Antagonism of vascular endothelial growth factor results in microvessel attrition and disorganization of wound tissue KRISHNAMURTHY P GUDEHITHLU, NAILA AHMED, HENRY WU, NATALIA O LITBARG, SANDRA L GARBER, JOSE A L ARRUDA, GEORGE DUNEA, and ASHOK K SINGH CHICAGO, ILLINOIS Vascular endothelial growth factor (VEGF) is a potent growth factor that is indispensable for the development of blood vessels in the fetus and for wound healing in adults VEGF likely plays a role in maintaining the blood vessels once they have been formed It is not clear, however, whether a low tissue VEGF (caused either by disease or by systemic administration of VEGF antagonists) can cause abnormalities in preexisting blood vessels, especially of wound tissue that requires high local levels of VEGF for healing The present study investigated the effect of VEGF antagonism on blood vessels of foreign-body granulomas (a model of wound-healing tissue) Granulomas were induced by implanting perforated polyvinyl tubes into the subcutaneous tissue of rats and allowed to develop for 14 days, at which time the implanted tubes were completely encapsulated by the subcutaneous tissue The encapsulated granulomas consisted of distinct histological layers, of which the middle layer was well perfused by a rich supply of microvessels Morphologically, the granuloma remained “stable” after developing for 14 days At week, VEGF levels in the granuloma fluid, which is in equilibrium with the interstitial fluid, were 25 times higher than in the plasma VEGF levels in the granuloma fluid continued to increase for up to weeks, reflecting the high dependence of the wound tissue on ambient VEGF levels After injection of the VEGF receptor antagonist in the fully formed granuloma, the preexisting blood vessels in the middle layer regressed and underwent apoptosis, accompanied by expansion of the extracellular matrix (predominately collagen I) into areas normally devoid of matrix We conclude that wound tissue is sensitive to ambient VEGF levels, and that a low VEGF condition resulting from VEGF receptor antagonism can disrupt the healing of wound tissue (J Lab Clin Med 2005;145: 194 –203) Abbreviations: ED50 ϭ median effective dose; ELISA ϭ enzyme-linked immunosorbent assay; FITC ϭ fluorescein isothiocyanate; PAS ϭ periodic acid-Schiff; PCNA ϭ proliferating cell nuclear antigen; TUNEL ϭ terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nickend labeling; VEGF ϭ vascular endothelial growth factor; vWF ϭ von Willebrand factor V ascular endothelial growth factor (VEGF), a powerful endothelial cell growth factor, promotes the development of new blood vessels during fetal development, as shown by its critical pres- ence in fetal tissues and by experiments in which its deletion proved to be lethal.1–3 VEGF is also critical in inducing new blood vessels in wound-healing tissue.4,5 That even blood vessels in the normal adult tissue From the Division of Nephrology, Cook County Hospital, Chicago, IL; Section of Nephrology, University of Illinois and the Chicago Veterans Administration Medical Center, Chicago, IL; and the Hektoen Institute for Medical Research, Chicago, IL Submitted for publication September 9, 2004; accepted for publication February 15, 2005 Reprint requests: Ashok K Singh, PhD, Hektoen Institute for Medical Research, 2100 West Harrison St., Chicago, IL 60612; e-mail: singhashok@comcast.net 0022-2143/$ – see front matter © 2005 Mosby, Inc All rights reserved doi:10.1016/j.lab.2005.02.007 Supported by the National Kidney Foundation of Illinois, the Juvenile Diabetes Foundation International (grant JDA 1-2000-241), and the U.S Department of Veteran Affairs Merit Review Program (J.A.L.A and R.B.R.) 194 J Lab Clin Med Volume 145, Number continue to be responsive to VEGF is indicated by experiments in which VEGF injection into the cornea or skeletal muscle of adult mice resulted in sprouting of new blood vessels.6,7 It is less clear, however, whether low tissue VEGF, caused either by disease or by systemic administration of VEGF antagonists for treatment of solid cancers,8,9 can cause abnormalities in preexisting blood vessels, especially of wound tissue that depends on high local levels of VEGF for growth and sustenance To address this issue, we induced a foreign-body granuloma by implanting a perforated polyvinyl tube in the subcutaneous tissue of rat By weeks, the implanted tube was completely covered by new tissue and blood vessels (granuloma) This tissue growth was dependent on high VEGF levels, as suggested by high levels of VEGF found in the granuloma fluid that accumulated inside the tube Thus, although induced by an external stimulus, a foreign-body granuloma represents new tissue growth, having similarities to wound healing, and thus is a suitable model for examining the effects of VEGF antagonism on a healing tissue We first measured VEGF level in the granuloma fluid during the development of the granuloma, to ascertain that VEGF level was associated with the formation of new blood vessels and associated tissues To study the role of low VEGF on granuloma blood vessels, we injected an antagonist against the VEGF receptor in the granuloma soon after tube implantation (to study the effects on new blood vessel formation) and, more important, after its complete encapsulation (to study the effects on preexisting blood vessels) The antagonist was injected at a dose that blocked VEGF binding to its receptor locally in the granuloma but was too low to affect systemic VEGF reactivity MATERIALS AND METHODS Construction of the perforated polyvinyl tube for subcutaneous implantation A piece of polyvinyl chloride