- Email: [email protected]
Human blood-brain barrier insulin-like growth factor receptor
Human -Blood-Brain
Barrier Insulin-Like
Growth Factor Receptor
Kent R. Duffy, William M. Pardridge, and Ron G. Rosenfeld Insulin-like growth factorB(1 and IGF-2. may be important regulatory molecules in the CNS. Possible origins of IGFs in brain include either de novo synthesis or transport of circulating IGFs from blood into brain via receptor mediated transcytosis mechanisms at the brain capillary endothelial wall, ie. the blood-brain barrier (EBB). In the present studies, isolated human brain capillaries are used as an in vitro model system of the human BBB and the characteristics of IGF-1 or IGF-2 binding to this preparation were assessed. The total binding of IGF-2 at 37°C exceeded 130% par mg protein and was threefold greater than the total binding for IGF-1. However, at 37°C nonsaturable binding equaled total binding, suggesting that endocytosis is rate limiting at physiologic temperatures. Binding studies performed at 4°C slowed andocytosis to a greater extent than membrane binding, and specific binding of either IGF-1 or IGF-2 was detectable. Scatchard plots for either peptide were linear and the molar dissociation constant of IGF-1 and IGF-2 binding was 2.1 f 0.4 and 1.1. f 0.1 nmol/L, respectively. Superphysiologic concentrations of porcine insulin inhibited the binding of both IGF-1 (ED, = 2 pg/mL) and IGF-2 (ED, = 0.5 ug/mL). Affinity cross linking of ‘*61-IGF-1, ‘261-lGF-2,and ‘261-insulinto isolated human brain capillaries was performed using disuccinimidylsuberate (DSS). These studies revealed a 141 kd binding site for both IGF-1 and IGF-2, and a 133 kd binding site for insulin. No high molecular weight binding site for IGF-2 was detected. In conclusion, these studies suggest that a type 1 IGF receptor is present on isolated human brain capillaries that binds both IGFs, with a slightly higher affinity for IGF-2. The threefold greater activity of this receptor for IGF-2 relative to IFG-1 parallels the much higher concentration of IGF-2 in human brain and CSF (relative to IGF-1). These results are consistent with the hypothesis that the human BBB IGF receptor is a transport system for the circulating peptides. particularly for IGF-2. B 1988 by Grune & Stratton, Inc.
I
NSULIN-LIKE growth factor (IGF)-1 and IGF-2, like insulin, have receptors on a wide variety of cells,“’ including brain cells. **9It has been reported that the brain concentration of IGF-2 is elevated in megaloencephaly’ and that brain IGF levels may be decreased in Down’s syndrome.” Thus, the origin of brain IGFs may have pathologic significance. Brain IGF- 1 and IGF-2 may arise from de novo synthesis in brain. A second possible source of brain IGFs is blood via transport through the brain capillary wall, ie, the blood-brain barrier (BBB).” Recent studies using isolated human brain capillaries as an in vitro model system of the human BBB have documented the presence of specific high affinity receptors for both insulin and transferrin.‘2.‘3 Since these peptides are both bound and endocytosed by human brain capillaries, it has been hypothesized that the human BBB receptors for insulin and transferrin are actually transport systems that mediate the blood to brain exchange of these peptides,“*13 similar to insulin transcytosis through the animal BBB, which has been documented by in vivo studies.14
From the Departments of Medicine/Endocrinology and Surgery/ Neurosurgery, UCLA School of Medicine, Los Angeles: and the Department of Pediatrics, Stanford University School of Medicine. Stanford, CA. Supported by the Juvenile Diabeles Foundation, NIH grants ROI-AM28229 and RCDA AM-01275, and the US Army Medical Research and Development Command under Contract No. DAMD17-87-C-7137. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the US Army. Presented at the Western Section American Federation of Clinical Research, Carmel, CA in February 1987 and at the American Society of Clinical Investigation, San Diego in May 1987. Address reprint requests to William M. Pardridge, MD, Department of Medicine, Division of Endocrinology, UCLA School of Medicine, Los Angeles, CA 90024. 0 1988 by Grune & Stratton, Inc. 0026/0495/88/3702-0007$03.00/O 136
Therefore, the present studies were designed to investigate whether specific receptors for IGF-1 and IGF-2 can be found on isolated human brain capillaries. The present studies specifically investigate whether the binding and endocytosis of IGF-2 is greater than that of IGF-1 in isolated human brain capillaries, since previous studies have shown that brain and CSF levels of IGF-2 are considerably higher than that of IGF- 1.” If the BBB IGF receptor acts as a transport system, then it would be expected that the receptor is more active for IGF-2 than for IGF- 1. MATERIALS AND METHODS
Porcine insulin was obtained from Mary Root (Eli Lilly, Indianapolis). Pure IGF-1 for iodination was prepared from human serum as described previously.‘6.‘7After acid-ethanol extraction, gel filtration chromatography, thin layer isoelectric focusing, and Sephadex G-SO chromatography, fractions were collected and assayed by a specific radioimmunoassay (RIA) using an antibody provided by Drs L.E. Underwood and J.J. Van Wyk (University of North Carolina at Chapel Hill), and distributed for research by the National Hormone and Pituitary Distribution Program. The isolated active peptide had a pI of 8.0 and no detectable IGF-2.” Pure IGF-2 for iodination was the generous gift of Enberg and Hall (Stockholm). The amino acid composition and sequence analysis up to position 39 of this peptide is identical to the previously reported analysis for IGF-2.19 Saturation of peptide binding was assayed with the partially purified IGF preparation isolated from an acid-ethanol extract of Cohn Fraction IV, which was partially purified by gel filtration chromatography. The IGF-I and IGF-2 contents of this preparation were 25 and 18 ng/pg dry weight as determined by RIA. Insulin, IGF-1 and IGF-2 were iodinated with ‘a?-iodine and chloramine T using methods previously described.” The peptides were periodically purified using Sephadex G-50 column chromatogacid precipitaraphy as described previously. ‘2.20The trichloroacetic bility of all iodinated peptides exceeded 97%. Isolated human brain capillaries were prepared as described previously.‘* During the course of this study, capillaries were isolated from four autopsy brains (four males, ages 52-64 years) that were delivered from the pathologist between 20 to 30 hours postmortem. None of the subjects had neurologic disease. Capillaries were Metabofism,
Vol37,
No 2
(February), 1988: pp 136-140
HUMAN BBB IGF RECEPTOR
137
obtained in high yield (4 to 8 mg protein capillary per 10 g cortex) and were pure as judged by light microscopy. Photomicrographs of the capillary preparation have been reported previously.‘z.2’ The isolated capillaries have very low adenosine triphosphate (ATP) levels,” but are positive for factor VIII, gamma-glutamyl transpeptidase, and have insulin receptor activity and phenylalanine transport activity that is indistinguishable from that found in capillaries isolated from fresh bovine, rat, or rabbit brain.‘2~2’ Time course and saturation analysis of “51-GGF-l and “‘1-IGF-2 binding to isolated human brain capillaries were performed as described previously at 37°C and 4OC in 0.01 mol/L Hepes buffered (pH = 7.4) Ringer’s buffer (RHB).!* Time courses were performed for up to three hours in the presence of either 1 or 200 ng/mL IGF. ‘H-inulin was used as a marker of nonspecific uptake, including fluid phase uptake by capillaries during the incubation period, or entrapment of fluid in the pellet formed by centrifugation. Scatchard analysis was performed using a nonlinear regression analysis.12 In some studies, 50% human serum was also added to the incubation medium to assess for the effects of IGF binding proteins in human serum. Previous studies with either ‘r51-insulin or “‘1-1GFs have shown that the nonsaturable uptake of peptide (eg, at 100 to 200 ng/mL concentrations) is equal to the uptake of hormone into a postmembrane pool of the capillary that is resistant to mild acid wash.‘***’ On this basis, it is assumed in the present studies that the component of capillary uptake of IGF that is nonsaturable represents endocytosed peptide. The possibility that the endocytosed ‘*‘I material represented degradation products of the IGFs was assessed by Sephadex G-50 gel filtration. After incubation of human brain capillaries at 4OC or 37°C for 30 minutes with ‘251-IGF-2, the cell pellet was solubilized in 2 mL of 6 mol/L guanidine and 0.1% trifluoroacetic acid at 4°C
14Or
.
