The Influence of Hyperthyroidism andHypothyroidism on the /8-Adrenergic Responsiveness

of the Turkey Erythrocyte

JOHNP. BILEZIKIAN, JOHNN. LOEB, and DONALDE. GAMMON,Department ofMedicine, College of Physicians and Surgeons, Columbia University,New York 10032

A B S T RA C T The mechanisms responsible for al-tered adrenergic tone in hyperthyroidism and hypo-thyroidism are not fully understood. To investigate thesemechanisms, the /8-adrenergic receptor-cyclic AMPcomplex of the turkey erythrocyte was studied amonggroups of normal, hyperthyroid, and hypothyroid tur-keys. In erythrocytes obtained from hypothyroid tur-keys, there were fewer,8/-adrenergic receptors than innormal cells as determined by the specific binding of[1251]iodohydroxybenzylpindolol, as well as associateddecreases both in catecholamine-responsive adenylatecyclase activity and in cellular cyclic AMPcontent.In contrast, erythrocytes obtained from hyperthyroidturkeys contained the same number of,8-receptors andhad the same catecholamine-responsive adenylatecyclase activity as cells from normal birds. Other char-acteristics of the ,8-receptors in cells from hyper-thyroid birds were indistinguishable from those pres-ent in normal erythrocytes. However, within the rangeof circulating catecholamine concentrations, 5-50 nM,the erythrocytes of the hyperthyroid turkeys generatedsubstantially more cyclic AMPafter exposure to iso-proterenol than did normal cells. These results suggestthat thyroid hormone affects /3-receptor-cyclic AMPin-terrelationships in the turkey erythrocyte by two dis-tinct mechanisms: (a) In hypothyroidism, both /8-re-ceptors and catecholamine-dependent cyclic AMPformation are coordinately decreased; (b) in hyperthy-roidism, /8-receptors are unchanged but there is anamplification of the hormonal signal so that occupation

A preliminary report of this work was presented at the60th Annual Meeting of the Endocrine Society, Miami Beach,Fla., 14 June 1978.

Dr. Bilezikian is Molly Berns Senior Investigator of theNew York Heart Association and the recipient of ResearchCareer Development Award 5 K04 HL-00383.

Received for publication 31 July 1978 and in revised form12 October 1978.

184

of a given number of receptors at physiological con-centrations of catecholamines leads to increased levelsof cyclic AMP.

INTRODUCTIONThyrotoxicosis is associated with cardiac, metabolic,and nervous system manifestations of increased adren-ergic activity (1). The remarkable efficacy of the /-adrenergic inhibitor propranolol in alleviating the signsand symptoms of thyrotoxicosis (2) without affectingcirculating thyroid hormone levels (3) has implicated aspecific influence of the thyroid hormones upon /-adren-ergic mechanisms of catecholamine action.

Numerous experimental models to study the bio-chemical bases for the apparent increase in adrenergicresponsiveness in hyperthyroidism and the conversein hypothyroidism have focused upon catecholamine-sensitive adenylate cyclase and phosphorylase activi-ties as well as upon direct physiological concomitantsof catecholamine actions (4, 5). An additional potentialsite of thyroid hormone influence which could mediatepossible effects at these more distal points is the /8-receptor itself, the putative cellular recognition site for,8-adrenergic catecholamines. The development of suit-able radioligands that can be used to detect specificbinding sites with the characteristics of /8-adrenergicreceptors (6) has recently permitted an inquiry into thispossibility (7-9). The purpose of the current investiga-tion was twofold: (a) to examine the hypothesis that,8-adrenergic receptors may be altered in states of thy-roid hormone excess and deficiency; and (b) to relatedata on the characteristics of 8-receptors in hypothy-roidism and hyperthyroidism to biochemical eventsthat follow their occupation. The turkey erythrocytewas used as the experimental model because both its,/-receptor (10, 11) and its physiological responses to,8-adrenergic agents (12, 13) have already been wellcharacterized, and because sustained hyperthyroidism

J. Clin. Invest. © The American Society for Clinical Investigation, Inc., 0021-9738179/0210184109 $1.00Volume 63 February 1979 184-192

and hypothyroidism were found to be convenientlyestablished in the turkey.

METHODSMaintenance of turkeys. White female turkeys weighing

12-15 lb were obtained from The Butcher, Bronx, N. Y.,banded, and kept in three separate cages. Normal turkeyswere fed a well-balanced diet (Layena Complete Animal Feed,Ralston Purina Co., St. Louis, Mo.) and were provided withwater ad libitum. Hyperthyroid turkeys were maintained onthe same diet, except for the addition of L-thyroxine (3 ug/ml)to the drinking water. Their daily intake of L-thyroxine wasin the range of 600-900 ,ug. Hypothyroidism was establishedin one of two ways: (a) by the provision of a low-iodine diet(14) fortified with vitamins (ICN Pharmaceuticals Inc., Cleve-land, Ohio) and the addition of 0.5% sodium perchlorate toto the drinking water; or (b) by the intravenous administra-tion of an ablative dose (10 mCi) of 1311-. When radioactivityin the serum of the latter birds was no longer detectable,they were maintained among the normal turkeys and fed anidentical diet. Each of the above groups of turkeys containedthree to four birds.