tubing (length, 20 mm; diameter, mm) (PVC 180; Nalge Nunc, Rochester, NY) was sealed at both open ends by heat application to create an enclosed chamber (inside volume, 0.5 mL) Eight holes (diameter, 0.5 mm) were drilled around the chamber to allow steady diffusion between the tube contents and the surrounding tissue The tubes were stored in 70% alcohol for sterility Before implantation in the rats, the tubes were washed vigorously with sterile saline solution and air-dried Surgical implantation of the polyvinyl tube in rats Animal experiments were conducted under the protocol approved by the institutional animal care and use committee SpragueDawley rats (males, 225–250 g) were anesthetized with a 1:5 mixture of ketamine 50 mg/mL and acepromazine 10 mg/mL at a dose of 0.1 mL/100 g The animals’ backs were shaved and cleaned with alcohol and povidone In each animal, two 1-cm incisions were made on either side of the lumbar region Gudehithlu et al 195 Fig Method of injecting test substances and aspirating granuloma fluid from the polyvinyl tube granuloma in live rats (under restraint) Using blunt dissection, a subcutaneous pocket was made around the incisions, into which the polyvinyl tubes were inserted The incisions were closed with silk sutures, and the animal was allowed to recover and form a granuloma around the tube Injection of VEGF receptor antagonist in the granuloma VEGF receptor antagonist was purchased from R & D Systems, Minneapolis, MN (cat # 471-F-1) The VEGF receptor antagonist is a soluble recombinant mouse VEGF R1 (Flt-1)/Fc chimera that is active in inhibiting the VEGFdependent proliferation of cultured human umbilical vein endothelial cells According to the manufacturer, the median effective dose (ED50) of this activity is 10 –30 ng/mL The VEGF receptor antagonist treatment was initiated following protocols The first protocol was designed to study the effect of the VEGF receptor antagonist on new blood vessel formation in the granuloma; the second protocol was designed to study its effects on preexisting blood vessels In the first protocol, administration of the VEGF receptor antagonist was initiated days after implantation of the plastic tube, and the injections were continued for weeks, after which the granulomas were harvested In the second protocol, administration of the VEGF receptor antagonist was initiated weeks after tube implantation (after complete encapsulation of the tube had occurred), and the injections were continued for weeks, after which the granulomas were harvested Animals were classified as either control or experimental in each protocol Each experimental animal received 0.5 mL sterile saline solution containing 100 ng of the antagonist/ granuloma on the first day and 0.1 mL containing 20 ng of the antagonist/granuloma on subsequent alternate days Controls received similar volumes of saline solution in the granuloma at the same times as the experimental animals The animals were injected under restraint by passing the 25-gauge syringe needle through the skin, the granuloma, and the polyvinyl wall of the tube into the cavity of the granuloma (Fig 1) Estimation of the intragranuloma and blood levels of VEGF receptor antagonist Because of the unavailability of an assay for VEGF receptor antagonist, we estimated the antagonist level by determining the clearance rate of the granuloma fluid This was carried out in a separate group of 196 Gudehithlu et al rats (n ϭ 5) implanted with polyvinyl tubes Two weeks after tube implantation, a trace amount of radioactive (125I) human albumin (prepared by the chloramine-T method10) was injected in the granuloma Samples (25 L) of granuloma fluid and serum were collected at time and every 30 minutes thereafter for hours The samples were precipitated with 10% trichloroacetic acid, and the radioactivity in the precipitates was measured in a gamma counter The data were plotted to obtain a clearance curve for injected (125I) human albumin Determination of VEGF levels in the granuloma fluid VEGF165 levels in the granuloma fluid and serum were determined by a sandwich enzyme-linked immunosorbent assay (ELISA) kit (murine; R&D Systems) The interference because of the presence of VEGF receptor antagonist (present in the granuloma fluid of experimental rats) in the VEGF ELISA was determined by adding increasing amounts of the antagonist to standard solutions of VEGF and comparing the consequent VEGF levels with the VEGF levels in samples not containing the antagonist ELISA found no difference in measured VEGF levels in the antagonist-free and antagonistcontaining solutions, confirming that the copresence of the VEGF receptor antagonist in the samples did not interfere with the VEGF measurements Processing of granulomas Each granuloma was surgically resected and its wet weight recorded About 1/4 of the granuloma was placed in a nonformalin fixative (Histochoice; Amresco, Solon, OH) for histological studies, and the remaining 3/4 was processed for hemoglobin determination Hemoglobin determination in the granulomas for determination of blood content in the granulomas Hemoglobin concentration was determined spectrometrically as a surrogate index of the blood content of the granuloma Each granuloma was placed in ice-cold water containing heparin to promote red cell lysis and avoid blood clotting The tissue was homogenized in a polytron homogenizer The homogenate was centrifuged at 15,000 g for 10 minutes The supernatant was filtered through 0.45- filters to obtain a clear filtrate The filtrate was scanned in a spectrophotometer in the 400 – 800 nm range to measure the typical absorption peaks of