- Email: [email protected]
Localization of binding sites for insulin-like growth factor-I (IGF-I) in the rat brain by quantitative autoradiography
Brain Research, 444 (1988) 205-213 Elsevier
205
BRE 13386
Research Reports
Localization of binding sites for insulin-like growth factor-I (IGF-I) in the rat brain by quantitative autoradiography •
?
Nancy J. Bohannon 3, Eric S. Corp 4"*, Barbara J. Wilcox-, Dianne P. Figlewicz4, Daniel M. Dorsal'5 and Denis G. Baskin 1-3 t Division of Endocrinology~Metabolism, Veterans Administration Medical Center, Seattle WA 98108 (U.S.A.) and Departments of 2Medicine, 3BiologicalStructure, 4Physiologyand SPharmacoiogy, University of Washington, Seattle, WA 98195 (U.S.A.) (Accepted 8 September 1987) Key words: Insulin-like growth factor-1; Quantitative autoradiography; Brain; Rat; Receptor
In vitro quantitative autoradiography was used to localize IGF-I binding sites in rat brain. Slide-mounted sections of frozen rat brain were incubated in 0.01 nM 125I[ThrSg]IGF-I, alone or mixed with 10 nM unlabeled [Thrsg]IGF-I or insulin, for 22 h at 4 °C and apposed to LKB Ultrofilm. Measurement of labeled [Thrsg]IGF-I binding by computer digital image analysis of the autoradiographic images indicated that high affinity IGF-I binding sites are widely distributed at discrete anatomical regions of the brain microarchitecture. The highest concentration of specific binding sites was in the choroid plexus of the lateral and third ventricles. Unlabeled porcine insulin was less potent than unlabeled IGF-I in competing for binding sites on brain slices. Regions of the olfactory, visual, and auditory, as well as visceral and somatic sensory systems were labeled, in particular the glomerular layer of the olfactory bulb, the anterior olfactory nucleus, accessory olfactory bulb, primary olfactory cortex, lateral-dorsal geniculate, superior colliculus, medial geniculate, and the spinal trigeminal nucleus. High concentrations of IGF-I-specific binding sites were present throughout the thalamus and the hippocampus, (dentate gyrus, Ca1, Ca2, Ca3). The hypothalamus had moderate binding in the paraventricular, supraoptic, and suprachiasmatic nucleus. Highest binding in the hypothalamus was in the median eminence. The arcuate nucleus showed very low specific binding, approaching the levels found in optic chiasm and white matter regions. Layers II and VI of the cerebral cortex also had moderate IGF-I binding. The results suggest that the development and functions of brain sensory and neuroendocrine pathways may be regulated by IGF-I. INTRODUCTION Insulin-like growth factor-I (IGF-I) is a somatomedin (moi. wt. approximately 7000) with insulin-like structure and biological properties, which is made in the liver and circulates in plasma bound to carrier proteins 17'22. A peptide with immunological properties of IGF-I has been extracted from brain and cerebrospinal fluid 1'20'28A0. Furthermore, receptors for IGF-I have been identified in membrane preparations of rat brain 22'26'27, and human adult 26'39 and fetal brain as. Synthesis of IGF-I by the brain is suggested by reports of IGF-I-like m R N A in rat brain 33'45,
and the production of IGF-I in cultures of brain cells9. These observations support the hypothesis that the brain is an IGF-I-sensitive organ. Indeed, IGF-I has been shown to stimulate the release of somatostatin from hypothalamic cultures 8 and to stimulate D N A synthesis in fetal rat brain cultures 3°. The sites of action of IGF-I in the brain are largely unknown, however. We have approached this problem by identifying the location of IGF-I specific binding sites with autoradiography. In an earlier study, we found a high density of specific binding sites for IGF-I in the rat median eminence ~3. In the present study, we have mapped the location of specific bind-
* Present address: Bourne Research Laboratory, Cornell University Medical Center, White Plains, NY 10605, U.S.A. Correspondence: D.G. Baskin, Division of Endocrinology/Metabolism (151), VA Medical Center, 1660 S. Columbian Way, Seattle, WA 98108, U.S.A. 0006-8993/88/$03.50 © 1988 Elsevier ~,. :nce Publishers B.V. (Biomedical Division)
206 ing sites for an iodinated analogue of IGF-I in slices of rat brain, using autoradiography with LKB Uitrofilm. The relative concentrations of IGF-I binding sites "n discrete anatomical locations were determined by computer digital image densitometry 4'5. A portion of these results were reported as preliminary data in an abstracd 2. MATERIALSAND METHODS Male Sprague-Dawley rats (200-250 g) were anesthetized with sodium pentobarbital and perfused with ice-cold saline via the left cardiac ventricle. The brains were removed, frozen in crushed dry ice and stored a t - 7 0 °C for up to 14 days. Coronal cryostat sections (15/~m) were mounted on gelatin coated slides and stored at-70 °C for a maximum of 4 weeks. Brain IGF-I receptors have been shown to be stable at-20 °C for up to a year 1°. The tracer peptide for these studies was a recombinant form of human IGF-I, [ThrSg]IGF-I (Amersham), which has been widely used for analysis of rat and human IGF-I receptors -;6. ~2-SI[Thr'~gllGF-I(spec. act. -- 267 pCi/pg) was dissolved to a final concentration of 0.01 nM (10 pmol/ml) in 10 mM Hepes assay buffer containing 0.5% BSA, 0.025% bacitracin, 0.0125% N-ethylmaleimide and 100 K.I.U./ml Aprotinin (Sigma), pH 7.64"23. Each brain slice was covered with 0.1 ml of labeled [ThrSg]IGF-I either alone or mixed with 10 nM unlabeled [Thr59]IGF-I (AmGen) or porcine insulin (Lilly), and incubated for 22 h at 4 °C. The slides were then rinsed in 3 one-rain
E• TG
~
changes of ice-cold assay buffer, immersed in a final one-min rinse of 0 °C distilled water, and rapidly dried. We have previously worked out equilibrium conditions using brain slices and found that specific binding of labeled [ThrS'~]IGF-I to brain slices reaches equilibrium by 12 h at 4 °C, is unchanged at 24 h, and is reversible II. The procedures for quantitative autoradiography are described in detail elsewhere 4'5'2z. Briefly, the labeled and dried brain slices were placed in direct contact with the emulsion of LKB Ultrofilm in an Xray film cassette for 72 h. The film was then developed for 4 min in D-19 at 20 °C. Quantitative analysis of optical density in small, defined anatomic regions of the autoradiographic images was made with the DUMAS computer digitizing system and BRAIN program for quantitative autoradiography 24. Optical 6ensities were converted to DPM/100 mm 2 of the section area using radioactivity standards calibrated for 1251radioactivity in tissue slices and a 3-day exposure period 4's. To obtain specific binding, radioactivity bound in the presence of 1000-fold excess unlabeled peptide (non-specific binding) was subtracted from binding with labeled peptide alone (total binding). Anatomical regions were identified with a stereotaxic atlas of rat brain 32. RESULTS qinding of 125I[Thrsg]lGF-I was heterogeneously distributed among discrete anatomical regions of the rat brain (Figs. 1-3). The pattern of localization that
AQ
'
G -
L
Fig. 1. Autoradiographic images of olfactory bulb coronal slices labeled with t2SI[Thr59]IGF-I.A: rostral olfactory bulbs. B: level of accessory olfactory bulbs.
207
,dl ,/I,' ,
,~
.
".~.~'~.V:
Fig. 2. Binding of 125I[Thr59]lGF-Ito coronal brain slices at levels of caudate-putamen (A), suprachiasmatic nucleus (B). paraventricular nucleus (C), median eminence (D), medial geniculate (E), and pontine nuclei (F). we describe in this paper has been observed in brains of more than 40 rats and with 6 separate batches of ]25I[ThrS9]IGF-I. The concentrations of IGF-l-specific binding in 46 different brain regions when 0.01 nM 1251[ThrSg]IGFI was used are shown in Table I. The quantitative results reported here represent measurements made on the brains of 3 rats which were processed and assayed together. Non-specific binding of 125I[ThrS9]IGF-I was uniformly distributed and averaged 4.8 + 1.0
fmol/100pm 2 for all regions combined. Thus the total binding shown in autoradiographic images (Figs. 1-3) reflects the relative differences in specific binding among the brain regions which were labeled. Throughout the cortex, layers II and VI were heavily labele0, with specific binding of 15.9 fmol/100/~m 2 and 12.2 fmol/100/tin 2, respectively. Layers I I I - V were more lightly labeled. The highest concentration of specific [ThrSg]IGF-I binding sites was in the choroid plexus of the lateral ventricles; the choroid plex-
208 dense in the glomerular layer (18.6 fmol/100 # m 2) and the inner portion of the external plexiform layer TABLE I Quantitative distribution of lG F-I binding sites in rat brain Specific binding of 0.01 nM 125I[ThrS9]IGF-Ito various regions of rat brain measured by densitometry of autoradiographic images on LKB Ultrofilm. Non-specific binding has been subtracted. Data (means + S.E.M.) were obtained from 3 rats, with 1-13 measuremen,s per animal, depending on the size of the region sampled.
