THE EFFECTS OF THYROID STATE ON BETA-ADRENERGIC AND SEROTONERGIC RECEPTORS IN RAT BRAIN

The effects of hyperthyroidism or hypothyroidism, alone or in combination with the tricyclic antidepressant desmethylimipramine (DMI), on brain beta-adrenergic and serotonin (5HT 2) receptors were studied in adult male Sprague - Dawley rats. Intraperitoneal administration of T 3 or T 4 for 7 days increased the number of cortical beta-adrenergic and 5HT 2 receptors. These increases were significant at levels of 250 ,.g/kg or above for both hormones. Neither thyroidectomy nor "reverse" T 3 (rT 3) (500 ,.glkg) produced an effect on either receptor type. The down-regulation of beta-adrenergic receptors produced by daily subcutaneous injections of 20 mg/kg of DMI for 7 days was partially offset by concurrent administration of T 4, whereas the down-regulation of 5HT 2 receptors produced by the drug was not affected by concurrent administration of either T 3 or T.- Hypothyroidism (thyroidectomy) did not significantly affect the adaptation of these receptor populations to DMI. As regards brain regions other than cortex, T 4 (250 ,.glkg) produced the same changes in hippocarnpus as in cortex, while thyroidectomy decreased beta-adrenergic receptors only in the cerebellum. Thyroxine also elevated 5HT 2 receptors in the hippocampus; thyroidectomy caused a significant decrease in 5HT 2 receptors in the striatum.


INTRODUCTION
THE INFLUENCE of endocrine status on behaviors and affect is exemplified by the high incidence of lethargy, confusion and depressive symptoms in clinical hypothyroidism. Clinical depression occurs in about 40% of hypothyroid patients, and in most of these patients replacement therapy produces prompt remission of depressive symptoms (Whybrow et al., 1969Sachar, 1975; the 5'-monodeiodinase which converts T 4 to T 3 , found in almost every peripheral organ (type I), is different from the enzyme (type II) present in brain, brown fat and pituitary (Visser et al., 1982). These enzymatic differences may be responsible for the marked differences in changes which occur in central and peripheral levels of T 3 during starvation, for example, in which peripheral T 3 is reduced and brain T 3 relatively unchanged (Larsen et al., 1981).
Beginning with what is now called the catecholamine hypothesis (Prange, 1964;Schildkraut, 1965;Bunney and Davis, 1965), data gathered from studies of neurochemical effects of antidepressants in animals have helped form the basis of a series of theories and modifications thereof, concerning the etiology of depression and the mechanisms of action of antidepressant therapy (for review, see Stone, 1983). Currently in focus is the concept of Sulser et al. (1978) which, like many of its predecessors, proposes an association between central noradrenergic activity and the depressed state: specifically, the mechanism underlying successful antidepressant therapy is desensitization of the central beta-adrenergic receptor-coupled adenylate cyclase system. Evidence that 5HT receptors are also down-regulated by TCA (Peroutka and Snyder, 1980;Sugrue, 1981) and that serotonergic innervation seems to play a permissive role in beta-adrenergic receptor desensitization (Brunello et al., 1982) supports the idea of an indoleamine-catecholamine link (Prange et al., 1974) and has resulted in further modification of this concept (Sulser, 1984a(Sulser, , 1984b. Catecholamine -indoleamine theories of the neurochemistry of depression, along with clinical findings linking thyroid states to affective disorders, have evoked considerable interest in the effects of thyroid status on monoamine turnover and central adrenergic receptors. Indeed, there have been several recent and, in some aspects, conflicting reports (Gross et al., l 980a;Perumal et al., 1984;Atterwill et al., 1984;Schmidt and Schultz, 1985) on the effects of thyroidal condition on the status of beta-adrenergic receptors in rat cerebral cortex and other brain regions (see below). Here we present data on the effects of thyroidectomy or administration of thyroid hormones on beta-adrenergic and 5HT 2 receptor binding in rat brain and how these effects interact with receptor effects produced by chronic treatment with desmethylimipramine (DMI).

