The Mini-TRH Test, Dopamine Transmission, and Schizophrenia Symptoms

Abstract

Studies in animals and humans suggested that the tonic dopamine inhibition of prolactin release may be estimated by submaximal prolactin stimulation by thyrotropin-releasing hormone (TRH), the mini-TRH test. Because patients with schizophrenia may be more vulnerable to stress-induced elevations of prolactin, great care was taken to avoid stress-induced increases in prolactin, including applying local anaesthesia before blood extraction in our psychotic patients. Basal prolactin levels were in the reference range in all psychotic patients studied by us and were not higher in male patients than in normal men. Results of the mini-TRH test suggested that in acute patients with non-affective psychoses, everyday memory problems, non-paranoid delusions, and first-rank symptoms, but not other Comprehensive Psychopathological Rating Scale (CPRS) positive symptoms, could correlate with decreasing dopamine transmission in lactotrophs. In acute patients with first-episode schizophrenia, increasing negative disorganisation symptoms might correlate with increasing dopamine transmission. In first-episode patients, a hypersensitivity of the TRH response was detected, which could indicate that variability in the basal prolactin levels may confound the interpretation of the mini-TRH response. To avoid that, a smaller dose of TRH was recommended in first-episode patients. Studies using other estimates of basal dopamine release suggested that striatal dopamine transmission reflected delusions and hallucinations but not other Positive and Negative Symptom Scale (PANSS) positive symptoms. Including a wide range of symptoms in the PANSS positive scale may reduce its specificity for assessing basal dopamine transmission, although the scale remains useful for tracking treatment response.

1. Introduction

The dopamine hypothesis of schizophrenia evolved from observations that amphetamine can induce psychotic symptoms and their improvement by dopamine antagonists. Estimating the basal dopamine level could provide more direct evidence for the hypothesis, which has evolved to include both increased and decreased dopamine transmission dependent on the site of the brain, as reviewed in this text. Much of previous work has concentrated on finding a difference in dopamine transmission between acutely stressed patients with schizophrenia and non-stressed controls, which complicates interpretation. Correlative associations of estimates of dopamine transmission with psychotic symptoms could provide further evidence not necessarily revealed in mean differences in patients and controls. In the following text, such correlative analyses of the prolactin responses to submaximal TRH stimulation are reviewed with an emphasis of minimizing confounders to aid future researchers. Other tests estimating basal dopamine transmission are reviewed as well. Currently, only a limited set of schizophrenia symptoms have been directly associated with basal dopamine transmission.

2. The Mini-TRH Test, Studies in Animals and Normal Men

Tuberoinfundibular dopamine (TIDA) is the main prolactin-inhibiting factor in the rat and humans. The medial basal hypothalamus secretes TIDA to stimulate dopamine-2 receptors on the pituitary lactotrophs via the pituitary stalk circulation, and in the rat pituitary plasma, dopamine contributed to an estimated 40–70% of the total prolactin-inhibiting activity [1]. In vitro, a small dose of thyrotropin-releasing hormone (TRH) almost completely antagonised dopamine inhibition of prolactin release, while a larger dose only partially antagonised it [2], suggesting that prolactin responses to smaller, more physiologic doses could be more specific than larger doses in reflecting dopamine transmission. Indeed, in stalk-sectioned monkeys, an increased prolactin response to submaximal stimulation by TRH suggested a role for decreased dopamine in the response [1]. The submaximal response at 15 min may estimate, predominantly or entirely, hypophyseal effects of TRH on prolactin release because TRH passes the blood–brain barrier poorly, and the prolactin short feedback is much too slow to increase TIDA release at 15 min [1].