(1251)-IGF-2
1
ng/mlIGF
0 200 ng/ml IGF
0
40
80
120
Minutes Fig 1. Time course of ‘z’I-IGF-l and ‘261-IGF-2 binding to isolated human brain capillaries in the presence of either 1 ng/mL IGF (closed circles) or 200 ng/mL IGF (open circles) at either 37°C (A) or 4°C 16). The uptake of ‘H-inulin, en extracellular marker, was not significantly different at either temperature and the values observed at 37% ere shown in the figure (closed squares). There was no significant difference between the uptake of the IGFs in the presence of 1 or 200 ng/mL IGF at 37°C. However, nonspecific binding was about half that of total binding et 4°C.
’
overnight. The solution was microfuged for 30 seconds and applied to a 1.6 x 60 cm column of G-SO (fine), which was eluted with 0.1 mol/L (NH&CO, (pH, 7.4) containing 0.1 g/dL bovine albumin. More than 90% of the cell radioactivity comigrated (fraction 38, 2 mL fractions) with the “‘1-IGF-2 standard after either the 37“C or 4°C incubations. The remaining ~10% of the ‘*‘I migrated at the void volume (fraction 22) and no radioactivity migrated near the salt volume (fraction 60) indicating no significant degradation of the “‘1-IGF-2 at either 37°C or 4°C. The molecular weight of the ‘251-IGF-l, ‘*51-1GF-2, and i*‘Iinsulin receptor binding subunit was assessed by affinity cross linking with disuccinimidylsuberate (DSS) as described previously.‘* Briefly, 2 &i/mL concentrations of labeled peptides in 200 PL of RHB containing approximately 350 rg protein of isolated capillaries was incubated for 180 minutes at 4°C. At the end of this incubation, IO PL of 25 mmol/L DSS in 100% dimethylsulfoxide was added, followed by a 15minute incubation on ice. The reaction was quenched by the addition of 1.5 mL 0.15 mol/L Tris (pH = 7.4) followed by centrifugation at 1,000 g at 4°C for five minutes. The cross-linked capillaries were solubilized in sodium dodecylsulfate (SDS) sample buffer, boiled for five minutes, and applied to a 4% stacking gel with a 7.5% separating gel and run at constant voltage in the presence of 5% &mercaptoethanol. The gels were then dried and exposed to Kodak XAR-5 film with a DuPont Cronex intensifying screen. High molecular weight markers (Bio Rad, Richmond, CA)
were used to calibrate the gel. The binding of ‘ZSI-IGF-I to human serum protein was determined by equilibrium dialysis at 4OC for 48 hours using Visking dialysis tubing with a pore size excluding compounds 2 12,000 daltons. RESULTS
Both ‘?-IGF-1 and ‘251-IGF-2 were bound by isolated human brain capillaries at 37OC at rates considerably faster than that for ‘H-inulin, a compound of comparable molecular weight (Fig 1). The uptake of 3H-inulin was independent of time and temperature (3.1 + 0.2%/mg protein at 37OC; 2.7 k 0.3%/mg protein at 4V). However, for both ‘*?IGF-1 and ‘Z51-IGF-2 the entire binding at 37OC was nonsaturable over the range of 1 to 200 ng/mL IGF (Fig 1). Assuming nonsaturable uptake of peptide represents endocytosis,“,*’ these data suggested that endocytosis is much faster than plasma membrane binding at physiologic temperatures. Therefore, the incubations were performed at 4OC (Fig 1). The lower temperatures lowered the nonspecific binding to a greater extent than it did the total binding, and allowed for detection of saturable binding of either ‘251-IGF-1 or lz51IGF-2 to the isolated human brain capillaries (Fig 1). Therefore, all saturation analyses were performed at 4OC. As shown in Fig 2, the specific binding of ‘251-IGF-l was 50% displaced by 19 ng/mL of IGF-1, whereas the specific binding of ‘251-IGF-2 to isolated human brain capillaries was 50% displaced by approximately 14 ng/mL IGF-2. Large concentrations of unlabeled insulin displaced the binding of both IGF-1 and IGF-2 to isolated human brain capillaries. The binding of IGF-1 and IGF-2 was 50% inhibited by 2 rg/mL and 0.5 I.rg/mL insulin, respectively. Scatchard analyses of the saturation data in Fig 2 are shown in Fig 3. The capacity (R,) of IGF-1 was not significantly different from that of IGF-2, but the affinity of IGF-2 binding to isolated human brain capillaries was approximately twofold greater than the affinity of IGF-1 binding. When human serum was
138
DUFFY, PARDRIDGE, AND ROSENFELD
nq/ml
added to the incubation medium, the total binding and uptake of IGF-1 or IGF-2 at 4‘C was greatly inhibited, as shown in Fig 4. The nonspecific uptake of ‘H-inulin, 2.7 f 0.3%/mg protein (Fig l), is significantly lower than the uptake of either IGF-1 (5.4 A 0.7%/mg protein) or IGF-2 (6.3 k OJ?%/mg protein) by isolated human brain capillaries in the presence of 50% human serum. This serum sample by bound more than 99% of ‘251-IGF-1 as determined equilibrium dialysis. The results of the affinity cross-linking experiments for
0.4
Fig 2. The specific binding of “‘IIGF-2 and ‘*‘I-IGF-1 by isolated human brain capillaries at 4°C is shown in the presence of varying concentrations of either IGF-1, IGF-2, or porcine insulin. IGF-1 binding was 50% displaced by 19 ng/mL IGF-1 or 2 pg/mL porcine insulin. The specific binding of ‘*slIGF-2 was 50% displaced by 14 ng/mL IGF-2 or 0.5 pg.mL porcine insulin.
