purification and characterization of the insulin-like growth factor

Purification and Characterization of the Insulin-LikeGrowth Factor-Binding Protein-1 Phosphoform Found inNormal Plasma*

MELISSA WESTWOOD, J MARTIN GIBSON, AND ANNE WHITE

Endocrine Sciences Group, Department of Medicine, University of Manchester, Manchester, M13 9PT,United Kingdom

ABSTRACTOur previous work has shown that, in the normal circulation,

insulin-like growth factor-binding protein-1 (IGFBP-1) is present asa single highly phosphorylated species. In this study, we have purifiedthis previously uncharacterized isoform of IGFBP-1 to determine itsligand-binding affinity and the potential significance of highly phos-phorylated IGFBP-1. Immunoaffinity chromatography was used toisolate IGFBP-1 from normal human plasma and from human hep-atoma (Hep G2) cell medium as an alternative source of the IGFBP-1phosphoform in the circulation. The affinity of this highly phosphor-ylated IGFBP-1 was compared with that of nonphosphorylated IG-FBP-1 and recombinant human (rh) IGFBP-3 by equilibrium bindingto IGF-I and IGF-II.

Anion-exchange (IEX) HPLC, nondenaturing electrophoresis, al-kaline phosphatase treatment, and ligand-binding studies indicatedthat the highly phosphorylated IGFBP-1 from HepG2 cells was com-parable with IGFBP-1 from plasma. In binding to IGF-I, the plasma

phosphoform of IGFBP-1 was found to have a higher affinity (2.3 61.1 3 1010 M21) than nonphosphorylated IGFBP-1 (2.5 6 1.7 3 109

M21, P , 0.002). However, when binding to IGF-II, phosphorylationhad no affect on the affinity of IGFBP-1 (3.6 6 2 3 109 M21 vs. 1.8 633 109 M21,P not significant). Therefore, in the circulation, IGF-I hasa considerably higher affinity than IGF-II for IGFBP-1 (P, 0.02). Theaffinity of phosphorylated IGFBP-1 from plasma (2.3 6 1.1 3 1010

M21) also was significantly higher than the affinity of IGFBP-3 forIGF-I (5.6 6 4.2 3 109 M21, P , 0.005).

These data suggest that the highly phosphorylated IGFBP-1 in thenormal circulation will preferentially bind IGF-I rather than IGF-II,whereas in pregnancy, the affinity of IGFBP-1 for IGF-I will be re-duced because of the appearance of non- and lesser-phosphorylatedforms. This lends support to the theory that changes in IGFBP-1phosphorylation may influence the modulatory effects of IGFBP-1 onIGF bioavailability. (Endocrinology 138: 1130–1136, 1997)

THE INSULIN-LIKE growth factors (IGF-I and -II) arepresent in extracellular fluids bound to specific, high

affinity, binding proteins (IGFBPs) that modulate their ac-tivity at the cellular level (1, 2). One of these binding proteins,IGFBP-1, has been shown to both inhibit and potentiate theactions of IGF-I (3–5). These seemingly discrepant findingsmay be explained by the existence of several IGFBP-1 iso-forms (5). Phosphorylated IGFBP-1 variants, as well as theunmodified peptide, have been identified in amniotic fluid(AF) (6), a commonly used source of IGFBP-1, and in otherbiological fluids and tissues (7, 8). These variants have beenfound to augment the effects of IGF-I on porcine smoothmuscle cells (4, 6), whereas more highly phosphorylatedisoforms, isolated from human hepatoma (Hep G2) cell con-ditioned medium (CM), did not (6). In vivo studies of woundhealing, performed both in rats (9, 10) and humans (11),reflect in vitro findings; nonphosphorylated IGFBP-1 en-hances IGF-I-stimulated wound repair, whereas phosphor-ylated IGFBP-1 is inhibitory of IGF trophic actions.We have shown that in the normal adult human circula-

tion, IGFBP-1 is present as a single, highly phosphorylated

species; this phosphoform is not found in AF (12). However,circulating IGFBP-1 phosphorylation status can be altered,and this is particularly evident during pregnancy, when non-and lesser-phosphorylated variants increase markedly in thematernal circulation (12). Phosphorylation of IGFBP-1 hasbeen reported to increase affinity for ligand (6); thus, changesin the phosphorylation status of IGFBP-1 may influence theability of IGF to interact with its cell surface receptors.In adult human serum, IGFBP-3 is thought to be saturated

with IGF-I or -II, forming the 150 kDa complex with theacid-labile glycoprotein (13, 14). In contrast, the low-molec-ular mass-binding proteins, which form the 40–50 kDa com-plex, are thought to be unsaturated. Evidence for this comesfrom cross-linking of iodinated IGF-I to serum, which dem-onstrates binding to these species but not to IGFBP-3 (15). Inaddition, gel filtration chromatography of serum from hu-mans injected with radiolabeled IGF-I shows that the labeledIGF-I is initially present in the 40–50 kDa complex (16). Thishas led some investigators to question the functional impor-tance of IGFBP-1 in terms of regulating IGF bioavailabilityand actions.The aim of this study was to determine the significance of

changes to IGFBP-1 phosphorylation status by comparingthe ligand-binding affinities of the normal circulating phos-phoformwith those of nonphosphorylated IGFBP-1. Becausecirculating IGFBP-1 is mainly derived from the liver, weinvestigated the possibility of using a liver cell line (Hep G2cells) as a source of the phosphorylated isoform of IGFBP-1

Received May 23, 1996.Address all correspondence and requests for reprints to: Melissa

Westwood, Endocrine Sciences Research Group, Department of Medi-cine, University of Manchester, Stopford Building, Oxford Road,Manchester, M13 9PT, United Kingdom. E-mail: [email protected].

