Protease-Resistant Insulin-Like Growth Factor (IGF)-Binding Protein-4 ...

47 downloads 0 Views 733KB Size Report
Jul 19, 2007 - IGF-I has been shown to play a role in the progression of atherosclerosis in experimental animal models. IGF-binding protein-4 (IGFBP-4) binds ...
0013-7227/07/$15.00/0 Printed in U.S.A.

Endocrinology 148(10):5002–5010 Copyright © 2007 by The Endocrine Society doi: 10.1210/en.2007-0571

Protease-Resistant Insulin-Like Growth Factor (IGF)-Binding Protein-4 Inhibits IGF-I Actions and Neointimal Expansion in a Porcine Model of Neointimal Hyperplasia T. C. Nichols, W. H. Busby, Jr., E. Merricks, J. Sipos, M. Rowland, K. Sitko, and D. R. Clemmons Department of Medicine, University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, North Carolina 27599 IGF-I has been shown to play a role in the progression of atherosclerosis in experimental animal models. IGF-binding protein-4 (IGFBP-4) binds to IGF-I and prevents its association with receptors. Overexpression of a protease-resistant form of IGFBP-4 has been shown to inhibit the ability of IGF-I to stimulate normal smooth muscle cell growth in mice. Based on these observations, we prepared a protease-resistant form of IGFBP-4 and infused it into hypercholesterolemic pigs. Infusion of the protease-resistant mutant inhibited lesion development by 53.3 ⴞ 6.1% (n ⴝ 6; P < 0.01). Control vessels that received an equimolar concentration of IGF-I and the protease-resistant IGFBP-4 showed no reduction in lesion size compared with control lesions that were infused with vehicle. Infusion of a nonmutated form of IGFBP-4 did not significantly inhibit lesion development. Proliferating cell nuclear antigen analysis showed that the mutant IGFBP-4 appeared to inhibit cell proliferation. The area occupied by extracellular

I

NCREASED EXPRESSION OF IGF-I has been shown to be a component of the vascular response to injury (1, 2). Arterial wall cell types such as smooth muscle cells (SMC) and endothelial cells possess IGF-I receptors (3). When these cells are grown in culture, they respond to IGF-I with increases in cell proliferation and migration (4). Studies in whole animals have indicated that IGF-I synthesis by vascular cell types is significantly increased after injury. After rat carotid artery injury, there is a major increase in IGF-I synthesis which occurs at the time of the maximal increase in SMC division (1, 5, 6). Other types of injuries, such as exposure to hyperglycemia or high cholesterol, stimulate activated macrophages to synthesize IGF-I (7, 8) Zhu et al. (9) demonstrated that targeted overexpression of IGF-I in vascular SMC in mice resulted in a hyperproliferative response to injury and increased neointimal formation. That this response is due to locally produced IGF-I has been suggested by studies in which hypophysectomized animals that have been injured show increased IGF-I gene expression in the vessel wall area of injury but no change circulating IGF-I (6).

First Published Online July 19, 2007 Abbreviations: IGFBP, IGF-binding protein; PCNA, proliferating cell nuclear antigen; SMC, smooth muscle cell. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

matrix was also reduced proportionally compared with total lesion area. Immunoblotting revealed that the mutant IGFBP-4 remained intact, whereas the wild-type IGFBP-4 that was infused was proteolytically cleaved. Further analysis of the lesions revealed that a marker protein, IGFBP-5, whose synthesis is stimulated by IGF-I, was decreased in the lesions that received the protease-resistant, IGFBP-4 mutant, whereas there was no change in lesions that received wildtype IGFBP-4 or the mutant protein plus IGF-I. These findings clearly illustrate that infusion of protease-resistant IGFBP-4 into the perilesion environment results in inhibition of cell proliferation and attenuation of the development of neointima. The findings support the hypothesis that inhibiting IGFBP-4 proteolysis in the lesion microenvironment could be an effective means for regulating neointimal expansion. (Endocrinology 148: 5002–5010, 2007)

IGF-binding protein (IGFBP)-4 is one member of a family of IGFBP that binds IGF-I with high affinity (10). Unlike several other members of the IGFBP family, IGFBP-4 does not attach to extracellular matrix or to cell surfaces, and it has been shown to inhibit the binding of IGF-I to cell surface receptors (11). To further determine the role of locally secreted IGF-I, Zhang et al. (12) prepared a protease-resistant mutant of IGFBP-4 and overexpressed it in vascular, bladder, and intestinal SMC. This resulted in SMC hypoplasia in the developing mice, whereas expression of nonmutated IGFBP-4 did not inhibit smooth muscle growth. Those results suggested that IGF-I is an important growth factor for fetal SMC proliferation. Previous studies that have attempted to inhibit IGF-I actions in blood vessels to determine its role in neointimal expansion have had mixed results. When a synthetic peptide that inhibited IGF-I binding to its receptor and IGFI-stimulated [3H]thymidine incorporation in vitro was administered to rats after carotid artery injury, it had no effect on inhibiting neointimal thickening (13). In contrast, when hypercholesterolemic pigs who had arterial lesions induced by placement of Goldblatt clamp received an infusion of an ␣V␤3 inhibitor that had been shown to block IGF-I actions, they had attenuation of neointimal expansion (14). However, that study did not exclude the possibility that the ␣V␤3 inhibitor had effects on neointimal expansion that were independent of inhibiting IGF-I ac-

5002

Nichols et al. • IGFBP-4 Inhibits Lesion Development

Endocrinology, October 2007, 148(10):5002–5010

5003

tions. Therefore this study was undertaken to determine whether infusion of a protease-resistant mutant of IGFBP-4, a very specific inhibitor of IGF-I actions, could inhibit the effects of locally secreted IGF-I on the development and expansion of neointima in pigs.

lesterol and triglycerides were 169 ⫾ 54 and 36 ⫾ 21 mg/dl, respectively. After feeding the high-fat diet for 4 wk (i.e. before treatments), cholesterol and triglycerides rose to 843 ⫾ 270 and 89 ⫾ 17 mg/dl, respectively. The high-fat diet was continued during the treatment phase of the study, and at the end of the treatment period, cholesterol and triglycerides were 741 ⫾ 90 and 88 ⫾ 19 mg/dl, respectively.