‘251-IGF- 1, ‘251-IGF-2, and ‘251-insulin are shown in Fig 5. The binding of ‘251-IGF-1 was nearly completely displaced by 250 ng/mL of IGF, but there was no visible displacement by 400 ng/mL insulin. Similarly, the binding of ‘251-IGF-2 was essentially completely displaced by 250 ng/mL of IGF, but there was a partial displacement of ‘251-IGF-2 by 400 ng/mL insulin (Fig 5). The binding of ‘*?-insulin was nearly completely displaced by 400 ng/mL insulin and approximately 50% displaced by 250 ng/mL of IGF. The molecular weight of the IGF-1 or IGF-2 binding subunit was equal at 141 kd, whereas the molecular weight of the insulin binding subunit was slightly lower at 133 kd. No high molecular weight (eg, 250 kd) receptor for IGF-2 was detected (Fig 5).
IGF-2
n
I\
KD = 7.9 t 0.6ng/ml R. = 1.6tO.lng/mgp
TEl
DISCUSSION
The mitogenic effect of insulin-like growth factors on peripheral tissues may also extend to the CNS.22 Therefore,
Kg = 15.9 + 2.9 ng/ml
.
Control
0
lOOng/ml
a 50% human
IGF serum
B
Bound
(ng /ml) Hours
Scatchard analysis of the saturation data shown in Fig Fig 3. 2. The maximal binding (l&I and dissociation constant (K,I were determined by nonlinear regression analysis using the method of least squares. The K, values are slightly lower than the ED, values shown in Fig 2, as the latter were determined by eye.
Total binding of ‘261-lGF-2 (Al and ‘261-IGF-1 (B) by Fig 4. isolated human brain capillaries at 4°C in the presence of 1 ng/mL IGF (closed circles), 100 ng/mL IGF (open circles). or 50% human serum (closed squares).
HUMAN
BBB IGF RECEPTOR
Fig 5. Affinity cross linking of ‘261-IGF-1,‘261-IGF-2,or “61-insulin to isolated human brain capillaries using DSS. The capillaries were incubated with labeled hormone in the presence of either no additive, 259 ng/mL IGF, or 409 ng/mL porcine insulin. The migration of molecular weight markers is shown on the right-hand side of the figure. The molecular weight of the insulin-binding site (133 kd) is shown on the right-hand side of the figure. The molecular weight of the IGF binding site (141 kd) is shown on the lehhand side of the figure. The SDS solubilized affinity cross linked capillary proteins were separated on a 7.5% SDS polyacrylamide gel before slab drying and autoradiography.
‘2”I-IGF-1
+
‘“%-IGF- 2 f25f- insutin IGFG!SO~) insulin (400
++
+ +
-
3) -
it is important to define the possible origins of the IGFs in brain. De novo synthesis and neurosecretion of the IGFs is one possible source of IGFs in brain. Another possible pathway is transport of circulating IGFs from blood into brain through receptor mediated transcytosis mechanisms at the BBB.” Indeed, previous studies with human brain capillaries have characterized the human BBB insulin and transferrin receptors.‘2’3 The present results extend these findings to the human BBB 1GF receptor. Studies of IGF receptors in peripheral tissues demonstrate a type 1 receptor with a 130 kd molecular weight for the binding subunit and a type 2 receptor with an approximately 260 kd molecular weight.*,” Although concentrations of approximately 1 to 10 pg/mL insulin are known to displace IGFs from the type 1 receptor, concentrations of insulin in excess of 100 wg/mL are believed not to displace IGFs from the type 2 receptor.5~7*23Previous studies with bovine brain are consistent with the presence of both the type 1 and type 2 receptor on bovine brain capillaries, since approximately 1 to 10 pg/mL insulin displaced IGF-1 from isolated bovine brain capillaries but 100 pg/mL of insulin exerted no displacement of IGF-2 from isolated bovine brain capillaries.*’ The pattern of peptide competition is different with isolated human brain capillaries, since 0.5 pg/mL concentrations of insulin displaced “‘1-IGF-2 from binding sites on isolated human brain capillaries (Fig 2). Another area of difference concerning the interaction of isolated IGFs with human brain capillaries, vis-a-vis bovine brain capillaries, is the fact that at 37%, binding, not internalization, is the rate limiting step in peptide uptake. This conclusion is drawn from the observation that the uptake of either IGF-1 or IGF-2 by isolated human brain capillaries is nonsaturable by 200 ng/mL IGF at 37°C (Fig 1). Previous studies with isolated human brain capillaries have shown that the non-
+
c
-
+
+m.-f--+-+
--
+-
saturable component of uptake is identical to the portion that is resistant to a mild acid wash,‘*,*’ which presumably represents either partial or complete internalization of the peptide within the endothelial cytoplasm. Conversely, in isolated bovine brain capillaries, internalization is always the slow step compared with receptor binding even at physiologic temperatures.” The affinity cross linking studies suggest only a type 1 receptor is present in human brain capillaries with a molecular weight of the binding subunit of 141 kd (Fig 5). The affinity cross linking studies are also consistent with the cross competition results shown in Fig 2, in that insulin is more effective in displacing ‘*‘I-IGF-2 from its binding site on human brain capillaries as compared with the binding of ‘251-IGF-1. Thus, the overall results suggest that only a type 1 receptor exists on isolated human brain capillaries and that this receptor has a higher affinity for IGF-2 than for IGF-1. This is compatible with the recent hypothesis of Casella et alz4 that there are two binding sites on the type 1 receptor in human placental membranes. Site A preferentially binds IGF-1 and site B preferentially binds IGF-2.24 Similarly, the human BBB IGF receptor appears to be a type 1 receptor that binds IGF-2 to a slightly greater extent than IGF- 1. The putative B site binds insulin weakly but fourfold greater than does the A (or IGF-1) binding site (Figs 2 and 5). However, the conclusions of the present study are provisional because a mixture of IGF-1 and IGF-2 was used in the displacement experiments (Figs 2, 3, and 5). The model of dual binding sites on the human BBB IGF receptor will be clarified by future studies using purified IGF-1 and IGF-2. The idea that the human BBB IGF receptor acts as a transport system is buttressed by the parallel observations made in the present study and in studies of IGF content in human brain. The concentration of IGF-2 in either human
140
DUFFY, PARDRIDGE, AND ROSENFELD
brain or in human CSF is much higher than the concentration of IGF-l.lS Moreover, the kinetics of either receptor binding to the surface of the isolated human brain capillary or the internalization step is at least three times as fast for IGF-2 as compared with IGF-1 (Fig 1). Endocytosis is only half of the overall transcytosis pathway and the demonstration of endocytosis is not proof that transcytosis of IGFs through the BBB occurs in vivo. However, the initial studies showing endocytosis of insulin” and transferrin” have been confirmed by in vivo experiments documenting transcytosis of insulinI and of transferrin” through the BBB. Finally, the limited study shown in Fig 4 regarding the effects of IGF binding proteins in human serum suggest that binding proteins greatly retard access of the peptide to the receptor. However, since more than 99% of serum IGF is
bound by the binding proteins:6 it is surprising that the presence of human serum in the incubation medium did not lower the uptake of the IGFs to the values seen for an extracellular marker such as ‘H-inulin (Fig 1). This suggests that there may be partial access of the bound IGF to the receptor, perhaps as a binding protein-IGF complex. In this regard, it is of interest that the concentration of the IGF-2 binding protein in human CSF is high compared with that for the IGF-1 binding protein.27
ACKNOWLEDGMENT
The authors are indebted to Jody Eisenberg and to Jing Yang for valuable technical assistance, and to Dawn Brown who skillfully prepared the manuscript.