* This work was supported by the Medical Research Council and theSalford Royal Hospitals Trust.

0013-7227/97/$03.00/0 Vol. 138, No. 3Endocrinology Printed in U.S.A.Copyright © 1997 by The Endocrine Society

1130

found in plasma. In addition, the affinity of the previouslyuncharacterized plasma phosphoform of IGFBP-1 was com-pared with that of IGFBP-3, the main IGF carrier in thecirculation.

Materials and MethodsSources of IGFBP-1

IGFBP-1 was purified from two sources: plasma and the CM of ahuman hepatocyte carcinoma cell line (Hep G2). Plasma was obtainedfrom healthy female volunteers who were taking a combined oral con-traceptive pill; we have shown previously that circulating IGFBP-1 iselevated in such subjects, but the phosphorylation status of the IGFBP-1is unchanged, because only the single, highly phosphorylated isoformis present (17). The origin of circulating IGFBP-1 is the liver, and there-fore, the Hep G2 cells were investigated as a potential source of thecirculating phosphoform.

Culture of Hep G2 cells

Hep G2 cells (85011430, passage number 90) obtained from the Eu-ropeanCollection ofAnimal Cell Cultures (PortonDown, Salisbury,UK)were cultured in DMEM supplemented with 10% FCS, 1% nonessentialamino acids, 0.04 mm L-glutamine, and 1 mm pyruvate. Biochemicalanalysis and RIA (see below) of the growth medium indicated thatbovine IGFBP-1 does not cross-react with the antibodies used in thisstudy. The cells were grown in 5% carbon dioxide at 37 C until confluent,at which point the medium was harvested and the cells passaged 1:4after trypsinization.

Immunoaffinity chromatography

Immunoaffinity chromatography was used to isolate IGFBP-1 fromnormal plasma (NP) or Hep G2-CM. Monoclonal antibody (MAb) 6303(a generous gift of Medix Biochemica, Kauniainen, Finland), whichrecognizes all IGFBP-1 variants (12), was coupled to Sephacryl S-300 (11)at 1 mg/ml to form the immunoaffinity matrix. A 10-ml column wasequilibrated for 24 h at 4 C by the application of PBS/0.25% BSA/0.1%Tween 20 at a flow rate of 5 ml/h.

Next, 250 ml plasma or 500 ml Hep G2 CMwas recirculated throughthe column for 72 h at a flow rate of 3.75 ml/h, the column was washedwith 100ml Tris buffer, pH 8.0 (50mm Tris/0.5mNaCl/0.1% Tween 20),and the bound peptide was eluted by application of 0.1 m hydrochloricacid. Then, 10 3 1-ml fractions were collected into tubes containing 200ml 1 m Tris pH 9.0 and analyzed for IGFBP-1 by RIA (see below).Fractions containing more than 100 mg/l IGFBP-1 were pooled andconcentrated by centrifugation through Centricon 10 filters (Amicon,Stonehouse, Gloucestershire, UK). The particular IGFBP-1 isoforms con-tainedwithin the concentrate were assessed by n-octyl glucoside (n-OG)electrophoresis and Western ligand blotting with 125I-IGF-I (see below).

HPLC

One hundred microliters of concentrate obtained from the 6303 im-munoaffinity column or 1 mg rhIGFBP-1 (kindly donated by Amgen,Boulder, Co) dissolved in 0.02 m Tris/0.1 m NaCl/2% isopropanol, (pH9.5) were applied to a 4.6 3 150 mm Hema Mono Q column (Alltech,Camforth, UK). Sample was eluted isocratically for 5 min with 0.02 mTris/0.1 m NaCl/20% isopropanol followed by a linear gradient to 0.02m Tris/0.3 mNaCl/20% isopropanol over 20 min. The flow rate was 1.5ml/min, and absorbance was monitored at 220 nm. Then, 50 3 30sfractions were collected and analyzed for IGFBP-1 by RIA and n-OGelectrophoresis/Western ligand blot (see below).

IGFBP-1 RIA

IGFBP-1 levels in the fractions from the immunoaffinity and IEXcolumnswere determined using our previously reportedRIAs, RIA 6303and RIA 6305 (12). The assays use rhIGFBP-1 (a kind gift of Dr. L.Fryklund, Pharmacia, Stockholm, Sweden), for standards (1–250 mg/l)and radiolabel, and either MAb 6303 or 6305 (generously provided byMedix Biochemica, Kauniainen, Finland). RIA 6303 recognizes all iso-

forms of IGFBP-1, including the phosphoform characteristic of NP,whereasMAb 6305 does not recognize the circulating phosphoform, andtherefore, the 6305 RIA only detects the non- and lesser-phosphorylatedisoforms (12).

