Thyroid hormone modulation of the hypothalamic growth hormone (GH)-releasing factor-pituitary GH axis in the rat.
N Miki, … , T Aoki, H Demura
J Clin Invest. 1992;90(1):113-120. https://doi.org/10.1172/JCI115823.
Both thyroid hormone and hypothalamic growth hormone (GH)-releasing factor (GRF) facilitate pituitary somatotroph function. However, the pathophysiological role of thyroid hormone in GRF secretion is less well understood.
Thyrotoxicosis, induced by administration of thyroxine (T4) in rats, inhibited both pituitary GH levels and immunoreactive GRF secretion from incubated hypothalamus. At the highest dose of T4 given for 12 d, GRF secretion and pituitary GH decreased by 50 and 39%, respectively. Hypothyroidism induced by thyroidectomy (Tx) enhanced GRF secretion approximately twofold while depleting pituitary GH by greater than 99%. Both of these hypothalamic and pituitary effects were reversed by replacement of T4 but not human GH for 7 or 14 d. Human GH was as potent as T4 in restoring decreased body weight gains or serum insulin-like growth factor-1 levels in Tx rats. These results indicate that at both physiological and pathological concentrations in serum, thyroid hormone acts as an inhibitory modulator of GRF
secretion, probably not involving a feedback mechanism through GH. A biphasic effect of thyroid hormone on pituitary GH levels appears to derive from the difference in primary target tissues of hyper- and hypothyroidism, the hypothalamus and the pituitary, respectively.
Research Article
Find the latest version:
https://jci.me/115823/pdf
Thyroid Hormone Modulation of the Hypothalamic Growth Hormone
(GH)-releasing Factor-Pituitary GH Axis in the Rat
Nobuhiro Miki,MasamiOno,NaomiHizuka, Tomoko Aoki, and HiroshiDemura
DepartmentofMedicine, InstituteofClinical Endocrinology, Tokyo Women'sMedicalCollege, Tokyo 162,Japan
Abstract
Both thyroid hormone and hypothalamic growth hormone (GH)-releasingfactor (GRF) facilitate pituitary somatotroph function. However, the pathophysiologicalrole ofthyroidhor- mone inGRFsecretion isless wellunderstood. Thyrotoxicosis, induced byadministrationof thyroxine (T4) in rats,inhibited both pituitary GH levelsand immunoreactive GRF secretion from incubated hypothalamus.At the highest dose ofT4given for 12 d,GRF secretion and pituitary GH decreasedby50and 39%, respectively. Hypothyroidism induced bythyroidectomy
(T.)
enhanced GRF secretion approximately twofold whilede- pleting pituitary GH by >99%. Both, ofthese hypothalamic andpituitary effectswerereversed by replacement of T4but not human GH for 7or 14 d.HumanGHwas as potent as T4in restoring decreased body weight gains or serum insulin-like growthfactor-i
levels inT.
rats.These results indicate thatat both physiological and pathological concentrationsin serum, thyroidhormone acts as aninhibitory modulator of GRFsecre- tion, probably not involving a feedback mechanism through GH. A biphasic effect of thyroid hormone on pituitary GH levels appearstoderive from the difference in primarytarget tissues of hyper- and hypothyroidism, the hypothalamus and thepituitary,respectively. (J.Clin.Invest.1992.90:113-120.) Key words: hyperthyroidism-hypothyroidism*somatotropin- somatotropin-releasingfactor *insulin-like growth factor-I IntroductionSecretionofpituitarygrowth hormone
(GH)'
ispositivelyregu- latedby thehypothalamic peptide, GH-releasing factor(GRF) ( 1-3). GRF generates spontaneous burstsofGHsecretion(4- 6)and also increases GHgenetranscription andGHmRNA levels(7, 8).Somatostatin (SRIF)is the hypothalamicpeptidePortionsofthis work were presented in abstract form at the1stInterna- tionalCongressofNeuroendocrinology,San Francisco, CA, 9-1 1 July 1986, andatthe 8th International Congress of Endocrinology, Kyoto, Japan, 17-23 July 1988.
Addressreprintrequests to N. Miki, Department ofMedicine, Insti- tuteofClinicalEndocrinology,Tokyo Women's Medical College, 8-1 Kawada-cho,Shinjuku-ku, Tokyo 162, Japan.
Receivedfor publication 23 September 1991 and in finalform 31 January1992.
1.Abbreviations used in this paper: BW, body weight; GH, growth hormone; GRF, GH-releasing factor; hGH, human GH; HKRB, Hepes-bufferedKrebs-Ringer bicarbonate solution;IGF-I,insulin-like growthfactor-I;SRIF, somatostatin;T3,triiodothyronine; T4, thyrox- ine;TX,thyroidectomyorthyroidectomized.
thatinhibitsGHsecretion(9, 10) andmaybe also involved in theregulation ofGRFsecretion withinthehypothalamus (11, 12). Although GRF and SRIF playaprimaryrole in GHsecre- tion, thereareotherimportanthormones that modulatepitu- itaryGHsecretion, probably througheffectsontheGH-regula- torypeptides,GRFand/orSRIF. These includethyroidhor- mone, adrenal glucocorticoid, and gonadal steroid of peripheral endocrine organs (13). The most important for growth and development of mammals is thyroid hormone.
Thyroid hormone stimulates synthesis andsecretion ofGH (14-16) and also increases GH gene transcription and GH mRNAlevels (17-19). HypothyroidismdepletespituitaryGH ( 14-16), inhibits GH secretion (20-22), and causes severe growthretardation(23).