REFERENCES
1. Adashi EY, Resnick CE, D’Ercole AJ, et al: Insulin-like growth factors as intraovarian regulators of granulosa cell growth and function. Endocr Rev 6:400-405, 1985 2. Bar RS, Boes M: Distinct receptors for IGF-I, IGF-II, and insulin are present on bovine capillary endothelial cells and large vessel endothelial cells. Biochem Biophys Res Commun 124:203207, 1984 3. Catanese VM, Grigorescu F, King GL et al: The human erythrocyte insulin-like growth factor I receptor: Characterization and demonstration of ligand-stimulated autophosphorylation. J Clin Endocrinol Metab 62:692-701, 1986 4. Hammerman MR. Gavin JR III: Binding of insulin-like growth factor II and multiplication-stimulating activity-stimulated phosphorylation in basolateral membranes from dog kidney. J Biol Chem 259:13511-13518,1984 5. Oppenheimer CL, Pessin JE, Massague J, et al: Insulin action rapidly modulates the apparent affinity of the insulin-like growth factor II receptor. J Biol Chem 258:4824-4832, 1983 6. Rechler MM, Nissley SP, Podskalny JM, et al: Identification of a receptor for somatomedin-like polypeptides in human fibroblasts. J Clin Endocrinol Metab 44:820-826, 1977 7. Rosenfeld RG, Ceda G, Wilson DM, et al: Characterization of high affinity receptors for insulin-like growth factors I and II on rat anterior pituitary cells. Endocrinology 114:1571-1576, 1984 8. Lesniak MA, Hill JM, Pert CB, et al: Autoradiographic localization of ‘2SI-IGFI and ‘“*I-rIGFI1 receptors in rat brain: Comparison with insulin receptors. Diabetes 35:58A, 1986 (suppl)
IGF II and of higher molecular USA 82:2153-2157, 1985
mass forms.
Proc Nat1 Acad
Sci
16. Rosenfeld RG, Dollar LA: Characterization of the somatomedin-C/insulin-like growth factor I (SM-C/IGF-I) receptor on cultured human fibroblast monolayers: Regulation of receptor concentrations by SM-C/IGF-I and insulin. J Clin Endocrinol Metab 55:434-441, 1982 17. Rosenfeld RG, Thorsson AV, Hintz RL: Increased somatomedin receptors in newborn circulating mononuclear cells. J Clin Endocrinol Metab 48:456-462, 1979 18. Hintz RL, Liu F: A radioimmunoassay for insulin-like growth factor-2 specific for the C-peptide region. J Clin Endocrinol Metab 54:442-449, 1982 19. Enberg G, Carlquist M, Joinvall H, et al: The chromatography of somatomedin A, isolated by microcomputer-controlled chromatography, reveals an apparent identity to insulin-like growth factor I. Eur J Biochem 143:117-120, 1984 20. Frank HJL, Pardridge WM, Morris WL, et al: Binding and internalization of insulin and insulin-like growth factors by isolated brain microvessels. Diabetes 35:654-661, 1986 21. Choi T, Pardridge WM: Phenylalanine transport at the human blood-brain barrier. Studies in isolated human brain capillaries. J Biol Chem 261:6536-6541, 1986 22. Sara VR, King, TL, Stuart MC, et al: Hormonal regulation of fetal brain cell proliferation: Presence in serum of a trophin responsive to pituitary growth hormone stimulation. Endocrinology 99:1512-1517, 1976
9. Sara VR, Hall K, Misaki M, et al: Ontogenesis of somatomedin and insulin receptors in the human fetus. J Clin Invest 71:10841089,1983 10. Wisniewski KE, Kamionowska ML, Wisniewski HM: Evidence of arrest of neurogenesis and synaptogenesis in brains of patients with Down’s syndrome. N Engl J Med 311:1187-l 191, 1984
23. DeVroede MA, Romanus JA, Standaert ML, et al: Interaction of insulin-like growth factors with a nonfusing mouse muscle cell line: Binding, action, and receptor down-regulation. Endocrinology 114:1917-1922, 1984
11. Pardridge WM: Receptor-mediated peptide transport through the blood-brain barrier. Endocr Rev 7:314-330, 1986
25. Fishman J, Rubin JB, Handarhan JV, et al: Receptormediated transcytosis of transferrin across the blood-brain barrier. J Neurosci Res (in press)
12. Pardridge WM. barrier insulin receptor.
Eisenberg J, Yang J: Human blood-brain J Neurochem 44:1771-1778, 1985
13. Pardridge WM, Eisenberg J, Yang J: Human blood-brain barrier transferrin receptor. Metabolism 36:892-895, 1987 14. Duffy KR, Pardridge WM: Blood brain barrier transcytosis of insulin in developing rabbits. Brain Res 420:32-38, 1987 15. Haselbacher GK, Schwab ME, Pasi A, et al: Insulin-like growth factor II (IGF II) in human brain: Regional distribution of
24. Casella SJ, Han VK, D’Ercole AJ, et al: Insulin-like growth factor II binding to the type I somatomedin receptor. Evidence for two high affinity binding sites. J Biol Chem 261:9268-9272, 1986
26. Smith GL: Somatomedin nol 34:83-90, 1984
carrier
proteins.
