人体生理学 (9).pdf
Abstract The possibility that the action of growth hor-mone(GH)on cartilage is mediated by a separate hor-monal agent found in serum was suggested by incubationwith hypophysectomized rat costal cartilage.The stabili-ty of this tissue permitted long incubations and the mea-surement of the uptake of35S-sulfate provided a conve-nient index of growth stimulation.Under the conditionsarbitrarily selected,normal rat serum,but not serumfrom hypophysectomized rats,induced a great stimula-tion of 35S uptake.In contrast,GH added directly to car-tilage in these incubations was virtually inactive.It wassuggested that a serum sulfation factor,now known asinsulin-like growth factor-I(IGF-I),was a mediator ofGH action.Recently it has been observed that addition of35S-sulfate after 24 h of preincubation with GH permit-ted the direct effect of GH to be recognized.Other obser-vations in intact hypophysectomized rats have estab-lished that GH can induce the expression of IGF-I in car-tilage that acts in an autocrine-paracrine manner.The rel-ative importance of the endocrine and autocrine-para-crine routes of IGF-I action on the growth of cartilage isin dispute.It is clearly established that serum IGF-I ex-erts a negative feedback on GH secretion by action onthe hypothalamus and pituitary.Serum IGF-I concentra-tions reflect GH action in postnatal life.Measurement ofserum IGF-I is the most-valuable index of GH hyperse-cretion in acromegaly and in conditions of growth im-pairment.GH receptor deficiency leads to a marked de-crease in circulating IGF-I.Hypernutrition and hyperin-sulinism of obesity directly promote hepatic IGF-I re-lease and inhibit GH secretion by the pituitary.Differ-ences in hepatic IGF-I synthesis in response to GH maycontribute to physiological differences in stature.Key words Growth hormone Insulin-like growth factor-I Cartilage GrowthInsulin-like growth factor-I is an endocrine mediatorof the action of growth hormone on cartilageIn 1957 we 1 began searching for a test system to studythe anabolic effects of growth hormone(GH)on tissuesin vitro.We chose rat cartilage because of the essentialrole that epiphyseal cartilage plays in skeletal growth,and because the action of GH on epiphyseal cartilage ofhypophysectomized rats had provided the most-sensitivebioassay for GH 2.Because epiphyseal cartilage cannotbe easily dissected,we chose to conduct our studies onisolated costal cartilage segments.This was a fortunatechoice because cartilage has a number of properties thatmake it ideal for in vitro studies.Costal cartilage con-tains essentially a single cell type.It is devoid of vascu-lar,lymphatic,and neural elements.Unlike dissociatedchondrocyte preparations,isolated costal cartilage seg-ments contain chondrocytes with their normal cell-celland cell-matrix connections.The dense cartilage matrix,composed of collagen and the huge hyaluronan complex,acts as a barrier to the diffusion of molecules of the sizeof hemoglobin 3,and probably would also hinder thediffusion of insulin-like growth factors(IGFs)bound totheir binding proteins,but provides little barrier to thediffusion of peptides such as the IGFs and insulin.Another favorable property of cartilage is that it has avigorous anaerobic metabolism that makes it virtually in-dependent of diffused oxygen.This is not true of liverslices and many other tissue preparations that becomeunstable after only a few hours of incubation,even in thepresence of oxygen.Cartilage contains ample substratefor prolonged anaerobic glycolysis,but amino acids inthe medium are required for hyaluronan synthesis.Despite its inactive appearance,cartilage maintainsvigorous protein synthesis.Collagen synthesis is slow,but turnover is extremely slow.The situation is quite dif-ferent with hyaluronan,which is a huge macromoleculeW.H.DaughadayDivision of Endocrinology and Metabolism,Department of Medicine,University of California,Irvine,California,USAW.H.Daughaday(u)Box 157,Balboa Island,CA 92662,USAW.H.Daughaday530 S.Bay Front,Balboa Island,CA 92662,USAPediatr Nephrol(2000)14:537540 IPNA 2000GROWTH HORMONE INSULIN-LIKE GROWTH FACTORAXIS/REVIEW ARTICLEWilliam H.DaughadayGrowth hormone axis overview somatomedin hypothesisReceived:5 March 1999/Revised:16 December 1999/Accepted:26 December 1999with a greatly extended hyaluronic acid chain to whichcore proteins are attached.Each core protein carries hun-dreds of chondroitin sulfate and keratin sulfate adducts4.The entire macromolecule has a size of 23 milliondaltons.Despite its size,this is not a static molecule.Inskin the half-life of the complex is 23 weeks 5.It islikely that the turnover is equally fast in cartilage.Sul-fate incorporation is dependent on the synthesis of theentire macromolecule 6.In conclusion,the measure-ment of sulfate incorporation provides a convenientquantitative measure of the synthesis of this importantconstituent of cartilage.The conditions,which we selected more or less arbi-trarily,included incubation in a simple