Our hypothesis was that submaximal prolactin responses to TRH could uncover continuing activity of dopamine-mediated inhibition of prolactin release, i.e., tonic dopamine inhibition of prolactin release in humans. We used the basal prolactin level and the increase in plasma prolactin 15 min after the administration of an i.v. bolus of 12.5 mcg (the mini-TRH test) and the maximal prolactin response to TRH to estimate endogenous dopamine transmission in a series of correlative studies in men. Reproducible doses could be administered using a small insulin needle. We compared correlations of the above tests to other dopamine tests: prolactin response to 0.5 mg i.m. haloperidol, which may reflect dopamine receptor sensitivity [3], 24 h urinary excretions of 17-ketogenic steroids (glucocorticoids), and homovanillyl acid (HVA). Blood samples for the combined TRH test assessing both submaximal and maximal prolactin responses to TRH and a separate haloperidol test [4] were drawn on separate weekends after one hour of rest in a quiet laboratory with no interruptions between 11 a.m. and 1 p.m. to reduce stress and the circadian variability of prolactin release [4]. None of the subjects had an endocrinological disease or were on continuous medication [5]. The subjects had not taken any medication for at least four weeks before the study [5]. The laboratory was familiar to most of the subjects, and only one to three subjects were simultaneously present [6]. They were asked to lie down and relax, and conversation was discouraged [6]. In humans, diet had no significant effects on urinary HVA unless it included large amounts of monoamines [5]. The subjects collected their 24 h urine specimens consuming their habitual diets under relaxed conditions during weekends [5,6]. Because the blood and the urine specimens were (with one exception) obtained on separate days [5], it is not likely that stress reactions associated with the prolactin tests could provoke significant correlations with the urinary test results. The number of days between the 24 h urine collection of HVA and the mini-TRH test was 57.4 ± 11.8 (mean ± SEM), and the number of days between the urine collection and the haloperidol test was 142 ± 23.8 (mean ± SEM) [5].

Because the United States Food and Drug Administration did not approve the use of i.v. haloperidol, we used prolactin responses to i.m. haloperidol, the relatively specific dopamine antagonist. Studies in normal men using 1 to 1.5 mg i.m. haloperidol produced maximal prolactin responses and, also, consistently provoked various degrees of sedation and/or restlessness [1]. On the other hand, 0.5 mg i.m. haloperidol produced a prolactin increase of at least 50% (mean increase 60%) of the maximal prolactin increase in 90 min with no noticeable side effects [1]. Prolactin-induced increase in TIDA turnover (the short feedback) started to increase TIDA release after 1 to 3 h, reaching significant levels thereafter [1]. At the 90-minute time point, prolactin feedback loops affecting TIDA release were only starting to become operative.

The Mini-TRH Test in Normal Men

The subjects were 28 male volunteers aged less than 50 years. We did not study normal females. After being provided information about the study, two of the subjects refused to participate in the haloperidol test, and in one, the prolactin response to 200 mcg TRH was not obtained. In 25 normal men, the mini-TRH test, but not the basal prolactin level or the prolactin increase to 200 mcg TRH, significantly correlated with prolactin response to i.m. haloperidol (r = +0.57, p = 0.003) [4] and with 24 h urinary excretion of glucocorticoids (r = −0.64, p < 0.001) [5]. The finding of no correlation between the basal prolactin level and plasma prolactin response to 0.5 mg i.m. haloperidol replicated findings from other studies [4]. Consistent with our results, pre-incubation with glucocorticoids decreased prolactin response to submaximal stimulation by TRH in the rat hypophysis [1]. In 28 normal men, the mini-TRH test, but not the basal prolactin level, negatively correlated with 24 h urinary excretion of HVA (r = −0.48, p = 0.010), even after controlling for activity of the sympathetic nervous system as assessed by 24 h urinary excretion of vanillylmandelic acid (r = −0.44, p = 0.019) [6]. In 27 men, there was no correlation between the maximal prolactin response to TRH and urinary HVA [7]. Our results suggested that factors other than dopamine inhibition of prolactin release determined the ceiling of TRH-induced prolactin release. Dopamine inhibition of prolactin release involves an integrated effect of TIDA levels and dopamine-2 receptor sensitivity on lactotrophs. Our results raised the possibility that a generalised dopamine turnover extended to TIDA. In 26 normal men with both the mini-TRH test and the haloperidol test performed in our laboratory, removing the linear effects of 24 h urinary corticosteroids reduced the correlation between the mini-TRH test and 24 h urinary HVA from r = −0.53, p = 0.0054 to r = –0.16, p = 0.43 and between the haloperidol test and 24 h urinary HVA from r = −0.25, p = 0.22 to r = +0.04 [3]. The widespread origin of urinary HVA raised the possibility that glucocorticoids may integrate TIDA release in other structures sensitive to these steroids [3]. In 25 of the normal men, the maximal prolactin response to TRH significantly correlated with the basal prolactin level in the TRH test (r = +0.52, p < 0.01, but not with the basal prolactin level drawn another day in the haloperidol test (r = +0.04) [5], suggesting that the maximal prolactin response to TRH may be associated with fluctuations in the basal prolactin level. These fluctuations may decrease the accuracy of the basal prolactin level and the maximal prolactin response to TRH when estimating the tonic inhibition of dopamine on prolactin secretion. Prolactin secretory bursts occur every 42 to 65 min, independent of prolactin levels and sex, and the bursts contribute at least as much to the prolactin level as the time-invariant mode of prolactin secretion [1]. A single prolactin level may be of limited value as a trait variable. On the other hand, the time-invariant mode could reflect the tonic dopamine inhibition of prolactin release.