Preparation of plasma samples, wvashed erythrocytes, anderythrocyte hemolysates. Heparinized blood was obtainedby syringe from a wing vein. Blood from at least three turkeysin each group was sampled in any given experiment. Aftercentrifugation at 400 g for 10 min, the plasma was removedand stored in aliquots at -20°C for subsequent assays. Eryth-rocytes were washed by suspension in incubation buffercontaining 150 mMNaCl, 10 mMKCl, 11.1 mMglucose, and10 mMTris at pH 7.4, and were washed three times beforeuse. For the adenylate cyclase assay, an aliquot of washed,packed erythrocytes was hemolyzed in 2 vol of buffer con-taining 50 mMTris at pH 7.5, 5 mMKCI, 1 mMMgSO4,and 20 mMdithiothreitol (11). After centrifugation at 6,400 gfor 15 min, the hemolyzed erythrocyte preparations were re-suspended in the same buffer, centrifuged again, and finallyresuspended in a volume corresponding to twice that of theoriginal packed cells. This crude, adenylate cyclase-rich frac-tion was stable for at least 3 mo when stored at -80°C. Forbinding studies, and for the determination of intracellularcyclic AMP, aliquots of washed cells were suspended to theiroriginal volume in incubation buffer, maintained at 4°C, andstudied on the same day.

Adenylate cyclase assay. Adenylate cyclase activity incrude erythrocyte hemolysates was determined in a well-established assay that depends upon the conversion of[a-32P]ATP (ICN Pharmaceuticals Inc.) to [32P]cyclic AMP(15).Data points for all experiments are the averages of triplicatedeterminations (coefficient of variation, 15%). Adenylatecyclase activities of different groups of turkey erythrocyteswere all measured on the same day.

Preparation of [125I]iodohydroxybenzylpindolol. Hy-droxybenzylpindolol (HYP)l was iodinated according to meth-ods previously described (10, 16). Before use, an aliquot of[125I]HYP was reduced to dryness under nitrogen and re-constituted in the assay buffer.

Binding assay. Freshly obtained, washed, intact turkeyerythrocytes (2.5 x 107/ml) and [1251]HYP (40 pMunless other-wise indicated) were incubated with other compounds asnoted in the individual experiments. After equilibrium wasreached at 37°C (30 min), 100-,ul samples were removed intriplicate and immediately filtered over Gelman A/E glassfilters (A. H. Thomas, Philadelphia, Pa.). These filters re-

1Abbreviations used in this paper: Gpp (NH)p, guanylyl 5'-imidophosphate; HYP, hydroxybenzylpindolol.

tained the erythrocytes and the ligand bound to them. Theywere washed with 12 ml of Tris buffer (10 mM, pH 7.5) atroom temperature, and the radioactivity retained on the filterswas determined in a Packard Auto-Gamma spectrometer(Packard Instrument Co. Inc., Downers Grove, Ill.). Underthese conditions the assay blank (radioactivity trapped on thefilter in the absence of cells) was 150/15,000 cpm filtered(1%), a negligible component of the 35-50% bound to theerythrocyte suspension. The results are reported in terms ofspecific binding defined as that component of total bindinginhibited by 0.1 ,uM unlabeled HYP. Specific binding wasgenerally 60% of the total binding. The coefficient of varia-tion among triplicate determinations was 5%. All experimentscomparing the binding characteristics of different groups ofturkey erythrocytes were performed on the same day.

Measurement of cyclic AMP. Intact, washed erythrocytes(7.5 x 108/ml) were incubated at 37°C for 30 min in the pres-ence of specific agents noted in the individual experiments,and the cells were processed according to a previously pub-lished protocol (12). Experiments comparing the amount ofcyclic AMPgenerated by erythrocytes from different groupsof turkeys were performed on the same day. Cyclic AMPwasdetected by radioimmunoassay according to the method ofSteiner et al. (17).

Determination of catecholamines. Plasma samples wereobtained from awake turkeys under the same conditions thatblood was obtained for other portions of this study. The tur-keys were gently held but were not otherwise restrained.Blood was immediately introduced into iced test tubes con-taining EGTAand glutathione and all samples assayed inquadruplicate according to the method of Passon and Peuler (18).

L-Thyroxine and L-triiodothyronine. L-Thyroxine and L-triiodothyronine were measured by radioimmunoassay (19).There was no significant overlap by either specific radioim-munoassay for the other hormone. Unextracted plasma wasused. The lower limit of detectability for L-thyroxine was 0.25,ug/ml, and, for L-triiodothyronine, 10 ng/ml. Plasma samplesfrom all turkeys drawn at a specified time were analyzedsimultaneously.

Other biochemical measurements. Uric acid, total protein,globulin, total lipids, cholesterol, serum glutamic-oxaloacetictransaminase, serum glutamic-pyruvic transaminase, andlactic dehydrogenase were all determined by standard auto-Analyzer techniques (Technicon Instruments, Tarrytown,N. Y.). Creatine phosphokinase was determined by a pre-viously published method (20).

RESULTS

Establishment of the hyperthyroid and hypothyroidstates. Turkeys developed chemical hyperthyroidismwithin 1-2 wk and hypothyroidism within 6-8 wk afterthe treatment protocols were initiated (see Methods).Over an 8-mo period of observation the hyperthyroidturkeys were noted to be excitable, hyperkinetic, andapprehensive, whereas the hypothyroid birds weregenerally docile and much less active. In addition, thehyperthyroid birds showed conspicuous signs of heatintolerance, including panting and standing with theirfeet in water, whereas hypothyroid turkeys oftenshivered and warmed their heads under their wings.Basal body temperature was significantly (P < 0.002)elevated in hyperthyroid turkeys (41.0+0.12°C) and de-creased in hypothyroid turkeys (39.8+0.10°C) as com-pared with that of normal birds (40.5-+0.06°C). The rest-