Biochemical characterization of IGFBP-1 isoforms

The IGFBP-1 isoforms in the NP, Hep G2 CM, and the fractions fromthe immunoaffinity and IEX columns were characterized by immuno-precipitation, n-OG electrophoresis, and Western ligand blotting, aspreviously described (6, 12). Samples were incubated at 4 C overnightwith 250 ml MAb 6303 or 6305 (1:1000 dilution). Precipitating antibody(250 ml antimouse coated cellulose suspension (Sac-Cel; IDS, Tyne andWear, Boldon, UK) was then added and incubated for 1 h at 37 C. Boundantibody was separated by centrifugation at 1000 3 g for 10 min. Theprecipitated proteins were washed (33) by the addition of 1 ml PBS/0.25% BSA/0.1% tween 20 and centrifuged at 10003 g for 10 min beforeresuspending in 100 ml gel loading buffer [170 mm Tris.HPO4 pH 5.5/90mm n-OG (Sigma, Poole, Dorset, UK)/40% glycerol/0.008% bromophe-nol blue]. All samples were boiled for 5 min before loading onto astacking gel of 4% acrylamide, which, like the resolving gel (15% acryl-amide), contained the nonionic detergent n-OG at 20 mm. The gels wererun at a constant voltage of 175V for approximately 6 h. After transferonto nitrocellulose membranes, the proteins were incubated with150,000 cpm/ml 125I-IGF-I for 4 h at 25 C, washed, and visualized byautoradiography (5 days exposure).

Treatment of purified plasma and Hep G2 phosphoformwith alkaline phosphatase

The highly phosphorylated IGFBP-1 isoform, obtained from IEXHPLC of plasma and Hep G2 CM, was incubated with 1 U calf intestinalalkaline phosphatase (Boehringer Mannheim, Indianapolis, IN) for 2 hat 37 C. Samples were then subjected to immunoprecipitation andWest-ern ligand blotting, as described above.

Assay of IGFBP binding to IGFs

Ligand-binding assays were performed with 125I IGF-I or -II in thepresence of unlabeled IGF-I or IGF-II to determine the relative affinitiesof rhIGFBP-1, the highly phosphorylated IGFBP-1 isoform, purifiedfromplasma andHepG2CMand rhIGFBP-3 (recombinant IGFBP-3wasused in these studies because Sommer et al. (18) and Mukku et al. (19)have shown that posttranslational modifications do not affect its affinityfor ligand.). The concentration yielding 25% specific binding was firstdetermined for each IGFBP preparation fromdilution curves establishedwith 125I IGF-I or -II.

125I-labeled rhIGF-I (18, 500 cpm; specific activity 185 mCi/mg) or125I-rhIGF-II (25,000 cpm; specific activity 255 mCi/mg) was incubatedwith 10 mg/liter IGFBP-1or 1 mg/liter IGFBP-3 in the presence of un-labeled IGF-I or -II (final concentration 0–40 ng/ml). The reactions wereperformed in triplicate overnight at 4 C in 0.25 ml 0.1 m HEPES/44 mmNaH2PO4/0.01%TritonX-100/0.25%BSA/0.02% sodiumazide, pH6.0).Bound label was separated from free by adding 250 ml 1% humang-globulin and 500 ml 25% polyethylene glycol (Mr 8000; Sigma) andcentrifugation at 1000 3 g for 15 min. The pellet was washed with 1 ml6.25% polyethylene glycol and the final pellet counted in a g spectrom-eter. Nonspecific binding was determined by measuring the amount of125I-IGF-I/-II that could be precipitated in the presence of 1.0 mg/lunlabeled IGF.

To confirm the effect of phosphorylation on the affinity of IGFBP-1for IGF, the dephosphorylated plasma phosphoform was subjected toligand-binding assays, as described above.

Statistical analysis

The Simfit program (20), kindly provided by Dr. W. G. Bardsley,University of Manchester, Manchester, UK, was used to analyze theligand-binding data, and the Mann-Whitney U test for nonparametricdata was used to compare the mean affinities of the various IGFBPpreparations for their ligands.

CHARACTERIZATION OF PLASMA IGFBP-1 1131

ResultsCharacterization of IGFBP-1 isoforms in Hep G2-CM

Because the highly phosphorylated circulating form ofIGFBP-1 is of hepatic origin, human hepatoma (HepG2) cellswere investigated as a potential source of the NP phospho-form. CM from HepG2 cells (HepG2 CM) was immunopre-cipitated with eitherMAb 6303 or 6305 and then analyzed byn-OG electrophoresis and Western ligand blotting (Fig. 1).Only MAb 6303 recognizes all IGFBP-1 variants. MAb 6305does not detect the isoform found in the circulation (lane 4),and this highly phosphorylated form of IGFBP-1 was notpresent in AF (lane 2). Hep G2 cells do produce the IGFBP-1phosphoform found in the normal circulation; however,multiple lesser-phosphorylated isoforms and nonphospho-rylated IGFBP-1 also were present (lane 5).