hemoglobin at 540 and 575 nm Hemoglobin concentration was determined from the absorption at 575 nm using a standard curve constructed from fresh diluted human blood with known hemoglobin concentration, also read at 575 nm Assuming the hemoglobin content in normal blood to be 15 g/dL, the hemoglobin concentration was converted to L blood/granuloma and L blood/g of tissue Histology and immunocytochemistry Tissues fixed in Histochoice were embedded in paraffin, and 4--thick sections were cut in a microtome Sections were deparaffinized and stained with trichrome and periodic acid-Schiff (PAS) stains for visualizing extracellular matrix and general tissue morphology The nature of the extracellular matrix was further examined by immunostaining, by first incubating with goat anti–type I collagen, goat anti–type III collagen (Southern Biotechnology Associates, Birmingham, AL) and rabbit anti-human fibronectin (Chemicon, Temecula, CA) antibodies, followed by J Lab Clin Med April 2005 washing and reincubating with the appropriate secondary anti-goat IgG-alkaline phosphatase conjugate (or anti-rabbit IgG-alkaline phosphatase antibodies) (Sigma, St Louis, MO) The slides were washed, and the bound alkaline phosphatase was developed with fast-red naphthol (Sigma) Slides were similarly immunostained for VEGF, VEGF-R1 and -R2, and proliferating cell nuclear antigen (PCNA) by first incubating with the primary antibody [rabbit anti-mouse VEGF (NeoMarkers, Fremont, CA), goat anti-mouse VEGF-R1 (Flt-1) and R2 (Flk-1) (R&D Systems), or antimouse PCNA (clone 19A2; Biogenex, San Ramon, CA)], followed by an appropriate second antibody, fluorescein isothiocyanate (FITC)-labeled in cases of VEGF, VEGF-R1, and VEGF-R2 and peroxidase-labeled in cases of PCNA FITClabeled slides were examined and photographed under epifluorescence (Nikon, New York, NY), and peroxidase labeled slides were developed with diaminobenzidine-H2O2 (brown color) and examined under a light microscope (Nikon) Microvessel density was studied by staining sections for type IV collagen Sections were first incubated with primary goat anti–type IV collagen (Southern Biotechnology Associates), followed by washing and reincubation with the secondary anti-goat IgG-alkaline phosphatase conjugate (Sigma) The slides were washed, and the bound alkaline phosphatase was developed with fast-red naphthol (Sigma) The organization of the microvessels (only those with Ͼ 3– endothelial cells) in the granuloma tissue was examined by double immunostaining for ␣-smooth muscle actin, to highlight pericytes, and von Willebrand factor (vWF; also known as factor VIII), to highlight endothelial cells (Smaller microvessels containing 1–3 endothelial cells could not be analyzed in this manner, because they were not stainable by factor VIII antibody.) The tissue was permeabilized by pretreatment with trypsin (10 g/mL; Sigma) at room temperature for 10 minutes, then reacted with monoclonal anti–␣smooth muscle actin (Sigma) antibody, followed by antimouse IgG conjugated to alkaline phosphatase The enzyme reaction was developed with fast-red naphthol (red color) Subsequently, the tissue was reacted with rabbit anti-vWF antibody (Sigma) and anti-rabbit IgG antibody conjugated to horseradish peroxidase and developed with diaminobenzidine-H2O2 (brown color) The tissue was counterstained with hematoxylin (blue color) Apoptosis in the nuclei of microvessels was visualized by double-staining for type IV collagen (to highlight microvessels) and DNA fragments (to highlight apoptotic nuclei in microvessels) The sections were first immunostained for type IV collagen (red color), as described earlier Subsequently, apoptosis was visualized by staining for DNA fragments by the reaction of tissue with dT terminal transferase in the presence of biotinylated DNA bases, followed by streptavidin-peroxidase and diaminobenzidine-H2O2 reaction (brown color) using reagents supplied by Travegen (Gaithersberg, MD) The sections were finally counterstained with hematoxylin (blue color) Quantification of immunocytochemical changes Representative areas of type I collagen, type III collagen, fibronectin, and VEGF-stained slides from several granulomas J Lab Clin Med Volume 145, Number were photographed (n Ͼ 10) and analyzed for intensity of appropriate color (red for fibronectin and types I and III collagens, green for VEGF) using the Image J software (JAVA imaging software inspired by the National Institutes of Health and available free at http://rsb.info.nih.gov) The percentage changes between the control and VEGF-receptor antagonist groups reported in the results and figure legends were derived from the arbitrary intensity units on a scale of (zero intensity)–255 (maximum intensity) P values of changes (Student t test) are indicated in the text Microvessels stained with type IV collagen were quantified by counting the number of vessels per high-power field Apoptosis was quantified by randomly examining 20 microvessels in the medial layer of each granuloma (n ϭ from each group), and approximately 300 healthy and apoptotic endothelial nuclei were scored (brown-stained) to arrive at a statistical evaluation of the extent of apoptosis in the microvessels PCNA-positive cells were counted from several (n ϭ 20) high-power fields of control and antagonist-treated granulomas (n ϭ of each group) All quantitative data are expressed as mean Ϯ standard error, and statistical comparisons were made using the Student t test RESULTS Induction of the subcutaneous foreign body granuloma In preliminary experiments, we studied the for- mation of granuloma after the surgical implantation of a