Fig. 3. Total binding of l:51[Thr59]IGF-I to hindbrain with labeled peptide alone (A) and in preseace of 1000-fold excess unlabeled [Thr59]IGF-I (I3), which shows non-specific binding typical of all brain regions. The image in B was overexposed compared to A in order to obtain a printable photo because the non-specific binding levels produced a faint film image. For this reason, tbe background is darker in B, and the actual image is much fainter relative to A.
us of the third ventricle was also heavily labeled (23.3 and 22.8 fmol/100/,tin 2, respectively). Specific binding was widespread in the limbic system. In particular, the nucleus accumbens, dentate gyrus, and C a ! , Ca2 and Ca3 regions of the hippocampus were moderately labeled (Fig. 2). Several circumventricular organs showed relatively high concentrations of [Thr59]IGF-I-specific bindins sites, including the median eminence (17.5 fmol/100/~m~), and subfornical organ (12.4 fmoi/100 g m 2) (Fig. 2). In the median eminence, the highest binding was in the external zone. Specific binding above non-specific levels was also observed ~,n the O V I , T and area postrema, but there were insufficient sections for accurate m e a s u r e m e n t s . A distinctive labeling pattern was observed in the olfactory bulb (Fig. 1). Specific binding was most
Region of the brain
Specific binding (fmol/ l OOl~mz)
Choroid plexus, lateral ventricle Choroid plexus third ventricle Glomerular layer, olfactory bulb Median eminence Tenia tecti (precommissural hippocampus) Frontal cortex, layer VI External plexiform layer, olfactory bulb Hippocampus, region Ca3 Accessory olfactory bulb, granular layer Anterior olfactory nucleus, external zone Anterior ventral thalamic nucleus Accessory olfactory bulb, principal Medial geniculate Olfactory cortex, primary Anterior olfactory nucleus, dorsal Subfornical orgon Frontal cortex layer II Gelatinosus thalamic nucleus Nucleus accumbens Hippocampus, dentate gyrus Accessory olfactory nucleus, intermediate Pet~ventricular nucleus, hypothalamus Laterodorsal thalamic nucleus Mediodorsal thalamic nucleus Ventroposterior thalamic nucleus, medial Anterior olfactory nucleus, lateral Ventroposterior thalamic nucleus, lateral Dorsolateral geniculate nucleus Ventrolateral thalamic nucleus Anterior olfactory nucleus, ventral Suprachiasmatic nucleus Supraoptic nucleus Pontine nuclei Spinal tract nucleus of trigeminal nerve Superior coiliculus, superficial gray Caudate-putamen Paraventricular nucleus, hypothalamus Inferior olive Mammalary nucleus Motor nucleus, trigeminal nerve Sensory nucleus, trigeminal nerve Superior olive Ventral cochlear nucleus, anterior Inferior colliculus Arcuate nucleus Optic chiasm
23.3 + 2.0 22.8 _+0.2 18.6 + 0.1 17.5 4. 1.0 17.5 +_0.1 15.9 + 2.0 15.2 _ 0.6 14.8 +_0.1 14.3 4. 0.9 14.1 _+2.0 13.9 +_0.1 13.6 _+0.3 13.4 4. 1.0 13.0 4. 1.0 12.7 4. 0.9 12.4 +_0.1 12.2 _+0.1 12.1 4- 0.4 11.8 +_ 1.0 11.4.4- 0.7 11.4 4- 0.3 11.3 _+0.1 11.2 4- 1.0 11.2 4- 0.1 11.1 4- 0.5 11.0 _+.1.0 11.0 4- 1.0 11.0 4- 0.2 10.0 4- 0.1 9.0 4- 0.5 8.6 _+0.6 8.4 4. 0.1 8.4 _ 0.2 8.3 +_0.1 8.2 _+2.0 7.6 4. 0.1 6.9 _ 2.0 4.8 _+0.1 4.8 +_0.1 3.9 _+0.1 3.8 4. 0.1 3.7 _+0.8 3.6 _+0.1 3.2 _+0.1 1.8 4. 0.1 1.5 _+0.1
209 (15.2 fmol/100/~m2)..The layer of tufted cells in the outer external plexiform layer had relatively light labeling. Also heavily labeled were the. . anterior olfactory nucleus, accessory Olfactory bulb, and the granule cell layer of the accessory olfact0ry bulb. Regions involved in visual pathways contained moderate concentrations of [ThrS9]IGF-I binding sites (Fig. 2). Specific binding was present in the dorsolateral geniculate (11.0 fmol/100/zm2), superficial grey of the superior colliculus (8.2 fmol/100/~m2), and the suprachiasmatic nucleus (8.6 fmol/100/zm2). Regions involved in somatosensory pathways which were moderately labeled included the spinal tract trigeminal nucleus (8.3 fmol/100/zm2), ventroposterior thalamic (medial) nucleus (11.1 fmol/100 ~m 2) and the ventral posterior thalamic (lateral) nucleus (11.0 fmol/!00/~m2). In addition to those regions of the thalamus previously mentioned, the gelatinosus nucleus, anteroventral nucleus, ventrolateral nucleus, and laterodorsal nucleus were moderately labeled. Unlike the thalamus, the hypothalamus was for the most part labeled relatively lightly (Fig. 2). Hypothalamic regions that were labeled included the paraventricular nucleus, supraoptic nucleus, and periventricular nucleus. Highest specific binding in the hypothalamus was in the median eminence (17.5 fmol/100/~m2). In contrast, the arcuate nucleus was among the regions of the brain which had the lowest specific binding for ~25I[ThrS9]IGF-I. Specific binding in the arcuate nucleus (1.8 fmol/100/zm 2) was only slightly above that Binding [fmol/I00pm2"I 30-T-
i
20~S-
T
i
DISCUSSION
Two types of IGF receptors are present in the brain membranes 26. The type 1 IGF receptor resembles the insulin receptor in its structure, and has highest affinity for IGF-I. It also recognizes !GF-ll and, at high concentrations, insulin. The type 2 IGF receptor is structurally unlike the insulin receptor, and has highest affinity for IGF-II. It recognizes IGF-I but does not bind insulin. In our study, a 1000-fold excess of insulin reduced binding of 1251[Thrsg]IGF-I to brain slices, although not as potently as equimolar amounts of [Thr59]IGF-I. (Substantially higher concentrations of insulin (in the micromolar range)
~1251[Thr59"]IGF_ I plus
251[ThrSg]IGF-I plus
~-~1251[Th r 59]IGF_ I alone
25-
in the optic chiasm (1.5 fmol/100 #tin2), which was typical for white matter regions. The specificity of binding was evaluated by measuring binding of 12SI[ThrS9]IGF-I mixed with 1000fold excess of unlabeled [Thrsg]IGF-I or porcine insulin. In all brain regions that were measured, unlabeled [Thrsg]IGF-I was more effective than insulin in reducing binding of labeled [ThrS9]IGF-I, and the non-specific binding in the presence of unlabeled [Thrsg]IGF-I was similar (Fig. 4). In addition, the labeled [ThrS9]IGF-I was mixed (50 pg/mi) wi.:h an antibody to the rat IGF-II receptor (provided by Dr. Ron Rosenfeld, Stanford University). This antibody, which blocks binding to the IGF-II receptor aT, did not inhibit binding of the labeled [ThrS9]IGF-I to rat brain slices.
~lOnM
lOn~ [ThrSg"hGF-I
Insulin
-J-
~
lOS-
i;:
o CP Lal.
Vent
ME
CC Layer Vl
AOB Gr. Layer
AON Ext. Layer
AOB OC Primary Primary
GN
NA
DG
VPM
AON Lat.
