Despite decades of research aimed at identifying the causes of postpartum depression (PPD) and developing effective methods of screening, prevention, and treatment, PPD remains common, affecting between 7% and 20% of women following delivery.Reference Gavin, Gaynes, Lohr, Meltzer-Brody, Gartlehner and Swinson 1 PPD is one of the most important public health problems that we can address: it not only affects women at a highly vulnerable time, but it also has deleterious effects on children and families. Many have speculated that PPD is caused, at least in part, by the rapid change in the reproductive hormones estradiol and progesterone before and immediately after delivery.Reference Hendrick, Altshuler and Suri 2 Although a number of human and nonhuman animal studies suggest that changes in reproductive hormone levels contribute to PPD,Reference Bloch, Schmidt, Danaceau, Murphy, Nieman and Rubinow 3 – Reference Suda, Segi-Nishida, Newton and Duman 8 several studies have failed to detect an association between hormone concentrations and PPD symptoms.Reference Buckwalter, Stanczyk and McCleary 9 – Reference O’Hara, Schlechte, Lewis and Varner 11 For example, cross-sectional human studies examining between-group differences in ovarian hormones levels and depressive symptoms during the postpartum period have failed to demonstrate an association between absolute estrogen and progesterone concentrations and PPD.Reference Buckwalter, Stanczyk and McCleary 9 – Reference O’Hara, Schlechte, Lewis and Varner 11 In contrast, studies in which PPD was treated with estradiol have successfully reduced depressive symptoms,Reference Gregoire, Kumar, Everitt, Henderson and Studd 5 , Reference Ahokas, Kaukoranta, Wahlbeck and Aito 12 and animal studies have demonstrated that estradiol and progesterone withdrawal provoke depression-like behavior.Reference Galea, Wide and Barr 4 , Reference Stoffel and Craft 7 , Reference Suda, Segi-Nishida, Newton and Duman 8
The mixed results regarding the role of estradiol and progesterone in PPD are likely due to 3 factors. First, the PPD diagnosis contains enormous variability. A postpartum depressive episode can meet the diagnostic criteria in a number of different ways, which results in women with very different symptom presentations receiving the same diagnosis. Two women could share only one symptom of major depression, experience timing of onset of the episode during very different hormonal conditions (eg, first trimester of pregnancy versus first week postpartum), and both receive a PPD diagnosis. Thus, PPD likely represents a number of depressive phenotypes, which may in large part account for the difficulty in identifying any one biological or hormonal factor central to the disorder.
Second, based on epidemiologic studies of risk, social and psychological factors play a large role in the pathogenesis of PPD. For example, decreased social support, poor quality social support, and poor marital satisfaction increase the risk of PPD.Reference Beck 13 – Reference O’Hara and Swain 15 The number of previous episodes of depression, a history of PPD, and depression during pregnancy are also significant risk factors for PPD.Reference O’Hara and Swain 15 – Reference O’Hara, Neunaber and Zekoski 17 PPD, like any mood disorder, is therefore best seen as a clinical integration of risk and protective factors that culminate in the triggering of a mood episode in the context of a biological (or reproductive) state.
Third, the existing studies have used widely diverging methods to examine how reproductive hormones influence depressive symptoms: some have examined absolute hormone concentrations in those with and without the disorder,Reference Buckwalter, Stanczyk and McCleary 9 – Reference O’Hara, Schlechte, Lewis and Varner 11 some have examined the change in hormone levels during pregnancy and the immediate postpartum period and the attendant changes in depressive symptoms,Reference Heidrich, Schleyer and Spingler 10 , Reference Schiller, O’Hara, Rubinow and Johnson 18 some have administered hormones to well individuals at high risk for PPD,Reference Bloch, Schmidt, Danaceau, Murphy, Nieman and Rubinow 3 and some have used hormones as a treatment for PPD.Reference Gregoire, Kumar, Everitt, Henderson and Studd 5 , Reference Ahokas, Kaukoranta, Wahlbeck and Aito 12 Any biological model of PPD must account for all 3 of these problems.
The purpose of this review is to examine the evidence for a reproductive hormone model of PPD in which fluctuating reproductive hormone levels trigger affective dysregulation. We will define PPD and discuss the diagnostic issues that contribute to difficulties in identifying a single biomarker for the disorder. We will discuss alternative biological models of PPD to demonstrate the potential for multiple PPD phenotypes and to describe the complex interplay of changing reproductive hormones and alterations in thyroid function, immune function, HPA axis function, lactogenic hormones, and genetic expression that may contribute to affective dysfunction. We will present animal models and human studies of reproductive hormones and PPD, and discuss methodological issues that have contributed to conflicting findings in the literature. We will provide evidence of a “hormone-sensitive” PPD phenotype, and discuss the potential neurobiological pathophysiology of PPD for this group of women. Finally, we will review human brain imaging and genetic studies as they pertain to the hormonal contribution to affective dysregulation during the perinatal period.
