Directed in vivo angiogenesis assay and the study of systemicneoangiogenesis in cancer

Claudio Napoli1,2, Antonio Giordano2,3,4, Amelia Casamassimi1, Francesca Pentimalli4, Louis J. Ignarro5

and Filomena De Nigris1

1 Department of General Pathology, Division of Clinical Pathology, 1st School of Medicine, II University of Naples, Naples, Italy2 Sbarro Institute for Cancer Research and Molecular Medicine, Center for Biotechnology, Temple University, Philadelphia, PA3 Department of Human Pathology and Oncology, University of Siena, 53100, Siena, Italy4 CROM - Center of Oncology Research Mercogliano, ‘‘Fiorentino Lo Vuolo’’, 83013, Mercogliano Avellino, Italy5 Department of Molecular and Medical Pharmacology, David Geffen School of Medicine, University of California, Los Angeles, CA

Targeting neoangiogenesis is a well-established anticancer strategy, however, one of the major problems in angiogenesis

research, both at the basic and applied levels, remains the development of suitable in vivo methods for assessing and

quantifying the systemic angiogenic response. Therefore, there is an urgent need to adopt technically simple and reproducible

methodologies which allow to easily quantify neoangiogenesis independently of morphological parameters. Recently, a

reproducible and quantitative method was developed, the directed in vivo angiogenesis assay (DIVAA) consisting of the

subcutaneous implantation of surgical grade silicone cylinders closed at one end, called angioreactors, into the dorsal flanks

of nude mice. In the past few years, DIVAA has been successfully used in evaluating the inhibition and or enhancement of

systemic perturbation of angiogenesis by several molecules. Thus, DIVAA studies systemic angiogenesis and its therapeutic

modulation associated to cancer progression and metastasis.

It is well-established that cancer progression is accompaniedby vascular events, such as the formation and remodelling ofnovel blood vessels.1,2 It is obvious that tumors need a greatbulk of oxygen and nutrients. For the first one million of ma-lignant cells, tumors get their oxygen and nutrients from thehost capillaries and extracellular fluid.1,2 However, as theyoutgrow the host supply they start making their own bloodvessels (the so-called ‘‘cancer neoangiogenesis.’’ Indeed, ma-lignant cells ‘‘persuade’’ the existing host capillaries to secretevascular endothelial growth factors, to sprout changing direc-

tion and growing throughout the tumor. De novo formedtumoral vessels may provide an entry point for metastasis.The process of metastasis formation per se could be linked toangiogenesis perturbing signalling phenomena. Cancer-associ-ated vessels are different from normal vessels and their endo-thelial cells could provide useful targets for treatment. More-over, both primary tumor and circulating cancer cells mayinfluence and prepare distant sites for engraftment throughchanges in systemic angiogenesis. It remains unclear whethervascular perturbing is simply induced by the conditions oftumor microenvironment, or it is a consequence of geneticchanges underlying the onset and progression of malignancy.

Angioreactors in the study of systemic neoangiogenesis

We considered of primary importance to study the perturba-tion of systemic angiogenesis during tumor growth and metas-tasis, whereas most studies so far have analyzed angiogenesisin the tumor microenvironment, focusing on the events thatlead to the development of a neovascular blood supply to thetumor mass. Thus, we analyzed systemic angiogenesis in amodel of metastatic osteosarcoma.3 We found that proangio-genic molecules released in the blood flow from tumor cellswere able to induce new vessel formation in distal area fromtumor site (systemic blood vessels). We restricted the areawhere systemic vessels grow and can be recovered at the end ofthe experiment without mice sacrifice. After recovering, weobtained a vessel mass with red blood flow, which could bequantified and molecularly analyzed. We compared the num-ber of systemic vessels vs. intrametastatic vessels and we

Key words: angiogenesis, cancer, endothelium

This work is dedicated in memory of Prof. Giovan Giacomo

Giordano, Oncologist (II University of Naples)

Grant sponsor: Progetto di Rilevante Interesse Nazionale Ministero

Italiano Universita e Ricerca 2006 and 2008; Grant number:

0622153_002; Grant sponsor: Meccanismi fisiopatologici di danno

vascolare/trombotico ed angiogenesi; Grant number:

2008T85HLH_002; Grant sponsor: Regolazione dell’espressione

genica della via SIRT1/FoxO1-dipendente in cellule endoteliali

progenitrici della nicchia vascolare

DOI: 10.1002/ijc.25743

History: Received 2 Sep 2010; Accepted 4 Oct 2010; Online 25 Oct

2010

Correspondence to: Claudio Napoli, Department of General

Pathology, Division of Clinical Pathology, 1st School of Medicine, II

University of Naples, Via Luigi de Crecchio, 7, Complesso S. Andrea

delle Dame, Naples 80138, Italy, Tel: þ390815667560, Fax:

390815667564, E-mail: claudio.napoli@unina2.it

MiniReview

Int. J. Cancer: 128, 1505–1508 (2011) VC 2010 UICC

International Journal of Cancer

IJC

observed a strong correlation between them. However, thequantification of systemic blood vessels was more efficient tomonitor small changes in response to genetic changes (such assilencing of the transcription factor YY1, in our case) or phar-macological antiangiogenic treatments in vivo (such as inhibi-tion of the chemokine receptor CXCR4 through the T22 pep-tide). The recovered systemic vessels were totally free fromsurrounding tumor cells, therefore an easy staining and quanti-fication of endothelial cells was possible. In our experimentalsetting, we used a relatively novel reproducible and quantitativemethod, the directed in vivo angiogenesis assay (DIVAA, Tre-vigen).3,4 This assay consists of the subcutaneous implantationof grade silicone cylinders closed at one end, called angioreac-tors, into the dorsal flanks of nude mice. Angioreactors arefilled with a small amount of extracellular matrix (18 ll) pre-mixed with or without modulating angiogenic factors (such asVEGF, FGF-2, etc.) and/or cancer cells. When filled withangiogenic factors vascular endothelial cells migrate and prolif-erate into the matrix of the angioreactor to form vessels. Fewdays after implantation (usually from 9 to 15 days) capillarysprouts originating from the host vessels invade the extracellu-lar matrix and form vessels in the angioreactor so that thereare enough cells to determine an effective dose response toangiogenic factors. Thus, the extent of vascularization withinthe angioreactors can be quantified by the i.v. injection of fluo-rescein isothiocyanate-dextran or lectin before recovery, fol-lowed by spectrofluorimetry. Values for cell invasion areexpressed in relative fluorescent units. The intensity of the sig-nal is proportional to the number of endothelial cells containedin each of the angioreactors. This provides a measurement ofblood volume within the angioreactor. Angioreactors examinedby immunofluorescence show cells and invading angiogenicvessels at different developmental stages. The minimally detect-able angiogenic response requires 9 days after implantation.Control angioreactors only occasionally show cellular invasionof the Matrigel without vascular invasion. Histological exami-nation of the angioreactor is also useful to reveal possible tissuegranulation and infiltrates of mononuclear cells.

In our study,3 we first intravenously injected cancer cellsinto the mice tails. At day 30th after cell injection, mice skinwas incised for 1.5 cm and two angioreactors were implantedsubcutaneously into their dorsal flank. Thus, the same animalhad also the positive control (angioreactor coated with FGF-2). At the end of the experiment the angioreactors were col-lected and the new vessel formation was determined by FITC-lectin staining. The fluorescence was measured in 96 multiwellplates using an HP spectrofluorimeter model (excitation 485nm, emission 510 nm; Perkin-Elmer, Boston, MA). Thus, wedirectly quantified the number of newly formed blood vesselsformed in subcutaneous angioreactors during the metastasisdevelopment (Fig. 1). Hence, by this method we were able toshow that the silencing of YY1 transcription factor in osteo-sarcoma cells could interfere with formation of new vessels.Indeed, mice inoculated with YY1 silenced cells producedfewer vessels with lower lumen than SaOs-2 cells bearing

mice. These experiments indicated that the implantation ofYY1 silenced cells was less effective for activating the tumorproangiogenic microenvironment than SaOs-2 cells.3 More-over, the fact that T22, a CXCR4 blocking peptide, was inef-fective to further reduce the new vessels formation in vivosuggested that YY1 interfered with CXCR4/angiogenesis path-way downstream the receptor activation. Thus, we evaluatedthe long-term effect of drug treatment on systemic neoangio-genesis compared to intrametastatic vascularization.3 As aresult, we studied the perturbing of systemic tumor-inducedneoangiogenesis, and demonstrated a role for the transcriptionfactor YY1 and VEGF/CXCR4 in the pathogenesis of the ma-lignant phenotype of osteosarcoma.3 Our findings are associ-ate to a novel CXCR4/YY1-dependent mechanism in whichangiogenesis is perturbed during malignancy of the bone5

