| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Radiology and Nuclear Medicine, Universitätsklinikum Benjamin Franklin, Free University of Berlin, Berlin, Germany
Address all correspondence and requests for reprints to: A. Baumgartner, M.D., Department of Radiology and Nuclear Medicine, Hindenburgdamm 30, 12200 Berlin, Germany. E-mail: abaum{at}cipmail.ukbf.fu-berlin.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As regards the regulation of the activities of 5'D-II and 5D-III in the CNS, there is good evidence that the activity of 5'D-II is inhibited and that of 5D-III stimulated by different iodothyronine compounds, whereas in hypothyroidism the reverse changes occur (12, 13). It is believed that the purpose of this "autoregulatory mechanism" is to protect the CNS against unphysiological changes in T3 concentrations in the case of hypo or hyperthyroidism.
However, recent evidence suggests that the concentrations of T3 in the CNS may vary substantially after pharmacological interventions and even under physiological conditions. There is, for example, a significant circadian rhythm in 5'D-II deiodinase activity in the rat CNS, with significant variations in tissue levels of T3 (14). We recently reported the surprising finding that even mild stress such as handling a rat causes dramatic increases in 5'D-II activities and T3 concentrations in specific regions of the rat brain (15).
As clinical studies support an effect of antidepressant drugs on thyroid hormone metabolism, in recent years we have investigated the effects of different antidepressant medications on thyroid hormone metabolism and concentrations in the rat CNS. We hypothesized that all of these drugs would enhance the tissue concentrations of T3 in at least one relevant brain region. Stimulation of 5'D-II was indeed found in 8 out of 11 brain regions following treatment with the norepinephrine reuptake inhibitor desipramine (16). However, T3 levels were elevated in only two of these regions (17). Further investigations revealed that many different drugs used in the treatment and prophylaxis of affective disorders do in fact affect deiodinase activities in rat brain. The results of these studies which are reported here showed effects that were specific for each type of treatment in terms of the isoenzyme affected, the direction of the change, the brain region involved and the time of day. No common effects on deiodinase activities of all the drugs investigated were evident.
These data are therefore disappointing from a psychopharmacological point of view, but raise the question as to whether thyroid hormone homeostasis in the CNS is more sensitive and its regulation more complex than generally thought. To learn more about this issue we also investigated the effects of a variety of nonpharmacological (sleep deprivation, 12 h fasting, and 14 days of calorie-reduced diet) and nonantidepressant, pharmacological (ethanol, neuroleptic drugs and one anticonvulsant) treatments on the activities of both 5'D-II and 5D-III and also on the tissue concentrations of T4 and T3. The results are reported together with those of investigations on four antidepressants and one prophylactic drug. The activities of 5'D-I in the CNS and liver and pituitary were determined in several selected subgroups only as the physiological relevance of this isoenzyme in rat CNS is questionable (5) and its activity has not been demonstrated in human brain (6).
The results obtained for Group 17 (treatment with DMI, Table 1
) and the deiodinase activities (but not
the tissue levels of T3 and
T4) of the groups treated subchronically with
fluoxetine, lithium and carbamazepine (Groups 15, 16, and 18) have been
previously reported in a psychopharmacological context (16, 17, 18, 19). All
other results presented in this study have not been previously
published.
|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Carbamazepine, desipramine, and clozapine were gifts from Novartis Pharmaceuticals (Basel, Switzerland). Fluoxetine was donated by Eli Lilly & Co. (Indianapolis). Lithium, tranylcypromine, haloperidol and mianserin were bought from Sigma. Pellets containing carbamazepine and lithium as well as the control pellets were prepared by Altromin (Lage, Germany).
Animal treatments
All of the animal experiments described in this study were
evaluated and approved by the Animal Protection Committee of the Berlin
Senate. Adult male euthyroid Sprague Dawley rats weighing 250 to
300 g were employed throughout. They were housed in individual
cages on a 12-h light, 12-h dark schedule (lights on at 0600 h)
and had access to food and water ad libitum. Unless stated
otherwise 24 rats were used for each experiment. Deiodinase activities
were measured in 6 rats in each group, while thyroid hormone
concentrations were determined in a further 6. The remaining 12 served
as controls (6 for deiodinase activities and 6 for thyroid hormone
concentrations). Experimental groups which received a drug for 14 days
were always decapitated 24 h after the last dose unless otherwise
stated. Drugs were usually administered and the rats decapitated at
about noon, unless otherwise stated. The following groups were
investigated (see Table 1
).
Group 1 (sleep deprivation). Each of the 12 rats was placed in one of 6 drums which were rotated at a speed of one revolution per 45 sec. The rats were placed in the drums at 1000 h and remained there for 24 h. Food and water were available ad libitum throughout the whole procedure. At between 1000 h and 1200 h on the next day they were killed by decapitation without anesthesia together with 12 control rats, which received no treatment at all.