Purification of IGFBP-1 isoforms

Immunoaffinity chromatography was used to isolateIGFBP-1 from Hep G2 CM; separation of the different phos-phoforms then was achieved by IEX HPLC. The resultingchromatogram is shown in Fig. 2A. The upper line representsthe elution profile and the lower line, a fitted baseline used forcalculation of peaks. A species (peak a) with identical reten-tion time to rh (nonphosphorylated) IGFBP-1 (Fig. 2B) wasapparent, and in addition, the IGFBP-1 prepared from HepG2 CM separated into several other discrete peaks (peaksb–e) with increased retention times, suggesting the separa-tion of variantly phosphorylated isoforms.The fractions collected from IEX chromatography of both

rhIGFBP-1 and IGFBP-1 purified from HepG2 CM were an-alyzed by RIA 6303 and RIA 6305, and the results are shownin Fig. 3. All fractions corresponding to chromatographicpeaks contained IGFBP-1. With rhIGFBP-1, both assays mea-sured approximately the same level of IGFBP-1 in each of thefractions (Fig. 3A); however, this was not true for the frac-tions obtained from IEX-chromatographic analysis ofIGFBP-1 purified from HepG2 CM. Unlike RIA 6303, RIA6305 failed to detect any IGFBP-1 in the fraction correspond-

ing to the peak e, suggesting that this fraction contained thecirculating IGFBP-1 phosphoform; because this is the mosthighly phosphorylated variant of IGFBP-1, this specieswould be expected to have the greatest retention time. Tosupport this data, an aliquot of the IGFBP-1 purified fromplasma by immunoaffinity chromatography also was ana-lyzed by IEX HPLC (Fig. 3C). A single peak of immunore-activity, which coincided with peak e in the Hep G2 chro-matogram, was observed when the fractions were analyzedwith RIA 6303; however, as anticipated, RIA 6305 did notdetect any IGFBP-1 in any of the fractions.

Characterization of IGFBP-1 isoforms in fractions fromIEX-HPLC of HepG2 CM

To provide further evidence that IEXHPLC had separatedthe isoform corresponding to the plasma phosphoform fromthe other variants in the Hep G2-derived IGFBP-1, the frac-tions corresponding to the peaks were immunoprecipitatedby MAb 6303 and analyzed by n-OG electrophoresis andWestern ligand blotting (Fig. 4). The phosphoform charac-teristic of NP was found in the fractions, coinciding with theimmunoreactivity peak (peak e) detected byRIA 6303 but notRIA 6305.

FIG. 1. Characterization of IGFBP-1 isoforms present in AF, NP, andHepG2-CM. Samples were immunoprecipitated with MAb 6303 orMAb 6305 and subjected to n-OG electrophoresis and Western ligandblotting with 125I-IGF-I.

FIG. 2. IEX HPLC of (A) IGFBP-1 purified from HepG2 CM and (B)rhIGFBP-1. The upper line represents the elution profile and thelower line represents a fitted baseline for calculation of peaks.

1132 CHARACTERIZATION OF PLASMA IGFBP-1 Endo • 1997Vol 138 • No 3

Figure 5 shows that the phosphoform purified from HepG2 comigrates with the phosphoform purified from plasmaand that both the Hep G2-derived and the plasma-derivedIGFBP-1 are recognized by MAb 6303 but not by MAb 6305.To confirm that the Hep G2- and plasma-derived IGFBP-1were phosphorylated, both isoforms were incubated withalkaline phosphatase for 2 h at 37 C. With the Hep G2-derived IGFBP-1, this resulted in the appearance of an iso-form that comigratedwith nonphosphorylated rhIGBP-1 andcould be detected by MAb 6305 and MAb 6303. Although a2-h incubation was insufficient for the complete dephos-phorylation of the plasma-derived IGFBP-1, alkaline phos-phatase did produce an isoform that comigrated withrhIGFBP-1 and was detected by MAb 6305.

Ligand-binding affinities of IGFBP-1 isoforms

To determine the effect of phosphorylation on IGFBP-1binding of IGF, equilibriumbinding studieswere performed.Nonspecific binding constituted less than 25% of bound ra-diolabel in the absence of unlabeled IGF and less than 18%of the total amount of radiolabel added to each tube. Scat-chard plots obtained for all isoforms of IGFBP-1 were linear,indicating the presence of a single binding site.

IGF-I. Figure 6 shows the Scatchard analysis of IGF-I bindingto rhIGFBP-1 and the phosphoform isolated fromNP (beforeand after treatment with alkaline phosphatase). The mean(n5 5) affinity constant of rhIGFBP-1was 2.56 1.73 109m21

(Table 1), which is in accordance with previous reports (6).However, the affinity of the highly phosphorylated IGFBP-1isolated from plasma was significantly greater at 2.3 6 1.1 31010 m21 (n 5 5; P , 0.002), as was that of the highly phos-phorylated isoform purified from HepG2 CM (P , 0.007;Table 1). There was no statistical difference between theaffinity of the plasma- and Hep G2-derived phosphoformsfor IGF-I, though the range in affinities determined for thelatter isoform (1.3 3 1010 m21 to 5.8 3 1011 m21) resulted ina mean of 1.6 6 2.8 3 1011 m21. Dephosphorylation of thephosphorylated IGFBP-1 from plasma with alkaline phos-phatase, though incomplete, resulted in a significant de-crease in the affinity for IGF-I (2.3 6 1.1 3 1010 m21 to 5.4 61.1 3 109 m21; P , 0.004).

FIG. 4. Characterization of IGFBP-1 isoforms contained in fractionsobtained from IEX-HPLC of IGFBP-1 purified fromHepG2 CM. Eachfraction was immunoprecipitated by MAb 6303 and the precipitatesanalyzed by n-OG electrophoresis and Western ligand blotting with125I-IGF-I. Unpurified Hep G2 medium and human NP are shown forcomparison.