KPS phosphatebuffer,with the addition of 35S at the start of a 24-h incu-bation 1.We found that as little as 5%of normal rat se-rum produced a 200%300%stimulation of uptake of ra-dioactivity.In contrast,serum from hypophysectomizedrats was virtually inactive.An unexpected but crucial finding was that bovineGH added to the incubation medium produced little orno simulation of uptake of radioactivity,even when add-ed to serum from hypophysectomized rats.GH treatmentof hypophysectomized rats restored the ability of theirserum to stimulate uptake of 35S-sulfate.We postulatedthat the simulation of sulfate uptake was due to a GH-dependent factor,which we called“sulfation factor.”Thename was later changed to“somatomedin”and finally to“insulin-like growth factor I”when its primary structurewas established 7,8.Our findings with hypophysec-tomized rat cartilage were confirmed and extended byAlmqvist et al.9,with chick embryo cartilage by Hall10,and with porcine cartilage by Van den Brande andDu Caju 11.It was fortunate that we chose to do our initial incuba-tions with rat serum rather than with human serum,be-cause rat serum contains far more IGF-I than human se-rum.Measurements of adult rat serum total IGF-I withhomologous radioimmunoassays have given average lev-els of 100 nmol/l120 nmol/l,which is 4 to 5 times theconcentration in young adult humans 12,13.The dif-ference in the concentration of free IGF-I between ratand human serum is even greater.Frystyk et al.14found the concentration of free IGF-I in rat serum to beabout 8 nmol/l,whereas in young adult human serum theconcentration is 0.2 nmol/l0.3 nmol/l 15,16.As freeIGF-I has the greater access to the IGF-I receptor onchondrocytes,it explains why rat serum proved to besuch a potent stimulator of sulfate uptake by rat carti-lage.The effect of rat serum is due almost entirely toIGF-I,because there is little IGF-II in rat serum.The conditions of incubation that were initially select-ed permitted recognition of the effects of IGF-I withoutsignificant effects of GH.In a recent report,Salmon andBurkhalter 17 reexamined the effects of GH onhypophysectomized rat cartilage.They preincubated car-tilage with either IGF-I or GH in a HEPES buffer en-riched by amino acids for 24 h before the addition of ei-ther 35S-sulfate or methyl-3H thymidine for 2 h.Underthese conditions GH stimulated both parameters,but theeffects were delayed in onset and smaller in magnitudethan observed with IGF-I.Both the effects of GH andIGF-I on hypophysectomized rat cartilage in vitro wereblocked by the addition of IGF-I antibody to the medi-um,indicating that GH was acting by the local produc-tion of IGF-I.These observations are consistent with a large body ofevidence that GH can stimulate cartilage growth by localaction in vivo and in vitro 18.Much of the local effectof GH is mediated by IGF-I,but it suggested that GHmight stimulate mitogenesis in chondrocyte progenitorsdirectly.Local production of IGF-I occurs in many othertissues of the body,and the relative roles of IGF-I as anendocrine agent and as an autocrine-paracrine agent mayvary from tissue to tissue.IGF-I and the endocrine regulation of GH secretionThe regulation of GH secretion is controlled by two prin-cipal neurohypophyseal hormones,GH releasing hor-mone(GRH)and somatostatin(SS).The secretion ofGRH and SS are further modified by many neuropep-tides,catecholamines,steroid hormones,and metabolitesacting on the hypothalamus.In humans,GH is secretedin six to eight bursts per day with intervening periods ofnegligible secretion.After fetal life,GH acts on the liverthrough its receptor to promote the synthesis and secre-tion of IGF-I into the bloodstream.Less IGF-I is secret-ed by other tissues.In turn IGF-I reaches the hypothala-mus and pituitary through the circulation and inhibitsGH secretion.This regulatory circuit has relatively slug-gish time characteristics that set the magnitude of the se-cretory bursts.GH can also inhibit its own secretion by ashort feedback loop acting on the hypothalamus.Tomake the regulation of GH secretion more complicated,both GH and IGF-I are synthesized in the hypothalamusand probably contribute to the regulation of GH secre-tion.During human fetal life,concentrations of serum GHare high and not under the feedback control of IGF-I,which is present in low concentrations and apparentlylargely independent of GH 19.Until the last weeks ofgestation,fetal growth is independent of fetal GH.Themechanisms controlling GH-independent IGF-I synthesisand secretion in fetal life remain to be determined.It isof interest that GH-independent secretion of IGF-I andgrowth persists throughout life in the guinea pig 20.Impairment of IGF-I secretion after the neonatal peri-od results in increased GH secretion.This is most evidentin children with a genetic absence of the GH receptor(Laron dwarfism)21,22.In the prepubertal children re-ported from Ecuador,serum IGF-I was very low(1 g/l7 g/l)despite greatly increased serum GH con-centrations.In postpubertal children,a modest increase inserum IGF-I was found(3.4 g/l86 g/l).It is possiblethat changes in sex steroids or body composition accountfor this modest increase in IGF-I secretion by the liver.538A more-selective defect in hepatic GH receptor ex-pression has recently been recognized in Bovis indicus.Short stature in these cows appears to be inherited as arecessive trait and results in a 30%decrease in size.Se-rum GH levels are greatly elevated and serum IGF-I isreduced 22,23,24,25.My laboratory found that hepat-ic membranes prepared from these small calves had a re-duced ability to bind labelled GH(unpublished observa-tions).Liu et al.