Towards Understanding the Biology of Premenstrual Dysphoric Disorder: From Genes to GABA

Abstract

Premenstrual dysphoric disorder (PMDD) is a severe mood disorder, with affective symptoms that rise and fall in concert with the hormonal fluctuations of the menstrual cycle. PMDD’s pathophysiology is poorly understood. This review describes recent research on potential biological contributors to PMDD, with a focus on neuroactive steroids, genetics, neuroimaging and cellular studies. Studies suggest that a key contributor is abnormal CNS response to fluctuations in neuroactive steroid hormones. Imaging studies are limited but support alterations in serotonergic and GABA transmission. Genetic studies suggest heritability, yet specific genetic contributors have not been characterized. Finally, recent cutting-edge cellular studies indicate an underlying vulnerability to the effect of sex hormones at a cellular level. Overall the findings across studies do not yet fit together into a complete description of the underlying biology of PMDD. It is possible that PMDD consists of biological subtypes, and future research may benefit from a subtyping approach.

Introduction

Premenstrual dysphoric disorder (PMDD) is a severe mood disorder, with affective symptoms that rise and fall in concert with the hormonal fluctuations of the menstrual cycle (Bäckström et al., 2014; Rubinow et al., 1988; Schmidt et al., 1998; Seippel & Bäckström, 1998; Wang et al., 1996). Typical PMDD symptoms include affective lability, irritability, depressed mood, anxiety, anhedonia, sense of overwhelm, difficulty concentrating, fatigue, changes in appetite or food cravings, sleep changes, and physical symptoms such as breast tenderness, bloating or headaches (C. Epperson et al., 2012). These symptoms emerge in the luteal phase of the menstrual cycle, one to two weeks before menses. Once menses begins, symptoms remit within a few days of bleeding, and the individual remains asymptomatic through the follicular phase of the menstrual cycle, between menses and ovulation. In order to meet the DSM-5 diagnostic criteria for PMDD, the monthly emergence of symptoms results in significant impairment in function, affecting relationships and work or school performance (C. Epperson et al., 2012). PMDD affects three to eight percent of women worldwide (Halbreich et al., 2003), a prevalence on par with that of generalized anxiety disorder or panic disorder (Altemus et al., 2014; Kessler et al., 2005; McLean et al., 2011).

PMDD’s pathophysiology is understudied relative to its fellow mood disorders, but likely involves altered central nervous system (CNS) sensitivity to the normal hormonal fluctuations of the menstrual cycle. It is important to note that the luteal phase of the menstrual cycle is also known to exacerbate other types of psychiatric symptoms including in those in major depressive disorder, panic disorders, eating disorders, psychotic disorder, and borderline personality disorder (Handy et al., 2022; Nolan & Hughes, 2022). Thus, the hormonal fluctuations associated with the luteal phase of the menstrual cycle may exacerbate psychiatric symptoms more generally and understanding the underlying biological basis of sensitivity to normal female hormonal fluctuations may help shed light on the underlying pathophysiology of psychiatric disorder more globally. In addition to premenstrual exacerbation, premenstrual mood symptoms not necessarily meeting criteria for PMDD are also common and a recent report, using a survey administered to over 238,000 participants in 140 countries via a mobile phone app, found 64% reported premenstrual mood symptoms (Hantsoo et al., 2022). A similar study in the Netherlands of over 40,000 women found that 77% reported some form of premenstrual mood symptoms (Schoep et al., 2019), emphasizing how common these symptoms are, even in milder forms than PMDD.

This review will describe recent research on the underlying biology of PMDD with a focus on genetic studies and neuroactive steroids, as well as imaging and cellular level studies.