FIG. 5. Dephosphorylation of the highly phosphorylated IGFBP-1isoform isolated from Hep G2 CM and human plasma. Samples wereimmunoprecipitated with MAb 6303 or MAb 6305 and subjected ton-OG electrophoresis and Western ligand blotting with 125I-IGF-I.Nonphosphorylated rhIGFBP-1 (lane 1), AF (lane 2), and NP (MAb6303; lane 3) were included as controls. The purified Hep G2 phos-phoform was immunoprecipitated with MAb 6303 and MAb 6305 inthe absence (lanes 4 & 5, respectively) and presence (lanes 6 & 7,respectively) of alkaline phosphatase. Lanes 8 and 9 show the im-munoprecipitation of the phosphoform purified from plasma withMAbs 6303 and 6305, respectively. Lanes 10 and 11 are the result ofimmunoprecipitation after treatment with alkaline phosphatase.

FIG. 3. IGFBP-1 in fractions collected from IEX-HPLC of (A) rhIG-FBP-1, (B) IGFBP-1 purified from HepG2 CM, and (C) IGFBP-1 pu-rified from plasma. Comparison of levels measured by RIAs 6303 (F)and 6305 (E).


IGF-II. The Scatchard analysis of IGF-II binding to rhIGFBP-1and the highly-phosphorylated IGFBP-1 purified fromplasma is shown in Fig. 7. The mean (n 5 5) affinity constantof rhIGFBP-1 for IGF-II was 1.8 6 3.0 3 109 m21, which wasnot significantly different from the affinity constant for IGF-I(Table 1). The affinity constant of the highly phosphorylatedisoform derived from plasma and Hep G2 medium (3.6 62.0 3 109 m21 and 5.7 6 6.3 3 109 m21, respectively) was nodifferent from that of rhIGFBP-1. Hence, phosphorylationdoes not seem to affect the affinity of circulating IGFBP-1 forIGF-II. Therefore, the circulating phosphoform has a signif-icantly higher affinity (P , 0.02) for IGF-I than for IGF-II(Table 1).

Ligand-binding affinity of rhIGFBP-3

The affinity of rh (nonglycosylated) IGFBP-3 for IGF-I and-II was in accordance with previous reports (11, 12). IGFBP-3had a slightly higher affinity for IGF-II than IGF-I (1.86 2.331010m21 vs. 5.66 4.23 109m21; Table 1), though this was notsignificant. However, there was a significant difference be-tween the affinity of IGFBP-3 and the circulating isoform ofIGFBP-1 for IGF-I; IGFBP-1 had a 4-fold higher affinity (P ,0.005) than IGFBP-3, the main IGF carrier in the circulation,though for IGF-II, the affinity of IGFBP-3 was no differentfrom that of the highly phosphorylated isoform of IGFBP-1.

Discussion

We have reported that, in the normal circulation, IGFBP-1is present as a single highly phosphorylated species. In thisstudy, we have purified this previously uncharacterized iso-form of IGFBP-1 to determine its ligand-binding affinity andthe functional significance of IGFBP-1 phosphorylation. Cir-culating IGFBP-1 is produced by the liver, and therefore,human liver carcinoma (HepG2) cells also were investigatedas a source of the plasma phosphoform of IGFBP-1. Theplasma form of IGFBP-1 is produced by Hep G2 cells; how-ever, unlike the liver, these cells also produce non- and lesser-phosphorylated variants. This may be a result of alteredintracellular kinase pathways or extracellular dephosphor-ylation of the highly phosphorylated isoform. Nevertheless,by using a combination of immunoaffinity chromatographyand IEX HPLC, we were able to purify the individualIGFBP-1 isoforms to homogeneity for use in ligand-bindingassays. The most highly phosphorylated IGFBP-1 isoform,produced by Hep G2 cells coeluted from IEX HPLC withplasma IGFBP-1, had identical mobility in the n-OG electro-phoresis system and had a similar affinity for both IGF-I andIGF-II as plasma-derived IGFBP-1. Therefore, we concludethat Hep G2 cells provide an alternative and more accessiblesource of the circulating isoform of IGFBP-1.The affinity of the plasma phosphoform for IGF-I was

much greater than that reported for the phosphoformspresent in AF (6) andwas approximately 10-fold greater thannonphosphorylated IGFBP-1 seen inmaternal plasmaduringpregnancy (12); thus, phosphorylation does increase the af-finity for IGF-I, as anticipated from other studies (6). De-phosphorylation of IGFBP-1 during pregnancy may result inIGF peptides, particularly IGF-I, being liberated from ormore weakly bound to this altered circulating IGFBP-1, lead-ing to increased IGF bioavailability for placental and fetalgrowth.The affinity of the highly phosphorylated form of IGFBP-1

for IGF-I was also significantly higher than the affinity ofIGFBP-3, the main carrier of IGF in the circulation. It isthought that the 150-kDa complex of IGFBP-3, IGF-I, or -II

FIG. 7. Scatchard analysis of IGF-II binding to two forms of IGFBP-1:nonphosphorylated rhIGFBP-1(å) and IGFBP-1 purified fromplasma (F). The slopes of the first order regression lines obtained fromfive experiments were used to determine mean Ka values.

FIG. 6. Scatchard analysis of IGF-I binding to three forms ofIGFBP-1: nonphosphorylated rhIGFBP-1 (å), IGFBP-1 purified fromplasma (F), and dephosphorylated plasma IGFBP-1 (■). The slopes ofthe first order regression lines obtained from five experiments wereused to determine mean Ka values.

