JOURNAL OF PATHOLOGY, VOL. 180: 90-94 (1 996)

THE ROLE OF VASCULAR ENDOTHELIAL GROWTH FACTOR IN A MURINE CHRONIC GRANULOMATOUS

TISSUE AIR POUCH MODEL OF ANGIOGENESIS

IAN APPLETON, NICOLA J . BROWN, DEAN WILLIS, PAUL R. COLVILLE-NASH, CHANDON ALAM, JOANNE R. BROWN AND DEREK A. WILLOUGHBY

Department of Experimental Pathology, St . Bartholomew's Hospital Medical College, Charterhouse Square, London, ECI A4 6BQ, U. K.

SUMMARY Chronic granulomatous inflammation may be considered an angiogenic-dependent process. Recently it has been demonstrated that

vascular endothelial growth factor (VEGF) or vascular permeability factor is essential for tumour angiogenesis. Its role in inflammation- mediated angiogenesis has yet to be determined. In this study, the murine chronic granulomatous air pouch model was used to investigate the role of VEGF in angiogenesis. Animals were treated twice weekly with 1Opg per animal of neutralizing antibody to rh VEGF and the vascularity and granuloma dry weight were assessed after 7 days. This resulted in significant suppression of both angiogenesis and granuloma dry weight. Western blot analysis demonstrated the presence of VEGF; the levels of protein paralleled the angiogenic response. These results demonstrate for the first time that VEGF may be an important regulator of angiogenesis in inflammation.

KEY WORDS-inflammation; angiogenesis; VEGF

INTRODUCTION

Increased vascular permeability is an important aspect of angiogenesis, or the growth of new blood vessels, a process crucial to a number of pathophysiological processes, including tumour growth, wound healing, and rheumatoid arthritis (RA). A number of cytokines have been implicated in this process. Vascular endothelial growth factor (VEGF), vascular permeability factor (VPF), or vasculotropin is a family of dimeric glyco- proteins, the structures of which are similar to platelet- derived growth factor. It is produced by aortic smooth muscle cells' and macrophages2.

It is unique among cytokines in that its mitogenic effects are restricted to endothelial cells,3 it is angiogenic in v ~ v o , ~ . ~ and is a potent vascular permeability factor. This and other observations have led to the suggestion that it is the tumour angiogenic f a ~ t o r . ~ Although increased levels of VEGF are observed in RA synovial fluid,6 a direct involvement of VEGF in inflammation- mediated angiogenesis has yet to be established.

We have previously demonstrated that in a murine model of chronic granulomatous inflammation, angio- genesis is maximal at day 7.7 In this study, therefore, we have used this model to examine the effects of neutral- izing antibody to VEGF on inflammation-induced angiogenesis.

MATERIALS AND METHODS Induction of the muvine air pouch

Female Tuck original mice (30 k 2 g), n= 15 per group, were injected with 3 ml of air into the dorsal

Addressee for correspondence: Dr Ian Appleton, The School of Biological Sciences, Rm 3.138, The Stopford Building, Manchester University, Oxford Road, Manchester, M13 9PT, U.K.

CCC 0022-3417/96/090090-05 $3 1996 by John Wiley & Sons, Ltd.

subcutaneous tissue. Twenty four hours later, 0.5 ml of 0.1 per cent v/v croton oil in Freund's complete adjuvant was injected into the air pouch.7 Animals were dosed twice weekly, intraperitoneally (i.p.), with 10 p g of poly- clonal goat anti-human VEGF (affinity purified IgG; Research and Development Systems Ltd., Abingdon, U.K.), 1Opg of control IgG (Sigma, Poole, Dorset, U.K.), or phosphate-buffered saline (PBS). The pouch was allowed to develop for 7 days.