Thyroid Status and /3-Adrenergic Responsiveness 185

ing heart rate was higher in the hyperthyroid (262+18/min) than in the euthyroid birds (217±+19/min). Theseclinical observations were confirmed by measuring thecirculating concentration of L-thyroxine, which was10.0+1.4 ,ug/dl in the hyperthyroid birds and 0.34+0.04,ug/dl in the hypothyroid birds-both significantly(P < 0.05) different from normal (0.57+±0.08 ,g/dl). The"low" thyroxine level in the normal turkey-relativeto the level in human subjects-reflects the absenceof specific thyroid hormone-binding proteins in avianspecies (21). The circulating L-triiodothyronine con-centration in the hypothyroid birds (22 ng/ml) was alsosignificantly lower than normal (111+20 ng/ml; P< 0.05). It has been previously reported in certain mam-mals that the levels of circulating catecholamines areinfluenced by thyroid hormone (22); among thesegroups of turkeys, however, the total circulating cate-cholamine levels did not significantly differ (range:14+4 to 23±6 nM). Hyperthyroidism was associatedwith other departures from normal as indicated bylower serum concentrations of uric acid, total lipids,globulins, and cholesterol. In hypothyroid turkeys, con-centrations of the following serum components wereabove the normal range: cholesterol, uric acid, serumglutamic-pyruvic transaminase, serum glutamic-oxalo-acetic transaminase, lactic dehydrogenase, and creatinephosphokinase. These observations were repeatedlyand reproducibly made over the 8 to 10-mo period ofstudy.

Binding of [1251]HYP to intact turkey erythrocytes.Intact erythrocytes from hypothyroid, normal, andhyperthyroid turkeys demonstrated specific binding ofthe radioligand [125I]HYP. In eight separate experi-ments in which erythrocytes from individual turkeyswere studied separately at a ligand concentration of 40pM, the average specific binding of [125I]HYP by hyper-thyroid turkey erythrocytes (2.95±0.33 fmol/107 cells)was not significantly different from the binding ob-served in euthyroid turkey erythrocytes (2.74 ±0.39fmol/107 cells). In contrast, hypothyroid turkey eryth-rocytes bound significantly less [1251]HYP (2.26±0.23fmol/107 cells) than did cells from normals (P < 0.05).The same decrease in erythrocyte binding capacity wasnoted for the radiothyroidectomized turkeys as for theturkeys in which hypothyroidism had instead been in-duced by diet. This reduction in ligand binding wasfurthermore observed at free concentrations of [125I]IHYPas low as 10 pM. Inasmuch as the mean corpus-cular volume for erythrocytes in all groups was similar(140±4 ,um3), the observed differences in ligand bind-ing could not be ascribed to differences in membranesurface area per cell but instead reflected differencesin membrane receptor density. Binding in each casewas saturable. Half-maximal occupation of specificbinding sites occurred at approximately 40 pM for allthree groups (Fig. 1). By Scatchard analysis (23) the

r-

0

'B 4.0

8

-E4-

O 3.0

010 2.0

1.0H

Z.s

0.4

0.3

0.2

A- ~0.1 AAA

0

2 3 4

E15 SY OUND(fmol/lI- cells)

20 40 60 80 100 150

(1251] HYP (pM)

FIGURE 1 Saturation offl-receptor sites in erythrocytes fromnormal, hyperthyroid, and hypothyroid turkeys as a functionof radioligand concentration. Turkey erythrocytes (3 x 107/ml)were incubated with increasing concentrations of [1251]HYPas noted on the ordinate. Binding was determined as describedin Methods. For every concentration of ligand, nonspecificbinding was determined by the amount of radioactivity boundin the presence of 0.1 ,uM unlabeled HYP. This value wassubtracted from the total binding at that concentration ofligand, resulting in the amount of [1251]HYP specifically boundwhich is depicted in the figure. The data shown are repre-sentative of four separate experiments. Individual data pointsare the means of triplicate determinations. (Inset) The dataare plotted according to Scatchard (23). (0), Normal; (A),hyperthyroid; (A), hypothyroid.

binding capacity of the hypothyroid turkey erythro-cytes was half that of normal (Fig. 1, inset). The slopeof the line for the hypothyroid birds was not significantlydifferent from the slopes for normal or hyperthyroidturkeys. Association kinetics were rapid, with half-maximal binding being established within 5 min at370C for all groups (data not shown). Binding was alsoreversible, with similar dissociation rate constants(k-,) in the range of 0.034 to 0.047/min for all threegroups (Fig. 2). Dissociation of bound ligand occurredat the same rate and to the same extent whether ex-cess unlabeled HYPor dilution alone was used to dis-sociate [1251]HYP from the receptor. These observa-tions suggest that negatively cooperative interactionsbetween receptor sites did not occur in any of the groupsstudied. The concentration at which the 3-adrenergicagonist isoproterenol was able to inhibit binding of[125I]HYP half-maximally was similar among the threegroups (Fig. 3). Prior exposure of turkey erythrocytesto isoproterenol did not alter the number of detectablebinding sites, nor did it change their affinity for,8-adrenergic agonists (data not shown). Guanylyl 5'-imidophosphate [Gpp(NH)p] did not alter the bindingaffinity for (-)-isoproterenol. Strict stereospecificity for(-)-antipodes of, -adrenergic agonists and antagonistswas observed in each case. The salient feature in thesebinding studies, therefore, was a decrease in the num-ber of 8-receptors in the hypothyroid turkey erythro-cyte. In all other respects, the characteristics of the,8-receptor in the hypothyroid erythrocyte were indis-

J. P. Bilezikian, J. N. Loeb, and D. E. Gammon186

20 40 60 80 100 120TIME (min-)

FIGURE 2 Dissociation of [125I]HYP from the (3-receptor ofhyperthyroid, normal, and hypothyroid turkey erythrocytes.Freshly obtained turkey erythrocytes were washed and resus-pended (3 x 107/ml) in incubation buffer containing 40 pM[125I]HYP at 250C. After 60 min, the cells were immediatelychilled, centrifuged at 3,000g and washed three times incold incubation medium. The cells were resuspended in a100-fold volume excess of incubation medium, and 10-mlaliquots were filtered in duplicate at the times indicated.The data are represented as the percent of [125I]HYP specifi-cally bound to normal erythrocytes at equilibrium, or, inset,as the natural logarithm (ln) of the difference between bindingat equilibrium and binding at t = 120 min. (0), Normal;(A), hyperthyroid; (A), hypothyroid. The lines were deter-mined by linear regression analysis.

tinguishable from normal. On the other hand, the ,3-receptor of the hyperthyroid turkey erythrocyte wasidentical in all respects to normal, including receptornumber.