Thus, boththyroidhormone and GRF directly stimulate pituitary somatotroph function,andtheirhypophysealactions aresimilarinmany aspects.However, the
pathophysiological
roleofthyroidhormone inhypothalamic
GRFsecretionis less well understood. It appears more complex because GRF re- ceivesanegative feedback signal fromGH(24-27),
sothyroid
hormone might influence GRF secretion indirectly through theeffectonpituitaryGH. Furthermore,thyroid
hormone has a biphasic effect on growth rate. In immature animals and youngpatients, growth is retarded byhypothyroidism,
isre- storedbyreplacementdosesofthyroxine
(T4),and is inhibited bythyrotoxic dosesof T4 (28).Weand otherspreviously
re- ported thatthyroidectomy(T,)
enhanced invitro secretion of hypothalamic GRF in rats (26, 29). Whether this effect is caused by thyroid hormone deficiency per se or mediated through associatedGHdeficiency remainsunknown. Alsoun- known is theeffect of hyperthyroidism onGRFsecretion.Inthepresentstudiesinrats, wehavecompared effects of thyroidhormone excess and deficiency in vivo on
hypotha-
lamicGRFsecretioninvitro. Toexaminethemechanism of action ofhypothyroidism,
we haveinvestigated
whether theTX-induced
rise in GRFsecretion is reversibleby replacement of T4or human GH(hGH). We have also studied effectof hGH in combination with T4, because GHgivenwithT4pro- motesthegrowthofthyroid-deficient
animalsmorethan GH alone(23, 30).Inexperiments with hGH,serumlevelsofinsu- lin-like growth factor-I(IGF-I)
aswell asbodyweightgains
weremeasured as anindex ofabiologicaleffectof hGH. Fi- nally,wehave characterizedGRF secreted in the mediumto examine whetheranabnormalthyroidstatusaltersamolecular form ofGRF.
Methods
Chemicals
SodiumL-thyroxine (T4)wasobtained fromSigmaChemical Co.(St.
Louis, MO). Recombinant hGHwasprovidedby Sumitomo Pharma- ceutical Co. Ltd. (Osaka, Japan). Acetonitrile (CH3CN)and2-pro- panolwereobtained from KantoChemical Co. Ltd.(Tokyo, Japan) ThyroidHormoneand GrowthHormone-releasingFactor 113 J. Clin. Invest.
©TheAmericanSociety for Clinical Investigation, Inc.
0021-9738/92/07/0113/08 $2.00 Volume 90, July 1992, 113-120
and trifluoroacetic acid(TFA)from Wako Pure ChemicalIndustries Ltd. (Osaka,Japan).SyntheticratGRF waspurchasedfrom Peninsula Laboratories,Inc. (Belmont, CA).
Animals and experimental protocols
MaleSprague-Dawleyrats wereobtainedfromJapan SLCInc. (Hama- matsu City, Shizouka, Japan)at5 wkofage.Theywerehoused in a room undercontrolledtemperature(22±10C), humidity (50-55%) andlight/darkconditions(lightson0800-2000h). Foodand water were availablead libitum.
Hyperthyroidismwasinduced at 10-1 1 wk of ageby subcutane- ously injecting T4oncedailyfor6 or 12 d. Animalswereweighed daily at1700-1730 h, and T4 wasgivenon adailyweight basisin dosesof25, 50,or100,g/100 gbodyweight (BW).Anintermediate50,ug/100 g BW doseofT4 was the dose thatinhibitedBWgainbutdidnot reduce pretreatment BW,accordingtotheresultsofapreliminaryexperiment.
T4 wasdissolvedin50%propyleneglycolanddiluted four timeswith 0.0033 N NaOH beforesubcutaneousadministration.Controlratsre- ceived vehicle injection. The animalswerekilled by decapitation20 h afterthe lastinjection. Hypothalamic tissueswereobtained forsubse- quentincubation experimentsand anteriorpituitaries forGH measure- ment.The trunk blood was alsocollectedinexperimentsinwhichT4 or vehiclewasadministered for 12 d.
Hypothyroidismwas producedbysurgicalthyroidectomy
(T.)
at6wkofage. BecauseparathyroidectomyattenuatespulsatileGH secre- tion in rats(31),atleastthreeparathyroid glandswereleft,and cal- ciumwas notsupplementedindrinkingwater.From 4 wkafteropera- tion,
T,
rats werereplacedoncedaily witha"physiological"doseofT4 (1 or 2,g/100gBW)ortreatedtwicedailywith hGH (100ig/rat)for 7 d. hGH wasdissolved in physiological saline. Inthe next experi- ments,T.
ratsreceived thesame treatmentsfor14d.Finally,weinves- tigatedwhether hGH could affectGRF secretionsynergisticallywith thyroidhormone.TXrats weresupplementedwithT4 alone (2,g/100 gBW, oncedaily)orin combination withhGH(100ag/rat,twice daily) for 7 d.T4wasinjectedsubcutaneouslyat 1730-1800 h and hGH at0830-0900hand 1730-1800 h.ControlTXandsham-operated rats were treatedwith vehiclesfor T4, hGH,orboth. Theanimalswere decapitated 18 hafterthe lastinjection;and thehypothalamus,ante- riorpituitary,and trunk blood were collected.Hypothalamic
incubationThemethodof hypothalamic incubationhas beendescribed elsewhere (24).Inbrief,thebrainwasobtainedafter cuttingthepituitarystalks carefully,and thehypothalamiwere cutsagittally throughthemedian eminence withthe ventralsurfaceup. The areabetweentheposterior borderof theopticchiasma, theanteriorborder of themammillary body,the lateralhypothalamic sulcus,and 2 mmdeep fromthe ventral surfacewasdissected intotwofragments.Thefragmentswereplacedin polyethylene vials containing0.6 ml of 11 mMHepes-bufferedKrebs- Ringerbicarbonate solution (HKRB),pH7.40,which waspregassed with95%02-5%CO2and supplementedwith5.5 mM glucose,0.1%
BSA,and500 mg/mlbacitracin.Fourhypothalamicfragments from two rats wereincubatedin one vial. Allincubationswere 30 min at 37°Cunder an atmosphere of 95%02-5%CO2in a shaking water bath (60cycles/min).Afterpreincubation,the tissueswere incubatedin HKRB (basal secretion) and then in 50 mMK+-containingHKRB (K+-stimulated secretion). The medium was collected every 30min and replaced with fresh buffer. The collected medium wascentrifuged at 3,000 rpm for 5 min at4°C, and thesupernatants were quickly frozen on dry ice and kept at-20°Cuntilassayed.Eachhypothalamic incubation used 18-24 animals, and asecretionexperiment consisted of two to three separateincubations.