METHODS
Intact, thyroidectomized and sham-operated adult male Sprague-Dawley rats (Charles River Breeding Laboratories, Wilmington, MA) weighing 125-150 g upon arrival were used in this study. Mature animals were used to avoid developmental effects of thyroid status on brain monoamine systems. Surgical thyroidectomies or sham-operations were performed by the vendor. No less than 5 days after arrival the animals were injected subcutaneously (s.c.) once daily at 1()00-1200 hr for 7 or l() consecutive days with desmethylimipramine HCl (DMI; 20 mg/kg) or vehicle (0.9% NaCl) and intraperitoneally with various doses of the free acids of T (15-1000 µg/kg), T 4 (15-500 µg/kg), rT 3 (500 µg/kg) or vehicle [0.9% NaCl:methanol:NHpH (8 N); 396:3:1 Cv/v)]. The TCA DMI was selected for use in this study because it reliably down-regulates both beta-adrenergic and 5HT 2 receptors in rat cerebral cortex (Sugrue, 1981). Throughout the treatment protocol, rats were housed two to a cage and given free access to water and laboratory chow. Animal quarters were kept at a temperature of 22-25°C, and the light-dark cycle was 12 hr.
All animals were weighed 24 hr after the last injections. They were then killed by decapitation, and blood was collected at that time for analysis of serum thyroid hormones by radioimmunoassay (T 3 , T 4 -Becton-Dickinson Immunoassay, Orangeburg, NY; rT 3 -Serono Laboratories, Braintree, MA). Brains were immediately removed and kept on ice while the corpus striatum, cerebellum and hippocampus and/or the frontal cerebral cortex were dissected out, weighed and frozen on dry ice. Brain tissues were kept frozen at -50°C until thawed for preparation of membranes for measurement of beta-adrenergic or serotonin 2 (5HT 2 ) receptor binding based on the procedures of Byland and Snyder (1976) or Leysen et al. (1982), respectively.
A crude membrane fraction was prepared from brain regions by homogenization of tissue in 19 vols 0.32 M sucrose followed by centrifugation at 3000 g for l() min. The pellet was discarded, and the supernatant was centrifuged at 48,000 g for l() min. The pellet was resuspended by homogenization in 40 mM Tris -HCl buffer, pH 7.4 and was recentrifuged (48,000 g for l() min). For beta-adrenergic receptor binding a portion of the final pellet was resuspended in 40 mM Tris-HCl (pH 7 .4) buffer at a concentration corresponding to 50 mg original tissue/ml. Binding incubations were carried out in triplicate in 1 ml of medium containing 40 mM Tris-HCl pH 7.4 and [ 3 H]dihydroalprenolol ([3H]DHA; 102.7 Ci/mmol) at various concentrations. The amount of tissue used per tube corresponded to 5 mg original wet weight and contained 300-400 µ.g protein as determined by the method of Lowry et al. (1951). At the end of a 25 min incubation at 25~C. samples were filtered on glass fiber discs (25 mm diameter, 0.3 µ.pore size, Gelman Inc., Ann Arbor, Ml) and washed twice rapidly with 5 ml Tris buffer. Filter discs were then dried and counted in 5 ml Aquasol scintillation mixture with a Beckman LS 7000 scintillation counter at an efficiency of 38-43%. Control incubations containing unlabeled competing ligand (10-6 M alprenolol) were carried out simultaneously in order to determine the extent of non-specific binding (about 30%).
Membrane fractions of different regions from each animal were first assayed at a [ 3 H]DHA concentration of 1 nM. These data were used to compare mean receptor binding of the different treatment groups. The unused portions of membranes from animals of the same treatment groups were then pooled to obtain sufficient material for additional binding assays, in triplicate, with seven different [ 3 H]DHA concentrations ranging from 0.1 to 5.0 nM. These data were used for Scatchard type analysis (Scatchard, 1949). Receptor binding of [ 3 H]DHA in brain samples from animals in the different treatment groups was analyzed statistically with Tukey 's range test (Miller, 1966) or Dunnett' s test for multiple comparisons (Dunnett, 1964) after ANOVA. Significant differences were defined asp values of0.05 or less.
Portions of the same membrane preparations used for studies of [ 3 H]DHA binding, except cerebellar membranes, were used for characterizing [ 3 H]ketanserin ([ 3 H]KET) binding: cerebellar membranes contain few 5HT 2 receptors. An assay procedure similar to the [ 3 H]DHA binding assay was used. Incubations were initially carried out for 15 min at 37°C in triplicate in 1 ml Tris-HCl buffers, pH 7.4, containing 1 nM [ 3 H]KET, and later in pooled preparations at various concentrations of the radioligand (0.5-20 nM). Non-specific binding was determined with 10-6 M methysergide.
All radioligands were purchased from New England Nuclear Corporation (Boston, MA); thyroid hormones were bought from either Sigma Chemical Co. (St. Louis, MO) or Hennings (Berlin, F .R.G.). DMI was a gift from the Revlon Health Care Group, Research and Development Division, Tuckahoe, NY.