Adenylate cyclase activity. Basal adenylate cyclaseactivity was found to be similar in crude hemolysatesof hyperthyroid and euthyroid erythrocytes, but it wassignificantly lower in hypothyroid cells (Table I). Inno group was basal activity decreased by the f8-adren-ergic inhibitor propranolol, suggesting that under theseexperimental conditions basal adenylate cyclase activ-ity cannot be attributed to adrenergic effects. (-)-Iso-proterenol by itself caused a significant four- to six-fold stimulation above basal levels in all groups. The(-)-isoproterenol-stimulated activity, in contrast tobasal activity, was completely inhibited by propranololbut was not affected by the a-adrenergic inhibitorphentolamine. Although hypothyroid turkey erythro-cytes showed the same relative stimulation over theirbasal level, the absolute activity of isoproterenol-sen-sitive adenylate cyclase was significantly lower thanin the normal cells. Gpp(NH)p caused a significant

a

D 80 w'40-

oLEE 60 -

40-

20-

-8 -7 -6 -5[(-)-ISOPROTERENOL] (LOGM)

FIGURE 3 Inhibition of [125I]HYP binding by (-)-isopro-terenol. Hyperthyroid (A), normal (0), and hypothyroid (A)turkey erythrocytes were incubated with 40 pM [125I]HYP andincreasing concentrations of (-)-isoproterenol. Specific bind-ing was determined as described in Methods.

stimulation of basal adenylate cyclase and togetherwith isoproterenol was responsible for a synergisticincrease in isoproterenol-responsive adenylate cyclaseactivity. Addition of Gpp(NH)p did not, however, affectthe comparative observations made in the presence ofisoproterenol alone, namely, that hyperthyroid andeuthyroid hemolysates had similar maximal-stimula-table adenylate cyclase activity, whereas hypothyroiderythrocytes had significantly less. The component ofadenylate cyclase activity responsive to fluoride wasalso decreased in the hypothyroid state (Table I). In-creasing concentrations of (-)-isoproterenol were ac-companied by increased adenylate cyclase activity, withhalf-maximal activation occurring at m2-4 ,uM for eachgroup (Fig. 4A). Similarly, half-maximal inhibition ofisoproterenol-responsive activity by (-)-propranolol

TABLE IAdenylate Cyclase Activity in Hemolysates from

Hyperthyroid, Normal, and HypothyroidTurkey Erythrocytes

Hyperthyroid Normal Hypothyroid

Basal 9.9±+1.6 10.5+2.2 4.9+1.6*Gpp(NH)p,

70 MM 36.4±7.9 38.5±5.1 24.5+5.4*Isoproterenol,

0.5 mM 51.9±5.6 45.4±5.7 29.4±5.1*Isoproterenol

+ Gpp(NH)p 136± 19 132± 14 81.7±12.3*Fluoride,8 mM 156+19 152±19 115±22*

Adenylate cyclase activity was determined according toMethods. The values are expressed as picomoles of cyclicAMPgenerated per 109 cell equivalents per 10 min (means± SEM). No significant differences were observed betweennormal and hyperthyroid erythrocyte hemolysates (P > 0.05).* P < 0.01.

Thyroid Status and /3-Adrenergic Responsiveness 187

occurred at the same concentration for each group (Fig.4B). In the hypothyroid turkey erythrocyte the mag-nitude of the decrease in catecholamine-sensitiveadenylate cyclase activity was similar to the degreeby which 8-receptor sites were decreased.

Cyclic AMPcontent. Dose-response relationshipsto /3-adrenergic catecholamines derived from particu-late broken-cell preparations generally exhibit less sen-sitivity than do those examined in intact cells usingeither cyclic AMPgeneration or physiological eventsas markers (24). It therefore was of importance to meas-ure the response of intact erythrocytes to isoproterenolat concentrations of catecholamines comparable to thecirculating levels determined in the course of this study.Accordingly, intracellular cyclic AMPwas determinedafter exposure to very low concentrations of (-)-iso-proterenol (5-50 nM). As shown in Fig. 5, at theselow concentrations of isoproterenol, hyperthyroidturkey erythrocytes-in marked contrast to their be-havior at higher concentrations-accumulated substan-tially more cyclic AMPthan did normal erythrocytes.At concentrations >50 nM, however, the augmentedactivity in the hyperthyroid erythrocytes decreased sothat, at maximal concentrations of isoproterenol, thecyclic AMPconcentration in hyperthyroid and normalerythrocytes was similar. In contrast, in the hypothy-

100 - A B

a0 A' --

0 80d

0

OO

-7 -6 -5 -4 -6 -5 -4(H)-ISOPROTERENOL] (LOGM) L(-)-PROPRANOLOL] (LOGM)

FIGURE 4 Stimulation and inhibition of catecholamine-sen-sitive adenylate cyclase in erythrocytes from hyperthyroid,normal, and hypothyroid turkeys. (A) Adenylate cyclase activ-ity in crude hemolysates prepared from hyperthyroid (A),normal (-), or hypothyroid (A) turkey erythrocytes was meas-ured in response to increasing concentrations of (-)-isopro-terenol. The data are expressed as a percentage of maximalstimulatable adenylate cyclase for each of the groups: hyper-thyroid erythrocytes (44±+7 pmol/109 cell equivalents); normal(38+6) or hypothyroid (26±6). The arrows depict half-maximalstimulation. (B) Hyperthyroid, normal, or hypothyroid eryth-rocyte hemolysates were exposed to (-)-isoproterenol (50,uM) and increasing concentrations of (-)-propranolol. Theresults are expressed as a percentage of maximal activity.Arrows depict half-maximal inhibition. Both panels show theresults of representative experiments performed on four sep-arate occasions. All data points are the means of triplicatedeterminations.