Tissue
extraction
Immediately afterthehypothalamic incubations,fourhypothalamic fragmentswerepooled andGRF in the tissueswasextractedasde- scribed(24).Thefragmentswereboiled for7min andhomogenized by
polytroninIN acetic acid and 0.02 N HCIcontaining10 mMEDTA, 1 gg/mlpepstatin,and 0.01%0-mercaptoethanol.Thehomogenates
werecentrifugedat 10,000g for 30 minat4VC,and thesupernatants
were frozen in adryice-methanol bath, lyophilized, and stored at
-20'Cuntilassayed.The anteriorpituitarieswerehomogenized by polytronin 0.05 MNaHCO3-Na2C03buffer, pH 9.96. Thehomoge-
natesweresupplementedwithanequalvolume of0.01 M PBS contain- ing 1% BSA, 0.1% TritonX-100, and 25 mM EDTA and storedat
-20'CuntilassayedforratGH.
Reverse-phase
HPLCA total of 10-12hypothalamiofhyper-orhypothyroidratsand their controlswereincubated inadepolarizing (50mMK+)condition for twosuccessive30-min periods.The medium(6.0-7.2 ml)waspooled, acidifiedbythe addition of 0.3 M TFAto afinal concentration of 0.06 M,and extractedondisposableSep-Pak C18 cartridges (Waters Chro- matographic Division, Millipore Corp.,Bedford, MA)asdescribedby Frohman and Downs(32).The cartridgeswerepreactivated bysucces- sive4-mlwashes of 0.01 M TFA, 80% (vol/vol) CH3CN/20%(vol/
vol)0.01 MTFA,and0.01MTFA. After the samples were applied, the cartridgeswerewashed with 4 ml of 0.01 MTFA,and GRFwaseluted with2 ml of 80%(vol/vol)CH3CN/20% (vol/vol) 0.01 M TFA. The eluantwasfreeze-dried and reconstituted inatwentieth theoriginal medium volume of0.01 M TFA for HPLC analysis. The HPLC system (WatersChromatographic Division,MilliporeCorp.) consisted of600 E multisolvent delivery system, U6K universal liquidchromatography injector,and programmable multiwavelength detector.Reverse-phase HPLC columnwasVydac218TP54(C 18,0.46X25 cm, 5-,Amparticle size) (The Separationgroup,Hesperia, CA). Solvent A was 0.1%(vol/
vol) TFA and solvent B was 80% (vol/vol) CH3CN/20%(vol/vol) 2-propanol/0. 1% (vol/vol) TFA. The samples were eluted with a lin- eargradient of 25-45% (vol/vol) solvent B in 32 min and then 45-80%
(vol/vol) solvent B in 28 min. Flow rate was 1 ml/mmin. -ml fractions werecollected, lyophilized, and assayed for GRF. Blankinjections
werecarriedoutbefore each sample injection, and fractions collected didnotcontain any detectable GRF-like immunoreactivity. The final recovery ofsyntheticGRFwas- 64%.
RIAsforratGRF, GH, T4, and IGF-I
GRF in the medium and tissue samples was determined by RIAas described elsewhere (24). The minimum detectable dose(BO-2SD) was 2pg/tube, with half-maximaldisplacementof22-25 pg/tube. The nonspecific binding was < 2%. The samples were measured directly without extraction or concentration. In assays ofthe medium samples, HKRB buffer was added to all standard curves in a volume of 100-200
tL equivalent to that ofthe unknown. The total incubation volumewas 500 ,u.PituitaryGHcontents weremeasured induplicate byaspecific ratGH RIA (5, 12,24) at two dilutions using materials provided by the Rat Pituitary Hormone DistributionProgram, National Institute of Diabetes and Digestive andKidneyDiseases(NIDDK)National Insti- tutesof Health. Results wereexpressedintermsof NIDDKratGH RP-2reference standard. Serum T4 levels were determined byadouble antibody method with a RIA kit purchased from Eiken ICL(Tokyo, Japan). The sensitivity of the T4 assay was 1Ag/dl.Serum IGF-I levels weremeasuredbyRIAasdescribed(33),using biosynthetichuman IGF-Iasstandards andanantiserum notcross-reactive with IGF-II.
Serum samples were extracted by the aceton-formic acid method of Bowsher et al. (34). The least detectable level of IGF-I assaywas 10 pg/tube.
Data analysis
Resultswereexpressedasmeans±SE.Analysisofvariancewasused for statisticalanalysis. RelationshipbetweenserumT4and GRF secretion inhyperthyroidrats wasanalysedusinglinearregression analysis.Sub- sequent toanalysis ofvariance,thesignificanceof differences between experimental groups was determined by theStudent-Newman-Keuls test.AP value of < 0.05wasconsideredstatisticallysignificant.
? 50
40
~
030
dered
6
a3 1
0 50 100 0 25 50 100
14
Dose(pg/b10g BW/Day)
Figure1. GRFsecretionfrom hypothalamic fragmentsofratsren-
deredhyperthyroid byadministration of 14 or6(left panel)or12 d (rightpanel). Closed and open columns indicate basal and 50 mM K+-stimulatedGRF release,respectively.Twohypothalamiwere in- cubated in onevial,and the secretion of GRF wasexpressedaspico- grams (means + SE) perhypothalamus/30 min. The numbers of ob- servationswereshownonthe top of closed columns. *P<0.05, **P
<0.01,***P<0.001vs. control indicated by a zero doseofT4.
