Historical Background
In 1892, Emil Kraepelin postulated links between hormones and dementia praecox and reviewed the endocrine status of his patients.
Other early researchers such as Hoskins (1929) studied endocrine changes in schizophrenia in postmortem studies. During this period, the discovery of insulin stimulated further interest in the interaction between behavior and metabolism. Hoskin’s work ultimately discounted the efficacy of treating schizophrenia by insulin-induced hypoglycemia and provided an important early focus on endocrine research in schizophrenia.
Between 1940 and 1970, there was considerable interest in the psychoendocrine characterization of schizophrenia. The mimicking of psychotic symptoms by the administration of high doses of steroid hormones by Mason (1975) and the demonstration of elevated levels of the steroid metabolic 17-hydroxycorticosteroid in acutely ill people with schizophrenia by Sachar et al. (1963) represent two important early neuroendocrine studies in schizophrenia. Brambilla and Penati (1978) reviewed the evidence for hypo- and hyperfunction of the adrenal, pituitary, and thyroid glands and of the gonads in patients with schizophrenia. Endocrinopathies were only rarely present, suggesting that endocrine abnormalities in these patients were due to the disease process of schizophrenia rather than being causally related.
With increasing clarity of the role of neurotransmitters in the regulation of pituitary hormone release via hypothalamic hormones, neuroendocrine studies in schizophrenia conducted in the early 1980s aimed to use the pituitary gland as the window to the brain, using probes that modified the secretion of anterior pituitary hor- mones to detect abnormalities in tuberoinfundibular-pituitary func-
Clinical Estrogen Trials in Patients with Schizophrenia
Jayashri Kulkarni
tion, reflecting similar abnormalities in the mesolimbic system.
Biological research in schizophrenia in the 1980s and 1990s largely excluded women from neuroendocrine studies, citing menstrual cycle effects as major confounding factors. More recently, the pioneering research of Häfner (1991), and Seeman and Lang (1990) has brought about a long overdue focus on gender differences in schizo- phrenia. This, in turn, has led to considerable attention being paid to the role of female hormones or functioning of the hypothalamic – pituitary – gonadal (HPG) axis in patients with schizophrenia, with the resultant “estrogen protection hypothesis” – that is well de- scribed in chap. 2 of this book.
Animal Studies
The effects of estrogen on the dopamine system are complex and the effects of estrogen on dopaminergic neurotransmission are believed to depend on the dose and length of administration. Studies have demonstrated that, within 24 h of low-dose estradiol being adminis- tered to rats, there is a significant decrease in the ratio of high-to- low-affinity agonist states of the striatal dopamine D2 receptors (Gordon et al. 1980; DiPaolo et al. 1982; Joyce et al. 1982); this has been described as an acute hyposensitive state. On the other hand, administration of high doses of estrogen results in the development of striatal dopamine receptor hypersensitivity (Perry et al. 1981;
Clopton and Gordon 1983). This can be expressed behaviorally in rats as an increase in dopamine agonist-induced stereotypes or biochem- ically as an increase in the density of [3H] spiroperidol-binding sites (Gordon et al. 1980; Gordon and Perry 1983). Ferretti et al. (1992) found that estrogen administration had little effect on dopamine D1
receptors, but that D2 receptor density fell in response to low-dose estradiol.
Behrens et al. (1992) studied the effects of estradiol and testos- terone on cataplexy induced by haloperidol in rats and concluded that estradiol downregulated dopamine transmission.
More recently, Fink and colleagues (1998; 1999) have shown that estrogen induces a significant increase in 5-HT2Areceptors and the serotonin transporter (SERT) in regions of the rat forebrain that, in humans, are associated with mental state, mood, cognition, memory, emotion, and neuroendocrine control. The precise mechanism of
estrogen-serotonin interaction is not clear. However, the forebrain receives a dense innervation of serotonergic projections from the midbrain raphe neurons, and the 5-HT2Areceptors are present at high concentrations in most areas of forebrain, in particular in the anterior and frontal cingulate cortex – regions of the forebrain associated with cognition and emotion. The highest density of these receptors is in laminae IV and VA of these regions of the cortex, where the receptors are located on the apical dendrites of the pyramidal neurons and on glutamatergic interneurons.