polyvinyl tube in the subcutaneous tissue of rat After implantation, animals were killed at week and 2, 3, and weeks, and granulomas were extracted The granulomas were examined, wet-weighed, homogenized to determine hemoglobin (or blood) content, and processed for histology At week, the polyvinyl tube was completely encapsulated with a thin layer of tissue, which was supplied by at least large blood vessels extending from the surrounding tissue By weeks, the granuloma appeared thicker than at week Also by weeks (not seen at week), the holes around the plastic tube were plugged by tissue, which appeared to be highly vascular The granulomas increased in weight until weeks, after which the weights remained constant for up to weeks The granuloma was organized in histologically demarcated layers (Fig 2) The inner layer in contact with the polyvinyl tube was approximately 200 thick and consisted of mononuclear epithelioid cells and fibroblasts, had little extracellular matrix, and was devoid of blood vessels The medial layer was thicker and consisted of tightly packed fibroblastic cells and extracellular matrix fibers (stained blue by trichrome) arranged parallel to the length of the granuloma This layer was also richly endowed with fine blood vessels The outer layer was more amorphous, with loose connective tissue and large blood vessels One-week-old Gudehithlu et al 197 Fig Histology of the foreign-body granuloma formed weeks after the subcutaneous implantation of a polyvinyl tube in the rat The granuloma appeared to be organized in histological demarcated layers The inner layer (ϳ 200 thick) contained mostly mononuclear epithelioid cells and fibroblasts The thicker medial layer consisted of tightly packed fibroblastic cells, extracellular matrix (blue stain), and microvessels (arrows) The outer layer was made up of loose connective tissue and large blood vessels (Trichrome stain; original magnification ϫ 50.) granulomas were as histologically organized as the older granulomas, except that each of the histological layers was thinner than in the older granulomas The organization of the microvessels was examined immunohistochemically by double-staining with anti– ␣-smooth muscle actin to visualize the pericytes and with anti-vWF to highlight the endothelial lining of the microvessels It was observed that 100% of the microvessels containing at least 3– endothelial cells also had pericytes surrounding the endothelial layer, indicating that the microvasculature in the granuloma was similar to that seen in normal adult state (not shown) The histological pattern including microvessels observed weeks after implantation of the polyvinyl tube remained unchanged during the next weeks, suggesting stabilization of the healing process Granuloma fluid Within days after implantation of the polyvinyl tube, an accumulation of a clear fluid inside the tube was seen, which was in a dynamic equilibrium with the interstitial fluid/plasma, being replaced by 50% within 90 minutes, as determined by radiotracer studies (Fig 3) It resembled plasma (or serum) This was confirmed by the similarity of its electrophoretic profile (sodium dodecyl sulfate–polyacrylamide gel electrophoresis) to that of rat plasma and serum (not shown) Fig shows the VEGF concentration in the granuloma fluid of control animals between 198 Gudehithlu et al Fig Clearance of 125I-albumin from the polyvinyl tube granuloma fluid As indicated, a half-life of approximately 90 minutes was calculated for the clearance of 125I-albumin J Lab Clin Med April 2005 nyl tube implantation (first protocol) resulted in incomplete encapsulation of the tube In contrast, control granulomas exhibited well-formed, complete encapsulation These results were reflected in significantly lower wet weights and blood content of the experimental granulomas, confirming the antiangiogenic activity of the VEGF receptor antagonist on new blood vessel formation (Table I) Note that even though the blood content/granuloma decreased in the antagonist-injected granulomas, the blood content/g of tissue remained the same (Table I) In the next set of experiments, we tested the effect of the VEGF receptor antagonist in formed granulomas by injecting the antagonist starting weeks after implantation of the polyvinyl tube (second protocol) Examination of these granulomas revealed well-formed granulomas that were more fibrous, less elastic, and more leathery to the touch than the control tissues Despite these qualitative differences, the wet weights and blood content of the VEGF receptor antagonist–injected granulomas were similar to those of the control granulomas (Table I) However, important histological changes were seen in the microvessels of the granulomas (see below) Because we were interested primarily in examining the effects of the VEGF receptor antagonist in preexisting blood vessels, we carried out further studies only in the granulomas from the second protocol experiment Estimation of the intragranuloma and blood levels of VEGF receptor antagonist We found the half-life of the Fig Comparison of VEGF levels in the granuloma fluid from control and VEGF receptor antagonist–injected animals at different times before and after injection of the antagonist Each bar represents a mean of samples *Denotes statistical difference at P Ͻ 05 compared with control sample at the same time VEGF levels were comparable between controls and experimental until weeks (ie, before injection of VEGF receptor antagonist) After injection of the receptor antagonist, VEGF levels decreased in the experimental granulomas compared with controls, suggesting inhibition of the VEGF