Psychobiology and Treatment of Perimenopausal Mood Disorders

National Institute Of Mental Health

Investigators

Linked publications & trials

Abstract

This report includes work arising from the following clinical protocols: NCT00030147, NCT00060736, NCT00001231, NCT03689543 and NCT00001322. We completed a prospective, longitudinal study of healthy asymptomatic women followed every 6-8 months during their transition through the menopause. At each semi-annual clinic visit, women were evaluated with diagnostic interviews and ratings including quality of life, experience of life events, marital satisfaction and social support. The presence of depression was monitored through clinical assessment, self-report, and review of standardized mood rating scales. Primary outcome measures included SCID-confirmed depressive episodes (major or minor), hot flushes, and circulating hormone levels. Depressive episodes were characterized in relation to the participant's STRAW stage, hot flush onset, and final menstrual period (FMP). Blood samples obtained at each 6-month clinic visit were assayed for several steroid hormones and gonadotropins by liquid chromatography tandem mass spectroscopy (LC-MS/MS) and chemiluminescent assays (respectively) under batched conditions within each woman, as well as pairing women who developed PMD with those who remained asymptomatic. Eighty-eight asymptomatic, premenopausal women, ages 41-55 years, were monitored longitudinally for an average of 6.1 years (range of 2-13 years) until 6-12 months after their FMP. Our results document the clustering of depressive episodes in proximity to the FMP. We prospectively identified 29 episodes of depression, affecting twenty-two women. Twenty of these episodes occurred within the 24 months surrounding the FMP (Figure 1). Over 80% of episodes (82.8%) occurred in the context of elevated mid-follicular FSH levels (defined as > 14 IU/L). Approximately 70% of women experiencing depression during the study did so for the first time in their lives. Ten of the depressive episodes (34.5%) met criteria for major and nineteen (65.5%) met criteria for minor depression. We compared the patterns of secretion of serial blood hormone levels in women who developed PMD and those who remained euthymic to characterize and compare the kinetics of hormone measures prior to the FMP. Anchoring hormone levels to each woman’s FMP, enabled direct comparisons of E2 and FSH trajectories between women who developed PMD and those women who remained euthymic. Linear mixed effects models were employed to examine the pattern of hormone secretion within each group of women. As women approached their respective FMPs, the patterns of their plasma E2 and FSH levels were best represented by quadratic equations. All women experienced an initial increase in mean E2 levels, paralleled by a reciprocal decrease in mean FSH levels, during the early perimenopausal stages compared with premenopausal levels. This was followed by a transition into an E2 withdrawal (and FSH rise) phase. Women with and without PMD did not differ in average or overall levels of E2 and FSH. However, we did identify differences in the rates of change for both E2 and FSH levels, which were higher in women with PMD (i.e., steeper slopes of the serial measures) compared to those who remained asymptomatic (p=0.04 and p=0.004, E2 and FSH, respectively). In the whole cohort, the inflection point for E2’s peak level and decline occurred approximately five years prior to the FMP, but in women who developed PMD, the peak occurred roughly eight months closer to the FMP (i.e., later) compared to those who remained asymptomatic. Several cross-sectional and longitudinal studies have also examined circulating androgen levels in relation to PMD. We controlled for baseline BMI (since BMI is physiologically linked to circulating androgen levels40), and we observed no differences in circulating testosterone or DHEA/S levels between women with and without PMD. Notably, PMD was associated with higher pregnenolone levels throughout the MT potentially consistent with our functional genomics studies in cell lines from women with PMD (see Research Focus #2), in which we observed trait-like (i.e., hormone-independent) alterations in expression of CYP7B1 (the enzyme metabolizing pregnenolone to the neurosteroid 7α-hydroxypregnenolone) in PMD compared with controls. These observed differences in the E2 and FSH secretory patterns and the circulating levels of pregnenolone provide important insights into the possible ontogeny of PMD. These differences in endocrine secretions emerged before the onset of depressive symptoms, suggesting