The net effect of 5-HT stimulation of the 5-HT2A receptors in frontal and cingulate cortex would probably be to increase the rate of firing of the pyramidal neurons, although this might be offset by inhibitory effects mediated by 5-HT1receptors. Since the functional density of 5-HT receptors is influenced by the concentrations of 5-HT to which they are exposed, the density and degree of activity of serotonin transporters (SERT) may also have a bearing on 5-HT2A receptor activity. However, the estrogen-induced increase in the density of SERT sites occurs in areas of the forebrain (amygdala, lateral septum, and venteromedial nucleus of the hypothalamus) in which estrogen has no apparent effect on the density of the 5-HT2A receptors.
The action of estrogen on 5-HT2Areceptors and the SERT could be mediated by both genomic and nongenomic mechanisms. Fink et al. (1998) reported that estrogen induced a two- to threefold in- crease in the amount of 5HT2Areceptor messenger RNA in the dor- sal raphe nucleus, suggesting that exposure to estrogen stimulated 5HT2Areceptor gene transcription. This is in contrast to the finding of an estrogen-induced rise in 5HT2A receptor density without a concordant rise in 5HT2A receptor messenger RNA in cortical neurons. In turn, this is analogous to the action of estradiol on D2
receptor-expressing neurons in the striatum, where there are few estradiol receptors. The action of estradiol in the striatum may be mediated by membrane receptors rather than classical cytoplasmic receptors. Mosselman et al. (1996) cloned a second estrogen receptor and termed it estrogen receptor-β (as distinct from estrogen recep- tor-α). Shughrue et al. (1997), Fink et al. (1998), and Sumner et al.
(1999) have provided data that suggest that any genomic action of estrogen on the 5HT2Areceptor and SERT genes may be mediated by estrogen receptor-β.
Estrogen, by way of its actions on the 5-HT2Areceptor, the SERT, and the D2 receptor, may protect against depressive and psychotic symptoms. This apparent psychoprotective effect of estrogen may have a major biological purpose in that, teleologically, the actions of estrogen may be related to its role as the major endocrine coordinator of events that lead to fertility and reproduction. Estrogen triggers the ovulatory luteinizing hormone-releasing hormone/luteinizing hor- mone surge which influences mating behavior in nonhuman pri- mates. In order for mating to occur, mood and mental state must be right in both humans and other animals. Hence, estrogen triggers and coordinates several biological events which are necessary for pro- creation.
The significance of these findings for schizophrenia lies in the fact that the 5HT2Areceptor is the target for the atypical antipsychotics, such as olanzapine and clozapine, which are particularly effective in the treatment of the negative symptoms of the illness (Fink 1995).
Thus, the effect of estrogen on 5-HT2Areceptors could be part of the explanation for the findings encompassed in the estrogen protection hypothesis. There is no conflict between the possible involvement of serotonin and the dopamine hypothesis of schizophrenia, since estrogen increased density of both the dopamine D2, and the 5HT2A receptor, such that the two mechanisms could operate in parallel (Fink 1995; Fink et al. 1998).
Clinical Studies
Following these epidemiological, clinical, and animal study results, we conducted an open-label pilot study (Kulkarni et al. 1996) in which 11 women of child-bearing age with schizophrenia were given 0.02 mg ethinyl estradiol orally as an adjunct to antipsychotic drug treatment for 8 weeks and their progress compared with a similar group who received antipsychotic drugs only. The group receiving estrogen made a significantly more rapid recovery from acute psy- chotic symptoms and also reported improvement in their general health status.
Subsequent to this, we conducted a dose-finding study for the optimal use of estradiol in women with schizophrenia (Kulkarni et al. 2001). This was a three-arm, double-blind, placebo-controlled, 28- day study in which 12 women received 50 µg transdermal estradiol
plus standardized antipsychotic drug, 12 women received 100 µg transdermal estradiol plus standardized antipsychotic drug, and 12 women received placebo plus standardized antipsychotic drug. All of the women had a diagnosis of DSM-IV schizophrenia.
Women were excluded from the trial if they had any known endocrine abnormalities, were pregnant or lactating, were currently taking synthetic steroids, including the oral contraceptive pill, or were using illicit drugs.