3. Controlling Stress-Induced Prolactin Release

In clinical practice, psychic stress at venipuncture and the inherent pulsatory nature of prolactin release may trigger elevations in blood prolactin levels, leading to a false diagnosis of hyperprolactinemia. The Endocrine Society recommends repeat prolactin testing on a different day at 15 to 20 min intervals to rule out pulsatory prolactin levels when in doubt [8]. Even one normal prolactin level in the reference range should rule out hyperprolactinemia [8]. Another way to control stress-induced prolactin pulses is intravenous cannulation followed by a rest period of 30 to 90 min [8]. A third way is to use local anaesthesia followed by a rest period, which was used in studies with the mini-TRH test in psychotic patients. We took great care to avoid not only stress provoked by venipuncture, but also by apprehension [8]. We used local anaesthetic cream on a vein in the morning of the mini-TRH test and allowed no interruptions during the rest periods before the mini-TRH test. The same investigator who had introduced himself the day before applied the anaesthesia on the day of the test and allowed no interruptions during the rest period [8]. In the first study in psychotic patients, the rest period before the mini-TRH test was one and a half to two hours after local anaesthesia [9]. I found only one study to examine physical stress-induced increases in prolactin levels in schizophrenia [10]. This study reported significantly higher prolactin increases in men with schizophrenia than in control subjects after immersion of a hand in ice water for 2 min [10]. The increase was the same duration as in the control subjects, and prolactin levels returned from stress-induced levels to pre-test levels in 30 min. Because of this study and to avoid an unnecessarily long rest period in a test with a potential use in clinical practice, we used a 30 min rest period in our second study [11].

4. The Mini-TRH Test in Psychotic Subjects

We conducted the mini-TRH test from 10:00 to 11:30 a.m. in all our psychotic patients to avoid the nocturnal rise in pulsatory prolactin secretion, which is stabilised in about 2 h after awakening [1]. In every psychotic patient, the basal prolactin level was in the reference range and, in male patients, not higher than in normal men [1]. In exacerbation of schizophrenia, a controlled study could not substantiate a difference between patients and controls in the mean prolactin response to dopamine agonists [1]. The finding of about two times higher mean mini-TRH test in 16 psychotic men than in 28 normal men could thus be associated with a decreased mean TIDA release rather than a different dopamine-2 receptor sensitivity in patients [11]. Because the method used to measure prolactin in normal men was different from the methods used in patients and because only patients received local anaesthesia, a precise comparison was not possible. These findings do not exclude the possibility that dopamine receptor sensitivity on lactotrophs could be associated with some symptoms of psychosis. Because the distribution of the mini-TRH and the basal prolactin level was skewed, we employed logarithmic transformation before calculating Pearson’s correlations in psychotic patients.