Defining PPD
The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) expanded the definition of PPD to include major depressive episodes with a perinatal onset as those beginning in either pregnancy or within the first 4 weeks postpartum. 19 Although PPD and non-perinatal major depressive disorder have the same DSM diagnostic criteria (ie, depressed mood, anhedonia, sleep and appetite disturbance, impaired concentration, psychomotor disturbance, lethargy, feelings of worthlessness or guilt, and suicidal ideation), 19 the symptoms of psychomotor agitation and lethargy are more prominent in PPD than MDD.Reference Bernstein, Rush and Yonkers 20 Additional symptoms of PPD include mood lability and preoccupation with infant well-being. PPD also is frequently associated with symptoms of anxiety, ruminative thoughts, and panic attacks.Reference Wisner, Sit and McShea 21 Indeed, most women with PPD have comorbid anxiety disorders.Reference Wisner, Sit and McShea 21 Recent estimates suggest that 7% of women experience an episode of major depression in the first 3 months following delivery, and the prevalence increases to 20% when episodes of minor depression are also included.Reference Gavin, Gaynes, Lohr, Meltzer-Brody, Gartlehner and Swinson 1 The majority of existing studies suggest that PPD is no more common than non-postpartum depressionReference O’Hara 22 ; however, the largest epidemiological study to date demonstrated an increased risk of depression during the postpartum period.Reference Vesga-Lopez, Blanco, Keyes, Olfson, Grant and Hasin 23
PPD is distinguished from the postpartum blues, which is defined by a normative, mild increase in negative affect in the days following delivery.Reference O’Hara 22 Also distinct from PPD is postpartum psychosis, which has a rapid onset associated with hallucinations or bizarre delusions, mood swings, disorganized behavior, and cognitive impairment.Reference Sit, Rothschild and Wisner 24 , Reference Wisner, Peindl and Hanusa 25 Many cases of postpartum psychosis are manifestations of bipolar disorder,Reference Brockington, Cernik, Schofield, Downing, Francis and Keelan 26 , Reference Kendell, Chalmers and Platz 27 which may present as mania for the first time during the postpartum period. The perturbation in mood, limited reality testing, and gross functional impairment make postpartum psychosis particularly dangerous for mothers and babies.Reference Sit, Rothschild and Wisner 24
An important limitation of the DSM criteria for PPD is that they are not mechanistically based, which is why the National Institute of Mental Health (NIMH) Research Domain Criteria (RDoC) project may be an ideal framework for studying PPD. The RDoC project advocates study of basic dimensions of functioning (eg, emotion processing) across multiple units of analysis (eg, genetic risk and epigenetic modification, limbic system, self-reported affective state) in a specific context (eg, reproductive hormonal state). The RDoC initiative, therefore, allows researchers to go beyond the DSM criteria to identify women who demonstrate patterns of affective dysregulation related to reproductive states and to examine the underlying neurobiological pathophysiology. For example, while some previous studies have strictly defined PPD according to the DSM criteria, most have used more inclusive criteria, including episodes of depression that began before or during pregnancy and carried over into the postpartum period and episodes with an onset several months following delivery. A study by Forty et al Reference Forty, Jones and Macgregor 28 demonstrated that defining PPD onset within 8 weeks of delivery is optimal for studying the biological triggering of affective dysregulation. Using this definition, Deligiannidis et al Reference Deligiannidis, Sikoglu and Shaffer 29 identified functional neural correlates of postpartum depressive symptoms that occur in the context of changing reproductive hormone and neurosteroid levels.
Biological Models of PPD in Humans
Reproductive hormone model of PPD
Many have hypothesized a role for reproductive hormones in PPD because of the temporal association between the substantial and rapid changes in hormone concentrations that occur at delivery and the onset of depressive symptoms.Reference O’Hara, Schlechte, Lewis and Varner 11 However, there are several more reasons for hypothesizing that reproductive hormones play a role in PPD. First, reproductive hormones play a major role in basic emotion processing, arousal, cognition, and motivation, and thus, may contribute to PPD indirectly by influencing the psychological and social risk factors. However, reproductive hormones also regulate each of the biological systems implicated in major depression, which suggests that hormones may impact a woman’s risk for PPD directly. In the forebrain and hippocampus, ovariectomy decreases and estradiol increases brain-derived neurotrophic factor (BDNF) levels,Reference Sohrabji, Miranda and Toran-Allerand 30 which are decreased by depression and stress and increased by antidepressants.Reference Shimizu, Hashimoto and Okamura 31 Estradiol also increases cAMP response element-binding (CREB) protein activityReference Zhou, Watters and Dorsa 32 and the neurotrophin receptor protein trkA,Reference Sohrabji, Greene, Miranda and Toran-Allerand 33 and it decreases GSK-3 beta activityReference Cardona-Gomez, Perez, Avila, Garcia-Segura and Wandosell 34 in the rat brain similar to antidepressant medications. Progesterone also regulates neurotransmitter synthesis, release, and transport.Reference Finocchi and Ferrari 35 For example, progesterone upregulates BDNF expression in the hippocampus and cerebral cortex.Reference Pluchino, Russo, Santoro, Litta, Cela and Genazzani 36 The relevance of gonadal steroids to affective regulation is further suggested by modulatory effects on stress and the HPA axis, neuroplasticity, cellular energetics, immune activation, and cortical activity,Reference Rubinow and Girdler 37 all processes that have been implicated as dysfunctional in depression.