TABLE 1. Comparison of the mean (6 SD) binding affinities ofphosphorylated plasma-derived IGFBP-1 (before and afterdephosphorylation with alkaline phosphatase), phosphorylatedHep G2-derived IGFBP-1, nonphosphorylated recombinant human(rh)IGFBP-1, and rhIGFBP-3 for IGF-I and -II

IGF-I IGF-II

Plasma IGFBP-1 2.3 6 1.1 3 1010 M21 3.6 6 2.0 3 109 M21

Dephosphorylatedplasma IGFBP-1

5.4 6 1.1 3 109 M21

Hep G2 IGFBP-1 1.6 6 2.8 3 1011 M21 5.7 6 6.3 3 109 M21

rhIGFBP-1 2.5 6 1.7 3 109 M21 1.8 6 3.0 3 109 M21

rhIGFBP-3 5.6 6 4.2 3 109 M21 1.8 6 2.3 3 1010 M21


and the acid-labile glycoprotein (13, 14) is saturated withIGFs and serves as their endocrine storage site. In this form,the half-life of the IGF peptides is increased (16), and theiraccess is limited to extravascular spaces (21) because the150-kDa complex does not readily cross the capillary barrier.The IGF bound to the IGFBP-1 and -2 in the 40- to 50-kDacomplex has a serum half-life of 20–30 min (16) because thecomplex can leave the vascular compartment (22); thereby itsIGF may reach and interact with tissue receptors. We andothers (23–25) have observed marked variations in IGFBP-1levels throughout the day. Some investigators (26–28) havesuggested that such fluctuations in plasma IGFBP-1 levels,particularly in response to insulin, imply a role for IGFBP-1in glucose homeostasis. However, the significance of this hasbeen questioned (29) because the concentration of IGFBP-1 innormal adult serum is reported to be approximately 10-foldlower than IGFBP-2 and around 100-fold less than that ofIGFBP-3 (29); thus, fluctuations of IGFBP-1 occur on top ofa high IGFBP background. The current findings indicate thattheNP phosphoform of IGFBP-1 has a slightly higher affinityfor IGF-I than IGFBP-3 and, indeed, than any of the otherIGFBPs (30); this suggests that circulating IGFBP-1 could befully saturated with IGFs and supports a role for IGFBP-1 inmodifying IGF bioavailability and contributing to glucosecounterregulation.In general, the IGFBPs are thought to inhibit IGF actions

(31) by preventing or attenuating IGF interaction with theircell surface receptors, which have a lower affinity for IGF(6.7 3 108 m21; 32). Enhancement of IGF action has beenobserved with IGFBP-1, -3, and -5. It is thought that this isfacilitated by posttranslational modifications of the IGFBPsleading to altered IGF affinities. Association of IGFBP-3 withthe cell surface (33) and binding of IGFBP-5 to the extracel-lular matrix (34) significantly lowers their affinities for IGFs,as compared with the soluble forms of these IGFBPs, there-fore favoring IGF/receptor interactions. Similarly, proteo-lytic cleavage of IGFBP-3 (35, 36) and IGFBP-5 (37) lowersIGF affinities. IGFBP-2 (38) and -4 (39) also are susceptible,though the exact roles of these proteases in controlling thedistribution of the IGFs in serum has not been determined.To date, no such protease has been described for IGFBP-1; weand others (12, 31) have suggested that changes in phos-phorylation status may represent the mechanism by whichIGFs are released by IGFBP-1. This hypothesis is supportedby our finding that the normal circulating form of IGFBP-1has a 10-fold higher affinity for IGF-I than the nonphospho-rylated IGFBP-1 that appears during pregnancy. Whetherchange in IGFBP-1 phosphorylation status (for example, atthe cell surface of IGF target tissues) represents a more gen-eralizedmechanism for controlling tissue IGF bioavailabilityremains to be determined, the analogy being that tissueIGFBP-3 proteolytic activity is reported to be 8-fold higherthan in serum (40).All IGFBPs bind both IGF-I and -II with high specific

affinity, and Rechler (30) has suggested that the affinity con-stants of the six IGFBPs are similar for IGF-I and IGF-II, withthe exception of IGFBP-6, which has a 20- to 70-fold higheraffinity for IGF-II, and that variations of some affinity con-stants found in the literature may be caused by differenttemperatures at which Ka values were determined. In the

present study, the binding affinity of the plasma form ofIGFBP-1 for IGF-I and IGF-II was determined under the sameconditions, and phosphorylationwas found to affect only theaffinity of IGFBP-1 for IGF-I. Although the explanation forthis phenomenon is not clear, it may involve different bind-ing sites for IGF-I and -II on IGFBP-1, as has been suggestedrecently for IGFBP-2 (41). These findings may also explainwhyKratz et al. (42)were able to showenhanced proliferativeresponse to IGF-I, but not IGF-II, in the presence of non-phosphorylated (recombinant) IGFBP-1 in human keratino-cytes and fibroblasts. Scatchard plots obtained by Roghani etal. (43) also imply two classes of binding site for IGF-I andIGF-II, one of high and one of low affinity, though this islikely to be because their IGFBP-1 preparation was purifiedfrom AF and would have contained several different iso-forms of IGFBP-1. Our Scatchard data gives linear plots,which would suggest that, when using a homogenous prep-aration of IGFBP-1, the binding of IGF-I or -II is at a singlesite.In summary, we have purified the highly phosphorylated

form of IGFBP-1 found in the normal circulation and haveshown that this species has a significantly higher affinity forIGF-I in comparison with the nonphosphorylated isoformthat appears under some physiological and pathological con-ditions. These data suggest that normally, IGFBP-1 in thecirculation would be inhibitory of IGF actions; however,changes in IGFBP-1 phosphorylation status may permit in-creased IGF, particularly IGF-I, bioavailability.