Results
Effects ofhyperthyroidism
onhypothalamic
GRF, pituitaryGH, and
serum T4Fig. 1 shows the results of secretion experiments in hyperthy- roidrats. Asshown intheleft panel, administration ofT4for6 d hadnoeffecton basalGRFsecretion (closed columns)but significantly inhibited K+-evoked GRF secretion (open col- umns). Theextent of GRF suppressionwas notdifferentbe- tween two doses ofT4: -28.9% (P < 0.05) by 50
qg
and-31.3% (P<0.01) by 100
jig
T4/ 100g perd.Administration ofT4 for 12dinhibitedGRFsecretion (right panel) and in- creasedserum T4 levels(Table I).Botheffectsweredosere- lated, and theinhibitory effectonGRF secretionwasgreater than that observed after 6 d of T4 administration. Basal GRF secretion was inhibited below the assay detection limit (P<0.01)by both of the 50 and 100 ug/ 100gdoses.K+-evoked GRF secretion was suppressed by 10.3% (NS), 27.3% (P
<0.0l),and5o.1%(P<0.00l)bythe25,50,and
l0,g/100
gdosesofT4,respectively. Linear regression analysisshoweda
significant negative relationship(r= -0.77,P< 0.0001) be- tween serumT4 levels and K+-evoked GRFsecretion (datanot illustrated).
The T4-induced suppression ofGRFsecretion after 12 d wasassociatedwith aparallel dose-related decline in pituitary GH(TableI).Thedeclinein GHwassignificantat aninterme- diate doseofT4(- 18.5%,P<0.01), and the minimum GH levelatthe highest dosewas60.7% (P<0.001) ofthecontrol value. T4 at thehighest dose also reducedpituitaryGHafter6 d, butto alesserextent(-10.6%, P< 0.05)than after 12 d (TableI). HypothalamicGRFcontents werediminishedonly afteradministration ofthe highestdosefor 12 d (-16.4%, P
<0.05)(Table I).
AtthebeginningofT4 treatmentfor 12 d,mean(±SE)BW of allanimalswas378±5g(n=90), and those aftertreatment were412±4 (n = 26), 399±6(n = 14), 366±7 (n = 26),or 329±6g(n =24) (P<0.001 vs.initial weight)inratsgiven0 (vehicle), 25, 50,or
100,ug
T4/100gperd,respectively. After administration ofT4 for 6 d, theinitial BW wassignificantly decreased alsoatthehighestdoseonly(-7.6%,P<0.01).Effects of hypothyroidism and
T4 orhGH replacement
onhypothalamic
GRF secretion in vitroAs shown in Fig. 2,
Tx
significantly enhanced both basal (closed columns)andK+-stimulatedGRFsecretion(opencol- umns)at5 wkafteroperation. ReplacementofT4for7dat2 ,ug/100 g per d partially reversed theT,-induced
rise inK+-stimulatedGRFsecretion(-53%,P<0.01), whereas the enhanced basal GRF secretionwas reversed (P < 0.01) to a level statistically indistinguishable from that of the
sham-Tx
rats(leftpanel). However, hGHtreatmentfor7dat200
jig/
ratperd hadnoinhibitory effectontheenhanced GRFsecretion either basallyorunderK+ stimulation in
Tx
rats(rightpanel).Tx
ratsweighed
30% less than shamTx
controlsat4wk afterTableI. Serum T4, PituitaryGH,and HypothalamicGRFLevelsinHyperthyroidRats
Number
Treatment Dose Length of rats SerumT4 Pituitary GH Hypothalamic GRF
Alg days sg/dl ug/gland pg/tissue
Vehicle 6 12 ND 459±15 858±53
T4 50 6 12 ND 459±23 817±47
T4 100 6 12 ND 410±22* 902±23
Vehicle 12 26 5.4±0.1 453±22 719±23
T4 25 12 14 12.7±0.5§ 408±20 775±41
T4 50 12 26 15.4±0.8§ 369±16t 781±32
T4 100 12 24 18.2±0.6§ 275±12§ 601±27*
Results are means + SE. ND, notdetermined. DosesofT4 are expressed asqg/100g bodyweight/day. Tissue GRF levels were measured by poolingtwohypothalamiandareexpressedaspicogramsper onehypothalamus. *P<0.05, *P <0.01, §P<0.001 vs. vehicle.
ThyroidHormoneand GrowthHormone-releasingFactor 115
1001
80
E
~60
-c
C
40
0
1120 6 6
U
66
Sham Tx Tx Sham Tx Tx
Tx T4 Tx GH
Figure 2. Effects ofthyroidectomy (Tx)and replacement of T4(left panel)orhGH(rightpanel)for7d onGRFsecretionfromincubated hypothalamicfragments.Tx rats were treatedoncedailywith T4 ( 2 yg/100g)ortwicedailywith hGH(100jig/rat)beginningat 4 wk afterTx.Closed and open columnsindicatebasal and 50 mM K+- evokedsecretion,respectively. The results were expressed aspico- grams(means +SE)ofGRF secretedfromonehypothalamusper30 min.The numbersof observationswereshown on the topofclosed columns. *P<0.05,** P<0.001vs.sham
T,.
tP<0.01 vs. Tx.operation. hGH increased theirBWsat a rate comparable to thatofT4. During7dof treatments, the mean (±SE) weight gainof sham
Tx, T.,
orTx
+T4 groups was29.2+2.3,
4.7+2.5,or 22.5±1.4 g/rat, and that ofsham
Tx,
Tx, orTx
+ hGH groups was 33.5±3.1, 4.2±1.5, or 25.2±2.7 g/rat (n = 12 each), respectively.Tx
rats were thenreplaced with T4 or hGHfor14d at the same dailydose (Fig. 3). T4 reversed theTx-induced
risein K+-stimulated GRFsecretion by 76% (P<0.01). Althoughv Sham Tx
Tx
Figure 3. Effects of replace- mentof T4orhGH for14 donGRFsecretion inTX
rats.T4wasgivenonce dailyat2,ug/100 g and hGHtwicedailyat100 ,ug/
** ratbeginningat4wkafter TX.ShamTXand controlTX
rats weretreated with vehi-
*+ clesfor both T4 and hGH.