Menstrual Cycle Hormonal Changes

It has long been known that women with PMDD have normal levels of estrogen and progesterone throughout the menstrual cycle (Rubinow et al., 1998). In an elegant study, Schmidt et al demonstrated that women with PMDD had an abnormal mood response to normal hormonal fluctuations (Schmidt et al., 1998). They recruited women with prospectively confirmed PMDD and women who were prospectively confirmed healthy controls and induced a hypogonadal state with a gonadotropin releasing hormone agonist. They found that only women with PMDD experienced depressive mood symptoms when either estradiol or progesterone were added back in a blinded fashion. Importantly, the healthy control group experienced no change in mood during the study, indicating that normal concentrations of gonadal steroids trigger an ―abnormal response,‖ i.e. mood symptoms, in susceptible women. Thus the brain of a woman with PMDD reacts in an abnormal way to the hormonal fluctuations that occur in the luteal phase of the menstrual cycle. Symptoms begin with the rise in progesterone and estrogen levels during the luteal phase, typically worsen in the late luteal phase when hormone levels drop, and end when menstruation begins (or shortly thereafter) (Cunningham et al., 2009; O’Brien et al., 2011). One important clue to the etiology of PMDD is that symptoms primarily occur in cycles with ovulation: symptoms disappear in anovulatory cycles (Hammarbäck et al., 1991) when estrogen and progesterone levels don’t reach high enough levels to trigger ovulation. Thus, estrogen and progesterone must reach a certain peak in order to trigger not only ovulation, but also mood symptoms in susceptible women.

Subtypes of PMDD

Much of the research to date, including most of what is reviewed below, assumes that PMDD is a singular disorder with one pathophysiology. However there is heterogeneity within PMDD in symptom types and timing, severity, and treatment response, which may indicate underlying subtypes (Eisenlohr-Moul et al., 2020). For example, as reviewed below, only about 50% of patients with PMDD respond to selective serotonin reuptake inhibitors (SSRIs) (Hantsoo & Riddle, 2021) and surprisingly only about 50% respond to gonadotropin receptor hormone agonists (Wyatt et al., 2004) which induce low, stable hormone levels that in theory should alleviate the offending hormonal fluctuations. Other work has attempted to examine clusters of symptoms that make up PMDD including psychological, physical, and a combination of symptoms (Freeman et al., 2011). This work found that psychological and mixed symptom clusters were more likely to respond to SSRI treatment than physical symptoms. Another study used group-based trajectory modeling to identify unique trajectory subgroups of mood and total PMDD symptoms in a sample of prospectively diagnosed individuals with PMDD and found that there were three groups: 1) a group with moderate symptoms only in the premenstrual week (65%), a group with severe symptoms across the entire luteal phase (17.5%) and 3) a group with severe symptoms in the premenstrual week that were slow to resolve in the follicular phase (17.5%) (Eisenlohr-Moul et al., 2020). While all three groups met DSM-5 criteria for PMDD, it is clear that the distinct patterns of symptoms may result from unique biological responses to hormonal fluctuations at different points in the menstrual cycle. For instance, some experience a brief symptom burst around ovulation that may follow the post-ovulatory fall in estradiol (Reid, 1983, 2000). Thus, future research should consider whether more homogenous groups of individuals with PMDD are needed in order to identify underlying biological processes.

Genetic Studies of PMDD

Family and twin studies support a genetic component in the etiology of PMDD, though not every study has supported this idea. Several studies have found that retrospectively reported premenstrual mood symptoms demonstrate familiality (i.e. that the trait runs in families) (Diraimondo et al., 1964; Widholm & Kantero, 1971) though at least one has not (Glick et al., 1993). Twin studies have generally found a higher concordance rate in monozygotic twins compared to dizygotic twins (reviewed in (McEvoy et al., 2017)). Heritability estimates range from 35.1% (Kendler et al., 1992) to 56% (Kendler et al., 1998; van den Akker et al., 1987). No studies have conducted a linkage analysis genetic study, likely because of the difficulty of obtaining multiple first-degree relatives with multiple generations of women with reliable reports of premenstrual mood symptoms (McEvoy et al., 2017). Interpreting the family and twin studies is also complicated by the fact that many of them use retrospective reporting rather than the gold standard of prospective monitoring and it remains unclear if the genetic basis is for premenstrual mood symptoms themselves or another genetic trait such as neuroticism (McEvoy et al., 2017). Regardless, there have been a number of candidate gene studies, most of which have been negative (reviewed in (McEvoy et al., 2017)) in studies that specifically looked at PMDD alone (and not PMDD plus another trait such as a personality trait). There have been three positive candidate gene studies: 1) the Estrogen Receptor-1 (ESR1) gene with prospective reporting (Huo et al., 2007), 2) the ESR2 gene with retrospective reporting (Takeo et al., 2005), and 3) the Serotonin 1A receptor C(-1019) G polymorphism with prospective reporting (Dhingra et al., 2007). Notably, each of these findings has a companion study that is negative for association with PMDD (reviewed in (McEvoy et al., 2017)). Thus, although there is some evidence in family and twin studies that the trait of premenstrual mood symptoms runs in families, there is no evidence to date of a clear genetic etiology.