Mol Cell Endocri-
27. Binoux M, Hardouin S, Lassarre C, et al: production by the liver of two IGF binding proteins molecular weights but different affinities for IGF I Their relations with serum and cerebrospinal fluid proteins. J Clin Endocrinol Metab 55:600-660, 1982
Evidence for with similar and IGF II. IGF binding
Barrier Insulin-Like
Growth Factor Receptor
Kent R. Duffy, William M. Pardridge, and Ron G. Rosenfeld Insulin-like growth factorB(1 and IGF-2. may be important regulatory molecules in the CNS. Possible origins of IGFs in brain include either de novo synthesis or transport of circulating IGFs from blood into brain via receptor mediated transcytosis mechanisms at the brain capillary endothelial wall, ie. the blood-brain barrier (EBB). In the present studies, isolated human brain capillaries are used as an in vitro model system of the human BBB and the characteristics of IGF-1 or IGF-2 binding to this preparation were assessed. The total binding of IGF-2 at 37°C exceeded 130% par mg protein and was threefold greater than the total binding for IGF-1. However, at 37°C nonsaturable binding equaled total binding, suggesting that endocytosis is rate limiting at physiologic temperatures. Binding studies performed at 4°C slowed andocytosis to a greater extent than membrane binding, and specific binding of either IGF-1 or IGF-2 was detectable. Scatchard plots for either peptide were linear and the molar dissociation constant of IGF-1 and IGF-2 binding was 2.1 f 0.4 and 1.1. f 0.1 nmol/L, respectively. Superphysiologic concentrations of porcine insulin inhibited the binding of both IGF-1 (ED, = 2 pg/mL) and IGF-2 (ED, = 0.5 ug/mL). Affinity cross linking of ‘*61-IGF-1, ‘261-lGF-2,and ‘261-insulinto isolated human brain capillaries was performed using disuccinimidylsuberate (DSS). These studies revealed a 141 kd binding site for both IGF-1 and IGF-2, and a 133 kd binding site for insulin. No high molecular weight binding site for IGF-2 was detected. In conclusion, these studies suggest that a type 1 IGF receptor is present on isolated human brain capillaries that binds both IGFs, with a slightly higher affinity for IGF-2. The threefold greater activity of this receptor for IGF-2 relative to IFG-1 parallels the much higher concentration of IGF-2 in human brain and CSF (relative to IGF-1). These results are consistent with the hypothesis that the human BBB IGF receptor is a transport system for the circulating peptides. particularly for IGF-2. B 1988 by Grune & Stratton, Inc.
I
NSULIN-LIKE growth factor (IGF)-1 and IGF-2, like insulin, have receptors on a wide variety of cells,“’ including brain cells. **9It has been reported that the brain concentration of IGF-2 is elevated in megaloencephaly’ and that brain IGF levels may be decreased in Down’s syndrome.” Thus, the origin of brain IGFs may have pathologic significance. Brain IGF- 1 and IGF-2 may arise from de novo synthesis in brain. A second possible source of brain IGFs is blood via transport through the brain capillary wall, ie, the blood-brain barrier (BBB).” Recent studies using isolated human brain capillaries as an in vitro model system of the human BBB have documented the presence of specific high affinity receptors for both insulin and transferrin.‘2.‘3 Since these peptides are both bound and endocytosed by human brain capillaries, it has been hypothesized that the human BBB receptors for insulin and transferrin are actually transport systems that mediate the blood to brain exchange of these peptides,“*13 similar to insulin transcytosis through the animal BBB, which has been documented by in vivo studies.14
From the Departments of Medicine/Endocrinology and Surgery/ Neurosurgery, UCLA School of Medicine, Los Angeles: and the Department of Pediatrics, Stanford University School of Medicine. Stanford, CA. Supported by the Juvenile Diabeles Foundation, NIH grants ROI-AM28229 and RCDA AM-01275, and the US Army Medical Research and Development Command under Contract No. DAMD17-87-C-7137. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the US Army. Presented at the Western Section American Federation of Clinical Research, Carmel, CA in February 1987 and at the American Society of Clinical Investigation, San Diego in May 1987. Address reprint requests to William M. Pardridge, MD, Department of Medicine, Division of Endocrinology, UCLA School of Medicine, Los Angeles, CA 90024. 0 1988 by Grune & Stratton, Inc. 0026/0495/88/3702-0007$03.00/O 136
Therefore, the present studies were designed to investigate whether specific receptors for IGF-1 and IGF-2 can be found on isolated human brain capillaries. The present studies specifically investigate whether the binding and endocytosis of IGF-2 is greater than that of IGF-1 in isolated human brain capillaries, since previous studies have shown that brain and CSF levels of IGF-2 are considerably higher than that of IGF- 1.” If the BBB IGF receptor acts as a transport system, then it would be expected that the receptor is more active for IGF-2 than for IGF- 1. MATERIALS AND METHODS
Porcine insulin was obtained from Mary Root (Eli Lilly, Indianapolis). Pure IGF-1 for iodination was prepared from human serum as described previously.‘6.‘7After acid-ethanol extraction, gel filtration chromatography, thin layer isoelectric focusing, and Sephadex G-SO chromatography, fractions were collected and assayed by a specific radioimmunoassay (RIA) using an antibody provided by Drs L.E. Underwood and J.J. Van Wyk (University of North Carolina at Chapel Hill), and distributed for research by the National Hormone and Pituitary Distribution Program. The isolated active peptide had a pI of 8.0 and no detectable IGF-2.” Pure IGF-2 for iodination was the generous gift of Enberg and Hall (Stockholm). The amino acid composition and sequence analysis up to position 39 of this peptide is identical to the previously reported analysis for IGF-2.19 Saturation of peptide binding was assayed with the partially purified IGF preparation isolated from an acid-ethanol extract of Cohn Fraction IV, which was partially purified by gel filtration chromatography. The IGF-I and IGF-2 contents of this preparation were 25 and 18 ng/pg dry weight as determined by RIA. Insulin, IGF-1 and IGF-2 were iodinated with ‘a?-iodine and chloramine T using methods previously described.” The peptides were periodically purified using Sephadex G-50 column chromatogacid precipitaraphy as described previously. ‘2.20The trichloroacetic bility of all iodinated peptides exceeded 97%. Isolated human brain capillaries were prepared as described previously.‘* During the course of this study, capillaries were isolated from four autopsy brains (four males, ages 52-64 years) that were delivered from the pathologist between 20 to 30 hours postmortem. None of the subjects had neurologic disease. Capillaries were Metabofism,
Vol37,
No 2
(February), 1988: pp 136-140
HUMAN BBB IGF RECEPTOR
137
obtained in high yield (4 to 8 mg protein capillary per 10 g cortex) and were pure as judged by light microscopy. Photomicrographs of the capillary preparation have been reported previously.‘z.2’ The isolated capillaries have very low adenosine triphosphate (ATP) levels,” but are positive for factor VIII, gamma-glutamyl transpeptidase, and have insulin receptor activity and phenylalanine transport activity that is indistinguishable from that found in capillaries isolated from fresh bovine, rat, or rabbit brain.‘2~2’ Time course and saturation analysis of “51-GGF-l and “‘1-IGF-2 binding to isolated human brain capillaries were performed as described previously at 37°C and 4OC in 0.01 mol/L Hepes buffered (pH = 7.4) Ringer’s buffer (RHB).!* Time courses were performed for up to three hours in the presence of either 1 or 200 ng/mL IGF. ‘H-inulin was used as a marker of nonspecific uptake, including fluid phase uptake by capillaries during the incubation period, or entrapment of fluid in the pellet formed by centrifugation. Scatchard analysis was performed using a nonlinear regression analysis.12 In some studies, 50% human serum was also added to the incubation medium to assess for the effects of IGF binding proteins in human serum. Previous studies with either ‘r51-insulin or “‘1-1GFs have shown that the nonsaturable uptake of peptide (eg, at 100 to 200 ng/mL concentrations) is equal to the uptake of hormone into a postmembrane pool of the capillary that is resistant to mild acid wash.‘***’ On this basis, it is assumed in the present studies that the component of capillary uptake of IGF that is nonsaturable represents endocytosed peptide. The possibility that the endocytosed ‘*‘I material represented degradation products of the IGFs was assessed by Sephadex G-50 gel filtration. After incubation of human brain capillaries at 4OC or 37°C for 30 minutes with ‘251-IGF-2, the cell pellet was solubilized in 2 mL of 6 mol/L guanidine and 0.1% trifluoroacetic acid at 4°C
14Or
.