Each subject was enrolled in the trial for 28 days (one menstrual cycle) and received a baseline psychopathology and hormone assess- ment followed by assessments at days 4, 7, 14, 21, and 28. At each assessment, psychopathology was evaluated using the Positive and Negative Syndrome Scale (PANSS; Kay et al. 1987). The PANSS is comprised of three subscales – the positive symptom subscale, the negative symptom subscale, and the general symptom subscale.
Hormone assays for serum estrogen (E2), progesterone (Prog), pro- lactin (PRL), luteinizing hormone (LH), follicle-stimulating hormone (FSH), and testosterone (Test) were performed. Separate radio- immuno-assay tests were conducted for each individual sample. A Menstrual Cycle Interview (MCI; Kulkarni et al. 1996b) was used to stage the patients’ menstrual cycle phase.
All patients were randomized into the active estradiol skin patch treatment groups or an identical placebo patch group. For trans- dermal delivery adhesive skin patches were utilized that contained either 4 mg estradiol per patch, with a release rate of 50 µg per 24 h, or 8-mg estradiol patches, with a release rate of 100 µg estradiol per 24 h. Both estrogen and placebo patches were changed every 4 days.
Placebo patches were also adhesive but had no active substance. All patients received antipsychotic drug treatment, which was adminis- tered according to a protocol that indicated doses of between 3 and 6 mg per day of risperidone. The dose of risperidone was dependent on the patient’s clinical state.
Results
The group receiving adjunctive 100 µg estradiol showed greatest improvement across the study compared to the other two groups, although the 50-µg adjunct group showed more improvement than the placebo adjunct group. Improvement scores differed significantly
between the three groups (p < 0.005). The 100-µg group improv- ed more than the placebo group (p = 0.001) and the 50-µg group (p = 0.001), whereas the latter two groups did not differ significantly.
The mean change from baseline for the positive symptom subscale of the PANSSdiffered between the three groups (p < 0.005).
The 100-µg group improved more than the placebo group (p = 0.002) and the 50-µg group (p = 0.001), whereas the latter two groups did not differ significantly from each other. The mean change from baseline for the negative symptom subscale of the PANSS showed no statis- tically significant difference between the three groups, although there was a significant difference between the 100-µg estrogen and placebo groups (p = 0.039). The mean change from baseline for the general symptom subscale of the PANSSrevealed a significant dif- ference between the groups (p = 0.001). The 100-µg group improv- ed more than the 50-µg group (p = 0.001) and the placebo group (p = 0.002), whereas the 50-µg and placebo groups did not differ signif- icantly from each other.
Hormone data analysis showed that the 100-µg group had signifi- cantly lower mean LH levels than the 50-µg group (p = 0.027), suggesting that 100 µg adjunctive estradiol had greater impact on the pituitary and perhaps, therefore, a direct neuroleptic effect on the dopamine and serotonin systems.
Follow-up of the 100-µg Estradiol Group
To further investigate the significant improvement in psychotic symptoms with the addition of 100 µg of estradiol to antipsychotic drug treatment, we studied 36 women with DSM-IV schizophrenia further.
In this 28-day, double-blind study, all women were randomized to the active 100-µg estradiol skin patch treatment group or an identi- cal placebo patch group. All patients received antipsychotic drug treatment of risperidone, according to a standardized protocol. Psy- chopathology was assessed by the PANSS rating scale and serum hormone levels of estrogen, progesterone, prolactin, LH, and FSH were measured. A battery of cognitive tests was also administered.
Results
There were no significant differences between the 100-µg and placebo adjunct groups in mean age, menstrual cycle phase, race, illness duration, or antipsychotic drug dose.
Mean Baseline PANSS for the two groups were: 86.3 +/– 27.7 for the 100-µg estradiol group and 73.3 +/– 12.3 for the placebo group.
This difference is statistically significant (p < 0.05). To allow for correction of this difference in the baseline scores, the change across time was calculated.
Change from baseline in the total PANSS and the positive, ne- gative, and general subscales across time are shown (Figs. 1–4).
The changes in hormone data across 28 days for the two groups revealed that the estrogen group had significantly higher mean estrogen levels and higher testosterone and lower LH levels. There were no differences in progesterone, FSH, prolactin, or testosterone levels between the estrogen adjunct and placebo adjunct groups.
