Influence of the high affinity growth hormone (GH) binding protein on plasma profiles of free and bound GH and on the apparent half life of GH Modeling analysis and clinical applications

Amendment history:

Correction

(August 1993)

Correction

(May 1993)

Influence of the high-affinity growth hormone

(GH)-binding protein on plasma profiles of free

and bound GH and on the apparent half-life of

GH. Modeling analysis and clinical applications.

J D Veldhuis, … , M Mercado, G Baumann

J Clin Invest.

1993;

91(2)

:629-641.

https://doi.org/10.1172/JCI116243

.

The discovery of a specific high-affinity growth hormone (GH) binding protein (GH-BP) in

plasma adds complexity to the dynamics of GH secretion and clearance. Intuitive

predictions are that such a protein would damp sharp oscillations in GH concentrations

otherwise caused by bursts of GH secretion into the blood volume, prolong the apparent

half-life of circulating GH, and contribute a reservoir function. To test these implicit

considerations, we formulated an explicit mathematical model of pulsatile GH secretion and

clearance in the presence of absence of a specific high-affinity GH-BP. Simulation

experiments revealed that the pulsatile mode of physiological GH secretion creates a highly

dynamic (nonequilibrium) system, in which the half-life of free GH, its instantaneous

secretion rate, and the GH-BP affinity and capacity all contribute to defining momentary

levels of free, bound, and total GH, the percentage of GH bound to protein, and the

percentage occupancy of GH-BP [corrected]. In contrast, the amount of free GH at

equilibrium is specified only […]

Research Article

Find the latest version:

Influence

of the High-Affinity Growth

Hormone

(GH)-Binding

Protein

on

Plasma

Profiles of Free and Bound GH and

on

the Apparent Half-Life of GH

Modeling Analysis and ClinicalApplications

Johannes D. Veldhuis,* Michael L.Johnson, LUndsay M. Faunt,t Moises Mercado, andGerhard Baumann

**Division of Endocrinology and Metabolism, DepartmentsofInternalMedicineandtPharmacology, UniversityofVirginiaHealth SciencesCenter, and NationalScience Foundation CenterforBiologicalTiming, Charlottesville, Virginia22908;andlCenterfor Endocrinology, Metabolism andNutrition, DepartmentofMedicine,Northwestern UniversityMedical School, Chicago, Illinois60611

Abstract

Thediscovery ofaspecifichigh-affinitygrowth hormone (GH)

binding protein (GH-BP) inplasma adds complexity to the

dynamics of GH secretion and clearance. Intuitive predictions

are that such aproteinwould damp sharp oscillations in GH

concentrations otherwisecaused by burstsof GHsecretion into

the blood volume, prolong theapparenthalf-lifeof circulating

GH, and contributeareservoir function.To testthese implicit

considerations,weformulatedanexplicit mathematicalmodel

of pulsatileGH secretion and clearancein the presence or ab-sence ofa specific high-affinity GH-BP. Simulation experi-ments revealed that the pulsatile mode of physiological GH

secretioncreates ahighly dynamic (nonequilibrium)system,in

which the half-life of free GH, its instantaneous secretion rate, andthe GH-BP

affinity

andcapacityallcontributetodefining

momentarylevelsoffree, bound, and total GH; the percentage

ofGHbound toprotein;and the percentage occupancy of GH-BP. In contrast, the amountof freeGHatequilibriumis

speci-fied only by the GH distribution volume and secretionrateand thehalf-life of freehormone.Weconclude thatthe invivo

dy-namics of GHsecretion, trapping, and clearance from the

circu-lationofferavariety of regulatory lociatwhichthetime struc-tureoffree, bound,and totalGH deliveryto targettissuescan becontrolled physiologically. (J. Clin.Invest. 1993.

91:629-641.) Keywords:diabetes mellitus* growth hormone * Laron

dwarfism*obesity

Introduction

Recently a

high-affinity, specific, low-capacity

growth

hor-mone-binding

protein

(GH-BP)'

thatcirculates in blood has

been

characterized, purified,

and clonedinhumans andseveral

animal

species

( 1-7). This

high-affinity

GH-BP correspondsto

the extracellulardomain of the tissue growth hormone (GH)

receptor(5-9), andcontrastswithan

incompletely

character-Addressreprintrequeststo Dr.JohannesD.Veldhuis,Division of En-docrinologyandMetabolism,Box202, Departmentof Internal Medi-cine, University of VirginiaHealthSciences Center, Charlottesville,

VA22908.

Receivedfor publication 17April1992and inrevisedform4 Sep-tember 1992.

1.Abbreviations used in thispaper:GH-BP,growthhormone-binding

protein;IGF-I,insulin-likegrowthfactor I.

ized low-affinity binding protein that is unrelatedtotheGH receptor (10, 11). High-affinity GH-BP concentrations are very low ordysfunctional in Laron-type dwarfism ( 12, 13) and decreased in fetal plasma and in certain GH-resistant popula-tions such as pygmies in Africa or the New Guinea highlands ( 10, 14-16). Studies in heterologous species (e.g., injection of human GH-BPcomplexed with human GH into the rat) indi-catethat the presence ofa high-affinity binding protein (BP) can prolong the apparent half-life of GH in plasma ( 17-19).

Moreover,anintuitivepredictionis thatplasmaGH-BP would damp oscillations engendered by the burstlike secretion of GH into the plasma compartment, and contribute to the retention ofGH in the circulation ( 17, 20, 21 ). Here, we have tested these intuitive inferencesby formulating an explicit mathemati-cal model to evaluate the predicted effects of the high-affinity GH-BP in plasma on the half-life of total, bound, and free GH, and the steady-state and nonequilibrium plasma concentra-tions offree, bound, andtotal GH as a function of the mode

(pulsatilevs.continuous)and amountof GHsecreted.

Methods

Computersimulations ofGH dynamics in vivo. As an initial approxi-mation of the dynamicsofGH secretion and removal in the absence and presence ofaGH-BP, we assumed that a single high-affinity, low-capacity,andspecificGH-BPcontributes predominantly to GH bind-ing in plasma, i.e., that any contribution of other low-affinity BPs is relatively small by comparison, as has been shown (10, 11, 22). In addition,weprovisionally assumed that (a) the turnover ofthe GH-BP is slow comparedtothatoffreeGHitself; (b) the intact GH-BP com-plex is removed at a rate that is slow compared to the clearance of free GH (18); (c) GH andBP mass areconserved;(d)thebinding stoichi-ometry of GHtoits BP isequimolar2; (e)thenative elimination func-tionfor removal ofGHirreversiblyfrom plasma isasingle exponential, which acts only onfree GH;(f)GHof 22,000-D molecular mass is measured andconstitutes theprincipal circulating form of GH (23, 24); (g) both bound and free GHaremeasuredasimmunoreactive GH (25);(h)GHis secreted in distinct pulses of finite half-duration (dura-tionathalf-maximalamplitude), amplitude,and mass;(i)BP

concen-tration/activity isconstant overtime(26, 27);and(j)thefateofa

molecule ofGHinjectedinto the bloodstream by the anteriorpituitary gland includes entry into the free(protein unbound) compartment, possible association withthe BP, dissociation from the BP, and/or irreversible elimination of free hormone by metabolic and excretory mechanisms.

2.Recentinvitrostudies indicate thata_singlemolecule of GHcan

capturetwomoleculesof recombinant GH-BP

through

twobinding

sites; seereference 53. It is notclearatpresent whether such a2:1 complexoccursinplasma.Evidencetodatehas indicateda predomi-nantly 1:1 stoichiometryfor the nativecomplexinplasma.

J.Clin.Invest.

© The AmericanSocietyfor ClinicalInvestigation,Inc. 0021-9738/93/02/0629/13 $2.00

Based on earlier in vitromeasurementsof theequilibrium associa-tion of human GH with its naturalhigh-affinity plasmaBP and values averaged from several laboratories(1,2, 9, 12, 20),weusedan appar-entequilibriumKd of 1.5 mMas anominal value inmostsimulations.3

Inaddition,weusedanominal V0for GH of7%body weight (28),a

putativetl/2oftotal GH of 18 min(29),andameanGH-BP

concentra-tionof 1.0 nM ( 1,2,9).From earlierestimates,weassumed that the daily total secretion of GH in humanscanspananapproximaterange of0.1 1-1.1 mg/d(5-50 nmol/d),and thatasingleGH secretory burst has a masseither 2.2or 22 gg (0.1 or 1.0 pmol), unless indicated otherwise(30-33).