Of particular note are the manifold effects of gonadal steroids on brain function as revealed by brain imaging studies. These studies, by employing positron emission tomography (PET) or functional magnetic resonance imaging (fMRI) in asymptomatic women, have demonstrated that physiologic levels of gonadal steroids modulate the neurocircuitry involved in normal affective states. In a study of healthy women, regional cerebral blood flow (rCBF) was attenuated in the dorsolateral prefrontal cortex, inferior parietal lobule, and posterior inferior temporal cortex during GnRH agonist-induced hypogonadism, whereas the characteristic pattern of cortical activation re-emerged during both estradiol and progesterone addback.Reference Berman, Schmidt and Rubinow 38 Studies of neural activity during the menstrual cycle have compared activation across menstrual phases within subjects. Goldstein et al Reference Goldstein, Jerram and Poldrack 39 found increased amygdala activity during the late follicular phase (higher estradiol levels) compared to the early follicular phase (lower estradiol levels), and Protopopescu et al Reference Protopopescu, Pan and Altemus 40 demonstrated increased activity in the medial orbitofrontal cortex (a region that exerts inhibitory control over amygdalar function) during the luteal phase (higher estradiol levels) compared with the follicular phase (relatively lower estradiol levels). The opposite was true for the lateral orbitofrontal cortex, suggesting that sensory and evaluative neural functions are suppressed in the days prior to menstruation.Reference Protopopescu, Pan and Altemus 40 Ovarian hormones also modulate neural reward function, with increased activation of the superior orbitofrontal cortex and amygdala during reward anticipation and of the midbrain, striatum, and left ventrolateral prefrontal cortex during reward delivery in the follicular phase (compared with the luteal phase).Reference Dreher, Schmidt, Kohn, Furman, Rubinow and Berman 41 Thus, there is evidence that reproductive hormones influence the biological systems and neural circuits implicated in depression directly, suggesting that the hormone instability inherent in the perinatal period could contribute to mood dysregulation in PPD.
Alternative biological models
The hormonal changes of pregnancy and the postpartum period do not occur in isolation: several other biological systems are altered during pregnancy and have been implicated in PPD. Alterations in any of these systems may provoke PPD independent of the changing hormonal milieu, which would suggest that there are a number of PPD phenotypes, each with their own relevant biomarkers. Thus far, the search for one biomarker for the general category of PPD has been elusive, and further research is needed to determine whether there are multiple PPD phenotypes with distinct etiologies. It also stands to reason that perturbations of other biological systems act in concert with rapidly changing hormone levels to contribute to affective dysregulation. Indeed, reproductive hormones have been shown modulate all of the other biological systems implicated in PPD: thyroid function,Reference Santin and Furlanetto 42 lactogenic function,Reference Schumacher, Coirini, Pfaff and McEwen 43 the hypothalamic–pituitary–adrenal (HPA) axis,Reference Walf and Frye 44 , Reference Roca, Schmidt and Altemus 45 and the immune system.Reference Butts and Sternberg 46 As such, we will discuss the potential contribution of each of these systems to affective dysregulation during pregnancy and the postpartum period, and we will discuss the evidence of a genetic susceptibility to PPD.
Thyroid hormones have been proposed as a biomarker of PPD in large part because of the presumed relationship between thyroid dysfunction and major depressionReference Bunevicius, Kusminskas, Mickuviene, Bunevicius, Pedersen and Pop 47 : depression accompanies thyroid pathologies,Reference Placidi, Boldrini and Patronelli 48 , Reference Gulseren, Gulseren, Hekimsoy, Cetinay, Ozen and Tokatlioglu 49 thyroid dysregulation accompanies depression,Reference Berent, Zboralski, Orzechowska and Gałecki 50 , Reference Nemeroff, Simon, Haggerty and Evans 51 and the administration of thyroid hormones is thought to augment and accelerate antidepressant treatment.Reference Cooper-Kazaz and Lerer 52 , Reference Cooper-Kazaz, Apter and Cohen 53 Estrogen increases thyroxine-binding globulin (TBG) and consequently increases circulating thyroxine (T4) levels.Reference Ben-Rafael, Struass, Arendash-Durand, Mastroianni and Flickinger 54 , Reference Arafah 55 Thyroid dysfunction is associated with pregnancyReference Vaidya, Anthony and Bilous 56 and may contribute to PPD in some women.Reference Pedersen, Johnson and Silva 57 , Reference Pedersen, Stern, Pate, Senger, Bowes and Mason 58 However, previous studies have failed to detect a clear association between thyroid hormone dysregulation and PPD in the majority of patients.Reference Bloch, Daly and Rubinow 59 – Reference Kent, Stuckey, Allen, Lambert and Gee 61