Imaging Studies

Given that the symptoms of PMDD seem to be due to an abnormal brain response to normal hormonal fluctuations, imaging studies using a variety of techniques have been completed in women with PMDD. In a recent review by Dubol and colleagues (Dubol et al., 2020), it was noted that greater cerebellar gray matter volume and metabolism have been observed in PMDD along with altered serotonergic and GABAergic neurotransmission. In addition, enhanced amygdala and diminished fronto-cortical activation have been found in response to emotional stimuli. This review also noted that there had been no white matter integrity studies - however a recent publication using diffusion tension imaging found greater fractional anisotropy in the left uncinate fasciculus (d = 0.69) in individuals with PMDD compared to controls (Gu et al., 2022). Moreover, the volume of the right uncinate fasciculus was higher in individuals with PMDD compared to controls (d = 0.40). As well, the severity of premenstrual depression was positively correlated with fractional anisotropy in the right superior longitudinal fasciculus (r = 0.35). This work was limited by the small size of the control group. Although the literature is limited, and more research and replication studies are needed, the available literature is consistent with the idea that there is a maladaptive neural response to normal hormonal fluctuations of the menstrual cycle.

Cellular Studies

Previous work has determined that messenger RNA (mRNA) expression in peripheral leukocytes is roughly similar to that in neurons (Tylee et al., 2013) and so can be used to approximate gene expression in the brain. One study attempted to determine if women with PMDD had altered gene expression in response to estrogen or progesterone (Dubey et al., 2017). Whole-transcriptome sequencing was used to compare RNA sequences and protein expression from lymphoblastoid cell lines from women with prospectively confirmed PMDD and asymptomatic controls after gonadotropin releasing hormone agonist (Lupron) treatment. All participants with PMDD responded symptomatically to the Lupron treatment. Cells were then exposed to either estrogen or progesterone and RNA sequencing and protein expression were again compared via whole-transcriptome sequencing (RNA sequencing) between the two groups. In women with PMDD, they found overexpression of the ESC/E(Z) gene complex across all conditions (Dubey et al., 2017). This complex is known to regulate epigenetic responses to the environment, including epigenetic responses to sex hormones (Marrocco et al., 2020). In contrast, protein expression of the genes making up the complex was reduced in cells from PMDD participants. Further, hormonal treatment of the cells differentially affected the transcription of genes involved in the complex in cells from participants with PMDD compared to healthy controls (Dubey et al., 2017). In a follow-up study in which control and PMDD participants underwent ovarian suppression via Lupron followed by ovarian hormone add-back, ESC/E(Z) gene expression correlated with change in resting regional cerebral blood flow from the ovarian suppression condition to the progesterone condition. These findings suggest that women with PMDD have an underlying vulnerability to the effect of sex hormones at a cellular level (Baller & Ross, 2019).

Neuroactive Steroids in PMDD

Neuroactive steroids (NASs) are steroid hormones produced in endocrine tissue or brain that interact with neuronal receptors, such as the gamma-aminobutyric acid (GABA)-A receptor (GABA_A-R). The key NASs that fluctuate across the menstrual cycle, relevant to PMDD, are progesterone, its metabolites, and estradiol. As described above, timing of symptom onset and offset in PMDD corresponds to fluctuations in NAS levels, yet levels of peripheral ovarian steroid hormones are generally similar between women with PMDD and asymptomatic controls (C. N. Epperson et al., 2002; Nguyen et al., 2017; Thys-Jacobs et al., 2008). (Although some studies have found differences in NAS levels between PMDD and controls, including higher allopregnanolone (ALLO) in the luteal phase among women with PMDD (Girdler et al., 2001) and lower ALLO in the luteal phase among women with premenstrual syndrome (PMS) (Rapkin et al., 1997)). The hormone sensitivity hypothesis proposes that PMDD represents an aberrant response to sex steroid hormones (Dubey et al., 2017; Schmidt et al., 1998), i.e. women with PMDD may have altered CNS receptor sensitivity to normal menstrual cycle hormone fluctuations (Bäckström et al., 2011; MacKenzie & Maguire, 2014).