Group 2 (12 h fasting). This group was completely deprived of food at 2000 h and decapitated at between 0800 and 0900 h on the next morning. The 12 control rats received no specific treatment.
Groups 3 and 4 (14 days on a calorie-reduced diet). Twelve rats received a diet adjusted on a daily basis to achieve a weight reduction of approximately 50% within a 2-week period. As control rats of the same age usually undergo a weight increase of approx. 30% within 14 days, the diet was adjusted so as to induce a weight loss of approx. 20% of the initial weight during a 2-week period. On Day 14, Group 3 had lost 16.2 ± 4.5% of their initial body weights and Group 4 15.3 ± 3.9%, while the two control groups had gained 29.5 ± 3.5% and 31.1 ± 4.5%, respectively.
Group 5 (ethanol, acute). Twelve rats received 1 g ethanol/kg body weight and 12 control rats received the same volume of saline by gavage at approximately 1200 h. Twelve rats were decapitated 30 min later, and the other twelve 120 min later.
Groups 6 and 7 (ethanol, 14 days). Twelve rats received a 5% solution of ethanol as sole fluid during a 14-day period. Six of these rats were decapitated at 0800 h and the remaining 6 at 2000 h, each group of 6 together with the corresponding controls, which received pure water ad libitum. The ethanol was not withdrawn before decapitation. In this group only deiodinase activities were determined.
Groups 8 and 9 (neuroleptic drugs). Group 8 was treated for 14 days with 1 mg/kg haloperidol, which was suspended in a 2% Tween 80 solution and administered by gavage. Haloperidol is a "classical" and potent neuroleptic drug used in the treatment of schizophrenia. It is strong antagonist of dopamine 2 receptors and induces extrapyramidal side effects. Group 9 was given 20 mg/kg clozapine ip once daily for 14 days. The clozapine was also suspended in a 2% Tween 80 solution. Clozapine is an atypical and potent neuroleptic drug that does not have extrapyramidal side effects. The control rats of Group 8 received the same volume (7.5 ml) of a 2% Tween 80 solution (dissolved in saline) by gavage. The controls of Group 9 received the vehicle ip.
Groups 10 and 11 (lithium, acute, I). Twelve rats each received a single ip injection of 7.5 mmol/kg lithium chloride at approximately 2000 h and 12 control rats received an injection of saline ip. Twelve rats were decapitated after 12 h, the remaining twelve 24 h later, each group together with 12 control rats.
Group 12 (lithium, acute, II). As the dose of lithium
employed for acute treatment in Groups 10 and 11 (7.5 mmol/kg) yielded
highly toxic serum concentrations of lithium (2.6 ± 0.1 and
1.9 ± 0.1 mmol/liter, see Table 1), a further group was treated
with one ip injection of 3 mmol lithium/kg at 2000 h and killed
12 h later. The control rats received an injection of saline
ip.
Groups 13 and 14 (carbamazepine, acute). Twelve rats received 40 mg/kg carbamazepine ip at 2000 h, 12 control rats received saline ip. Twelve rats were killed after 12 h, the remaining twelve 24 h later.
Group 15 (lithium, subchronic). Twelve rats received a 0.15% lithium diet and the other 12 a 0.3% lithium diet for two weeks. Twelve control rats received the same pellets without lithium.
Group 16 (carbamazepine, subchronic). Twelve rats received a 0.4% carbamazepine diet for 2 weeks. Twelve controls received the pellets without carbamazepine.
Groups 17 to 20 (antidepressants, subchronic). Different
groups of rats were treated with the drugs shown in Table 1 for 14
days. Desipramine is a norepinephrine reuptake inhibitor, fluoxetine a
serotonin reuptake inhibitor, tranylcypromine an MAO inhibitor and
mianserin an atypical antidepressant. All these drugs were dissolved in
NaCl. Desipramine and fluoxetine were administered by gavage,
tranylcypromine, and mianserin ip. The control rats received saline by
the same route as the respective drug-treated group.
Group 21 (age effects). Six rats each weighing approximately 200 g were decapitated together with 6 rats aged 2 yr. In this group only deiodinase activities were determined.
All rats were decapitated without anesthesia. Their brains were dissected according to Glowinski and Iversen (21). The pituitaries and livers were also removed and all tissues stored immediately at -70 C. Blood was drawn from the decapitation wound, centrifuged and the serum stored at -20 C.
Iodothyronine deiodinase assays
The deiodinase activities were measured as previously described
(16). Tissue samples were homogenized individually on ice in 5 to 6 vol
of 0.25 M sucrose, 10 mM HEPES (pH 7.0)
containing 10 mM dithiothreitol (DTT) and immediately
frozen in a dry ice/acetone bath and stored at -80 C until assay. The
measurement of 5'D-I, 5'D-II and 5D-III was based on the release of
radioiodide from the
125I-labeled substrates
(22).