The lactogenic hormones oxytocin and prolactin have been implicated in PPD.Reference Stuebe, Grewen, Pedersen, Propper and Meltzer-Brody 62 Failed lactation and PPD commonly co-occur, and lactogenic hormones regulate not only the synthesis and secretion of breast milk, but also maternal behavior and mood. Oxytocin, in particular, may account for the shared pathogenesis of unplanned early weaning and PPD.Reference Stuebe, Grewen and Meltzer-Brody 63 Estrogen and progesterone modulate oxytocin mRNA expression in brain regions associated with maternal behavior and lactation.Reference Amico, Crowley, Insel, Thomas and O’Keefe 64 , Reference Broad, Kendrick, Sirinathsinghji and Keverne 65 Lower oxytocin levels during the third trimester are associated with increased depressive symptoms during pregnancyReference Stuebe, Grewen and Meltzer-Brody 63 and the immediate postpartum period.Reference Skrundz, Bolten, Nast, Hellhammer and Meinlschmidt 66 In a recent study by Stuebe et al,Reference Stuebe, Grewen and Meltzer-Brody 63 oxytocin secretion during breastfeeding was inversely associated with depression and anxiety symptoms at 8 weeks postpartum. Although depression and anxiety symptoms were not associated with breastfeeding success in this study, reduced oxytocin may predispose women to PPD and subsequently lead to unsuccessful breastfeeding. Moreover, low oxytocin levels in mothers with PPD are associated with low oxytocin levels in fathers and their children, suggesting a potential neuroendocrine mechanism for the increased risk of depression in children of depressed mothers.Reference Apter-Levy, Feldman, Vakart, Ebstein and Feldman 67 Last, oxytocin has also been examined as a potential treatment for a wide range of psychiatric disorders, including PPD, but with inconsistent findings to date.Reference Kim, Soeken, Cromer, Martinez, Hardy and Strathearn 68 , Reference Mah, Van IJzendoorn, Smith and Bakermans-Kranenburg 69
HPA axis dysfunction has also been implicated in the pathogenesis of PPD. HPA axis hyperactivity is one of the most consistent findings in the neuroendocrinology of depression.Reference Nestler, Barrot, DiLeone, Eisch, Gold and Monteggia 70 Hypercortisolism is associated with depressive symptoms and is corrected with antidepressant treatment.Reference Nestler, Barrot, DiLeone, Eisch, Gold and Monteggia 70 Additionally, the HPA axis is dysregulated by stress and trauma,Reference Heim, Newport, Wagner, Wilcox, Miller and Nemeroff 71 both of which are known precipitants of PPD.Reference Beck 13 , Reference Records and Rice 72 , Reference Ross and Dennis 73 Levels of corticotropin-releasing hormone (CRH), adrenocorticotropic hormone (ACTH), and cortisol increase substantially during pregnancy and drop 4 days following delivery.Reference Mastorakos and Ilias 74 HPA axis function normalizes at approximately 12 weeks postpartum.Reference Mastorakos and Ilias 74 The effects of pregnancy on HPA axis function may be at least partially attributable to the effects of estrogen on corticosteroid binding globulin,Reference Young 75 CRH gene expression,Reference Vamvakopoulos and Chrousos 76 and circulating corticotropin concentrations.Reference Walf and Frye 44 Similar to the HPA axis dysregulation seen in nonpuerperal depression, basal concentrations of plasma cortisol are increased in women with PPD, and suppression of cortisol by dexamethasone is blunted.Reference Bloch, Daly and Rubinow 59 In one study, for women with PPD, there was no association between ACTH and cortisol levels in response to a stress test, whereas in nondepressed control women, there was a more regulated association with cortisol levels rising following the increase in ACTH.Reference Jolley, Elmore, Barnard and Carr 77 Some evidence suggests that higher cortisol levels at the end of pregnancy are associated with increased blues symptoms.Reference Handley, Dunn, Waldron and Baker 78 However, it remains unclear whether HPA dysregulation contributes to the onset of PPD or occurs as an epiphenomenon.
Immune dysregulation has been hypothesized to contribute to the development of PPD.Reference Corwin and Pajer 79 During pregnancy, anti-inflammatory cytokines responsible for immunosuppression are elevated and promote pregnancy maintenance, whereas proinflammatory cytokines are downregulated. Delivery abruptly shifts the immune system into a proinflammatory state, which lasts for several weeks. Patients with depression tend to have higher levels of the proinflammatory cytokines tumor necrosis factor (TNF)-α and interleukin (IL)-6,Reference Dowlati, Herrmann and Swardfager 80 and administration of cytokines is associated with the onset of depression.Reference Raison, Capuron and Miller 81 The immune axis is regulated by estradiol. Estradiol modulates cytokine production, cytokine receptor expression, activation of effector cells, both the number and function of dendritic cells and antigen presenting cells, and monocyte and macrophage immune function.Reference Cunningham and Gilkeson 82 Differential patterns of gene expression that are functionally related to differences in immunity have been found to distinguish women with PPD from those without.Reference Segman, Goltser-Dubner and Weiner 83 Although one recent study identified several prenatal immune markers of PPD,Reference Krause, Jobst and Kirchberg 84 other studies have failed to detect an association between immune dysfunction and postpartum depressive symptoms.Reference Blackmore, Moynihan, Rubinow, Pressman, Gilchrist and O’Connor 85 – Reference Okun, Luther, Prather, Perel, Wisniewski and Wisner 87 Thus, the role of immune function in PPD remains unclear.
