Introduction
Tuberoinfundibular dopamine (TIDA) is the major prolactin-inhibiting factor in both
the rat and in man; it is secreted in the medial basal hypothalamus via the
pituitary stalk circulation to stimulate dopamine-2-receptors on pituitary
lactotrophs [1]. It was estimated that
40–70% of the total prolactin-inhibiting activity in the pituitary
stalk plasma in the rat could be directly attributed to dopamine [2]. Prolactin short feedback, i. e.,
prolactin-induced increase in TIDA turnover, started to increase TIDA release after
1 to 3 h to reach significant levels thereafter [3]. In normal men, peak prolactin levels were
reported to occur within 15 min after the i. v. administration of
thyrotropin-releasing hormone (TRH) and maximal prolactin levels were produced by 25
to 100 µg i. v. as reviewed earlier [4]. A mutual antagonism was reported between
TRH and dopamine on prolactin secretion in men [5]. In stalk-sectioned monkeys, an increased prolactin response to
submaximal stimulation by TRH was demonstrated [6], suggesting a role of decreased dopamine in the response. TRH, at
doses lower than that needed to stimulate prolactin secretion directly, can almost
completely antagonize the inhibitory effect of dopamine on pituitary prolactin
release in vitro [7]. This finding is the
rationale for determining the prolactin increase 15 min following an
i. v. bolus of 12.5 µg TRH (the mini-TRH test) to estimate
endogenous dopamine transmission avoiding direct effects of non-dopamine prolactin
inhibiting or releasing factors. The mini-TRH test was extended to include prolactin
response to 200 µg TRH (max-TRH test) 15 min after the
mini-TRH test prolactin response [4]. TRH
passes the blood-brain barrier poorly [8], and
this enables the mini-TRH test to estimate predominantly or entirely hypophyseal
effects of TRH on prolactin secretion. In small doses, TRH antagonizes dopamine
signal transduction downstream of the membrane site of dopamine receptor at the
guanine nucleotide-binding protein level [9].
Therefore, an integrated effect of the level of TIDA and dopamine-2 receptor
sensitivity contributes to the mini-TRH test. As far as I know, utilizing a
neuroimaging technique or some other hormonal test fails to assess this kind of
integrated dopamine transmission, although two tests were recently integrated in an
attempt to achieve this [10].
In addition to small doses of TRH, neuroleptics antagonize the effect of dopamine on
prolactin release. Because i. v. haloperidol was not approved by the United
States Food and Drug Administration, we used intramuscular (i.m.) responses to
haloperidol as a safety measure, although the individual release of the drug into
the circulation was expected to increase the variability of the response. For a
near-maximal prolactin response to haloperidol, i.m. doses of 1 or 1.5 mg
were required in normal men [11]. However,
these doses consistently produced various degrees of sedation and/or
restlessness; 0.5 mg of haloperidol i.m. caused no side effects and produced
a prolactin increase of at least 50% (mean 60%) of the maximal
prolactin increase in 90 min [11]. At
that time point, prolactin feedback loops are only starting to become operative. Low
dopamine levels reaching the hypophysis would be expected to result in a relative
preponderance of haloperidol over endogenous dopamine and produce an enhanced
prolactin response to the drug. To avoid the apprehension-stress-induced increases
in prolactin release [12] provoked by side
effects, prolactin response to 0.5 mg i.m. haloperidol at 90 min
(hal-test) was used as a reference estimate of dopamine inhibition of prolactin
release. The combined mini-TRH test/max TRH test was done at least one week
before or after the hal-test. These tests were done in normal men after 1 h
of rest in a quiet laboratory, during weekends, and with no interruptions [13]. We used standard blood extraction between
11 a.m. and 1 p.m. to avoid circadian variability of basal [14] and of TRH-induced prolactin release [15]. The nocturnal rise in pulsatory prolactin
is stabilized in about 2 h after awakening [16].
The activity of brain dopaminergic structures may be reflected in urinary dopamine
metabolites [17]. In normal women,
dexamethasone-induced inhibition of prolactin secretion was not accompanied by
blunting of maximal prolactin response to TRH [18], but submaximal stimulation by TRH was not studied. There is a lack
of studies investigating the association of endogenous glucocorticoids and dopamine
inhibition of prolactin release in normal men. Our cohort of healthy men collected
their 24-h urinary samples of 17-ketogenic steroids (u-17-KGS) and homovanillic acid
(u-HVA) during weekends under relaxed conditions. Urinary excretions were correlated
with prolactin tests. Blood and urine specimens were (with one exception) obtained
on separate days to minimize the potentially blurring effects of stress reactions on
correlations between prolactin tests and urinary excretions.