Theassociation ofGH with its BP has beenpreviously reported (1); itsrapidity precludes preciseestimation ofthe associationrate constant

(k/.)

with availabletechnology.Therefore,the much slower dissocia-tion ofthecomplexwasmeasured. To estimate the unidirectional disso-ciationrate constant(offrate,koff)of GH from thehigh-affinity GH-BP,pooled adult humanserum(9 ml)waspreincubatedwith - 9 ng

ofmonomeric [12251]-humanGHfor 2 hat370Casdescribed(1).An excessof unlabeled human GH(1 mg/liter)wasthen added in0.1ml ofphosphate-buffered saline, pH 7.4, and incubation continued at

370C.Periodically, aliquotswereremoved,immediatelychilled inan ice-water bath, and free GHwasseparatedfrom bound GHon Sepha-dexG-100 columns at4VC. Peakscorrespondingtothehighaffinity

GH-BP complexand free GH wereintegrated (1),and boundGH expressedas afunction of time after initiation of dissociation. Asingle

rateconstantofdissociationwasestimated from thesemeasurements

as0.037/min (67%confidence interval0.030-0.043);seeFig. 1.

(At-temptstodetermine thedissociation of the lowaffinityGH-BP

com-plex[ 10] in wholeplasmawereunsuccessfulowingtothesmall percent-age of GH boundtothatBP, which renderedprecisekinetic estima-tionsdifficult.)

The following differentialequationswere formulatedtodescribe thebinding, dissociation, and removal of secretedGHfrom human plasma.Specifically, givenaconstantsecretion rate, theratesofchange

of concentrations of free and boundGHovertime(t)aregivenby:

d[free] _=_(k +

_kn

.([BP]

-

[bound])).[free]

+

_kfff.

[bound]+-V (1)

dlboundl~d=

k

n.([BP]

-[bound])

.[free]

-

kff- [bound]

(2)

dt

And, ifweassumethatapulsedinjectionorsecretoryburst ofGH entersthe bloodstreamby way of the free compartment, then d

[free]

=

_-(k,

₊

_k.

(

_[BP]

-

_[bound]))

* [free]

dt

+

k~ff-

[bound] +

_VS)

(3)

d

[free]

_-(k,

_{+ k . ([BP]}- [ bound]))-

[free]

dt

+

koff.[bound]

+ ->

_ampl(i)

_exp I 1 (4)

These constructsassume that in plasmaGH is either bound or free, where[free], [bound], and [BP] represent the concentrations (M) of

3.Arelatively wide range of Kd values is reported for different in vitro incubation and separation conditions ofhuman GH binding in plasma or various buffer systems: viz. from 0.25 nM (Michael J. Waters, Queensland, Australia)to5.0nM(reviewed in reference 24).Wehave thenominal value of 1.5 nMformostof the present work.

a- 28

o

20-O

12-04

E _{0 400 800 1200}

Time (min)

Figure 1. Timecourseof in vitrodissociation ofGH from the human high-affin-ity GH-BP in plasma at 37°C. Observationswerefit to afirst-order (monoex-ponential)decay.

free GH, bound GH, and BP, respectively; S designates thesecretory pulse function withahalf-duration ofsigmatimes 2.354, and ampli-tudeofampl(i);SR denotes the basal secretionrateof GH(mol/min), if any; V0 denotes thedistributionvolume(liters);

k,

(min') is therate constantfor GHto BP(M'min);koffis thedissociationrateconstant (which is related to itst1/2 byln

2/Ik,

min);

kI,,

is theassociation rate constantfor GHto BP(M

-'/

min);koffis the dissociationrate constant

(min-');andkdis theequilibrium dissociation constant, which equals

koff/k0n

(M)

-The above differential equations were solved using the Runge-Kutta formulas as implemented for first-order differential equa-tions (34).

Results

Equilibrium conditions.As afirst approximation, we investi-gated GH kinetics under equilibrium conditions, although under

physiological

circumstances suchconditionsrarely ap-ply. As shown in Table I, different mean daily GH secretion

rates within and exceeding the approximately physiological

range observed in the human across various age spans and

body weights yieldacorrespondingly broadrangeoftotal, free,

and boundGHconcentrations in plasmaattheoretical equilib-rium. Varyingthe mean secretion rateofGHinfluencesthe apparent equilibrium plasma concentrations of free, bound,

and totalGH,and the percentage occupancyofthe BP,butless

markedly the percentage ofGH bound to the BP. The latter decreasesslightly (from 39%to33%) when themeanGH secre-tionrateis increased 10-foldabove thenominalnormal value

inhumans.

Table I alsoshows theinfluenceof altering the equilibrium

dissociationconstant. Usingtheapproximate literature value

forthe equilibrium dissociation constant, namely a kd of 1.5

nM,3

and our measuredkff of0.037/min,weobserved that BP

affinity

for GH influencesthe total andboundconcentrations ofGH inplasma,the percentageofGH bound, and the percent-age occupancyoftheGH-BP,but not theconcentrationof free GH. Alternatively, altering the apparent capacity of the BP

modifiestheconcentrations oftotal and boundGHand per-centageofGH bound at anygivenmoment, but does not influ-encetheconcentration of freeGH or the percentage occupancy

ofthe BP.

The

t1/2

of free GH within the physiological range influ-encestheconcentrations of freeand boundGHinplasma, the

t1/2

oftotalGH,and thepercentage of BP that is occupied, but

TableI.Estimated Molar Concentrations and Percentages ofFree and Bound GH in Plasma at Equilibrium in thePresence of Varying Amounts and/or Affinities ofGH-BP and Different Simulated Conditions

Percentages Predicted steady-state

serumGHconcentrations GH

boundto BP

Simulated BPproperties GHsecretionrate Total Free Bound protein occupied

pmol/min pM

Nominal* 10 34.2 20.6 13.6 39.8 1.3

Double the BP affinity 10 47.4 20.6 26.8 56.5 2.6

Double the BP concentration 10 47.4 20.6 26.8 39.8 1.3

Halve the BP concentration 10 27.4 20.6 6.8 39.8 1.3

Complete absence ofBP 10 20.6 20.6 0 0 0

Nominal but secretion rate doubled 20 68 41.2 26.8 39.4 2.6

Nominal but secretion rate increased 10-fold 100 327 206 121 37.0 12.1

Nominal but GHhalf-life doubled 10 68.0 41.2 26.8 39.4 2.6

Nominal but GHhalf-life halved 10 17.1 10.3 6.8 39.4 0.65

Volume of distribution doubled 10 17.1 10.3 6.8 39.8 0.65

*_{Assuming: a nominal free}

GH

_{half-life of7min in the absenceof BP, which correspondsto atotal GHhalf-life of 18 min in the presence of BP;}

aGHdistribution volume of 4.9 litersor7.0% ina70-kgman;aplasmaGH-BP concentration of 1 nMwithanequilibriumbindingconstant (Kd) of 1.5 nM andanoff-rate constant

(kff)

of 0.037min-';anominaldailyGH secretionrateof 314_,g,which correspondsto - 10pmol/min

of GH secretion.GivenaGH molecularmassof22,000 D,then 45 pM is 1 ug/liter.

t112

ofGH andnegativelyby the value ofthe V0, butis not

influencedbythebinding propertiesofthe GH-BP. The

con-centrations of boundand totalGHarecontrolledbythe GH secretionrateand t

12,

and thebindingaffinityandcapacity of

the BP. The percentageof GH boundatequilibriumis

speci-fied by the BP affinity and

capacity,

but not the

t112

or SR

withinthephysiological range.Finally,the percentageof GH-BPthat is occupied at steady state is stipulated by the GH

secretionrate,

t112,

and

VO

andby the

affinity

of theBPbutnot

its capacity.

Theabovecomputer-simulation dataapply to steady-state

conditions,

whichcanbeachieved by infusions ofexogenous

GH, butarerarely ifeverattained spontaneously under physio-logical conditionsinvivo (30-33, 35).Aplausiblepathological

exception

atleastovershorttimespans

(minutes)

may be

pa-tients with

GH-secreting

pituitarytumors,sincethelatter ap-pear toexhibitintervals of tonicorbasalsecretionuponwhich

may be superimposed high

frequency

bursts of GH release (36). Asshown in

Fig. 2, decay

fromsome

given steady-state

plasmaconcentration of

free,

bound,ortotal GH isinfluenced bythe presenceofa

GH-binding

elementin

plasma.