Assessment of angiogenesis

At the end of the dosing period, the vascular index was determined by a modification of the method accord- ing to Kimura et al.* Animals were anaesthetized with 0.5ml of hypnorm/hypnovel i.p. (1 part hypnorm'O, 1 part hypnovel(@, 2 parts diluted water). One millilitre of carmine solution (10 per cent carmine red in 5 per cent gelatin solution) at 37°C was injected intravenously (i.v.), and the animals were refrigerated to solidify the gelatin, thus forming a vascular cast. The pouch was dissected and the dry weight determined. The tissue was then digested (330 mg 1- N-acetyl cysteine, 12 units ml- papain, I mM EDTA in 0.05 M PBS, pH 7.0) at 56°C. The dye was solubilized by the addition of 5~ NaOH and the amount of carmine present was assessed spectrophotometrically at a wavelength of 490 nm. Results are expressed as the ratio of total carmine content b g ) per mg of tissue dry weight.

Whole aiv pouch cleaving

At the end of 7 days, the pouch was injected with carmine as above and dissected. Samples were then dehydrated in ascending concentrations of alcohol and cleared in cedar wood oil for 2 weeks.

Accepted I2 Fehruury I996

VASCULAR ENDOTHELIAL GROWTH FACTOR AND INFLAMMATION 91

a 0 PBScontrol 70

60

w 50

f' 40 M

1z] Igc: control

anti-VEGF E

.3

0 30

4 10 h 20

0

b

.. I

.3 c)

0 c

T

** -

Fig. I-The effects of neutralizing antibody to VEGF on granuloma dry weight (a) and the ratio of dry weight per mg carmine content (b). Results are expressed as the mean =k SE mean for n= 15 animals per group. Statistical significance was tested by the Mann-Whitney West with P<0,05 considered significant

Immunolocalization of platelet endothelial cell adhesion molecule {PECAM, CD31)

The pouch was dissected and immediately placed in n-hexane precooled in liquid nitrogen to - 70°C. Ten- micrometre cryostat sections were cut through the skin and granulomatous tissue, thaw-mounted on poly-L- lysine-coated slides, and air-dried. Prior to im- munolabelling, the sections were fixed in acetone for 10 min. Endogenous peroxidases were quenched with 0.3 per cent H202 in methanol and sections washed with 0.1 per cent Triton XlOO in PBS. Non-specific binding of IgGs was blocked using normal goat serum 1:50 in 0.1 per cent essentially globulin-free bovine serum albumin. The sections were incubated with a rat anti-mouse CD31 (1 : 100) for 1 h (Pharmingen, U.S.A.), washed, and incu- bated for a further 30 min with biotinylated goat antirat secondary antibody. Following a further 30 min incu- bation with Vectastain ABC horseradish peroxidase (Vector Laboratories, Peterborough, U.K.), the substrate (0.05 per cent 3,3-diaminobenzidine tetrahydrochloride in Tris buffer) was added for the appropriate time period (5-10 min). This resulted in positive immunoreactivity labelling brown. Primary antiserum was replaced with normal serum as a negative control.

Western blot analysis The air pouch was set up as described above and

samples were taken at 6 , 12, and 24 h, and 3, 7, 14, 21, and 28 days. Tissue samples were homogenized in 1 mM phenylmethylsulphonyl fluoride, 70 mg ml - pepstatin

A, and 0.01 per cent leupeptin (w/v) in 5 0 m ~ Tris- HCl, pH 7.4, and boiled for 10 min with gel loading buffer (Tris, 5 0 m ~ ; SDS, 10 per cent; glycerol, 10 per cent; 2-mercapthoethanol, 10 per cent; bromophenol blue, 2 mg/ml) in a ratio of 1:l and centrifuged at 10 000 g for 10 min. Protein concentrations of the superna- tants were determined by the Bradford assay,9 and total protein-equivalents for each sample were separated on 15 per cent SDS-polyacrylamide mini-gels (Biorad, Hemel Hempstead, U.K.) using the Laemmli buffer system and transferred to nitrocellulose membranes (Biorad, Hemel Hempstead, U.K.). Non-specific IgGs were blocked with 5 per cent dried milk protein (0.1 per cent bovine serum albumin) and incubated with a polyclonal rabbit anti-human VEGF, 1 pg ml- (Onco- gene Sciences, U.S.A.). Bands were detected with an amplified alkaline phosphatase kit (Sigma, Poole, Dorset, U.K.). Rainbow marker (Amersham Inter- national, Buckinghamshire, U.K.) and prestained blue (Sigma Chemical Company, Poole, Dorset, U.K.) protein markers were used for molecular weight determinations.