.........

---6-

a./

2500

-J

-8 -7 -6 -5C-I ISOPROTERENOL) (LOG M)

FIGURE 5 Accumulation of cyclic AMPin the erythrocytesof hyperthyroid, normal, and hypothyroid turkeys as a functionof isoproterenol concentration. Erythrocytes were exposed toincreasing concentrations of isoproterenol, and cyclic AMPwas determined as described in Methods. The data representthe means+ 1 SEMfrom three separate experiments in whichthe three groups of turkey erythrocytes were each studied inthe same experiments. Arrows in the gray area of the curverepresent the concentrations of isoproterenol required for half-maximal accumulation of cyclic AMPalong the entire rangeof isoproterenol concentrations employed. The arrows in thewhite area of the curve represent the concentrations of iso-proterenol required for half-maximal accumulation of cyclicAMPwithin the physiological range of catecholamine con-centrations as shown in the white area. Hyperthyroid (A),normal (0), and hypothyroid (A).

roid turkey erythrocyte cellular cyclic AMPwas lowerat both the "physiological" and "pharmacological"ends of the dose-response curve. Conventional deter-mination of the isoproterenol concentration requiredfor half-maximal accumulation of cyclic AMPyieldedvalues of approximately 0.15 ,tM for all three groups(arrows in the gray area of Fig. 5). The similarity ofthese values despite the apparent major differencesin sensitivities to isoproterenol at the lower end ofthe dose-response relationship is the result of the ob-servation that 75% of the ultimate cyclic AMPgen-erated occurs in response to unphysiologically highconcentrations of (-)-isoproterenol (>50 nM). Therelative catecholamine sensitivities of these threegroups of erythrocytes are hence better appreciatedby determining the concentration of isoproterenol re-quired for 25% maximal stimulation, resulting inamounts of cyclic AMPgenerated at the physiologicalend of the dose-response curve. These concentrationsare 0.9, 15, and 32 nM for hyperthyroid, normal, andhypothyroid turkey erythrocytes, respectively (arrows inwhite area of Fig. 5). The hemolysates had negligiblephosphodiesterase activity when measured directly,and similar observations were also made when experi-ments were carried out in the presence of the phos-phodiesterase inhibitor theophylline (data not shown).

J. P. Bilezikian, J. N. Loeb, and D. E. Gammon188

It would thus appear that these differences are not to beaccounted for by differences in cellular phosphodies-terase activity. Propranolol was found to prevent theisoproterenol-sensitive accumulation of cyclic AMPinall three groups with equal potency (Fig. 6).

DISCUSSION

The avian erythrocyte provides an opportunity tostudy the biochemical basis for thyroid hormone-in-duced changes in catecholamine sensitivity affectinga number of well-defined physiological responses to,8-adrenergic agents. Many of the clinical and biochem-ical features of altered thyroid function generally ob-served in human subjects were found to be presentin the turkey, and it is also noteworthy that the chronicnature of this model-as opposed to the acute, moreshort-term studies undertaken in previous investiga-tions-is analogous to what is observed in disorders ofhuman thyroid function.

The erythrocyte of the normal turkey contains a classof binding sites with extremely high affinity for theradiolabeled /8-adrenergic inhibitor [1251]HYP (10). Thestereospecificity, saturability, and reversibility of bind-ing, as well as the order of interaction of these sites with,8-agonists and antagonists all conform to those charac-teristics expected for a /3-adrenergic receptor. Occupa-tion of these /3-receptor sites by agonists and antag-onists is functionally linked to the activation or inhibi-tion of catecholamine-sensitive adenylate cyclase andto subsequent effects on cellular cyclic AMPlevels.

a.4

Q-J:U

4t-J

-JUj

c

E0

0ol

0Ec

3.0 \

A. I

F.'II A

II

-7 -6 -5

t(-) - PROPRANOLOLU(LOG M)

FIGURE 6 Inhibition of isoproterenol-sensitive cyclic AMPaccumulation by (-)-propranolol. Hyperthyroid (A), normal(@), or hypothyroid (A) turkey erythrocytes were exposed to50 ,LM (-)-isoproterenol and increasing concentrations of(-)-propranolol. Cyclic AMPwas determined as describedin Methods. Arrows denote the concentrations of (-)-pro-pranolol required for half-maximal inhibition.

The presence of these characteristics has made it pos-sible to examine the hypothesis that thyroid hormonemay influence the response of the turkey erythrocyteat the /8-receptor itself or at more distal biochemicalsites within the cell.

One possibility is that the observed thyroid hormone-specific effects on the ,8-adrenergic system are attribut-able to changes at the level of the /8-receptor itself,the initial site of catecholamine action. Such an influ-ence has been demonstrated in the acutely thyrotoxicrat heart, in which /3-receptors have been shown tobe increased in number but otherwise indistinguish-able from those identified in the euthyroid rat heart(7, 8). Although changes in ,8-receptor number are some-times associated with concomitant changes in receptor-mediated functions (25-27), it remains to be estab-lished that the increase in /8-receptors observed inthese thyrotoxic models is linked to changes in catechol-amine-related biochemical or physiological properties.This question assumes greater importance in light ofearlier reports suggesting that the catecholamine-sen-sitive adenylate cyclase activities of normal and hyper-thyroid rat hearts are similar (28, 29). Furthermore,the /3-receptor-adenylate cyclase complex has beenfound to be dissociated in some systems with the resultthat,8-receptors do not always reflect changes in recep-tor-mediated cellular properties (30-32).