VPL
Sp 5
AN
Fig. 4. Binding of 0.01 nM 12~l[Thr~'~]lGF-I to selected regions of rat brain, showing competition between [Thr~"]IGF-! and insulin for 12"~l[ThrS')]IGF-Ibinding sites. Bars indicate total binding in fmol/100 ,urn2 of section area (means + S.E.M.) in presence of 12~l[ThrS')]lGF-I alone (clear bars) and mixed with 10()0-fold excess (10 nM) unlabeled porcine insulin (halchcd bars) or IThr~"IIGF-! (stippled bars),
210 would be required to displace binding of labeled IGFI to the background (non-specific binding) levels shown in Fig. 4.) This result suggests that the 125I[Thr59]IGF-I bound principally to receptors for IGFI, since insulin does not bind to the brain IGF-II receptor26. Furthermore, it has been shown that the [Thr59]IGF-I recombinant form of IGF-I, which we used in these studies, binds relatively poorly to the !GF-II receptor compared with natural IGF-I (ref. 36). In addition, preliminary studies in our laboratory indicate that most IGF-II binding sites in the brain do not coincide with the location of IGF-I receptors 2, and we found that binding was not blocked by an antibody which inhibits binding of IGF-II to its receptor. Therefore we conclude that our autoradiogrephic studies with labeled [Thr59]IGF-I identified primarily type 1 IGF receptors (highest affinity for IGF-I) and that these receptors are widely distributed at anatomically discrete sites in the functional architecture of the brain. In related quantitative autoradiographic studies, we have also shown that the principal locations of insulin receptors in the rat brain3,4,ts are different from the regions which label heavily with n5I[Thr59]IGF-I. Insulin receptors are highly concentrated in the external plexiform layer of the olfactory bulb and are relatively scarcer in the glomerular and nerve fiber layers, which have high concentrations of IGF-I binding sites. In the hypothalamus, insulin binding is highest in the arcuate nucleus, which has very few IGF-I binding sites. Likewise, the median eminence, which has a very high density of IGF-I binding sites, contains relatively few insulin receptors. IGF-I binding is strikingly high in the thalamus and medial geniculate, whereas insulin binding is at background levels in these regions. In the cerebral cortex, IGF-I binding is high in layers I, II and VI; insulin receptors are concentrated in layers III-V. The lack of concordance in the locations of binding sites for insulin and IGF-I suggests that these peptides may have different functions in the CNS. A notable exception to the dissimilar distribution of IGF-I and insulin receptors in the brain is the choroid plexus, which has high concentrations of binding sites for both insulin3,21 and IGF-I. We have consiste~tly found that t]~e choroid plexus is the brain region with the highest concentrations of insulin and IGF-I binding sites. The precise cellular location of
these binding sites in the choroid plexus is not known, however. The high concentrations of receptors for these peptides within the choroid plexus suggests that IGF-I and insulin in plasma and/or CSF regulates choroid plexus functions, perhaps influencing choroid plexus transport and the composition of CSF. These IGF-I binding sites may also indicate that the choroid plexus may be a site of transport of insulin and IGF-I from the plasma to CSF. In this study, as well as in previous work 13we have found that IGF-I binding sites are concentrated in the median eminence. This observation raises the hypothesis that the median eminence IGF-I binding sites may play a role in regulation of growth hormone secretion. IGF-I in plasma inhibits growth hormone secretion by a little understood feedback mechanism ~4. Evidence supports IGF-I feedback on growth hormone secretion at both the level of the anterior pituitary gland and hypothalamus. The anterior pituitary gland has receptors for IGF-134,35, and growth hormone release from cultured pituitary adenomas is inhibited by IGF-116. In cultured hypothalamus slices, IGF-I stimulates the release of somatostatin 7`s which is known to inhibit release of pituitary growth hormone. The precise sites where IGF-I acts within the hypothalamus to initiate somatostatin release are unknown, however. We previously suggested that the median eminence may be a major site where IGF-I acts to indirectly inhibit GH secretion by stimulating the release of somatostatin into the hypophyseal portal circulation 13. The concentration of the IGF-I binding sites in the external zone of the median eminence where somatostatin terminals are known to be situated, supports this hypothesis. We have recently found that IGF-I receptors in the median eminence become more numerous during a prolonged fast which is accompanied by reduced plasma IGF-I levels ~1. This suggests that median eminence IGF-I receptors may play a role in regulation of growth hormone levels during episodes of nutritional deprivation. It is noteworthy that olfactory, visual and auditory, as well as somatic afferent systems, are rich in IGF-I binding sites. This is reflected by high concentrations of IGF-I binding sites in the glomerular and external plexiform layer of the olfactory bulb, anterior olfactory nucleus, accessory olfactory nucleus, primary olfactory cortex, lateral geniculate, superior coiliculus,
211 suprachiasmatic nucleus, medial geniculate, inferior colliculus, superior olive, vestibuiochochlear nucleus, and the spinal trigeminal nucleus. These resuits suggest that IGF-I may modulate synaptic transmission in somatosensory pathways. The thalamus is especially rich in IGF-I binding sites. The ventrolateral (VPL) and ventromedial (VPM) nuclei, thalamic regions of high IGF-I binding density, receive the termination of the somatosensory pathways from the head and neck (VPM) and the body, limbs and tail (VPL). These in turn are reciprocally connected in a topographically organized manner with the somatosensory cortex 42'43. The gelatinosus nucleus is also well represented with IGF-I binding sites. Studies of Craig and Burton t9 have suggested that, in contrast to the mainly 'tactile' nature of the VPM and VPL, the gelatinosus nucleus may be involved chiefly with processing nociceptive information from the periphery. The mediodorsal and laterodorsal thalamic nuclei also contain high concentrations of IGF-I binding sites. Taken together, these results suggest that the integration of sensory pathways in the thalamus may be influenced by ikJt'-l. The cellular location of brain IGF-I receptors is largely unknown. The in vitro method of detecting IGF-I binding sites with LKB Ultrofilm does not have the resolution to determine which cell types bear IGF-I receptors. A portion of the binding that is visualized by this method may ~'epresent binding of labeled IGF-I to brain microvessels, which have IGFI receptors 25. Evidence suggests that neurons in culture have IGF-I receptors 15'3°, but the relative distribution of IGF-I receptors among microvesse!s, glia, and neurons in situ remains an unsolved problem. The functions of IGF-I in the CNS are largely un-
known. In addition to a feedback effect on release of growth hormone, IGFs have been proposed to have a CNS action on food intake. Mixtures of IGF-I and IGF-II injected into brain CSF reduce food intake and body weight 4~. Although a recent report indicates that IGF-II may be more potent than IGF-I in this action 29, a satiety role for IGF-I in the CNS ca~,not be ruled out. Insulin infused into brain CSF has a similar effect on food intake ~ and high concentrations of insulin receptors are located throughout the brain 6'18. Since insulin and IGF-I interact with each other's receptors 26,35, it is possible that some of the IGF-I binding sites that we identified in the brain may also represent potential sites of insulin action as well. Furthermore, the well described growth promoting effects of IGF-I on peripheral tissues 22, together with data indicating that IGF-I stimulates nucleic acid synthesis in cultured neurons ~5,a° and induces oligodendrocyte development in vitro 3~, suggest that IGF-I has a major role in brain development. Identification of the anatomical location of IGF-I binding sites may therefore provide clues to neural pathways whose development may be regulated by IGF-I. ACKNOWLEDGEMENTS This work was supported by NIH Grants AM 12829, AM 17047, NS 24809, and by the Veterans Administration. We are grateful to Dr. Ronald Chance of the Lilly Corp. for providing unlabeled porcine insulin, and to Dr. Ronald Rosenfeld of Stanford University for providing antibody to the rat IGF-II receptor. We thank Chare Vathanaprida and Debra Felt for technical assistance and William Hintz for editorial assistance.
ABBREVIATIONS Figs. 1-3 AM amygdalacomplcx AN arcuatenu. AOB accessoryolfactory bulb AOE anteriorolf. nu., external AOL anterior olf. nu., lateral AV anteroventralthalamic nu. Ca 1 Ca I region, hippocampus Ca2 Ca2 region, hippocampus Ca3 Ca3 region, hippocampus CG centralgray
CP CPu DG DLG E EC EN EPL FN GI
choroidplexus caudate-putamen dentate gyrus dorsolatera! geniculate ependymallayer, oil. bulb endopyriformcortex entorhinal cortex ext. plexiform layer, olf. bulb facialnuclear group glomerularcell layer, oil. bulb
GN Gr GrA IAM MD ME MG MN MOL NA
gelatinosusnu~ granulecell layer, main olf. bulb granulecell layer, accessolf. bulb interanteromedial thai. nu. mediodorsalthalamic nu. medianeminence medialgeniculate mammalarynuclear group cerebellum, molecular nuclear nucleusaccumbens
212 NF OC OpC PN Po PVN
nerve fiber layer, oil. bulb olfactory cortex optic chiasm pontine nuclear group posterior thai. nuclear group paraventricular nu.
Fig. 4 AN arcuate nucleus AOB accessory olfactory bulb (granular layer) AON anterior olfactory nu. (external. lateral)
RN RT SC SCN SON
reticular nu. complex reticularis thalamic nu. superior colliculus suprachiasmatic nu. supraoptic nu.
"FC VMN VPT II VI
outer tufted cell layer, olf. bulb ventromedial hypothalamicnu. ventroposterior thai. nuclear group cerebral cortex, layer 2 cerebral cortex, layer 6
CP CC DG GN ME
choroid plexus (lateral ventricle) cerebral cortex (layer VI) dentate gyrus gelatinosus nucleus median eminence
NA OC Sp5 VPL VPM
nucleus accumbens olfactory cortex (primary) spinal nucleus of trigeninai nerve ventroposteriorlateraithalamic nu. ventroposteriomedialthalamicnu.