Materials and Methods Animal preparation

Morphometric analysis of neointimal area

Male and female pigs, spotted Poland/China cross, age 12 months, were treated according to the Guide for Care and Use of Laboratory Animals (National Institutes of Health Publication 85-23) using an approved protocol. The animals were fed a diet containing 1% cholesterol, 20% beef tallow, and 0.75% cholic acid for 4 wk before the initiation of the study (14). This diet has been shown to result in the appearance of cholesterol deposition and foam cells after 4 wk (15–19). Under anesthesia, a 0.5-cm Goldblatt steel clamp was applied to both common carotid and to both femoral arteries in each of seven animals. The clamps were closed to produce an 80% stenosis. The clamp technique results in the formation of neointimal lesions within 3 wk that are composed of predominantly SMC (14, 20 –23). However, because the animals have also been placed on a high-fat diet, these lesions also contain macrophages (14, 20, 23). Alzet pumps (2.0 ml; Alzet Corp., Cupertino, CA) were placed in the tissue adjacent to the artery, and infusion catheters were placed within the vessel wall under the Goldblatt clamp (14). The pump reservoirs were filled with 2.0 ml PBS that contained the specific proteins that were used in each treatment group. Four vessels from each animal (e.g. two carotids and two femorals) were randomized so that each received one of the four treatments. That is, each of the four treatments was administered to each animal. One set of vessels (n ⫽ 7) received the mutant form of IGFBP-4 (48 ␮g/ml) alone. The IGFBP-4 mutant protein had been shown to inhibit SMC migration in vitro (24). It had also been shown to be resistant to proteolytic cleavage in vitro (12). A second set of vessels received an identical concentration of nonmutated IGFBP-4. A third set of vessels received the mutant IGFBP-4 protein plus an equimolar concentration of IGF-I (10 ␮g/ml). The recombinant human IGF-I was a gift from Genentech Inc. (South San Francisco, CA). The fourth set received PBS alone. The infusion pumps delivered the experimental and control compounds at a rate of 4 ␮l/h from a 2.0-ml reservoir for 21 d. At the end of the infusion period, the animals were killed. Arterial segments were fixed in 10% buffered paraformaldehyde and paraffin embedded. Six cross-sections per vessel were obtained at 50-␮m intervals beginning at 1.9 mm and extending to 2.2 mm from the edge of the clamp (Fig. 1). The total distance covered by the sections was 0.3 mm. This area was adjacent to the catheter insertion site. Two vessels could not be analyzed for technical reasons. In one a thrombosis had occurred, and in the other the catheter had become dislodged.

Serum cholesterol and triglycerides Cholesterol and triglyceride values were determined as described previously (25, 26). At baseline while ingesting low-fat pig chow, cho-

FIG. 1. The proximal portion (Prox) of the artery encompassed by the 5-mm Goldblatt clamp was processed by histomorphometry beginning at 1.9 mm from the edge of the clamp and extending to 2.2 mm (indicated by double-headed arrow). The distal half (⬃2.5 mm, shaded) was processed for biochemical studies (modified from Ref. 14).

Arterial sections were mounted on glass slides, deparaffinized in xylene, hydrated in a series of graded alcohol solutions, stained by two methods (hematoxylin and eosin and Verhoff van Gieson), and visualized using a Nikon microphot-FXA microscope with Optronics TEC470CCD video camera system (14, 23, 27, 28). Images of the Verhoff van Gieson-stained sections (this stain is used to identify the internal and external elastic lamina) were captured by NIH Image (14, 23, 26). They were analyzed for neointimal area using the following equations: intimal area (square micrometers) ⫽ area within the internal elastic laminalumenal area. The neointimal area (circumferentially) is included in each analysis. A mean of six sections were used for each vessel. Sections stained with hematoxylin and eosin were used to clarify details in any regions that were not clear on Verhoff van Gieson. The observers were blinded as to treatment group. Measurements were made by two blinded observers. The inter- and intra-observer variability was between 4 and 7%.

Histochemical and immunohistochemical detection of macrophages, mucopolysaccharides, collagen, and proliferating cell nuclear antigen (PCNA) The paraffin-embedded sections were fixed onto Probe-On Plus charged microscopic slides, deparaffinized in xylene, and hydrated in a series of graded alcohol solutions. Endogenous peroxidases were quenched by exposure of the tissue sections to a 0.03% hydrogen peroxide in absolute methanol before antigen retrieval was performed. Trypsin (1.5 mg/ml; LabVision, Fremont, CA) was used for antigen retrieval (10 min at 37 C) for macrophage and PCNA detection but not for detection of mucopolysaccharides or collagen. Nonspecific binding was blocked by exposure of all sections to 2% horse serum. To detect macrophages, the tissue sections were then incubated with MAC387 (NeoMarkers, Fremont, CA), a monoclonal antibody that detects pig macrophages using a 1:200 dilution for 1 h at room temperature (29). This was followed by incubation with a horse antimouse secondary antibody and Vectastain R.T.U. Elite ABC reagents supplied in the Vectastain Universal Elite ABC kit (Vector Laboratories, Burlingame, CA). The stain was visualized using NovaRed (Vector) and counterstained with hematoxylin for enhanced contrast. An average of five sections were counted per pig. The average number of positively stained cells in the neointima was determined. The area of neointimal extracellular matrix staining positive for mucopolysaccharides and collagen was measured using the modified Movat’s procedure (30, 31) according to the manufacturer’s specifications (Movat’s Staining kit; Electron Microscopy Sciences, Hatfield, PA). This staining procedure stains mucopolysaccharides blue-green and collagen yellow. Images were captured with the computer-driven photomicroscope system described above. The total neointimal area was measured as described above and NIH program Image J was used with the threshold color plug-in to measure the area of the neointima that stained for mucopolysaccharides or collagen. The results for all three individual areas (neointimal, mucopolysaccharides, and collagen) are expressed as square micrometers. An average of three to five sections per vessel were analyzed. To detect PCNA, an index of cell proliferation (14, 23, 32), tissue sections were incubated with a 1:100 dilution of PC-10, a polyclonal antibody to PCNA (Dako Corp. Carpinteria, CA) for 1 h at 22 C and then exposed to the ABC Vectastain secondary horse antirabbit antibody and the chromogen diaminobenzidine. An average of three to five sections were counted per vessel. The artery was divided into four quadrants, and at least 100 neointimal and 100 medial cells were counted per quadrant. The number of cells that stained for PCNA in the neointima or media was then determined by direct counting and the result divided by the total number of cells counted in the respective regions of each

5004

Endocrinology, October 2007, 148(10):5002–5010

Nichols et al. • IGFBP-4 Inhibits Lesion Development

FIG. 2. Carotid arteries from the four treatment groups. Arterial cross-sections from the four treatment groups were stained with Verhoff van Gieson. The portions of the neointima outlined by the black box inserted on the left panels are shown at higher magnification in the right panels. A and B, IGFBP-4 wild-type; C and D, IGFBP-4 mutant; E and F, IGFBP-4 mutant plus IGF-I; G and H, vehicle. Magnification bars apply to all of the panels in each respective column. The degree of magnification was identical from panels A, C, E, and G and for panels B, D, F, and H. The neointimal areas for this specific figure are as follows (in ␮m2): A, 722,000; C, 47,850; E, 586,125; G, 257,450.