GABAergic Neuroactive Steroids

Progesterone is a NAS with low levels during the follicular phase that rise across the luteal phase and drop rapidly in the late luteal phase just prior to menses. Progesterone is converted to metabolites including ALLO, pregnanolone, and their isomers. The conversion of progesterone to its metabolites requires neurosteroidogenic enzymes; 5α-reductase converts progesterone to 5α-dihydroprogesterone (5α-DHP), then 3α hydroxysteroid dehydrogenase (3α-HSD) converts 5α-DHP to ALLO (Agís-Balboa et al., 2006). It is possible that women with PMDD have alterations in the enzymatic pathway from progesterone to its neuroactive metabolites. For instance, women with PMDD had higher plasma 5α-DHP in the mid-luteal phase than control women, potentially suggesting limited conversion to ALLO (C. N. Epperson et al., 2002). However, when ovarian function was suppressed and estradiol or progesterone were administered, women with PMDD showed no differences in metabolic processing of these exogenous hormones compared with controls (Nguyen et al., 2017). Blocking 5-α reductase’s conversion of progesterone to ALLO reduced premenstrual mood symptoms in PMDD (Martinez et al., 2016); this may have been due to stabilizing ALLO levels, and eliminating the normal fluctuation in ALLO levels from follicular to luteal phases of the menstrual cycle, as opposed to neurosteroidogenic enzyme effects. Indeed, there are no studies to date directly demonstrating altered neurosteroidogenic enzymes in PMDD.

ALLO is a key metabolite of progesterone, and its levels fluctuate in tandem with progesterone across the menstrual cycle (Timby, 2012). ALLO has potent anxiolytic, anesthetic and sedative properties, similar to barbiturates or benzodiazepines (Schüle et al., 2014). This is because ALLO is a positive allosteric modulator (PAM) of the GABA_A-R, potentiating the effect of GABA, the main inhibitory neurotransmitter in the CNS. Upon binding to its site on the GABA_A-R, ALLO potentiates function by increasing Cl⁻ ion flux when GABA binds to the receptor (Chen et al., 2019; Lambert et al., 2009). While ALLO and pregnanolone are positive allosteric modulators of GABA_A-R and enhance the inhibitory effects of GABA, their isomers isoallopregnanolone and epipregnanolone are negative allosteric modulators that inhibit GABAergic neurotransmission. More research is needed on whether the balance between positive and negative allosteric modulators of GABA_A-R plays a role in PMDD.

Some studies have examined levels of GABA itself in PMDD. One found lower levels of GABA in mood-related brain regions of women with PMDD compared with controls (Liu et al., 2015). Another study found that cortical GABA levels increased from the follicular to luteal phase in women with PMDD and declined across the menstrual cycle in healthy women (C. N. Epperson et al., 2002), while another conversely found that GABA levels decreased from the follicular phase to luteal phase in women with PMDD (Halbreich et al., 1996). More work needs to be conducted in this area.

Neuroactive Steroid Fluctuations

Animal models of PMDD support the hypothesis that NAS fluctuations likely contribute to PMDD symptomatology. Indeed, rodent models of PMDD are based on withdrawal from progesterone or ALLO (Y. Li et al., 2012; Smith et al., 2006), wherein rapid withdrawal produces PMDD-like symptoms including anxiety behavior (Gulinello & Smith, 2003; Smith et al., 2006) and depressive behavior (Y. Li et al., 2012). In an animal model of PMDD based on progesterone withdrawal, rats that rapidly withdrew from physiological doses of progesterone exhibited PMDD-like symptoms, including social withdrawal and anhedonia (Y. Li et al., 2012). Similarly, when rate of progesterone withdrawal was manipulated in rodents, a rapid decline in plasma progesterone increased anxiety-like behavior, while a gradual progesterone decline did not (Doornbos et al., 2009).

Human studies also suggest a role for NAS withdrawal in PMDD. In a study that observed progesterone levels across the luteal phase, progesterone levels declined gradually in the eight days prior to menses in controls, but declined rapidly in the three days prior to menses in women with premenstrual symptoms (T. A. Lovick et al., 2017). In a randomized, double-blind, placebo-controlled crossover study, dutasteride, a 5-alpha reductase inhibitor, was administered, blocking 5-alpha reductase’s conversion of progesterone to ALLO (Martinez et al., 2016). This resulted in stable ALLO levels from the follicular to luteal phase rather than the normal increase in ALLO levels, and in women with PMDD, reduced premenstrual mood symptoms.