Closedand open columns indicate basal and 50 mM K+-evoked GRFsecretion, 6 respectively. The secretion
of GRF was expressed as
Tx - ** picograms(means+ SE)
Tx -
_per
hypothalamus/30min.
Tx Tx _ *P<0.01, **P<0.001vs.
T4 GH shamTx.tP<0.01vs.Tx.
notcomplete, this reversibility was greater than that observed after 7 d (Fig. 2). T4 suppressed the enhanced basal GRFsecre- tion completely (P < 0.01 ), as it did after 7 d. At a half daily dose of 1
jtg/
100 g, T4 replaced for 14 dwas also effective, though to a lesser extent, in inhibiting the enhanced K+- evoked GRF secretion (data not shown). In contrast, hGH did not significantly reduce either basal or K+-stimulated GRF secretion inTX
rats. However, hGH restored their diminished weight gains to normal. During 14 d of treatments, themean (±SE) weight gain of shamTx,
T1,Tx
+ T4, orTx
+ hGH groups was48.8±5.0, 5.3±1.2, 49.4±2.8, or 51.1±2.8 g/rat (n= 12-16), respectively.
The thirdexperiment with hGH was carried out to study the effect ofcombined hGH and T4 replacement in
Tx
rats (Fig.4).The doses of T4 and hGH were the same with those shown inFigs. 2 and 3, and the length of treatments was 7 d. T4 alone partially reversed the enhancedK+-evokedGRF secretion (P
<0.01)whilenormalizingtheincreased basal secretion. hGH incombination withT4similarlyreversed theenhanced K+- evoked GRF secretion significantly (P<0.01), but the level reversedby the combined hGH and T4 therapy was statistically not different from that attained by T4 alone. hGH and T4 re- placedtogether increased the mean (±SE) weight gainof
Tx
ratsfrom 3.0±2.1 to 33.4±0.9 g/rat during 7 d of treatment, which was significantly greater than that of shamTx
rats(26.7±3. 1,P<0.05)
orT4-treatedTx
rats(21.6±3.5,P<0.01) (n = 12 each).Serum T4, serum
IGF-I,
pituitaryGH,
andhypothalamic
GRF levels inhypothyroid
ratsand effects of
T4 orhGH replacement
Asshown in Table II,
Tx
reduced serum T4to undetectable levels(< 1 ,g/dl) (P<0.001),serumIGF-Iby 36.5-38.6% (P<0.001),andhypothalamicGRF contentsby25.7-34.3% (P
<0.05-0.01). Replacement ofT4for7and 14d restored the decreased serum T4 and hypothalamicGRF to levels higher than those in sham
Tx
rats, whereas hGH treatment waswith- outeffectoneither parameter. The decreasedserumIGF-I lev- elsweresignificantlyrestoredbyT4 (P<0.01 vs.Tj)
to74.3and 85.6% of the value of sham
Tx
animals after7and 14d, respectively. hGH treatmentfor7dhad noeffectontheTx-in-
100 Figure 4. Effects of replace-
.' mentofhGH in combina-
E ** tion with T4 for 7
d'on
GRFsecretioninTXrats.
8 The dosesof T4and hGH
E
werethesamewith those
° 60 **+ shown in Figs. 2 and 3: T4,
QT1 *t 2
12g/
100g perd inasingleT1 dose and hGH, 200 ,g/rat
I--, perd intwodivided doses.
.0 ShamTXandcontrolTX
H animals received vehicles
°)
20 ii 6 ~ for bothT4and hGH.1 * 6 6 Treatments started at 4 wk
0ix
*L-
6 - *_-+ after T.. Closed and opencolumnsindicate basal andSham Tx Tx Tx 50 mMK+-evoked GRF
TX T4 T4 +GH secretion, respectively.
Thesecretion of GRFwas expressedaspicograms (means+SE) per hypothalamus/30 min.
*P<0.01, **P<0.001 vs.shamTx.tp<0.01 vs.Tx.
116 N.Miki,M.Ono,N.Hizuka, T.Aoki,and H.Demura
- 80
._CE
1-U)
E 60
s
-C
z40
g 20
u)0
vx
lL**
ii
Table II. Serum T4, SerumIGF-I,andHypothalamicGRF Levels in Thyroidectomized(Tx)Rats Treated with T4 orhGH
Number
Group Length ofrats Serum T4 Serum IGF-I Hypothalamic GRF
days ug/dl ng/ml pg/tissue
ShamTX 7 12 5.9±0.2 441±6 803±35
TX 7 12 <1.0 280±8§ 596±58*
TX+T4 7 12 8.6±0.7*1 328±7*1 1068±481
ShamTX 7 12 5.6±0.1 391±14 726±65
TX 7 12 <1.0 243±7§ 501±24*
TX+hGH 7 12 <1.0§ 261±8§ 543±52*
ShamTX 14 16 6.1±0.1 396±14 1023±68
TX 14 12 <1.0§ 243±9§
672±740
TX+T4 14 12 8.9±1.211 339±15*' 1359±37*1
TX+hGH 14 12
<1.0§
329±8$' 734±85*Resultsarethemeans± SE.Tissue GRF levelsweremeasuredby poolingtwohypothalami andareexpressedaspicograms peronehypothala- mus. *P<0.05, tP<0.01, §P<0.00Ivs.shamTX. IP<0.01, 'P<0.001 vs.TX.
duceddecrease inserumIGF-Ilevels. However,hGH given for 14dincreasedit significantly (P < 0.001 vs.