5'D-I and 5'D-II assay. The activity of 5'D-I was determined
by measuring the release of radioiodide from 100,000 cpm (
2.5 kBq)
(5'-125I)-rT3 at 5 nM
rT3, 20 mM DTT, in the presence (for
5'D-II) and absence (5'D-I + 5'D-II) of 5'D-I inhibiting
6-n-porpyl-2-thiouracil (PTU) (4). 5'D-II activity was
determined using (5'-125I)-T4 as substrate in the
presenceof 6 nM T4, 30
mM DTT, 1 mM PTU, and 1
µM T3, to inhibit the
innerring deiodination of T4 in those tissues
containing significant 5D-III activity (4).
The measurement was conducted after 45- to 90-min (usually 60 min)
incubation at 37 C with 50–100 µg of protein from the crude
homogenate in 100 µl of 0.1 M potassium phosphate buffer
(pH 7.0), 1 mM EDTA. The reaction was started by the
addition of the tissue homogenate and stopped by adding 50 µl
ice-cold 5% BSA and 10 mM PTU, followed by 400 µl of
10% ice-cold trichloroacetic acid. After centrifugation at 4000
x g for 30 min, the supernatant containing the
125I- was further purified
by cation exchange chromatography on 1.6-ml Dowex 50 WX 8 columns
(100–200 mesh) (Serva GmbH and Co., Heidelberg, Germany). The iodide
was then eluted with 2 1-ml aliquots of 10% acetic acid and the
radioactivity was counted in a 
-counter.
5D-III assay. For determination of 5D-III (inner-ring
deiodinase) 20–70 µg protein were incubated in a final volume of 100
µl 0.1 M potassium phosphate buffer (pH 7.4), 1
mM EDTA with approximately 1.2 kBq(50,000 cpm)
inner-ring labeled
[5-125I]-T3, at 50
nM T3, 20 nM DTT, and 1
mM PTU for 60 min at 37 C. Radioiodide release was measured
as described above.
Preliminary experiments established that for each tissue (a) the reaction rates were constant over time for up to 120 min in the presence and absence of PTU; (b) the reaction rates were proportional to protein concentrations in the ranges used (50–100 µg/tube in the 5'D-I and 5'D-II assays; 20–70 µg/tube in the 5D-III assay) and (c) after incubation, equal amounts of 3,3'-T2 and I- were produced from rT3 (5'D I+II assay) or from T3 (5D-III assay) in homogenates from each of the different tissues, as determined by reversed-phase HPLC of the incubation extracts. Likewise, it was established that equal amounts of T3 and I- were produced from T4 in the T4 5'D-II assay.
In all assays, control incubations substituted homogenization buffer for tissue homogenates and the amount of 125I- produced in the tissue-free controls (usually 0.3–0.5% of the total radioactivity added) was then subtracted from the sample results.
Because the substrates were randomly labeled with 125I at the equivalent 3' or 5' positions of the phenolic ring (for rT3 and T4) or at the equivalent 3 or 5 positions of the tyrosyl ring (for inner-ring labeled T3), the labeled iodide release was half that of the degraded iodothyronines. This was accounted for in the analysis of the data. The reaction conditions selected were such that <10–15% of the substrate was consumed by enzymatic deiodination. Each experimental point was determined in triplicate with coefficients of variation of less than 5%.
Determination of tissue concentrations of T4 and
T3
The tissue concentrations of T4 and
T3 were determined by RIA as previously described
(17, 23). Briefly, tissue samples were homogenized in 100%
methanol-containing 1 mM PTU and 100 µM
phloretin, extracted into chloroform-methanol and back extracted into
an aqueous phase, which was then purified through Bio-Rad Laboratories, Inc. AG 1 x 2 resin columns (Bio-Rad Laboratories, Inc. Richmond, CA). The iodothyronines were eluted
with 70% acetic acid, evaporated to dryness and taken up in the RIA
buffer. The limits of sensitivity were 3.0 and 2.5 pg for
T4 and T3, respectively.
All the samples of a particular tissue were processed individually and
extracted and assayed together at the same time. Each sample was
determined in triplicate. Intraassay variations were 5.1 and 7.6% for
the T4 and T3 RIAs,
respectively; the corresponding interassay variations were 5.9 and
8.2%. Molar cross-reactivities of the anti-T4
and anti-T3 sera have been reported previously
(20). The results were corrected on the basis of individual recovery
data obtained after addition of maximal specific activity
[125I]T3 and
[131I]T4 to every sample
during the initial extraction process. The quantities of radioactive
T3 and T4 added (1,000
cpm per sample) represented approximately 0.02–0.5% of the amount of
endogenous hormone. Recovery ranged between 80 and 90% for extracted
T3 and between 70 and 80% for
T4. Tissue concentrations are given in nanograms
of T4 or T3 per gram wet
weight.
