Assays for angiogenesis: A review

Pharmac. Ther. Vol. 51, pp. 1-11, 1991 Printed in Great Britain. All rights reserved

0163-7258/91 $0.00 + 0.50 © 1991 Pergamon Press pie

Specialist Subject Editor: R. AUERBACH

ASSAYS FOR ANGIOGENESIS: A REVIEW ROBERT AUERBACH, WANDA AUERBACH a n d IGOR POLAKOWSKI Laboratory o f Developmental Biology, University o f Wisconsin, Madison, 141153706, U.S.A. Abstract--Accurate, reliable quantitation of the neovascular (angiogenic) response, both in vitro and in vivo, is an essential requirement for the study of new blood vessel growth. Over many years, ingenious ways have been developed for measuring this process, and they have contributed much to our present understanding of the vasculogenesis and angiogenesis that accompany normal embryonic development, lactation and wound healing, as well as tumor growth and a variety of other disease states ranging from diabetic retinopathy to autoimmune vasculitis. In this review we describe and evaluate the methodology and specific features of some of the most frequently used of these assays.

CONTENTS 1. Introduction 2. In Vivo Tests of Angiogenesis 2.1. Light and electron microscopy: qualitative studies 2.2. Angiogenesis induction in the cornea 2.3. Intradermal assays for angiogenesis 2.4. The chicken chorioallantoic membrane (CAM) assay 2.5. The disc angiogenesis assay 2.6. Other assays 3. In Vitro Assays for Angiogenesis 3.1. Endothelial cell migration 3.1.1. In vitro 'wound healing' 3.1.2. Chemokinesis 3.1.3. Chemotaxis 3.2. Endothelial cell proliferation 3.3. Endothelial cell functions 3.4. Three-dimensional differentiation: In vitro capillary formation 3.4.1. Three-dimensional substrates 3.4.2. Chick embryo culture systems 4. Concluding Remarks Acknowledgements References

1. I N T R O D U C T I O N Angiogenesis, the formation of new blood vessels, is a key element of a large number of normal and pathological processes. New blood vessel formation, or neovascularization, is seen most dramatically in tumor development, where the growth of solid tumors and their established metastases are dependent on the induction of an adequate blood supply (Folkman, 1975, 1985; D ' A m o r e and Thompson, 1987; Zetter, 1988; Blood and Zetter, 1990). Angiogenesis is associated with sites of immunological activity ranging from autoimmune reactions to inflammation (Auerbach, 1981). Neovascularization in the eye is an almost universal accompaniment of ocular diseases and of ocular injury (BenEzra et al., 1987).

1 2 2 3 3 4 5 5 6 6 6 6 7 7 7 7 7 8 8 8 8

Neovascularization is a prominent feature of psoriasis and scleroderma (Kaminski et al., 1984; Majewski et al., 1987). Angiogenesis plays an important part in normal processes as well. For example, the generation of blood vessels is a primary event in the establishment of the placenta and of the extraembryonic membranes during early embryonic development (Feinberg et al., 1991). In the absence of a vascular system, growth of an embryo is limited in mass to what can be nourished by diffusion. Angiogenesis thus acts as an important regulator of development by providing the means by which an individual structure of the embryo as a whole can expand by three-dimensional growth (Folkman, 1985). Neovascularization is a prominent feature of m a m m a r y gland changes associated with lactation.

2

R. AUERBACHet al.

Angiogenesis is also an essential component of wound healing and of repair processes (Knighton et al., 1990). Many disciplines of medicine and biology are therefore concerned with angiogenesis, with the consequence that the literature is widely scattered, appearing in basic science journals covering fields such as immunology, biochemistry, molecular biology, endocrinology, cell biology, anatomy or developmental biology, and in clinical journals directed towards study of diabetes, eye diseases, atherosclerosis, heart diseases, rheumatology or transplantation. Studies of drugs affecting angiogenesis may appear in the pharmacological literature while the problems of angiogenesis attendant to implantation of vascular prostheses may be discussed in the literature of biomedical engineering. Angiogenesis-inducing substances may be sought for stimulation of collateral heart vessels, increased vascularization of hair follicles, acceleration of engraftment of transplants, or promotion of wound healing; inhibitors may be tested for efficacy in reducing periodontal inflammation, inhibiting embryo implantation, inducing the regression of psoriatic lesions or preventing the growth of tumors or their metastatic foci. Since the development of assay procedures is a direct consequence of the need of an investigator to obtain specific information, the interests of that investigator have dictated both the type of assay developed and the nature of the data generated by the assay. The developmental biologist has been seeking assays for extension of blood vessels into embryonic rudiments; the ophthalmologist has sought information on choroidal or cornea vascularization; the oncologist has looked for a means of assessing angiogenic factor production by tumor cells; the immunologist has focussed on inflammatory mediators of local vasculogenesis; the reproductive biologist has sought for means of influencing blood vessel formation in the uterus. Perhaps the most consistent limitation in all these studies and approaches has been the availability of simple, reliable, reproducible, quantitative assays of the angiogenic response. In vivo tests have been difficult to perform and exceedingly variable; in vitro tests have been overly general and poorly validated. It is the purpose of this review to describe and evaluate the many in vivo and in vitro assay procedures that have been used in the study of angiogenesis.

2. I N V I V O TESTS OF A N G I O G E N E S I S

2.1. LIGHT AND ELECTRONMICROSCOPY; QUALITATIVESTUDIES

The oldest and in many ways still the most reliable assessment of angiogenesis has been that of simple, direct observation over a period of time. Even in ancient times there were descriptions of blood vessel development, as documented most clearly in the writings of Aristotle who examined chick embryo blood vessel development from the time of laying to the time of hatching (Aristotle, transl., 1945). Tumor vascularization was well discussed by Jones in 1850 who reported on his observations made on cancer

patients. With the establishment of pathology as a distinct discipline the number of careful descriptions of vascular manifestations of disease increased dramatically. Perhaps the most careful early study of embryonic angiogenesis was made by His, as summarized in his extensive monograph published in 1868 (His, 1868). He gave a precise description of chick embryo formation, staging embryos largely on the basis of the formation of the heart and blood islands. Using a combination of histology and traditional microscopy he used camera lucida drawing to reconstruct the events underlying vasculogenesis. His conclusion that blood vessel growth and development occurred by a centripetal movement of endothelial cells from their primary, central location to the periphery is still considered largely correct and it has served as basis for interpretation of many of the results obtained using experimental tests for angiogenesis. Contemporary methods using scanning or transmission microscopy are usually also supported by classic light microscopy observations (Cliff, 1963). Electron microscopy, however, provides resolution with respect to junctional complexes and microarchitecture that, in combination with light microscopy, can provide precision in identifying microvascular endothelial cells as these participate in the extension, regeneration or neoformations associated with angiogenesis (cf. Simionescu and Simionescu, 1991; Ausprunk et al., 1991). Studies performed by Rhodin and Fujita 0989) were based on a two-step procedure, combining light microscopy to select material and then using electron microscopy to gain the necessary precision. Selected segments of the mesenteric microvascular bed were examined and mapped for the presence of capillary sprouts using a videotape recording system. Besides selection of areas with sprouts, taped material later enabled these investigators to perform directional analysis of blood flow in newly formed capillaries. Selected areas of the microvascular bed were fixed by superfusion of glutaraldehyde solution, excised from anesthesized animals, embedded in epoxy resin and sectioned for transmission electron microscopy. Careful examination at 250 35,000 x magnification permitted the investigators to describe stages of elongation (extension) of endothelial cells involved in sprouting while simultaneously providing information on accessory cells including fibroblasts, pericytes, erythrocytes and platelets, all of which might be involved indirectly in the process of neovascularization. Similar findings were described in a study carried out by Wakui and his colleagues (Wakui, 1988; Furusato et al., 1990) on a sample of human granulation tissue. In this study capillary sprouting analysis involved three-dimensional reconstructions of serially sectioned material. Using micrographs of these sections, the cell profiles were traced on plastic sheets of a thickness adjusted to magnification, and the plastic replicas were glued to each other to result in a complete three-dimensional model. Horn et al. (1988) employed a more contemporary method to obtain three-dimensional reconstructions of dog mesenteric arterioles. High voltage electron microscopy (HVEM) was utilized permitting examin-