4.1. The Mini-TRH Test in Acute Non-Affective Psychosis

In our study of 20 patients admitted to the University Hospital of Helsinki with acute. non-affective psychosis, 9 were diagnosed using DSM-IIIR criteria in the schizophrenia spectrum [9]. Using many sources, we obtained a reliable drug history of all the subjects. The drug-free group included 13 patients, 9 of whom were neuroleptic-naïve, 3 had not taken neuroleptic drugs for at least 4 months before the study, and 1 had not taken neuroleptic drugs for 38 days. After admission to the hospital, a washout period of at least four days was executed in all the patients before the mini-TRH test and rating of the Comprehensive Psychopathological Rating Scale (CPRS) psychosis subscale. Removing the effects of age, sex, and recent drug use revealed a significant correlation between the mini-TRH test and the score for the psychosis subscale (rp = +0.71, p < 0.01) [9], more precisely p = 0.001 [1]. The partial correlation remained the same even when patients in the drug withdrawal group were excluded (r = +0.69) [9]. These results showed an association of decreasing dopamine transmission in regulating prolactin with increasing CPRS psychosis subscale scores. Of the item-specific correlations, the strongest was between the mini-TRH test and the score for other delusions (rp = +0.73, rp < 0.001). This item excludes delusions of control, pessimistic thoughts, hypochondriasis, ideas of persecution, ideas of grandeur, delusional mood, and morbid jealousy [9]. It thus rates non-paranoid, non-affective delusions [9]. The mini-TRH test also marginally correlated with the score for the item for disrupted thoughts, which rated first-rank symptoms (rp = +0.51, p < 0.05) but not with the score for ideas of persecution (rp = +0.05, p > 0.05) or with the total score for various hallucinations (rp = +0.36, p > 0.05) [9]. Removing the effects of age, sex, and recent drug use did not reveal a significant partial correlation between the basal prolactin level and the rating for the psychosis subscale (rp = −0.20, p > 0.05) [9]. In the same 20 patients with non-affective psychoses, 40% of the patients demonstrated memory dysfunction on the CPRS item for failing memory, which is a simple test to estimate everyday memory dysfunction [7]. After controlling effects for age, sex, and recent drug use, the mini-TRH test significantly correlated with the CPRS item for failing memory (rp = +0.67, p = 0.003) [7]. Increasing memory function with advancing age could not explain this finding because, after controlling for sex and recent drug use only, the correlation remained about the same (rp = +0.60, p = 0.008). These findings suggested that the mini-TRH test could uncover age-independent memory dysfunction in association with decreased pituitary dopamine transmission. There was no correlation between the basal plasma prolactin level and ratings for failing memory (rp = +0.07, p = 0.78) [7]. The large fluctuation in the basal prolactin can reduce the accuracy of associations between basal prolactin blood levels and symptoms. Consistent with our results, in 91 patients with drug-naïve schizophrenia, the basal serum prolactin level was not associated with the severity of cognitive symptoms, including memory dysfunction [12].

4.2. The Mini-TRH Test in Acute First-Episode Schizophrenia

Over a four-year period, acute psychotic patients with a presentation of first-episode schizophrenia and no previously known treatment were admitted to the Moisio Hospital, which provides all inpatient psychiatric care for the catchment area of approximately 215,000 inhabitants [11]. We reviewed medical records potentially relevant for diagnostic purposes and possible earlier administration of psychotropic medication. We excluded patients who had previously taken any antipsychotic, antidepressant, or mood-stabilising drugs, as well as patients who, based on their history or examination, appeared to have any clinically significant medical condition or who were suspected to have abused alcohol or drugs. Twenty of the subjects met DSM-IV clinical criteria for schizophrenia, one withdrew her participation in the study. The mean age of the remaining 19 participants was 27.2 years (SD ± 7.7 years), of whom 14 were female and 5 were male [11]. The literature on symptom clusters of the defect state and symptomatic worsening after methylphenidate administration had suggested that the items for poverty of content of speech and inattention in the Scale for the Assessment of Negative Symptoms (SANS) could be particularly related to dopamine activity, more closely than negative or positive symptom clusters [11]. The difference in the mini-TRH test in 14 women and 5 men was not significant, and there was no significant correlation between the mini-TRH test and patient age (r = −0.22). In 19 patients, the mini-TRH test correlated negatively with the SANS scores for poverty of content of speech (r = −0.55, p = 0.01) and objective scores for inattention (r = −0.52, p = 0.02) [11]. The scores for these items significantly correlated with each other (r = +0.74, p < 0.0001). There were no significant correlations between the mini-TRH test and the rest of the SANS symptoms, i.e., negative symptoms not associated with disorganisation. In the patients with first-episode schizophrenia, there was a significant correlation between the mini-TRH test and the basal plasma prolactin level (r = +0.61, p = 0.006), not discovered in normal men [5] or in patients with non-affective psychosis [9]. This correlation raised the possibility that the mini-TRH test might have resulted in maximal prolactin responses in some patients with first-episode schizophrenia because in normal men, the basal prolactin level correlated with the maximal prolactin response to TRH [5]. This could be a factor limiting the specificity of the mini-TRH test. To reduce the effects of maximal prolactin responses, a dose of 6.25 mcg in the mini-TRH test was recommended in first-episode patients [1].