Determinations of the serum concentrations of T4,
T3, and TSH
The serum levels of T4 and
T3 were determined by a slightly modified
double-antibody RIA as previously described for human serum (20). For
assaying total T4 and T3 in
the rat sera, standards were set up in iodothyronine-free rat serum.
The serum levels of TSH were measured by a specific RIA developed for
the rat, using immunoreactants kindly supplied by the National
Institute of Arthritis, Diabetes & Digestive and Kidney Diseases of the
National Institutes of Health (Bethesda, MD).
Statistical analysis
The data are given as means ± SEM.
P values of less than 0.05 were considered significant.
Individual comparisons were performed using the Mann-Whitney
U test. Altogether, 267 statistical comparisons have been
calculated in Table 1 and
Figs. 1–8. When a significance level of
P < 0.05 was applied, approx. 13 "significant"
results would be expected to occur by chance. Strict application of
Bonferroni’s correction of the P value would result in a
level of significance of P > 0.00025, in which case
the results of only 12 statistical calculations would remain
significant. However, in our study we found 98 significant results.
Thus, if Bonferroni’s correction were strictly applied, it is likely
that several relevant findings would be lost. We therefore discuss
critically whether all results yielding P values of less
than 0.05 seem plausible or may reflect statistical artefacts.
|
|
|
|
|
|
|
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The tissue levels of
T3 were elevated in all brain regions except the
hypothalamus. 5D-III activity and the tissue levels of
T4 were not significantly influenced. 5'D-I
activity was measured in the cortex, hippocampus, and liver and found
to be unchanged (data not shown). The serum concentrations of
T4 and T3 were also
significantly higher than in the controls (Table 1). However, the TSH
levels remained unchanged (2.12 ± 1.02 ng/ml and 2.18 ±
0.99 ng/ml for the controls and experimental rats, respectively).
Group 2
The effects of 12-h food deprivation on thyroid hormone parameters
are presented in Fig. 2. As in Group 1,
significant increases in 5'D-II activity were noted in four brain
regions; however, in contrast to the results for Group 1, the levels of
T4 in the groups having fasted were dramatically
lower than in the controls in three out of four brain regions, whereas
the tissue levels of T3 remained unchanged. The
serum levels of T4 remained unchanged, but those
of T3 were significantly lower in the fasting
group than in the controls (Table 1). The serum levels of TSH were not
significantly affected in this or any of the remaining groups (data not
shown).
Groups 3 and 4
In contrast to the results after 12 h fasting, 14 days on a
calorie-reduced diet induced a pronounced reduction in 5'D-II activity
in the frontal cortex in the rats killed at 0500 h, but not in
those killed at 2000 h (Fig. 2). In addition, the group
decapitated in the early morning also had significantly lower 5D-III
activities and tissue concentrations of T3 and
T4 than the controls. In the group decapitated in
the evening, a nonsignificant trend in the same direction was noted for
the tissue levels of T4 and
T3. No significant effects were noted for the
serum concentrations of T4,
T3, or TSH.
Group 5
Administration of ethanol significantly reduced the activity of
5D-III in the frontal cortex after 120 min and in the amygdala after
both 30 and 120 min (Fig. 3, left). All other tissue and serum parameters were
unaffected. The mean blood concentrations of ethanol were 0.84 ±
0.05
and 0.39 ± 0.03 after 30 and 120 min,
respectively.
Groups 6 and 7
After 14 days of ethanol treatment, 5'D-II activity was enhanced
in the cortex at 2000 h only and in the amygdala at both measuring
times (Fig. 3, right). 5D-III activity was reduced in the
cortex at 0800 h and in the amygdala at 2000 h only. The
tissue levels of T4 and T3
were not measured in this group and the serum levels were not
significantly affected (Table 1).
Group 8
Fourteen days of haloperidol treatment induced an increase in
5'D-II activity in the cortex. All other parameters shown in Fig. 7
remained unchanged. Also, 5'D-I activity was not affected in the
cortex, pituitary or liver (data not shown).
Group 9
Fourteen days of treatment with the atypical neuroleptic drug
clozapine inhibited 5'D-II activity in 5 out of 10 brain regions and
5D-III activity in 4 out of 8 brain regions (Fig. 4). 5'D-I activity was unchanged in all
these brain regions and also in the pituitary, liver, and kidney. The
tissue concentrations of T4 and
T3 were measured in the cortex. They were not
significantly affected by clozapine (data not shown). The serum
concentrations of T4 and T3
rose significantly (Table 1), whereas the TSH levels remained unchanged
(data not shown).