We have also examined lung metastases thus revealingthat YY1 silencing could also reduce the metastatic implanta-tion and tumor growth (i.e., 9 out of 10 SaOs-2 bearing micedeveloped metastases while only 4 out of 10 mice bearingYY1 silenced cells did).3 Consistently, the mean area ofSaOs-2 metastases were 10-fold bigger than that observed inYY1 silenced mice. The administration of the T22 peptidealso reduced significantly the size of metastasis in SaOs-2bearing mice (10-fold) whereas it was ineffective on meta-static size in mice bearing YY1 silenced cells.

Hence, although several in vitro and in vivo assays havebeen developed so far to study tumor-induced angiogenesis,they still show some limitations (high costs, technical difficul-ties, most methods rely on selective quantitative and morpho-metric analysis, and require substantial amounts of test com-pound).6 Compared to similar in vivo assays, the use ofangioreactors shows some advantages. For instance, thismethod reduces assay errors mainly due to absorption of thematrix by the mouse. Indeed, the geometry and configurationof the angioreactor maintains constant and constrained matrixvolume, fixed concentration of angiogenic factor(s), as well asdefined local concentration of test compounds. Furthermore,compared to other in vivo assays, only a small amount of testsubstance is used in each angioreactor and up to four angior-eactors can be implanted in the same mouse thus giving moredata for analysis.4,7 To this regard, it is also possible to reducethe angiogenesis response variation among mice by introduc-ing appropriate controls in each mouse. Moreover, this systempermits to perform a time course of systemic angiogenesisoccurring during tumor progression or drug treatment in thesame animal, and to study several parameters of these newvessels. The DIVAA technique also allows material containedin the angioreactor to be recovered for additional biochemical,cellular, or molecular analysis. Noteworthy, it is also an ob-server-independent, reproducible and quantitative assay.7 Inthe past few years, DIVAA has been successfully used in eval-uating both the role of several genes and their products in themultistep process of neoangiogenesis and also the potentialinhibition and/or enhancement of angiogenesis by a numberof novel synthetic molecules.8–20 Basically, there are several

MiniReview

1506 Systemic neoangiogenesis in cancer

Int. J. Cancer: 128, 1505–1508 (2011) VC 2010 UICC

pathophysiological mechanisms linked to tumoral neoangio-genesis8–15 and therapeutic approaches using already marketeddrugs16,17 or new pipeline compounds.18–20 For instance,DIVAA analysis was used to demonstrate the role of adreno-medullin secreted from mast cells as a proangiogenic factor,8,9

as well a critical role of cripto-1 in regulating new blood ves-sel formation.10 Moreover, angioreactors were useful also toclarify the mechanism by which a member of a family of met-alloproteinases, the membrane Type 1-MMP (MT1-MMP,also called MMP14), controls tumor-induced angiogenesis.11

A further study involving the use of DIVAA demonstratedthat the member of the inhibitor of apoptosis protein family,Survivin DEx3, is involved in neoangiogenesis.12 Similarly,CD97, which is highly expressed on various inflammatorycells and some carcinomas, has been shown to act as a proan-giogenic factor and contributes to inflammation-mediatedangiogenesis.13 Furthermore, a small GTPase, Rap1b, has beenidentified as a positive regulator of angiogenesis with putativeinvolvement in all angiogenesis-related diseases including can-cer.14 On the other side, it has been recently utilized to dem-onstrate the role of aryl hydrocarbon receptor repressor as aputative tumor suppressor gene with antiangiogenic potential