Evidence of a genetic vulnerability to PPD has emerged from family, candidate gene, genome-wide, and gene manipulation studies. Family and twin studies suggest that PPD aggregates in families,Reference Forty, Jones and Macgregor 28 , Reference Murphy-Eberenz, Zandi and March 88 is heritable,Reference Treloar, Martin, Bucholz, Madden and Heath 89 and may be genetically distinct from nonpuerperal depression.Reference Treloar, Martin, Bucholz, Madden and Heath 89 Although multiple genes likely contribute to PPD, the role of specific genetic variations remains unclear. Candidate gene studies of PPD have identified several of the same polymorphisms implicated in nonpuerperal depression, including the Val66Met polymorphism of the BDNF gene,Reference Figueira, Malloy-Diniz and Campos 90 , Reference Comasco, Sylvén, Papadopoulos, Oreland, Sundström-Poromaa and Skalkidou 91 the Val158Met polymorphism of the COMT gene,Reference Comasco, Sylvén, Papadopoulos, Sundström-Poromaa, Oreland and Skalkidou 92 , Reference Alvim-Soares, Miranda, Campos, Figueira, Romano-Silva and Correa 93 the BcII polymorphism of the glucocorticoid receptor and the rs242939 polymorphism of the CRH receptor 1,Reference Engineer, Darwin, Nishigandh, Ngianga-Bakwin, Smith and Grammatopoulos 94 the short version of the serotonin-transporter linked polymorphic region (5-HTTLPR) genotype,Reference Binder, Newport and Zach 95 , Reference Mitchell, Notterman and Brooks-Gunn 96 3 polymorphisms in the serotonin 2A receptor (HTR2A) gene,Reference El-Ibiary, Hamilton, Abel, Erdman, Robertson and Finley 97 and 3 polymorphisms at protein kinase C-β (PRKCB).Reference Costas, Gratacòs and Escaramís 98 There is also evidence of PPD biomarkers that are theoretically distinct from those of MDD and that implicate reproductive hormones. For example, polymorphisms in the estrogen receptor α gene (ESR1) have been found to be associated with PPD.Reference Costas, Gratacòs and Escaramís 98 , Reference Pinsonneault, Sullivan, Sadee, Soares, Hampson and Steiner 99 However, to date, the results of candidate gene studies of MDD and PPD have failed to replicate,Reference Pinheiro, Coelho and Silva 100 and have not been statistically significant after correcting for multiple comparisons.Reference El-Ibiary, Hamilton, Abel, Erdman, Robertson and Finley 97 , Reference Costas, Gratacòs and Escaramís 98 Also, there is little consistency in the specific polymorphisms tested and identified across studies, which means that any one genetic variant or set of genetic variants is of limited utility as a diagnostic indicator. Genomic studies aim to address some of these shortcomings, and there have been a few small genomic studies of PPD to date. In a genome-wide linkage study of 1,210 women, researchers identified genetic variations on chromosomes 1q21.3-q32.1 and 9p24.3-p22.3 that may increase susceptibility to PPD.Reference Mahon, Payne and MacKinnon 101 Of particular relevance here, the strongest implicated gene was Hemicentin 1 (HMCN1), which contains multiple estrogen binding sites. Although the results were no longer significant after accounting for multiple comparisons,Reference Mahon, Payne and MacKinnon 101 the association between the rs2891230 polymorphism of the HMCN1 gene and PPD was confirmed by a subsequent candidate gene study.Reference Alvim-Soares, Miranda, Campos, Figueira, Correa and Romano-Silva 102 Similarly, a genome-wide association study yielded a third-trimester biomarker panel of 116 transcripts that predicted PPD onset with 88% accuracy in both the discovery sample of 62 women and the independent replication sample of 24 women.Reference Mehta, Newport and Frishman 103 Of these transcripts, ESR1 was the only enriched transcription factor binding site, again potentially implicating estrogen in the pathogenesis of PPD.Reference Mehta, Newport and Frishman 103 Estrogen-induced DNA methylation change has also been implicated in PPD, which suggests that women with PPD have an enhanced sensitivity to estrogen-based DNA methylation reprogramming.Reference Guintivano, Arad, Gould, Payne and Kaminsky 104 In order to serve as reliable biomarkers of PPD, these genetic variants will require replication in larger, independent samples, which is currently an active area of investigation in the field.
Reproductive Hormone Models of PPD in Rodents
Nonhuman animal studies largely support the role of reproductive hormones in PPD. Ovariectomized rats treated with 17β-estradiol and progesterone followed by vehicle only, in order to induce a hormone withdrawal state similar to the rodent postpartum period, show increased immobility during the forced swim test,Reference Galea, Wide and Barr 4 , Reference Stoffel and Craft 7 a behavioral indicator of despair, and decreased sucrose consumption and preference,Reference Green, Barr and Galea 105 a behavioral indicator of anhedonia. One recent study demonstrated that estradiol supplementation and withdrawal alone was sufficient to provoke immobility during the forced swim test and anhedonic behavior during lateral hypothalamic self-stimulation.Reference Schiller, O’Hara, Rubinow and Johnson 18 Increased depression-like behavior during the “postpartum” demonstrated in previous studies could therefore be attributed to estradiol withdrawal alone.