that a steeper and later E2 withdrawal prior to the onset of depression may be a key factor in PMD vulnerability. Clinically, our findings reinforce the potential benefits of interventions that stabilize E2 dynamics. Initiating estradiol therapy during the five-year window prior to the FMP—when women often begin to experience irregular or skipped menstrual periods—may be beneficial. Mechanistically, our results underscore the importance of E2 withdrawal in PMD etiology, although the precise nature of this relationship requires further investigations in future studies. Finally, we will employ this longitudinal sample to evaluate potential markers of PMD including several measures identified in our functional genomics studies. In a separate clinical trial, we attempt to dissect possible receptor signaling mechanisms by which E2 withdrawal (or acute E2 deficiency) destabilizes affective state in women with PMD. The tissue effects of E2 primarily occur through activation of two receptor subtypes, often with opposing outcomes: ER alpha and ER beta. We focus on ER beta because it is reported to mediate both the effects of E2 on the serotonergic system and the antidepressant-like effects of E2 in the forced-swim test. Moreover, selective agonists of ER beta attenuate the behavioral and hypothalamic-pituitary-adrenal (HPA) axis response to stress in animal models. We employ the ER beta receptor agonist LY500307, and a similar design as described above (double-blind, placebo-controlled E2 withdrawal), in asymptomatic postmenopausal women with past PMD (who reported beneficial effects of estradiol therapy on their PMD). Based on our previous work, approximately 70% of the women with past PMD will develop mood symptoms during E2 withdrawal (i.e., placebo). Women with past PMD receive an open-label standard dose (100 mcg per day) of transdermal estradiol (TE). After three weeks of TE, all women were randomly assigned under double-blind conditions to one of three medication arms: 1) LY500307 high dose (75 mg per day), 2) LY500307 low dose (25 mg per day), or 3) placebo (i.e., E2 withdrawal). We employ the higher 75 mg dose to ensure that adequate blood levels of LY500307 are achieved to enhance the likelihood of this compound penetrating the CNS (with the caveat that the specificity of LY500307 for ER beta is reported to decrease with higher doses – although the dosage in women above which ER alpha is activated has not been established). Our study hypothesis is that in women with a past PMD, affective symptoms consequent to experimentally induced E2 withdrawal will be prevented by an ER beta agonist but not by placebo. In forty-six postmenopausal women with a past PMD, our findings suggest that LY500307 alone does not have the same beneficial effects on PMD as TE. Women with a past PMD who were crossed over from TE to placebo experienced an increase in depression symptom severity (CES-D and HDRS p<0.001), similar to our findings in the original E2 withdrawal study. Women with PMD who were randomized to either dose of LY500307 did not experience a significant recurrence of mood symptoms, and CESD/HDRS scores were not different from those on TE. However, symptom severity scores during double-blind conditions did not differ between the randomization groups. Thus, LY500307 appeared to attenuate the symptom recurrence in these women after E2 withdrawal compared with TE but not significantly better than placebo. Symptom severity during TE was greater in the women randomized to LY500307 compared with those randomized to placebo, which could confound our comparisons during double blind conditions. Thus, in a secondary analysis we adjusted for symptom severity at the end of the open label TE. The change in CESD (but not HDRS) scores in those women on LY500307 were lower (p=0.02) compared with those on placebo (i.e., E2 withdrawal). Nonetheless, both the lack of significant differences in depression severity scores between either dose of the ERbeta agonist and placebo, as well as the increased depression severity compared to open label TE, suggest that the beneficial effects of E2 are mediated through a combination of both ER beta and ER alpha. It also is possible that the dose of ERbeta agonist we administered, resulted in a lower occupancy of the ERbeta receptor than was present during TE. Finally, it is possible that LY500307 at either dose did not penetrate the brain sufficiently to impact mood and fully counteract the mood destabilizing effects of E2 withdrawal. Thus, these data do not support LY500307 as a therapeutic agent for PMD, albeit potentially safer and more acceptable than estrogen therapy.