Assays for angiogenesis ation of thick (0.5 #m) sections. Rather than tracing the outlines on plastic sheets, each subsequent picture was digitized with a GTCO digipad and processed by image analysis software to generate three-dimensional images of small vessels and capillaries. Confocal microscopy, with its laser-light generated thin plane imaging, will permit a quantitative description of thrce-dimensional microvascular changes without the time-consuming need for thin-section preparation nor the highly limited access to equipment for HVEM studies (Cavanagh et al., 1990). 2.2. ANGIOGENESISINDUCTION IN THE CORNEA

The elicitation of an angiogenic reaction in the cornea must be considered the most convincing demonstration of true neovascularization, since the cornea is normally completely avascular (Henkind, 1978). The induction of new blood vessels requires that a stimulus be sensed by cells in the limbal vasculature at the edge of the cornea (Gimbrone et al., 1974). Endothelial cells from these vessels then migrate into the subepithelial space between the corneal epithelium and the stromal cells, forming a large number of initial sprouts all directed towards the source of angiogenic stimulation. Secondary sprouts may develop from the growing tip of the initial outgrowth (e.g. tumor-induced angiogenesis) leading to a 'brush border' morphology (Muthukkaruppan and Auerbach, 1979; Muthukkaruppan et al., 1982). Alternatively, there may be no further increase in the number of these vascular projections (e.g. lymphocyte-induced angiogenesis). In either case, once some of the new vessels reach the source of the stimulus, blood flow through those vessels increases. Concomitantly there is a decrease of flow in the less advanced vessels which then regress. The earliest use of intracorneal grafts to assess angiogenesis was carried out in rabbits (Gimbrone et al., 1974) where large pockets could be made for the insertion of tissues or slow release pellets containing angiogenic or antiangiogenic substances. When tumor cells or tumor extracts were placed within 2 mm of the corneoscleral junction, vascular sprouting could be detected within 36 hr. Although the rabbit can be used for the study of xenografts (Auerbach et al., 1975) major disadvantages of the rabbit as an experimental animal include absence of genetically uniform strains, difficulties in handling, amount of space needed and expense. Several other animal species have been substituted, including the guinea pig, rat and mouse (Fournier et al., 1981; Muthukkaruppan and Auerbach, 1979). The mouse corneal assay system in particular has overcome the previous limitations because of the wide range of inbred strains of mice available for study, the relative ease of handling and examination of the animals, and the low cost of their maintenance. On the negative side, the small size of the mouse eye makes it difficult to introduce slow release polymers, and the method of intracorneal implantation in this small animal requires greater skill in microsurgical techniques to obtain reproducible, reliable results. When slow release polymers such as Elvax or Hydron are used, the rabbit is still considered the animal of choice.

Attempts to achieve quantitation of corneal neovascularization have in the past met with limited success, although new methods of image analysis suggest that such quantitation is now feasible. Initially, the rate of vessel penetration into the cornea was taken as a measure of strength of the induced angiogenesis reaction but there was so much variability, largely due to an inability to achieve topographically precise placement of test substances (uniform distance from the limbus), that these measures provided only a crude indication of the strength of the reaction. Measurement of the number of vessels entering the cornea was also of marginal value for similar reasons. In our own laboratory we have introduced radioactive microspheres into the arterial circulation in an effort to quantitate blood flow in the cornea, but once again there was too much variability to place reliance in the results obtained. Image analysis methods have been reported by Klintworth and his colleagues that may improve quantitation (Proia et al., 1988; Culton et al., 1990; Haynes et al., 1989). Basically, vector analysis has been used to assess vascularization in response to a centrally-placed corneal insult. To date these methods require strong responses, and still have failed to negate the variability inherent in the procedure itself, especially since India ink infusion must be complete to achieve adequate contrast. Falkvoll (1991) has applied automated image analysis methods to the mouse cornea, and the newer imaging techniques combined with high-resolution video cameras are now able to resolve vessels in situ. This makes possible the sequential monitoring of neovascularization in individual animals, thus permitting study of progressive changes accompanying the induction of and inhibition of angiogenesis reactions in the cornea. Czegledy and his colleagues (unpublished observations) most recently have applied fractal mathematical modelling techniques developed for large blood vessels to analyze vessel penetration into the mouse cornea. Whether these methods can achieve more accurate quantitation, however, still needs to be demonstrated. 2.3. INTRADERMALASSAYS FOR ANGIOGENESIS

The intradermal assay for angiogenesis was originally developed in the mouse as a measure of the normal lymphocyte transfer reaction (Sidky and Auerbach, 1975). First described in 1975, the method was used to assess a local graft-versus-host reaction elicited when allogeneic lymphocytes are inoculated into the skin of an immunologically unresponsive host animal. The host response is marked by an increase in the number and tortuosity of visible vessels surrounding the inoculation site within 48-72 hr. Sidky and Auerbach achieved quantitation by enumerating the number of new vessels obtained, and demonstrated a strict relation between the number of effector cells injected and the number of vessels elicited at the injection site (Sidky and Auerbach, 1975, 1976, 1979; Auerbach and Sidky, 1979). Although counting of vessels is tedious and subject to some observer judgment, this kind of assay has been successfully used in a variety of studies. For example, vessel counts have led to delineation of

