Estrogen is a potent gonadal steroid that can have dynamic effects in the brain. A rich literature encompassing both preclinical and clinical studies describes the effect that estrogen can exert on cognition, mood, and behaviour. The past two decades have seen an increased interest in the role of estrogen in the pathophysiology and treatment of schizophrenia [1–3]. This chapter aims to critically analyse the current evidence for the utility of 17β-estradiol and its analogues as a form of therapy for schizophrenia, with particular focus on the feasibility of the treatment for the cognitive deficits associated with the disorder.

Schizophrenia is a complex neuropsychiatric disorder that will affect approximately seven to eight individuals per 1,000 during their lifetime [4]. The disorder is characterised by three broad categories of symptoms: positive, negative, and cognitive symptoms [5]. The positive and negative symptoms are often considered the most debilitating; however, cognitive deficits are the best predictor of functional outcome [6, 7]. Current antipsychotic drugs are not suitable for the effective treatment of the cognitive dysfunctions associated with schizophrenia; thus, novel treatments require exploration [8]. It has previously been theorised that second-generation antipsychotics can provide some cognitive benefit in patients [9]. However, more recent research suggests observed outcomes could be attributed to poor study design, practise effects, and inappropriate doses of medication [10].

Cognition is a broad term referring to the mental processes related to acquiring knowledge and understanding [11]; measurable areas of cognition can include learning, processing speed, memory, and reasoning. In this chapter we focus on information processing and learning and memory. It is well established that individuals with schizophrenia suffer from working memory problems [12]. A meta-analysis by Forbes et al. in 2009 [12] concluded there are large deficits in all domains of working memory (central executive, visuospatial, phonological) in schizophrenic patients compared with healthy controls.

Cognitive and intellectual underperformance has been consistently identified as a risk factor for schizophrenia. Decline in cognitive ability precedes the onset of clinical symptoms by nearly a decade [13]. A meta-analysis of neurocognitive function has found schizophrenia patients perform on average 1.5–2.5 standard deviations below the norm on neurocognitive tests of attention, motor performance, memory, and general intelligence [10]. Importantly, even when the positive and negative symptoms of schizophrenia are in remission, cognitive deficits remain [14]. This demonstrates the robustness of the cognitive dysfunctions associated with the disorder, in addition to the limited capability of current pharmacological treatments to improve function [15].

Estrogen can exert potent effects in numerous regions of the brain, consequently affecting mood, cognition, and behaviour [28, 29]. Physicians first noticed the benefits of hormones for psychiatric illness and menopausal symptoms over 100 years ago [30]. In the past 50 years, research into estrogen therapy has markedly advanced. Its medical utility has been explored in breast [31] and prostate cancer [32], and osteoporosis [33]. Moreover, researchers continue to study the use of estrogens for the delay or alleviation of neurodegenerative diseases, and treatment of psychiatric disorders including treatment-resistant depression [34], perimenopausal and postpartum depression [35, 36], bipolar affective disorder [37], and schizophrenia [2].

Evidence suggests that steroidal hormones, such as estrogen, exert their effects over the entire lifetime, protecting the brain from certain insults [38]. Accumulating evidence has led to the hypothesis that reoccurring hormone influxes in women serve as a protective factor in the initial development of schizophrenia [39]. The hypothesised role of estrogen in schizophrenia is not a recent theory. In 1961, Diczfalusy and Lauritzen [40] reviewed studies measuring estrogen concentration in the blood and urine of women with schizophrenia [40, 41]. In seven of the eight studies examined, low estrogen levels were detected [40]. This has contributed to the hypoestrogenism hypothesis, which posits that low 17β-estradiol in schizophrenia is either a trigger or initial vulnerability leading to development of the mental illness, or conversely, an outcome of the disease itself [41]. While these studies are over 70 years old, admittedly used small sample sizes, and applied laboratory methods now considered outdated, the research is of significance due to its occurrence during the pre-antipsychotic era [41]. The current use of antipsychotic treatment in women with schizophrenia, particularly typical antipsychotics, affords difficulty in reliably assessing hormone levels due to hyperprolactinemia [42]. A review of research into antipsychotic-induced hyperprolactinemia in schizophrenia patients found 67% of the sample of women had abnormal levels of prolactin [42]. Elevated prolactin levels occur when antipsychotics block dopamine D₂ receptors on the anterior pituitary gland [42], and consequently can suppress gonadal function.

Despite the potential confounding effect of antipsychotics, research continues to demonstrate that low estrogen plasma concentration correlates with an increased risk of symptoms of schizophrenia [41]. For example, in pre-menopausal hospital patients with schizophrenia, Bergemann et al. [43] found a significant increase in admissions during the 3 days prior to and following the first day of menses. Further, in a more recent study Bergemann et al. [44] assessed plasma 17β-estradiol concentration in schizophrenia patients, while also controlling for antipsychotic treatment; hypoestrogenism occurred in approximately 60% of their sample. In addition, there was a significant difference in prolactin levels between the patients in the typical and atypical antipsychotic groups, however, no difference between the two atypical antipsychotic groups (clozapine and olanzapine). Ultimately, Bergemann et al. concluded that low serum levels of 17β-estradiol exist independent of antipsychotic-induced elevation of prolactin levels [44].