production in the granuloma by the antagonist VEGF levels in serum were Ͻ 0.1 ng/mL at all times (not shown) weeks and VEGF levels in the serum of control animals remained below 0.1 ng/mL at all times (data not shown) The VEGF levels in the granuloma fluid at week were 25 times higher than that in the serum and continued to increase to 50 times higher than that in the serum by the third week, after which they remained unchanged Effect of injecting VEGF receptor antagonist during granuloma formation Initiating the VEGF-receptor an- tagonist injections in the granuloma soon after polyvi- granuloma fluid to be 90 minutes (Fig 3) Given the initial concentration of the antagonist injected in the granuloma of 200 ng/mL (10 ϫ ED50) on the first day and 40 ng/mL on subsequent days (2 ϫ ED50) (see Materials and Methods), one can see that these concentrations will decline to below ED50 within a matter of a few hours after injection as the antagonist diffuses out of the granuloma into the blood Assuming rapid clearance of the antagonist from the plasma, one could then estimate that the drug would be at its highest concentration in the blood soon after the intragranuloma injection The blood concentration could be estimated to be 300-fold less than the intragranuloma concentration, because the whole body fluid content of the animal (150 mL for a 250-g rat) is 300 times that of the intragranuloma volume (0.5 mL) Therefore, over the period of weeks of antagonist injections, the antagonist must have reached the 10-fold ED50 level only once for a few hours after the first injection and, subsequently, the 2-fold ED50 for a few hours every other day in the granuloma For most of the remaining time, the antagonist must have been less than the ED50 level in the granuloma The blood levels must be less than the ED50 levels at all times J Lab Clin Med Volume 145, Number Gudehithlu et al 199 Table I Wet weights and blood content of granulomas injected with VEGF receptor antagonist Wet weight (g) Blood content (L/granuloma) VEGF receptor antagonist injected days after tube implantation (first protocol) Control (n ϭ 4) 0.52 Ϯ 0.02 Antagonist (n ϭ 4) 0.31 Ϯ 0.01* VEGF receptor antagonist injected 14 days after tube implantation (second protocol) Control (n ϭ 8) 0.51 Ϯ 0.08 Antagonist (n ϭ 8) 0.62 Ϯ 0.05 Blood content (L/g of tissue) 41.9 Ϯ 3.1 22.0 Ϯ 2.2* 80.6 Ϯ 6.6 70.9 Ϯ 8.6 43.0 Ϯ 6.7 50.3 Ϯ 5.6 84.3 Ϯ 6.7 81.2 Ϯ 9.1 Granulomas were harvested for wet weights and blood content (by hemoglobin determination) at the end of the VEGF receptor antagonist injections (17 days after tube implantation in the first protocol and 28 days after tube implantation in the second protocol) *P Ͻ 05 compared with its respective control Fig Immunofluorescent localization of VEGF in control and experimental (treated with the VEGF receptor antagonist) granulomas In the control granuloma, VEGF was localized predominantly to the inner layer, which decreased significantly after treatment with the VEGF receptor antagonist (Control VEGF intensity ϭ 91.1 Ϯ 6.3 versus antagonist-treated ϭ 51.0 Ϯ 4.6 green intensity units/unit area; n ϭ 12; P Ͻ 05; original magnification ϫ 100.) VEGF in the granuloma fluid and tissue after VEGF receptor antagonist injection Fig shows VEGF concen- trations in the granuloma fluid from weeks 1– in the experiment where the VEGF receptor antagonist was injected after weeks of tube implantation (second protocol) For the first weeks of granuloma formation and before VEGF receptor antagonist injection, VEGF levels were similar in the control and experimental granulomas, as expected After the start of the VEGF receptor antagonist injection, VEGF levels in the experimental granulomas became significantly lower than in the control granulomas injected with saline solution The lower VEGF level was not because of interference in the immunoassay due to the copresence of the VEGF receptor antagonist in the granuloma fluid, because the antagonist per se did not interfere in the assay for VEGF (see Materials and Methods) By immunocytochemical staining, VEGF was localized predominantly to the inner layer of the granuloma, and consistent with the decreased fluid levels of VEGF, a significant decrease in tissue-associated VEGF was also observed in this area (Fig 5) Extracellular matrix changes in the granuloma after treatment with VEGF receptor antagonist Granulomas from VEGF receptor antagonist–injected animals (second protocol) were histologically distinguishable from the control granulomas by the expansion of the extracellular matrix (PAS-positive material) from the medial layer to the inner layer, resulting in disorganization of the inner layer Moreover, the extracellular matrix in these granulomas appeared to lose its orderly orientation compared with that in controls (Fig 6) The nature of the PAS-positive material was further examined by immunostaining of the granuloma sections for collagen I, collagen III, and fibronectin The medial and outer layers of the control granulomas were filled predominately with collagen I, which increased in amount and expanded into the inner layer after injection of the VEGF receptor antagonist (Fig 7A, B) (control ϭ 35.2 Ϯ 3.3 vs antagonist treated ϭ 55.1 Ϯ 2.8 red intensity units/unit area; n ϭ 12; P Ͻ 05) Small amounts of fibronectin were diffusely distributed in the inner and medial layers of the control granulomas, and these appeared to decrease slightly after VEGF receptor antagonist injection (data not shown) Collagen III was present in minor amounts throughout the granuloma and did not appear to change after the antagonist treatment (data not shown) 200 