REFERENCES 1 B~ickstrom,M., Hall, K. and Sara, V., Somatomedin levels in cerebrospinal fluid from adults with pituitary disorders, Ac~ ~ndocrinol., 107 (1984) 171-178. 2 Baskin, D. and Bohannon, N., Localization of binding sites for insulin-like growth factor-lI (IGF-II) in rat brain by in vitro quantitative autoradiography, Endocrinology, Suppl., 120 (1987) 42. 3 Baskin, D., Brewitt, B., Davidson, D., Corp, E., Paquette, T., Figlewicz, D., Graham, M., Woods, S. and Dorsa, D., Quantitative autoradiographic evidence for insulin receptors in the choroid plexus c,f the rat brain, Diabetes, 35 (1986) 246-249. 4 Baskin, D., Davidson, D., Corp, E., Lewellen, T. and Graham, M., An inexpensive microcomputer digital imaging system for densitometry: quantitative autoradiography of insulin receptors with 125Iand LKB Ultrofilm, J. Neurasci. Meth., 16 (1986) 119-129. 5 Baskin, D. and Dorsa, D., Quantitative autoradiography and in vitro radioligand binding. In D. McEachron (Ed.), Functional Mapping in Biology and Medicine: Computer Assisted Autoradiography, Experimental Biology and Medicine, Vol. II, Karger, New York, 1986, pp. 204-234. 6 Baskin, D., Figlewicz, D., Woods, S., Porte, Jr., D., and Dorsa, D., Insulin in the brain, Annu. Rev. Physiol., 49 (1987) 335-347. 7 Berelowicz, M., Firestone, S., and Frohman, L., Effects of growth hormone excess and deficiency on hypothalamic somatostatin content and release and on tissue somatostatin distribution, Endocrinology, 109 (1981) 714-719. 8 Berelowicz, M., Szabo, M., Frohman, L., Firestone, S., Chu, L. and Hintz, R., Somatomedin-C mediates growth hormone negative feedback by effects on both the hypothalamus and pituitary, Science, 212 (1981) 1279-1281. 9 Binoux, M., Faivre-Bauman, A., Lassarre, C., Barret, A. and Tixier-Vidal, A., Triiodothyronine stimulates production of insulin-like growth factor (IGF) by fetal hypothalamus cells cultured in serum-free medium, Dev. Brain Res., 21 (1985) 319-321. 10 Binoux, M., Hossenlopp, P., Lassarre, C. and Hardouin, N., Production of insulin-like growth factors and their carrier by rat pituitary gland and brain explants in culture, FEBS Lett., 124 (1981) 178-184. 11 Bohannon, N., Corp, E., Wilcox, B., Figlewicz, D., Dorsa, D. and Baskin, D., Insulin-like growth factor-I (IGF-I) binding in the median eminence of the rat is increased by
food restriction, Endocrinology, Suppl., 120 (1987) 42. 12 Bohannon, N., Corp, E., Wilcox, B., Porte, D., Jr., Figlewicz, D. and Baskin, D., Identification of specific binding sites for insulin-like growth factor-I (IGF-I) in the rat brain by in vitro quantitative autoradiography, Appetite, 7 (1986) 243-244. 13 Bohannon, N., Figlewicz, D., Corp, E., Wilcox, B., Porte, !r., D., and Baskin, D., Identification of binding sites for an insulin-like growth factor (IGF-I) in the median eminence of the rat brain by quantitative autoradiography, Endocrinology, 119 (1986) 943-945. 14 Brazeau, P., Guillemin, R., Ling, N., Van Wyk, J. and Humbel, R., Inhibition par les somatomedines de la secretion de l'hormone de croissance stimulee par le facteur hypothalamique somatocrinine (GRF) ou le peptide de synthese hpGRF, C.,'~. Seances Acad. Sci. (IH) (Paris), 295 (1982) 651-654. 15 Burgess, S., Jacobs, S., Cuatrecasas, P. and Sahyoun, N., Characterization of a neuronal subtype of insulin-like growth factor I receptor, J. Biol. Chem., 262 (1987) 1618-1622. 16 Ceda, G., Hoffman, A., Silverberg, G., Wilson, D. and Rosenfeld, R., Regulation of growth hormone release from cultured human pituitary adenomas by somatomedins and insulin, J. Clin. Endocrinol. Metab., 60 (1986) 1204-1209. 17 Clemmons, D. and Van Wyk, J., Factors controlling blood concentrations of somatomedin C Clinics, In W.H. Daughaday (Ed.), Clinics in Endocrinology and Metabolism, Vol. 13, No. 1, Tissue Growth Factors, Saunders, New York, 1984, pp. 113-144. 18 Corp, E., Woods, S., Porte, Jr., D., Figlewicz, D. and Baskin, D., Localization of 125I-insulinbinding sites in the rat hypothalamus by quantitative autoradiograhy, Neurosci. Lett., 70 (1986) 17-22. 19 Craig, A. and Burton, H., Spinal and medullary lamina I projection to nucleus submedius in medial thalamus: a possible pain center, J. Neurophysiol., 45 (1981) 443-466. 20 D'Ercole, A., Applewhite, G. and Underwood, L., Evi. dence that somatomedin is synthesized by multiple tissues in th,.-fetus, Dev. Biol., 75 (1980) 315-328. 21 Dorsa, D.M. and Baskin, D.G., Identification of neuropeptide receptors. In A. Boulton, G. Baker and Q. Pittman (Eds.), Neuromethods, Vol. 6, Humana, Clifton, NJ, 1987, pp. 203-244. 22 Froesch, E. and Zapf, J., Insulin-like growth factors and insulin: comparative aspects, Diabetologia, 28 (1985) 485-493.
213 23 Figlewicz, D., Baskin, D., Woods, S. and Porte, Jr., D., Insulin and insulin-like growth factor-I (IGF-I) binding is decreased in the olfactory bulb (OB) but not the cerebral cortex (C) of the obese Zucker (fa/fa) rat, Int. J. Obesity, 9 (1985) 233. 24 Finegold, E., Seshadri, S.B. and Tretiak, O.J., Hardware and software design considerations in engineering an image processing workstation: autoradiographic analysis with DUMAS and the BRAIN autoradiograph analysis software package. In D. McEachron (Ed.), Functional Mapping in Biology and Medicine: Computer Assisted Autoradiography, Experimental Biology and Medicine, Vol, II, Karger, New York, 1986, pp. 175-201. 25 Frank, H., Pardridge, W., Morris, W., Rosenfeld, R. and Choi, T., Binding and internalization of insulin-like growth factors by isolated brain microvessels, Diabetes, 35 (1986) 654-661. 26 Gammeltoft, S., Hasselbacher, G., Humbel, R., Fehlmann, M. and Van Obberghen, E., Two types of receptor for insulin-like growth factors in mammalian brain, EMBO J., 4 (1985) 3407-3412. 27 Goodlier, C., de Stephano, L., Lai, W., Guyda, H. and Posne~, B., Characterization of insulin-like growth factor receptors in rat anterior pituitary, hypothalamus, and brain, Endocrinology, 114 (1984) 1187-1195. 28 Hasselbacher, G., Schwab, M., Pasi, A. and Humbel, R., Insulin-like growth factor II (IGF-II) in human brain: regional distribution of IGF-II and of higher molecular mass forms, Proc. Natl. Acad. Sci. U.S.A., 82 (1985) 2153-2157. 29 Lauterio, T., Marson, L., Daughaday, W. and Baile, C., Intracerebroventricular injections of insulin-like growth factor II, but not insulin-like growth factor I, decrease food intake and body weight in rats, Soc. Neurosci. Abstr., 12 (1986) 795. 30 Lenoir, D. and Honegger, P., Insulin-like growth factor I (IGF-I) stimulates DNA synthesis in fetal rat brain cell cultures, Dev. Brain Res., 7 (1983) 205-213. 31 McMorris, F., Smith, T., DeSalvo. S. and Furlanetto, R.. Insulin-growth factor I/Somatomedin C: a potent inducer of oligodendrocyte development, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 822-826. 32 Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Coordinates, Academic, New York, 1982. 33 Rail, L., Falloon, L., Styne, D. and Bell, G., Insulin-like growth factor I and II mRNA are present in numerous tissues of the primate, Ciin. Res., 34 (1986) 135A. 34 Rosenfeld, R., Ceda, G., Cutler, C. and Hoffman, A., Insulin and insulin-like growth factor (somatomedin) recep-
tors on cloned rat pituitary tumor cells, Endocrinology, 117 (1985) 2008-2016. 35 Rosenfeld, R., Ceda, G., Wilson, D., Dollar, L. and Huffman, A., Characterization of high affinity receptors for insulin and insulin-like growth factors I and II on rat anterior pituitary cells, Endocrinology, 114 (1985) 1571-1575. 36 Rosenfeld, R.G., Conover, C.A., Hodges, D., Lee, P.D.K., Misra, P., Hintz, R.L. and Li, ~.H., Heterogeneity of insulin-like growth factor-I affinity for the insulin-like growth factor-II receptor: comparison of natural, synthetic and recombinant DNA-derived insulin-like growth factorI, Biochem. Biophys. Res. Comm., 143 (1987) 199-205. 37 Rosenfeld, R.G., Hodges, D., Pham, H., Lee, P.D.K. and Poweli, D.R., Purification of the insulin-like growth factorlI (IGF-II) receptor from an IGF-II-producing cell line, and generation of an antibody which both immunoprecipitates and blocks the type 2 IGF receptor, Biochem. Biophys. Res. Comm., 138 (1986) 304-311. 38 Sara, V., Hall, K., Misaki, M., Fryklund, L , Christensen, N. and Wetterberg, L., Ontogenesis of somatomedin and insulin receptors in the human fetus, J. Clin. Invest., 71 (1983) 1084-1094. 39 Sara, V., Hall, K., Von Holz, H., Humbei, R., Sjogren, B. and Wetterberg. L., Evidence for the presence of specific receptors for insulin-like growth factors I (IGF-I) and II (IGF-2) and insulin throughout the adult human brain, Neurosci. Lett., 34 (1982) 39-44. 40 Sara, V., Uvnas-Moberg, K., Uvnas, B., Hall, K., Wetterberg, L., Posloncec, B. and Goiny, M., The distribution of somatomedins in the nervous system of the cat and their release following neural stimulation, Acta. Physiol. Scand., 115 (1982) 467-470. 41 Tannenbaum, G., Guyda, H. and Posner, B., Insulin-like growth factors: a role in growth hormone negative feedback and body weight regulation via brain, Science, 220 (1983) 77-79. 42 Wise, S.P., The laminar organization of certain afferent and efferem fiber systems in the rat somatosensory cortex, Brain Research, 90 (1975) 139-142. 43 Wise, S.P. and Jones, E.G., Developmental studies of thalamocortical and commissurai connections in the rat somatosensory cortex, J. Comp. Neurol., 178 (1978) 187-208. 44 Woods, S. and Porte, Jr., D., The role of insulin as a satiety factor in the central nervous system. In A. Szabo (Ed.), CNS Regulation of Carbohydrate Metabolism, Academic Press, New York, pp. 1983, 457-468. 45 Zangger, I., Eberwine, J., Barchas, J., Jansen, J., Rosenfeld, R., Hoffman, A. and Hintz, R., Tissue distribution of IGF-I mRNA, Clin. Res., 34 (1986) 68A.