Gender differences in schizophrenia, and comparison of endocrinological function in schizophrenic women compared to the healthy population, provides a case for the estrogen hypothesis. Evidence including lower baseline levels of circulating estrogen, amenorrhea, [45], the association between earlier onset of puberty and later onset of the disorder [27], and superior response to antipsychotic medication in women compared to men, suggests estrogen can exert a protective effect in schizophrenia, and consequently may serve as an appropriate form of treatment. Molecular findings further strengthen the clinical observations and evidence for the estrogen hypothesis [46]. Most notably, Weickert et al. [47] discovered that estrogen receptors are altered in the brains of individuals with schizophrenia, consequently affecting their ability to respond to endogenous estrogen. Clinical and molecular findings over the past 50 years provide a substantial argument for the use of estrogen therapy in women with schizophrenia; however, the most appropriate estrogenic compound for long-term use is yet to be determined.

There exist numerous forms of endogenous estrogen; however, 17β-estradiol is considered the most potent form. Although it is often considered the primary ‘female sex hormone’, it is present in both sexes [48]. While 17β-estradiol is predominantly produced in the ovaries to regulate the menstrual cycle in females, it is also created by non-endocrine tissues, including fat, breast, and neural tissues [49]. It is important to note that reference to estrogen treatment, particularly in early research, can broadly refer to numerous estrogenic compounds including estrone, diethylstilbestrol, equilin, 17β-estradiol, and ethinylestradiol. From here on, this chapter will specifically focus on the estrogen 17β-estradiol, unless otherwise stated.

Generally, there is no clinical necessity for healthy women of childbearing age to be treated with 17β-estradiol. Although the effects of estrogen-based contraceptives in healthy young women have been investigated [67, 68], the contraceptive formulations do not include 17β-estradiol, but rather synthetic derivatives (e.g., ethinylestradiol). Therefore, research examining the effect of 17β-estradiol treatment has primarily been concerned with neurodegenerative diseases [69], psychiatric disorders [2], and the postmenopausal population [70]. Perhaps the most notable study in this area of research is the Women’s Health Initiative (WHI) memory study, a large double-blind, randomised placebo-controlled trial, which investigated cognitive function in a postmenopausal sample [71]. Following cognitive testing of 1,416 women receiving hormone replacement therapy, researchers found there was no beneficial effect of estrogen on cognition [71]. Importantly, participants partaking in the WHI study were treated with conjugated equine estrogens (CEE). Conjugated estrogens primarily contain estrone and equilin, with lesser quantities of 17β-estradiol [56]. This is important to note, as 17β-estradiol and the main component of CEE, estrone, have shown opposing effects on cognition in preclinical research [56]. Estrone has been found to impair spatial working memory [72], while 17β-estradiol has the opposite effect [73]. The discrepancy between CEE and 17β-estradiol is often not outlined in research administering CEE as estrogen therapy, which has consequently contributed to the incorrect conclusion that 17β-estradiol is not beneficial, or is even harmful, for cognitive function in postmenopausal women. Research concerning estrogen treatment and CEE in the postmenopausal population will not be detailed further in this chapter; however, see Fischer et al. [28] for an overview of hormone therapy relevant to cognition in postmenopausal women. Similarly, Gogos et al. [68] recently reviewed literature concerning ethinylestradiol-based oral contraceptives and their effect on cognition in pre-menopausal women.

The beneficial effect of 17β-estradiol on cognition has been consistently replicated in animal studies. It should be noted that this section of the chapter only refers to studies of 17β-estradiol in rat cognition; see Gibbs [84] for research relevant to mice and non-human primates. In rats (Table 26.1), the beneficial effect of estrogen has been seen in learning and memory including spatial working, recognition, and reference memory domains [53, 73, 131, 135]. Luine et al. [73] examined spatial memory in OVX rats in an eight-arm radial maze; 3 days of 17β-estradiol treatment via subcutaneous implant did not enhance memory performance compared to the untreated OVX rats; however, 12 days of treatment significantly improved memory. Using the novel-object and placement-recognition paradigm, Luine et al. [53] demonstrated that 17β-estradiol treatment enhanced visual and place memory in OVX rats compared to controls. In addition, estrogen treatment enhanced memory when given prior to or immediately after the sample trial, but not, however, when administered 2 h later. Thus, it is theorised that 17β-estradiol treatment affects memory encoding or consolidation, rather than retrieval. Interestingly, authors also determined that compared to novel-object recognition, a different dose of 17β-estradiol for novel-place recognition was necessary to see a significant effect [53].

Dose-dependent effects of 17β-estradiol have also been demonstrated in reference memory [55]. In their 2010 review, Barha and Galea concluded that high levels of 17β-estradiol can impede working and reference memory, whereas low levels of 17β-estradiol have no significant effect on reference memory, however can facilitate working memory [55]. Similarly, contextual fear conditioning can be facilitated by a low dose of 17β-estradiol, however, a high dose can impair [55, 115]. This demonstrates the capacity of 17β-estradiol to differentially affect forms of memory, all of which are hippocampus-dependent. Collectively, the literature indicates an inverted U-shaped dose–response curve [54]; lower and higher doses of 17β-estradiol can often inhibit or impair cognition [56].

Numerous studies have shown that behavioural tasks employing the hippocampus can be altered by 17β-estradiol [88, 92, 99, 102, 107, 112]; however, fewer have investigated cortical-dependent tasks [29]. Wide et al. [150] examined the effect of 17β-estradiol in OVX rats in the non-spatial delayed alternation task, mediated by the integrity of the prefrontal cortex. A lower dose of 17β-estradiol was most effective for facilitation of non-spatial working memory; subjects receiving a high dose made significantly more errors compared to the controls, demonstrating that a task considered primarily prefrontal cortical-dependent can also be affected by 17β-estradiol [150].