Nichols et al. • IGFBP-4 Inhibits Lesion Development

Endocrinology, October 2007, 148(10):5002–5010

5005

TABLE 1. Changes in neointimal area Neointimal area ⫾ (␮m2 )

Treatment

IGF-I IGFBP-4 (wild type) (n ⫽ 6) IGFBP-4 (mutant) (n ⫽ 6) IGFBP-4 (mutant) ⫹ IGF-I (n ⫽ 7) Vehicle (n ⫽ 7)

SD

P value compared with vehicle

211,371 ⫾ 46,670 100,618 ⫾ 42,604 279,522 ⫾ 39,887 209,593 ⫾ 41,143

P value compared with wild type

0.862 0.032 0.296

0.046 0.224 0.862

P value compared mutant plus IGF-I

0.224 0.002 0.296

n, The number of vessels that were analyzed in each treatment group. arterial cross-section to calculate the percentage of cells that were labeled (13, 14, 23, 32). Endothelial cells were excluded from this analysis. For immunohistochemical methods, detecting macrophages or PCNA, positive controls consisted of detection of staining of the respective antigens in paraffin-embedded intestine and pig coronary and aortic atherosclerotic plaques that had been obtained in other studies, (14, 25, 33). These tissues were also used to confirm that the Movat’s staining reagents were functioning properly. Negative controls were processed as described above except that the primary antibody was not used. The analyses were evaluated by two blinded observers who had a 5% inter-observer variability.

Preparation and purification of protease-resistant IGFBP-4 Because proteolysis of IGFBP-4 results in a marked reduction in its affinity for IGF-I (24), we expressed a mutant form of IGFBP-4 that had been shown to be resistant to proteolytic cleavage (12). The molecular cloning and mutagenesis methods necessary to prepare the IGFBP-4 mutant and cDNA vector have been described previously (12). The vector containing this mutant was provided by Dr. James Fagin, University of Cincinnati, Cincinnati, OH. This was transfected into B104 cells using a previously described method (27). Cells were then grown to confluency in medium containing G-418 (500 ␮g/ml), and conditioned medium was collected from confluent cultures after a 48-h incubation. Approximately 2 liters of conditioned medium was collected and then purified as described previously for IGFBP using phenyl-Sepharose hydrophobic interaction chromatography, IGF affinity chromatography, and reverse-phase HPLC, C4 column (24). The purified fractions were analyzed by SDS-PAGE with immunoblotting for c-myc. To remove residual nonmutated IGFBP-4, the mutant protein was passed over a reverse-phase C4HPLC column and eluted with an acetonitrile gradient from 20 – 40% over 30 min. The final preparation was concentrated and then reconstituted in 17.5 ml PBS. Control lesions received the same concentration of nonmutated, human IGFBP-4 that had been purified by a previously described method (24). When the affinities of the mutant and nonmutated forms of IGFBP-4 for IGF-I were analyzed, they were 1.8 and 1.6 ⫻ 10⫺9 liters/mole, respectively. In addition, both forms of IGFBP-4 have similar affinities for IGF-II (e.g. 3.2 and 3.7 ⫻ 10⫺9 liters/ mole). The IGF-I that was used in these studies was provided by Genentech, and it had been purified to homogeneity.

Quantitation of IGFBP-4 For all biochemical analyses, the arterial tissue that was obtained within 2.5 mm distal to the catheter insertion site was analyzed (Fig. 1). The tissues were removed rapidly by dissection, wet weight was determined, and then they were frozen immediately in liquid nitrogen. All extractions were performed on the liquid nitrogen frozen tissue. To determine the abundance of intact and cleaved IGFBP-4, lesion segments were homogenized in 1.0 ml/100 mg tissue of 100 mm TrisTABLE 2. Changes in SMC PCNA Labeling Treatment

% labeling in neointima ⫾ SD

% labeling in media ⫾ SD

IGFBP-4 (wild type) (n ⫽ 6) IGFBP-4 (mutant) (n ⫽ 6) IGFBP-4 (mutant) ⫹ IGF-I (n ⫽ 7) Vehicle (n ⫽ 7)

17.5 ⫾ 3.8 11 ⫾ 2.7a 17.8 ⫾ 3.7 15.7 ⫾ 3.7

1.8 ⫾ 1.1 1.5 ⫾ 0.9 4.3 ⫾ 0.6 2.2 ⫾ 1.6

a

P ⬍ 0.05 compared with vehicle.

HCl (pH 7.2) that contained 5 mm EGTA, 2% Triton X-100, 10 ␮g/ml aprotinin, 2 ␮g/ml leupetin, 1 mm phenylmethylsulfonyl fluoride, and 1 mm pepstatin for 3 min at room temperature. The extract was centrifuged at 14,000 ⫻ g for 20 min. IGFBP-4 was quantified by direct immunoblotting of the supernatant. Twenty-five microliters of the extract (40 ␮g protein) were mixed 1:1 with 2⫻ Laemmli sample buffer, and the proteins were separated by SDS-PAGE, 9% gel. Protein concentrations were determined by bicinchoninic acid analysis (24). After transfer to Immobilon membranes, the membranes were washed as described previously and then incubated with a 1:5000 dilution of anti-IGFBP-4 antiserum (34). The immune complexes were visualized using enhanced chemiluminescence. After autoradiographic analysis, the gels were scanned by scanning densitometry (AGFA Scanner, Ridgefield Park, NJ), and the data were analyzed using NIH Image version 1.6. Each lesion extract was analyzed three times, and the results represent the mean of three separate analyses. Intact IGFBP-4 was run as a control, and the intensity of that band was used to standardize the measurements of band intensity. In some experiments, direct immunoblotting of the myclabeled IGFBP-4 mutant was undertaken to determine whether the mutated protein remained intact. After electrophoresis, the labeled proteins were transferred to Immobilon membranes and immunoblotted using a 1:1000 dilution of antihuman myc antiserum. The immune complexes were visualized by enhanced chemiluminescence. The myc IGFBP-4 was run in a parallel lane as a standard for comparison.