Whenexamined in other models of hyperthyroidism,an increase in the number of,/3-receptors has not beenconsistently observed. For example, thyrotoxic rat skeletalmuscle and fat cells, as well as human lymphocytes fromthyrotoxic subjects, all contain /3-receptors which appearto be identical in number and affinity to their normalcounterparts (8, 33, 34). Similarly, in the studies re-ported here, the /8-receptor in the turkey erythrocytewas found to be unaffected by the hyperthyroid state.Maximal binding capacity, association and dissociationkinetics, and binding affinity were indistinguishablein normal and hyperthyroid turkeys. Agonists were ableto compete with [125I]HYP for binding sites in erythro-cytes from either normal or toxic birds equally well.There were no Gpp(NH)p-dependent shifts in agonist-specific interactions, nor was there negative or positivecooperativity among the binding sites of hyperthyroidor normal erythrocytes. The number of /3-receptorscould not be reduced by subsensitization after expo-sure to isoproterenol in either group.

Athough detailed studies of adenylate cyclase activ-ity of the turkey erythrocyte demonstrated no signifi-cant differences between thyrotoxic and normal eryth-rocytes, it was necessary to interpret these findingswith regard to the limitations of the adenylate cyclaseassay, which requires broken-cell preparations inwhich hormone sensitivity may be diminished (24).This is particularly pertinent for the turkey erythrocyte,in which the sensitivity of adenylate cyclase to isopro-

Thyroid Status and /3-Adrenergic Responsiveness 189

terenol does not correlate well with the much lowerconcentrations associated with both cyclic AMPaccumu-lation and the stimulation of ion transport in the in-tact cell (12, 35). Because, for this reason, adenylatecyclase activity per se could not be used to answerquestions related to the responsiveness of the turkeyerythrocyte at concentrations of catecholamines withinthe physiological range, cyclic AMPitself was meas-ured in intact cells after exposure to very low levelsof isoproterenol. At concentrations of only 5-50 nMit was found that hyperthyroid turkey erythrocytes infact generated substantially more cyclic AMPthan didnormal cells. Thus, despite no change in number oraffinity of 18-receptors, there was a significant ampli-fication of the adrenergic signal in cells of the hyper-thyroid birds. A similar set of observations has veryrecently been made in the hyperthyroid rat adipocyte,in which it has been shown that catecholamine-in-duced lipolysis is greater than normal despite the ab-sence of changes in 13-receptor number (9). Whetheror not the amplified cyclic AMP responsiveness tocatecholamines in the hyperthyroid turkey erythrocyteis associated with a comparable augmentation ofa phys-iological response, such as catecholamine-dependentsodium transport, is currently under investigation. Inany event, it would appear that one mechanism bywhich thyroid hormone may affect,8-adrenergic eventsis by amplification of the hormonal signal at pointsdistal to the initial cellular recognition site for cate-cholamines.

The capacity of many cells, including the turkeyerythrocyte, to respond to maximal concentrations ofa hormone with maximal cyclic AMPgeneration farexceeds the amount of cyclic AMPrequired to inducea maximal physiological response (12, 24, 35). It istherefore probably of little physiological relevance thatthe difference in cyclic AMPformation between nor-mal and hyperthyroid turkey erythrocytes is lost whenpharmacological levels of isoproterenol are employed.These results underscore the importance of studyingmechanisms of hormone-dependent events at concen-trations of hormone within the physiological range.

In hypothyroidism-as in hyperthyroidism-the re-sults suggest that a change in thyroid hormone levelaffects the actions of catecholamines. In contrast tohyperthyroidism, however, the means by which adren-ergic responsiveness is altered in hypothyroidismappears, at least in part, to involve changes inf1-adren-ergic receptor number. Hypothyroid turkey erythro-cytes demonstrated a coordinate decline in ligand bind-ing, in isoproterenol-dependent adenylate cyclase ac-tivity, and in the generation of cyclic AMP. The degreeto which all of these decreases were observed wassimilar. Although changes in guanyl nucleotide regu-latory sites leading to altered coupling between 13-re-ceptors and adenylate cyclase could account for some

(though not all) of these observations, it is not yetpossible to address this question directly.

Except for the number of specific ,8-receptor sites,binding characteristics of the hypothyroid turkey eryth-rocyte were indistinguishable from those of normaland hyperthyroid cells. It is noteworthy that the in-crease in cyclic AMPat very low concent-rations ofisoproterenol was reduced in hypothyroid cells, andhence that, in the physiological range of catecholamineconcentrations, major differences in cyclic AMP re-sponsiveness were observed between all three groupsof erythrocytes. Three recent studies have reporteddecreased cardiac and skeletal muscle 8-receptors inhypothyroidism, but no further catecholamine-depend-ent parameters were explored (7, 36, 37). Previousinvestigations, antedating the development of 8-recep-tor-specific ligands now available, have, however,shown a decrease in catecholamine-sensitive adenylatecyclase and associated physiological events in hypo-thyroid target tissues (38-40). These results suggestthat one mechanism for decreased adrenergic respon-siveness in hypothyroidism common to many systemsmay be a reduction in the number of 8-receptors. Itis also possible that decreased adrenergic responsive-ness in hypothyroidism occurs by additional mecha-nisms involving modulation of events beyond the p-receptor. In short-term studies of the hypothyroid fatcell, for example, catecholamine-dependent lipolysishas been found to be diminished without any apparentdecrease in 13-receptor number (9).