The number of animals is in parentheses.
We observed that neither 10-s M T 3 nor T 4 had an effect on the binding of [3H]DHA or [3H]KET. The effects of thyroid hormones on beta-adrenergic and 5HT 2 receptor binding in the striatum, cerebellum (beta-adrenergic receptor binding only) and hippocampus are described in Table IV. Administration of thyroid hormones significantly increased binding to both types of receptors only in the hippocampus, whereas hypothyroidism only reduced beta-adrenergic receptors in the cerebellum. Scatchard analysis of [3H]DHA binding in pooled hippocampal membrane preparations indicated that the increase in binding in animals that received 250 The number of animals is in parentheses.
The effects on receptor binding of phannacological agents with and without thyroid hormones or thyroidectomy are presented in Table V. The tricyclic antidepressant DMI produced a downregulation of both beta-adrenergic and 5HT 2 receptors. As expected, an opposite effect was produced by 7 days of reserpine treatment. The down-regulation of beta-adrenergic receptors produced by DMI was significantly attenuated by 500 µg T 4 ; the two treatments produced opposite effects that were offsetting. However, the significant but slightly less potent downregulation of 5HT 2 receptor binding produced by DMI was apparently not antagonized by concurrent administration of either T 3 or T 4 • The effect of DMI in thyroidectomized rats was not different from that in sham-operated or intact controls. DISCUSSION Our study demonstrates that peripheral administration of thyroid hormones produces specific dose-dependent changes in two populations of cerebral cortical receptors in neurotransmitter pathways affected by antidepressant therapy and thought to be involved in the etiology of depression. Similar studies of thyroidally induced changes on beta-adrenergic receptors have been conducted by other investigators, often using different strains of rats, different manipulations of the thyroid axis and different methods of measuring receptor binding. Understandably, the effects of thyroid status on this receptor type are not completely clear, and there are areas of seeming contradiction.
Our data indicate that chronic thyroid hormone administration produces a dose-related increase in [ 3 H]DHA binding to cortical beta-adrenergic receptors. We observed that significant differences from vehicle-Injected controls occurred consistently only in animals that received a relatively large dose of T 3 or T 4 ( > 100 µ.g/kg). In general, this finding is consistent with that of Perumal et al. (1984), who observed a significant up-regulation of cortical beta-adrenergic receptors after dosing rats with 375 µ.g/kg T 4 for 7 consecutive days, and with the work of Atterwill et al. (1984), who reported no significant change in this receptor after giving Porton strain rats 100 µ.g/kg T 3 for 14 days. Our findings are, however, apparently discordant with those of Schmidt and Schultz (1985), who reported that small doses of T 3 (15 µ.g/kg) or T 4 (50 µ.g/kg) down-regulated the noradrenergic cyclic AMP generating system and the number of beta-adrenergic receptors in the cerebral cortex of Sprague-Dawley and Long -Evans, but not of Wistar and F-344, rats.
We also observed that the affinity of beta-adrenergic receptors for [ 3 H]DHA decreased with increasing doses of T 3 and T 4 • Although we were unable to provide an adequate explanation for this binding decrease, it does not appear to be an artifact of interference by T 3 or T 4 in the binding assay, and the modest increases in receptor number suggest that steric hindrance is not involved. It is more likely that thyroid hormones produce some membrane change(s) which preferentially affect the affinity of the beta-receptor.
We did not find any significant change in rat cortical beta-adrenergic receptor binding 28-42 days after thyroidectomy. This is also in agreement with the finding of Atterwill et al. (1984) in rats made hypothyroid by treatment with 6-propyl-2-thiouracil (PTU) for 2 weeks, but not in agreement with Gross et al. (1980a), who reported a significant reduction of cortical betaadrenergic receptors in Wistar rats after 6 weeks of PTU treatment. Atterwill et al. (1984) attributed this difference to morphological changes in the brains of chronically hypothyroid rats.
