Proadrenomedullin NH2-Terminal 20 Peptide Is a Potent Angiogenic Factor, and Its Inhibition Results in Reduction of Tumor Growth

[CANCER RESEARCH 64, 6489–6494, September 15, 2004]

Alfredo Martõ«nez,1,3 Enrique Zudaire,1,3 Sergio Portal-Nu«n˜ez,1,3 Liliana Gue«dez,1,3 Steven K. Libutti,2,3 William G. Stetler-Stevenson,1,3 and Frank Cuttitta1,3 1Cell and Cancer Biology Branch, 2Surgery Branch, 3Vascular Biology Faculty, National Cancer Institute, NIH, Bethesda, Maryland

ABSTRACT Despite the growing interest in the angiogenic effects of adrenomedullin and its influence in tumor biology, no attention has been We have found through ex vivo and in vivo angiogenesis models that the paid to the actions of proadrenomedullin NH -terminal 20 peptide adrenomedullin gene-related peptide, proadrenomedullin NH -terminal 2 2 (PAMP) in this field, and we decided to investigate them. Here, we 20 peptide (PAMP), exhibits a potent angiogenic potential at femtomolar concentrations, whereas classic angiogenic factors such as vascular endo- report that PAMP is a very potent angiogenic factor, able to induce thelial growth factor and adrenomedullin mediate a comparable effect at neovascularization in animal models at concentrations 6 orders of nanomolar concentrations. We found that human microvascular endothe- magnitude lower than other classic proangiogenic factors such as lial cells express PAMP receptors and respond to exogenous addition of vascular endothelial growth factor (VEGF) and adrenomedullin. We PAMP by increasing migration and cord formation. Exposure of endo- demonstrate that human microvascular endothelial cells have recep- thelial cells to PAMP increases gene expression of other angiogenic factors tors for PAMP and respond to it by increasing migration and cord such as adrenomedullin, vascular endothelial growth factor, basic fibro- formation in Matrigel assays. In addition, PAMP stimulation induces blast growth factor, and platelet-derived growth factor C. In addition, the expression of classic angiogenic factors in endothelial cells. We use a peptide fragment PAMP(12-20) inhibits tumor cellÐinduced angiogenesis fragment of PAMP that specifically inhibits PAMP-induced effects as in vivo and reduces tumor growth in xenograft models. Together, our data a tool to define for the first time the important role of PAMP in demonstrate PAMP to be an extremely potent angiogenic factor and implicate this peptide as an attractive molecular target for angiogenesis- angiogenesis. This fragment acts as an inhibitor of tumor cell–induced based antitumor therapy. angiogenesis and is able to delay tumor growth in xenograft models of tumor progression. INTRODUCTION The formation of new blood vessels from preexisting ones, or MATERIALS AND METHODS angiogenesis, is necessary for normal physiology in processes such as Chemicals. Synthetic human adrenomedullin, PAMP, and PAMP(12-20) embryogenesis, growth, wound healing, and the endometrial cycle. were purchased from Bachem (King of Prussia, PA). Recombinant human However, when the balance between proangiogenic and antiangio- VEGF and basic fibroblast growth factor (bFGF) were obtained from R&D genic regulators gets distorted, the resulting lack or excess of blood Systems (Minneapolis, MN). vessels can lead to numerous pathological conditions (1). Angiogen- Chick Embryo Aortic Arch Assay. The chick embryo aortic arch assay is esis is a multistep process that requires at least endothelial cell an ex vivo angiogenesis assay that was performed as described previously (14, proliferation, production of extracellular proteases, migration of en- 15). In brief, aortic rings of approximately 0.8 mm in length were prepared dothelial cells, tube formation, development of loops, and recruitment from the five aortic arches of 13-day-old chicken embryos (CBT Farms, of pericytes and smooth muscle cells for larger vessels (2). Chestertown, MD), and the soft connective tissue of the adventitia layer was Like normal tissues, tumors require an adequate blood supply to carefully removed with tweezers. Each aortic ring was placed in the center of a well in a 48-well plate and covered with 10 ␮L of Matrigel (BD Biosciences, maintain their metabolic needs. As the tumor grows, the cells located San Jose, CA). After the Matrigel solidified, 300 ␮L of growth factor–free at the center of the tissue mass experience hypoxia, a physiologic human endothelial-SFM basal growth medium (Invitrogen, Carlsbad, CA) stimulus for the expression and secretion of numerous proangiogenic containing the proper concentration of the test substances were added to each factors, which in turn will promote enough blood vessel formation to well. The plates were kept in a humid incubator at 37°C in 5% CO2 for 24 to support additional tumor growth (3). Hence the hypothesis that de- 36 hours. Microvessels sprouting from each aortic ring were photographed in priving tumors from their proangiogenic factors would stop tumor an inverted microscope, and the area covered by the newly formed capillaries progression (4, 5). In the last few years, antiangiogenic therapies have was estimated as reported previously (14). been found to successfully delay tumor growth in animals, and several Directed In vivo Angiogenesis Assay. Analysis and quantitation of angio- antiangiogenic factors are currently undergoing clinical trials (6, 7). genesis was done using directed in vivo angiogenesis assay as described The regulatory peptide adrenomedullin is a multifunctional mole- previously (10, 16). In brief, 10-mm-long surgical-grade silicone tubes with only one end open (angioreactors) were filled with 20 ␮L of Matrigel alone or cule (8) that has been recently characterized as a proangiogenic factor mixed with adrenomedullin, bFGF, VEGF, PAMP, and/or PAMP(12-20) at the with the help of ex vivo and in vivo animal models (9–12; among indicated concentrations. Human lung cancer cell lines (see below) were also others). In addition to inducing angiogenesis, adrenomedullin func- premixed with Matrigel alone or in combination with PAMP(12-20) at 10,000 tions in cancer cells as an autocrine growth factor, enhances thymidine cells per angioreactor. After the Matrigel solidified, the angioreactors were incorporation, reduces apoptosis, and is induced by hypoxia, therefore implanted into the dorsal flanks of athymic nude mice (NCI colony). After 11 suggesting that this peptide may be an important tumor cell survival days, the mice received i.v. injections of 25 mg/mL FITC-dextran (100 factor and a potential target for antitumor therapy (13). ␮L/mouse; Sigma, St. Louis, MO) 20 minutes before removing angioreactors. Photographs of the implants were taken for visual examination of angiogenic response. Quantitation of neovascularization in the angioreactors was deter- Received 1/13/04; revised 6/28/04; accepted 7/12/04. The costs of publication of this article were defrayed in part by the payment of page mined as the amount of fluorescence trapped in the implants and was measured charges. This article must therefore be hereby marked advertisement in accordance with in a HP Spectrophotometer (Perkin-Elmer, Boston, MA). This protocol was 18 U.S.C. Section 1734 solely to indicate this fact. approved by the internal NIH animal committee. Requests for reprints: Alfredo Martı´nez, Cell and Cancer Biology Branch, National The human cancer cell lines used, A549 and H1299, were obtained from the Cancer Institute, NIH, Building 10, Room 13N262, Bethesda, MD 20892. Phone: 301- 402-3308; Fax: 301-435-8036; E-mail: [email protected]. American Tissue Culture Collection (Manassas, VA) and fed with RPMI 1640 ©2004 American Association for Cancer Research. containing 10% fetal bovine serum (Invitrogen). Before they were used in 6489