Groups 10 and 11
Twelve hours after a single administration of the highly toxic
dose of 7.5 mmol/kg lithium (i.e. at 0800 h) a
dramatic, almost 6-fold increase in 5'D-II activity was found in the
frontal cortex. The increase was also significant in the group killed
24 h after administration of lithium (2000 h, Fig. 5). This increase was specific for the
cortex because no effects were seen on the 5'D-II activities in the
hypothalamus and pituitary (data not shown). The 5D-III activities
remained unchanged, but cortical levels of T4
fell sharply to about 25% of the original values at both measuring
times. The tissue concentrations of T3 were also
significantly reduced, but to a lesser extent. The serum concentrations
of T4 and T3 were
drastically reduced at 0800 h and still significantly lower than
in the controls at 2000 h (Table 1). The TSH levels remained
unaffected (data not shown). The mean serum concentrations of lithium
were in the toxic range at both 0800 h (2.6 ± 0.1
mmol/liter) and 2000 h (1.9 ± 0.1 mmol/liter).
Group 12
The effects of a single dose of the low dose of 3 mmol/kg lithium
on thyroid hormone parameters after 12 h are shown in Fig. 6. 5'D-II activity was significantly
elevated in all four brain regions. The tissue levels of
T4 were distinctly lower than in controls,
whereas cortical concentrations of T3 were
significantly enhanced in three out of four areas, respectively. The
serum levels of thyroid hormones were not affected. The mean serum
concentration of lithium was 0.44 ± 0.02 mmol/liter,
i.e. below the range considered to be clinically effective
in patients with affective disorders.
Groups 13 and 14
The only effect of an acute dose of carbamazepine was a
significant increase in 5D-III activity at 2000 h in the frontal
cortex (Fig. 5). The serum levels of carbamazepine were not detectable
at either 0800 h or 2000 h.
Group 15
Fourteen days of administration of two different dosages of
lithium had opposing effects on 5'D-II activities (Fig. 5, right). The 0.15% lithium diet reduced 5'D-II activity in
the cortex, whereas the 0.3% lithium diet enhanced it.
T4 levels were significantly elevated after the
low dose and significantly reduced after the high dose. Likewise, the
serum concentrations of T4 and
T3 were enhanced after the 0.15% diet and
lowered after the 0.3% diet (Table 1). The activities of 5D-III were,
however, significantly reduced after administration of both
dosages.
Group 16
Fourteen days of consumption of a 0.4% carbamazepine diet raised
the activity of 5'D-II and reduced the tissue levels of
T4 and T3 in the cortex
(Fig. 5, right). The serum concentrations of
T4 and T3 were also
significantly lowered (Table 1).
Groups 17–20 (antidepressant drugs)
Fourteen days of treatment with the low dose of desipramine (5
mg/kg) raised the levels of T3 in cortical
tissue. However, the activities of both deiodinases remained unchanged
(Fig. 7). The higher dose of 20 mg/kg
induced a significant increase in 5'D-II activity and also in the
tissue concentrations of T3. After the toxic dose
of 50 mg/kg, there was a pronounced increase in 5'D-II activity and the
tissue concentrations of both T4 and
T3 were significantly lower than in the controls.
Fluoxetine induced an increase in 5'D-II activity and a reduction in
5D-III activity. Tranylcypromine significantly enhanced 5D-III activity
without inducing any changes in thyroid hormone concentrations. None of
these treatments significantly affected 5'D-I activity (data not
shown). Fourteen days of treatment with the atypical antidepressant
drug mianserin led to a rather irregular pattern of changes in
deiodinase activities and thyroid hormone concentrations (Fig. 8). Lowered concentrations of 5'D-II were
seen in the cortex, hippocampus, and limbic forebrain and decreases in
5D-III concentrations in the hippocampus and amygdala. However, the
T4 levels were elevated in the hippocampus and
reduced in the limbic forebrain, while those of
T3 remained unaffected. No effects were seen on
5'D-I activity in any brain region. Scrutiny of Table 1 reveals that
several T4 concentrations were significantly
lowered after administration of all four antidepressant drugs. The
serum levels of T3 were significantly reduced
only after the toxic dose of desipramine.