The point in the avian erythrocyte's history at whichchanges in thyroid hormone level induce the changesin 13-adrenergic responsiveness reported here is notknown. Current evidence indicates that the matureavian erythrocyte, despite the presence of a nucleus,engages in negligible RNAand protein synthesis (41).In this regard the observed changes in 13-adrenergicreceptor characteristics most probably reflect modifi-cations induced during the course of erythropoiesis it-self, or, less likely, subsequent changes in the matureerythrocyte induced under the influence of either in-sufficient or excessive hormone. The latter changes,however, would have to occur by a means independentof transcriptional mechanisms currently believed tomediate most if not all of the known effects of thethyroid hormones. Differences in ,8-adrenergic respon-siveness were observed to be fully established within6 wk of the onset of the hypothyroid or hyperthyroidstate, a time interval comparable to the known life-span of avian erythrocytes (30-45 d [42]) and hencealso consistent with the time necessary for the estab-lishment of a pure population of cells produced underconditions of either insufficient or excessive thyroidhormone.

These studies have not established and indeed donot imply that the marked changes in catecholamine

190 J. P. Bilezikian, J. N. Loeb, and D. E. Gammon

action in hyperthyroid and hypothyroid turkey erythro-cytes are necessarily linked to changes in catecholam-ine-dependent ion transport or to other cyclic AMP-dependent physiological events. Additional studies arerequired to explore this possibility. Nevertheless, itis clear in this experimental model that hyperthyroid-ism is associated with increased levels of catecholam-ine-dependent cyclic AMP, and that in hypothyroidismcatecholamine-dependent cyclic AMP is decreased.The decreased levels of the cyclic nucleotide in hypo-thyroidism appear to be mediated at least in part bya reduction in,t8-adrenergic receptor number, whereasthe increased levels in hyperthyroidism appear to occurby a process distal to the ,8-receptor itself. Furtherstudies of these two mechanisms by which alterationsin thyroid function can modify physiological adrener-gic events may be of relevance to other target tissues,including those involved in disorders of human thyroidfunction as well.

ACKNOWLEDGMENTS

Wethank Dr. Qais Al-Awqati for helpful suggestions.This work was supported in part by grants HL-20859 and

HD-05506 from the National Institutes of Health.

REFERENCES

1. Werner, S. C. 1978. Hyperthyroidism. In The Thyroid.S. C. Werner and S. H. Ingbar, editors. Harper & Row,Publishers, New York. 4th edition. 591-603.

2. Irvine, W. J., and A. D. Toft. 1976. The diagnosis andtreatment of thyrotoxicosis. Clin. Endocrinol. 5: 687-707.

3. Spaulding, S. W., and R. H. Noth. 1975. Thyroid-catecholamine interactions. Med. Clin. N. America. 59:1123-1131.

4. Levey, G. S. 1971 Catecholamine sensitivity, thyroidhormone and the heart. Am. J. Med. 50: 413-420.

5. Jones, J. K., F. Ismail-Beigi, and I. S. Edelman. 1972Rat liver adenyl cyclase activity in various thyroid states.

J. Clin. Invest. 51: 2498-2501.6. Maguire, M. E., E. M. Ross, and A. G. Gilman. 1977.

,3-Adrenergic receptor: Ligand binding properties and theinteraction with adenyl cyclase. Adv. Cyclic NucleotideRes. 8: 1-83.

7. Ciraldi, T., and G. V. Marinetti. 1977. Thyroxine andpropylthiouracil effects in vivo on alpha and beta adren-ergic receptors in the heart. Biochem. Biophys. Res.Commun. 74: 984-991.

8. Williams, L. T., R. J. Lefkowitz, A. M. Watanabe, D. R.Hathaway, and H. R. Besch, Jr. 1977. Thyroid hormoneregulation ofB-adrenergic receptor number.J. Biol Chem.252: 2787-2789.

9. Malban, C. C., F. J. Morano, R. J. Cabelli, and J. N.Fain. 1978. Fat cell adenylate cyclase and ,8-adrenergicreceptors in altered thyroid states. J. Biol. Chem. 253:671-678.

10. Brown, E. M., D. Hauser, F. Troxler, and G. D. Aurbach.1976. ,8-Adrenergic receptor interactions: Characteriza-tion of iodohydroxybenzylpindolol as a specific ligand.

J. Biol. Chem. 251: 1232-1238.11. Bilezikian, J. P. 1976. The interrelationship between 38-

adrenergic receptors and adenylate cyclase activity in the

turkey erythrocyte. In Modern Pharmacology. Gerald S.Levy, editor. Marcel Dekker, Inc., New York. 349-371.

12. Gardner, J. D., H. L. Klaeveman, J. P. Bilezikian, andG. D. Aurbach. 1974. Stimulation of sodium transport inturkey erythrocytes by cyclic 3', 5'-AMP. Endocrinology.95: 499-507.

13. Rudolph, S. A., D. E. Schafer, and P. Greengard. 1977.Effects of cholera enterotoxin on catecholamine-stimu-lated changes in cation fluxes, cell volume, and cyclicAMPlevels in the turkey erythrocyte.J. Biol. Chem. 252:7132-7139.

14. Remington, R. E. 1937. Improved growth in rats oniodine-deficient diets. J. Nutr. 13: 223-233.

15. Salomon, Y., C. Londos, and M. Rodbell. 1974. A highlysensitive adenylate cyclase assay. Anal. Biochem. 58:541-548.

16. Maguire, M. E., R. A. Wiklund, H. J. Anderson, and A. G.Gilman. 1976. Binding of [25I]iodohydroxybenzylpindololto putative B3-adrenergic receptors of rat glioma cells andother cell clones. J. Biol. Chem. 251: 1221-1231.

17. Steiner, A. L., C. W. Parker, and D. M. Kipnis. 1972.Radioimmunoassay for cyclic nucleotides. I. Preparationof antibodies and iodinated cyclic nucleotides. J. Biol.Chem. 247: 1106-1113.