More studies were needed to study the correlations of the mini-TRH test with CPRS hallucinations (rp = +0.36, p > 0.05) and with disorganisation symptoms other than those rated by the SANS. We administered the Positive and Negative Symptom Scale (PANSS) in the last 13 of the 19 patients recruited with first-episode schizophrenia during the 4-year period after the training for the PANSS was completed by the rater [3]. We first correlated ratings of positive and disorganisation symptoms of schizophrenia based on the consensus five-factor model introduced by Wallwork and Fortgang of the PANSS with ratings of the mini-TRH test [3,13]. The requirement that the strongest loading in at least 24 of 29 factor analytic studies ensured that only the most specific symptoms entered the model [13]. In a total of 33 patients (20 with non-affective psychosis and 13 with first-episode schizophrenia), we calculated the difference in correlations in 2 independent samples. Among the specific PANSS positive symptoms, only the correlation between the mini-TRH test and the rating for hallucinatory behaviour (p < 0.0125), but not delusions, unusual thought content, or grandiosity (each p > 0.125), differed significantly from the correlation between the mini-TRH test and the rating for the CPRS psychosis subscale [3]. Although hallucinations and hallucinatory behaviour are not synonymous, the results may not support a predominant role of hallucinations in reflecting TIDA transmission. The corresponding correlative analysis with specific PANSS disorganisation symptoms showed that correlations of the mini-TRH test with difficulty in abstract thinking and poor attention differed significantly from the correlation between the mini-TRH test and the rating for the CPRS psychosis subscale (r < 0.005 each) [13]. There was no significant difference in the corresponding association with conceptual disorganisation (p > 0.05), which is the only PANSS disorganisation symptom that rates positive formal thought disorder [1].

5. Other Tests of Dopamine Transmission in Psychotic Subjects

Amphetamine-induced increase in intrasynaptic dopamine competes with radioligand occupancy to dopamine receptors, and the relative decrease from pre-amphetamine scan in radioligand binding can be quantified. In nonhuman primates, in the frontal cortex, there was a linear association between amphetamine-induced decrease in the dopamine-2/3 antagonist radiotracer [11C]FLB 457 and increase in extracellular dopamine concentration [14], and correlations were found of amphetamine-induced striatal intrasynaptic dopamine release with corresponding decreases in binding of dopamine-2/3 antagonist radiotracers [123I]IBZM and [11C]raclopride [15,16]. In healthy adults, cortisol levels were positively associated with ventral striatal amphetamine-induced decrease in [11C]raclopride binding [17]. As opposed to physiologic dopamine release, amphetamine releases dopamine to a major degree from both cytosolic and vesicular stores [18]. This may limit the accuracy of the amphetamine challenge tests, i.e., amphetamine release capacity for a given dose of amphetamine, in reflecting physiologic processes. Two important studies investigated amphetamine-induced dopamine release in different extrastriatal regions in schizophrenia. In 20 patients, the release was defective in many cortical and extrastriatal regions, including the midbrain, and did not significantly correlate with the basal PANSS positive scale score [19]. The mean PANSS positive score was 15.1 and comparable to the score in patients stabilised with drugs [1]. This could be a limiting factor because in some extrastriatal structures, the defective amphetamine-induced dopamine release may be more apparent in the acute phase of psychosis, as suggested by a strong inverse correlation between the amphetamine tests in the associative striatum and the prefrontal cortex (r = 0.71, p = 0.005) [20]. Dopamine release in these areas could translate into different symptoms. In a mixed population of 34 patients with schizophrenia, including patients both in the active phase and in remission, a significant correlation was found between striatal amphetamine-induced displacement in the radiotracer [123I]IBZM and the change in PANSS positive scale score (r = 0.54, p < 0.001) [21]. This result was not verified in smaller studies, with less variability in the clinical status utilizing [11C]NPA [22] or [11C]-(+)-PHNO [23], in patients with schizophrenia and first-episode psychosis, respectively. Two major studies showed an abnormally enhanced striatal amphetamine-induced dopamine release [21,23] in acute patients with PANSS positive scale scores of 18.9 ± 6.5 and 21.4 ± 6.8, respectively, but not in another study with a PANSS positive scale score of 19.5 ± 5.3 [22] or in patients in remission with a PANSS positive score of 16.1 ± 5.8 [21]. On the other hand, none of the studies reported a correlation between amphetamine-induced dopamine release and pre-amphetamine PANSS positive scale, with the largest study reporting the coefficient of correlation (r = 0.01, p = 0.98) [21]. This was the same study that reported the significant correlation between amphetamine-induced striatal dopamine release and the increase in positive symptoms [21]. The magnitude of the change in PANSS positive symptoms was not associated with the baseline severity of these PANSS symptoms (r = 0.10, p = 0.54) [21]. There was an antipsychotic-free washout period of at least 2 weeks in all of the above studies before the amphetamine challenge.