The

abso-lute valueofthe rate constant foran

approximated

monoex-ponential

decaystructure

depends

upon theamountand bind-ingproperties oftheGH-BP,the

_tl/2

of freeGH

removal,

and

the

starting

concentrations of freeandbound GHin

plasma.

For

example,

for threedifferent

starting

concentrations of total

GHinplasma, namely 11,

34,

and 101

pM, generated

respec-tively

bymean

daily

secretionratesofGH of0.1 1,

0.33,

and1.1

mg/24h(4, 15, and 40nmol/24 h),wefind by

monoexponen-tial

approximation

the apparent

t1/2

valuesshownin Table II. These values fall within thewiderange of measured GH

ki-netics reported in theavailable literature

(reviewed

in

refer-ences29 and33).

Fig.

3

gives

the

monoexponential

approxi-mation for bound, total,andfree plasmaGHhalf-livesin the presence vs.absenceof BP,

assuming

arangeof freeGH

half-lives in the absence ofa BP. Note that afreeGHt12of- _{7 min}

in the absence of BPcorrespondstoa

t,2

of 18 minfortotal GH when the BP is present in physiological concentrations.

Nonequilibrium conditions. We next simulatedthe more complex nonequilibrium conditions that are more akin those occurringphysiologically.The predicted dynamic effects ofthe BP on the plasmaconcentrations of free,bound, and total GH overtime generated by a secreted pulse ofGH are shown inFig.

4.In response to a single pulse of GH, the plasma concentra-tion of free GH increases rapidly, and decays rapidlyinitially

because of removal by irreversible metabolic clearance as well as by association with BP in plasma. The concentration of bound GH increases almost as rapidly, but decays considerably moreslowly because molecules ofGHdissociate fromthe BP at afinite rate while some additional molecules ofGH continue toassociate with theBP. The plot oftotalGHconcentrations

overtime is also considerably more extended than the profile of free hormone for the same reason. Ofinterest,the percentage of GH that is bound to plasma BP begins at - 39% (assumed

equilibriumapproximationofprevioussecretionevents),and declinesrapidlyasthe freeGHconcentrationincreasesduring apulseofGHdeliveryinto thebloodstream.Thepercentageof GH bound to BP thenincreasesoverthe

succeeding

minutes,

andthereafter beginstofallasGHdissociates from the specific

BPreservoir.

Fig.5 (leftpanels) illustrates the dynamicsof two different GH pulse masses, namely 2.2

Ag

(0.1 nmol) and 22 ug (1.0

nmol). Table III givesthetimes to maximal values forfree, bound,and totalplasmaGHconcentrations in relationtothe threedifferent amounts(nanomolesper

burst)

of GHinjected intothebloodstream. Aspredicted

intuitively,

freeplasmaGH concentrationpeaks first,followedbythe bound and total GH

Total

80-40

0

80

40

0

80

40

0

80

40

0

0 100 200 300

Bound

0 100 200

80

40

80

40

0

80

40

0

80

40

0

300

Free

.(7 MIN)

, i,"

0 100 200 300

Time

(min)

Figure2. Predicted timecoursesoftotal,bound,and freeplasma GH concentrations after decayfromasteady-state GH infusionrateof10 pmol/min (314gg/d)assuminga

V.

of4.9LandaGHmolecularmassof 22,000D.Foursimulationsareshown: (top) withnoBPpresent andafree GHt1 of7min(whichwouldcorrespondtoatotalGH t1/2of18minin thepresenceofBP);(secondrowfromtop)in thepresence

ofaBPwith Kd= 1.5nM,capacity 1.0 nM,and

koff

=0.037/min; (secondrowfrombottom)asimilarBPbutwitha_{t1/2 of free GH of}14min

(totalGHt1231 min);and(bottom)asimilarBPwithaKdof 0.75 nM (double the affinity). The_t11 values shownonthefigurearefrom

approximated monoexponential fits of the simulateddecay data. Thesetheoreticalcurves aresubjecttotheassumptions of conservation ofGH mass,irreversiblemetabolicremovaloffree hormone only, anda1:1 stoichiometryof GH bindingtoitsBPataconcentration of 1.0nM. As

afirstapproximation,the GH-BP complexwasassumedtobeformed(kI. =2.47 X 10'M-'min), remain intact,ordissociate(k0ff=0.037/

min), butnotundergo irreversible metabolicremoval.

amplitude) ofoneminute.One-tenth of the daily production rateofGHwas"injected" in this single pulse (22jIgor 1nmol, as above). Note that the plasma concentrations of free and bound GH under these conditions increase to considerably higher absolute values, and that the duration of the total

plasma GH concentration pulse is abbreviated. Comparison of a 1- and 30-min secretory burst is shown in Fig. 5 (right panels).

Wenextevaluatedthe effects of four different half-lives of free GHontheprofiles of free, bound, and total GH concentra-80

40

NO BINDING PROTEIN

(7 MIN)

NOMINAL BINDINS

PROTEIN

(18 MEN)

(27 MIN) (11 MIN)

0

80

40

0

80

40

0

80

-J

E

a

%0. c

0

co

._

4"

-0 0

CD

co

E

0

co

DOUBLE HALF-LIFE

(31 MEN)

(40 MIN) (25 MEN)

DOUBLE AFFINITY

(27 MIN)

(33 MEN)

40

0

(16 MEN)

Table I.Approximated MonoexponentialGHHal/-Livesand

Steady-State

Plasma GH ConcentrationsafterContinuous SecretionofGH Estimated half-lives* Concentrationsatsteady-state

Total Bound Free Total Bound Free

min pM

Secretionratest

314,gg/d

(10pmol/min) 18 29 9 [34] [13] [21]

104 ug/d (3.33pmol/min) 18 29 9 [11] [4.5] [6.5]

942

,gg/day

(30pmol/min) 18 29 9 [101] [39] [62]

Half-livest

3.5 min 13 25 4.3 [17] [7] [10]

7.0min 18 29 9 [34] [14] [21]

14min 29 43 22 [68] [27] [41]

GH-BP

Absent 7 None 7 [21] [0] [21]

Twice the capacity 25 35 16 [48] [27] [27]

Twice thenominalaffinityt 25 35 16 [48] [27] [21]

* Values are fitted GH half-lives in the indicated compartments after abrupt cessation of equilibrated continuous secretion of GH at the indicated rate(orotherwise,314Mg/d= 10pmol/min).t Theoretical decay half-time of free hormone in the absence of BP was assumed to be 7 min or asindicated. Also assumed are a distribution volume of 4.9 liters, a GH-BP concentration of 1.0 nM and Kd of 1.5 nM. Note that once a BP is introduced, thehalf-lifeoffree GH is controlled by both the removal rate of unbound GH and the binding properties of the BP.

tions in plasma. We used nominal half-lives

of

free GH of4, 9,

18, and 36(Fig. 6). The corresponding approximate half-lives oftotalGHare

16, 21, 37,

and63min,valueswhichcoverthe range reportedintheliteraturefor physiological and

pathologi-cal conditions. One-tenth of the daily GH secretion rate was

injected as a single pulse in these simulations (22 ,ug or 1

nmol). Note that the times to maximum and the maximal valuesofthe plasmaconcentrations of free, bound, and total

GHarestrongly controlled by theapparenthalf-life offreeGH

inplasma; see Table III.

Thepathophysiological impact of decreased plasma GH-BPconcentrationsonpulsatile free,total, and bound GH

con-centrationsovertime is illustrated in Fig.7(left panels),which comparesGHprofiles over24h ina normal male(nominal GH-BPconcentration 1.0 nM) andapatient with insulin-de-pendent diabetes mellitus(meanGH-BP concentration0.725 nM).Nonlinear curve-fitting ofthe total plasma GH

concen-trations measured in samples collectedat 10-minintervals for

24 hshowedsignificant increases inGHsecretionrates andfree GHconcentrations in diabetes mellitus, andadecrease intotal

60 BOUND

E ," a/ Figure

3.

Monoexponential

C _{40, approximation of GH}

half-., Art' / / livesforbound,total,and I

0,," ,/ / ,,t'' free GH given thenominal 20 ,vS

/;FREE

SS GH-BPmodel summarized

X20- '^ //° ares'- in Fig. 2, and assuming

a z

.-SSNo

BINDINGPROTEIN various

theoretical

half-livesof free GH in the

ab-00 0 senceofaBP.The line with

0l10 20 30 _{the filled squares} corre-Theoretical Half-life ofFree GH(min) spondsto total GH.

but not free GHhalf-life as predicted by the kinetic model.