IGFBP-5 analysis To determine the effect of inhibiting IGF-I actions in SMC, the amount of IGFBP-5 in the lesion microenvironment was quantified by immunoblotting for human IGFBP-5. IGFBP-5 synthesis has been shown to be directly stimulated in this cell type by IGF-I (35); therefore, induction of IGFBP-5 serves as a marker of IGF-I action. Porcine IGFBP-5 reacts well with the antihuman IGFBP-5 antiserum, and therefore, this antiserum was used to detect the porcine form of the protein. Approximately 50 ␮g arterial protein extract was separated by SDS-PAGE (7.5% gel) and transferred to Immobilon filters. The filters were exposed to a 1:2000 dilution of antihuman IGFBP-5. Pure human IGFBP-5 was run as a standard in a parallel lane. Estimates of intact IGFBP-5 (30 kDa) band intensity were determined by scanning densitometry and analyzed using NIH Image. The results are the mean ⫾ sem of three separate determinations of each lesion.

Determination of IGF-I IGF-I was quantified by a RIA using reagents supplied by ALPCO (22 IGF-I R20; American Laboratory Products Co., Windham, NH). Briefly, the samples were solubilized by homogenizing the tissue (100 mg) in the buffer listed previously and adding 13 ␮l of the homogenate (20 ␮g protein) to 0.5 ml assay buffer that contained 0.05 mm Na acetate (pH 3.1). An excess of IGF-II (0.2 ␮g/ml) was added to neutralize binding protein interference (36). Cross-reactivity of the antibody for IGF-II is less than 0.05%. After IGF-II addition, the pH of the assay buffer was neutralized to pH 7.0, and [125I]IGF-I (20,000 cpm/tube, specific activity 130 mCi/mg) plus a 1:1000 dilution of anti-IGF-I antisera were added. After a 48-h incubation at 4 C, the immune complexes were precipitated after the addition of goat antirabbit antiserum. The results were read against a known standard curve using human IGF-I as a standard. The lower limit of detection is 0.2 ng/ml. The interassay coefficient of variation is 5.5%.

5006

Endocrinology, October 2007, 148(10):5002–5010

Nichols et al. • IGFBP-4 Inhibits Lesion Development

TABLE 3. Extracellular matrix content Area (␮m2)

Treatment

Neointimal

Vehicle (n ⫽ 7) IGFBP-4 mutant (n ⫽ 6) IGF-I ⫹ IGFBP-4 mutant (n ⫽ 7) Nonmutated IGFBP-4 (n ⫽ 6)

191,619 ⫾ 27,100 76,875 ⫾ 15,506a (⫺60%) 186,285 ⫾ 33,355 (⫺10%) 188,992 ⫾ 47,678 (⫺1.4%)

Mucopolysacchrides

Collagen

Total

5,206 ⫾ 3,915 1,909 ⫾ 1,187a (⫺63%) 4,727 ⫾ 2,358 (⫺9%) 3,067 ⫾ 2,407a (⫺41%)

3,634 ⫾ 1,959 1,497 ⫾ 732a (⫺58%) 2,067 ⫾ 1,014a (⫺43%) 7,880 ⫾ 3,152 (⫹108%)

8,840 ⫾ 2,214 3,406 ⫾ 1,721a (⫺61%) 6,794 ⫾ 2,422 (⫺24%) 10,947 ⫾ 5,844 (⫹12%)

The percentage change compared with control is shown in parentheses. n, The number of vessel areas in each treatment group that were analyzed. a P ⬍ 0.05 compared with vehicle.

Statistical analysis Comparison of lesion sizes, PCNA labeling, and extracellular matrix area was analyzed by four-way ANOVA followed by multivariate analysis and t tests for some individual comparisons. When the changes in carotid and femoral arteries were compared, there were no statistically significant differences in lesion area in the vessels that were exposed to vehicle only (Ref. 14) and present study); therefore, data from both vessels were pooled. The differences in IGFBP-4 or -5 band intensities were compared using Student’s t test.

Results

Representative arterial sections for each of the treatments are shown in Fig. 2. It can be seen that the vessel that received the IGFBP-4 mutant (Fig. 2, C and D) had a substantial reduction in neointimal area compared with the vessels that received the other three treatments. The lesion areas in the neotinima for the various treatments are shown in Table 1. The initial four-way ANOVA compared individual animal, treatment group, artery (e.g. femoral vs. carotid), and right or left side. Only the treatment group showed a significant effect (P ⬍ 0.04). The six lesions that received the protease-resistant form of IGFBP-4 had a marked reduction in lesion area. Direct comparison showed that there was a 53% reduction in neointimal area compared with saline control. The difference in neointimal area was highly significant when compared with the areas of the lesions that received PBS, wild-type IGFBP-4, or the IGFBP-4 mutant plus IGF-I (Table 1). When the neointimal area was quantified in the lesions that received native IGFBP-4, there was no change in lesion area compared with the PBS control. As an additional control, an equimo-

lar concentration of IGF-I was infused with the proteaseresistant mutant form of IGFBP-4. These lesions also showed no change in neointimal area compared with the PBS control. Therefore, only the vessels that received the protease-resistant mutant of IGFBP-4 showed a significant decrease in neointimal area (Table 1). These findings strongly suggest that IGFBP-4 is binding to IGF-I in the extracellular fluid and preventing it from interacting with receptors and thus reducing neointimal thickening. They further suggest that proteolysis of IGFBP-4 is reducing its capacity to inhibit the effect of IGF-I. To further define the mechanism by which IGFBP-4 was functioning to reduce neointimal area, PCNA labeling was determined. As shown in Table 2, the IGFBP-4 mutant resulted in a significant decrease in the number of PCNApositive cells within the neointima. In contrast, there was no significant change in the number of positive cells in the media. This finding suggests that the reduction in lesion size is due to inhibition of cell proliferation, although an additional effect on cell migration or on cell size cannot be excluded. To determine the degree to which lesion size might have been altered by changes in extracellular matrix, the lesions were stained for mucopolysaccharides and collagen. The results in Table 3 show that the mutant IGFBP-4 reduced mucopolysaccharides by 63% and collagen by 58% compared with vehicle. These changes were comparable to the changes in neointimal area. In contrast, lesions that received the combined treatment had a 24% reduction in total extracellular matrix staining that was not significant. The lesions that received nonmutated IGFBP-4

FIG. 3. Immunohistochemical detection of macrophages in neointima using the monoclonal antibody MAC387. A, Portion of the artery with neointima; B, area enclosed by the black rectangle in A shown at higher magnification. Macrophages stain deep red with this detection method.