Statistics Results are expressed as the mean f SE mean for

at least n= 10 separate determinations. Statistical sig- nificance was tested using the Mann-Whitney U-test with a P value of less than 0.05 being considered significant.

RESULTS Eflects of anti- VEGF on angiogenesis

Treatment with anti-VEGF resulted in a significant (P<O.OO 1) decrease in granulomatous tissue dry weight (43.07 f 1.32 mg) in comparison with the control IgG (62.5 f 2.85 mg) and control PBS (59.82 f 4.23 mg) (see Fig. la). This decrease in dry weight was also accompanied by a significant (P<O.O 1) suppression of the vascular index. For control PBS-treated, the ratio of carmine (pg) to tissue dry weight (mg) was 0.78 f 0.08; for the control IgG group 0.7 f 0.09 and the anti-human VEGF-treated group 0.42 * 0.07, see Fig. lb.

Whole air pouch cleaving The results after treatment with cedar wood oil are

illustrated in Fig. 2. In (a), a control PBS-treated sample is shown. The carmine is retained within the vasculature. No leakage into the surrounding tissue occurred. In (b), the profile of the IgG-treated group is shown, the appearance being identical to that of the PBS group. The effects of neutralizing antibody to VEGF are illustrated in (c). Note the lack of capillaries, the carmine being retained within the pre-existing blood vessels of the dermis.

Immunolocalization of PECAM Treatment with antibody to VEGF resulted in a sig-

nificant decrease in the number of blood vessels in the

92 1. APPLETON ET AL.

Fig. 2-The effects of antibody to VEGF on blood vessel development. In (a), the control PBS-treated group is illustrated and in (b), the IgG-treated group. Note the large number of tortuous capillaries present in the granulomatous tissue. The anti-VEGF group is shown in (c); virtually no capillaries are seen in the granulomatous tissue. The large vessels are the pre-existing blood vessels in the dermis. x 4

granulomatous tissue (see Fig. 3) . Furthermore, there was a marked decrease in the density of the granulo- matous tissue in the anti-VEGF-treated group. The blood vessels in the skeletal muscle juxtaposed to the granulomatous tissue were unaffected by treatment with antiVEGF.

Western blot analysis of VEGF

The time course of the profile for VEGF is shown in Fig. 4. VEGF protein occurred at approximately 23 kD. In the acute stages of the inflammation, VEGF was present up to 24h. In the chronic phase VEGF was

VASCULAR ENDOTHELIAL GROWTH FACTOR AND INFLAMMATION 93

Fig. 3-Immunolocalization of PECAM in the control PBS (a) and anti-VEGF (b)-treated groups. Numerous capillaries are observed in the control group but in the anti-VEGF- treated group virtually none is present. Note also the reduction in density of the granulomatous tissue in the anti-VEGF-treated group. The pre-existing blood vessels in the skeletal muscle (top part of each panel) were not affected by treatment. x 80

Fig. &Western blot analysis of VEGF in the murine chronic granulomatous tissue air pouch model at 6 , 12, and 24 h, and 3, 7, 14, 21, and 28 days. The VEGF band occurred at approximately 23 kD

present, at a reduced level, at day 3, but the levels post day 3 were significantly reduced. No VEGF was detected in normal skin.

in vivo3s4 and treatment with anti-VEGF antibodies significantly reduces angiogenesis in a number of tumour models.10 The growth of the invasive tissue, pannus, in rheumatoid arthritis has been suggested to be angiogenic-dependent." However, the role of VEGF in inflammation-induced angiogenesis has not been estab- lished. In this study, we have shown that treatment with anti-VEGF antibodies not only will significantly reduce the granulomatous tissue dry weight, but will also, more

DISCUSSION

There is an increasing body of evidence to suggest that VEGF may be the tumour angiogenic factor. It is the only endothelial cell-specific m i t ~ g e n , ~ it is angiogenic

94 1. APPLETON ET AL.

importantly, significantly suppress the vascular index, ACKNOWLEDGEMENTS i .e, angiogenesk Thus; as the growth of an inflam- matory tissue can be considered angiogenic-dependent,' by reducing the vascularity it is also possible to decrease the granulomatous tissue dry weight.