Gudehithlu et al J Lab Clin Med April 2005 Fig Histological changes in control and experimental (rats treated with the VEGF receptor antagonist) granulomas Experimental granulomas were distinguishable from the control granulomas by the expansion of the extracellular matrix (pink) from the medial layer to the inner layer Furthermore, the extracellular matrix in the experimental granulomas appeared disorganized compared with that of the control granulomas (PAS stain; original magnification ϫ 150.) Microvascular changes in granuloma injected with VEGF receptor antagonist The microvasculature of the granulomas was examined by immunostaining for type IV collagen, which highlighted the endothelial basement membrane of the microvessels (Fig 7C and D; low power) Whereas the control tissue exhibited a high density of microvessels in the medial layer (53 Ϯ capillaries/high-power field; n ϭ 12), the experimental tissue demonstrated significantly fewer and weaker microvessels in this layer (23 Ϯ capillaries/high-power field; n ϭ 12; P Ͻ 05), suggestive of degenerating vessels (Fig 7E and F; high power) Apoptosis in the microvessels of the granuloma The granuloma sections were double-stained for DNA fragmentation and type IV collagen to examine the endothelial cells for signs of apoptosis (Fig 8) There was a statistically significant 8-fold increase in the degree of apoptosis in the microvessels of the granulomas treated with VEGF receptor antagonist compared with controls Cell proliferation in the microvessels of the granuloma To examine cell proliferation in the granuloma, tissues were immunostained for PCNA, a marker of cell proliferation Most of the proliferative activity was present in the medial layer, which is rich in microvessels Compared with controls, the VEGF receptor antagonist–injected granulomas demonstrated 70% fewer PCNA positive cells in the medial layer, suggesting that the antagonist exerted a significant antiproliferative effect on the microvessels (Fig 9) DISCUSSION When a perforated polyvinyl tube was implanted subcutaneously into rats, it was surrounded by a gran- uloma amply sustained by a new vascular network derived from surrounding blood vessels The histological pattern observed in this granuloma appeared to be very similar to other foreign-body granulomas described in the literature.4,10 –13 The polyvinyl tube granuloma model was preferred because it created an enclosed wound tissue that allowed us to inject the drug (VEGF receptor antagonist) directly into the confined wound tissue and to conveniently harvest the solid wound tissue for histological evaluation After encapsulation (1 week after implantation), the implanted tube collected a plasma-like fluid that could be sampled for measuring VEGF levels The origin of the granuloma fluid is not completely understood Based on the immunocytochemical studies, it appears that VEGF was secreted primarily by the inner layer of the granuloma, then diffused and accumulated in the polyvinyl tube Being derived from a rapidly growing wound tissue, this fluid is rich in VEGF and other growth factors The fluid is in equilibrium with the interstitial fluid (fluid bathing the granuloma tissue), and as such reflects the fluid environment of the granuloma tissue Clearance experiments with radioactive albumin showed that the fluid was replaced by 50% in 90 minutes (half-life) Injecting the VEGF receptor antagonist into the granuloma soon after implantation of the polyvinyl tube (after days) prevented the formation of a complete granuloma, as expected based on the critical role of VEGF in new blood vessel formation in wound healing.4,5 This resulted in decreased weight and lower blood content of the granulomas This experiment confirmed the bioactivity of the VEGF receptor antagonist used in this study All subsequent J Lab Clin Med Volume 145, Number Fig Distribution of extracellular matrix (type I collagen) and microvessels (type IV collagen) by immunostaining (red) in control and experimental (rats treated with the VEGF receptor antagonist) granulomas A and B, Control granulomas showed well-organized type I collagen mostly in the extracellular regions of the medial layer (A) In granulomas injected with the VEGF-receptor antagonist, type I collagen appeared less well organized, increased in the medial layer, and extended into the inner layer (B) (Uncounterstained; original magnification ϫ 50.) C and D, Architecture of granuloma blood vessels visualized by type IV collagen immunostaining (red) at low power Well-formed robust blood vessels and microvessels appeared in the medial layer of control granulomas (C) In granulomas treated with the VEGF receptor antagonist, the blood vessels were fewer in number and appeared to be degenerating (D) (Uncounterstained; original magnification ϫ 50.) E and F, Microvascular changes in the medial layer of granulomas visualized by type IV collagen immunostaining (red) at high power Whereas well-formed capillaries are visible in the control granulomas (E), there appear to be fewer and weaker microvessels in the medial layer of granulomas injected with the VEGF receptor antagonist, suggestive of degenerating vessels (F) [Uncounterstained (background color adjusted to highlight red); original magnification ϫ 200.] studies were performed in pre-formed granulomas to allow us to study the effect of VEGF receptor antagonist on preexisting blood vessels Injecting the VEGF receptor antagonist after complete formation of the granuloma had a dramatic effect on the already-formed blood vessels and matrix The blood vessels in the middle layer appeared to have atrophied, and there was a