Neuromelanin gradually accumulates in the form of neuromelanin–iron complexes, which are paramagnetic and can be estimated by magnetic resonance imaging across midbrain dopaminergic nuclei, reflecting long-lasting dopamine turnover [24]. In thirteen patients with schizophrenia, the scores for the Scale for the Assessment of Positive Symptoms (SAPS) showed a negative correlation with ventral tegmental area (VTA) signalling (r = −0.69, p = 0.012) [25], consistent with involvement of the VTA-prefrontal cortical tracts. The signal intensity at VTA was significantly lower and did not increase at substantia nigra in patients with schizophrenia compared to normal subjects [25]. The patients were on a relatively long-lasting antipsychotic drug, which, according to the authors, might have affected the results. In nine unmedicated patients with schizophrenia and nine healthy subjects, after controlling for diagnosis and age, a positive correlation was reported between neuromelanin signalling in the substantia nigra and amphetamine-induced striatal dopamine release [26], indexing storage and release capacity [27]. Nine patients with a PANSS positive scale scores of more than 19 showed higher neuromelanin signalling than twenty-four patients with scores of 19 or less, suggesting that patients disposed to react with positive symptoms might show more neuromelanin signalling in the substantia nigra [26]. In 42 unmedicated patients with schizophrenic psychoses, after controlling for age, PANSS negative and general psychopathology scores, neuromelanin contrast in the substantia nigra correlated weakly with the PANSS positive scale score (rp = +0.35, p = 0.03) driven by delusions and hallucinatory behaviour [28]. In a model controlling only for age, the effect of PANSS delusions on neuromelanin signalling was robust (rhop = +0.42, p = 0.006), but the effect of hallucinatory behaviour was not (rhop = +0.07, p = 0.66).

Both extrastriatal and striatal structures contribute to cerebrospinal fluid (CSF) HVA levels [29]. There may be a major contribution of prefrontal rather than striatal HVA to CSF HVA in psychotic patients [29]. This could explain the low CSF HVA reported in patients in association with more positive symptoms [30]. In chronic treatment-refractory schizophrenia, only non-paranoid delusions were significantly associated with reduced 24 h urinary excretion of dopamine [31]. In acute schizophrenia, first-rank symptoms were associated with reduced CSF HVA [32]. In patients with chronic schizophrenia with a high Brief Psychiatric Rating Scale score of 46.3 ± 9.5, low CSF HVA correlated with poor ability to recall visuospatial information and poor results in the Wisconsin Card Sorting Test (WCST) [33]. In all of the above three studies employing 24 h urinary or CSF HVA, the antipsychotic-free washout period lasted at least 2 weeks. I have not found in the literature increased 24 h urinary or CSF HVA levels in association with the above symptoms in patients with schizophrenia.