References

1. Sara VR, Hall K 1990 Insulin-like growth factors and their binding proteins.Physiol Rev 70:591–614

2. Rosenfeld EG, LamsonG, PhamH,OHY,Conover C,De LeonDD,DonovanSM, Ocrant I, Giudice L 1990 Insulin-like growth factor binding proteins.Recent Prog Horm Res 46:99–163

3. Liu L, BrinkmanA, Blot C,Harel L 1991 IGFBP-1, an insulin-like growth factorbinding protein is a cell growth inhibitor. Biochem Biophys Res Commun174:673–679

4. Elgin RG, Busby WH, Clemmons DR 1987 An IGFBP enhances the biologicresponse to IGF-I. Proc Natl Acad Sci USA 84:3254–3258

5. Busby WH, Klapper DG, Clemmons DR 1988 Purification of a 31 000-daltoninsulin-like growth factor binding protein from amniotic fluid. J Biol Chem263:14203–14210

6. Jones JI, D’Ercole AJ, Camacho-Hubner C, Clemmons DR 1991 Phosphor-ylation of insulin-like growth factor (IGF)-binding protein-1 in cell culture andin vivo: effects on affinity for IGF-I. Proc Natl Acad Sci USA 88:7481–7485

7. Frost RA, Tseng L 1991 Insulin-like growth factor binding protein-1 is phos-phorylated by cultured human endometrial cells and multiple protein kinasesin vitro. J Biol Chem 266:18082–18088

8. Koistinen R, Angervo M, Leinonen P, Hakala T, Seppala M 1993 Phosphor-ylation of insulin-like growth factor binding protein-1 increases in humanamniotic fluid and decidua from early to late pregnancy. Clin Chim Acta215:189–199

9. Jyung RW, Mustoe TA, Busby WH, Clemmons DR 1994 Increased wound-breaking strength induced by insulin-like growth factor I in combination withinsulin-like growth factor binding protein-1. Surgery 115:233–239

10. Tsuboi R, Shi C-M, Sato C, Cox GN, Ogawa H 1995 Co-administration ofinsulin-like growth factor (IGF)-I and IGF binding protein-1 stimulates woundhealing in animal models. J Invest Dermatol 104:199–203

11. KratzG, LakeM,GidlundM 1994 Insulin-like growth factor-I and -II and theirrole in the re-epithelialisation of wounds; interactions with insulin-like growthfactor binding protein type 1. Scand J Plast Reconstr Hand Surg 28:107–112

12. Westwood M, Gibson JM, Davies AJ, Young RJ, White A 1994 The phos-phorylation pattern of insulin-like growth factor binding protein-1 in normalplasma is different from that in amniotic fluid and changes during pregnancy.J Clin Endocrinol Metab 79:1735–1741

13. Martin JL, Baxter RC 1986 Insulin-like growth factor-binding protein fromhuman plasma. Purification and characterization. J Biol Chem 261:8754–8760

14. Baxter RC, Martin JL 1989 Structure of the Mr 140,000 growth hormone


dependent insulin-like growth factor binding protein complex: determinationby reconstitution and affinity labeling. Proc Natl Acad Sci USA 86:6898–6902

15. McCusker RH, Campion DR, Clemmons DR 1988 The ontogeny and regu-lation of a 31,000 Mr insulin-like growth factor/somatomedin (IGF) bindingprotein in fetal porcine plasma and sera. Endocrinology 122:2071–2079

16. Guler H-P, Zapf J, Schmid C, Froesch ER 1989 Insulin-like growth factors Iand II in healthy man: estimations of half-lives and production rates. ActaEndocrinol 121:753–758

17. Westwood M, Gibson JM, Williams AC, Clayton PE, Hamberg O, FlyvbjergA,White A 1995 Hormonal regulation of circulating insulin-like growth factorbinding protein-1 phosphorylation status. J Clin Endocrinol Metab80:3520–3527

18. Sommer A, Spratt SK, Tatsuno GP, Tressel T, Lee R, Maack CA 1993 Prop-erties of glycosylated and non-glycosylated human recombinant IGF bindingprotein-3 (IGFBP-3). Growth Regul 1:46–49

19. Mukku VR, Chu H, Baxter R 1991 Insulin-like growth factor binding protein,IGFBP-3 is secreted as a phosphoprotein. 2nd International Symposium onIGFs/Somatomedins, San Francisco CA, p 237

20. Bardsley WG, Bukhari NAJ, Ferguson MWJ, Cachaza JA, Burguillo FJ 1995Evaluation of model discrimination, parameter estimation and goodness of fitin nonlinear regression problems by test statistics distributions. Comput Chem19:75–84

21. Binoux M, Hossenlopp P 1988 Insulin-like growth factor (IGF) and IGF bind-ing proteins: comparison of human serum and lymph. J Clin EndocrinolMetab67:509–514