GABA_A-R Structure and Function

In the NAS withdrawal rodent models of PMDD mentioned above, NAS withdrawal also alters GABA_A-R structure. GABA_A-Rs are composed of five subunits of nineteen (α1–α6, β1–β3, γ1–γ3, δ, ε, θ, π, ρ1–ρ3) (Chua & Chebib, 2017; Simon et al., 2004). GABA_A-R subunit expression is dynamic, varying by sex and hormonal status (Bhandage et al., 2015; J. Maguire & Mody, 2008; Sanna et al., 2009). Different subunit combinations produce different sensitivity to pharmacologic compounds and NASs (Knoflach et al., 2018), and subunit expression is responsive to NAS fluctuations. In female rats, progesterone exposure and withdrawal upregulated the α4 subunit of GABA_A-R (Smith, Gong, Li, et al., 1998). In a mouse model of PMDD, withdrawal from ALLO increased expression of the GABA_A-R α4 subunit eight-fold, accompanied by anxiety-like behaviors (Smith et al., 2006). Female rats withdrawn from chronic progesterone exhibited upregulation of GABA_A-R α4 subunit expression by 2–3 fold within 24 hours of withdrawal, which was paralleled by increased startle response, a behavioral marker of anxiety (Gulinello et al., 2003). Administration of pregnanolone increased α4 subunit expression in female rats over a 48-hour period (Shen et al., 2005). Indeed, increases in the α4 subunit are associated with increased anxiety behavior (Gulinello et al., 2001; Smith, Gong, Li, et al., 1998). Thus, changes in expression of GABA_A-R subunits, such as α4, may be a key piece of PMDD pathophysiology. The α subunits of GABA_A-Rs are sensitive to NAS; GABA_A-Rs containing α1 and α3 subunits are sensitive to low concentrations of ALLO, while those containing α2-, α4-, α5- and α6 subunits are sensitive to higher ALLO concentrations (Belelli et al., 2002).

The δ subunits of GABA_A-R are also sensitive to NASs (Belelli et al., 2002). The δ subunit is involved in tonic inhibition (i.e. inhibition that occurs via extrasynaptic GABA_A-Rs (Farrant & Nusser, 2005)), and plays a role in regulating anxiety-like behavior (J. L. Maguire et al., 2005). Administration of pregnanolone to female rodents over 48 hours increased δ subunit expression four- to five-fold, in addition to α4 subunits (Shen et al., 2005). ALLO withdrawal failed to elicit anxiety behavior in δ subunit knockout mice (Smith et al., 2006).

GABA_A-R Interactions with Neuroactive Steroids

Impaired GABA_A-R plasticity in response to NAS fluctuations may also play a role in reproductive affective disorders (Gallo & Smith, 1993; Gulinello et al., 2001; MacKenzie & Maguire, 2014; Reddy et al., 2012; Smith, Gong, Hsu, et al., 1998; Sundstrom-Poromaa et al., 2002). For instance, in postpartum depression, another reproductive affective disorder, rodent models show that mice unable to regulate GABA_A-R response to changes in ALLO levels across pregnancy and postpartum showed postpartum depression-like behaviors (J. Maguire & Mody, 2008). Conversely, the rapid postpartum NAS withdrawal in control mice was followed by an appropriate adjustment in number of functional GABA_A-Rs (J. Maguire & Mody, 2008).

Women with PMDD also show evidence of altered GABA_A-R response to NAS changes across the menstrual cycle. Intravenous administration of ALLO to healthy controls increased sedation and decreased (slowed) saccadic eye velocity (SEV), which is regulated by the GABA_A-R (Ball et al., 1991; Bixo et al., 2018). Women with PMDD had increased SEV and subjective sedation in the luteal phase compared with the follicular phase when administered ALLO, suggesting an increased sensitivity to ALLO in the luteal phase in PMDD (Timby et al., 2016). Intravenous administration of pregnanolone, another positive modulator of the GABA_A-R, reduced SEV in PMS patients, but only when administered with the SSRI citalopram in the luteal phase, with the authors suggesting that SSRIs increase sensitivity to NASs (Sundström & Bäckström, 1998). Some studies show a paradoxical effect of ALLO, in which ALLO induces irritability or aggression instead of sedation. This inverted U-shaped curve relationship comprises negative mood symptoms when ALLO levels are moderate, yet low and high ALLO concentrations do not have a detrimental effect on mood (Andréen et al., 2009; Bäckström et al., 2011). For instance, when progesterone was administered to postmenopausal women with a history of PMDD, mood symptoms worsened in women with moderate serum ALLO levels, but mood symptoms were minimal in women with low and high ALLO levels (Andréen et al., 2006). Data from our laboratory also indicates that women with PMDD have elevated anxiety-potentiated startle in the luteal phase relative to controls in the luteal phase, suggestive of altered GABA_A-R function in the context of fluctuating ALLO levels (although this was only true in SSRI responders) (Hantsoo et al., 2021). Together, findings in rodents and humans suggest that PMDD involves a failure in GABA_A-R plasticity in response to changing ALLO levels across the menstrual cycle.