Western blot analysis showed high levels of VEGF in the acute stages of the inflammatory response, consistent

Dr I. Appleton is a recipient of a Royal Society, Smith and Nephew Foundation Fellowship. We are indebted to The Hyal Foundation, Toronto, Canada, and O N 0 Pharmaceutical Company Ltd., Osaka, Japan for support.

with its role as a vascular permeability factor. For example, it is approximately 50 000 times more potent than histamine.12-13 VEGF also has effects on other mechanisms central to angiogenesis. It can induce the production of interstitial collagenase, an enzyme essen- tial to basement membrane degradation and endothelial cell rnigration.l4 Taken together, VEGF would seem to be a prime candidate in modulating angiogenesis. It may, however, also act in synergy with other cytokines.

We have previously demonstrated the distribution pattern of a number of cytokines in the murine air pouch.15 It was found that high levels of immunoreac- tivity to basic fibroblast growth factor (bFGF) were present throughout the time course. bFGF is a potent angiogenic factor both in vitro and in vivo.16-'8 In addition to direct effects on angiogenesis, bFGF can upregulate the expression of VEGF in vascular smooth muscle cells.I9 Furthermore, in vitro studies have shown that not only are VEGF and bFCF synergistic on endothelial cell proliferation, but also that their effects on capillary tube formation are more rapid when used in combination than either cytokine alone.20

We have also shown in the murine air pouch that in the acute stages of the inflammatory response, low levels of transforming growth factor-beta (TGFP) are present, whereas after day 7 the levels of immunoreactivity to TGFP are markedly increased. It was demonstrated by Pepper et that low levels of TGFP potentiated capillary tube formation induced by bFGF and VEGF, whereas high levels attenuated the response of both cytokines. Therefore TGFP may be a prime candidate for modulating angio- genesis in inflammation. In the early stages of inflamma- tion, the actions of low levels of TGFP in conjunction with bFGF and VEGF may drive the angiogenic response. However, the subsequent rise in TGFP levels may then inhibit further angiogenesis. As demonstrated here, post day 3 , virtually no VEGF protein was observed.

Human umbilical vein endothelial cells will actively take up VEGF.22 Termination of endothelial cell pro- liferation results in a decrease of VEGF production.*3 Hence the findings of VEGF protein up to day 3 suggest active angiogenesis. This is corroborated by macro- and microscopic observations, as treatment with anti-VEGF resulted in almost complete inhibition of new blood vessel growth. These results indicate that the role of VEGF in inflammation is possibly two-fold. In the acute stages of inflammation it may act as a vascular per- meability factor, whereas as the inflammation progresses to the chronic stage it acts as a potent pro-angiogenic factor. This seems to indicate that similar to its pivotal role in tumour blood vessel growth, VEGF may be a prime candidate for inflammation-mediated angio- genesis. Our results may therefore help to explain some of the complex mechanisms of angiogenesis in chronic inflammatory diseases.

REFERENCES

I . Ferra N, Winer J, Burton T. Aortic smooth inuscle cells express and secrete vascular endothelid growth factor. Growth FucforJ 1991; 5 141-148.

2. Berse B, Brown LF, Van De Water L, Dvorak HF, Senger DR. Vascular permeability factor (vascular endothelial growth factor) gene is expressed differentially in normal tissues, macrophages and tumors. Mol Bid CeN 1992; 3: 211-220.

3. Leung DW, Cachianes G, Kuang W-J, Goeddel DV, Ferrara N. vascular endothelial growth factor is a secreted angiogenic mitogen. Science 1989; 246 1306-1309.

