMI subjects. One limitation of the study is that HCs in PPMI were required to have MoCA ≥26, which may have created a larger difference in cognitive function between iRBD and HC than a randomly selected and matched HC group without a minimum MoCA requirement. However, a HC group with a more similar cognitive performance to iRBD would potentially be affected by other age-related pathologies. Therefore, we chose to use the HC group as our comparison group. Although efforts were made to account or adjust for the heterogeneity of scanner types and specifications, this is a limitation of this data set. Another limitation is that we did not adjust for multiple comparisons in our secondary analyses, so these results require future replication. In summary, our study found greater Ch4 degeneration in iRBD patients compared to controls, and Ch4 GMD was associated with performance on a task assessing working memory. These findings support the idea that Ch4 degeneration is associated with cognitive impairment in iRBD. Future longitudinal studies are needed to determine whether iRBD patients with reduced Ch4 volume are at increased risk of future cognitive decline.