concomitant expansion of the extracellular matrix, consisting of predominantly type I collagen (interstitial collagen), into the inner layer that was in contact with the polyvinyl tube This resulted in disruption of the inner layer, which now looked craggy and disorganized The extracellular matrix itself, originally organized in parallel to the length Gudehithlu et al 201 Fig Representative photomicrographs showing endothelial nuclei of a microvessel undergoing apoptosis (stained for TUNEL) in the medial layer of control (A) and experimental (rats treated with the VEGF receptor antagonist) (B) granulomas Normal nuclei are stained blue, and apoptotic nuclei are stained brown (Blue counterstain; original magnification ϫ 600.) (C) Quantitative comparison of apoptotic nuclei found in the microvessels of control and VEGF receptor antagonist–injected granulomas There was an 8-fold increase in the number of apoptotic endothelial nuclei in the VEGF receptor antagonist group compared with controls *P Ͻ 05 compared with controls (n ϭ 300 nuclei) of the granuloma, lost its directional arrangement so that the tissue was less elastic and more leathery to the touch Neither the weight of the granuloma nor its blood content (as measured by tissue hemoglobin) changed, possibly because the changes were limited only to the fine capillaries of the middle layer In these capillaries, the blockade of VEGF caused inhibition and apoptosis of the endothelial cells, as detected by decreased PCNA staining and increased terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL) reaction It should be noted that although in a disease state (like cancer), any VEGF antagonist would have to be delivered systemically, we preferred to apply the antagonist directly in the wound, to avoid indirect and confounding effects of the antagonist on the systemic vascular tissue and other vital organs The VEGF-receptor antagonist that we used in the study is a soluble receptor that blocks VEGF action by preventing VEGF from binding to its natural receptor on the endothelial cell Unexpectedly, the antagonist treatment reduced VEGF levels in the granuloma fluid The reason for this decrease in VEGF is not known, but could be related to accelerated catabolism of VEGF or excessive turnover of the granuloma fluid Further studies are needed to investigate these possibilities Because an assay for VEGF receptor antagonist was not available, we estimated the antagonist level in granuloma and blood by determining the clearance rate of 202 Gudehithlu et al J Lab Clin Med April 2005 Fig A and B, Representative photomicrographs showing cells immunostained (brown) for PCNA in the medial layer of control and experimental (rats treated with the VEGF receptor antagonist) granulomas (Blue counterstain blue; original magnification ϫ 200.) C, Compared with controls, VEGF receptor antagonist treated granulomas displayed 70% fewer PCNA-positive cells in the medial layer, suggesting that the antagonist exerted a significant antiproliferative effect on the microvessels *P Ͻ 05 compared with controls the granuloma fluid using radioactive albumin as a tracer Based on the experimentally determined clearance rate of the granuloma fluid, we showed that the intragranuloma antagonist dose used produced antagonist levels in the pharmacological range in the granulomas and only trace levels in blood Low tissue VEGF levels have been found in diabetic nephropathy,14,15 other chronic glomerular diseases,16 experimental models of reduced renal mass,17 mesangiolysis (induced by anti-Thy antibody),18 diabetic peripheral vascular damage,19,20 and hyperoxia-induced lung endothelial damage.21 The harmful effects of such low VEGF levels has been reversed by the systemic administration of VEGF or by gene transfer, an effect due largely to new blood vessel formation but not necessarily exerted on the preexisting blood vessels and the surrounding matrix.19,20,22 Although a strong association has been documented between disease and low VEGF, the specific pathological effects exerted on the preexisting blood vessels by a low VEGF environment have not yet been clarified Our results suggest that in a healing tissue (and possibly even in the normal adult tissue), whereas the major blood vessels seem to be refractory to low VEGF, the microvessels and the surrounding matrix are sensitive to the ambient VEGF levels Low VEGF levels caused attrition and disorganization of the pre-existing microvessels We also showed that the vessel regression after administration of the VEGF receptor antagonist was due to an antiproliferative effect on the endothelial cells, resulting in cell death by apoptosis So far, a direct effect of low VEGF levels on preexisting blood vessels has been reported only in experiments conducted in tumors In such experiments, decreasing VEGF caused selective endothelial cell death in blood vessels that lacked pericytes but spared those blood vessels that had pericytic support.23,24 Recently, however, Maynard et al25 found that in pregnant mice, a soluble VEGF receptor (sFlt1) released from the uterus reduced systemic VEGF levels and caused glomerular endothelial cell changes similar to those seen in the renal endotheliosis of preeclampsia These results would also indicate that VEGF continues to be an important endothelial growth factor in the adult life and that its reduction can result in endothelial cell death in existing blood vessels We conclude that VEGF, a classical angiogenic factor critical during fetal development, also regulates the maturation of wound tissue Low VEGF levels induced by disease or therapeutic manipulation may cause