205
BRE 13386
Research Reports
Localization of binding sites for insulin-like growth factor-I (IGF-I) in the rat brain by quantitative autoradiography •
?
Nancy J. Bohannon 3, Eric S. Corp 4"*, Barbara J. Wilcox-, Dianne P. Figlewicz4, Daniel M. Dorsal'5 and Denis G. Baskin 1-3 t Division of Endocrinology~Metabolism, Veterans Administration Medical Center, Seattle WA 98108 (U.S.A.) and Departments of 2Medicine, 3BiologicalStructure, 4Physiologyand SPharmacoiogy, University of Washington, Seattle, WA 98195 (U.S.A.) (Accepted 8 September 1987) Key words: Insulin-like growth factor-1; Quantitative autoradiography; Brain; Rat; Receptor
In vitro quantitative autoradiography was used to localize IGF-I binding sites in rat brain. Slide-mounted sections of frozen rat brain were incubated in 0.01 nM 125I[ThrSg]IGF-I, alone or mixed with 10 nM unlabeled [Thrsg]IGF-I or insulin, for 22 h at 4 °C and apposed to LKB Ultrofilm. Measurement of labeled [Thrsg]IGF-I binding by computer digital image analysis of the autoradiographic images indicated that high affinity IGF-I binding sites are widely distributed at discrete anatomical regions of the brain microarchitecture. The highest concentration of specific binding sites was in the choroid plexus of the lateral and third ventricles. Unlabeled porcine insulin was less potent than unlabeled IGF-I in competing for binding sites on brain slices. Regions of the olfactory, visual, and auditory, as well as visceral and somatic sensory systems were labeled, in particular the glomerular layer of the olfactory bulb, the anterior olfactory nucleus, accessory olfactory bulb, primary olfactory cortex, lateral-dorsal geniculate, superior colliculus, medial geniculate, and the spinal trigeminal nucleus. High concentrations of IGF-I-specific binding sites were present throughout the thalamus and the hippocampus, (dentate gyrus, Ca1, Ca2, Ca3). The hypothalamus had moderate binding in the paraventricular, supraoptic, and suprachiasmatic nucleus. Highest binding in the hypothalamus was in the median eminence. The arcuate nucleus showed very low specific binding, approaching the levels found in optic chiasm and white matter regions. Layers II and VI of the cerebral cortex also had moderate IGF-I binding. The results suggest that the development and functions of brain sensory and neuroendocrine pathways may be regulated by IGF-I. INTRODUCTION Insulin-like growth factor-I (IGF-I) is a somatomedin (moi. wt. approximately 7000) with insulin-like structure and biological properties, which is made in the liver and circulates in plasma bound to carrier proteins 17'22. A peptide with immunological properties of IGF-I has been extracted from brain and cerebrospinal fluid 1'20'28A0. Furthermore, receptors for IGF-I have been identified in membrane preparations of rat brain 22'26'27, and human adult 26'39 and fetal brain as. Synthesis of IGF-I by the brain is suggested by reports of IGF-I-like m R N A in rat brain 33'45,
and the production of IGF-I in cultures of brain cells9. These observations support the hypothesis that the brain is an IGF-I-sensitive organ. Indeed, IGF-I has been shown to stimulate the release of somatostatin from hypothalamic cultures 8 and to stimulate D N A synthesis in fetal rat brain cultures 3°. The sites of action of IGF-I in the brain are largely unknown, however. We have approached this problem by identifying the location of IGF-I specific binding sites with autoradiography. In an earlier study, we found a high density of specific binding sites for IGF-I in the rat median eminence ~3. In the present study, we have mapped the location of specific bind-
* Present address: Bourne Research Laboratory, Cornell University Medical Center, White Plains, NY 10605, U.S.A. Correspondence: D.G. Baskin, Division of Endocrinology/Metabolism (151), VA Medical Center, 1660 S. Columbian Way, Seattle, WA 98108, U.S.A. 0006-8993/88/$03.50 © 1988 Elsevier ~,. :nce Publishers B.V. (Biomedical Division)
206 ing sites for an iodinated analogue of IGF-I in slices of rat brain, using autoradiography with LKB Uitrofilm. The relative concentrations of IGF-I binding sites "n discrete anatomical locations were determined by computer digital image densitometry 4'5. A portion of these results were reported as preliminary data in an abstracd 2. MATERIALSAND METHODS Male Sprague-Dawley rats (200-250 g) were anesthetized with sodium pentobarbital and perfused with ice-cold saline via the left cardiac ventricle. The brains were removed, frozen in crushed dry ice and stored a t - 7 0 °C for up to 14 days. Coronal cryostat sections (15/~m) were mounted on gelatin coated slides and stored at-70 °C for a maximum of 4 weeks. Brain IGF-I receptors have been shown to be stable at-20 °C for up to a year 1°. The tracer peptide for these studies was a recombinant form of human IGF-I, [ThrSg]IGF-I (Amersham), which has been widely used for analysis of rat and human IGF-I receptors -;6. ~2-SI[Thr'~gllGF-I(spec. act. -- 267 pCi/pg) was dissolved to a final concentration of 0.01 nM (10 pmol/ml) in 10 mM Hepes assay buffer containing 0.5% BSA, 0.025% bacitracin, 0.0125% N-ethylmaleimide and 100 K.I.U./ml Aprotinin (Sigma), pH 7.64"23. Each brain slice was covered with 0.1 ml of labeled [ThrSg]IGF-I either alone or mixed with 10 nM unlabeled [Thr59]IGF-I (AmGen) or porcine insulin (Lilly), and incubated for 22 h at 4 °C. The slides were then rinsed in 3 one-rain
E• TG
~
changes of ice-cold assay buffer, immersed in a final one-min rinse of 0 °C distilled water, and rapidly dried. We have previously worked out equilibrium conditions using brain slices and found that specific binding of labeled [ThrS'~]IGF-I to brain slices reaches equilibrium by 12 h at 4 °C, is unchanged at 24 h, and is reversible II. The procedures for quantitative autoradiography are described in detail elsewhere 4'5'2z. Briefly, the labeled and dried brain slices were placed in direct contact with the emulsion of LKB Ultrofilm in an Xray film cassette for 72 h. The film was then developed for 4 min in D-19 at 20 °C. Quantitative analysis of optical density in small, defined anatomic regions of the autoradiographic images was made with the DUMAS computer digitizing system and BRAIN program for quantitative autoradiography 24. Optical 6ensities were converted to DPM/100 mm 2 of the section area using radioactivity standards calibrated for 1251radioactivity in tissue slices and a 3-day exposure period 4's. To obtain specific binding, radioactivity bound in the presence of 1000-fold excess unlabeled peptide (non-specific binding) was subtracted from binding with labeled peptide alone (total binding). Anatomical regions were identified with a stereotaxic atlas of rat brain 32. RESULTS qinding of 125I[Thrsg]lGF-I was heterogeneously distributed among discrete anatomical regions of the rat brain (Figs. 1-3). The pattern of localization that
AQ
'
G -
L
Fig. 1. Autoradiographic images of olfactory bulb coronal slices labeled with t2SI[Thr59]IGF-I.A: rostral olfactory bulbs. B: level of accessory olfactory bulbs.
207
,dl ,/I,' ,
,~
.
".~.~'~.V:
Fig. 2. Binding of 125I[Thr59]lGF-Ito coronal brain slices at levels of caudate-putamen (A), suprachiasmatic nucleus (B). paraventricular nucleus (C), median eminence (D), medial geniculate (E), and pontine nuclei (F). we describe in this paper has been observed in brains of more than 40 rats and with 6 separate batches of ]25I[ThrS9]IGF-I. The concentrations of IGF-l-specific binding in 46 different brain regions when 0.01 nM 1251[ThrSg]IGFI was used are shown in Table I. The quantitative results reported here represent measurements made on the brains of 3 rats which were processed and assayed together. Non-specific binding of 125I[ThrS9]IGF-I was uniformly distributed and averaged 4.8 + 1.0
fmol/100pm 2 for all regions combined. Thus the total binding shown in autoradiographic images (Figs. 1-3) reflects the relative differences in specific binding among the brain regions which were labeled. Throughout the cortex, layers II and VI were heavily labele0, with specific binding of 15.9 fmol/100/~m 2 and 12.2 fmol/100/tin 2, respectively. Layers I I I - V were more lightly labeled. The highest concentration of specific [ThrSg]IGF-I binding sites was in the choroid plexus of the lateral ventricles; the choroid plex-
208 dense in the glomerular layer (18.6 fmol/100 # m 2) and the inner portion of the external plexiform layer TABLE I Quantitative distribution of lG F-I binding sites in rat brain Specific binding of 0.01 nM 125I[ThrS9]IGF-Ito various regions of rat brain measured by densitometry of autoradiographic images on LKB Ultrofilm. Non-specific binding has been subtracted. Data (means + S.E.M.) were obtained from 3 rats, with 1-13 measuremen,s per animal, depending on the size of the region sampled.