Estrogens

As described above, among PMDD patients whose hormonal fluctuations were suppressed with a gonadotropin releasing hormone agonist, addback of either estradiol or progesterone elicited depressive symptoms, suggesting that estradiol may also play a role in PMDD (Schmidt et al., 1998). Other early work suggested that PMS patients had higher estradiol levels (and lower progesterone levels) in the luteal phase compared with controls, and when examining within-participant PMS symptom scores across cycles, cycles with higher luteal estradiol levels were characterized by more severe symptoms, while cycles with higher luteal 5-alpha-DHP and 5-alpha-THP levels were characterized by improved PMS symptoms (Wang et al., 1996). Recent research found that at the cellular level, estradiol induced changes in calcium homeostasis and endoplasmic reticulum function in PMDD (H. J. Li et al., 2021). Another recent study found that luteal estradiol level was negatively correlated with anxiety symptoms in women with PMDD, but only among G allele carriers of the ESR α-Xbal polymorphism (Yen et al., 2018). Similarly, in a rodent study, ovariectomized BDNF Met allele heterozygotes received estradiol, which induced anxiety-like and depressive behavior, not in wild-type mice (Marrocco et al., 2020). The authors proposed that estradiol may produce these effects via epigenetic mechanisms. Among women with a menstrually-related mood disorder, estradiol increases predicted greater anxiety symptoms, but only among those with a history of sexual abuse (Eisenlohr-Moul et al., 2016). Estradiol also influences neurotransmitter function (Amin et al., 2005), particularly serotonin (reviewed in (Schiller et al., 2016)). Indeed, women with PMDD show some differences in serotonergic function, including differences in 5-HT(1A) binding potential (Jovanovic et al., 2006), differences in brain serotonin precursor trapping (Eriksson et al., 2006), lower whole blood serotonin levels in PMS (Rapkin et al., 1987), exacerbation of premenstrual mood symptoms with tryptophan depletion (Menkes et al., 1994), and blunted serotonin production in response to tryptophan challenge in PMS (Rasgon et al., 2000). More work is needed to unravel the complex connections between estradiol and serotonin function in PMDD.

Selective Serotonin Reuptake Inhibitors (SSRIs)

SSRIs are the first-line treatment for PMDD, an approach supported by the International Society of Premenstrual Disorders (ISPMD) (Ismaili et al., 2016; T. Lovick, 2013). In randomized, double blind placebo-controlled trials, luteal phase administration of sertraline (Yonkers et al., 2015) and paroxetine (Steiner et al., 2008) were superior to placebo in reducing PMDD symptoms. In PMDD, when administered in the luteal phase, SSRIs typically reduce symptoms in one to three days (Landén & Thase, 2006; Steinberg et al., 2012). SSRIs are also effective at low doses in PMDD, for instance, 25–50 mg sertraline (Kornstein et al., 2006) or 20 mg fluoxetine (Steinberg et al., 2012; Steiner et al., 2003). This rapid therapeutic effect and efficacy at low doses suggests a non-serotonergic mechanism of action in treating PMDD. However, serotonergic mechanisms have not been ruled out. In one study, participants with PMDD took the SSRI fluoxetine for three months. Those who responded to fluoxetine were then given metergoline, a serotonin receptor antagonist. When the serotonin receptor was blocked by metergoline, participants who were euthymic on fluoxetine showed rapid re-emergence of PMDD symptoms. Participants who received placebo instead of metergoline maintained euthymic mood. The authors concluded that the serotonin system may convey some of the efficacy of SSRIs in PMDD (Roca et al., 2002).