Conclusion
The results to date supply further evidence suggesting that the addition of 100 µg estradiol transdermally provides greater im-
Fig. 1. Mean change from baseline in total PANSS scores for both groups
Mean change from baseline
provement in the treatment of psychotic symptoms in women with DSM-IV schizophrenia than standardized antipsychotic drug treatment alone. The impact of administering 100 µg of estradiol transdermally, as indicated by its effect on the pituitary measured by LH assay, suggests that this dose and type of unconjugated estrogen
Fig. 3. Mean change from baseline in PANSS Negative Subscale scores for both groups
Fig. 2. Mean change from baseline in PANSS Positive Subscale scores for both groups. The group receiving estrogen (100 µg) had a more significant decrease
in psychotic symptoms as measured by the PANSS rating scale
Mean change from baselineMean change from baseline
affects CNS neurotransmitter systems positively. This is consistent with the “estrogen hypothesis” as formulated by Häfner (1991).
Clinical Adjunctive Estrogens Trial in Men with Schizophrenia
In one of our studies (Kulkarni et al. 1995), we measured baseline gonadotropin and sex steroid hormones in men and women suffering from schizophrenia compared with age- and sex-matched controls. A major finding was that 19 men with schizophrenia had average testosterone levels of 26.6 mg compared with 17.7 mg in 18 healthy age-matched male controls. The male patients had significantly higher testosterone LH and FSH levels than controls. In effect, the male patients had a “puberty-like” gonadal axis profile. However, the patients had normal secondary sexual characteristics and no endo- crine-related illnesses. The patients had all been off neuroleptics for at least 3 months at the time of testing. We also correlated testos- terone levels with psychopathology scores and found a significant, positive correlation between testosterone levels and increasing SAPS and SANS score. The exact opposite results were found in the women we studied. Female patients had lower estrogen levels and higher LH
Fig. 4. Mean change from baseline in PANSS General Subscale scores for both groups
Mean change from baseline
levels than age- and cycle phase-matched female controls. In effect, the female patients had a “menopause-like” gonadal axis profile – even though the average age for the patient group was 26.4 years and they were not menopausal. Estrogen levels correlated inversely with psychopathology scores.
According to these data from clinical and preclinical studies, estrogen appears to have a potentially positive effect on psychotic symptoms, most likely due to its dopamine-modulating effect. In view of the positive response found in our adjunctive estrogen trials in women, we believe that adding short-term, low-dose estradiol to standardized antipsychotic drug treatment in men with schizo- phrenia could improve their response to neuroleptics.
Results of a Pilot Study
In a pilot study conducted in 1999, 2 mg estradiol valerate was given as an adjunct to six men who were receiving antipsychotic drugs and five men received oral placebo plus their standard antipsychotic medication. Oral estrogen, rather than transdermal estrogen was used to ensure compliance in acutely psychotic men. Side effects of estrogen therapy in men such as gynecomastia, decreased libido, and fluid retention are not reported in studies using less than 1.25 mg estrogen per day for less than 4 weeks.
Psychopathology was assessed using PANSS and the Brief Psychiatric Rating Scale (BPRS). High scores on both these scales indicate more severe symptoms. Both groups commenced antipsy- chotic drugs plus estrogen or placebo on day 1 of the trial. However, the dose and type of antipsychotic drug was not standardized in this small pilot study.
Summary of Pilot Study Results
The t-tests results from the pilot study show no significant difference in psychopathology scores between the two groups at the start of the study. At day 5, the adjunctive estrogen group had a significantly lower positive PANSS, BPRS, score (p < 0.05) than the placebo adjunct group. The estrogen group had a significantly higher negative PANSS score (p < 0.05). There were significant differences in mean hormone levels between the two groups at the start of the study. At day 7, the
estrogen group had significantly higher mean estrogen levels, higher LH levels, and higher prolactin levels. There was no difference in testosterone levels of FSH at day 7 between the groups.
Current Study
We are currently investigating the effect of adding 2 mg oral estradiol velate to standardized antipsychotic drug treatment in a group of men with schizophrenia for a 2-week, double-blind trial. A target number of 60 patients will be recruited over a 3-year period, with equal numbers being allocated to either the adjunctive estradiol or adjunc- tive place group, for 2 weeks.