We were the first to study submaximal prolactin responses to TRH and correlate
mini-TRH test results with key psychotic symptoms, including memory dysfunction in
inpatients hospitalized for acute psychosis [19]
[20]
[21]. Only one study examined physical
distress-induced prolactin increase in schizophrenia and found it to be
significantly greater than, but of the same duration as in control subjects [22]. Having a protein-rich lunch stimulates
prolactin release, which occasionally results in mild, short-lasting
hyperprolactinemia [23]. We took much care to
avoid confounding factors, especially stress. My coworker (P-E B) applied an
anesthetic cream on an antebrachial vein on the morning of the mini-TRH test and
secured a rest period with no interruption, even by lunch. Immediately thereafter,
he performed mini-TRH tests between 10:00 and 11:30. In our study with nonaffective
psychosis, the rest period was 90 to 120 min [19]. As prolactin levels in schizophrenia
returned from acute stress-induced levels to pretest levels in 30 min [22], we shortened the rest period in our study
with drug-naïve first episode schizophrenia to 30 min [20]. In addition to painless blood extraction,
rest periods were used for the general calming down of psychotic patients. Because
the distribution of prolactin response and basal prolactin were skewed, the obtained
data were log-transformed before statistical calculations.
Studies in healthy subjects
The subjects were 28 normal men [4], of
whom two refused to participate in the hal-test, and in one, max-TRH was not
obtained. Any correlations of≤.18 are reported as no correlation. There
was a positive correlation between the mini-TRH test and the hal-test,
rp(23)=+ 0.57, p=0.003 [4], and a negative correlation between the
mini-TRH test and u-HVA, rp(26)=− 0.48,
p =0.010 even after controlling for 24-h urinary
vanillylmandelic acid excretion to decrease the effect of HVA derived from the
sympathetic nervous system, rp(25)=− 0.44,
p=0.022 [13]. On the other hand,
the max-TRH test did not significantly correlate with the hal-test
rp(23)=+ 0.29, p=0.16. [4] and there was no correlation between the
max-TRH test and u-HVA [21]. The results
suggest that a dose of 12.5 µg TRH was small enough to
antagonize TIDA in normal man and that the mini-TRH test is positively
associated with dopamine receptor-mediated inhibition of prolactin secretion as
well as negatively associated with widespread dopamine turnover/release,
which could extend to TIDA as well.
As opposed to the mini-TRH test, basal prolactin did not correlate with the
hal-test [4], confirming previous findings
[11]. Neither did basal prolactin
correlate with u-HVA [13]. The results
suggest that the mini-TRH test may be a better estimate than basal prolactin of
dopamine inhibition of prolactin release. No correlations were found between the
mini-TRH test and basal prolactin obtained either in the TRH test or in the
hal-test [24], further emphasizing the
difference in these tests.
A significant negative correlation was detected of u-17-KGS with the mini-TRH
test, rp(23)=− 0.64, p<0.001, but with the
hal-test it was marginal, rp(23)=− 0.39, p=0.054
[24]. There were no correlations of
u-17KGS with the max-TRH test or with basal prolactin [24]. The absence of any correlation between
endogenous cortisol levels and basal prolactin was also reported in women [25]. Glucocorticoids may amplify various
dopamine-mediated processes [24]. Indeed,
in 26 men with both the mini-TRH test and hal-test performed, controlling for
u-17-KGS reduced the correlation between the mini-TRH test and u-HVA from
− 0.53 to − 0.16 and between the hal-test and
u-HVA from − 0.25 to+0.04 [26]. Preincubation with glucocorticoids
decreased prolactin response to submaximal stimulation by TRH in the rat
hypophysis [27], in line with our results.
The widespread origin of urinary HVA raises the possibility that glucocorticoids
might integrate TIDA release and dopamine release in other dopamine structures
sensitive to them [26].
The max-TRH test significantly correlated with basal prolactin
rp(23)=+ 0.52, p=0.008 [24]; an association also found in
depressive women [28]. However, no
correlation was detected between the max-TRH test with basal prolactin obtained
in the hal-test [24], suggesting that the
max-TRH test may be associated with fluctuations in basal prolactin. These
fluctuations may be induced by non-dopamine factors because the results of the
hal-test and u-HVA did not significantly correlate with the max-TRH test or with
basal prolactin. Absent of correlation was also reported between cerebrospinal
fluid (CSF) HVA and basal prolactin level [29].