Nichols et al. • IGFBP-4 Inhibits Lesion Development

had a 41% reduction in mucopolysaccharides but a 108% increase in collagen; therefore, the total extracellular matrix staining was increased minimally (e.g. 12%), and this change was not significant. To further characterize the lesions, they were stained for the presence of macrophages. As shown in Fig. 3, macrophages could be detected at the base of the neointima. When the number of macrophages that were present in the neointimal areas that received each of the four treatments was determined, there were no significant differences. To determine that the infused mutant form of IGFBP-4 remained intact, both immunoblotting of intact IGFBP-4 as well as immunoblotting for myc-labeled IGFBP-4 were undertaken. Immunoblotting of IGFBP-4 showed that the lesions that received the mutant form of IGFBP-4 had easily detectible intact IGFBP-4 bands (Fig. 4, left side, lanes 1– 4, A–C). In contrast, the lesions that received native, nonmutated IGFBP-4 or PBS showed minimal intact IGFBP-4. Scanning densitometry revealed that the arteries that received nonmutated IGFBP-4 had a 78 ⫾ 14% reduction in the intensity of the intact IGFBP-4 band compared with the arteries that received the IGFBP-4 mutant (Table 4). The coinfusion of IGF-I did not result in any

FIG. 4. A–C. IGFBP-4 immunoblot (left panel) and myc-tagged proteins (right panel). The vessels that received various treatments were extracted, and the extracts immunoblotted for IGFBP-4 (lanes 1– 4, left panel) or the same extracts were immunoblotted for the myc tag (lanes 1– 4, right panel). The treatments were as follows: lane 1, PBS control; lane 2, nonmutated IGFBP-4; lane 3, IGFBP-4 mutant plus IGF-I; lane 4, IGFBP-4 mutant alone. The arrow denotes the position of intact IGFBP-4. The lower band detected in the lanes in the right panel is a nonspecific band detected by the myc antibody that does not react with the IGFBP-4 antibody. Lane 5 in panel A shows the myctagged protein standard that was added to a lesion extract in vitro before immunoblotting. Panels B and C show the same four treatments from two different animals. To control for loading differences, the blots were stripped and reprobed for ␣-smooth muscle actin, which is shown as the lower band in the left panel. The results show that there were no significant loading differences in the 12 samples that were immunoblotted for IGFBP-4.

Endocrinology, October 2007, 148(10):5002–5010

5007

TABLE 4. IGFBP-4 and IGFBP-5 band intensities Treatment

IGFBP-4 (wild type) (n ⫽ 6) IGFBP-4 (mutant) (n ⫽ 6) IGFBP-4 mutant ⫹ IGF-I (n ⫽ 7) Vehicle (n ⫽ 7)

Scanning units ⫾ SD IGFBP-4

IGFBP-5

3,079 ⫾ 487 14,022 ⫾ 1,399a 13,337 ⫾ 1,524a 2,664 ⫾ 505

7,344 ⫾ 901 1,603 ⫾ 330a 10,664 ⫾ 1,336b 8,822 ⫾ 1,114

n ⫽ the number of vessel areas in each treatment group that were analyzed. a P ⬍ 0.01 compared with vehicle. b P ⬍ 0.05 compared with vehicle.

change in the amount of intact IGFBP-4. The results were further confirmed by immunoblotting for the myc tag (Fig. 3, right side, lanes 1– 4, A–C). These results showed that there was an easily detectible, intact IGFBP-4 band in lesions that received the mutant form of IGFBP-4 and that no myc-containing protein was detected in the lanes that received vehicle or IGF-I alone. These results also confirm that the infused mutant IGFBP-4 remained in an intact form. The results strongly support the conclusion that mutation of IGFBP-4 prevented its degradation in porcine arteries as had been shown previously in cultured cells and that despite the duration of the infusion, most of the mutant IGFBP-4 that was infused that was present in the vessels at the conclusion of the experiment was in an intact form and thus capable of binding to IGF-I. To further confirm that the infused mutant form of IGFBP-4 could inhibit IGF-I actions, the abundance of the marker protein IGFBP-5, whose synthesis is directly stimulated in SMC by IGF-I (35), was quantified by immunoblotting. The results showed almost no detectible IGFBP-5 in the lesions that received the mutant form of IGFBP-4 (Fig. 5). Quantification of the IGFBP-5 band intensities showed that there was an 82 ⫾ 9% (P ⬍ 0.01) decrease in the abundance of IGFBP-5 in lesions that received mutant IGFBP-4 compared with PBS control (Table 4). Importantly, the lesions that received the nonmutated form of IGFBP-4 showed a 17 ⫾ 7% decrease in IGFBP-5 abundance compared with lesions that received saline, and this change was not significant. The lesions that received the mutant IGFBP-4 plus IGF-I showed a 21 ⫾ 7% (P ⬍ 0.05) increase in IGFBP-5. Therefore, using this marker protein, we conclude that IGF-I action was markedly inhibited in the lesions that received the IGFBP-4 mutant compared

FIG. 5. IGFBP-5 immunoblot. The extracts obtained from four arterial samples from two pigs are shown. Lane 1, IGFBP-5 standard added to PBS control extract in vitro. Lanes 2–5 were obtained from the same animal, and lanes 6 –9 were from a different animal. Lanes 2 and 6, Mutant IGFBP-4; lanes 3 and 7, wild-type IGFBP-4; lanes 4 and 8, mutant IGFBP-4 plus IGF-I; lanes 5 and 9, PBS. The arrow denotes the position of IGFBP-5. The lower panel shows the SMC actin control.