18. Passon, P. G., and J. D. Peuler, 1973. A simplified radio-metric assay for plasma norepinephrine and epinephrine.Anal. Biochem. 51: 618-631.

19. Hollander, C. S., and L. Shenkman. 1974. Radioimmuno-assays triiodothyronine and thyroxine. In Nuclear Medi-cine In Vitro. B. Rothfeld, editor. J. B. Lippincott Co.,Philadelphia. 136-149.

20. Rosalki, S. B., 1967. An improved procedure for serumcreatine phosphokinase determination.J. Lab. Clin. Med.69: 696-705.

21. Ringer, R. K., 1976. Thyroids. In Avian Physiology. P. D.Sturkie, editor. Springer-Verlag NewYork Inc., NewYork.348-358.

22. Stoffer, S. S., N. Jiang, C. A. Gorman, and G. M. Pikler.1973. Plasma catecholamines in hypothyroidism andhyperthyroidism. J. Clin. Endocrinol. Metab. 36: 587-589.

23. Scatchard, G. 1949. The attractions of proteins for smallmolecules and ions. Ann. N. Y. Acad. Sci. 51: 660-672.

24. Robison, G. A., R. W. Butcher, and E. W. Sutherland.1968. Cyclic AMP. Ann. Rev. Biochem. 31: 149-173.

25. Kebabian, J. W., M. Zatz, J. A. Romero, and J. Axelrod.1975. Rapid changes in rat pineal /3-adrenergic receptor:Alterations in 1-[3H] alprenolol binding and adenylatecyclase. Proc. Natl. Acad. Aci. U. S. A. 72: 3734-3739.

26. Wolfe, B. B., T. K. Harden, and P. B. Molinoff. 1976.,3-Adrenergic receptors in rat liver: Effects of adrenalec-tomy. Proc. Natl. Acad. Sci. U. S. A. 73: 1343-1347.

27. Mickey, J., R. Tate, and R. J. Lefkowitz. 1975. Subsen-sitivity of adenylate cyclase and decreased ,8-adrenergicreceptor binding after chronic exposure to (-)-isoprotere-nol in vitro. J. Biol. Chem. 250: 5727-5729.

28. Levey, G. S., C. L. Skelton, and S. E. Epstein. 1969. In-fluence of hyperthyroidism on the effect of norepine-phrine on myocardial adenyl cyclase activity and con-tractile state. Endocrinology. 85: 1004-1009.

29. McNeil, J. H., L. D. Muschek, and T. M. Brody. 1969.The effect of triiodothyronine on cyclic AMP, phosphory-lase, and adenyl cyclase in the rat heart. Can. J. Physiol.Pharmacol. 47: 913-916.

30. Maguire, M. E., L. L. Brunton, R. A. Wiklund, H. J.Anderson, P. M. VanArsdale, and A. G. Gilman. 1976.Hormone receptors and the control of cyclic AMPmetab-

Thyroid Status and /3-Adrenergic Responsiveness 191

olism in parental and hybrid somatic cells. Recent Prog.Horm. Res. 33: 633-667.

31. Charness, M. E., D.. B. Bylund, B. S. Beckman, M. D.Hollenberg, and S. H. Snyder. 1976. Independent varia-tion of 3-adrenergic receptor binding and catecholamine-stimulated adenylate cyclase activity in rat erythrocytes.Life Sci. 19: 243-250.

32. Bilezikian, J. P., A. M. Spiegel, E. M. Brown, and G. D.Aurbach. 1977. Identification and persistence of beta-ad-renergic receptors during maturation of the rat reticulo-cyte. Mol. Pharmacol. 13: 775-785.

33. Williams, R. W., C. E. Guthrow, and R. J. Lefkowitz.1978. Beta-adreneigic and prostaglandin receptors inlymphocytes from hyperthyroid patients. Clin. Res. 26:315A. (Abstr.)

34. Virenda, S. K., and S. P. Banerjee. 1978. /3-Adrenergicreceptors in rat skeletal muscle: Effects of thyroidectomy.Biochim. Biophys. Acta. 539: 538-542.

35. Gardner, J. D., H. L. Klaeveman, J. P. Bilezikian, andG. D. Aurbach. 1973. Effect of 3-adrenergic catechola-mines upon sodium transport in the turkey erythrocyte.

J. Biol. Chem. 248: 5590-5596.

36. Banerjee, S. P., and L. S. Lung. 1977. /3-Adrenergic re-ceptors in rat -heart: Effects of thyroidectomy. Eur. J.Pharmacol. 43: 207-208.

37. Smith, R. M., W. S. Osbornewhile, and R. A. King. 1978.Changes in the sarcolemma of the hypothyroid heart.Biochem. Biophys. Res. Commun. 80: 715-721.

38. Krishna, G., A. Hynie, and B. B. Brodie. 1968. Effectsof thyroid hormones on adenyl cyclase in adipose tissuesand on free fatty acid mobilization. Proc. Natl. Acad. Sci.U. S. A. 59: 884-889.

39. Levey, G. S., C. L. Skelton, and S. E. Epstein. 1969.Decreased myocardial adenyl cyclase activity in hypo-thyroidism.J. Clin. Invest. 48: 2244-2250.

40. Rosenqvist, U. 1972. Noradrenaline-induced lipolysis insubcutaneous adipose tissue from hypothyroid subjects.Acta Med. Scand. 192: 361-369.

41. Williams, A. F. 1972. DNAsynthesis in purified popula-tions of avian erythroid cells. J. Cell. Sci. 10: 27-43.

42. Sturkie, P. D., and P. Griminger. 1976. Blood: Physicalcharacteristics, formed elements, hemoglobin, and coag-ulation. In Avian Physiology. P. D. Sturkie, editor.Springer-Verlag New York, Inc. New York. 53-75.

192 J. P. Bilezikian, J. N. Loeb, and D. E. Gammon