The effects of estradiol withdrawal on depressive behavior in nonhuman animals are well documented. Following bilateral ovariectomy, rats demonstrate increased immobility during the forced swim test, and these effects are reversed by treatment with estradiol alone.Reference Bekku and Yoshimura 106 , Reference Bernardi, Vergoni, Sandrini, Tagliavini and Bertolini 107 In addition, reduced immobility following a single administration of estradiol lasts 2–3 days, and the behavioral effects are the same as those following fluoxetine treatment.Reference Estrada-Camarena, Fernandez-Guasti and Lopez-Rubalcava 108 The antidepressant effects of estradiol during the forced swim test appear to involve selective actions at intracellular estrogen receptor-β (ERβ) in the ventral tegmental areaReference Walf, Rhodes and Frye 109 and, in fact, may require ERβ.Reference Walf and Frye 110 In addition, abrupt estradiol withdrawal following sustained high estradiol levels results in elevated brain cortical dehydroepiandrosterone sulfate (DHEA-S), a neuroactive steroid synthesized endogenously in the brain that attenuates GABA-ergic activity and may be relevant to postpartum depressive symptoms.Reference Maayan, Strous, Abou-Kaoud and Weizman 111 Chronic administration of estradiol leads to dopamine receptor upregulation and increased presynaptic dopamine activity in the striatum,Reference Bossé, Rivest and Di Paolo 112 – Reference Di Paolo, Poyet and Labrie 114 which, when followed by abrupt estradiol withdrawal, leads to dysregulation in brain dopaminergic pathways and depressive symptoms.Reference Byrnes, Byrnes and Bridges 115
Estradiol-withdrawal models of PPD have two weaknesses: (1) they have low face validity as models of PPD, given that the human postpartum period is characterized by a drop in both estradiol and progesterone (whereas progesterone drops before delivery in rodents), and (2) they result in depression without the attendant anxiety often seen in women with PPD.Reference Wenzel, Haugen, Jackson and Brendle 116 The addition of progesterone to hormone withdrawal models of PPD is relevant, given that progesterone withdrawal provokes anxiety. As noted above, progesterone metabolites act on GABA receptors in the brain, producing sedative-like effects by enhancing γ-aminobutyric acid (GABA) neurotransmission.Reference Beckley and Finn 117 Abrupt decreases in progesterone are associated with anxiety,Reference Smith, Gong and Li 118 and treatment with progesterone reduces anxiety.Reference Dennerstein, Spencer-Gardner, Gotts, Brown, Smith and Burrows 119 The anxiolytic effects of progesterone appear to be mediated by the progesterone metabolite allopregnanolone (ALLO).Reference Frye and Walf 120 Indeed, postpartum rats show increased depressive behavior (increased immobility, decreased struggling and swimming) compared with pregnant rats, and this effect appears to be mediated by low hippocampal ALLO levels during the postpartum period.Reference Frye and Walf 120
To examine the effects of concurrent estradiol and progesterone withdrawal, Suda et al Reference Suda, Segi-Nishida, Newton and Duman 8 created a novel rodent model of PPD by administering hormone levels that are more consistent with human pregnancy than rat pregnancy. The concurrent withdrawal of estradiol and progesterone resulted in decreased immobility during the forced swim test (ie, less depression-like behavior); however, it also resulted in learned helplessness, which was indicated by a failure to avoid repeated foot shocks.Reference Suda, Segi-Nishida, Newton and Duman 8 Animals in this study also showed increased anxiety. Taken together, the existing animal models suggest that the abrupt withdrawal of estradiol alone produces behavioral despair and anhedonia, whereas the concurrent withdrawal of progesterone and estradiol produces learned helplessness and anxiety. However, these studies do not explain how the same putative stimulus (ie, hormone change) is capable of causing depression in some women and not others.
Fluctuating Reproductive Hormone Levels Trigger PPD
There is no consistent or convincing evidence that women who develop PPD experience more rapid postpartum hormone withdrawal, have lower reproductive hormone concentrations during the postpartum period, or experience greater reductions in hormone levels from pregnancy to the postpartum than women without PPD.Reference Buckwalter, Stanczyk and McCleary 9 – Reference O’Hara, Schlechte, Lewis and Varner 11 , Reference Deligiannidis, Sikoglu and Shaffer 29 , Reference Chatzicharalampous, Rizos and Pliatsika 121 The onset of depressive symptoms, however, is temporally coincident with the rapid changes in estradiol and progesterone levels that occur at delivery, leading some researchers to view the change in reproductive hormones as a potent stressor in susceptible women.Reference O’Hara, Schlechte, Lewis and Varner 11
Evidence that a subgroup of women is vulnerable to perinatal changes in reproductive hormones comes from treatment studies that have examined the effects of administering exogenous estradiol to women at high risk for PPD or those with active PPD symptoms. In a pilot study of 11 women with a history of PPD and no other history of affective disorder, participants were prophylactically administered oral Premarin, a conjugated estrogen, immediately following delivery to prevent estrogen withdrawal and the onset of depressive symptoms.Reference Sichel, Cohen, Robertson, Ruttenberg and Rosenbaum 6 Ten of the 11 women remained well during the postpartum period and for the first year following delivery.Reference Sichel, Cohen, Robertson, Ruttenberg and Rosenbaum 6 A later double-blind, placebo-controlled study of 61 women with PPD that began within 3 months following delivery showed that women treated with estradiol (n=34) (delivered via a transdermal patch) improved significantly more than women who received placebo (n=27), although nearly half of the women in both groups were also taking antidepressant medication.Reference Gregoire, Kumar, Everitt, Henderson and Studd 5 A subsequent study examined the effects of estradiol treatment on a group of 23 women with severe postpartum depression, many of whom had attempted treatment with antidepressant medication or psychotherapy without effect.Reference Ahokas, Kaukoranta, Wahlbeck and Aito 12 At baseline, 16 of the 23 patients had serum estradiol concentrations consistent with gonadal failure. All women in the study received sublingual estradiol treatment for 8 weeks. After the first week, depressive symptoms significantly decreased, and by the end of the 8 weeks, all patients had achieved depressive symptom scores consistent with clinical recovery. Although Ahokas et al Reference Ahokas, Kaukoranta, Wahlbeck and Aito 12 suggest that postpartum “gonadal failure” is a risk factor for PPD, they did not compare estradiol levels in women with and without PPD. Instead, their data support the notion that, in susceptible women, low or declining estradiol levels may trigger PPD, while stable or increasing estradiol levels may ameliorate depressive symptoms. Although these treatment studies suggest a role for estradiol in the pathogenesis of PPD, they are limited by small sample sizes and lack of control groups, and are confounded by the effects of stress, lack of sleep, and homeostatic shifts attendant to childbirth.