Conversely, the effects of increasedserumGH-BP

concentra-tionsareshownforcomparison in Fig.7(right panels), which

depicts

24-h serum GH concentration

profiles

obtained by

sampling

bloodat10-minintervalsfor24 hinanobese middle-aged mananda nonobeseage-matched control.

Assuming

a

blood GH-BP concentration of1.7 nMin the obese (vs. 1.0nM

in the normalweight) subject,weuseddeconvolution analysis

todetermine the half-lifeof total and free GH, the time-course of plasma GH

secretion,

and the resultantplasma

concentra-tions offree GH, bound GH, and total GH. The

last-men-tioned valueswerecompared to measuredconcentrations of totalGH;seeFig.7(right panels).

Thedynamicchangesin GHbindingover 24 h in the above

subjects

areillustratedin

Fig.

8. Note that in thehealthy, obese, anddiabeticmen, the percentageof GH boundtohigh-affinity

BPvariesfrom 10%to80%over the courseof 24 h. Of particu-lar

interest, during

intervals of secretory

quiescence (i.e.,

be-tweensuccessive GH secretory bursts), the percentageof GH bound increases

progressively,

whichreflectsthereservoir-like and

buffering capacity

of GH-BP.

Conversely, during

individ-ual GH secretory

bursts,

the free concentration of GH in-creases

rapidly

andthe percentageof bound GH falls with a

consequently increased

availability

of GH to metabolic

re-moval. These

dynamics

apply to the normal,

diabetic,

and obese

profiles.

Discussion

Toevaluatethe

dynamics

of GH secretion

into,

removal

from,

and

partitioning

within

plasma,

we formulateda model that

assumed eithertheattainment of

equilibrium plasma

GH

con-centrations (e.g., achieved

by

constant infusionofexogenous

pitu-37.5

BOUND HORMONE

E 25.0 0. a,

12.5-0.0

75

-J

0

E

C.

50-25

-o

60

-40

20

-

0--80

20 -J

E

a 10

0

-i

E

0. 50

25

o

2

o -8

0 80 160 240

Time

_(O

itarytumors), or aphysiologically pulsatile mode ofGH re-lease, characterized by abrupt bursts of GH release into the bloodstream.To thisend,wemeasuredthe unidirectional rate constantforGHdissociation from its BP in vitro (- 0.037/

min), andusedavailable earlierestimates oftheequilibrium dissociation constant of GH from its BP,3 the total binding capacity of the GH-BPin humanplasma,andanassumed 1:1 stoichiometryfor GHbinding.2Forquantitativepurposes,we carriedoutcomputersimulationstodefine theequilibriumas wellasthe nonequilibrium characteristics of GH binding and

disposalovertimeinthepresenceorabsenceofaspecific

high-affinity GH-BP, given various assumed half-lives for the re-moval of freeGH fromplasma, plausible GH secretion rates

andsecretorypulse waveforms, differentBPaffinitiesand ca-pacities,anddifferentapparent V0's for GH.

Ouranalyses demonstrate that certain intuitive predictions ofthe effects ofGH-BP on the dynamics of GH secretion, distribution, and clearancecanbecorroborated quantitatively. Indeed, weidentifiedspecific predictors of plasma

concentra-tions of free, bound, and total GH, the percentage of GH boundtoplasmaBPatanygiven instant, and thepercentageof BP occupied by GH. The steady-state predictions given here correspond simply to those of a two-compartment kinetic model, andaretobe anticipated only during continuous (exog-enous) GH infusions, or in selected pathophysiology (e.g.,

FREE HORMONE Figure 4. Timecoursesof calculated GHsecretoryrate, HALF-LIFE-7 MIN total,bound and freeplasma

GH concentrations, and per-centagesof GH bound and protein occupied assuming: (a) ahigh-affinityplasma GH-BP(kd= 1.5 nM,

capac-ity= 1.0nM);(b)a distri-bution volume of 4.9 liters; (c) a total GH

_t1,2

of 18min;

PROTEIN CCUPIED (d)amolecularmassof GH

PROTEIN OCCUPIED _{of22,000D;(e) a}

_koff

_of 0.037/min; and (f) a sym-metric (Gaussian) GH secre-tory burst with a half-dura-tion (durahalf-dura-tion at half-maxi-malamplitude) of 30min centeredabout zero and a

massof 1 nmol(22

Mg);

see

Methods. For ease of visual-o0 0 80 160 240 ization,thecurveshavebeen

elevated toreflectalow rate of constant basal secretion ( 1

uin)

pmol/min).

somepatients with acromegalywho have highratesofbasal GH secretionwith no orfewsuperimposed pulses).In addi-tion,wepositedamodelcombiningpulsatile (i.e., nonequilib-rium) GH secretion and stable amounts of the GH-BP in plasma in orderto mimicphysiologicalconditions of burstlike pituitary GH secretion.This constructofpulsatile ligand secre-tion and associasecre-tionwithanddissociation fromahigh-affinity BP inplasma is schematizedinFig. 9,andoffersnovel predic-tionsthatarerelevant toanyprotein, metabolite,orsubstrate injected into the bloodstream withadefined timecourse,when thatligandcanbindtoone or morespecific (highorlow affin-ity) BPs in the plasmacompartment.Moreover,ourdynamic modelcanbeexpandedtoconditions in which either the unoc-cupied BP or ligand-BP complexesare themselves removed fromthecirculation, and/ortosituationswherehigher-order exponential kinetics exist for the removal offree, bound, or totalligand from plasma.

Under the simplifying assumptions of full equilibration, conservationofmass,and clearanceoffreehormoneonly,we foundthattheV0,meanGHsecretory rate, and thet1/2of free hormone allsubstantially affect the steady-state concentrations offree and bound(and total) GH,aswellasthepercentage of BP occupied. These parameters influence the percentage of GHboundconsiderablyless. Inaddition,whereasthe BP affin-ity influences theamount(concentration and percentage) of TOTAL HORMONE

HORMONE BOUND

Pulse of 1 nmol (22 p9)

I00_100K

,-Free

...J 50

0.

a ₀

100

50

0' 100

50

0f

-40 40 120 200 280

Pulseof 0.1 nmol

(2.2 pg)

200

Bound

100

0

_*_.

-40 40 1 20

Time (min)

1 MINUTE

200

100

0

30 MINUTES

r

_

Bound

dI

Figure5. (Left panels) Computer simulation ofahypothetical GHsecretoryburstcontaining specifiedamountsofGHwithinthepresumptive physiologicalrange.Asingle burst of GH secretionwassimulatedandcenteredabouttimezero(abscissa).The totalmassof GH delivered in

the burstwaseither 1 nmol(22 Mg, left column),or0.1 nmol (2.2 Mg,right column). The GHsecretoryburstwassimulated algebraically bya

Gaussian distribution of GH secretionrates,inwhich the half-duration (timeinminutes elapsinginhalf-maximal amplitude)was30min. The vertical axis givesthecalculatedplasmaGHconcentration inpmol/ L (pM). In each column threecurves areshowntodenote the expectedtime coursesoffree, bound,and totalGH assuming the nominalparametersgiveninFig.4.(Right panels) Timecourseof predicted plasma GH

concentration profiles afterapresumptively physiologicsecretoryburst (right column)comparedtoanearlyinstantaneouslysecretory event(left column),whichwereassigned respective half-durations (durationinminutes ofthesecretion episodeathalf-maximalamplitude)of30minor

1 min. Resultsarepresentedotherwiseasdescribed for left panels.