4

R. AUERBACHet al.

the relative role played by different genes of the major histocompatibility complex in the induction of immune-mediated angiogenesis (Auerbach and Sidky, 1979). It has been applied successfully to identify the role of CF4 + helper cells (Kaminski and Auerbach, 1988; Kreisle and Ershler, 1988), it has been applied in studies of psoriasis, sarcoidosis and systemic sclerosis (Majewski et al., 1987; Kaminski et al., 1984; Meyer et al., 1989; Wolf, 1989), and has been used to evaluate the effect of age of the responding animal to angiogenesis-inducing stimuli (Kreisle and Ershler, 1988; Hadar et aL, 1988). Efforts to obtain alternate measures of intradermal angiogenesis have until recently met with limited success. We have found that measurement of regional blood flow using radioactive microspheres is feasible, but the technique of intracardiac injection is too laborious to invite routine use. Tests using chromium labelled red blood cells or total hemoglobin content are usually inadequate since a single large vessel near the test site can unduly influence the measured regional blood flow. Recently, however, the intradermal assay has also become amenable to image analysis techniques. Photographs of vessels can be digitized and both the number of vessels and the number of divarication points can be obtained using standard computer programs. Even so, the three-dimensional growth of vessels in the skin represents an inherent problem since it is difficult to obtain a two-dimensional photograph in which an adjustment for depth of field does not at the same time reduce the physical length and detectable convoluted pattern of the newly formed blood vessels. A colorimetric assay for vascularization has also been described recently (Lee et al., 1990). 2.4. ThE CHICKEN CHORIOALLANTOICMEMBRANE (CAM) ASSAY Grafting of tissue onto the chick embryo chorioallantoic membrane (CAM) is a procedure that has been used by embryologists for more than 50 years. By transplanting embryonic rudiments onto the CAM, investigators were able to analyze the developmental potential of these isolated fragments over several days, using a dissecting microscope to monitor progress and histological analysis to determine histogenesis. Vascularization of grafts on the CAM was rapid, and because the early chick embryo lacks a mature immune system, even xenogeneic grafts from mammalian species were able to become established and grow. The study of Sorgente et al. (1975; cf. Eisenstein, 1991) who described the inhibitory effects of cartilage grafts on vascular development of the chick embryo, prompted Folkman and his associates to use CAM grafts for a direct study of tumor-induced anglogenesis (cf. Folkman, 1975). Eggs were incubated for 72 hr at which time they were prepared for subsequent grafting by withdrawing enough albumen to facilitate later graft placement without concomitant problems of adhesion to the shell membrane. A rectangular window in the shell served as an access point to place grafts or test material on the CAM at a time after the CAM vessels themselves had already been formed and their growing tips had progressed beyond the graft area. Angiogenesis was scored 3-4

days after grafting, and a 'spoke-wheel' type arrangement of vessels, directed towards the graft as the hub, was considered evidence of angiogenesis. A clear increase of vessels around the graft, even without the typical radial arrangement of vessels, was also considered a positive sign of neovascularization. In addition to studying tissue grafts, Folkman and his colleagues also examined the effect of chemical isolates by placing them on plastic coverslips, permitting the test substances to dry, then everting the coverslips onto the CAM (Folkman, 1985). The method has subsequently been widely applied in studies of angiogenesis inducers and angiogenesis inhibitors (Folkman and Klagsbrun, 1987). Quantitation of this assay was initially done by scoring the extent of vascularization on a graded scale of 0-4. Subsequently, it was found that serial dilution assays could be performed by scoring the number of positives at any particular dilution, using 4 eggs per assay point. High concentrations of a test substance would show 3/4 or 4/4 positive tests; as the concentration was reduced the number of positives would decrease until an endpoint (0/4) was reached. A more precise quantitation was achieved by Voss and his colleagues (Voss et al., 1984; Jakob and Voss, 1984) who used computer-assisted tracking of images to obtain both directional vector values and absolute values for the number of individual vessels. By use of vector values they were able to subtract the background 'noise' of resident vessels in the CAM, thus achieving a meaningful assessment of neovascularization as distinguished from normal vascularization. Their method has served as the prototype for the development of other methods of quantitation, none of which has yet been fully validated in the laboratory. Perhaps the most significant progress in the use of the chick embryo CAM assay has come from the development of in vitro methods for chick embryo culture (Auerbach et al., 1974). These are described in Section 3.4.2. below. Problems with the CAM assay, in addition to quantitation, have resulted from false positives produced by almost any irritant or wound. This is not surprising in the sense that wound healing and inflammatory responses include an angiogenic component (cf. Mahadevan et al., 1989). Publications in which investigators have questioned the CAM method as promoting the scoring of artifacts (e.g. egg shell dust) have legitimately alerted researchers to the need for care, but have missed the biological significance stemming from the fact that the positives are not 'false' but are biologically relevant reactions. Indeed, the additive or synergistic interactions between specific and more general stimuli of neovascularization may well be a rule rather than an exception, so that, for example, the testing of drugs or factors must consider the vehicle for administering a potential amplifier or inhibitor of the measured reaction. Several aspects of the CAM assay need to be considered both by those who select this assay for their analysis and by those who need to interpret the reported observations. The CAM is a growing, developing structure. It is heterogenous, with proximodistal differences in growth rates, differences in the size and nature of the vessels, and differences in the extent and direction of blood flow. The CAM in

Assays for angiogenesis the early chick embryo (days 7-9) is a rapidly growing membrane, whereas after day 11 it has virtually ceased to grow (Ausprunk et al., 1974; Ausprunk and Folkman, 1977). It is subject to modification by environmental factors ranging from gas content to pH, with the most pronounced variation being in the amount of keratinization, which has a significant effect on the CAM response to stimulation. 2.5. THE D~SC ANGIOGENESISASSAY Recently, a new assay has been described by Fajardo and his colleagues (1988) in which polyvinyl alcohol sponges are introduced subcutaneously and are then evaluated for penetration by host-derived blood vessels and/or other cell infiltrates. In brief, sponges made of polyvinyl alcohol foam are cut into 11 mm discs and their fiat sides are sealed with millipore filters. Before sealing a hole is cut, to generate a 'core', each core is impregnated with a test solution (e.g. FGF), sealed with a slow-release polymer (ethylene-vinyl acetate copolymer, Elvax) (cf. Langer et al., 1985) and then reinserted into the sponge. Reassembled sponges are then implanted subcutaneously and recovered after 1-3 weeks. Discs can be fixed, sectioned and stained with eosin and hematoxylin. Histological examination of sections of the discs, as reported by Fajardo and his coworkers, showed a distinctive cellular infiltration at the edges, and these cells included fibroblasts, endothelial cells and leukocytes. In some instances, complex blood vessels were observed. Although the authors suggest assessing the angiogenic response on the basis of the length of the longest blood vessel that has penetrated into the implanted disc, an alternative means of assessing the reaction is to stain the disc and measure the percent of the total cross-sectional area of the disc that shows penetration of ceils. Once again, this latter method permits the recording and analysis of digitized data. Using this latter approach we have been able to document both the enhanced penetration of cells in response to fibroblast growth factor (flFGF), a potent inducer of angiogenesis, and the inhibition of penetration in response to RNasin, an inhibitor of angiogenin and other ribonucleases (Auerbach and Polakowski, submitted for publication). One of the most important features of the disc angiogenesis assay is that the use of Elvax to coat sponge cores which leads to a slow release of test material. As presently designed, the coating by Elvax is imprecise, and the rate of release of such material following transfer into patients is not known. A similar approach was used by Orosz, Bishop and their colleagues (Bishop et al., 1989, 1990) who used subcutaneous implants of polyurethane foam cylinders. In these studies cells were introduced into the cylinders prior to closing of the incision. Ten days after implantation, the cylinders were recovered and processed for immunohistological examination. Antibodies against high endothelial venules served to delineate blood vessels within the polyurethane chambers. Various other comparable implant methods have been applied successfully to the study of angiogenesis in vivo in mice, rats and other experimental animals (Andrade et al., 1987; Mahadevan et al., 1989).