5008

Endocrinology, October 2007, 148(10):5002–5010

with PBS control and that the lesions that received mutant plus IGF-I had a significant stimulation of IGFBP-5. As an additional control to confirm that the differences were not simply due to differences in the total IGF-I concentration, total IGF-I was quantified after removal of binding protein interference (36). The lesions that received IGF-I plus mutant IGFBP-4 showed a 43 ⫾ 9% increase in the total IGF-I concentration (93 ⫾ 30 ng/50 mg tissue) compared with saline control lesions (65 ⫾ 18 ng/50 mg tissue) (P ⬍ 0.05). The lesions that received only the mutant IGFBP-4 (63 ⫾ 18 ng/50 mg tissue) or wild-type IGFBP-4 (62 ⫾ 17 ng/50 mg tissue) showed no difference compared with saline control (P value not significant). Therefore, the effect of IGFBP-4 mutant was not due to reduction in total IGF-I concentration but rather to a reduction in IGF-I action as assessed by the decrease in IGFBP-5. Discussion

IGF-I interacts cooperatively with other growth factors to stimulate SMC proliferation and migration in vitro, both processes that are believed to contribute to cellular accumulation within lesions (37, 38). In addition, targeted overexpression of IGF-I in arteries of transgenic mice increased neointimal area (9). Studies that have investigated the role of IGFBP in regulating IGF-I action both in arterial wall cell types and in other tissues have shown that these proteins can either inhibit or stimulate IGF-I actions depending upon the particular protein that is being synthesized, the cellular location of the protein, whether or not it has been proteolytically cleaved, and whether or not it is associated with the cell surface or extracellular matrix (10, 11, 39 – 42). Most in vitro and a few in vivo studies have shown that if a large molar excess of IGFBP is added to extracellular fluids and their affinity is high enough to inhibit IGF-I binding to receptors, they will inhibit IGF-I actions (11, 39). Pertinent to this study, overexpression of a protease-resistant form of IGFBP-4 during postnatal development in mice resulted in SMC hypoplasia (12, 43). Expression of native IGFBP-4 that was proteolytically cleaved did not result in growth inhibition, and ligand-blotting analysis of tissues showed that the animals that overexpressed the protease-resistant mutant form of IGFBP-4 had substantially higher intact IGFBP-4 concentrations, leading the authors to conclude that inhibition of IGF-I actions resulted in inhibition of SMC growth. More recently, mice that did not express pregnancy-associated plasma protein A (a protease that has been shown to cleave IGFBP-4) were shown to have decreased neointima formation after mechanical injury (44). However, in that study, no controls were provided to show that the effect of eliminating pregnancy-associated plasma protein A resulted in increased intact IGFBP-4 in blood vessels; therefore, it is difficult to conclude that the result obtained was specific for inhibition of IGF-I actions. Our findings suggest that IGFBP-4 is functioning primarily by inhibiting IGF-I binding to its receptor rather than a direct inhibitory effect that is mediated through a non-IGF-dependent mechanism. Because a high concentration of the infused mutant form of IGFBP-4 remained

Nichols et al. • IGFBP-4 Inhibits Lesion Development

intact, it is likely that it prevented IGF-I binding and thus inhibited receptor activation. Because infusion of mutant IGFBP-4 was associated with no change in the total IGF-I concentration compared with lesions that received vehicle alone, the findings strongly suggest that there was a reduction in the concentration of IGF-I that was available to receptors, thus resulting in decreased SMC stimulation. Although several binding proteins, principally IGFBP-3 and IGFBP-5, have been shown to have non-IGF-dependent actions in cell culture models, direct addition of IGFBP-4 to SMC cultures was shown to inhibit IGF-Istimulated DNA synthesis, but it had no effect that was independent of IGF-I (24). Some studies have shown that fragments of IGFBP-4 can have direct effects (45). However, some of these fragments have been shown to bind IGF-I, suggesting that part of their ability to inhibit IGF-I action is through inhibiting IGF-I binding to its receptor (46). In this study, we included a group of vessels that were treated with an IGFBP-4 plus an equimolar concentration of IGF-I. Vessels that received this combination had no reduction in lesion area and no attenuation of the ability of IGF-I to stimulate IGFBP-5. Therefore, we conclude that IGFBP-4 in this model system is working primarily by inhibiting IGF-I association with its receptor and thus inhibiting the ability of IGF-I to stimulate SMC growth and migration. It is also possible that IGFBP-4 is acting to inhibit the actions of IGF-II. Although we did not measure changes in IGF-II expression in lesions, others have shown that targeted overexpression of IGF-II in mice leads to focal intimal thickening that was attributed to an increase in SMC proliferation (47). Because IGFBP-4 binds IGF-II with high affinity, its ability to inhibit neointimal formation in animal models of atherosclerosis may also be due to inhibition of IGF-II actions as well as IGF-I. Although our studies did not definitively identify the mechanism by which IGF-I is stimulating neointimal expansion, they suggest that cell proliferation is a major component of the process because administration of protease-resistant IGFBP-4 resulted in a decrease in PCNA labeling. This finding does not exclude the possibility that IGF-I also is increasing neointimal area by stimulating SMC migration or by increasing SMC size. Additionally, there was no change in macrophage abundance, suggesting that IGF-I does not function by altering macrophage recruitment. Our analysis of changes in extracellular matrix was consistent with these results. Specifically, infusion of IGFBP-4 was associated with a reduction in extracellular matrix area that stained for mucopolysaccharides and collagen, and this reduction was similar to the change in total neointimal area. This is consistent with the changes obtained in neointimal area, but it does not completely exclude the possibility that there was a preferential decrease in extracellular matrix synthesis per cell because this was not measured directly. Because IGF-I is a known stimulant of extracellular matrix protein synthesis, these results are also consistent with the conclusion that mutant IGFBP-4 inhibited IGF-I actions. These studies confirm the findings in the transgenic mice that a protease-resistant form of IGFBP-4 inhibits IGF-I actions in arterial smooth muscle and extend them

Nichols et al. • IGFBP-4 Inhibits Lesion Development

to a model of neointimal hyperplasia. Infusion of the protease-resistant mutant form of IGFBP-4 resulted in reduction in neointimal expansion and inhibition of IGF-I action as determined by assessing the abundance of IGFBP-5 in the arterial tissues. The findings are consistent with a role for IGF-I in stimulating neointimal expansion and support the conclusion that attempts to target IGF-I action during the proliferative phase of atherosclerosis may result in a decrease in neointimal thickening. Acknowledgments We gratefully acknowledge Ms. Laura Lindsey for her help in preparing the manuscript. We thank Dr. James Fagin for his gift of the plasmid containing the mutant IGFBP-4 c-DNA. Received April 30, 2007. Accepted July 11, 2007. Address all correspondence and requests for reprints to: David R. Clemmons, M.D., CB 7170, 8024 Burnett-Womack, Division of Endocrinology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7170. E-mail: [email protected]. This work was supported by National Institutes of Health Grant HL-56850. T.C.N., W.H.B., E.M., J.S., M.R., K.S., and D.R.C. have nothing to declare.