TableIII.PeakGHSecretoryRates, Times toMaximal Values, and Maximal PlasmaGHConcentrations afteraSingleIntravenousPulseofGH

Times to maximal values Maximalplasma GH concentrations Peaksecretory

rate Free Bound Total Free Bound Total

pmol/min min pM

Massof GH pulse,_,gg(nmol)

22 (1.0) 32 7 23 10 [50] [21] [67]

2.2(0.1) 3.2 7 23 10 [5] [2] [7]

0.22(0.01) 0.32 7 23 10 [0.5] [0.2] [0.7]

FreeGHhalf-life (total GHhalf-life)

4(16) 32 4 19 7 [33] [13] [43]

7(18) 32 7 23 10 [50] [21] [67]

9(21) 32 9 25 11 [59] [25] [79]

18(37) 32 12 31 15 [86] [39] [117]

36 (63) 32 16 38 20 [114] [54] [158]

GHsecretorybursthalf-duration(min)

1 880 1 16 1 [180] [26] [190]

30 32 7 23 10 [50] [21] [67]

50 18 9 27 13 [34] [17] [49]

Values arepeaksecretoryrates andtimes (min)tomaximalvaluesoffree,bound,ortotalplasma GH concentrationsassuminga_singleGaussian secretoryburstcenteredattimezerowith ahalf-duration(durationathalf-maximal _amplitude)of30 min, a massof22

jtg

(1 nmol),a distribution volumeof4.9liters,ahalf-life of total GH of 18 min(correspondingto afree GHhalf-lifein theabsenceofBPof7

_min),

andaBP concentrationof 1.0nMandequilibrium Kdof1.5nM,exceptwhereindicated otherwise.Note that the

time-to-peak

valuesdiffer forfree, bound, andtotalGH,andhence themaximal valuesgiveninthelast three columns donotsimplysum.

Bound

a

0 Ca

-Ca

E

ca

Total 200Total

100

200 280 -40 40 120 200 280 -40 40 120 200 280

Half -life

36

min

200

00 k

-J 0

E

a

0.

a

0

cow

4-.Cu L.

0 0

0

E

co a.

01

200

100

0

200

100 t

Half -life

18

min

Half-life

9 min

Half-life

4 min

U ,... . . . .

-100 0 100 200 300400 -100 0 100 200300 400-100 0 100 200 300 400-100 0 100 200 300 400

Time

(min)

Figure 6.Influence of differentfree GH half-livesontheprofilesofplasma

free,

bound,and totalGH concentrations

following

a

hypothetical

GH secretoryburst. ThehypotheticalGHsecretoryburstwitha massof1 nmol(22

gg)

is described in

Fig.

5(left panels).The four columns of panelsdepicttheanticipated plasma concentrationsof

free,

bound,and totalGH

assuming

monoexponential decay

ratesof free GH in the

ab-senceofBPcorrespondingtohalf-lives of36, 18, 9,or4min whichcorrespondtototal GHhalf-livesof

63, 37, 21,

and 16min,respectively.

The lasttwovaluesbracket thephysiologicalrange forpublishedhalf-livesfortotal GHin humans.

hormone bound and alsothe percentage ofBP occupied by

hormone,thefree concentration ofligandatequilibrium isnot

controlled by the BP

affinity.

In contrast, the capacity (i.e., concentration) ofthe BPaffectstheconcentrationand percent-age of hormone bound, butdoes notmodifythe steady-state

free hormone concentrationor thepercentage of BP occupied

by its

ligand.

These general

inferences

applyqualitatively to

other

ligand-circulating

receptor interactions, under similar

simplifying

assumptions.

If the concentration of free hormone is primarily responsi-blefor hormone action intargettissues,thenfactors that con-trol thefree hormoneconcentration inplasmaare ofparticular interest.Weobserved that only the secretion rate,

_t1/2,

and

_VO

of the hormone contribute substantially to the equilibrium concentration of freeligand in plasma. BP affinity or capacity do not markedlyinfluencethe freehormone concentration at

steadystatewithinphysiological ranges of hormone secretion ratesorhalf-lives.

Indeed, complete

absenceof the BP doesnot

modifythefreehormone concentration at steady-state for any

given

secretionrate. Thisobservation has interesting

implica-tionsin theinterpretation ofhormoneconcentrations in

condi-tions ofinborndeficiencyofa BP. Forexample,inLaron

dwarf-ism, there is virtual absence of the GH-BPin plasma,while serumGH concentrations are very high ( 12, 13, 37, 38). This increase in mean serum GH concentrations indicates

mathe-maticallythat either the secretion rateand/orthe

_t1/2

ofGH are increased (orthe

VO

forGH isdecreased) in Laron dwarfs. Available data suggestthatincreased secretion ratherthan a

prolonged

t1/2

ofGHoccursinthis clinical model of GH recep-tor/BPdeficiency(39), in that the predicted (see Fig. 3) and measuredhalf-livesofGH are actually decreased in the absence of BP( 17-19, and Z.Laron, M. L.Johnson,and J. D.

Veld-huis, unpublished data).Such anincreasein GHsecretionrate presumably resultsfromthefailureof GH-dependent negative feedback mechanisms (such as generation of insulin-like growthfactorI, orIGF-I)tosuppress GH secretion (40). Of note, Laron-typedwarfism is notsimplyaGH-BP deficiency state. The fact that GH receptors are alsoabsent (38, 41) is

likelytoinfluence free and total GH

t112

independently ofthe

BP,as atleast partoftheclearanceof GH proceedsthrough receptor-mediated internalization into cellsandlysosomal

deg-radation. Thus, the combined absence of plasma GH-BP

Veldhuis aL

Free

Ja

Bound

Total

j

J.y

1

J\-h.

-j

-~~~~~~~~..

r

i

DIABETES MELLITUS CONTROL

150

100

50

0

7 5 A__

50 ~ 1

so

F

0

OBESE

A

30

20

10

L/

I

' '

A

1600 0 400 800 1200 1600 200

150 100

0

c_{0 400 800 1200 1600}

201]

II

10

0 -. ._

0 400 800 1200 1600

TIME(MIN)

Figure7.Dynamic features of GH secretion into freeandboundcompartmentsand theresultanteffectsontotalserumGHconcentrations(left panels)inahealthyyoungmale(control)anda manwithtype 1 diabetes mellitus withreduced plasmaGH-bindingprotein concentrations,

or(right panels)amiddle-agedmalecontrol and obesecounterpart. Each subjectunderwent bloodsamplingat 10-minintervals for 24 h, and the resultantsera weresubjectedtoimmunoradiometricassaytoestimate totalserumGH concentrations (lowermost panels).The GHsecretory rateovertimewasthendetermined by multiparameter deconvolution analysis(33) (uppermostpanels).Thefree and boundserumGH

con-centrationsweredefinedbyconservation ofmass,estimatedGH-binding protein capacities of respectively1.0, 0.725, and 1.7 nM in the control, diabetic,andobeseindividuals,aGH distribution volumeof 4.9liters,andanequilibriumKdof 1.5 mM.Nonlinearcurve-fitting ofthe observed totalserumGHconcentrationswascarriedout toestimate the half-life of free GH, whichwas - 3.25 mininthe_youngnormal anddiabetic

mencomparedto2.0 and 1.75 mininthemiddle-agednormaland obesemen.(Theseindividual values fall within thebroad normalrangeof

2-12min predictedinFig. 3.)The total GH half-lifewas24minin the normalyoungsubject and 16 minutes inthediabeticman,which

com-pareswith 14and 10mininthe nonobese and obesemiddle-agedvolunteers. Note that the numbers of detectableGHsecretoryepisodeswere

9 and17, respectively (control vs.diabetesmellitus),withcorrespondingGH interburstintervals of 123 and 84min,anddailyGH secretion ratesof0.573vs. 1.42mg.Similar deconvolutionestimatesofepisodicGHsecretionwereobtained in six othercontroland diabetic individuals. In the obese and normalweight middle-agedmen,therewere,respectively,5 and 9 detectable GHsecretoryburstsper24 h(interburstintervals

of 156 and 137min),anddailyGH secretion ratesof 0.013 and 0.22mg.The continuouscurvesdrawnthroughthe totalserumGH

concen-trationsarepredicted values,whichagreewell withthe observedmeasurements(also shown).

(which normally prolongst12) and tissue GHreceptors(which normallyshorten

_tl/2)

mayhave unpredictableneteffects on GH clearance in that condition(42-45).