damaging effects in a healing tissue, and possibly also in normal adult vasculature The authors thank Linda Wanna and P Sethupathi for help with the animal surgery and Bhakti Patel, Sreya Patri, Valentina Svoren, and Lev Rappaport for help with the immunocytochemistry J Lab Clin Med Volume 145, Number Gudehithlu et al REFERENCES Carmeliet P, Ferreira V, Brier G Abnormal blood vessel development and lethality in embryos lacking a single vascular endothelial growth factor allele Nature 1996;380:435–9 Ferrara N, Carver MK, Chen D Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene Nature 1996;380:439 – 42 Eremina V, Sood M, Haigh J, Nagy A, Lajoie G, Ferrara N, et al Glomerular-specific alterations of VEGF-A expression lead to distinct congenital and acquired renal diseases J Clin Invest 2003;111:707–16 Howdieshell TR, Riegner C, Gupta V, Callaway D, Grembowicz K, Sathyanarayana MD, et al Normoxic wound fluid contains high levels of vascular endothelial growth factor Ann Surg 1998;228:707–15 Kampfer H, Pfeilschifter J, Frank S Expressional regulation of angiopoietin-1 and -2 and the Tie-1 and -2 receptor tyrosine kinases during cutaneous wound healing: a comparative study of normal and impaired repair Lab Invest 2001;81:361–73 Asahara T, Chen D, Takahashi T, Fujikawa K, Kearney M, Magner M, et al Tie-2 receptor ligands angiopoietin-1 and angiopoietin-2 modulate VEGF induced postnatal neovascularization Circ Res 1998;83:233– 40 Springer ML, Chen AS, Kraft PE, Bednarski M, Blau HM VEGF gene delivery to muscle: potential role for vasculogenesis in adults Mol Cell 1998;2:549 –58 Hasan J, Kayson GC VEGF antagonists Expert Opin Biol Ther 2001;1:703–18 Folkman J Models of anti-cancer therapy Angiogenesis inhibitors: a new class of drugs Cancer Biol Ther 2003;2(Suppl 1):S127–33 10 McConahey PJ, Dixon FJ A method of trace iodination of proteins for in vivo studies Int Arch Appl Immunol 1996;29: 185–9 11 Teixeira AS, Andrade SP Glucose-induced inhibition of angiogenesis in the rat sponge granuloma is prevented by aminoguanidine Life Sci 1999;64:655– 62 12 Li Y, Metori K, Koike K, Kita F, Che QM, Sato T, et al Granuloma maturation in the rat is advanced by the oral administration of Eucommia ulmoides OLIVER leaf Biol Pharm Bull 2000;23:60 –5 13 Rodriguez MC, Saiz MP, Carbonell T, Mitjavila MT Local and systemic responses to iron dextran injected in a granuloma pouch in the rat J Lab Clin Med 1999;134:42– 14 Bailey E, Bottomley MJ, Westwell S, Pringle JH, Fumess PN, Feehally J, et al Vascular endothelial growth mRNA expression in minimal change, membranous, and diabetic nephropathy dem- 15 16 17 18 19 20 21 22 23 24 25 203 onstrated by non-isotopic in-situ hybridization J Clin Pathol 1999;52:735– Singh AK, Gudehithlu KP, Pegoraro AA, Singh GK, Basheeruddin K, Robey RB, et al Vascular factors altered in glucosetreated mesangial cells and diabetic glomeruli Changes in vascular factors impair endothelial cell growth and matrix Lab Invest 2004;84:597– 606 Shulman K, Rosen S, Tognazzi K, Manseau EJ, Brown LF Expression of vascular permeability factor (VPF/VEGF) is altered in many glomerular diseases J Am Soc Nephrol 1996;7: 661– Kang D-H, Joly AH, Oh S-W, Hugo C, Kerjaschi D, Gordon KL Impaired angiogenesis in the remnant kidney model: I Potential role of vascular endothelial growth factor and thrombospondin-1 J Am Soc Nephrol 2001;12:1434 –7 Wada Y, Morioka T, Oyanagi-Tanaka Y, Yao J, Suzuki Y, Gejyo F, et al Impairment of vascular regeneration precedes progressive glomerulosclerosis in anti-Thy glomerulonephritis Kidney Int 2002;61:432– 43 Rivard A, Silver M, Chen D, Kearney M, Magner M, Annex B, et al Rescue of diabetes-related impairment of angiogenesis by intramuscular gene therapy with adeno-VEGF Am J Pathol 1999;154:355– 63 Schratzberger P, Walter DH, Rittig K, Bahlmann FH, Pola R, Curry C, et al Reversal of experimental diabetic neuropathy by VEGF gene transfer J Clin Invest 2001;107:1083–92 Klekamp JG, Jarzecka K, Perkett EA Exposure to hyperoxia decreases the expression of vascular endothelial growth factor and its receptors in adult rat lungs Am J Pathol 1999;154:823–31 Kang D-K, Hughes J, Mazzali M, Schreiner GF, Johnson RJ Impaired angiogenesis in the remnant kidney model: II Vascular endothelial growth factor administration reduces renal fibrosis and stabilizes renal function J Am Soc Nephrol 2001;12:1448 –57 Benjamin L, Keshet E Conditional switching of vascular endothelial growth factor (VEGF) expression in tumors: induction of endothelial cell shedding and regression of hemangioblastomalike vessels by VEGF withdrawal Proc Natl Acad Sci U S A 1997;94:8761– Benjamin L, Golijanin D, Itin A, Pode D, Keshet E Selective ablation of immature blood vessels in established human tumors follows vascular endothelial growth factor withdrawal J Clin Invest 1999;103:159 – 65 Maynard SE, Min J-Y, Merchan J, Lim KH, Li J, Mondal S, et al Excess placental soluble fms-like tyrosine kinase (sFlt 1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia J Clin Invest 2003;111:649 –58 ... local levels of VEGF for growth and sustenance To address this issue, we induced a foreign-body granuloma by implanting a perforated polyvinyl tube in the subcutaneous tissue of rat By weeks, the... analyzed for intensity of appropriate color (red for fibronectin and types I and III collagens, green for VEGF) using the Image J software (JAVA imaging software inspired by the National Institutes of. .. Potential role of vascular endothelial growth factor and thrombospondin-1 J Am Soc Nephrol 2001;12:1434 –7 Wada Y, Morioka T, Oyanagi-Tanaka Y, Yao J, Suzuki Y, Gejyo F, et al Impairment of vascular