References 1. Cerecek B, Fishbein MC, Forrester JS, Helfant RH, Fagin JA 1990 Induction of insulin-like growth factor I messenger RNA in rat aorta after balloon denudation. Circ Res 66:1755–1760 2. Delafontaine P 1998 Growth factors and vascular smooth muscle cell growth responses. Eur Heart J 19(Suppl G):G18 –G22 3. Johansson GS, Arnqvist HJ 2006 Insulin and IGF-I action on insulin receptors, IGF-I receptors and hybrid insulin/IGF-I receptors in vascular smooth muscle cells. Am J Physiol Endocrinol Metab 291:E1124 –E1130 4. Imai Y, Clemmons DR 1999 Roles of phosphatidylinositol 3-kinase and mitogen-activated protein kinase pathways in stimulation of vascular smooth muscle cell migration and deoxyribonucleic acid synthesis by insulin-like growth factor-I. Endocrinology 140:4228 – 4235 5. Khorsandi MJ, Fagin JA, Giannella-Neto D, Forrester JS, Cercek B 1992 Regulation of insulin-like growth factor-I and its receptor in rat aorta after balloon denudation. Evidence for local bioactivity. J Clin Invest 90:1926 –1931 6. Khorsandi M, Fagin JA, Fishbein MC, Forrester JS, Cercek B 1992 Effects of hypophysectomy on vascular insulin-like growth factor-I gene expression after balloon denudation in rats. Atherosclerosis 93:115–122 7. Banskota NK, Taub R, Zellner K, Olsen P, King GL 1989 Characterization of induction of protooncogene c-myc and cellular growth in human vascular smooth muscle cells by insulin and IGF-I. Diabetes 38:123–129 8. Kirstein M, Aston C, Hintz R, Vlassara H 1992 Receptor-specific induction of insulin-like growth factor I in human monocytes by advanced glycosylation end product-modified proteins. J Clin Invest 90:439 – 446 9. Zhu B, Zhao G, Witte DP, Hui DY, Fagin JA 2001 Targeted overexpression of IGF-I in smooth muscle cells of transgenic mice enhances neointimal formation through increased proliferation and cell migration after intraarterial injury. Endocrinology 142:3598 –3606 10. Jones JI, Clemmons DR 1995 Insulin-like growth factors and their binding proteins: biological actions. Endocr Rev 16:3–34 11. Qin X, Strong DD, Baylink DJ, Mohan S 1998 Structure-function analysis of the human insulin-like growth factor binding protein-4. J Biol Chem 273: 23509 –23516 12. Zhang M, Smith EP, Kuroda H, Banach W, Chernausek SD, Fagin JA 2002 Targeted expression of a protease-resistant IGFBP-4 mutant in smooth muscle of transgenic mice results in IGFBP-4 stabilization and smooth muscle hypotrophy. J Biol Chem 277:21285–21290 13. Hayry P, Myllarniemi M, Aavik E, Alatalo S, Aho P, Yilmaz S, RaisanenSokolowski A, Cozzone G, Jameson BA, Baserga R 1995 Stabile D-peptide analog of insulin-like growth factor-1 inhibits smooth muscle cell proliferation after carotid ballooning injury in the rat. FASEB J 9:1336 –1344 14. Nichols TC, du Laney T, Zheng B, Bellinger DA, Nickols GA, Engleman W, Clemmons DR 1999 Reduction in atherosclerotic lesion size in pigs by ␣V␤3 inhibitors is associated with inhibition of insulin-like growth factor-I-mediated signaling. Circ Res 85:1040 –1045 15. Gerrity RG, Naito HK, Richardson M, Schwartz CJ 1979 Dietary induced atherogenesis in swine: morphology of the intima in prelesion stages. Am J Pathol 95:775–792