A suggested non-serotonergic mechanism is that SSRIs alter the conversion of progesterone to ALLO (Devall et al., 2015; Pinna et al., 2009), acting as selective brain steroidogenic stimulants (SBSS) (Pinna et al., 2006, 2009). Herein SSRIs act on enzymes that catalyze the conversion of 5α-DHP to ALLO (Griffin & Mellon, 1999; Trauger et al., 2002). Research in animals (Matsumoto et al., 2007; Pinna et al., 2006) and humans with major depression (Pinna et al., 2009; Uzunova et al., 1998) suggest that SSRIs enhance ALLO biosynthesis. In an open-label study, sertraline increased peripheral ALLO levels in women with PMDD who had low baseline ALLO, and decreased ALLO in those with high baseline levels, acting to normalize ALLO (Gracia et al., 2009). As described above, only about 50% of women with PMDD are responsive to SSRI treatment, perhaps indicating that there are two types of PMDD from a biological perspective - those that respond to SSRIs and those that do not. Future research should take this possibility into account.

GABA-Modulating Drugs

Recently, GABA-modulating drugs have been developed for reproductive affective disorders. Postpartum depression is similar to PMDD in that ALLO fluctuations may contribute to its etiology; ALLO levels rise steadily through pregnancy, then drop rapidly following parturition (Osborne et al., 2017). Brexanolone is a synthetic version of ALLO, recently FDA-approved for postpartum depression treatment. In brexanolone’s phase II (Kanes et al., 2017) and phase III clinical trials (Meltzer-Brody et al., 2018), brexanolone reduced PPD symptoms within 60 hours of administration. Similar to PMDD treatment with SSRIs, this is a rapid response compared with typical PPD treatments that can take weeks to relieve symptoms (Hantsoo et al., 2014; Wisner et al., 2006). This is likely due to brexanolone’s action as a positive allosteric modulator of GABA_A-Rs.

Sepranolone is synthetic isoallopregnanolone, an ALLO isomer and GABA-A modulating steroid antagonist (GAMSA). It inhibits the effect of ALLO on GABA_A-Rs. In an initial placebo-controlled trial, sepranolone reduced PMDD mood symptom scores by 75% when administered in the luteal phase in women with PMDD (Bixo et al., 2017). However, in a larger double-blind randomized controlled trial, sepranolone did not outperform placebo (Bäckström et al., 2021). Nevertheless, post-hoc analyses found that Sepranolone at a low dose improved symptoms more than placebo in the 9 days pre-menses (Bäckström et al., 2021).

Discussion

Most authorities agree that the primary underlying pathophysiology of PMDD is an abnormal CNS response to normal hormonal fluctuations, including estradiol, progesterone and progesterone metabolites, principally ALLO, across the menstrual cycle, particularly in the luteal phase. Imaging studies are limited but support abnormal metabolism and serotonergic and GABA transmission as well as enhanced amygdala activity and decreased frontal-cortical activity in response to emotional stimuli. These studies support an abnormal neural response in the pathophysiology of the illness. Genetic studies are conflicting but generally indicate that PMDD may run in families, yet it is unclear if there is a true genetic basis versus an environmental influence. Recent cellular studies in cells from individuals with PMDD indicate an overexpression of the ESC/E(Z) gene complex but an underexpression of proteins of the complex (Dubey et al., 2017). This complex is known to regulate epigenetic responses to the environment, including epigenetic responses to sex hormones and thus suggests that women with PMDD have an underlying vulnerability to the effect of sex hormones at a cellular level. Finally, NAS and the GABA-A-R are also thought to play a role. Fluctuations of positive and negative allosteric modulators of GABA-A-R, such as ALLO, pregnanolone, and their isomers, may contribute to PMDD via their interaction with the GABA-A-R. Relatedly, there may be deficits in neurosteroidogenic enzymes that facilitate the conversion of progesterone to these metabolites.

Overall the findings across studies are tantalizing but not definitive, and do not yet fit together into a complete description of the underlying biology of PMDD. It is likely, based on differential treatment responses to SSRIs and GnRH agonists, that PMDD consists of biological subtypes, and future research should begin to use a subtyping approach. For example, the biological dysfunction in individuals who quickly and definitively respond to SSRI treatment may be very different from the pathophysiology observed in individuals with PMDD who do not respond to SSRI treatments. Therefore, future studies could use a ―challenge‖ of treatment with SSRIs to discriminate between the two groups and study and compare biological responses between the two groups. Identifying more homogeneous subtypes of PMDD may lead to clearer and improved understanding of the biological process that produce the clinical syndrome described as PMDD.