Prolactin secretory bursts occur every 42 to 65 min. Their duration is
about 24 min, independent of prolactin levels and sex [30]. The prolactin level is determined at
least as much by secretory pulses as by the time-invariant mode of prolactin
secretion [16]. We determined in 26 normal
men under unstressed conditions; two basal prolactin levels were drawn with an
interval of at least one week. The correlation between the two prolactin
determinations was only+0.21, although the mean prolactin level was
almost exactly the same in the two prolactin tests [4]. In a strictly controlled study in
normal women, a straightforward correlation between the logarithm of two
prolactin levels one year apart was+0.40 in 60 premenopausal
and+0.18 in 47 postmenopausal subjects [31]. In the former 13 samples and in the latter 40 samples, the
reliability of a single prolactin determination was required to be raised to a
target level near 0.90 [31]. These studies
suggest that the widely held view that basal prolactin could be used as a trait
variable to estimate dopamine-mediated tonic inhibition of prolactin release,
even in unstressed subjects, is surprisingly poorly grounded.
Studies in schizophrenia disorders
CSF HVA is an estimate of widespread dopamine release/turnover. In one
meta-analysis, it was found to decrease in schizophrenia [32]. In post-mortem samples of psychiatric
patients and controls, CSF HVA correlated with prefrontal HVA content but not
with striatal HVA [33]. A widespread
defect in amphetamine-induced dopamine release, indicating a reduced capacity
for dopamine release, was discovered in the prefrontal cortex and most
extrastriatal regions in 20 drug-free patients with schizophrenia [34]. The positive symptom mean score was
15.1 on the Positive and Negative Symptom Scale (PANSS) [34], which is very similar to a score of
14.9 reported in 1538 patients with schizophrenia stabilized with drugs [35], suggesting that the defect in
extrastriatal dopamine release was not studied in the most acute phase of the
disease. Even in the acute phase of psychosis, scores on the mini-TRH test were
about two times higher in two populations of acutely psychotic men than in
normal men [20]. In schizophrenic
exacerbation and in control subjects, no difference was found in
dopamine-agonist-induced decrease in serum prolactin [36]. These results suggest that decreased
TIDA release could not be fully compensated by increased sensitivity of dopamine
receptors [26], suggesting a predominant
role for TIDA release in determining the mini-TRH test. In 20 patients with
nonaffective psychoses after controlling for sex, age, and drug use, the
mini-TRH test significantly correlated with the total rating for the
Comprehensive Psychiatric Rating Scale (CPRS) psychosis subscale
rp(15)=+ 0.71, p=0.001 [19]. Because the mini-TRH test reflects
widespread dopamine release, the above result is consistent with a negative
correlation between CSF HVA and key positive symptoms in schizophrenia [37]
[38]. In every psychotic patient studied by us, the basal prolactin
level was below the hyperprolactinemic level [19]
[20], indicating that local
anesthesia (applied by the same investigator who performed the mini-TRH test)
protected from the most stress-induced increase of prolactin. I suggest that
this routine is followed in future studies.
The correlation between the mini-TRH test and the CPRS psychosis subscale score
was mostly driven by the correlation between the test and the score for other
delusions, rp(15)=+ 0.73, p<0.001 [19]. This item excludes paranoid delusions
and many other delusional states not specific to schizophrenia [19]. Negative correlations were reported
between nonparanoid delusions and 24-h urinary excretion of dopamine and its
metabolites in chronic schizophrenia [17].
Because the mini-TRH test reflects urinary excretion of HVA, our results may
extend the association of decreased dopamine release and nonparanoid delusions
to newly admitted patients in the acute phase of psychosis [19]. The mini-TRH test marginally
correlated with non-hallucinatory first-rank symptoms,
rp(15)=+ 0.51, p=0.036 [19]. Although such symptoms are
heterogenous in nature [39], a negative
correlation was reported between the number of first-rank symptoms and CSF HVA
in acute schizophrenia [40], which is
consistent with our results. As far as I know, there are no studies reporting
positive correlations between urinary or CSF HVA and nonparanoid delusions or
first-rank symptoms.
We found no significant correlation between the mini-TRH test and scores for
ideas of persecution rp(15)=+ 0.05 [19] or hallucinations
rp(15)=+ 0.36, p=0.15 [19]. In an independent sample of acutely
psychotic patients with first-episode schizophrenia, there was no correlation
between the test and PANSS hallucinatory behavior,
rp(11)=− 0.12 [26]. Hallucinations and persecutory delusions have been suspected to
represent unspecific [41] and emotional
[42] reactions in psychosis. In the
setting of decreased prefrontal activity, these symptoms have been associated
with increased mesotemporal and ventral striatal activity in response to
threatening or even seemingly neutral stimuli in schizophrenia [43]. The lack of correlations between these
symptoms and the mini-TRH test suggests that the test may not be sensitive in
reflecting the activity of the above structures. In the acute phase of
schizophrenic psychosis, amphetamine-induced striatal dopamine release was
increased, and it was suspected that activation of preexisting dysfunctional
circuits was involved [44]. On the other
hand, reduced activation of several frontal and parietal areas in schizophrenia
was found during probabilistic reasoning [45], which might not only trigger but create new delusions. Future
studies are required to find out if probabilistic reasoning is reflected in the
mini-TRH test. We discovered no significant associations between basal prolactin
and positive symptoms. Mixed results in associating basal prolactin and
individual positive symptoms may be explained by differences in patient
populations and the measures used to assess symptoms [46] but also by heterogeneous variables
that affect prolactin levels in addition to tonic dopamine inhibition [1].