Group 21
The activities of 5'D-II and 5D-III were measured in the cortex
and amygdala in young and old rats. 5'D-II activity was significantly
higher in the young rats than in the old rats, in both regions
(111.8 ± 6.6 vs. 54.0 ± 7.6 fmol
I-/mg protein/h, P = 0.001 and
52.4 ± 5.6 vs. 32.7 ± 1.8 fmol
I-/mg protein/h, P = 0.01 for the
cortex and amygdala, respectively). There was no significant difference
between the activities of 5D-III in cortex in the young and old rats
(78.6 ± 4.4 vs. 86.2 ± 5 fmol
I-/mg protein/min). In the amygdala, however,
5D-III activity was again significantly higher in the younger rats than
in the old rats (568 ± 36 vs. 301 ± 25 fmol
I-/mg protein/min, P = 0.01). A
nonsignificant trend toward lower serum levels of
T4 and T3 was found in the
old rats (Table 1).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As regards the validity of our significant results, as we have
already commented in Materials and Methods, owing to the
large number of statistical calculations (267) about 13 significant
results will have occurred by chance. We cannot therefore rule out the
possibility that some "isolated" significant results may indeed
have occurred by chance. However, almost all of our significant
findings were obtained either in different brain regions, after
different dosages or at two different measuring times, or they were so
pronounced that they even fulfilled the P = 0.00025
criterion. The great majority of our significant findings can therefore
be regarded as valid. One potential methodological shortcoming should,
however, be mentioned. It was recently shown that studies using broken
cell deiodinase assays may not adequately predict the rate of
deiodination in vivo in intact cells (24). The same may well
apply to the activities of 5'D-II and 5D-III. In light of these
new findings our results on deiodinase activities must be confirmed by
in-vivo experiments before they can be regarded as valid.
Furthermore, several of the changes found to be significant in our
study were quantitatively small and therefore of unknown physiological
significance.
Only a few of our significant findings are easy to interpret. Sleep deprivation, for example, induced an enhancement in 5'D-II activity in four brain regions and an increase in the tissue concentrations of T3 in the same areas. As the serum concentrations of T4 were enhanced, the brain levels of T4 probably did not fall, despite a simultaneous increase in 5'D-II activity. It is, however, unclear why the T3 levels were elevated in the cerebellum, although the activity of 5'D-II was not affected. The possibility that the increase in serum levels of T3 leads to direct uptake of T3 in this and also in the other brain regions cannot be excluded. Why the serum levels of T4 and T3 were significantly elevated, despite the fact that the TSH concentrations remained unchanged, is also unexplained. A direct stimulatory effect on the thyroid, such as stimulation of the sympathetic nervous system (25) should be considered.
One of the most unexpected findings of this study was the dramatic
decrease in tissue levels of T4 after different
kinds of intervention such as 12 h fasting, a calorie-reduced diet
for 14 days, and a single, relatively low dose and a single toxic dose
of lithium (Groups 2, 3, and 10–12). After 12 h fasting and acute
administration of lithium, 5'D-II activities were enhanced, while after
14 days on a calorie-reduced diet they were lowered (Figs. 2, 4, and 6). It also seems doubtful whether the relatively small increases of
approx. 30–50% in 5'D-II activity after 12 h fasting and the low
dosage of lithium (Fig. 6) were responsible for the sharp decline in
tissue concentrations of T4 to between 10 and
50% of the initial levels, particularly as these were evident after
only 12 h. As the tissue concentrations of
T3 remained unchanged after a 12-h fast (Fig. 2)
and were only slightly enhanced after injection of lithium (Fig. 6), it
is unclear to which iodothyronine metabolite T4
was metabolized during fasting. As the activity of 5D-III remained
unchanged, enhanced degradation to reverse T3
seems unlikely. Other candidates are T3 sulfate
and 3,3'-T2 sulfate, which have been detected in
rat brain and astrocytes in culture, respectively (26, 27).
Deiodination of T3 to
3,5-T2 would also seem conceivable.
3,5-T2 has recently been detected in human
(28) and rat brain (29), but the enzyme catalyzing this reaction
has not yet been characterized.
Another open question is the mechanism underlying the enhanced 5'D-II activities seen in many of our experimental groups (Groups 1, 2, 8, 10–12, 15, 16, 17, and 18). The increases in 5'D-II activity in most of our experimental groups might be explained on the basis of a fall in tissue concentrations of T4 [e.g. Group 2 (except for the cerebellum) and Groups 10, 11, 12, and 15–18]. However, such an interpretation would not explain the mechanism underlying the fall in tissue levels of T4, particularly in the groups in which the serum levels of T4 remained unchanged (e.g. Groups 2 and 12). Furthermore, after sleep deprivation, the activity of 5'D-II was enhanced despite the fact that the tissue levels of T4 remained unaltered, and after a 14-day calorie-reduced diet the activity of 5'D-II was diminished although the tissue concentrations of T4 had fallen. It therefore seems likely that these treatments influence the activity of 5'D-II directly in the CNS. Many of our data may therefore be interpreted in the following ways. First, conditions such as administration of a toxic lithium dose or chronic carbamazepine treatment, where the experimental pattern of deiodinase activities and tissue hormone levels fit an understandable pattern: in these conditions, depressed tissue levels of T4 may have triggered an increase in 5'D-II activity with maintenance of relatively normal tissue concentrations of T3. Second, in conditions such as sleep deprivation, 14 days’ fasting or desipramine treatment the observed changes in deiodinase activity do not appear to have been triggered by changes in tissue levels of thyroid hormone, but would rather seem to be responsible for alterations in tissue concentrations of T3. One may speculate that in these circumstances the unexplained changes in deiodinase activity may be a direct or indirect effect of the treatment itself. Conceivably this is of physiological benefit (e.g. in the case of sleep deprivation) or related to the therapeutic effects of drugs (e.g. desipramine). Third, in some experiments the observed changes in deiodinase activities are small in magnitude (e.g. Groups 18 to 20) or are not consistent with any of the above-mentioned patterns (e.g. Groups 5 and 9).
