Effects of estrogen and progesterone on neuroactive steroids and cytokines in patients with suicidality

Background: In ovulating psychiatric patients experiencing suicidality, suicidal ideation (SI) often peaks perimenstrually. Our recent double-blind, placebo-controlled, crossover randomized clinical trial (RCT; NCT03720847) showed that perimenstrual administration of estradiol and progesterone (EP) can prevent this peak in SI and depressed mood. In this pre-registered follow-up analysis, we studied how the menstrual cycle and experimental manipulation affected two neurobiological systems associated with the menstrual cycle and suicide risk: GABAergic neuroactive steroids (NAS) and peripheral cytokines. Methods: In 26 psychiatric outpatients with natural menstrual cycles and past-month SI, we analyzed serum samples from three blood draws (midluteal, perimenstrual, midfollicular) per experimental condition (EP vs placebo) timed to a luteinizing hormone-surge ovulation test. Using gas chromatography/mass spectrometry (GC/MS), we measured the progesterone (P4)-derived pregnane NAS (3 & alpha;,5 & alpha;)- 3-hydroxypregnan20-one (3 & alpha;,5 & alpha;THP), (3 & alpha;,5 & beta;)- 3-hydroxypregnan-20-one (3 & alpha;,5 & beta;-THP), (3 & alpha;,5 & alpha;)- 3,21-dihydroxypregnan-20-one (3 & alpha;,5 & alpha;THDOC), (3 & alpha;,5 & alpha;)- 3-hydroxyandrostan-17-one (3 & alpha;,5 & alpha;-A), the androstane NAS (3 & alpha;,5 & beta;)- 3-hydroxyandrostan-17one (3 & alpha;,5 & beta;-A), (3 & alpha;,5 & alpha;,17 & beta;)-androstane-3,17-diol (3 & alpha;,5 & alpha;-A-diol), (3 & alpha;,5 & beta;,17 & beta;)-androstane-3,17-diol (3 & alpha;,5 & beta;-A-diol), and their precursor pregnenolone. High sensitivity multiplex assay kits quantified peripheral cytokines IL-1 & beta;, IL6, and TNF-& alpha;. Results: P4-derived NAS fluctuated in parallel with P4 and increased with exogenous perimenstrual administration of EP. Conversely, androstane NAS either did not fluctuate or fluctuated inversely from P4, and these NAS decreased with exogenous EP. Peripheral cytokines did not show cyclical patterns, but each significantly predicted SI, depressed mood, or anxiousness. Concomitant SSRI medication use predicted lower androstane NAS. Conclusions: While preliminary and exploratory, our findings provide critical descriptive context for future studies. Further, our work presents menstrual cycle-related patterns for ten frequently-studied biomarkers, allowing for improved quality of comparisons involving naturally-cycling populations in research.