TX)
to 83.0% ofthe euthyroid value, whichwas comparable to that in hypothyroid rats treatedwithT4 forthe same duration.TX
depletedpituitary GHcontents to < 1% ofthat of euthy- roidrats(P<0.001)at 5-6 wkafter operation,and the reduced pituitaryGH wassignificantly(P <0.001)reversed by T4 (Ta- ble III). Thisreversibilitywas greater byreplacement of T4 at a higher dose (2vs. 1 ,ug/100 g per d) when the length of treat- mentwas same, or by replacementfor a longer duration ( 14 vs.7d) when the doseofT4 was same.Withthereplacement of T4 at 2,ug/100g perd for 14 d, the decreasedpituitary GH was restored tothemaximumlevelof 82% of that of sham
TX
rats, although this level remained significantly lower (P < 0.001) than theeuthyroidvalue.Characterization ofhypothalamic
GRF inthe medium
Fig.5showsreverse-phase HPLCelution profiles ofimmunore- activeGRFinhyper- and hypothyroid animals and theircon- TableIII. Effectsof Thyroidectomy(TxJ
and Replacementof T4 orhGHonPituitary GH ContentsNumber
Group Treatment Dose Length of rats Pituitary GH
Ag days Mg/gland
ShamTX Vehicle 7, 14 80 393±5 (100)
TX Vehicle 7, 14 64 2.1±0.1(0.5)
Tx T4 1 7 14 142±16*(36)
Tx T4 2 7 12 183±11*(46)
Tx T4 1 14 14 243±12*(61)
Tx T4 2 14 12 324±10*(82)
Tx hGH 200 7 12 1.7±0.4(0.4)
Tx hGH 200 14 12 2.8±0.6(0.7)
Results are means±SE; values in parentheses arepercentages. Pitu- itaryGHcontentsin shamTxorTxrats treatedwith vehicle for 7 or
14dayswerecombined. *P <0.001 vs. shamTx+vehicle orTx+ vehicle.
trols, which wassecreted in vitro inresponse to 50 mMK+.
Rats weregivenvehicleorT4at100,ug/ 100g perdfor 6 d(left panels),orreceived sham
TX
orT.
5wkago(right panels). In all ofthese groups, a major proportion of immunoreactive GRF was eluted ata position identical to that of synthetic GRE. This indicates thatbyHPLCcriteria, ratGRF secreted into the mediumwasindistinguishable fromsynthetic peptide.A smallportion ofGRF-like immunoreactivity also was de- tectedatlaterfraction(s)than that of authenticratGRF,but the natureofthese smallimmunoreactive peaks is unknown.
Anelutionprofile of GRF of
TX
ratstreated with T4(2,gg/
100g perd)orhGH(200,ug/ratperd)for7dwassimilartothat of sham
TX
orTX
rats(datanotillustrated).Discussion
Boththyroidhormone andGRFarethephysiologicalfactors thatactdirectlyatthepituitary somatotrophand facilitate its functionalactivity.Their actionsweresimilar inmany aspects, which include stimulation ofGH genetranscription, mRNA levels,biosynthesis,andsecretion (4-8, 14-19).Theresults of thepresentstudiesdemonstrate thatthyroidhormone also af- fectsthesecretion of hypothalamic GRF, theGH-regulatory peptide.Theeffects of thyroidhormoneexcessanddeficiency onGRF secretionwereopposite,and therewasaninverserela- tionshipbetweenserumT4 levels and GRF secretion. This find- ing clearly indicatesaninhibitory role ofthyroidhormonein GRF secretion. Characterization ofGRF in the medium by reverse-phase HPLC revealed that the molecular form of GRF peptidewasnotaltered byanabnormalthyroidstatus.
Theinhibition of GRF secretion by thyroidhormone oc- curs atphysiological concentrationsin serum.Thisisindicated bythefindingsthatthyroidgland ablation enhancedGRFse- cretionand thatthis effectwashighlyreversible by replacement of T4atphysiologicaldoses.Also,atthyrotoxic doses,T4sup- pressed GRF secretioninnormalrats.Theseeffects ofthyroid hormone status onGRF secretionareprobablynotsimplythe result ofweightchanges or altered food intake. First, hyper- andhypothyroidism similarlydiminished BWs, while exhibit- inganopposite effectonGRFsecretion. Second,theinhibition ofGRF secretionbythyrotoxicosiswas alreadysignificantat ThyroidHormoneand GrowthHormone-releasingFactor 117
c0
U _
m
L-
ck:-
1 10 20 30 40 60 1 1 0 20 30 40 60
Fraction (number)
theintermediatedoseofT4,whichdidnotreduceinitialBWs ofanimals.Third, hyperthyroidismpromotesandhypothyroid- ismdepressesfood intake(35), whereasfood deprivationde- creasesGRF mRNAlevels (36) and GRFsecretionin therat (unpublished observation). Therefore, if thyroid hormone actedbyaltering food intake,it would have been the hypothy- roid but not hyperthyroid rats who showed the diminished GRFsecretion.
Consistent with a modulatory role in GH secretion, the effectthatthyroidhormone exertedonGRFsecretionwasofa delayed type at both physiological and thyrotoxic doses. Al- though the decreased GRFcontentsinhypothyroid rats were fullyrecoveredafter7dofT4replacement, the enhanced GRF secretionwas notcompletely reversedafter either7 or 14dof replacement. T4 was withouteffect on GRFsecretion when replacedfor4d in
TX
rats(26). Thesuppression ofGRFsecre- tioninhyperthyroidrats wasgreaterafteradministration ofthe highestdoseofT4 for 12than 6 d, and the dose-response rela- tionship becameapparentafter12d. These resultssuggestthat theeffectiveness ofthyroidhormoneonGRFsecretion is de- termined by the duration of itstreatment aswellasthe dose administered.Itis well establishedthathypothyroidism causes a severe depletion ofpituitary GH ( 14-16).Assoon as6 hafterreplace- ment, T4 can start to increase the diminished GH (16), whereaslittle isknown on aprecise timecourseofits recovery toeuthyroid level. Thepresentstudies givesomeinformation on thisissuebyreplacingT4 atphysiological dosesof1 or 2 ,g/ 100g,eachfor7and14d. However, replacement ofT4 for 14dwas notsufficientto recover thediminishedGH content completely. This is consistent witha morphological studyof DeFesi et al. that 17-24 d oftriiodothyronine (T3) replace- ment wasrequiredtonormalize decreasedsomatotroph num- ber in
TX
rats(37).In contrast,diminishedGH secretion inTX rats was normalizedwithin 4 dafter initiation of T4 replace- ment(22).Thesefindings,togetherwiththe present results on GRF, indicate that it takes > 14 d for replacement of T4 toFigure 5. Reverse-phase HPLC of a immunoreactivehypothalamicGRF
0 secreted invitro fromhypothalami
* of hyper- and hypothyroid rats and E their controls.Leftpanels, elution
0
a} profiles of GRF ofratsadministered
< with vehicle
(upper
panel)orT4 (100,ug/100 g per d)(lowerpanel) for 6 d.Rightpanels, elution profiles of GRFofsham Tx (upper panel) orT1 (lowerpanel) rats5wkafter operation.Arrowsindicate elution positions of syntheticratGRF. Re- sults wereexpressed aspicograms/hypothalamus per 30 min perfraction. Brokenlinesindi- catethedetectionlimit.