In order to assess the role of reproductive hormones in PPD directly, Bloch et al Reference Bloch, Schmidt, Danaceau, Murphy, Nieman and Rubinow 3 created a scaled-down hormonal model of the puerperium wherein euthymic women with or without a history of PPD were blindly administered high-dose estradiol and progesterone during ovarian suppression and then abruptly withdrawn. Women with a history of PPD showed increasing depressive symptoms during hormone addback and further exacerbation during hormone withdrawal, but women lacking a history of PPD experienced no perturbation of mood despite identical hormonal conditions. Increasing depressive symptoms during both hormone addback and withdrawal among those with a history of PPD is consistent with research demonstrating that one of the biggest risk factors for PPD is depression during pregnancy.Reference O’Hara and Swain 15 The advantage of this design is that the effects of reproductive hormones on mood were examined without the confounding biological and psychosocial stressors associated with childbirth. The results provide support for a hormone-sensitive PPD phenotype in which reproductive hormone change alone is sufficient to provoke mood dysregulation in otherwise euthymic women.
Some have hypothesized that the source of PPD vulnerability is in abnormal neural responses to the normal perinatal fluctuations in reproductive hormones. PPD is characterized by abnormal activation of the same brain regions that are implicated in nonpuerperal major depression: the amygdala, insula, striatum, orbitofrontal cortex, and dorsomedial prefrontal cortex.Reference Silverman, Loudon and Safier 122 – Reference Moses-Kolko, Fraser and Wisner 124 PPD is also associated with reduced connectivity between the amygdala and prefrontal regions, which implicates dysregulation of the limbic system in the neural pathophysiology of PPD.Reference Moses-Kolko, Perlman, Wisner, James, Saul and Phillips 123 Despite similar levels of circulating progesterone and ALLO to controls, women with PPD also show reduced resting state functional connectivity between the anterior cingulate cortex, amygdala, hippocampus, and dorsolateral prefrontal cortex in the context of the postnatal decline progesterone and ALLO.Reference Deligiannidis, Sikoglu and Shaffer 29 These neuroimaging studies suggest that the neural abnormalities associated with PPD are unique to the perinatal period and may be unmasked by changes in circulating reproductive hormone concentrations. Taken together, the results of the human studies are suggestive of a hormone-sensitive PPD phenotype characterized by neural abnormalities that are present during the puerperium when reproductive hormone concentrations change rapidly.
One potential mechanism by which changing reproductive hormone levels trigger PPD involves neurosteroid regulation of affect. Neurosteroids are metabolites of steroid hormones that are synthesized in the brain and nervous system and modulate GABA and glutamate. Two neurosteroids in particular play a role in affective dysregulation: dehydroepiandrosterone (DHEA) and ALLO. Abnormal DHEA secretion has been implicated in major depression,Reference Goodyer, Herbert, Altham, Pearson, Secher and Shiers 125 – Reference Schmidt, Murphy, Haq, Danaceau and St Clair 129 and DHEA is an effective antidepressant in both men and women.Reference Wolkowitz, Reus and Keebler 130 , Reference Schmidt, Daly and Bloch 131 The majority of research on neurosteroids in reproductive mood disorders, however, has focused on the progesterone metabolite ALLO. There are several reasons to speculate that ALLO plays a role in PPD. ALLO modulates the GABA receptor, which mediates anxiolysis.Reference Majewska, Harrison, Schwartz, Barker and Paul 132 ALLO supplementation has anxiolytic effects,Reference Bitran, Hilvers and Kellogg 133 – Reference Bitran, Purdy and Kellog 135 whereas ALLO withdrawal produces anxiety and insensitivity to benzodiazepines.Reference Smith, Gong and Li 118 , Reference Smith, Gong, Hsu, Markowitz, ffrench-Mullen and Li 136 ALLO levels are decreased in depression and increased with successful antidepressant treatment.Reference Uzunova, Sheline and Davis 137 – Reference Schüle, Baghai and di Michele 142 ALLO also modulates the biological processes that are dysregulated in major depressive disorder, including HPA axis regulation,Reference Patchev, Shoaib, Holsboer and Almeida 143 – Reference Kehoe, Mallinson, McCormick and Frye 146 neuroprotection,Reference Djebaili, Guo, Pettus, Hoffman and Stein 147 , Reference Sayeed, Parvez, Wali, Siemen and Stein 148 and immune function.Reference He, Evans, Hoffman, Oyesiku and Stein 149 ALLO also regulates the neural circuits that are implicated in depression, including the limbic system.Reference Bixo, Andersson, Winblad, Purdy and Bäckström 150 , Reference Akwa, Purdy, Koob and Britton 151
Cortical GABA and ALLO are reduced in postpartum women, regardless of the presence of PPD, compared with healthy women in the follicular phase.Reference Epperson, Gueorguieva and Czarkowski 152 Although there is no evidence of abnormalities in basal circulating ALLO levels in PPD, women with PPD show reduced resting state functional connectivity between the anterior cingulate cortex, amygdala, hippocampus, and dorsolateral prefrontal cortex in the context of the postnatal decline in ALLO.Reference Deligiannidis, Sikoglu and Shaffer 29 In addition, we recently reported an association between changes in ALLO levels and depressive symptoms during GnRH agonist-induced ovarian suppression and ovarian steroid addback in women with a history of PPD, but not in those without such a history.Reference Schiller, Schmidt and Rubinow 153 These studies suggest that, even in the presence of normal absolute levels, perinatal fluctuations in reproductive hormones may precipitate symptoms in a vulnerable subpopulation of women as a result of changing ALLO levels.