Poor performance on everyday memory was reported in schizophrenia [47]. Among our patients with nonaffective
psychosis [19], 8/20 scored memory
dysfunction on an estimate of everyday memory dysfunction, the CPRS item failing
memory [21]. After controlling for age,
sex, and drug use, the score for failing memory correlated with the mini-TRH
test, rp(15)=+ 0.67, p=0.003 [21], associating increasing memory
dysfunction with decreasing hypothalamic-pituitary dopamine transmission. There
was no correlation between basal prolactin level and score for failing memory
[21].
In 19 patients with acutely psychotic drug-naïve first-episode
schizophrenia, negative correlations were detected between the mini-TRH test and
the disorganization symptoms in the Scale for the Assessment of Negative
Symptoms (SANS) poverty of content of speech
rp(17)=− 0.55, p=0.014, and objective
inattention rp(17)=− 0.52, p=0.022 [20]. There were no significant correlations
between the rest of the SANS symptoms (actual negative symptoms) and the
mini-TRH test, suggesting that they are not reflected in TIDA during the acute
stage of psychosis. No significant correlations were detected between basal
prolactin and any SANS symptoms [20].
Differences in correlations of the mini-TRH test with CPRS subscale score and
the test with PANSS disorganization symptoms in 13 available patients with
first-episode schizophrenia utilizing the Wallwork model [48] showed significant separation of CPRS
psychotic subscale from the difficulty of abstract thinking, z=2.89,
p=0.004, and poor attention, z=2.97, p=0.003, but no
significant separation from conceptual disorganization, z=1.68,
p=0.093 [49]. The latter is the
only PANSS disorganization symptom that rates positive formal thought disorder
[50].
In patients with first-episode schizophrenia, the mini-TRH test significantly
correlated with basal prolactin level, rp(17)=+ 0.61,
p=0.006 [20]. No such correlation
was present in patients with nonaffective psychosis [19]. In normal subjects, the max-TRH test
significantly correlated with basal prolactin. These findings suggest a shift of
the TRH-induced prolactin dose-response curve to the left in patients with
first-episode schizophrenia. The same kind of shift was observed in pituitary
stalk-sectioned monkeys with a smaller dose of TRH than in normal monkeys needed
to induce a prolactin response [6]. To
reduce the effects of maximal prolactin responses, a 6.25 μg
dose could be used in the TRH test, especially in first-episode patients. This
is allowed by the increased sensitivity of the present prolactin tests.
It is unlikely that TIDA is responsible for symptoms of schizophrenia. A more
plausible explanation for our findings is a synchronized action of
dopamine-mediated inhibition of prolactin release and the activity of cortical
dopamine structures involved in psychotic symptoms. The thalamus is able to
mediate communication between various cortical areas [51]. In schizophrenia, reduced thalamic
connections with frontal regions were associated with PANSS delusions [52], and increased connections with
parietal regions in association with PANSS difficulty in abstract thinking [52] were found.
Cerebellum-thalamic-cortical hyperconnectivity was associated with
state-independent disorganized thought and speech [53]. In line with the above findings, the
mini-TRH test revealed decreased dopamine transmission in association with PANSS
delusions and increased dopamine transmission with SANS and PANSS negative
disorganization symptoms. As far as I know, the association between nonparanoid
delusions and functional thalamic connections has not been investigated as
yet.
A hypoactive extrastriatal dopaminergic state may not be confined to
schizophrenia-related disorders. In unipolar depression, significantly blunted
amphetamine-induced activation in prefrontal and frontal regions was reported
[54]. A recent meta-analysis revealed
that only CSF HVA levels but not CSF 5-HIAA or MHPG levels were decreased in
depressive disorder [55]. The mini-TRH
test results suggest that a hypodopaminergic state in depression extends to the
hypothalamic-pituitary dopamine system [20]. Antipsychotic drugs increase prefrontal cortical and striatal
dopamine and noradrenaline release [56].
In an animal model depression, the mildly stressed rats, acute repeated
quetiapine at antidepressant doses increased ventral tegmental area dopamine
neuron population activity to normal levels [57]. In future studies, the mini-TRH test could be correlated with
antidepressant responses to antipsychotic medications.