4. Connolly DT. Heuvelman DM, Nelson R, et a / . Tumor vascular per- meability factor stimulates endothelial cell growth and angiogenesis. J Clin Invest 1989; 8 4 1470-1474.

5. Klagsbrun M, Soker S. VEGFNPT: the angiogenesis factor found? Curr Biol 1993; 3 699-702.

6. Koch AE, Harlow LA, Haines GK, et ul. Vascular endothelial growth factor. A cytokine modulating endothelial function in rheumatoid arthritis. J Immunol 1994; 152 41494156.

7. Colville-Nash PR, Alam CAS, Appleton I, Brown JR, Seed MP, Willoughby DA. The pharmacological modulation of angiogenesis in chronic granulomatous inflammation. J Phurm Exp Ther 1995; 274 1463- 1472.

8. Kimura M, Amerniya K, Yamada T, Suzuki J. Quantitative method for measuring adjuvant-induced granuloma angiogenesis in insulin-treated diabetic mice. J Pharmucobiodyn 1986; 9 442446.

9. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976; 7 2 248-254.

10. Kim KJ, Li B, Winer J, er al. Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumour growth in vivo. Nature (London) 1993; 362 841-844.

11. Folkman J, Shing Y . Angiogenesis. J Biol Chem 1992; 267: 10931. 12. Senger DR, Connolly DT, Van De Water L, Feder J, Dvorak HF.

Purification and NH,-terminal amino acid sequence of guinea pig tumor- secreted vascular permeability factor. Cancer Res 1990; 5 0 17761778.

13. Keck PJ, Hauser SD, Krivi G, et a/. Vascular permeability factor, an endothelial cell mitogen related to PDGF. Science (Wash., DC) 1992; 246 1309-13 12.

14. Unemori EN, Ferrara N, Bauer EA, Aniento EP. Vascular endothelial growth factor induces interstitial collagenase expression in human endo- thelial cells. J Cell Physiol 1992; 153: 557-562.

15. Appleton I, Tomlinson A, Colville-Nash PR, Willoughby DA. Temporal and spatial immunolocalization of cytokines in murine chronic granuloma- tous tissue. Implications for their role in tissue development and repair processes. Lab Invest 1994; 6 9 405414.

16. Schweigerer L, Neufeld-Friedman J, Abraham JA, Fiddes JC, Gospodaro- wicz D. Capillary endothelial cells express basic fibroblast growth factor, a mitogen that promotes their own growth. Nature 1987; 325 257-259.

17. Lobb RR, Alderman ER, Fett JW. Induction of angiogenesis by bovine brain-derived class 1 heparin binding growth factor. Biochemistry 1985; 2 4 49694973.

18. Gospodarowicz D. Humoral control of cell proliferation: the role of fibroblast growth factor in regeneration, angiogenesis, would healing and neoplastic growth. Prog Clin Bid Res 1976; 9 1-19.

19. Stavri GT, Zachary IC, Baskerville PA, Martin JF, Erusalimsky JD. Basic fibroblast growth factor upregulates the expression of vascular endothelial growth factor in vascular smooth muscle cells. Synergistic interaction wit hypoxia. Circularion 1995; 9 2 11-14.

20. Pepper MS, Ferrara N, Orci L, Montesano R. Potent synergism between vascular endothelial growth factor and basic fibroblast growth factor in the induction of angiogenesis in vitm. Bioclzem Biophys Kes Comniun 1992: 189 824831.

21. Pepper MS, Vassalli JD, Orci L, Montesano R. Biphasic effect of transform- ing growth factor beta 1 on in v i m angiogenesis. Exp Cell Res 1993; 204 356-363.

22. Bikfalvi A, Sauzeau C, Moukadiri H, et ul. Interaction of vasculotropinl vascular endothelial cell growth factor with human umbilical vein endo- thelial cells: binding, internalization, degradation and biological effects. J . Cell Physiol 1991; 149 50-59.

23. Breier G , Albrecht U, Sterrer S, Risau W. Expression of vascular endo- thelid1 growth factor during ernbryonic angiogenesis and endothelial cell ditferentiation. Develupment 1992; 114 521-532.