Although most authors have focused their studies of drug-or hormone-induced adaptation of beta-adrenergic receptors on the cerebral cortex, others have noted changes in other brain regions. Like Perumal et al. (1984) and Atterwill et al. (1984), we observed some regional differences in the effects of thyroid status on beta-adrenergic receptors. We observed no significant effects on beta-adrenergic receptor binding in striatal tissue after administration of relatively large doses of thyroid hormones or after thyroidectomy, but we did find a 16% decrease in binding in cerebellar membranes of thyroidectomized rats. Large doses of T 3 or T 4 increased [ 3 H]DHA binding in the hippocampus. Binding of [3H]KET in the striatum was increased by T 4 and decreased by thyroidectomy. A similar increase in [3H]KET binding was observed in hippocampal membranes of hyperthyroid rats; tissue from this region from thyroidectomized animals was not studied. We also observed the previously described (Banerjee et al., 1977) down-regulation of betaadrenergic receptors after chronic administration ofDMI. As expected on the basis of our results with thyroid hormone administration alone, we found that concurrent administration of only large doses of thyroid hormones had a significant antagonistic effect on the down-regulation produced by the drug. In a study of the interactions of DMI and T 3 , Atterwill et al. (1986) noted that a single dose of T 3 (100 µ.g/kg) on day 1 slightly decelerated the rate of betaadrenergic receptor down-regulation induced by daily injections of DMI (5 mg/kg) for 14 days. It is interesting that DMI was effective in down-regulating beta-adrenergic receptors of thyroidectomized rats; it has been suggested, though only by anecdotal evidence, that TCAs are not as effective in hypothyroid patients (Avni et al., 1967). If down-regulation occurs in hypothyroid depressed patients, even as these patients show a poor or null antidepressant response, the down-regulation cannot be a sufficient explanation for the usual therapeutic action ofTCAs.
The dose-related increase in 5HT 2 receptor binding in the cortex, striatum and hippocampus produced by peripheral administration of thyroid hormones has not been reported previously. Although in most cases the thyroidally induced changes in the 5HT 2 and beta-adrenergic receptors were in the same direction and of similar degree, thyroidectomy reduced the number of 5HT 2 receptors in the striatum but produced no effect on beta-adrenergic receptor binding. Furthermore, we saw no significant antagonism of the DMI-inducd 5HT 2 receptor downregulation in the cerebral cortex by either T, or T 4 • Our general finding that significant changes in either beta-adrenergic or 5HT 2 receptor binding are induced only by large doses of thyroid hormones is consistent with evidence, also in rats, that the thyroid status of the brain remains relatively stable in the face of marked changes in peripheral thyroid hormone levels (Dratman et al., 1983). Of course, this does not preclude the possibility that small changes may have substantial functional effects. The directions of change in beta-adrenergic receptors which we report here are generally consistent with our previous work (Lipton et al., 1968;Prange et al., 1970) and the work of others (Parker, 1972), indicating that catecholamine synthesis and turnover in brain are reduced in hyperthyroidism and increased in hypothyroidism. However, opposite effects also have been reported (Engstrom et al., 1974. Jacoby et al., 1975Strombom et al., 1977), and it is not known whether the effects of thyroid hormones on this receptor or the 5HT 2 receptor population result from an adaptation to changes in synaptic level of endogenous ligands, stimulation of protein synthesis or other mechanisms.
Although the major emphasis of the work of Schmidt and Schultz ( 1985) is on cortical cyclic AMP generating systems, and not the beta-adrenoceptor per se, their finding that T, or T 4 down-regulates cortical beta-receptors in some strains of rats seems discordant with the present report and that of Perumal et al. (1984), as well as with earlier studies by Frazer et al. (1974) and Gross et al. (1980b) on the effects of thyroid status on the cortical adenylate cyclase system. Nevertheless, in the light of theories that advocate down-regulation ofbeta-adrenergic receptorcoupled adenylate cyclase systems as the mechanism of action of antidepressant drugs, their findings are consistent with the clinical observation that small doses of thyroid hormones can potentiate the action of antidepressant drugs. Although it is speculative to infer from studies of rats what occurs or does not occur in man, our data suggest that mechanisms other than changes in postsynaptic noradrenergic or serotonergic receptors may well be involved in the potentiation of antidepressants by small amounts of thyroid hormones, and that behavioral and affective changes associated with small disturbances in peripheral levels of thyroid hormones are probably not produced by or reflected in conspicuous changes in the binding of beta-adrenergic or 5HT 2 receptors.
A problem inherent in studying brain effects of thyroid states as defined by peripheral criteria is that, in rats at least, thyroid states of the brain and periphery may be dissociated. For example, it has been shown that DMI reduces the levels of T 4 in synaptosomes without altering plasma T 4 levels (Dratman and Crutchfield, 1979), while in starvation brain T, remains relatively unchanged as plasma T, levels decrease sharply (Larsen et al., 1981). Elucidating the neuronal effects of different thyroid states and their relationships to behavior and affect remain an important challenge.