normalize the whole ofthe altered hypothalamic GRF-pitu- itaryGHaxisinhypothyroidrats.
In additionto depleting pituitary GHcontents, hypothy- roidism also abolishes spontaneous GH secretion (20-22), suppressesGHresponse toGRF(38-40), andcausesgrowth retardation (23). Qualitatively similar effects can be repro- ducedby GRF antibodies(4-6,41)oritsantagonists (42, 43).
Thesefindings initiallyledustothehypothesisthat hypothy- roidism might cause somatotroph failure, at least in part, throughinhibiting hypothalamicGRF secretion.By observing adecrease in GRFcontentafter
TX
anditsrapidrecoverybyreplacementofT4, Katakami et al.(22)also have suggested thatacriticalroleofthyroidhormone in GHsecretion isre- latedtoits effectonhypothalamicGRF. Later, weandothers demonstrated that GRF secretion in vitrowasrather enhanced inhypothyroid, GH-deficientrats(26, 29).We conclude, there- fore,that thepituitary somatotroph failureinhypothyroidism is not causedby GRF deficiencyand so most probably isa directconsequenceofthyroidhormonedeficiency.Thesame
conclusionwasobtained also inaninvivostudy withneonatal hypothyroidrats(44).
Theeffectofhyperthyroidism on thehypothalamic-pitu- itaryGH axis has been less clear. Thereissome evidenceto suggestthat
thyrotoxicosis
hasaninhibitoryinfluenceonGH.Hyperthyroidism causes amoderatedeclineinbothpituitary GH andBWsin the neonatal rat(45,46).Physiological GH secretionwasdiminished in humansubjectswith
thyrotoxico-
sis andwasnormalized aftertheybecameeuthyroid (47).The present studies have demonstrated that hyperthyroidism in- hibits GRFsecretion, concomitantly decreasingbothBWs and pituitary GH in the adultrat. Thisfinding, together with the report of Clark and Robinson (48) that GRF accelerates growth and increases GH content in rats, indicates that the decreased invitrorelease of GRF inhyperthyroid
ratsisnot dueto anincreasedturnover rateof GRF in vivo.Theparallel
decline in GRFsecretion,GHcontent, andpossibly
GHsecre- tionaswellsuggests that thehypothalamusisaprimary
targettissue ofhyperthyroidism in GH secretion. Thisview is sup- ported bythe finding that pituitary GH responsivenesstoGRF invivo was notaffected or only minimally decreased in hyper- thyroid rats (49, 50). Thus,atpathological conditions, thyroid hormone appears to negatively influence pituitary somato- troph functionby two pathways: oneinhyperthyroidism via the hypothalamus and the other in hypothyroidism leading directly tothepituitary. Thisdifference in the primarytarget tissue could explain biphasic effects of thyroid hormone on growthrate(28) andpituitaryGHcontents.
At a supraphysiological concentration of 1.55 X 10-6 M, theaddition of T3 stimulateshypothalamic SRIFsecretion in vitro (51 ). This effect is inhibitory for pituitary GH secretion, as wastheeffect thatthyroidhormoneexcessexertedonGRF inthe presentstudies. Hyperthyroidism may depresspituitary GH secretion through adualeffectonGRF and SRIF in the hypothalamus. ModulationofSRIF bythyroidhormonedefi- ciencyisconflicting. Hypothalamic SRIF secretion in vitro is reported to be decreased (51), unaltered (21), orenhanced (29) inhypothyroidrats.Thismay be due in parttothediffer- ence in theinduction time orlength ofhypothyroidism orin the methodofoperation, thyroidectomyvs.thyroparathyroid- ectomy. Anyway, thesediscrepantstudiesseemtoagreeonthat hypothyroidrats cancontinueto secrete aconsiderable propor- tion of SRIF, if itscapacityisevendecreased.However, itre- mains unknown whether SRIF inhibition of GHsecretionisof anysignificanceinseverehypothyroidism,inwhich GH is vir- tually absent in the bloodstream.
With replacement ofhGH, we could not reversethe en- hancedGRFsecretion associatedwith GHdeficiencyin hypo- thyroidrats.Atthe dose anddurationused, hGHwasaspotent asT4 inrestoringdecreasedbodyweight gainsof
T.
rats.hGHalso restored thediminishedserumlevelsof IGF-I,amediator ofanabolic effects of GH(52). Theeffect ofhGHon IGF-I induction wascomparableto thatofT4after 14dof replace- ment, although not significant after 7 d. This finding agrees wellwith the earlier reports on hypothyroid adult rats treated withbovineGHfor7d(53)orhGHfor 14d (54). Further- more, unlike the effect on growth (23, 30, 55), a combined hGH and T4 therapywas not moreeffectivethan T4 alonein inhibiting GRF secretion in hypothyroid rats. These results suggest that the increased GRF secretion inhypothyroidism is notmediated through the feedback mechanism involving ei- ther GHorIGF-Ideficiency.