2.6. OTHEI~ASSAYS The hamster cheek pouch (HCP) is a special site, considered as 'immunologically privileged', i.e. it is possible to grow allogeneic or xenogeneic grafts in this area without immediate inflammatory reactions and with a reduced degree of cell-mediated immune response. At the same time the HCP provides a ready source of vessels from which neovascularization can be induced. Aside from tissue grafts, the HCP has also found use as a place for testing angiogenesisinducing factors in slow release formulations. For example, Schreiber et al. (1986) used the HCP assay to determine the angiogenic potential of TGF~t. The angiogenic reaction in this study was measured by examining histological sections stained with toluidine blue. In our own laboratory, the HCP assay was used to examine vascular reactions induced by tumor cells, scoring being accomplished by enumeration of vessels in the region surrounding the implanted tumor. The anterior eye chamber, like the HCP, has been considered immunologically privileged, serving as a site for xenogeneic growth of tumors since the early 1950s when Greene implanted human tumors into the rat anterior eye chamber (Greene, 1943). This site has been particularly useful in studies of neovascularization of preneoplastic mammary tumor cells transplanted into the rat anterior eye chamber. These studies have recently become the basis for a supplemental diagnostic protocol where biopsy pathology is considered suggestive (Folkman et al., 1989; Weidner et al., 1991). The dorsal air sac of the rat represents a classic site that was developed by Selye as a means of monitoring vascularization of tumor grafts (Selye, 1953). Air is injected under the dorsal skin of rats (or mice), permitting the introduction of reasonably large fragments of test tissues. Modifications of the air pouch method have included the introduction of transparent plastic that serves as a window under which neovascularization can be monitored. The omentum and other mesenteries, although used only sparingly to date, serve as a ready site for seeding of isolated tumor cells. For example, Williams et al. (1989) injected tumor cells into the peritoneal cavity where many of them adhered to the omentum. Subsequently, the omentum was excised and stained to visualize vascular sprouts. Histochemical tests could then be performed to characterize the newly formed vessels surrounding individual tumor islands. Norrby et al. (1986, 1990a,b) induced angiogenesis in the peritoneal membrane of rats and mice by activating resident mast cells with compound 40/80. Since this membrane is essentially avascular, the assay has some of the advantages described for the cornea. Angiogenesis was scored by enumeration of blood vessels. To achieve precision, blood vessels were detected on the basis of gray level analysis of digitized images obtained by acquiring a defined area of the membrane. The specially designed software created numerical representations (capillarization frequency) that depended on the average length of newly formed blood vessels and coefficients reflecting their directionality.

R. AUERBACHet al.

The chick embryonic yolk sac has been frequently used as an indicator of primitive vascular development (Taylor and Folkman, 1982; Rosenbruch, 1989; Takigawa et al., 1990). The chick yolk sac forms within 48 hr following incubation of fertilized eggs, and the development of blood vessels in the yolk sac is rapid and reasonably uniform, affecting every area of this large extraembryonic membrane. Thus the addition of test substances in slow release formulations can be studied by identifying zones of vessel inhibition in an otherwise uniformly developing vascular network. Fibrin, gelatin and other matrices have been used to provide a three-dimensional mesh into which yessels can be directed. For example, Dvorak and his associates demonstrated that fibrin introduced into the peritoneal cavity in a retaining cylinder acts as an angiogenesis initiator, such that endothelial cells that enter the clot form small patent vessels in this matrix (Dvorak et al., 1987). Effects of angiogenesis inducers and inhibitors could be studied by introducing test substances into the fibrin mixture before the congealing of the clot. Other matrix culture systems have been described (e.g. Nicosia and Ottinetti, 1990). Another useful method that has recently been developed is the use of ionic polysaccharides such as sodium alginate, which can be used to trap drugs and cells and can be injected subcutaneously (Bodmeier and Paeratakul, 1989; Piunkett and Hailey, 1990; Robertson et al., 1991). Alginate beads maintain part of their structure for as long as one month, during which time endothelial cell penetration into the beads can be monitored. Quantitation of angiogenesis has been achieved by determination of total hemoglobin as well as by measurement of blood flow using isotope-labelled red blood cells. Xenon clearance has been shown to be a useful means of quantitating new blood vessel formation (Andrade et al., 1987). This assay utilizes a polyester sponge implant placed under the skin. The sponge is penetrated by a cannula which extends to the surface of the skin and which is capped when not in use. The cannula provides access to the sponge so that test factors (e.g. angiogenesis factors) can be added. When angiogenesis is to be assessed radioactive 133Xe is introduced through the cannula and the xenon clearance is measured over a 6 h period. Increased clearance rate is correlated with vascularization of the sponge implant.

ASSAYS FOR ANGIOGENESIS

3. I N V I T R O

Since new blood vessels are comprised of vascular endothelial cells, the in vitro models for angiogenesis have all focussed on responses of endothelial cells that would provide in vitro correlates of in vivo angiogenesis. During the formation of new blood vessels, endothelial cells migrate away from a previously formed vessel in response to stimulation. That migration includes both an increased motility of the stimulated endothelial cells, as well as their migration toward the site from which angiogenic factors are released. While distally endothelial cells are seen to be

migrating, there is behind these an area where rapid cell proliferation takes place, providing the increased cell numbers essential for the formation of previously nonexistent blood vessels. Finally, the endothelial cells organize, form a lumen and establish a patent vascular connection leading to blood flow.

3.1. ENDOTHELIAL CELL MIGRATION

3.1.1. In Vitro ' Wound Healing' Since endothelial cells that were previously quiescent become active as they migrate to heal wounds, in vitro models of wound healing were developed. Typically, an endothelial cell monolayer is prepared which becomes static when the cells reach confluence (contact inhibition). If at any time, subsequent to contact inhibition, an area of the dish is denuded, cells at the edge of the cleared surface will begin to migrate outward to fill the empty space. The rate of recovering of the denuded surface can be measured to provide a quantitative assessment of the angiogenic response (Pepper et al., 1987, 1989, 1990). Several devices have been described that provide a uniform 'wounding' area, thus helping to achieve reproducibility of the method. As an alternative to wounding of a preformed confluent monolayer, endothelial cells are allowed to cover only a small circular area in the center of a culture dish. This is achieved by blocking access to the remainder of the dish surface by use of a teflon barrier. Once cells are confluent the barrier is removed and cells can now migrate outward from the central area of growth. Once again, angiogenesis is measured by determining the rate of outgrowth from the periphery of the preseeded endothelial cell (confluent) monolayer. All of these models of wound healing are based on the central thesis that endothelial cell migration into a denuded area is a pivotal event in wound healing in vivo. That this is a reasonable thesis is shown by the fact that the events following in vivo denuding through introduction of a balloon catheter into a large vessel are in large part similar to what is observed in these in vitro models. 3.1.2. Chemokinesis Since endothelial cells migrate in response to angiogenic factors, the simple measurement of increased cell movement provides a useful assay of endothelial cell reactivity. Zetter and his colleagues applied the phagokinetic track assay of AlbrechtBuehler to the study of endothelial cells (cf. Zetter, 1987). In that assay, colloidal gold is plated onto coverslips. Test cells are then added to the culture, so that individual cells that move on the gold monolayer phagocytose and/or push aside the colloidal particles, leaving migration tracks. These tracks can be measured to provide information on the extent of movement over time as well as on the polarized vs random motility of those ceils. Simple imaging techniques can be used to provide quantitative data on the formed tracks (cf. Weber et al., 1989; Rupnick et al., 1988).