Fig. 3. Total binding of l:51[Thr59]IGF-I to hindbrain with labeled peptide alone (A) and in preseace of 1000-fold excess unlabeled [Thr59]IGF-I (I3), which shows non-specific binding typical of all brain regions. The image in B was overexposed compared to A in order to obtain a printable photo because the non-specific binding levels produced a faint film image. For this reason, tbe background is darker in B, and the actual image is much fainter relative to A.
us of the third ventricle was also heavily labeled (23.3 and 22.8 fmol/100/,tin 2, respectively). Specific binding was widespread in the limbic system. In particular, the nucleus accumbens, dentate gyrus, and C a ! , Ca2 and Ca3 regions of the hippocampus were moderately labeled (Fig. 2). Several circumventricular organs showed relatively high concentrations of [Thr59]IGF-I-specific bindins sites, including the median eminence (17.5 fmol/100/~m~), and subfornical organ (12.4 fmoi/100 g m 2) (Fig. 2). In the median eminence, the highest binding was in the external zone. Specific binding above non-specific levels was also observed ~,n the O V I , T and area postrema, but there were insufficient sections for accurate m e a s u r e m e n t s . A distinctive labeling pattern was observed in the olfactory bulb (Fig. 1). Specific binding was most
Region of the brain
Specific binding (fmol/ l OOl~mz)
Choroid plexus, lateral ventricle Choroid plexus third ventricle Glomerular layer, olfactory bulb Median eminence Tenia tecti (precommissural hippocampus) Frontal cortex, layer VI External plexiform layer, olfactory bulb Hippocampus, region Ca3 Accessory olfactory bulb, granular layer Anterior olfactory nucleus, external zone Anterior ventral thalamic nucleus Accessory olfactory bulb, principal Medial geniculate Olfactory cortex, primary Anterior olfactory nucleus, dorsal Subfornical orgon Frontal cortex layer II Gelatinosus thalamic nucleus Nucleus accumbens Hippocampus, dentate gyrus Accessory olfactory nucleus, intermediate Pet~ventricular nucleus, hypothalamus Laterodorsal thalamic nucleus Mediodorsal thalamic nucleus Ventroposterior thalamic nucleus, medial Anterior olfactory nucleus, lateral Ventroposterior thalamic nucleus, lateral Dorsolateral geniculate nucleus Ventrolateral thalamic nucleus Anterior olfactory nucleus, ventral Suprachiasmatic nucleus Supraoptic nucleus Pontine nuclei Spinal tract nucleus of trigeminal nerve Superior coiliculus, superficial gray Caudate-putamen Paraventricular nucleus, hypothalamus Inferior olive Mammalary nucleus Motor nucleus, trigeminal nerve Sensory nucleus, trigeminal nerve Superior olive Ventral cochlear nucleus, anterior Inferior colliculus Arcuate nucleus Optic chiasm
23.3 + 2.0 22.8 _+0.2 18.6 + 0.1 17.5 4. 1.0 17.5 +_0.1 15.9 + 2.0 15.2 _ 0.6 14.8 +_0.1 14.3 4. 0.9 14.1 _+2.0 13.9 +_0.1 13.6 _+0.3 13.4 4. 1.0 13.0 4. 1.0 12.7 4. 0.9 12.4 +_0.1 12.2 _+0.1 12.1 4- 0.4 11.8 +_ 1.0 11.4.4- 0.7 11.4 4- 0.3 11.3 _+0.1 11.2 4- 1.0 11.2 4- 0.1 11.1 4- 0.5 11.0 _+.1.0 11.0 4- 1.0 11.0 4- 0.2 10.0 4- 0.1 9.0 4- 0.5 8.6 _+0.6 8.4 4. 0.1 8.4 _ 0.2 8.3 +_0.1 8.2 _+2.0 7.6 4. 0.1 6.9 _ 2.0 4.8 _+0.1 4.8 +_0.1 3.9 _+0.1 3.8 4. 0.1 3.7 _+0.8 3.6 _+0.1 3.2 _+0.1 1.8 4. 0.1 1.5 _+0.1
209 (15.2 fmol/100/~m2)..The layer of tufted cells in the outer external plexiform layer had relatively light labeling. Also heavily labeled were the. . anterior olfactory nucleus, accessory Olfactory bulb, and the granule cell layer of the accessory olfact0ry bulb. Regions involved in visual pathways contained moderate concentrations of [ThrS9]IGF-I binding sites (Fig. 2). Specific binding was present in the dorsolateral geniculate (11.0 fmol/100/zm2), superficial grey of the superior colliculus (8.2 fmol/100/~m2), and the suprachiasmatic nucleus (8.6 fmol/100/zm2). Regions involved in somatosensory pathways which were moderately labeled included the spinal tract trigeminal nucleus (8.3 fmol/100/zm2), ventroposterior thalamic (medial) nucleus (11.1 fmol/100 ~m 2) and the ventral posterior thalamic (lateral) nucleus (11.0 fmol/!00/~m2). In addition to those regions of the thalamus previously mentioned, the gelatinosus nucleus, anteroventral nucleus, ventrolateral nucleus, and laterodorsal nucleus were moderately labeled. Unlike the thalamus, the hypothalamus was for the most part labeled relatively lightly (Fig. 2). Hypothalamic regions that were labeled included the paraventricular nucleus, supraoptic nucleus, and periventricular nucleus. Highest specific binding in the hypothalamus was in the median eminence (17.5 fmol/100/~m2). In contrast, the arcuate nucleus was among the regions of the brain which had the lowest specific binding for ~25I[ThrS9]IGF-I. Specific binding in the arcuate nucleus (1.8 fmol/100/zm 2) was only slightly above that Binding [fmol/I00pm2"I 30-T-
i
20~S-
T
i
DISCUSSION
Two types of IGF receptors are present in the brain membranes 26. The type 1 IGF receptor resembles the insulin receptor in its structure, and has highest affinity for IGF-I. It also recognizes !GF-ll and, at high concentrations, insulin. The type 2 IGF receptor is structurally unlike the insulin receptor, and has highest affinity for IGF-II. It recognizes IGF-I but does not bind insulin. In our study, a 1000-fold excess of insulin reduced binding of 1251[Thrsg]IGF-I to brain slices, although not as potently as equimolar amounts of [Thr59]IGF-I. (Substantially higher concentrations of insulin (in the micromolar range)
~1251[Thr59"]IGF_ I plus
251[ThrSg]IGF-I plus
~-~1251[Th r 59]IGF_ I alone
25-
in the optic chiasm (1.5 fmol/100 #tin2), which was typical for white matter regions. The specificity of binding was evaluated by measuring binding of 12SI[ThrS9]IGF-I mixed with 1000fold excess of unlabeled [Thrsg]IGF-I or porcine insulin. In all brain regions that were measured, unlabeled [Thrsg]IGF-I was more effective than insulin in reducing binding of labeled [ThrS9]IGF-I, and the non-specific binding in the presence of unlabeled [Thrsg]IGF-I was similar (Fig. 4). In addition, the labeled [ThrS9]IGF-I was mixed (50 pg/mi) wi.:h an antibody to the rat IGF-II receptor (provided by Dr. Ron Rosenfeld, Stanford University). This antibody, which blocks binding to the IGF-II receptor aT, did not inhibit binding of the labeled [ThrS9]IGF-I to rat brain slices.
~lOnM
lOn~ [ThrSg"hGF-I
Insulin
-J-
~
lOS-
i;:
o CP Lal.
Vent
ME
CC Layer Vl
AOB Gr. Layer
AON Ext. Layer
AOB OC Primary Primary
GN
NA
DG
VPM
AON Lat.