The identification of biomarkers in humans is difficult because of a lack of experimental control over the patient’s environment and genetic background, as well as inaccessibility of brain tissue required for analysis. Gene manipulation studies in nonhuman animals can help model how genetic variants and the environment interact to yield a distinct behavioral phenotype.Reference Tarantino, Sullivan and Meltzer-Brody 154 Animal models, which have demonstrated that the behavioral effects of maternal care are associated with gene expression changes that persist into adulthood and can be transmitted across generations, provide a potent epigenetic model of PPD.Reference Tarantino, Sullivan and Meltzer-Brody 154 For example, estradiol withdrawal is clearly associated with estradiol-reversible anxiety in a strain-dependent fashion (Schoenrock et al, unpublished manuscript). One genetic knockout model potentially explains both the specificity of affective dysregulation during the perinatal period and also the variation in susceptibility to PPD among women.Reference Maguire and Mody 155 In this model, Maguire and ModyReference Maguire and Mody 155 demonstrated a GABAA receptor subunit knockout that is behaviorally silent until an animal is exposed to pregnancy and the postpartum state, following which the dam displays depression-like behavior and cannibalizes its young. Thus, reproductive events may unmask the genetic susceptibility to affective dysregulation. Maguire and ModyReference Maguire and Mody 155 , Reference Maguire, Stell, Rafizadeh and Mody 156 , Reference Maguire, Ferando, Simonsen and Mody 157 observed that alterations in the GABAA receptor δ-subunit occur as ovarian hormone levels fluctuate during the menstrual cycle, pregnancy, and the postpartum period. During pregnancy, the expression of the GABAA receptor δ-subunit is downregulated as ALLO levels increase, and at parturition, the expression of the GABAA receptor δ-subunit recovers in response to rapidly declining neurosteroid levels.Reference Maguire, Ferando, Simonsen and Mody 157 The failure to regulate these receptors during pregnancy and the postpartum, consequent to the knockout of the GABAA receptor δ-subunit, appears to provoke behavioral abnormalities consistent with PPD. Thus, as noted above, GABAA receptor δ-subunit deficient mice exhibit normal behaviors prior to pregnancy, but they show insensitivity to ALLO during pregnancy, depression-like and anxiety-like behavior, and abnormal maternal behavior.Reference Maguire and Mody 155 This model suggests that changes in reproductive hormone concentrations during pregnancy and the postpartum period are capable of provoking affective dysregulation, particularly in those with a genetically determined susceptibility.
Conclusion
The cross-species role of reproductive hormones in depressive behavior suggests a neuroendocrine pathophysiology for PPD. PPD, as defined in contemporary research, includes depression that began during or before pregnancy; depression that occurred in the context of a childhood trauma history, traumatic labor or delivery, subthreshold thyroid dysfunction, psychosocial stress, or sleep deprivation; and depression that co-occurred with obsessive-compulsive disorder, posttraumatic stress disorder, generalized anxiety disorder, or personality pathology. Logic would preclude consideration of all of these as the same disorder; consequently, when attempting to understand the contribution of hormonal signaling to postpartum affective dysregulation, it is therefore necessary to carefully define the study population and attempt to characterize and disentangle individual PPD phenotypes. The extant literature supports the existence of a hormone-sensitive PPD phenotype.Reference Bloch, Schmidt, Danaceau, Murphy, Nieman and Rubinow 3 In order to study the clinical and neuroendocrine correlates of this phenotype, some researchers have selected women with a history of PPD and without a history of nonpuerperal depressive episodes.Reference Bloch, Schmidt, Danaceau, Murphy, Nieman and Rubinow 3 , Reference Schiller, O’Hara, Rubinow and Johnson 18 Although these studies are primarily relevant for understanding the risk of PPD recurrence, they represent the first step toward identifying factors that predict first-onset PPD. There is sufficient evidence to suggest that reproductive hormone fluctuations trigger affective dysregulation in sensitive women. Even within the hormone-sensitive phenotype, alterations in multiple biological systems—the immune system, HPA axis, and lactogenic hormones—likely contribute to the pathophysiology of PPD. Studies are underway to disentangle the complex interplay of fluctuating reproductive hormones, neurosteroids, HPA axis reactivity, neural dysfunction, and genetics with a specific focus on identifying genomic, brain, and behavior relationships that contribute to affective dysfunction in the context of specific reproductive states. Consistent with the RDoC mission, this line of research represents not only an opportunity to identify novel treatment targets for PPD but also—critically—the potential to prevent PPD in susceptible women.
Disclosures
Crystal Schiller has the following disclosures: UNC Building Interdisciplinary Careers in Women’s Health (BIRCWH) Career Development Program (K12 HD001441), National Institute of Mental Health of the National Institutes of Health—Award Number R21MH101409, and Brain and Behavior Research Foundation. Samantha Meltzer-Brody has the following disclosures: Foundation of Hope, research, grant. David Rubinow has the following disclosures: active grants from NIH/NIMH: “Postdoctoral Training in Reproductive Mood Disorders,” T32 MH93315; “Identifying Biomarkers for Post-Partum Depression in African-American Women.” R01MH095992; “Neuroendocrine Mechanisms of Reproductive Hormone Related Affective Dysfunction,” R21MH101409; “Depression, Estrogen Replacement, and Cardiovascular Health in the Perimenopause,” 1R01MH087619-01A1; “Continuous OC Treatment in PMDD: Steroid Hormone Mechanisms in Premenstrual Dysphoric Disorder,” R01MH081837-01.