Assays for angiogenesis Obeso and Auerbach (Obeso and Auerbach, 1984; Auerbach et aL, 1991) modified the phagokinetic track assay by substituting polystyrene beads for the colloidal gold. Beads can readily be affixed to microwell culture wells, and thus they were able to set up multiple 96-well culture plates. Such mass chemokinesis assays are adequate for dose/response studies and hybridoma screening. By adding automated image analysis methodology to the culture system, the assay now provides a means of obtaining reliable information on large numbers of assay wells to describe chemokinetic responses of endothelial cells. Using this mass screening assay for evaluating endothelial cell movement, Auerbach and his colleagues have reported the presence of increased amounts of angiogenic cytokines in the bronchoalveolar lavage fluids of patients with granulomatous pathologies as well as the production of such angiokines by immune-stimulated T-lymphocytes (Weber et al., 1989). The method has most recently found application in the screening of inhibitors of angiogenesis, providing a rapid assay for obtaining dose/response kinetics on these inhibitors (Auerbach and Polakowski, submitted for publication). In his recent review, Zetter (1987) pointed out that "to date the ability to stimulate the migration of capillary endothelial cells has been a property of every angiogenic factor studied. Thus while both cell migration and proliferation are normal components of the angiogenic process, only cell migration appears to be obligatory. This makes the in vitro assay of capillary endothelial cell migration an extremely useful indicator of angiogenic potential." 3.1.3. Chemotaxis Chemotaxis includes not only cell movement but the directionality of that movement. As such, enumerating the cells that move preferentially toward a chemical stimulus may well be the most reliable measure of a direct response of endothelial cells to angiogenic factors. Since the classic Boyden chamber assay is laborious, several modifications have been described that provide multichamber assemblies to increase the efficiency of the assay system (e.g. Stokes et al., 1990). While the basic Boyden chamber methodology requires the counting of cells that have traversed a filter barrier, the original assessment by inspection under a microscope is being replaced by a variety of methods that permit more convenient quantitation (Tsuboi et al., 1990; Taraboletti et al., 1990). As pointed out by Zetter, where there is a clear indication of chemokinesis, chemotaxis studies may not be essential. However, a number of factors (e.g. TGFfl) have been identified whose effects are too subtle to be detected by increased motility, and it is only by the addition of a directional measure that these factors have demonstrated their angiogenic properties. Ideally, the convenience of measuring chemokinesis should be combined with the sensitivity of the chemotaxis assay, and several laboratories are now trying to achieve such a combination.

3.2. ENDOTHELIALCELL PROLIFERATION

Endothelial cell proliferation has been used more extensively than any other in vitro assay to assess angiogenic activity (cf. Ryan, 1988; Simionescu and Simionescu, 1988, 1991; Folkman and Ingber, 1987; Folkman and Klagsbrun, 1987). Generally, endothelial cells are cultured in the presence of test factors and proliferation is determined either by cell counts or by use of radioactive nucleotides whose incorporation is used to determine the extent of DNA synthesis (cf. Watt and Auerbach, 1986; Folkman and Klagsbrun, 1987). Other standard methods (e.g. nuclear staining, flow cytometric determination of DNA content) have also been used. There is little doubt that many angiogenic factors do induce endothelial cell proliferation. It must, however, be kept in mind that not all factors that induce such proliferation induce angiogenesis in vivo, and that, moreover, a fairly large number of angiogenesis-inducing factors do not evoke a proliferative reaction from cultured endothelial cells. On the other hand, since the assay is readily performed, quantitative, and frequently correlated with angiogenesis, it is a reasonable first test for validating an angiogenesisinducing or angiogenesis-inhibiting factor. 3.3. ENDOTHELIALCELLFUNCTIONS Many functional attributes of endothelial cells have been used to monitor endothelial cell activity in vitro (cf. Simionescu and Simionescu, 1988, 1991; Ryan, 1988). Many endothelial cell activities are altered as an adjunct of their reaction to angiogenesisinducing agents. Among the most useful attributes at alterations in the expression of von Willebrand factor antigen (Mourad et al., 1990), changes in fibrinolytic activity (Sueishi et al., 1989), and modulations in the production of cytokines (Shepro, 1988).

3.4. THREE-DIMENSIONALDIFFERENTIATION: I N V I T R O CAPILLARYFORMATION 3.4.1. Three-Dimensional Substrates One of the most intriguing three-dimensional assay systems is that of tube formation in vitro. Initially, it was observed that when endothelial cell cultures were maintained as confluent monolayers for prolonged periods of time without replenishment of nutrients three-dimensional structures would be formed whose morphology resembled that of true blood vessels (Folkman et al., 1979). Subsequently, it was shown that tube-like formations can be enhanced by providing a variety of substrates (Nicosia and Ottinetti, 1990; Madri and Pratt, 1986; Maciag et al., 1982; Maciag, 1984). For example, vessel formation is more rapid and reliably obtained when endothelial cells are seeded into collagen gels or plated on a layer of laminin (Yasunaga et al., 1989; Sueishi et al., 1989; Madri and Pratt, 1986). Addition of angiogenesisinducing substances such as sodium orthovanadate show enhanced and accelerated vessel formation in these gels. Moreover, the three-dimensional structures generated in vitro show typical vessel patency. While

8

R. AUERBACHet al.

formation of these vessels is not in itself a quantitative measure of angiogenesis, the method is nevertheless a useful extension of other in vitro methods, providing a more direct in vitro evidence of vasculogenesis than can be achieved by documentation of either endothelial cell migration or endothelial cell proliferation. 3.4.2. Chick Embryo Culture Systems The chick chorioallantoic membrane assay carried out on explanted chick embryos was developed as an alternative to the use of embryonated eggs, as already described above. Although technically this is indeed an in vitro assay, it comes closest to a whole animal test, since the entire embryo and its membranes are maintained intact. In this assay (Auerbach et al., 1974), the egg content is transferred to a petri dish where development continues to take place. Under these culture conditions, the CAM develops on top as a flat membrane, reaching to the edge of the petri dish to provide a two-dimensional monolayer onto which grafts can be placed. Because the entire membrane is visible, rather than only a small portion exposed in the shell window, multiple grafts can be placed on individual CAMs and the grafts can be photographed periodically to document vascular changes over time. Scoring initially was accomplished by grading as described above (cf. Auerbach et al., 1976; Form and Auerbach, 1983; cf. Folkman, 1985), but once again, image analysis now provides increased accuracy in quantitation. Moreover, in contrast to the vessels induced in the intradermal assay, the in vitro CAM assay yields a two-dimensional set of images not subject to the distortions of three-dimensional growth. Several modifications of this method have subsequently been described (Dunn et al., 1981; Jakobson et al., 1989; Stewart et al., 1990). The simplest of these utilize a container such as a beaker or plastic cup into which a piece of plastic wrap is suspended which holds the egg content in a more rounded position. The advantages include somewhat increased viability and decreased cost, but these are offset by the difficulty of monitoring angiogenesis during the period of incubation and by an inability to obtain twodimensional photographs suitable for image analysis. On the other hand, the modifications are suitable for preparing large numbers of embryos in order to carry out structure-function and dose-response endpoint assessments for efficacy of inducers and inhibitors of angiogenesis. As yet unexploited are extension of embryo culture methods to the study of mammalian angiogenesis. Techniques for embryo explantation including the maintenance of extraembryonic membranes are available and these offer the potential for threedimensional assessment of the effects of specific factors on mammalian vasculogenesis.