6. Factor Analytic Studies of the PANSS

In a meta-analysis of 5-factor studies of the PANSS in schizophrenia, excluding one loading close to 0.4, which was likely to be within the range of noise in the data, the positive dimension consisted of delusions, unusual thought content, suspiciousness/persecution, hallucinatory behavior, and grandiosity [34]. This does not allow the conclusion that all these symptoms could be primarily associated with overall dopamine transmission in the brain. The same investigators who introduced the PANSS recognised the incompleteness of the positive–negative model and published a factor-analytic study of the PANSS, which suggested a pyramidal triangular model with seven components [35]. A syndrome at the vertex in the positive corner was composed of the particular positive symptoms of unusual thought content, delusions, and grandiosity [35], the same PANSS symptoms that could, potentially, be associated with the mini-TRH test [3]. Two of the seven symptoms in the original PANSS positive scale were assigned to the PANSS positive vertex component, and unusual thought content was assigned to it from the general psychopathology scale. Suspiciousness/persecution formed a separate component, and the loading of hallucinative behaviour on the positive component was only 0.43, about the same as the loading of 0.39 on the depressive component. In first-episode psychosis, hallucinations loaded only on a positive dimension, and hallucinatory behaviour loaded on both a positive symptom dimension and a symptom dimension with a mixture of negative and disorganisation items [36].

7. Discussion

Comparisons of dopamine transmission in acutely psychotic patients with non-stressed control subjects may be subject to stress-induced bias in results. In psychotic patients, the assessment of unstimulated stress is not easy. We took great care to reduce the effect of stress in psychotic patients on the mini-TRH test, including the use of local anaesthesia to control stress induced by blood extraction. In first-episode psychosis, a disruption was reported of daily cortisol levels with severity of psychotic symptoms, number of stressful life events, and perceived stress [37]. In schizophrenia, a disruption was observed in the association between stress-induced increase in salivary cortisol and increase in frontal dopamine release [38]. Even with careful routines to reduce stress, the glucocorticoid level may have a limited value to assess stress-induced dopamine transmission in schizophrenia.

Given the strong negative correlation between amphetamine-induced dopamine release in the prefrontal cortex and the associative striatum [20], the interpretation of either of the above is not straightforward. Strong correlations of symptoms with dopamine release at one of the sites might be reflected in weak or even moderate correlations in the opposite direction of these symptoms at the other site, even if dopamine release could translate into different symptoms in the prefrontal cortex and the striatum.

In rats, spontaneous dopamine transients may constitute a major component of extracellular dopamine levels in the brain [39]. In primates, a greater increase and a shorter half-life of striatal than cortical response to amphetamine resulted in more fluctuations in amphetamine-induced striatal than cortical extracellular dopamine levels [40]. In psychotic patients, a similar difference might indicate more fluctuations in striatal than cortical dopamine release and contribute to the lack of correlation between amphetamine-induced striatal dopamine release and pre-amphetamine PANSS positive scale. Even the correlation between the more stable neuromelanin signalling and the basal PANSS positive scale was weak. Inclusion of a broad range of positive symptoms may limit the specificity of the PANSS positive scale to assess basal dopamine transmission, although this scale can track drug-induced changes in dopaminergic symptoms.

It is unlikely that TIDA could mediate symptoms of schizophrenia. A more plausible explanation for the findings with the mini-TRH test is a synchronised action of dopamine inhibition of prolactin release and the activity of cortical structures involved in symptoms of schizophrenia. The thalamus can integrate communication between brain networks [1]. Reduced thalamic connections with frontal regions were associated with PANSS delusions, and increased thalamic connections with parietal regions were associated with a negative disorganisation symptom, PANSS difficulty in abstract thinking [1]. These associations might have been reflected in decreasing TIDA transmission in association with CPRS non-paranoid and non-affective delusions and increasing TIDA transmission in association with increasing negative disorganisation symptoms.

In schizophrenia-related disorders and in first-episode psychosis, studies using the mini-TRH test, extrastriatal amphetamine-induced dopamine release, the MRI neuromelanin signal in the VTA, and urinary or CSF HVA suggested that dopamine transmission might be reduced at extrastriatal sites, including in the VTA. In schizophrenia-related disorders, studies using striatal amphetamine-induced dopamine release and the MRI neuromelanin signal in the substantia nigra suggested almost invariably that dopamine transmission may be increased in the striatum and the substantia nigra. A different radiotracer or the small population of patients studied might contribute to the divergent result [22].