Results
A total of 16 men with DSM-IV schizophrenia have been recruited since May 2001. Eight received 2 mg adjunctive estradiol orally and eight received adjunctive oral placebo. All patients received standard- ized antipsychotic drugs.
Demographic data revealed no differences between the two groups in terms of age, race, diagnosis, illness duration, or antipsychotic drug treatment.
The mean PANSS data for the estrogen and placebo groups as shown in Table 1 reveals that the estrogen adjunct group was not as well at the start of the trial.
To compare the improvement in the two groups across the 14-day trial, changes from baseline measures were calculated (Table 2).
The results show that the estrogen adjunct group made a more significant improvement over the 14-day trial as shown by the decrease in PANSS scores from baseline. In particular the most sig- nificant differences were seen in the PANSS general subscale symptoms (see Figs. 5–8).
Hormone Levels
There were significant changes between the two groups: the estrogen group had significantly higher day 14 estrogen levels (p = .002) and lower testosterone levels (p = 0.04). The change in LH levels from baseline measurements was not significant across time.
This study is in progress and is also a dose-finding study. Patients are closely monitored and no serious adverse effects have been noted:
Conclusions to Date
The use of estrogen as a potential treatment in schizophrenia opens up exciting new avenues of preventive and acute treatment of schizo- phrenia in both men and women. The rapid development of new estrogen compounds, selective estrogen receptor modulators – or
“brain estrogens,” has further expanded the area of hormone treatment. Studying the mechanisms by which estrogen potentially
Table 1. Mean PANSS data for the estrogen and placebo groups
E2 group Placebo group
Baseline Day 7 Day 14 Baseline Day 7 Day 14 Total mean
PANSS ± SD 75.4±9.6 68.0±13.8 63.0±9.0 67.6±3.6 64.6±8.4 63.2±7.1 Total mean
positive Sx 21.0±5.7 18.3±6.4 16.3±5.1 21.7±3.5 19.0±3.1 19.2±3.1 Total mean
negative Sx 16.6±4.8 15.5±2.4 14.3±2.0 13.2±2.9 12.4±2.3 13.0±1.8 Total mean
general Sx 37.7±4.3 34.3±6.7 30.3±4.1 32.6±1.7 32.0±5.1 31.5±3.6 (PANSS scores in bold denote significant differences p < 0.05)
Table 2. Mean PANSS change from baseline scores for estrogen and placebo groups
E2 group Placebo group pvalue
PANSS change from
baseline to day 14 -12.38+/6.4 -4.38+/6.1 0.04
Positive symptom
change 0-day 14 -3.6+/2.6 -2.5+/4.4 NS
Negative symptom
change 0 –day 14 -2.3+/5.1 -0.75+/4.1 NS
General symptom
change 0 – day 14 -6.7+/4.3 -1.1+/4.1 0.02
Fig. 5. Mean change from baseline for total PANSS scores for the two groups
Fig. 6. Mean change from baseline for PANSS positive subscale for the two groups
Mean change from baseline
Mean change from baseline
Total PANSS
Time (days)
Treatment
Mean change from baseline
Mean change from baseline
PANSS Positive Subscale
Time (days)
Treatment
2 mg Estradiol N = 8
Placebo N = 8
2 mg Estradiol N = 8 Placebo N = 8
Fig. 7. Mean change from baseline for PANSS negative subscale for the two groups
Fig. 8. Mean change from baseline for PANSS general subscale for the two groups
Mean change from baseline
Mean change from baseline
PANSS Negative Subscale
Time (days)
Treatment
Mean change from baseline
Mean change from baseline
PANSS General Subscale
Time (days)
Treatment
2 mg Estradiol N = 8
Placebo N = 8
2 mg Estradiol N = 8
Placebo N = 8
represents an option for treating psychotic symptoms may also promote our understanding of the etiological aspects of schizo- phrenia, in particular, the reasons why schizophrenia is a postpu- bertal disease.
References
Behrens S, Häfner H, De Vry J, Gattaz WF (1992) Estradiol attenuates dopamine- medicated behavior in rats. Implications for sex differences in schizophrenia.