4. C O N C L U D I N G REMARKS It is becoming increasingly apparent that neovascularization or angiogenesis represents a process of major significance both during normal development and as a cause of or accompaniment of disease processes. Recognizing this, many recent studies have

focussed on finding factors that induce or inhibit this reaction. Increased vessel formation would promote wound healing, accelerate engraftment, and provide a means of furnishing new collateral blood flow to areas of vascular insufficiency. Similarly, inhibition of angiogenesis could lead to a regression of solid tumors, reduction of metastatic lesions, and a diminishing of immune-mediated or inflammationassociated vasculitis. As was stated at the beginning of this review, "perhaps the most consistent limitation (to progress) has been the availability of simple, reliable, reproducible, quantitative assays of the angiogenic response." There have, however, been many recent technical advances, including the use of sophisticated image analysis techniques and in vitro methods for growing and assessing the responses of microvascular endothelial cells. This leads to the prediction that in the next few years much new information will be gained both on the factors that induce and those that inhibit the formation of new blood vessels. Acknowledgements--We thank Nova Pharmaceuticals, Inc.

for a gift in support of slow release polymer drug delivery research in our laboratory, and Promega Corporation and the State of Wisconsin Development Fund for underwriting the costs of exploring new in vivo assays of angiogenesisand antiangiogenesis. IJP was supported by an NIH Fogarty Center fellowship.

REFERENCES ANDRADE, S. P., FAN, T.-P. D. and LEWIS, G. P. (1987) Quantitative in-vivo studies on angiogenesis in a rat sponge model. Br. J. exp. Path. 68:755 766. ARISTOTLE384--322B. C. (1945) Parts of Animals, Harvard

University Press, Cambridge, MA. AUERBACn, R. (1981) Angiogenesis-inducing factors: A review. Lymphokines 4:69 88. AUERBACH, R. and SIDKY, Y. A. (1979) Nature of the stimulus leading to lymphocyte-induced angiogenesis. J. Immun. 123:751 754. AUERBACH,R., KUBAI,L., KNIGHTON,D. and FOLKMAN,J. (1974) A simple procedure for the long-term cultivation of chicken embryos. Devl Biol. 41:391 394. AUERBACH, R., ARENSMAN,R., KUBAI, L. and FOLKMAN,J.

(1975) Tumor-induced angiogenesis:lack of inhibition by irradiation. Int. J. Cancer 15: 241-245. AUERBACH,R., KUBAI,L. and SIDKY,Y. (1976) Angiogenesis induction by tumors, embryonic tissues, and lymphocytes. Cancer Res. 36: 3435-3440. AUERBACH, R., BIELICH,H., OBESO,J. and WEBER, J. (1991) Quantitation of endothelial cell movement. An in ~itro approach to vasculogenesis and angiogenesis. In: The Development of the Vascular System, pp. 180 189, FEINBERG, R. N., SHERER, G. K. and AUERaACH, R. (eds)

Karger, Basel. AUSPRUNK, D. H. and FOLKMAN, J. (1977) Migration

and proliferation of endothelial ceils in preformed and newly formed blood vessels during tumor angiogenesis. Microvasc. Res. 14: 53-65. AUSPRUNK, D. H , KNIGHTON,D. R. and FOLKMAN,J. (1974)

Differentiation of vascular endothelium in the chick chorioallantois: a structural and autoradiographic study. Devl Biol. 38:237 249. AUSPRUNK, D. H., DETHLEFSEN, S. M. and HIGG~NS,E. R. (1991) Distribution of fibronectin, laminin and type IV collagen during development of blood vessels in the chick chorioallantoic membrane. In: The Development of the

Assays for angiogenesis Vascular System, pp. 93-108, FEINBERG, R. N., SHERER, G. K. and AUERBACH,R. (eds) Karger, Basel. BEN EZRA, D., RYAN, S. J., GLASER,B. N. and MURPHY, R. P. (eds) (1987) Ocular Circulation and Neovascularization, Martinus Nijhoff/Junk, Amsterdam. BISHOP, n . K., JUTILA, M. A., SEDMAK,n . n., BEATTIE,M. S. and OROSZ, C. G. (1989) Lymphocyte entry into inflammatory tissues in vivo. Qualitative differences of high endothelial venule-like vessels in sponge matrix allografts vs isografts. J. Immun. 142: 4219-4224. BISHOP, D. K., SEDMAK,n . n., LEPP1NK, D. M. and OROSZ, C. G. (1990) Vascular endothelial differentiation in sponge matrix allografts. Hum. Immun. 28: 128-133. BLOOD, C. H. and ZE~rER, B. R. (1990) Tumor interactions with the vasculature: angiogenesis and tumor metastasis. Biochem. biophys. Acta 1032: 89-118. BODMEmR, R. and PAERATAKUL, O. (1989) Spherical agglomerates of water-insoluble drugs. J. Pharmac. Sei. 78: 964-967. CAVANAGH, H. D., JESTER, J. V., ESSEPIAN, J., SHIELDS, W. and LEMP, M. A. (1990) Confocal microscopy of the living eye. CLAO J. 16: 65-73. CLIFF, W. J. (1963) Observations on a healing tissue, a combined light and electron microscopic investigation. Trans. R. Soc. Lond. (Biol. Sci.) 248: 305-325. CULTON, M., CHANDLER, D. B., PRom, A. D., HICKINGBOTHAM, D. and KLINTWORTH, G. K. (1990) The effect of oxygen on corneal neovascularization. Invest. Ophthal. Vis. Sci. 31: 1277-1281. D'AMORE, P. and THOMPSON, R. W. (1987) Mechanisms of angiogenesis. A. Rev. Physiol. 49: 453-464. DUNN, B. E., FITZHARRIS,T. P. and BARNETT, B. D. (1981) Effects of varying chamber construction and embryo pre-incubation age on survival and growth of chick embryos in shell-less culture. Anat. Rec. 199: 33-43. DVORAK, H. F., HARVET, V. S., ESTRELLA,P., BROWN, L. F., McDONAGH, J. and DVORAK, A. M. (1987) Fibrin containing gels induce angiogenesis. Implications for tumor stroma generation and wound healing. Lab. Invest. 57: 673-686. EISENSTEIN, R. (1991) Angiogenesis in arteries: Review. Pharmac. Ther. 49: 1-19. FAJARDO, L. F., KOWALSKI, J., KHWAN, H., PRIONAS, S. D. and ALLISON, A. C. (1988) The disc angiogenesis system. Lab. Invest. 58: 718-724. FALKVOLL, K. H. (1991) A method to quantify neovascularization in the mouse cornea. Ophthal. Res. 112, in press. FEINBERG, R. N., SHERER, G. K. and AUERBACH, R. (eds) (1991) The Development of the Vascular System, S. Karger, Basel. FOLKMAN, J. (1975) Tumor angiogenesis. Adv. Cancer Res. 43: 175-203. FOLKMAN, J. (1985) Toward an understanding of angiogenesis: Search and discovery. Perspect. Biol. Med. 29: 10-36. FOLKMAN, J. and INGBER,D. E. (1987) Angiostatic steroids. Method of discovery and mechanism of action. Ann. Surg. 206: 374-383. FOLKMAN, J. and KLAGSBURN, M. (1987) Angiogenic factors. Science 235: 442-447. FOLKMAN,J., HAUDENSCHILD,B. C. and ZETTER,B. R. (1979) Long-term culture of capillary endothelial cells. Proc. hath. Acad. Sci. U.S.A. 76: 5217-5221. FOLKMAN, J., WATSON, K., INGBER, D. and HANAHAN, D. (1989) Induction of angiogenesis during the transition from hyperplasia to neoplasia. Nature 339: 58-61. FORM, D. M. and AImRBACH, R. (1983) PGE2 and angiogenesis. Proc. Soc. exp. Biol. Med. 172: 214-218. FOURNIER, G. A., LUTTY, G. A., WATT, S., FENSELAU,A. and PATZ, A. (1981) A corneal micropocket assay for angiogenesis in the rat eye. Invest. Ophthal. Vis. Sci. 21: 351-354.