(1251)-IGF-2
1
ng/mlIGF
0 200 ng/ml IGF
0
40
80
120
Minutes Fig 1. Time course of ‘z’I-IGF-l and ‘261-IGF-2 binding to isolated human brain capillaries in the presence of either 1 ng/mL IGF (closed circles) or 200 ng/mL IGF (open circles) at either 37°C (A) or 4°C 16). The uptake of ‘H-inulin, en extracellular marker, was not significantly different at either temperature and the values observed at 37% ere shown in the figure (closed squares). There was no significant difference between the uptake of the IGFs in the presence of 1 or 200 ng/mL IGF at 37°C. However, nonspecific binding was about half that of total binding et 4°C.
’
overnight. The solution was microfuged for 30 seconds and applied to a 1.6 x 60 cm column of G-SO (fine), which was eluted with 0.1 mol/L (NH&CO, (pH, 7.4) containing 0.1 g/dL bovine albumin. More than 90% of the cell radioactivity comigrated (fraction 38, 2 mL fractions) with the “‘1-IGF-2 standard after either the 37“C or 4°C incubations. The remaining ~10% of the ‘*‘I migrated at the void volume (fraction 22) and no radioactivity migrated near the salt volume (fraction 60) indicating no significant degradation of the “‘1-IGF-2 at either 37°C or 4°C. The molecular weight of the ‘251-IGF-l, ‘*51-1GF-2, and i*‘Iinsulin receptor binding subunit was assessed by affinity cross linking with disuccinimidylsuberate (DSS) as described previously.‘* Briefly, 2 &i/mL concentrations of labeled peptides in 200 PL of RHB containing approximately 350 rg protein of isolated capillaries was incubated for 180 minutes at 4°C. At the end of this incubation, IO PL of 25 mmol/L DSS in 100% dimethylsulfoxide was added, followed by a 15minute incubation on ice. The reaction was quenched by the addition of 1.5 mL 0.15 mol/L Tris (pH = 7.4) followed by centrifugation at 1,000 g at 4°C for five minutes. The cross-linked capillaries were solubilized in sodium dodecylsulfate (SDS) sample buffer, boiled for five minutes, and applied to a 4% stacking gel with a 7.5% separating gel and run at constant voltage in the presence of 5% &mercaptoethanol. The gels were then dried and exposed to Kodak XAR-5 film with a DuPont Cronex intensifying screen. High molecular weight markers (Bio Rad, Richmond, CA)
were used to calibrate the gel. The binding of ‘ZSI-IGF-I to human serum protein was determined by equilibrium dialysis at 4OC for 48 hours using Visking dialysis tubing with a pore size excluding compounds 2 12,000 daltons. RESULTS
Both ‘?-IGF-1 and ‘251-IGF-2 were bound by isolated human brain capillaries at 37OC at rates considerably faster than that for ‘H-inulin, a compound of comparable molecular weight (Fig 1). The uptake of 3H-inulin was independent of time and temperature (3.1 + 0.2%/mg protein at 37OC; 2.7 k 0.3%/mg protein at 4V). However, for both ‘*?IGF-1 and ‘Z51-IGF-2 the entire binding at 37OC was nonsaturable over the range of 1 to 200 ng/mL IGF (Fig 1). Assuming nonsaturable uptake of peptide represents endocytosis,“,*’ these data suggested that endocytosis is much faster than plasma membrane binding at physiologic temperatures. Therefore, the incubations were performed at 4OC (Fig 1). The lower temperatures lowered the nonspecific binding to a greater extent than it did the total binding, and allowed for detection of saturable binding of either ‘251-IGF-1 or lz51IGF-2 to the isolated human brain capillaries (Fig 1). Therefore, all saturation analyses were performed at 4OC. As shown in Fig 2, the specific binding of ‘251-IGF-l was 50% displaced by 19 ng/mL of IGF-1, whereas the specific binding of ‘251-IGF-2 to isolated human brain capillaries was 50% displaced by approximately 14 ng/mL IGF-2. Large concentrations of unlabeled insulin displaced the binding of both IGF-1 and IGF-2 to isolated human brain capillaries. The binding of IGF-1 and IGF-2 was 50% inhibited by 2 rg/mL and 0.5 I.rg/mL insulin, respectively. Scatchard analyses of the saturation data in Fig 2 are shown in Fig 3. The capacity (R,) of IGF-1 was not significantly different from that of IGF-2, but the affinity of IGF-2 binding to isolated human brain capillaries was approximately twofold greater than the affinity of IGF-1 binding. When human serum was
138
DUFFY, PARDRIDGE, AND ROSENFELD
nq/ml
added to the incubation medium, the total binding and uptake of IGF-1 or IGF-2 at 4‘C was greatly inhibited, as shown in Fig 4. The nonspecific uptake of ‘H-inulin, 2.7 f 0.3%/mg protein (Fig l), is significantly lower than the uptake of either IGF-1 (5.4 A 0.7%/mg protein) or IGF-2 (6.3 k OJ?%/mg protein) by isolated human brain capillaries in the presence of 50% human serum. This serum sample by bound more than 99% of ‘251-IGF-1 as determined equilibrium dialysis. The results of the affinity cross-linking experiments for
0.4
Fig 2. The specific binding of “‘IIGF-2 and ‘*‘I-IGF-1 by isolated human brain capillaries at 4°C is shown in the presence of varying concentrations of either IGF-1, IGF-2, or porcine insulin. IGF-1 binding was 50% displaced by 19 ng/mL IGF-1 or 2 pg/mL porcine insulin. The specific binding of ‘*slIGF-2 was 50% displaced by 14 ng/mL IGF-2 or 0.5 pg.mL porcine insulin.