(personal communication)found no se-quence abnormalities in the GH receptor gene.However,the total hepatic GH receptor mRNA was greatly de-creased as a result of the virtual absence of the normallydominant isoform,GH-R 1A.These findings suggestthat the inducible hepatic GHR-1A isoform is essentialfor normal IGF-I synthesis,and provide further supportfor the view that IGF-I acting as an endocrine agent isessential for normal growth in Bovis indicus.Functional defects in hepatic IGF-I synthesis occur inmalnutrition 26 and uncontrolled diabetes mellitus27,and lead to low serum IGF-I and increased serumGH concentrations.Estrogens,even at physiologicalconcentrations in menstruating females,have an inhibi-tory effect on IGF-I secretion and account for the slight-ly higher serum GH concentrations in females 28.Inpregnancy,despite the greatly elevated concentrations ofplacental GH in the serum 29,serum IGF-I concentra-tions are only modestly elevated because of the extreme-ly high concentrations of estrogens in pregnant women.Only very mild clinical manifestations of GH excess areobserved in pregnant women.In severe loss of liver pa-renchyma in cirrhosis,serum IGF-I concentrations areoften reduced and GH concentrations are elevated 30.Many studies have shown that GH secretion is greatlysuppressed in human obesity,yet serum IGF-I levels arenormal or slightly increased 31.It is likely that GH-independent secretion of IGF-I is promoted by the hyper-insulinism and hypernutrition of obesity.This in turnleads to suppression of GH secretion.Androgens as opposed to estrogens appear to promoteIGF-I secretion.During puberty there is an increase inGH secretion in boys,as measured by mean GH serumconcentrations,but an even greater relative increase inserum IGF-I concentration 32.Serum IGF-I in disorders of growthSerum IGF-I is an important determinant of normalgrowth.The role of serum IGF-I in growth was studiedby Hall and Filipsson 33 in 1975 using the chick em-bryo cartilage bioassay.They noted a significant rela-tionship between height velocity and somatomedin levelsin 263 children and adolescents with normal GH secre-tion after provocative testing.At a given bone age or pu-bertal stage,tall children had significantly higher IGF-1levels than short subjects.Within each height group therewas a significant variance of somatomedin levels,sug-gesting differences in tissue responsiveness.Binoux andGourmelen 34 noted that growth in height parallelledIGF-I levels in subjects of constitutionally variant stat-ure.Another large study of the relationship between se-rum IGF-I and growth in healthy children,adolescents,and adults was published by Juul et al.32 in 1994.Se-rum IGF-I was significantly correlated with height ve-locity in the following year for groups of comparable pu-bertal stage,and the age at which a given pubertal stagewas reached.An interesting study by Kao et al.35found that serum IGF-I concentrations of identical twinswere highly correlated and were correlated to a lesser ex-tent with difference in height of the twins.Measurements of serum IGF-I concentrations in shortyoung children with suspected GH deficiency are fre-quently within normal limits.The discrimination is bet-ter in older children.It is likely that most children who are diagnosed asGH deficient retain some ability to secrete GH.The di-agnosis of GH deficiency is based on an arbitrary inter-pretation of GH provocative tests.Retesting many suchpatients after puberty has shown that a substantial num-ber regain“normal”GH secretion 36.Serum IGF con-centrations are more uniformly low in patients with inac-tive GH gene or GH receptor genes.Serum IGF-I measurements have provided the best in-dication of iatrogenic and acromegalic hypersomatotro-pism.There is a predictable relationship between the logof the mean GH concentration and the serum IGF-I con-centration 37.Determination of the mean GH concen-tration requires many measurements over time and re-mains essentially a research procedure.In conclusion,the evidence cited supports the somato-medin hypothesis that IGF-I is an important endocrineregulator of skeletal growth.References1.Salmon WD,Daughaday WH(1957)A hormonally controlledserum factor which stimulates sulfate incorporation by carti-lage in vitro.J Lab Clin Med 49:8258362.Greenspan FS,Li CH,Simpson ME,Ev