FURUSATO, M., WAKUI, S., SUZUKI, M., TAKAG1,K., HORI, M., ASARI, M., KANO, Y. and USHIGOME,S. (1990) Threedimensional ultrastructural distribution of cytoplasmic interdigitation between endothelium and pericyte of capillary in human granulation tissue by serial section reconstruction method. J. Electron Microsc., Tokyo 39: 86-91. GIMBRONE, M. A. JR, COTRAN, R. A., LEAPMAN, S. B. and FOLKMAN, J. (1974) Tumor growth and neovascularization: An experimental model using the rabbit cornea. d. hath. Cancer Inst. 52: 413-427. GREENE,H. S. N. (1943) The heterologous transplantation of embryonic mammalian tissues. Cancer Res. 3: 809-822. HADAR, E. J., ERSHLER, W. B., KREISLE, R. A., HO, S. P., VOLK, M. J. and KLOPP, R. G. (1988) Lymphocyteinduced angiogenesis factor is produced by LDT4 + murine T lymphocytes, and its production declines with age. Cancer lmmun. Immunother. 26: 31-34. HAYNES, W. L., PROIA,A. n . and KLINTWORTH, G. K. 0989) Effect of inhibitors of arachidonic acid metabolism on corneal neovascularization in the rat. Invest. Ophthal. ISis. Sci. 30: 1588-1593. HENKIND, P. (1978) Ocular neovascularization, The Krill memorial lecture. Am. J. Ophthal. 85:287 301. HIS, W. (1868) Untersuchungen ueber die erste Anlage des Wirbelthierleibes, F. C. W. Vogel, Leipzig. HORN, L., KRAJEWSKI,W. S., PAUL, P. K., SONG,M. J. and SYDOR, M. J. (1988) Computerized 3-D reconstruction of small blood vessels. Adv. exp. Biol. Med. 242: 35-42. JAKOB,W. and Voss, K. (1984) Utilization of image analysis for the quantification of vascular responses in the chick chorioallantoic membrane. Exp. Path. 26: 93-99. JAKOBSON, A. i . , HAHNENBERGER,R. and MAGNUSSON,A. (1989) A simple method for shell-less cultivation of chick embryos. Pharmac. Toxic. 64:193 195. JONES, T. W. (1850) On the state of the blood and the blood-vessels in inflammation. Guy's Hosp. Reports, 2nd Set. 7: 1-94. KAMINSKI, M. and AUERBACH, R. (1988) Angiogenesis induction by CD-4 positive lymphocytes. Proc. Soc. exp. Biol. Med. 188: 440-443. KAMINSKI, M., MAJEWSKI, S., JABLONSKA,S. and PAW1NSKA, M. (1984) Lowered angiogenic capability of peripheral blood lymphocytes in progressive systemic sclerosis (scleroderma). J. Invest. Derm. 82: 239-243. KNIGHTON, D. R., PHILLIPS, G. n . and FIEGEL, V. D. (1990) Wound healing angiogenesis: indirect stimulation by basic fibroblast growth factor. J. Trauma 30 (Suppl. 12): S134-S144. KREISLE, R. A. and ERSHLER, W. B. (1988) Investigation of tumor angiogenesis in an id mouse model: role of hosttumor interactions. J. hath. Cancer Inst. 26: 849-854. LANGER, R., BREM, H. and TAPPER, n . (1985) Biocompatibility of polymeric delivery systems for macromolecules. d. Pharmac. Sci. 72: 1181-1185. LEE, K. E., IWAMURA, i . and COCKETT, A. T. (1990) Cortisone inhibition of tumor angiogenesis measured by a quantitative colorimetric assay in mice. Cancer Chemother. Pharmac. 26: 461-463. MACIAG,T. (1984) Angiogenesis. Prog. Haemost. Thromb. 7: 167-182. MAC1AG,T., KADISH,J., WILKINS, L., STEMERMAN,M. B. and WEINSTEIN, R. (1982) Organizational behavior of human umbilical vein endothelial cells. J. Cell Biol. 94:511-520. MADRI, J. A. and PRATT, B. M. (1986) Endothelial cell-matrix interactions: in vitro models of angiogenesis. J. Histochem. Cytochem. 34: 85-91. MAHADEVAN, V., HART, I. R. and LEWIS, G. P. (1989) Factors influencing blood supply in wound granuloma quantitated by a new in vivo technique. Cancer Res. 49: 415-419. MAJEWSKI, S., TIGALONOWA, i . , JABLONSKA, S., POLAKOWSKI, I. and JANCZURA,E. (1987) Serum samples

10

R. AUERBACH et al.

from patients with active psoriasis enhance lymphocyteinduced angiogenesis and modulate endothelial cell proliferation. Arch. Derm. 123: 221-225. MEYER, K. C., KAMINSKI, M. J., CALHOUN, W. J. and AUERBACH, R. (1989) Studies of bronchoalveolar lavage cells and fluids ion pulmonary sarcoidosis. I. Enhanced capacity of bronchoalveolar lavage cells from patients with pulmonary sarcoidosis to induce angiogenesis in vivo. Am. Rev. resp. Dis. 140: 1446-1449. MOURAD, M. M., GIDDINGS, J., MARKS, R., BARTON, S. P. and BAVERSTOCK,J. (1990) An immunoradiometric assay for von Willebrand factor antigen to assess angiogenesis. Br. J. Derm. 123: 21--28. MUTHUKKARUPPAN, V. and AUERBACH, R. (1979) Angiogenesis in the mouse cornea. Science 205:1416 1418. MUTHUKKARUPPAN, V. R., KUBAI, L. and AUERBACH, R. (1982) Tumor-induced neovascularization in the mouse eye. J. num. Cancer Inst. 69: 699-708. NICOSIA, R. F. and OTTINETTI, A. (1990) Growth of microvessels in serum-free matrix culture of rat aorta. A quantitative assay of angiogenesis in vitro. Lab. Invest. 63:115-122. NORRBY, K., JAKOBSSON,A. and SORBO,J. (1986) Mast-cellmediated angiogenesis: a novel experimental model using the rat mesentery. Virchows Arch. (B) 52: 195-206. NORRBY, K., JAKOBSSON,A. and SORBO, J. (1990a) Quantitative angiogenesis in spreads of intact rat mesenteric windows. Microvasc. Res. 39: 341-348. NORRBY, K., JAKOBSSON, A. and SORBO, J. (1990b) On mast-cell mediated angiogenesis in the rat mesentericwindow assay. Agents Actions 30: 231-233. OBESO, J. L. and AUERBACH, R. (1984) A new microtechnique for quantitating cell movement in vitro using polystyrene bead monolayers. J. Immun. Meth. 70: 141-152. PEPPER, M. S., VASSALLI,J. D., MONTESANO,R. and ORCI, L. (1987) Urokinase-type plasminogen activator is induced in migrating capillary endothelial cells. J. Cell Biol. 105: 2535 2541. PEPPER,M. S., SPRAY,D. C., CHANSON, M., MONTESANO,R., ORCl, L. and MEDA, P. (1989) Junctional communication is induced in migrating capillary endothelial cells. J. Cell Biol. 109: 3027-3038. PEPPER, M. S., BELIN, D., MONTESANO, R., ORC1, L. and VASSALL1,J. D. (1990) Transforming growth factor-beta 1 modulates basic fibroblast growth factor-induced proteolytic and angiogenic properties of endothelial cells in vitro. J. Cell Biol. 111:743 755. PLUNKETT, M. L. and HAILEY, J. A. (1990) An in vivo quantitative angiogenesis model using tumor cells entrapped in alginate. Lab. Invest. 62: 510-517. PROIA, A. D., CHANDLER, D. B., HAYNES,W. L., SMITH,C. F., SUVARNAMANI,C., ERKEL, F. H. and KLINTWORTH, G. K. (1988) Quantitation of corneal neovascularization using computerized image analysis. Lab. Invest. 58: 473479. RHODIN, J. A. G. and FUJITA, H. (1989) Capillary growth in the mesentery of normal young rats. Intravital video and electron microscope analyses. J. submicrosc. Cytol. Path. 21: 1-34. ROBERTSON,N. E., DISCAFANI,C. M., DOWNS, E. C., HAILEY, J. A., SARRE, O., RUNKLE, R. L. JR, POPPER, T. L. and PLUNKETT, M. L. (1991) A quantitative in vivo mouse model used to assay inhibitors of tumor-induced an#ogenesis. Cancer Res. 51: 1339-1344. ROSENBRUCH, M. (1989) Granulation tissue in the chick embryo yolk sac blood vessel system. J. comp. Path. 101: 363-373. RUPNICK, M. A., STOKES, C. U, WILLIAMS, S. K. and LAUFFENBURGER, D. A. (1988) Quantitative analysis of random motility of human microvessel endothelial cells using a liner under-agarose assay. Lab. Invest. 59: 363-372.

