👤 Lynn Marie Trotti

Neuroprotective properties of estrogen have poorly translated to reduced neurodegeneration in clinical trials of systemic estrogen replacement therapy. To more directly assess biological processes associated with brain estrogen (estrone, estradiol) levels, we recruited 81 women (42 non-white) and 28 men (13 non-white) for cerebrospinal fluid (CSF) hormone, targeted proteomic, and volumetric brain analysis. In the mostly post-menopausal women, we found CSF estrogen levels to only modestly correlate with their corresponding plasma levels, and were additionally influenced by body mass index or age. CSF estrone was also correlated with a marker of Alzheimer’s disease (AD) neuropathologic change (CSF Aβ42/Aβ40), but this was not the case for the more biologically active CSF estradiol. Aptamer-based proteomic analysis of 1,075 CSF markers for inflammation, proteolysis, signaling, and DNA/RNA regulation revealed CSF estrogen levels to associate with alternative complement pathway proteins, and shifts observed in AD (apoE, RAGE). Parallel MRI analysis correlated higher CSF estrogen with smaller volumes of the brain somatosensory and posterior-medial networks without influence from cognition or neurodegeneration. Analysis using plasma estrogens only partially reproduced CSF estrogens’ biochemical correlates but provided inconclusive relationships with brain volume correlates. These findings highlight the association between CSF levels of the more biologically active estradiol and CSF inflammatory pathways involving AD risk genes as potential mechanisms linking hormone status to AD risks, and suggest caution in using CSF estrone or plasma estrogens when interpreting treatment or preventive studies. The online version contains supplementary material available at 10.1186/s12974-025-03657-3. Show less

Restless legs syndrome (RLS) is a neurological disorder with a strong genetic susceptibility. A painful RLS sub-phenotype has been described previously but the neurobiological basis for this phenotypic variant remains unknown. This study investigated whether any of the six initially discovered genomic loci associating with RLS (BTBD9, MEIS1, PTPRD, MAP2K5/SKOR1, TOX3, and an intergenic region on chromosome 2), were more strongly associated with complaints of painful versus non-painful RLS. RLS patients (N = 199; Age = 53.1 ± 16.8; 100% Caucasians; 57% women) diagnosed clinically were genotyped for known variants associating with RLS. Definition of painful RLS required that subjects selected "painful" from a list of 14 adjectives to describe their RLS sensory experience and answered positively to a separate question that queried specifically as to whether they perceived their RLS sensations as painful. Genotype association tests employed logistic regression analyses with assumption of an additive genetic model. Analyses were performed using PLINK software v1.07. We identified two RLS patient subgroups: a painful (n = 41) and non-painful (n = 158). Among 10 tested SNPs, only rs3104767 (related to the TOX3 gene locus) was more associated with painful RLS. The minor allele T of SNP rs3104767 was associated with an increased risk of RLS being perceived as painful with an OR of 1.67 [CI = (1.01-2.74); p = 0.049]. Notably, this minor T allele associated with pain sensation in RLS patients in this study was the non-risk allele for RLS in the original RLS genome wide association study, but a similar trend was observed in a recent Parkinson disease sample study. This study might suggest the TOX3 gene variant as a potential genetic substrate for the painful RLS sub-phenotype. This was an exploratory small study and correction for multiple comparisons would have rendered the results not significant. Therefore, the above findings require replication in larger clinical as well as population-based samples of RLS subjects. Show less