in several translational models of human cancer.15 DIVAA hasbeen used successfully also to test antiangiogenesis factors, suchas fasudil,16 bevacizumab,17 anthrax lethal toxin (a MMP-acti-vated antitumor toxin),18 pentastatin-119 and CC-507920 in pre-clinical in vivo models of the disease. Although in the vast ma-jority of these studies tumor xenografts were used to analyzeand measure the loco-regional tumor-induced angiogenesisDIVAA has been also used as a disease independent assay, asin the study where the high potency of pentastatin-1 was dem-onstrated thereby suggesting a strong potential for applicationto various angiogenesis-dependent disease models, includingcancer.19 Interestingly, a recent paper showed that CC-5079, asmall molecule inhibitor of tubulin polymerization and phos-phodiesterase-4 activity, inhibited microvessel formation in alldifferent ex vivo and in vivo assay systems of angiogenesisused: the CAM, the rat aortic ring, and the DIVAA.20 Thesestudies support the concept that DIVAA may constitute a validpreclinical tool to test therapeutic efficiency against cancer-asso-ciated neoangiogenesis.

Advanced cancer is also associated with several alterationsin the systemic circulation, such as alterations of the coagula-tion cascade, and particularly with a hypercoagulable state,21

Figure 1. Schematic representation of DIVAA in the study of loco-regional and systemic angiogenesis. The mouse designed on the left of

the figure is directly injected with tumoral cells in the tail to give lung metastasis. Angioreactors are implanted subcutaneously into the

dorsal flanks of the mouse and vessels allowed to infiltrate. On the lower panel representative pictures of the implantation procedure and

angioreactors after recovering are also showed. On the right, loco-regional angiogenesis surrounding the lung metastasis can be analysed

directly through angioreactors analysis after mouse sacrifice together with perturbed signalling linked systemic neoangiogenesis.

MiniReview

Napoli et al. 1507

Int. J. Cancer: 128, 1505–1508 (2011) VC 2010 UICC

whose contribution in promoting angiogenesis could be bet-ter understood with the aid of this in vivo system.

ConclusionsThus, the study of systemic neoangiogenesis, through the useof angioreactors or alternative similar methods, which allowa quantitative and reproducible analysis, might shed light onthe processes of neoangiogenesis, cancer progression and me-

tastasis. Thus, DIVAA may help to develop therapeutic tar-geting of cancer-associated neaoangiogenesis.

AcknowledgementsSupport from the Progetto di Rilevante Interesse Nazionale Ministero Ital-iano Universita e Ricerca 2006 and 2008 to the II University of Naples(C.N.), Fondation Jerome Leyeune (Paris, France) (C.N.) and Regione Cam-pania 2008 (C.N) was gratefully acknowledged.

References

1. Folkman J Angiogenesis: an organizingprinciple of drug discovery? Nat Rev DrugDiscov 2007;6:273–86.

2. Folkman J. Angiogenesis. Annu Rev Med2006;57:1–18.

3. de Nigris F, Rossiello R, Schiano C, ArraC, Williams-Ignarro S, Barbieri A, LanzaA, Balestrieri A, Giuliano MT, Ignarro LJ,Napoli C. Deletion of Yin Yang 1 proteinin osteosarcoma cells on cell invasion andCXCR4/angiogenesis and metastasis.Cancer Res 2008;68:1797–1808.

4. Guedez L, Rivera AM, Salloum R, Miller ML,Diegmueller JJ, Bungay PM, Stetler-StevensonWG. Quantitative assessment of angiogenicresponse by the directed in vivo angiogenesisassay. Am J Pathol 2003;162:1431–1439.

5. de Nigris F, Crudele V, Giovane A,Casamassimi A, Giordano A, Garban HJ,Cacciatore F, Pentimalli F, Marquez-Garban DC, Petrillo A, Cito L, Sommese L,et al. C. CXCR4/YY1 inhibition impairsVEGF network and angiogenesis duringmalignancy. Proc Natl Acad Sci USA 2010;107:14484–9.

6. Balestrieri ML, Napoli C. Novel challengesin exploring peptide ligands andcorresponding tissue-specific endothelialreceptors. Eur J Cancer 2007;43:1242–50.

7. Norrby K. In vivo models of angiogenesis.J Cell Mol Med 2006;10:588–612.

8. Zudaire E, Martınez A, Garayoa M, Pıo R,Kaur G, Woolhiser MR, Metcalfe DD,Hook WA, Siraganian RP, Guise TA,Chirgwin JM, Cuttitta F. Adrenomedullinis a cross-talk molecule that regulatestumor and mast cell function duringhuman carcinogenesis. Am J Pathol 2006;168:280–91.