The GH-BPcomplexis knowntoexertcomplexbiological effectson target tissues. Forexample, addition ofunoccupied BP in vitroimpairsGH action(44, 45),yetinvivo the BPcan enhance GH action inadose-dependent fashion (46,47).This

seeming paradoxis mostlikely explained bytheopposing ef-fects ofcompetition forligandon theonehand(42-44) and

prolongationof_t,,2onthe other( 17-19).Hence, subjecttothe

assumption thatunoccupied BPhasnointrinsic effects itself (44), regulationof theamountofunoccupiedBP inplasmaor intissuescanprovide aplausiblemechanism for modulating theplasmaretention time and local tissue actions of GH. Our

3

CONTROL

1^'1

A .,

A'1

2 GH

SECRETORY RATE (NMOL/MIN)

o

FREE GH

CONCENTRATION (NMOL/L)

2.5 r

2.0 1.5 1.0 0.5

0.0

L

.tA/

A

A[A/

A ,i

IK

i

-A

!~

0.50

BOUND

OH CONCENTRATION

(NMOL/L)

Ik;

0.25

0.00

TOTAL

OH CONCENTRATION

(NMOL/L)

3

2

0

0 4000 800 1200

TIME(MIN)

.,

V,

CONTROL

3 TOTAL

SERUM

GH

CONCENTRATION

(NMOL/L)

2

0

DIABETES MELLITUS

3 __.

2 I.I1

11

i~ -~~~~

PERCENTAGE

OF

GH

BOUND

100 _

80ISOI ll

60

40

20

0

100

80

60

40

20

0

0 400 800 1200 1600 0 400 800 1200 1600 0 400 800 1200 1600

TIME (MIN)

Figure8.TotalserumGH concentrationsandpercentageGH boundtoGH-BPinahealthy (control),aninsulin-dependent diabetic patient,and

anobeseman,all ofwhomunderwentblood samplingasdescribedin the legendtoFig.7.(Upperpanels)Thepredictedcurvesformeasured

totalserumGH concentrationsarereproduced from Fig.7.(Lowerpanels)ThepercentageofGH boundtoitshigh-affinityplasmaBPis de-picted.Notethat thepercentageof GHboundtothetransportprotein isnotinsteadystatebutvaries from - 10%_to80%_over24 hs. Occasional

high"plateau" regions oftheplot ofpercentageGH boundovertime denote indeterminateestimates,in thatserumGH concentrationsat these times fell below the limit ofdetection of theparticularGH IRMA(0.5 ng/ml, =0.0225 nM). This alsooccurredattheoutsetof sampling

inthe obese subject.

GH

SECRETION RATE

FREE GH

BURST-LIKE SECRETION

TIME

I _{(PLASMA VOLUME)}

METABOLIC REMOVAL FREE GH T 1/2 FREE AND/OR RECEPTOR

BINDING

+

BP

TIME

GH-BP

COMPLEX

TIME

BINDING PROTEIN (BP) KON KOFF

GH-BP COMPLEX

C? SLOW REMOVAL OF COMPLEX AND/OR BPALONE)

1

[KD

Figure9. Schemaillustrating thenonequilibrium kinetic behaviorof theGHaxis,

as-sumingthepresenceofa

high-affinity GH-BP in plasma, metabolicclearance of GHfrom thefree

com-partment,andanepisodic

burstlikemode ofGH

secre-tion.

OBESE

0.50

0.25

0.00

e"

-1

computer simulations indicate that

ceterasparibus

the amount of unoccupiedBPis controlledbythe V0 and the secretion rate of GH (at high secretion rates, there is a reduced amount of unoccupied BP), as well as the free GH half-life (if

t112

is pro-longed, fewer binding sites on BP will remain unoccupied).

Similarly, the affinity of the BP regulates the availability of

unoccupied binding sites.

Clinical correlations indicatethat the GHsecretionrateis inverselyrelated to the level of GH-BP in plasma in healthy

boys (46). Thus,children withlowBPlevels tend to havehigh GHsecretionrates, andviceversa. Our computersimulations

predict that this physiological relationship in a group of

sub-jectswouldtendtodamppulsatile freeGHavailabilityto target

tissues. Abuffering effect ofconstant amountsof high-affinity GH-BPalso occursinindividualsoverminutesand hours, as

showninFig. 8 inanormal,diabetic,and obesesubjects.In all threeconditions,the percentageofGH bound tohigh-affinity GH-BP varies remarkably over 24 h; viz., from 10% to 80%

during the course of minute-to-minute pulsatile GH release

(Fig. 8). Of particular interest,therelative retention ofGHin

plasmaby GH's complexing withBP prevents plasma GH

con-centrations fromfallingto zerobetween successive GH secre-tory bursts. Indeed, the dissociation rate constantof 0.037/

min equates to a half-life ofGH dissociation of 18.7 min, which isconsiderablylonger than ourestimated freeGH

half-life of - 7(2-12)min.Thus, the percentageof GHbound to

BP willtendtorise followingasecretory burst, asfreeGHis removed morerapidlythan bound. DissociationofGHfrom itsBP atthese times of loworabsent GHsecretion maintains

some free GH in plasmabathing thetarget tissues, which in circumstances of decreased GH secretion (obesity) may be

physiologically relevant. However, whether such low but non-zeroconcentrations of free GHorGH-BP complexare

impor-tantin

maintaining

tissuesresponsiveness toGH isnot known.

Assuming

thatfree hormone isthemajor (butnot

necessar-ilyexclusive) moiety subjecttoirreversiblemetabolicremoval, then differences in the

tp12

offree hormone among different individuals will influence the absolute amounts of free and

bound GH inplasma, and the percentage occupancy ofthe

GH-BP. Uptofourfold

variability

has been

recognized

inthe

t1/2

ofexogenously infused GH, andmorerecently of endoge-nously secreted GHinhealthy individuals (reviewed in

refer-ences 29 and33).Based on

Fig.

3,wethereforecan

predict

free GHhalf-lives of2-12min innormaladults. Thus, whether free and/or bound GH acts on target

tissues,

this

variability

in

GHt,/2

has thepotentialto

modify

tissue actions of GH

signifi-cantly.

Inobese men, thecalculated

11/2

ofendogenous GHappears to besomewhat reduced(31 ). Such calculationsassumethat the V0for GH is not

significantly

increased in

obesity.

Simi-larly, in obese orchidectomized Rhesus

monkeys,

the

meta-bolic clearancerate

of

human GH infusedat a constant rateis

significantly

higher (47).Adecrease inGHt

_1/2

(orincrease in

metabolic clearance rate)would decrease the amount of free

andboundGHbathingtargettissues,andthereby contributeto

tissueGH

deficiency

in

obesity.

DecreasedGH

11/2

isnotdueto evident alterations in BP

affinity

or capacity, sincethe

esti-mated serum GH binding

capacity

is either high-normal or moderatelyincreasedin obese men(31, 48).Thus,alternative mechanisms for the decreasein apparentGH

1/2

in obese indi-vidualsshould beconsidered,andmight include increased

re-movalofGHby receptorand/or nonreceptor-mediated

mech-anisms and/oranincreased V0 for GH.

OurmeasurementsofGH-BPconcentrationsin type I

(in-sulin-dependent) diabetes mellitus reveal a significantly re-duced GH-BP capacity (49). In conjunction with increased

GHsecretion intype Idiabetic patients, the decrease inGH-BP

capacity amplifiesthe alreadyhighfreeGHconcentration in

plasma and would tend to shorten thehalf-lifeoftotal GH (see

Fig.7,left panels). Theseobservationsareconsistentwith the

free GH hypothesis, i.e., metabolicremoval ofGH proceeds

largelyor predominantly via freeGH rather than by way of

complexes ofGH-BP.

Ourcomputersimulationsalsoprovideapossible basis for interpretingsome apparentdifferencesin thereportedGHt

1/2

in the literature.Forexample,estimatingGH

11/2

before steady stateisattainedorinresponse toinfusions of pharmacological

amountsof GH sufficient to exceed BPcapacity would

pro-duceapreferential increase in free hormoneand yield initial phase kinetic estimates reflecting predominantlythe short

t1/2

of free hormone. However, measurements of the GH

t1/2

withinthephysiologicalGH concentration range of 50 ng/liter to50

_,g/L

(- 2.3 pMto2.3 nM)wouldreflect free GHt1/2 only if the laboratory technique didnotmeasure bound

hor-mone-an unlikely possibility in the case of immunoassays (25).Indeed standardimmunochemicalassaystypically mea-suretotalGH in plasma(25). Thus,thehalf-lives derived from suchmeasurementsreflectthedamping effect oftheGH-BPs, theconcentration of whichvariesamongindividuals( 1 1, 14). Thevariability inBPlevelsmayinpartexplain individual dif-ferences in GH clearance. Ofcourse, otherfactorscan

intro-ducevariability in estimates of GH

1/2,

including measure-menttechnique,assayspecificity, bufferandmatrix, choice of

GHstandards, sampling frequency, and the extent to which endogenous GH secretion is suppressed whenthe

t1/2

of exoge-nousGH iscomputed.