22. Bar RS, Clemmons DR, Boes M, Busby WH, Booth BA, Dake BL, Sandra A1990 Transcapillary permeability and subendothelial distribution of endothe-lial and amniotic fluid insulin-like growth factor binding proteins in the ratheart. Endocrinology 127:1078–1086

23. Gibson JM, Westwood M, Crosby SR, Gordon C, Holly JMP, Fraser W,Anderson C, White A, Young RJ 1995 Choice of treatment affects plasmalevels of insulin-like growth factor binding protein-1 (IGFBP-1) in noninsulin-dependent diabetes mellitus (NIDDM). J Clin Endocrinol Metab80:1369–1375

24. Yeoh SI, Baxter RC 1988 Metabolic regulation of the growth hormone inde-pendent IGFBP in human plasma. Acta Endocrinol 119:465–473

25. Holly JMP, Biddlecombe RA, Dunger DB, Edge JA, Amiel SA, Howell R,ChardT,Rees LH,Wass JAH 1988Circadian variation ofGH-independent IGFbinding protein in diabetes mellitus and its relationship to insulin. A new rolefor insulin? Clin Endocrinol 29:667–675

26. Lee PDK, Conover CA, Powell DR 1993 Regulation and function of insulin-like growth factor binding protein-1. Proc Soc Exp Biol Med 204:4–29

27. Holly JMP 1991 The physiological role of IGFBP-1. Acta Endocrinol [Suppl 2]124:55–62

28. Lewitt MS, Baxter RC 1991 Insulin-like growth factor binding protein-1: a rolein glucose counter regulation. Mol Cell Endocrinol 79:C147–C152

29. Zapf J 1995 Physiological role of the insulin-like growth factor binding pro-teins. Eur J Endocrinol 132:645–654

30. Rechler MM 1993 Insulin-like growth factor binding proteins. Vitam Horm47:1–114

31. Jones JI, Clemmons DR 1995 Insulin-like growth factors and their bindingproteins: biological actions. Endocr Rev 16:3–34

32. Steele-Perkins G, Turner J, Edman JC, Hari J, Pierce SB, Stover C, RutterWJ,Roth RA 1988 Expression and characterization of a functional human insulin-like growth factor I receptor. J Biol Chem 263:11486–11492

33. McCusker RH, Camacho-Hubner C, BayneML, CascieriMA, ClemmonsDR1990 Insulin-like growth factor (IGF) binding to human fibroblasts and glio-blastoma cells: the modulating effect of cell released IGF binding proteins(IGFBPs). J Cell Physiol 144:244–253

34. Jones JI, Gockerman A, Busby WHJ, Camacho-Hubner C, Clemmons DR1993 Extracellularmatrix contains insulin-like growth factor binding protein-5:potentiation of the effects of IGF-I. J Cell Biol 121:679–687

35. Guidice LC, Farrell EM, PhamH, Lamson G, Rosenfeld RG 1990 Insulin-likegrowth factor binding proteins in maternal serum throughout gestation andin the puerperium: effects of a pregnancy-associated serum protease activity.J Clin Endocrinol Metab 71:806–816

36. Davies SC, Wass AH, Ross RJM, Cotterill AM, Buchanan CR, Coulson VJ,Holly JMP 1991 The induction of a specific protease for insulin-like growthfactor binding protein-3 in the circulation during severe illness. J Endocrinol130:469–473

37. Andress DL, Loop SM, Zapf J, KeiferMC 1993 Carboxytruncated insulin-likegrowth factor binding protein-5 stimulates mitogenesis in osteoblast cells.Biochem Biophys Res Commun 195:25–30

38. Pucilowska JB, Davenport ML, Kabir I, Clemmons DR, Thissen JP, ButlerT, Underwood LE 1993 The effect of dietary protein supplementation oninsulin-like growth factors (IGFs) and IGF binding proteins in children withshigellosis. J Clin Endocrinol Metab 77:1516–1521

39. Conover CA, Kiefer MC, Zapf J 1993 Posttranslational regulation of insulin-like growth factor binding protein-4 in normal and transformed human fi-broblasts. Insulin-like growth factor dependence and biological studies. J ClinInvest 91:1129–1137

40. Lalou C, Binoux M 1993 Evidence that limited proteolysis of insulin-likegrowth factor binding protein-3 (IGFBP-3) occurs in the normal state outsideof the bloodstream. Regul Pept 48:179–188

41. Han VKM, Coulter C, Hayatsu J, Nygard K, Tanswell B 1995 Structure-function relationship of insulin-like growth factor binding proteins (IGFBPs).3rd International Symposium on Insulin-like Growth Factor Binding Proteins,Tuebingen Germany (Abstract)

42. Kratz G, Lake M, Ljungstrom K, Forsberg G, Haegerstrand A, Gidlund M1992 Effect of recombinant IGF binding protein-1 on primary cultures ofhuman keratinocytes and fibroblasts: selective enhancement of IGF-I but notIGF-II induced cell proliferation. Exp Cell Res 202:381–385

43. Roghani M, Lassarre C, Zapf J, Povoa G, Binoux M 1991 Two insulin-likegrowth factor (IGF)-binding proteins are responsible for the selective affinityfor IGF-II of cerebrospinal fluid binding proteins. J Clin Endocrinol Metab73:658–666


purification and characterization of the insulin-like growth factor

Documents