One finding, in particular, is currently rather difficult to interpret, namely, the decreases in both 5'D-II and 5D-III activities in the same brain region in several treatment groups (14 days on a low-calorie diet, low-dose, subchronic administration of lithium, and administration of the antidepressant mianserin and the neuroleptic clozapine). The physiological meaning of these changes is not clear because they result in a decrease in both the production and the degradation of T3. 5'D-II has recently been located in astrocytes and tanycytes (30), whereas the 5D-III isoenzyme has recently been located in neurons in vivo (31), but has also been demonstrated in astrocytes in culture (32). Even if both deiodinase isoenzymes were in fact located in different types of cell or subcellular compartments, the relevance of such constellations is not readily apparent. It is also hard to understand why we found lowered tissue concentrations of T4 in an experimental group in which the activities of 5'D-II and 5D-III were both significantly reduced (e.g. 14 days of calorie-reduced diet), whereas the T4 levels were elevated in other groups in which the activities of the two deiodinases were also reduced (Groups 15 and 20).
The prompt effect of ethanol on 5D-III activity (Fig. 3) was
surprising, as this enzyme tends to react very slowly to peripheral
hypo or hyperthyroidism, i.e. over a period of several days
(12, 32). The effects of ethanol would therefore seem to be more direct
and it is unlikely that they are mediated by changes in tissue
concentrations of T3.
With respect to the mechanisms underlying the decreases in serum levels of T4 seen after subchronic treatment with different psychotropic drugs (15, 16, 17, 18, 19, 20), both enhanced tissue uptake with subsequent deiodination to T3 and an increase in peripheral enzyme induction seem possible, as, for example, a rise in UDP-glucuronyl transferase activity in the liver was reported after 3 weeks of treatment with the anticonvulsant diphenylhydantoin (33). Furthermore, it should be noted that we measured protein-bound serum levels of T4 and T3. However, the serum concentrations of free thyroid hormone may be more relevant for the supply of thyroid hormones in the brain. We cannot, therefore, exclude the possibility that our results do not adequately reflect the relationships between the serum levels of free T4 on the one hand, and tissue levels of T4 and the deiodinase activities, on the other.
The activity of the 5'D-I isoenzyme was not affected by any of the treatments. This is consistent with the finding of a previous study that different forms of stress had no effect on 5'D-I activity (15). In light of these results, and also of the fact that this isoenzyme has a considerably higher substrate preference for rT3 than T4 (5) and is not found in the human brain (6), it seems doubtful that it is of physiological importance in the rat CNS.
Several of our findings show that factors such as drug dosage, length of treatment, and time of decapitation strongly influence the effects on thyroid hormone homeostasis. Effects of subchronic treatments on thyroid hormone homeostasis may slowly develop in some cases, reflecting underlying neuronal adaptation mechanisms. Moreover, a complete and valid evaluation of thyroid hormone metabolism and probably also function in the CNS should take into account the circadian components, as the results of a previous study revealed circadian variations both in 5'D-II activity and in the effects of desipramine, lithium, and carbamazepine on this activity (14, 16, 17, 18).
The biochemical mechanisms underlying the different changes in deiodinase activities seen in these experiments remain unexplained. Although it is well known that the thyroid hormone concentrations regulate deiodinase activities (12, 13, 32), it is unclear how all the different forms of treatment described in this study induced the observed changes in tissue concentrations of thyroid hormones. Furthermore, little is known about the neurotransmitter-related regulation of deiodinase isoenzymes in the CNS in vivo. It has repeatedly been demonstrated in astrocytes in culture that cAMP, isoproterenol, norepinephrine, and other factors such as TSH, FGF glucocorticoids, or protein kinase C activators stimulate 5'D-II and/or 5D-III activity (34, 35, 36, 37). Our data suggest that the regulation of deiodinase activities in vivo is highly complicated. The norepinephrine reuptake inhibitor desipramine enhanced 5'D-II activity and the serotonin reuptake inhibitor fluoxetine both enhanced 5'D-II activity and inhibited 5D-III activity. Tranylcypromine is an MAO inhibitor which enhances norepinephrine, serotonin, and dopamine concentrations and the corresponding neuronal activities in the CNS. This drug should therefore have at least some similar effects to desipramine and fluoxetine. However, it had no effect on 5'D-II activity and enhanced that of 5D-III. In light of these results, it seems most likely that the deiodinase isoenzymes are not regulated by a single specific transmitter system, e.g. 5'D-II by norepinephrine, etc. The regulation of these enzymes is probably more complex and as yet not at all understood.