Endocrinology, October 2007, 148(10):5002–5010

5009

16. Cornhill JF, Barrett WA, Herderick EE, Mahley RW, Dry DL 1985 Topographic study of sudanphilic lesions in cholesterol-fed minipigs by image analysis. Arteriosclerosis 5:415– 426 17. Thomas WA, Lee KT, Kim DN 1985 Pathogenesis of atherosclerosis in the abdominal aorta and coronary arteries of swine in the first 90 days of hyperlipidemic diet. In: Tumbleson ME, ed. Swine in biomedical research. Vol 3. New York: Plenum Press; 1151–1525 18. Griggs TR, Bauman RW, Reddick RL, Read MS, Koch GG, Lamb MA 1986 Development of coronary atherosclerosis in swine with severe hypercholesterolemia: lack of influence of von Willebrand factor or acute intimal injury. Arteriosclerosis 6:155–165 19. Lamb MA, Manning JE, Reddick RL, Griggs TR 1984 Smooth muscle cell proliferation in response to endothelial injury in coronary arteries of normal and von Willebrand’s disease swine. Arteriosclerosis 4:84 –90 20. Booth RFG, Martin JF, Honey AC, Hassall DG, Beesley JE, Moncada S 1989 Rapid development of atherosclerotic lesions in the rabbit carotid artery induced by perivascular manipulation. Atherosclerosis 76:257–268 21. Kockx MM, De-Meyer GR, Jacob WA, Bult H, Herman AG 1992 Triphasic sequence of neointimal formation in the cuffed carotid artery of the rabbit. Arterioscler Thromb 12:1447–1457 22. Kockx MM, De-Meyer GR, Andries LJ, Bult H, Jacob WA, Herman AG 1993 The endothelium during cuff-induced neointima formation in the rabbit carotid artery. Arterioscler Thromb 13:1874 –1884 23. Nichols TC, Bellinger DA, Reddick RL, Koch GG, Sigman JL, Erickson G, du Laney T, Johnson T, Read MS, Griggs TR 1998 von Willebrand factor does not influence atherogenesis in arteries subjected to altered shear stress. Arterioscler Thromb Vasc Biol 18:323–330 24. Rees C, Clemmons DR, Horvitz GD, Clarke JB, Busby WH 1998 A proteaseresistant form of insulin-like growth factor (IGF) binding protein 4 inhibits IGF-1 actions. Endocrinology 139:4182– 4188 25. Nichols TC, Bellinger DA, Tate DA, Reddick RL, Read MS, Koch GG, Brinhous KM, Griggs TR 1990 von Willebrand factor and occlusive arterial thrombosis: a study in normal and von Willebrand’s disease pigs with dietinduced hypercholesterolemia and atherosclerosis. Arteriosclerosis 10:449 – 461 26. Hasler-Rapacz, JO, Nichols TC, Griggs TR, Bellinger DA, Rapacz J 1994 Familial and diet-induced hypercholesterolemia is swine: lipid, ApoB, and ApoA-I concentrations and distributions in plasma and lipoprotein subfractions. Arterioscler Thromb 14:923–930 27. Brodala N, Merricks EP, Bellinger DA, Damrongsri D, Offenbacher S, Back J, Madianos P, Sotres D, Chang YL, Koch G, Nichols TC 2005 Porphyromonas gingivalis bacteremia induces coronary and aortic atherosclerosis in normocholesterolemic and hypercholesterolemic pigs. Arterioscler Thromb Vasc Biol 25:1446 –1451 28. Dumont D, Behler RH, Nichols TC, Merricks EP, Gallippi CM 2006 ARFI imaging for noninvasive material characterization of atherosclerosis. Ultrasound Med Biol 32:1703–1711 29. Yorozuya M, Suzuki H, Iso Y, Shibata M, Nakatani M, Koba S, Murakami M, Katagiri T, Takeyama Y 2002 Comparison of the morphological changes of restenosis after the implantation of various types of stents in a swine model. Coron Artery Dis 12:305–312 30. Movat HZ 1955 Demonstration of all connective tissue elements in a single section: pentachrome stains. AMA Arch Pathol 60:289 –295 31. Roman N, Perkins SF, Perkins Jr EM, Dolnick EH 1967 Orcein-hematoxylin in iodized ferric chloride as a stain for elastic fibers, with metanil yellow counterstaining. Stain Technol 42:199 –202 32. Azrin MA, Mitchel JF, Fram DB, Pedersen CA, Cartun RW, Barry JJ, Bow LM, Waters DD, McKay RG 1994 Decreased platelet deposition and smooth muscle cell proliferation after intramural heparin delivery with hydrogel-coated balloons. Circulation 90:433– 441 33. Nichols TC, Bellinger DA, Davis KE, Koch GG, Reddick RL, Read MS, Rapacz J, Hasler-Rapacz J, Brinkhous KM, Griggs TR 1992 Porcine von Willebrand disease and atherosclerosis: influence of polymorphism in apolipoprotein B100 genotype. Am J Pathol 140:403– 415 34. Parker A, Gockerman A, Busby WH, Clemmons DR 1995 Properties of an insulin-like growth factor-binding protein-4 protease that is secreted by smooth muscle cells. Endocrinology 136:2470 –2476 35. Duan C, Hawes SB, Prevette T, Clemmons DR 1996 Insulin-like growth factor-I (IGF-I) regulates IGF-binding protein-5 synthesis through transcriptional activation of the gene in aortic smooth muscle cells. J Biol Chem 271: 4280 – 4288 36. Blum WF, Breier BH 1994 Radioimmunoassays for IGFs and IGFBPs. Growth Regul 1:11–19 37. Clemmons DR, Van Wyk JJ 1985 Evidence for a functional role of endogenously produced somatomedinlike peptides in the regulation of DNA synthesis in cultured human fibroblasts and porcine smooth muscle cells. J Clin Invest 75:1914 –1918 38. Jawien A, Bowen-Pope DF, Lindner V, Schwartz SM, Clowes AW 1992 Platelet-derived growth factor promotes smooth muscle migration and intimal thickening in a rat model of balloon angioplasty. J Clin Invest 89:507–511 39. Imai Y, Busby Jr WH, Smith CE, Clarke JB, Garmong AJ, Horwitz GD, Rees

5010

40.

41.

42.

43.

Endocrinology, October 2007, 148(10):5002–5010

C, Clemmons DR 1997 Protease-resistant form of insulin-like growth factorbinding protein 5 is an inhibitor of insulin-like growth factor-I actions on porcine smooth muscle cells in culture. J Clin Invest 100:2596 –2605 Parker A, Rees C, Clarke J, Busby Jr WH, Clemmons DR 1998 Binding of insulin-like growth factor (IGF)-binding protein-5 to smooth-muscle cell extracellular matrix is a major determinant of the cellular response to IGF-I. Mol Biol Cell 9:2383–2392 Gockerman A, Prevette T, Jones JI, Clemmons DR 1995 Insulin-like growth factor (IGF)-binding proteins inhibit the smooth muscle cell migration responses to IGF-I and IGF-II. Endocrinology 136:4168 – 4173 Conover CA, Durham SK, Zapf J, Masiarz FR, Kiefer MC 1995 Cleavage analysis of insulin-like growth factor (IGF)-dependent IGF-binding protein-4 proteolysis and expression of protease-resistant IGF-binding protein-4 mutants. J Biol Chem 270:4395– 4400 Wang J, Niu W, Witte DP, Chernausek SD, Nikiforov YE, Clemens TL, Sharifi B, Strauch B, Fagin JA 1998 Overexpression of insulin-like growth factor-binding protein-4 (IGFBP-4) in smooth muscle cells of transgenic mice

Nichols et al. • IGFBP-4 Inhibits Lesion Development

44.

45. 46.

47.

through a smooth muscle ␣-actin-IGFBP-4 fusion gene induces a smooth muscle hypoplasia. Endocrinology 139:2605–2614 Resch ZT, Simari RD, Conover CA 2006 Targeted disruption of the pregnancy-associated plasma protein-A is associated with diminished smooth muscle cell response to insulin-like growth factor-I and resistance to neointimal hyperplasia after vascular injury. Endocrinology 147:5634 –5640 Sivwanowicz I, Popowicz GM, Wisniewska M, Huber R, Kuenkele KP, Lang K, Engh RA, Holak TA 2005 Structural basis for the regulation of insulin-like growth factors by IGF binding proteins. Structure 13:155–167 Fernandez-Tornero C, Lozano RM, Rivas G, Jimenez MA, Standker L, Diaz-Gonzalez D, Forssmann WG, Cuevas P, Romero A, Gimenez-Gallego G 2005 Synthesis of the blood circulating C-terminal fragment of insulin-like growth factor (IGF) binding protein-4 in its native conformation. Crystallization, heparin and IGF binding, and osteogenic activity. J Biol Chem 280:18899 –18907 Zaina S, Pettersson L, Ahren B, Branen L, Hassan AB, Lindholm M, Mattsson R, Thyberg J, Nilsson J 2002 Insulin-like growth factor II plays a central role in atherosclerosis in a mouse model. J Biol Chem 277:4505– 4511

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.