9. Martınez A, Zudaire E, Portal-Nunez S,Guedez L, Libutti SK, Stetler-StevensonWG, Cuttitta F. Proadrenomedullin NH2-Terminal 20 Peptide Is a potent angiogenicfactor, and its inhibition results inreduction of tumor growth. Cancer Res2004;64:6489–94.

10. Bianco C, Strizzi L, Ebert A, Chang C,Rehman A, Normanno N, Guedez L,Salloum R, Ginsburg E, Sun Y, Khan N,Hirota M, et al. Role of human Cripto-1 intumor angiogenesis. J Natl Cancer Inst2005;97:132–41.

11. Basile JR, Holmbeck K, Bugge TH,Gutkind JS. MT1-MMP Controls tumor-induced angiogenesis through the release ofsemaphorin 4D. J Biol Chem 2007;282:6899–905.

12. Caldas H, Fangusaro JR, Boue DR,Holloway MP, Altura RA. Dissecting therole of endothelial SURVIVIN {Delta}Ex3in angiogenesis. Blood 2007;109:1479–89.

13. Wang T, Ward Y, Tian L, Lake R, GuedezL, Stetler-Stevenson WG, Kelly K. CD97,an adhesion receptor on inflammatorycells, stimulates angiogenesis throughbinding integrin counterreceptors onendothelial cells. Blood 2005;105:2836–44.

14. Chrzanowska-Wodnicka M, Kraus AE,Gale D, White GC 2nd, Vansluys J.Defective angiogenesis, endothelialmigration, proliferation, and MAPKsignaling in Rap1b-deficient mice. Blood2008;111:2647–56.

15. Zudaire E, Cuesta N, Murty V, WoodsonK, Adams L, Gonzalez N, Martınez A,Narayan G, Kirsch I, Franklin W, HirschF, Birrer M, et al. The aryl hydrocarbonreceptor repressor is a putative tumor

suppressor gene in multiple humancancers. J Clin Invest 2008;118:640–50.

16. Yin L, Morishige K, Takahashi T,Hashimoto K, Ogata S, Tsutsumi S, TakataK, Ohta T, Kawagoe J, Takahashi K,Kurachi. H. Fasudil inhibits vascularendothelial growth factor-inducedangiogenesis in vitro and in vivo. MolCancer Ther 2007;6:1517–25.

17. Mabuchi S, Terai Y, Morishige K, Tanabe-Kimura A, Sasaki H, Kanemura M,Tsunetoh S, Tanaka Y, Sakata M, BurgerRA, Kimura T, Ohmichi M. Maintenancetreatment with bevacizumab prolongssurvival in an in vivo ovarian cancermodel. Clin Cancer Res 2008;14:7781–9.

18. Liu S, Wang H, Currie BM, Molinolo A,Leung HJ, Moayeri M, Basile JR, AlfanoRW, Gutkind JS, Frankel AE, Bugge TH,Leppla SH. Matrix metalloproteinase-activated anthrax lethal toxin demonstrateshigh potency in targeting tumorvasculature. J Biol Chem 2008;283:529–40.

19. Koskimaki JE, Karagiannis ED, Tang BC,Hammers H, Watkins DN, Pili R, PopelAS. Pentastatin-1, a collagen IV derived20-mer peptide, suppresses tumor growthin a small cell lung cancer xenograftmodel. BMC Cancer 2010;10:29.

20. Vu HN, Miller WJ, O’Connor SA, He M,Schafer PH, Payvandi F, Muller GW,Stirling DI, Libutti SK. CC-5079: a smallmolecule with MKP1, antiangiogenic, andantitumor activity. J Surg Res 2010;164:116–125.

21. Rak J, Milsom C, Yu J. Vasculardeterminants of cancer stem celldormancy—do age and coagulation systemplay a role? APMIS 2008;116:660–76.

MiniReview

1508 Systemic neoangiogenesis in cancer

Int. J. Cancer: 128, 1505–1508 (2011) VC 2010 UICC