Schizophr Res 6: 114
Brambilla F, Penati G (1978) Perspectives in endocrine psychobiology. In: Brambilla F and Bridges P (ed). Perspectives in endocrine psychobiology. Witey, London, 309-422
DiPaolo T, Payet P, Labrie F (1982) Effect of prolactin and estradiol on rat striated dopamine receptors. Life Sci 31: 2921-2929
Ferretti C, Blengio M, Vigna I, Ghia P, Gerazzani E (1992) Effects of estradiol on the ontogenesis of striatial dopamine D1 and D2 receptor sites in male and female rats. Brain Res 571: 212-217
Fink G (1995) The psychoprotective action of estrogen is mediated by central serotonergic as well as dopaminergic mechanisms. In: Takada A and Curzon G (eds). Serotonin in the central nervous system and periphery. Elsevier Science, Amsterdam, 175-187
Fink G, Sumner B, McQueen JK, Wilson H, Rose R (1998) Sex steriod control of mood, mental state and memory. Clin Exp Pharmacol Physiol 25: 764-765 Fink G, Sumner B, Rosie R, Wilson H, McQueen J (1999) Androgen actions on central
serotonin neurotransmission: relevance for mood, mental state and memory.
Behav Brain Res 105: 53-68
Gordon HH, Borison RL, Diamond BI (1980) Modulation of dopamine receptor sensi- tivity by estrogen. Biol Psychiatry 15: 389-396
Gordon JH, Perry KO (1983) Pre- and postsynaptic neurochemical alteration following estrogen-induced striatal dopamine hypo-and hypersensitivity. Brain Res Bull 10: 425-428
Häfner H (1991) The epidemiology of beginning schizophrenia. Presented at the WPA Section of Epidemiology and Community Psychiatry Symposium, June 14-16, Oslo
Hoskins RG (1920) Endocrine factors in dementia praecox. N Engl J Med 200: 361 Joyce JN, Smith RL, Van Hartesveldt C (1982) Estradiol suppresses then enhances intracaudate dopamine-induced contralateral deviation. Eur J Pharm 81: 117-122 Kraepelin E (1982) Psychiatrie, vol 3, part 2, 8th ed. Translated (1919). Dementia
praecox and paraphrenia. Livingstone, Edinburgh
Kulkarni J, de Castella A, Riedel A, Taffe J, Fitzgerald P, Burger H (2001) Estrogen – a potential new treatment in schizophrenia. Schizophr Res 48: 137-144
Kulkarni J, Smith D, McKenzie D, Hill C, Keks N, Singh B, Copolov D (1995) Donadotrophin response to naloxone challenge in female and male psychotic patients. Biol Psychiatry 38: 701-703
Kulkarni J, de Castella A, Smith D, Taffe J, Keks N, Copolov D (1996a) A clinical trial of the effects of estrogen in acutely psychotic women. Schizoph Res 20:
247-252
Kulkarni J, Gostt K, de Castella A (1996b) The menstrual cycle in women with schizophrenia. Schizophr Res 18: 254
Mason JW (1975) Emotion as reflected in patterns of endocrine investigation. In: Levi L (ed) Emotions: their parameters and measurement. Raven Press, New York, 143-181
Mosselman S, Polman J, Dukema R (1996) ER beta: identification and character- ization of a novel human estrogen receptor. FEBS Lett 392: 49-53
Perry KO, Diamond BI, Fields JZ, Gordon JH (1981) Hypophysectomy induced hyper- sensitivity to dopamine: antagonism by estrogen. Brain Res 226: 211-219 Sachar EJ, Mason JW, Kolmer HS, Arfess KL(1963) Psychoendocrine aspects of acute
schizophrenic reactions. Psychosom Med 25: 510
Seeman MV, Lang M (1990) The role of estrogens in schizophrenia gender differ- ences. Schizophr Bull 16: 185-195
Shughrue PJ, Lane MV, Merchenthaler I (1997) Comparative distribution of estrogen receptor and X and B MRNA in the rat central nervous system. J Compr Neurol 388: 507-525
Sumner BEH, Grant DE, Rosie R, Hegele-Hartung Ch, Fritzemeier KH, Fink G (1999) Effects of tamoxifen on serotonin transporter and 5-hydroxytryptamine 2A receptor binding sites and mRNA levels in the brain of ovariectomized rats with or without acute estradiol replacement. Mol Brain Res 73: 119-128