VPL
Sp 5
AN
Fig. 4. Binding of 0.01 nM 12~l[Thr~'~]lGF-I to selected regions of rat brain, showing competition between [Thr~"]IGF-! and insulin for 12"~l[ThrS')]IGF-Ibinding sites. Bars indicate total binding in fmol/100 ,urn2 of section area (means + S.E.M.) in presence of 12~l[ThrS')]lGF-I alone (clear bars) and mixed with 10()0-fold excess (10 nM) unlabeled porcine insulin (halchcd bars) or IThr~"IIGF-! (stippled bars),
210 would be required to displace binding of labeled IGFI to the background (non-specific binding) levels shown in Fig. 4.) This result suggests that the 125I[Thr59]IGF-I bound principally to receptors for IGFI, since insulin does not bind to the brain IGF-II receptor26. Furthermore, it has been shown that the [Thr59]IGF-I recombinant form of IGF-I, which we used in these studies, binds relatively poorly to the !GF-II receptor compared with natural IGF-I (ref. 36). In addition, preliminary studies in our laboratory indicate that most IGF-II binding sites in the brain do not coincide with the location of IGF-I receptors 2, and we found that binding was not blocked by an antibody which inhibits binding of IGF-II to its receptor. Therefore we conclude that our autoradiogrephic studies with labeled [Thr59]IGF-I identified primarily type 1 IGF receptors (highest affinity for IGF-I) and that these receptors are widely distributed at anatomically discrete sites in the functional architecture of the brain. In related quantitative autoradiographic studies, we have also shown that the principal locations of insulin receptors in the rat brain3,4,ts are different from the regions which label heavily with n5I[Thr59]IGF-I. Insulin receptors are highly concentrated in the external plexiform layer of the olfactory bulb and are relatively scarcer in the glomerular and nerve fiber layers, which have high concentrations of IGF-I binding sites. In the hypothalamus, insulin binding is highest in the arcuate nucleus, which has very few IGF-I binding sites. Likewise, the median eminence, which has a very high density of IGF-I binding sites, contains relatively few insulin receptors. IGF-I binding is strikingly high in the thalamus and medial geniculate, whereas insulin binding is at background levels in these regions. In the cerebral cortex, IGF-I binding is high in layers I, II and VI; insulin receptors are concentrated in layers III-V. The lack of concordance in the locations of binding sites for insulin and IGF-I suggests that these peptides may have different functions in the CNS. A notable exception to the dissimilar distribution of IGF-I and insulin receptors in the brain is the choroid plexus, which has high concentrations of binding sites for both insulin3,21 and IGF-I. We have consiste~tly found that t]~e choroid plexus is the brain region with the highest concentrations of insulin and IGF-I binding sites. The precise cellular location of
these binding sites in the choroid plexus is not known, however. The high concentrations of receptors for these peptides within the choroid plexus suggests that IGF-I and insulin in plasma and/or CSF regulates choroid plexus functions, perhaps influencing choroid plexus transport and the composition of CSF. These IGF-I binding sites may also indicate that the choroid plexus may be a site of transport of insulin and IGF-I from the plasma to CSF. In this study, as well as in previous work 13we have found that IGF-I binding sites are concentrated in the median eminence. This observation raises the hypothesis that the median eminence IGF-I binding sites may play a role in regulation of growth hormone secretion. IGF-I in plasma inhibits growth hormone secretion by a little understood feedback mechanism ~4. Evidence supports IGF-I feedback on growth hormone secretion at both the level of the anterior pituitary gland and hypothalamus. The anterior pituitary gland has receptors for IGF-134,35, and growth hormone release from cultured pituitary adenomas is inhibited by IGF-116. In cultured hypothalamus slices, IGF-I stimulates the release of somatostatin 7`s which is known to inhibit release of pituitary growth hormone. The precise sites where IGF-I acts within the hypothalamus to initiate somatostatin release are unknown, however. We previously suggested that the median eminence may be a major site where IGF-I acts to indirectly inhibit GH secretion by stimulating the release of somatostatin into the hypophyseal portal circulation 13. The concentration of the IGF-I binding sites in the external zone of the median eminence where somatostatin terminals are known to be situated, supports this hypothesis. We have recently found that IGF-I receptors in the median eminence become more numerous during a prolonged fast which is accompanied by reduced plasma IGF-I levels ~1. This suggests that median eminence IGF-I receptors may play a role in regulation of growth hormone levels during episodes of nutritional deprivation. It is noteworthy that olfactory, visual and auditory, as well as somatic afferent systems, are rich in IGF-I binding sites. This is reflected by high concentrations of IGF-I binding sites in the glomerular and external plexiform layer of the olfactory bulb, anterior olfactory nucleus, accessory olfactory nucleus, primary olfactory cortex, lateral geniculate, superior coiliculus,
211 suprachiasmatic nucleus, medial geniculate, inferior colliculus, superior olive, vestibuiochochlear nucleus, and the spinal trigeminal nucleus. These resuits suggest that IGF-I may modulate synaptic transmission in somatosensory pathways. The thalamus is especially rich in IGF-I binding sites. The ventrolateral (VPL) and ventromedial (VPM) nuclei, thalamic regions of high IGF-I binding density, receive the termination of the somatosensory pathways from the head and neck (VPM) and the body, limbs and tail (VPL). These in turn are reciprocally connected in a topographically organized manner with the somatosensory cortex 42'43. The gelatinosus nucleus is also well represented with IGF-I binding sites. Studies of Craig and Burton t9 have suggested that, in contrast to the mainly 'tactile' nature of the VPM and VPL, the gelatinosus nucleus may be involved chiefly with processing nociceptive information from the periphery. The mediodorsal and laterodorsal thalamic nuclei also contain high concentrations of IGF-I binding sites. Taken together, these results suggest that the integration of sensory pathways in the thalamus may be influenced by ikJt'-l. The cellular location of brain IGF-I receptors is largely unknown. The in vitro method of detecting IGF-I binding sites with LKB Ultrofilm does not have the resolution to determine which cell types bear IGF-I receptors. A portion of the binding that is visualized by this method may ~'epresent binding of labeled IGF-I to brain microvessels, which have IGFI receptors 25. Evidence suggests that neurons in culture have IGF-I receptors 15'3°, but the relative distribution of IGF-I receptors among microvesse!s, glia, and neurons in situ remains an unsolved problem. The functions of IGF-I in the CNS are largely un-
known. In addition to a feedback effect on release of growth hormone, IGFs have been proposed to have a CNS action on food intake. Mixtures of IGF-I and IGF-II injected into brain CSF reduce food intake and body weight 4~. Although a recent report indicates that IGF-II may be more potent than IGF-I in this action 29, a satiety role for IGF-I in the CNS ca~,not be ruled out. Insulin infused into brain CSF has a similar effect on food intake ~ and high concentrations of insulin receptors are located throughout the brain 6'18. Since insulin and IGF-I interact with each other's receptors 26,35, it is possible that some of the IGF-I binding sites that we identified in the brain may also represent potential sites of insulin action as well. Furthermore, the well described growth promoting effects of IGF-I on peripheral tissues 22, together with data indicating that IGF-I stimulates nucleic acid synthesis in cultured neurons ~5,a° and induces oligodendrocyte development in vitro 3~, suggest that IGF-I has a major role in brain development. Identification of the anatomical location of IGF-I binding sites may therefore provide clues to neural pathways whose development may be regulated by IGF-I. ACKNOWLEDGEMENTS This work was supported by NIH Grants AM 12829, AM 17047, NS 24809, and by the Veterans Administration. We are grateful to Dr. Ronald Chance of the Lilly Corp. for providing unlabeled porcine insulin, and to Dr. Ronald Rosenfeld of Stanford University for providing antibody to the rat IGF-II receptor. We thank Chare Vathanaprida and Debra Felt for technical assistance and William Hintz for editorial assistance.
ABBREVIATIONS Figs. 1-3 AM amygdalacomplcx AN arcuatenu. AOB accessoryolfactory bulb AOE anteriorolf. nu., external AOL anterior olf. nu., lateral AV anteroventralthalamic nu. Ca 1 Ca I region, hippocampus Ca2 Ca2 region, hippocampus Ca3 Ca3 region, hippocampus CG centralgray
CP CPu DG DLG E EC EN EPL FN GI
choroidplexus caudate-putamen dentate gyrus dorsolatera! geniculate ependymallayer, oil. bulb endopyriformcortex entorhinal cortex ext. plexiform layer, olf. bulb facialnuclear group glomerularcell layer, oil. bulb
GN Gr GrA IAM MD ME MG MN MOL NA
gelatinosusnu~ granulecell layer, main olf. bulb granulecell layer, accessolf. bulb interanteromedial thai. nu. mediodorsalthalamic nu. medianeminence medialgeniculate mammalarynuclear group cerebellum, molecular nuclear nucleusaccumbens
212 NF OC OpC PN Po PVN
nerve fiber layer, oil. bulb olfactory cortex optic chiasm pontine nuclear group posterior thai. nuclear group paraventricular nu.
Fig. 4 AN arcuate nucleus AOB accessory olfactory bulb (granular layer) AON anterior olfactory nu. (external. lateral)
RN RT SC SCN SON
reticular nu. complex reticularis thalamic nu. superior colliculus suprachiasmatic nu. supraoptic nu.
"FC VMN VPT II VI
outer tufted cell layer, olf. bulb ventromedial hypothalamicnu. ventroposterior thai. nuclear group cerebral cortex, layer 2 cerebral cortex, layer 6
CP CC DG GN ME
choroid plexus (lateral ventricle) cerebral cortex (layer VI) dentate gyrus gelatinosus nucleus median eminence
NA OC Sp5 VPL VPM
nucleus accumbens olfactory cortex (primary) spinal nucleus of trigeninai nerve ventroposteriorlateraithalamic nu. ventroposteriomedialthalamicnu.