‘251-IGF- 1, ‘251-IGF-2, and ‘251-insulin are shown in Fig 5. The binding of ‘251-IGF-1 was nearly completely displaced by 250 ng/mL of IGF, but there was no visible displacement by 400 ng/mL insulin. Similarly, the binding of ‘251-IGF-2 was essentially completely displaced by 250 ng/mL of IGF, but there was a partial displacement of ‘251-IGF-2 by 400 ng/mL insulin (Fig 5). The binding of ‘*?-insulin was nearly completely displaced by 400 ng/mL insulin and approximately 50% displaced by 250 ng/mL of IGF. The molecular weight of the IGF-1 or IGF-2 binding subunit was equal at 141 kd, whereas the molecular weight of the insulin binding subunit was slightly lower at 133 kd. No high molecular weight (eg, 250 kd) receptor for IGF-2 was detected (Fig 5).
IGF-2
n
I\
KD = 7.9 t 0.6ng/ml R. = 1.6tO.lng/mgp
TEl
DISCUSSION
The mitogenic effect of insulin-like growth factors on peripheral tissues may also extend to the CNS.22 Therefore,
Kg = 15.9 + 2.9 ng/ml
.
Control
0
lOOng/ml
a 50% human
IGF serum
B
Bound
(ng /ml) Hours
Scatchard analysis of the saturation data shown in Fig Fig 3. 2. The maximal binding (l&I and dissociation constant (K,I were determined by nonlinear regression analysis using the method of least squares. The K, values are slightly lower than the ED, values shown in Fig 2, as the latter were determined by eye.
Total binding of ‘261-lGF-2 (Al and ‘261-IGF-1 (B) by Fig 4. isolated human brain capillaries at 4°C in the presence of 1 ng/mL IGF (closed circles), 100 ng/mL IGF (open circles). or 50% human serum (closed squares).
HUMAN
BBB IGF RECEPTOR
Fig 5. Affinity cross linking of ‘261-IGF-1,‘261-IGF-2,or “61-insulin to isolated human brain capillaries using DSS. The capillaries were incubated with labeled hormone in the presence of either no additive, 259 ng/mL IGF, or 409 ng/mL porcine insulin. The migration of molecular weight markers is shown on the right-hand side of the figure. The molecular weight of the insulin-binding site (133 kd) is shown on the right-hand side of the figure. The molecular weight of the IGF binding site (141 kd) is shown on the lehhand side of the figure. The SDS solubilized affinity cross linked capillary proteins were separated on a 7.5% SDS polyacrylamide gel before slab drying and autoradiography.
‘2”I-IGF-1
+
‘“%-IGF- 2 f25f- insutin IGFG!SO~) insulin (400
++
+ +
-
3) -
it is important to define the possible origins of the IGFs in brain. De novo synthesis and neurosecretion of the IGFs is one possible source of IGFs in brain. Another possible pathway is transport of circulating IGFs from blood into brain through receptor mediated transcytosis mechanisms at the BBB.” Indeed, previous studies with human brain capillaries have characterized the human BBB insulin and transferrin receptors.‘2’3 The present results extend these findings to the human BBB 1GF receptor. Studies of IGF receptors in peripheral tissues demonstrate a type 1 receptor with a 130 kd molecular weight for the binding subunit and a type 2 receptor with an approximately 260 kd molecular weight.*,” Although concentrations of approximately 1 to 10 pg/mL insulin are known to displace IGFs from the type 1 receptor, concentrations of insulin in excess of 100 wg/mL are believed not to displace IGFs from the type 2 receptor.5~7*23Previous studies with bovine brain are consistent with the presence of both the type 1 and type 2 receptor on bovine brain capillaries, since approximately 1 to 10 pg/mL insulin displaced IGF-1 from isolated bovine brain capillaries but 100 pg/mL of insulin exerted no displacement of IGF-2 from isolated bovine brain capillaries.*’ The pattern of peptide competition is different with isolated human brain capillaries, since 0.5 pg/mL concentrations of insulin displaced “‘1-IGF-2 from binding sites on isolated human brain capillaries (Fig 2). Another area of difference concerning the interaction of isolated IGFs with human brain capillaries, vis-a-vis bovine brain capillaries, is the fact that at 37%, binding, not internalization, is the rate limiting step in peptide uptake. This conclusion is drawn from the observation that the uptake of either IGF-1 or IGF-2 by isolated human brain capillaries is nonsaturable by 200 ng/mL IGF at 37°C (Fig 1). Previous studies with isolated human brain capillaries have shown that the non-
+
c
-
+
+m.-f--+-+
--
+-
saturable component of uptake is identical to the portion that is resistant to a mild acid wash,‘*,*’ which presumably represents either partial or complete internalization of the peptide within the endothelial cytoplasm. Conversely, in isolated bovine brain capillaries, internalization is always the slow step compared with receptor binding even at physiologic temperatures.” The affinity cross linking studies suggest only a type 1 receptor is present in human brain capillaries with a molecular weight of the binding subunit of 141 kd (Fig 5). The affinity cross linking studies are also consistent with the cross competition results shown in Fig 2, in that insulin is more effective in displacing ‘*‘I-IGF-2 from its binding site on human brain capillaries as compared with the binding of ‘251-IGF-1. Thus, the overall results suggest that only a type 1 receptor exists on isolated human brain capillaries and that this receptor has a higher affinity for IGF-2 than for IGF-1. This is compatible with the recent hypothesis of Casella et alz4 that there are two binding sites on the type 1 receptor in human placental membranes. Site A preferentially binds IGF-1 and site B preferentially binds IGF-2.24 Similarly, the human BBB IGF receptor appears to be a type 1 receptor that binds IGF-2 to a slightly greater extent than IGF- 1. The putative B site binds insulin weakly but fourfold greater than does the A (or IGF-1) binding site (Figs 2 and 5). However, the conclusions of the present study are provisional because a mixture of IGF-1 and IGF-2 was used in the displacement experiments (Figs 2, 3, and 5). The model of dual binding sites on the human BBB IGF receptor will be clarified by future studies using purified IGF-1 and IGF-2. The idea that the human BBB IGF receptor acts as a transport system is buttressed by the parallel observations made in the present study and in studies of IGF content in human brain. The concentration of IGF-2 in either human
140
DUFFY, PARDRIDGE, AND ROSENFELD
brain or in human CSF is much higher than the concentration of IGF-l.lS Moreover, the kinetics of either receptor binding to the surface of the isolated human brain capillary or the internalization step is at least three times as fast for IGF-2 as compared with IGF-1 (Fig 1). Endocytosis is only half of the overall transcytosis pathway and the demonstration of endocytosis is not proof that transcytosis of IGFs through the BBB occurs in vivo. However, the initial studies showing endocytosis of insulin” and transferrin” have been confirmed by in vivo experiments documenting transcytosis of insulinI and of transferrin” through the BBB. Finally, the limited study shown in Fig 4 regarding the effects of IGF binding proteins in human serum suggest that binding proteins greatly retard access of the peptide to the receptor. However, since more than 99% of serum IGF is
bound by the binding proteins:6 it is surprising that the presence of human serum in the incubation medium did not lower the uptake of the IGFs to the values seen for an extracellular marker such as ‘H-inulin (Fig 1). This suggests that there may be partial access of the bound IGF to the receptor, perhaps as a binding protein-IGF complex. In this regard, it is of interest that the concentration of the IGF-2 binding protein in human CSF is high compared with that for the IGF-1 binding protein.27
ACKNOWLEDGMENT
The authors are indebted to Jody Eisenberg and to Jing Yang for valuable technical assistance, and to Dawn Brown who skillfully prepared the manuscript.