Thedynamicmodelof GH secretion and GH partitioning

between BP and metabolic clearance presented here can be used topredictthetimecourseof plasma GH-BPcomplex and

free GH. Althoughthetimecourse of total, free, and bound

GHin blood can bepredicted, theexacttemporal profiles of freeand boundGHtowhich tissue GHreceptors areexposed

are notknown, sincetherelativeextent towhich free GHand

GH-BPcomplexpermeatevarioustargettissueshasnotbeen

defined. An additional unanswered

question

is: "What

time-configuration

of free GH (or GH-BP complex) inbloodmost

effectively

initiates ormodifies tissue responses toGH?" For

example,undersome

conditions,

a

pulsatile

rather thana

con-tinuous mode of GH

delivery preferentially

enhances tissue

responses, such aslinear growthandweight gainintheratand theinductionof

epiphyseal cartilage

and muscle IGF-I mRNA

(50). In contrast, in

acromegaly,

thenadir (ormean or

inte-grated)

serumGH concentration isthe strongest

single

corre-lateofserum IGF-Iconcentrations (36).Moreover, the induc-tionofa"feminized"patternof certain

hepatic

enzymes

(e.g.,

carbonic anhydrasein therat) isbestachieved

by

continuous

GHdelivery (51 ). Near-continuous GH

replacement

may also be

optimal

for

generating

IGF-I inGH-deficienthumans

(52).

In conclusion,

assuming

that plasma GH concentration profiles are transferred to tissue receptor

sites,

the

dynamic

metabolic clearance offers significant regulatory controlover tissue GH actions within the somatotropic axis in health and disease.

Acknowledgments

Wethank Patsy Craigfor her skillfulpreparationof themanuscript, Drs.W. S.Evansand C.M.Asplinforprovidingthe 24-hserumGH profilesin type I diabetesmellitus,and Paula P. Azimi for the artwork. This workwassupportedin partby National Institutes of Health grant RR 00847tothe Clinical Research Center of theUniversityof Virginia, RCDA 1 K04 HD00634 (to Dr. Veldhuis),GM-28928 (to Dr. Johnson), DK-38128 (toDr.Baumann),the Diabetesand Endo-crinologyResearch Center grant NIH DK-38942,theNational

Insti-tutes-supported Clinfo DataReduction Systems, the Pratt Founda-tion, the UniversityofVirginia Academic Enhancement Fund, the National Science Foundation Science Center for Biological Timing, and the NorthwesternMemorialFoundation.

References

1.Baumann,G., M. W. Stolar, K. Amburn, C. P. Barsano, and B. C. De Vries. 1986.Aspecificgrowthhormone-bindingprotein in human plasma: initial char-acterization.J. Clin.Endocrinol. Metab. 62:134-141.

2.Herington, A. C., S. Ymer, and J. Stevenson. 1986. Identification and

characterisationofspecific bindingproteinsforgrowthhormone in normal

hu-man sera.J.Clin.Invest.77:1817-1823.

3.Ymer,S. I., and A. C. Herington. 1985.Evidencefor thespecific binding of

growth hormonetoareceptor-like proteinin rabbitserum.Mol.Cell. Endocrinol.

41:153-161.

4.Spencer,S. A., R. G.Hammonds, andW.J.Henze,H.Rodriquez,M. J. Waters, andW.I. Wood. 1988. Rabbit liver growth hormone receptor andserum bindingprotein: purification, characterization and sequence. J. Biol. Chem.

263:7862-7867.

5. Leung, D. W., S. A. Spencer, G. Cachianes, R. G. Hammonds, C. Collins, W. J.Henzel, R. Barnard,M.J. Waters, and W. I. Wood. 1987. Growth hormone receptor andserumbindingprotein: purification, cloning,andexpression.Nature (Lond.).330:537-543.

6.Smith,W.C.,J.Kuniyoshi,and F.Talamantes.1989. Mouseserumgrowth

hormone(GH) bindingproteinhasGH receptor extracellular and substituted transmembrane domains. Mol.Endocrinol.3:984-990.

7.Baumbach,W.R.,D. L.Horner,and J.S. Logan. 1989. Thegrowth

hor-mone-binding proteinin rat serum isanalternativelyspliced formof therat

growth hormone receptor. GenesDev.3:1199-1205.

8.Baumann,G.,and M.A.Shaw.1988. Immunochemicalsimilarity ofthe humanplasma growthhormone-binding proteinandtherabbit liver growth hor-mone receptor.Biochem. Biophys.Res.Commun. 152:573-578.

9.Barnard, R., P. Quirk, and M.J.Waters. 1989.Characterization of the growth hormone-bindingprotein ofhuman serumusingapanelofmonoclonal antibodies.J. Endocrinol. 123:327-332.

10. Baumann, G., andM.A.Shaw. 1990. A second, lower affinity growth

hormone-bindingprotein in human plasma. J.Clin. Endocrinol.Metab. 70:680-686.

11.Baumann, G.,M.A.Shaw, and K. Amburn. 1989. Regulation of plasma growth hormone-binding proteins in health and disease. Metab. Clin. Exp. 38:683-689.

12. Daughaday, W. H., and B. Trivedi. 1987. Absence of serum binding protein in patients with growth hormone receptor deficiency (Laron dwarfism). Proc.Natl. Acad.Sci. USA. 84:4636-4640.

13. Baumann, G., M. A. Shaw, and R. J. Winter. 1987. Absence ofthe plasma growthhormone-binding protein in Laron-type dwarfism.J.Clin. Endocrinol. Metab. 65:814-816.

14.Daughaday, W. H., B. Trivedi, and B. A. Andrews. 1987. The ontogeny of

serumGHbindingprotein in man: a possible indicator of hepatic GH receptor development.J.Clin. Endocrinol. Metab. 65:1072-1074.

15.Baumann, G., M. A. Shaw, and T. J. Merimee. 1989. Low levels of

high-affinitygrowth hormone-binding protein in African pygmies. N. Engl. J. Med.

320:1705-1709.

16. Baumann, G., M. A. Shaw, R. C. Brumbaugh, and J. Schwartz. 1991.

Short stature anddecreased growthhormone-binding protein in the Mountain Ok People of Papua New Guinea.J.Clin. Endocrinol. Metab. 72:1346-1349.

17.Baumann, G., M. W. Stolar, and T. A. Buchanan. 1987. The effect of

circulatinggrowth hormone binding protein on metabolic clearance, distribution

anddegradation ofhuman growth hormone. J. Clin. Endocrinol. Metab. 64:657-660.

18. Baumann,G., M. A. Shaw, and T.A.Buchanan.1989. In vivo kineticsof

acovalent growth hormone-binding protein complex. Metab. Clin. Exp. 38:330-333.

19. Moore, J., R. Vandlen, P. MacKay, and S. A. Spencer. 1988. Serum clearance of human growth hormone boundtogrowth hormone bindingprotein. Endocrinology.122(Suppl.):121.(Abstr.)

20. Barsano, C. P., and G. Baumann. 1989.Simple algebraicandgraphic

methods for the apportionmentofhormone(andreceptor) into bound and free fractions in binding equilibria; or how to calculate bound and free hormone?

Endocrinology. 124:1101-1106.

21.Veldhuis, J. D., andM.L.Johnson. 1992.Computer-assisted algorithms

toanalyzepulsatilegrowth hormone release. Excerpta Med. Int.Congr.Ser. In press.

22. Baumann,G., K. Amburn, andM. A.Shaw. 1988. Thecirculatinggrowth hormone (GH)-bindingproteincomplex: amajorconstituentofplasma GH in man.Endocrinology. 122:976-984.

23. Baumann,G., J. G.MacCart, and K. Amburn. 1983.Themolecular natureofcirculating growth hormone in normal andacromegalicman:evidence foraprincipaland minor monomericforms.J.Clin. Endocrinol. Metab. 56:946-952.

24. Baumann, G. 1991. Growth hormoneheterogeneity:genes,isohormones,

variants and binding proteins.EndocrineRev. 12:424-449.

25. Jan, T., M. A.Shaw, and G. Baumann. 1990.Effectsofgrowth

hormone-binding proteinson serumgrowth hormonemeasurements.J.Clin. Endocrinol. Metab.72:387-391.