Thereis biochemical evidence to support that thyroid hor- mone actsdirectlyatthe levelof the hypothalamus. The con- version of T4toitsactive metabolite T3 is an important step in mediatingtheaction of thyroid hormone on its nuclear recep- torand has beenconsideredthemainsourceofT3in thebrain (56,57). Type II 5 ' -monodeionase isanenzyme thatdeiodin- atesT4toT3 in the brain (56), and the activity of this enzyme is highest in the arcuate nucleus and median eminence in brains of hypothyroidrats (58). Because thesehypothalamic regionspossessbothcellbodiesand nerveterminals ofthe hy- pophysiotrophic GRFneurons(59, 60), T3 generated locally in thehypothalamus could directly influence the functional activity of GRF neurons. This viewmay be supported by the significantcorrelationbetween serumT4levelsand GRF secre- tionthat weobserved inhyperthyroidrats. Alternatively, thy- roid hormone may act indirectly through altering metabolic and/orotherfactors ofthe centralorperipheral origin.
Other laboratories have
recently reported
thathypotha-
lamicGRFmRNAlevels inratswereincreased in
hypothyroid-
ism
(26, 27,
61)
anddecreased inhyperthyroidism (61 ).
Thisnegative relationship
betweenGRFmRNAandthyroid
func-tionissimilartothat whichwefoundbetween
GRF
secretion andserumT4.
Intheir studiesonhypothyroid
rats,however,
GH showed an
inhibitory
effect on GRF mRNAlevels,
whereasweobserved that GHwaswithouteffecton
GRF
se- cretion.Theincreased GRFmRNA levelswerecompletely
re-versed
by
5 dofratGHreplacement
in adulthypothyroid
rats,andratGHwas morepotentthan
T4 (26).
Itwasalso reversedby
hGH inprepubertal hypothyroid
rats,although
14d ofre-placement
wasrequired
forcomplete
reversel(27). Thus, thy-
roid hormone modulates GRFmRNA levels and
peptide
se-cretioninasimilar
inhibitory
mode.However,
theunderlying
mechanismsare
probably different, involving
the GH feedback for GRFmRNAbut notfor GRFpeptide
secretion. Alterna-tively,
GRF secretionmight
be lesssensitivethanGRF
mRNAto the GH feedback
signal
orcouldrespond
toGH adminis- tered inaphysiological pulsatile
patternonly.
Acknowledgments
Wethank the NationalInstituteof Diabetes and
Digestive
andKidney
Diseasesforsupplyingmaterials forratGH RIA andSumitomoPhar- maceuticalCo. Ltd. for the generous
gift
ofrecombinanthGH.This work was supported in partby a grant from the Growth ScienceFoundation.
References
1. Guillemin, R., P. Brazeau, P. Bohhlen, F. Esch, N. Ling,and W. B.
Wehrenberg.1982.Growthhormone-releasingfactor fromahumanpancreatic
tumorthat causedacromegaly.Science(Wash.DC).218:585-587.
2.Rivier,J.,J.Spiess,M.Thorner,and W. Vale.1982. Characterization ofa
growthhormone-releasingfactorfromahumanpancreaticislettumor.Nature (Lond.).300:276-278.
3.Spiess,J.,J.Rivier,and W. Vale. 1983.Characterization ofrathypotha- lamicgrowthhormone-releasingfactor.Nature(Lond.). 303:532-535.
4.Wehrenberg,W.B.,P.Brazeau,R.Luben,P.Bohlen,and R.Guillemin.
1982.Inhibition ofthepulsatilesecretion ofgrowthhormonebymonoclonal antibodiestothehypothalamic growthhormonereleasingfactor(GRF).Endocri- nology. 111:2147-2148.
5. Miki,N.,M.Ono,and K.Shizume. 1984.Evidencethatopiatergicand a-adrenergicmechanisms stimulateratgrowthhormone release viagrowthhor- mone-releasingfactor(GRF).Endocrinology. 114:1950-1952.
6.Ono, M.,N.Miki,and H. Demura. 1991.Effectof antiserumto ratgrowth
hormone(GH)-releasingfactoronphysiologicalGHsecretionin thefemalerat.
Endocrinology. 129:1791-1796.
7. Barinaga,M.,G.Yamonoto,C.Rivier,W.Vale,R.Evans,andM.G.
Rosenfeld. 1983.Transcriptional regulationofgrowthhormone geneexpression bygrowth hormone-releasingfactor.Nature(Lond.).306:84-85.
8. Gick,G.G., F. N.Zeytin,P. Brazeau,N. C. Ling,F. S.Esch,and C.
Bancroft. 1984.Growth hormone-releasingfactorregulatesgrowthhormone
mRNA inprimaryculturesofratpituitarycells. Proc.Natl. Acad. Sci. USA.
81:1553-1555.
9.Brazeau, P.,W.Vale,R.Burgus,N.Ling, M.Butcher,J.Rivier,andR.
Guillemin. 1973.Hypothalamicpolypeptidethatinhibits the secretion of immu- noreactivepituitary growthhormone. Science(Wash. DC). 179:77-79.
10.Reichlin,S. 1983. Somatostatin.N.Engl.J.Med.309:1495-1501.
11.Plotsky,P.M.,andW.Vale.1985. Patterns ofgrowth hormone-releasing factorandsomatostatin secretion into thehypophysial-portalcirculationin the rat.Science(Wash. DC).230:461-463.
12.Miki,N.,M.Ono,and K. Shizume. 1988.Withdrawalofendogenous
somatostatin induces secretion ofgrowthhormone-releasingfactor inrats. J.
Endocrinol.117:245-252.
13. Reichlin, S. 1974. Regulation of somatotrophic hormone secretion. In HandbookofPhysiology,Section7:Endocrinology,Vol.IV:ThePituitaryGland
ThyroidHormoneandGrowthHormone-releasingFactor 119