In conclusion, thyroid hormone metabolism and concentrations in the CNS are affected by many physiological or pharmacological influences that may change neuronal activity. Each specific effect on brain function seems to induce a specific "pattern" of changes in thyroid hormone homeostasis in terms of the direction and type of isoenzyme and hormone involved and the brain region affected. The data presented here also indicate that mechanisms other than the two deiodinase isoenzymes may be operating in the regulation of thyroid hormone metabolism in the CNS. Sulfur transferases, for example, are possible candidates.
As regards the physiological significance of these findings, it is very unlikely that all the highly complex and specific changes in thyroid hormone homeostasis in the CNS are not functionally important. In recent years, effects of thyroid hormones on the expression of a large number of genes have also been reported in the adult CNS (e.g. 38–40). It is, however, as yet unclear whether T3 has a direct or indirect effect on these genes. Furthermore, it has been demonstrated that thyroid hormones have a large number of effects on a large variety of parameters of the adult CNS. These effects range from direct modulation of neuronal activity (e.g. 41) to numerous influences on the characteristics of many kinds of G protein-coupled neurotransmitter receptors (e.g. 42) or even morphological changes (43).
The question that currently remains unanswered is by what mechanisms
thyroid hormones affect all these CNS parameters. Such mechanisms may
well include the "classical" effects of T3
and nuclear receptors. However, it is not unlikely that others are
mediated directly at the mitochondrial level, since 
and ß
T3 receptors have recently been identified in
mitochondria (44), a thyroid hormone response element has been
identified in mitochondrial genes (45), and effects of hypothyroidism
on the expression of mitochondrially encoded genes have been reported
in the adult CNS (46). The hypothesis that thyroid hormones may be
critically involved in the adjustment of the mitochondrial production
of ATP to the current ATP requirement seems worthy of further
investigation. The same applies to that of a possible specific effect
of thyroid hormones directly at synaptic membranes, because effects of
both T4 (e.g. 47) and
T3 (41) at plasma membranes of CNS tissue have
been reported.
|
Footnotes |
|---|
Received June 9, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ß-related
T3 receptors in rat liver mitochondria. Eur
J Cell Biol 62:105–113[Medline]
This article has been cited by other articles:
 |
|
 |
|
  J. P Brouwer, B. C Appelhof, R. P Peeters, W. J G Hoogendijk, J. Huyser, A. H Schene, J. G P Tijssen, R. Van Dyck, T. J Visser, W. M Wiersinga, et al. Thyrotropin, but not a polymorphism in type II deiodinase, predicts response to paroxetine in major depression. Eur. J. Endocrinol., June 1, 2006; 154(6): 819 - 825. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. C. Appelhof, J. P. Brouwer, R. van Dyck, E. Fliers, W. J. G. Hoogendijk, J. Huyser, A. H. Schene, J. G. P. Tijssen, and W. M. Wiersinga Triiodothyronine Addition to Paroxetine in the Treatment of Major Depressive Disorder J. Clin. Endocrinol. Metab., December 1, 2004; 89(12): 6271 - 6276. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Broedel, M. Eravci, S. Fuxius, T. Smolarz, A. Jeitner, H. Grau, G. Stoltenburg-Didinger, H. Plueckhan, H. Meinhold, and A. Baumgartner Effects of hyper- and hypothyroidism on thyroid hormone concentrations in regions of the rat brain Am J Physiol Endocrinol Metab, September 1, 2003; 285(3): E470 - E480. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Everson and T. S. Nowak Jr. Hypothalamic thyrotropin-releasing hormone mRNA responses to hypothyroxinemia induced by sleep deprivation Am J Physiol Endocrinol Metab, July 1, 2002; 283(1): E85 - E93. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Pinna, O. Brodel, T. Visser, A. Jeitner, H. Grau, M. Eravci, H. Meinhold, and A. Baumgartner Concentrations of Seven Iodothyronine Metabolites in Brain Regions and the Liver of the Adult Rat Endocrinology, May 1, 2002; 143(5): 1789 - 1800. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||