RYAN, U. S. (ed) (1988) Endothelial Cells, CRC Press, Boca REtort, LA. SCHREIBER, A. B., WINKLER, M. E. and DERYNCK, R. (1986) Transforming growth factor alpha: a more potent an#ogenie factor than epidermal growth factor. Science 232: 1250-1253. SELYE, H. 0953). On the mechanism through which hydrocortisone affects the resistance of tissues to injury. J. Am. med. Ass. 152: 1207-1213. SH~PRO, D. (1988) Endothelial cells, inflammatory edema, and the microvascular barrier: Comments by a "free radical". Microvasc. Res. 35: 247-264. SIDKY, Y. A. and AUERBACH, R. (1975) Lymphocyteinduced angiogenesis: a quantitative and sensitive assay of the graft-vs.-host reaction. J. exp. Med. 141:1084-1100. SIDKY, Y. A. and AUERBACH, R. (1976) Lymphocyteinduced angiogenesis in tumor-bearing mice. Science 192: 1237-1238. SIDKY, Y. A. and AUEREACH, R. (1979) Response of the host vascular system to immunocompetent lymphocytes: effect of preimmunization of donor or host animals. Proc. Soc. exp. Biol. Med. 161: 174-177. SIMIONESCU,N. and SIM~OrCESCU,M. (eds) (1988)Endothelial Cell Biology in Health and Disease, Plenum, New York. SIMIONESCU,N. and SIMIONESCU,M. (eds) (1991) Endothelial Cell Dysfunction, Plenum, New York. SORGENTE, N., KUETTNER, K. E., SOBLE, L. W. and EISENSTEIN, R. (1975) The resistance of certain tissues to invasion. II. Evidence for extractable factors in cartilage which inhibit invation by vascularized mesenchyme. Lab. Invest. 32: 217-222. STEWART, R., NELSON, J. and WILSON, D. J. (1990) Growth of the chick area vasculosa in ovo and in shell-less culture. J. Anat. 172: 81-87. STOKES, C. L., RUPNICK, M. A., WILLIAMS, S. K. and LAUFFENBURGER, D. A. (1990) Chemotaxis of human microvessel endothelial cells in response to acidic fibroblast growth factor. Lab. Invest. 63: 657-668. SUEISm, K., YASUNAGA, C., NAKASHIMA, Y., TSUTSUI, H. and Isnn, Y. (1989) Endothelium-fibrinolysis system interaction. Nippon Ketsueki Gakkai Zasshi 52: 1350 1358. TAKIGAWA, M., NISHIDA, Y., SUZUKI, F., KIsm, J., YAMASHITA, K. and HAYAKAWA,T. (1990) Induction of angiogenesis in chick yolk-sac membrane by polyamines and its inhibition by tissue inhibitors of metalloproteinases (TIMP and TIMP-2). Biochem. biophys. Res. Commun. 171: 1264-1271. TARABOLETTI,G., ROBERTS,D., LIOTTA,L. A. and GIAVAZZI, R. (1990) Platelet thrombospondin modulated endothelial cell adhesion, motility, and growth: a potential angiogenesis regulatory factor. J. Cell Biol. 111: 765-772. TAYLOR, S. and FOLKMAN, J. (1982) Protamine is an inhibitor of angiogenesis. Nature 297: 307-312. TSUBOI, R., SATO, Y. and RIFKIN, D. B. (1990) Correlation of cell migration, cell invasion, receptor number, proteinase production, and basic fibroblast growth factor levels in endothelial cells. J. Cell Biol. 110: 511-517. Voss, K., JAKOB, W. and ROTH, K. (1984) A new image analysis method for the quantification of neovascularization. Exp. Path. 26: 155-161. WAKUI, S. (1988) Two- and three-dimensional ultrastructural observation of two cell angiogenesis in human granulation tissue. Virchows Arch (B) 56: 127-139. WATT, S. L. and AUERBACH, R. (1986) A mitogenic factor for endothelial cells obtained from mouse secondary mixed leukocyte cultures. J. Immun. 136: 197-202. WEBER,J., MEYER, K. C., BANDA, P., CALHOUN,W. J. and AUERBACH, R. (1989) Studies of bronehoalveolar lavage cells and fluids in pulmonary sarcoidosis. II. Enhanced capacity of bronchoalveolar lavage fluids from patients with pulmonary sarcoidosis to induce cell movement in vitro. Am. Rev. res. Dis. 140: 1450-1454.

Assays for angiogenesis WEIDNER,N., SEMPLE,J. P., WELCH,W. R. and FOLKMAN,J. (1991) Tumor angiogenesis and metastasis---correlation in invasive breast carcinoma. New Engl. J. Med. 324: 1-8.

WILLIAMS, R. J., ROBERTSON,D. and DAVIES,A. J. (1989) Identification of vascular endothelial cells in murine omentum using the lectin, Dolichos biflorus agglutinin: possible applications in the study of angiogenesis. Histochem. J. 21: 271-278.

11

WOLF, J. E. JR. (1989) Angiogenesis in normal and psoriatic skin. Lab. Invest. 61: 139-142. YASUNAGA, C., NAlC~,SHIMA,Y. and SUEISHI, K. (1989) A role of fibrinolytic activity in angiogenesis. Quantitative assay using in vitro method. Lab. Invest. 61: 698-704. ZETTER, B. R. (1987) Assay of capillary endothelial cell migration. Meth. Enzym. 147: 135-144. ZETTER, B. R. (1988) Angiogenesis. State of the art. Chest. 93: 159S-166S.