26. Snow, K. J., M. A. Shaw, L.M.Winer, and G. Baumann. 1990. Diurnal patternofgrowth hormonebinding proteininman.J.Clin.Endocrinol. Metab. 70:417-420.

27. Carlsson, L., S. Rosberg, W. L. Wong, and K. Albertsson-Wikland. 1991. Analyses of 24-hour plasmaprofiles ofgrowth hormone (GH) binding protein in healthychildren.Horm. Res.35:63.(Abstr.)

28.Owens, D., M. C. Srivastava, C.V.Tompkins,J. D.N.Nabarro, and P. H. Sonksen. 1973. Studiesonthemetabolic clearance rate, apparent distribution space and plasmahalf-disappearancetime of unlabelled humangrowth hormone in normalsubjectsandinpatientswith liverdisease,renaldisease, thyroiddisease anddiabetes mellitus. Eur. J.Clin.Invest.3:284-294.

29.Faria,A.C.S.,J. D.Veldhuis,M.0.Thorner,and M. L. Vance. 1989.

Half-timeof endogenous growth hormone (GH)disappearancein normalman after stimulation of GH secretion byGH-releasinghormone andsuppression

withsomatostatin. J. Clin.Endocrinol.Metab. 68:535-541.

30.Iranmanesh,A., G.Lizarralde,and J. D.Veldhuis. 1991. Age and relative

adiposityarespecific negativedeterminantsof thefrequencyandamplitudeof GHsecretorybursts andthehalf-life of endogenousGH inhealthymen.J.Clin. Endocrinol. Metab.73:1081-1088.

31.Veldhuis, J. D.,A.Iranmanesh,K. K. Y.Ho, G. Lizarralde, M. J. Waters, and M.L.Johnson. 1991. Dualdefectsinpulsatile growthhormone secretion and clearance subserve thehyposomatotropism of obesityinman. J. Clin. Endo-crinol. Metab.72:51-59.

32.Ulloa-Aguirre,A., R. M.Blizzard,E.Garcia-Rubi,A.D.Rogol,K.Link,

C.M.Christie,M. L.Johnson,and J. D.Veldhuis. 1990. Testosterone and

oxan-drolone,anon-aromatizable androgen,specifically amplifythemassandrateof

growthhormone(GH)secreted per burstwithoutalteringGHsecretoryburst durationoffrequencyorthe GHhalf-life. J.Clin. Endocrinol. Metab. 71:846-854.

33. Hartman, M. L.,A.C. S. Faria, M. L. Vance, M. L. Johnson, M.0.

Thorner, and J. D.Veldhuis. 1991. Temporal structure of in vivo growth hor-monesecretoryevents in man.Am. J.Physiol.260:E101-E110.

34.Shampine,L.F., H.A.Watts, and S. M. Davenport. 1976. Solving

non-stiffordinarydifferentialequations-thestateof theart.SIAMRev. 18:376-441. 35.Winer,L.M., M.A.Shaw, and G. Baumann.1990. Basal plasma growth hormonelevels in man: newevidencefor rhythmicity of growth hormone secre-tion.J.Clin.Endocrinol. Metab. 70:1678-1686.

36. Hartman, M. L.,J. D.Veldhuis,M. L.Vance, A. C. S. Faria,R. W.

Furlanetto, andM.0.Thorner.1990. Somatotropin pulse frequency and basal growth hormone concentrations are increased in acromegaly.J.Clin. Endocrinol. Metab. 70:1375-1384.

37.Silbergeld,A.,I.Lazar,B.Erster, R. Keret, R. Tepper, and Z. Laron. 1989. Serum growth hormonebindingprotein activity in healthy neonates, children and young adults: correlation with age, height and weight. Clin. Endocrinol.

31:295-303.

38. Eshet, R.,Z.Laron,A.Pertzelan,R.Arnon,and M.Dintzman. 1984. Defectof human growth hormone receptors in the liver of two patients with Laron-type dwarfism.Isr.J.Med.Sci. 20:8-11.

39.Keret, R.,A.Pertzelan,A.Zeharia,Z.Zadik,andZ.Laron. 1988. Growth hormone(hGH) secretionandturnover inthreepatients with Laron-type dwarf-ism.Isr.J.Med. Sci. 24:75-79.

40.Berelowitz, M.,M.Szabo,L. A.Frohman,S.Firestone,and L.Chu. 1981.

Somatomedin-Cmediatesgrowthhormonenegativefeedback by effects on both thehypothalamus and thepituitary.Science(Wash.DC). 212:1279-1281.

41.Godowski,P.J., D. W. Leung, L. R. Meacham, J. P.Galgani,R.Hellmiss, R.Keret,P.S.Rotwein,J.S.Parks, Z. Laron, and W. I. Wood. 1989. Character-izationof the human growth hormone receptor gene and demonstration of a partial gene deletion intwopatientswith Laron-type dwarfism. Proc. Natl. Acad. Sci. USA.86:8083-8087.

42. Lim, L., S. A. Spencer, P. McKay, and M. J. Waters. 1990. Regulation of growth hormone (GH)bioactivitybyarecombinant human GH-binding pro-tein.Endocrinology.127:1287-1291.

43. Mannor, D. A., L. M. Winer, M. A. Shaw, and G. Baumann. 1991. Plasma growth hormonebinding proteins: effectongrowth hormonebindingtoreceptors and on growth hormoneaction.J.Clin. Endocrinol.Metab.73:30-34.

44. Mannor, D. A., M. A. Shaw, L. M. Winer, and G. Baumann. 1988. Circulating growth hormone-bindingproteins inhibit growth hormone (GH) binding to GH receptors but not in vivo GH action. Clin. Res. 36:870a. (Abstr.) 45.Clark, R. G., B.Cunningham,J. A.Moore, M. G. Mulkerrin, L. M. S. Carlsson, S. A. Spencer, W. I. Wood, and M. J. Cronin. 1991. Growth hormone

binding proteinenhances thegrowth promoting activity ofGH in the rat. In 1991 Program, 73rdMeetingof the Endocrine Society. 1611. (Abstr.)

46. Martha,Jr., P. M.,A.D. Rogol, R. M. Blizzard, M. A. Shaw, and G. Baumann. 1991. Growth hormone-binding protein activity is inversely related to 24 hour growth hormone release in normal boys. J. Clin. Endocrinol. Metab. 73:175-181.

47.Dubdey, A. K.,A.Ahanukoglu, B. C. Hansen, and A. A. Kowarski. 1988.

Metabolicclearance ratesofsynthetic human growth hormone in lean and obese malerhesusmonkeys. J.Clin.Endocrinol. Metab.67:1064-1067.

48.Hochberg,Z., P. Hertz, V. Colin,S.Ish-Shalom,D. Yeshurun, M. B. Youdim, and T. Amit. 1992. The distalaxis ofgrowth hormone (GH) in nutri-tional disorders: GH-binding protein, insulin-like growth factor-I (IGF-1), and

IGF-Ireceptors in obsesity and anorexia nervosa.Metab. Clin.Exp.41:106-112. 49. Mercado, M., M. E. Molitch, and G. Bauman. 1992. Low plasma growth hormone binding protein in insulin-dependent diabetes mellitus. Diabetes.

41:605-609.

50. Isgaard, J., L. Carlsson,0.G. P. Isaksson, and J.0.Jansson.1988. Pulsa-tile intravenousgrowth hormone (GH) infusion inhypophysectomizedrats in-creases insulin-likegrowth factorImessenger RNA in skeletal tissuesmore effec-tively thancontinuousGH infusion.Endocrinology. 123:2605-26 10.

51.Jeffrey,S., N. D. Carter, R. G. Clark, and I. C. Robinson. 1990. The

episodic secretorypatternofgrowthhormoneregulateslivercarbonic anhydrase. III. Studies in normal and mutant growthhormone-deficient dwarfrats.

Bio-chem. J.266:69-74.

52. Jorgensen, J.0.L., N. Moller, T.Lauritzen,and J. S.Christiansen. 1990. Pulsatile versus continuous intravenous administration of growth hormone (GH) inGH-deficient patients: effectsoncirculatingIGF-I and metabolic

indi-ces.J.Clin.Endocrinol. Metab.70:1616-1623.

53.Cunningham,B.C.,M.Ultsch,A.M.DeVos,M.G.Mulkerrin, K. R.

Clauser, and J.A.Wells.1991.Dimerization of the extracellular domain of the human growth hormone receptor by a single hormone molecule. Science(Wash.