REFERENCES 1 B~ickstrom,M., Hall, K. and Sara, V., Somatomedin levels in cerebrospinal fluid from adults with pituitary disorders, Ac~ ~ndocrinol., 107 (1984) 171-178. 2 Baskin, D. and Bohannon, N., Localization of binding sites for insulin-like growth factor-lI (IGF-II) in rat brain by in vitro quantitative autoradiography, Endocrinology, Suppl., 120 (1987) 42. 3 Baskin, D., Brewitt, B., Davidson, D., Corp, E., Paquette, T., Figlewicz, D., Graham, M., Woods, S. and Dorsa, D., Quantitative autoradiographic evidence for insulin receptors in the choroid plexus c,f the rat brain, Diabetes, 35 (1986) 246-249. 4 Baskin, D., Davidson, D., Corp, E., Lewellen, T. and Graham, M., An inexpensive microcomputer digital imaging system for densitometry: quantitative autoradiography of insulin receptors with 125Iand LKB Ultrofilm, J. Neurasci. Meth., 16 (1986) 119-129. 5 Baskin, D. and Dorsa, D., Quantitative autoradiography and in vitro radioligand binding. In D. McEachron (Ed.), Functional Mapping in Biology and Medicine: Computer Assisted Autoradiography, Experimental Biology and Medicine, Vol. II, Karger, New York, 1986, pp. 204-234. 6 Baskin, D., Figlewicz, D., Woods, S., Porte, Jr., D., and Dorsa, D., Insulin in the brain, Annu. Rev. Physiol., 49 (1987) 335-347. 7 Berelowicz, M., Firestone, S., and Frohman, L., Effects of growth hormone excess and deficiency on hypothalamic somatostatin content and release and on tissue somatostatin distribution, Endocrinology, 109 (1981) 714-719. 8 Berelowicz, M., Szabo, M., Frohman, L., Firestone, S., Chu, L. and Hintz, R., Somatomedin-C mediates growth hormone negative feedback by effects on both the hypothalamus and pituitary, Science, 212 (1981) 1279-1281. 9 Binoux, M., Faivre-Bauman, A., Lassarre, C., Barret, A. and Tixier-Vidal, A., Triiodothyronine stimulates production of insulin-like growth factor (IGF) by fetal hypothalamus cells cultured in serum-free medium, Dev. Brain Res., 21 (1985) 319-321. 10 Binoux, M., Hossenlopp, P., Lassarre, C. and Hardouin, N., Production of insulin-like growth factors and their carrier by rat pituitary gland and brain explants in culture, FEBS Lett., 124 (1981) 178-184. 11 Bohannon, N., Corp, E., Wilcox, B., Figlewicz, D., Dorsa, D. and Baskin, D., Insulin-like growth factor-I (IGF-I) binding in the median eminence of the rat is increased by
food restriction, Endocrinology, Suppl., 120 (1987) 42. 12 Bohannon, N., Corp, E., Wilcox, B., Porte, D., Jr., Figlewicz, D. and Baskin, D., Identification of specific binding sites for insulin-like growth factor-I (IGF-I) in the rat brain by in vitro quantitative autoradiography, Appetite, 7 (1986) 243-244. 13 Bohannon, N., Figlewicz, D., Corp, E., Wilcox, B., Porte, !r., D., and Baskin, D., Identification of binding sites for an insulin-like growth factor (IGF-I) in the median eminence of the rat brain by quantitative autoradiography, Endocrinology, 119 (1986) 943-945. 14 Brazeau, P., Guillemin, R., Ling, N., Van Wyk, J. and Humbel, R., Inhibition par les somatomedines de la secretion de l'hormone de croissance stimulee par le facteur hypothalamique somatocrinine (GRF) ou le peptide de synthese hpGRF, C.,'~. Seances Acad. Sci. (IH) (Paris), 295 (1982) 651-654. 15 Burgess, S., Jacobs, S., Cuatrecasas, P. and Sahyoun, N., Characterization of a neuronal subtype of insulin-like growth factor I receptor, J. Biol. Chem., 262 (1987) 1618-1622. 16 Ceda, G., Hoffman, A., Silverberg, G., Wilson, D. and Rosenfeld, R., Regulation of growth hormone release from cultured human pituitary adenomas by somatomedins and insulin, J. Clin. Endocrinol. Metab., 60 (1986) 1204-1209. 17 Clemmons, D. and Van Wyk, J., Factors controlling blood concentrations of somatomedin C Clinics, In W.H. Daughaday (Ed.), Clinics in Endocrinology and Metabolism, Vol. 13, No. 1, Tissue Growth Factors, Saunders, New York, 1984, pp. 113-144. 18 Corp, E., Woods, S., Porte, Jr., D., Figlewicz, D. and Baskin, D., Localization of 125I-insulinbinding sites in the rat hypothalamus by quantitative autoradiograhy, Neurosci. Lett., 70 (1986) 17-22. 19 Craig, A. and Burton, H., Spinal and medullary lamina I projection to nucleus submedius in medial thalamus: a possible pain center, J. Neurophysiol., 45 (1981) 443-466. 20 D'Ercole, A., Applewhite, G. and Underwood, L., Evi. dence that somatomedin is synthesized by multiple tissues in th,.-fetus, Dev. Biol., 75 (1980) 315-328. 21 Dorsa, D.M. and Baskin, D.G., Identification of neuropeptide receptors. In A. Boulton, G. Baker and Q. Pittman (Eds.), Neuromethods, Vol. 6, Humana, Clifton, NJ, 1987, pp. 203-244. 22 Froesch, E. and Zapf, J., Insulin-like growth factors and insulin: comparative aspects, Diabetologia, 28 (1985) 485-493.
213 23 Figlewicz, D., Baskin, D., Woods, S. and Porte, Jr., D., Insulin and insulin-like growth factor-I (IGF-I) binding is decreased in the olfactory bulb (OB) but not the cerebral cortex (C) of the obese Zucker (fa/fa) rat, Int. J. Obesity, 9 (1985) 233. 24 Finegold, E., Seshadri, S.B. and Tretiak, O.J., Hardware and software design considerations in engineering an image processing workstation: autoradiographic analysis with DUMAS and the BRAIN autoradiograph analysis software package. In D. McEachron (Ed.), Functional Mapping in Biology and Medicine: Computer Assisted Autoradiography, Experimental Biology and Medicine, Vol, II, Karger, New York, 1986, pp. 175-201. 25 Frank, H., Pardridge, W., Morris, W., Rosenfeld, R. and Choi, T., Binding and internalization of insulin-like growth factors by isolated brain microvessels, Diabetes, 35 (1986) 654-661. 26 Gammeltoft, S., Hasselbacher, G., Humbel, R., Fehlmann, M. and Van Obberghen, E., Two types of receptor for insulin-like growth factors in mammalian brain, EMBO J., 4 (1985) 3407-3412. 27 Goodlier, C., de Stephano, L., Lai, W., Guyda, H. and Posne~, B., Characterization of insulin-like growth factor receptors in rat anterior pituitary, hypothalamus, and brain, Endocrinology, 114 (1984) 1187-1195. 28 Hasselbacher, G., Schwab, M., Pasi, A. and Humbel, R., Insulin-like growth factor II (IGF-II) in human brain: regional distribution of IGF-II and of higher molecular mass forms, Proc. Natl. Acad. Sci. U.S.A., 82 (1985) 2153-2157. 29 Lauterio, T., Marson, L., Daughaday, W. and Baile, C., Intracerebroventricular injections of insulin-like growth factor II, but not insulin-like growth factor I, decrease food intake and body weight in rats, Soc. Neurosci. Abstr., 12 (1986) 795. 30 Lenoir, D. and Honegger, P., Insulin-like growth factor I (IGF-I) stimulates DNA synthesis in fetal rat brain cell cultures, Dev. Brain Res., 7 (1983) 205-213. 31 McMorris, F., Smith, T., DeSalvo. S. and Furlanetto, R.. Insulin-growth factor I/Somatomedin C: a potent inducer of oligodendrocyte development, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 822-826. 32 Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Coordinates, Academic, New York, 1982. 33 Rail, L., Falloon, L., Styne, D. and Bell, G., Insulin-like growth factor I and II mRNA are present in numerous tissues of the primate, Ciin. Res., 34 (1986) 135A. 34 Rosenfeld, R., Ceda, G., Cutler, C. and Hoffman, A., Insulin and insulin-like growth factor (somatomedin) recep-
tors on cloned rat pituitary tumor cells, Endocrinology, 117 (1985) 2008-2016. 35 Rosenfeld, R., Ceda, G., Wilson, D., Dollar, L. and Huffman, A., Characterization of high affinity receptors for insulin and insulin-like growth factors I and II on rat anterior pituitary cells, Endocrinology, 114 (1985) 1571-1575. 36 Rosenfeld, R.G., Conover, C.A., Hodges, D., Lee, P.D.K., Misra, P., Hintz, R.L. and Li, ~.H., Heterogeneity of insulin-like growth factor-I affinity for the insulin-like growth factor-II receptor: comparison of natural, synthetic and recombinant DNA-derived insulin-like growth factorI, Biochem. Biophys. Res. Comm., 143 (1987) 199-205. 37 Rosenfeld, R.G., Hodges, D., Pham, H., Lee, P.D.K. and Poweli, D.R., Purification of the insulin-like growth factorlI (IGF-II) receptor from an IGF-II-producing cell line, and generation of an antibody which both immunoprecipitates and blocks the type 2 IGF receptor, Biochem. Biophys. Res. Comm., 138 (1986) 304-311. 38 Sara, V., Hall, K., Misaki, M., Fryklund, L , Christensen, N. and Wetterberg, L., Ontogenesis of somatomedin and insulin receptors in the human fetus, J. Clin. Invest., 71 (1983) 1084-1094. 39 Sara, V., Hall, K., Von Holz, H., Humbei, R., Sjogren, B. and Wetterberg. L., Evidence for the presence of specific receptors for insulin-like growth factors I (IGF-I) and II (IGF-2) and insulin throughout the adult human brain, Neurosci. Lett., 34 (1982) 39-44. 40 Sara, V., Uvnas-Moberg, K., Uvnas, B., Hall, K., Wetterberg, L., Posloncec, B. and Goiny, M., The distribution of somatomedins in the nervous system of the cat and their release following neural stimulation, Acta. Physiol. Scand., 115 (1982) 467-470. 41 Tannenbaum, G., Guyda, H. and Posner, B., Insulin-like growth factors: a role in growth hormone negative feedback and body weight regulation via brain, Science, 220 (1983) 77-79. 42 Wise, S.P., The laminar organization of certain afferent and efferem fiber systems in the rat somatosensory cortex, Brain Research, 90 (1975) 139-142. 43 Wise, S.P. and Jones, E.G., Developmental studies of thalamocortical and commissurai connections in the rat somatosensory cortex, J. Comp. Neurol., 178 (1978) 187-208. 44 Woods, S. and Porte, Jr., D., The role of insulin as a satiety factor in the central nervous system. In A. Szabo (Ed.), CNS Regulation of Carbohydrate Metabolism, Academic Press, New York, pp. 1983, 457-468. 45 Zangger, I., Eberwine, J., Barchas, J., Jansen, J., Rosenfeld, R., Hoffman, A. and Hintz, R., Tissue distribution of IGF-I mRNA, Clin. Res., 34 (1986) 68A.