In a review of 36 studies observing acute effects of amphetamine or methylphenidate in schizophrenia, 43% of the patients with active psychotic symptoms worsened as opposed to 35% with nonactive symptoms, and 36% showed no significant change. A total of 21% of patients with active psychotic symptoms showed improvement as opposed to only 6% without active symptoms, suggesting that the improvement was greater in patients with positive symptoms [41]. A possible explanation for these findings is that the net effect of psychostimulant drugs could depend on the worsening of positive symptoms mediated by the dorsal striatum and improving them by effects on extrastriatal structures, such as the prefrontal cortex. The improvement in active symptoms in 21% of patients is considerable, considering the diminished responsivity to dopamine agonists in mesocortical neurons compared to nigrostriatal neurons [29], which could lead to classifying some of the change in mesocortical neurons in the group with no significant change. In the largest study using the amphetamine challenge, there was no correlation between dopamine release in the striatum and the basal PANSS positive scale score [21], suggesting that the latter may not be mainly determined by dopamine transmission in the striatum. The finding of no correlation between the magnitude of change in the PANSS positive scale provoked by amphetamine and the baseline severity of the PANSS positive scale [21] could indicate that amphetamine-induced changes in positive symptoms might differ from basal positive symptoms.

In nonhuman primates, chronic blockade of dopamine-2 receptors downregulates dopamine-1 receptors in the prefrontal cortex and produces severe impairments in working memory [42]. In drug-treated patients with schizophrenia, this could weaken recovery from impaired working memory. In patients with schizophrenia on antipsychotic medication during probabilistic decision-making under uncertainty, reduced activation was reported in the prefrontal cortex and other associated networks [43]. Because of the considerable similarity between the extrastriatal networks associated with probabilistic decision-making and working memory [43], it is possible that drug-induced reduction in extrastriatal dopamine activity could weaken the recovery from impaired probabilistic assessment as well.

8. Conclusions

In conclusion, correlative analyses of symptoms may provide information on symptom-specific alterations in dopamine transmission in non-affective psychotic symptoms. Among symptoms of schizophrenia, decreasing effects of endogenous dopamine on prolactin release as estimated by the mini-TRH test may reflect increasing memory problems, non-paranoid delusions, and first-rank symptoms. On the other hand, increasing endogenous striatal dopamine transmission may reflect increasing delusions and hallucinations. Amphetamine administration is known to increase delusions, paranoid ones in particular [8]. The mini-TRH test and MRI neuromelanin signalling in the substantia nigra are affordable and readily available tests to study basal dopamine transmission, not associated with rapid fluctuations in dopamine release. Using them together would provide a broader view of central dopamine transmission.

9. Future Directions

Most of the studies in this review suggested that derangements in dopamine transmission may be most pronounced in patients in the acute phases of psychosis or in patients prone to experiencing them frequently. A study investigating amphetamine-induced dopamine release at major extrastriatal sites and their associations with psychotic symptoms in the acute phase of schizophrenia is needed. This could help localise sites that are reflected in the mini-TRH test. I have not found a study in the literature comparing dopamine transmission in acutely psychotic patients to dopamine transmission in control subjects experiencing stress of a comparable degree to that found in acutely psychotic patients.

The correlation of the mini-TRH test with 24 h urinary excretion of HVA might associate the mini-TRH test with prefrontal dopamine transmission. In patients with prominent non-paranoid delusions or everyday memory problems, strong correlations of the mini-TRH test with these symptoms could be nonspecifically reflected in weak or even moderate correlations in the opposite direction in the striatum, and great care should be exercised in the interpretation of such results.

It remains to be established whether the mini-TRH test could reflect working memory dysfunction and/or impaired probabilistic assessment in antidopaminergic drug-free or drug-treated psychotic patients. In drug-free patients with schizophrenia-related diseases, it has not been studied whether the mini-TRH test could predict nonresponse to antidopaminergic drugs. Even in the framework of lower dopamine transmission in therapy-resistant patients compared to other patients, there could be a correlation between the mini-TRH test and non-paranoid delusions and/or unusual thoughts represented as first-rank symptoms.

Studying basal dopamine release using the SAPS with its items that rate bizarre delusions or the PANSS positive vertex component with its item for unusual thought content instead of the PANSS positive scale could prove beneficial in schizophrenia, the DSM-V diagnosis of which requires either unwanted obsessive or strange thoughts.