REFERENCES
1. Adashi EY, Resnick CE, D’Ercole AJ, et al: Insulin-like growth factors as intraovarian regulators of granulosa cell growth and function. Endocr Rev 6:400-405, 1985 2. Bar RS, Boes M: Distinct receptors for IGF-I, IGF-II, and insulin are present on bovine capillary endothelial cells and large vessel endothelial cells. Biochem Biophys Res Commun 124:203207, 1984 3. Catanese VM, Grigorescu F, King GL et al: The human erythrocyte insulin-like growth factor I receptor: Characterization and demonstration of ligand-stimulated autophosphorylation. J Clin Endocrinol Metab 62:692-701, 1986 4. Hammerman MR. Gavin JR III: Binding of insulin-like growth factor II and multiplication-stimulating activity-stimulated phosphorylation in basolateral membranes from dog kidney. J Biol Chem 259:13511-13518,1984 5. Oppenheimer CL, Pessin JE, Massague J, et al: Insulin action rapidly modulates the apparent affinity of the insulin-like growth factor II receptor. J Biol Chem 258:4824-4832, 1983 6. Rechler MM, Nissley SP, Podskalny JM, et al: Identification of a receptor for somatomedin-like polypeptides in human fibroblasts. J Clin Endocrinol Metab 44:820-826, 1977 7. Rosenfeld RG, Ceda G, Wilson DM, et al: Characterization of high affinity receptors for insulin-like growth factors I and II on rat anterior pituitary cells. Endocrinology 114:1571-1576, 1984 8. Lesniak MA, Hill JM, Pert CB, et al: Autoradiographic localization of ‘2SI-IGFI and ‘“*I-rIGFI1 receptors in rat brain: Comparison with insulin receptors. Diabetes 35:58A, 1986 (suppl)
IGF II and of higher molecular USA 82:2153-2157, 1985
mass forms.
Proc Nat1 Acad
Sci
16. Rosenfeld RG, Dollar LA: Characterization of the somatomedin-C/insulin-like growth factor I (SM-C/IGF-I) receptor on cultured human fibroblast monolayers: Regulation of receptor concentrations by SM-C/IGF-I and insulin. J Clin Endocrinol Metab 55:434-441, 1982 17. Rosenfeld RG, Thorsson AV, Hintz RL: Increased somatomedin receptors in newborn circulating mononuclear cells. J Clin Endocrinol Metab 48:456-462, 1979 18. Hintz RL, Liu F: A radioimmunoassay for insulin-like growth factor-2 specific for the C-peptide region. J Clin Endocrinol Metab 54:442-449, 1982 19. Enberg G, Carlquist M, Joinvall H, et al: The chromatography of somatomedin A, isolated by microcomputer-controlled chromatography, reveals an apparent identity to insulin-like growth factor I. Eur J Biochem 143:117-120, 1984 20. Frank HJL, Pardridge WM, Morris WL, et al: Binding and internalization of insulin and insulin-like growth factors by isolated brain microvessels. Diabetes 35:654-661, 1986 21. Choi T, Pardridge WM: Phenylalanine transport at the human blood-brain barrier. Studies in isolated human brain capillaries. J Biol Chem 261:6536-6541, 1986 22. Sara VR, King, TL, Stuart MC, et al: Hormonal regulation of fetal brain cell proliferation: Presence in serum of a trophin responsive to pituitary growth hormone stimulation. Endocrinology 99:1512-1517, 1976
9. Sara VR, Hall K, Misaki M, et al: Ontogenesis of somatomedin and insulin receptors in the human fetus. J Clin Invest 71:10841089,1983 10. Wisniewski KE, Kamionowska ML, Wisniewski HM: Evidence of arrest of neurogenesis and synaptogenesis in brains of patients with Down’s syndrome. N Engl J Med 311:1187-l 191, 1984
23. DeVroede MA, Romanus JA, Standaert ML, et al: Interaction of insulin-like growth factors with a nonfusing mouse muscle cell line: Binding, action, and receptor down-regulation. Endocrinology 114:1917-1922, 1984
11. Pardridge WM: Receptor-mediated peptide transport through the blood-brain barrier. Endocr Rev 7:314-330, 1986
25. Fishman J, Rubin JB, Handarhan JV, et al: Receptormediated transcytosis of transferrin across the blood-brain barrier. J Neurosci Res (in press)
12. Pardridge WM. barrier insulin receptor.
Eisenberg J, Yang J: Human blood-brain J Neurochem 44:1771-1778, 1985
13. Pardridge WM, Eisenberg J, Yang J: Human blood-brain barrier transferrin receptor. Metabolism 36:892-895, 1987 14. Duffy KR, Pardridge WM: Blood brain barrier transcytosis of insulin in developing rabbits. Brain Res 420:32-38, 1987 15. Haselbacher GK, Schwab ME, Pasi A, et al: Insulin-like growth factor II (IGF II) in human brain: Regional distribution of
24. Casella SJ, Han VK, D’Ercole AJ, et al: Insulin-like growth factor II binding to the type I somatomedin receptor. Evidence for two high affinity binding sites. J Biol Chem 261:9268-9272, 1986
26. Smith GL: Somatomedin nol 34:83-90, 1984
carrier
proteins.
Mol Cell Endocri-
27. Binoux M, Hardouin S, Lassarre C, et al: production by the liver of two IGF binding proteins molecular weights but different affinities for IGF I Their relations with serum and cerebrospinal fluid proteins. J Clin Endocrinol Metab 55:600-660, 1982
Evidence for with similar and IGF II. IGF binding