animals, both cell lines were tested for a panel of human and murine pathogens intratumoral injection, according to their group. Group 1 (control) received 100 and found to be pathogen-free. ␮L of PBS; group 2 received 100 ␮L of 10 nmol/L PAMP(12-20) in PBS; and Ca2؉ Measurements. Human dermal microvascular endothelial cells were group 3 received 100 ␮Lof1␮mol/L PAMP(12-20) in PBS. When the tumor cultured in 96-well plates at 1.0 ϫ 105 cells per well. The cells were loaded for burden became unbearable (larger than 2000 mm3), the mice were sacrificed. 60 minutes at room temperature with the fluorescent dye FLIPR (Molecular This schedule was chosen based on previous antiangiogenic treatments in Devices, Sunnyvale, CA) and then transferred to the FlexStation II (Molecular xenograft studies (20). In a second experiment, another 30 mice that received Devices) for analysis. The test compounds were prepared in another plate at injections of A549 cells were divided in three groups. Group 1 (control) 5ϫ concentration and added to the proper wells by the robotic arm of the received 100 ␮L of PBS; group 2 received 100 ␮L of 100 nmol/L PAMP in FlexStation II. Fluorescence was measured every 5 seconds in each well and PBS; and group 3 received 100 ␮L of 1 nmol/L PAMP in PBS. As with the recorded. ATP (1 mmol/L; Sigma) was used as a Ca2ϩ influx agonist (17). other animal experiments, these studies were done under a NIH-approved Proliferation Assay. The same microvascular endothelial cells were protocol. seeded in 96-well plates at a density of 2.0 ϫ 105 cells per well in serum-free Statistical Analysis. When appropriate, data were compared by two-tailed medium containing different concentrations of the test peptides. After 3 days Student’s t test. P values lower than 0.05 were considered statistically signif- in culture, the number of viable cells per well was estimated by the 3-(4,5- icant. dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay as reported previously (10). Results are represented as the percentage of growth over the RESULTS untreated control. Migration Assay. Cell motility was measured as described previously (10). Angiogenic Potential of Proadrenomedullin NH2-Terminal 20 Test peptides were placed at various concentrations at the bottom of a Peptide in Ex vivo and In vivo Assays. There are many ways of ChemoTx chamber (NeuroProbe, Inc., Gaithersburg, MD). The intermediate testing angiogenesis (15), and we chose a chick embryo aortic arch ␮ membrane was coated with 10 g/mL fibronectin, and in the upper chamber, assay for a preliminary assessment of the angiogenic properties of 5.0 ϫ 105 human endothelial cells were added. After a 4-hour incubation at PAMP. Angiogenic factors such as adrenomedullin and VEGF in- 37°C, the membrane was fixed and stained (Protocol Hema3; Biochemical Sciences, Inc., Bridgeport, NJ). The cells trapped in the porous membrane were duced a statistically significant increase of the number of sprouting photographed through a ϫ25 microscope objective, and the number of cells per blood vessels over untreated controls at concentrations of 100 nmol/L photographic field was counted. and higher (results not shown). PAMP induced comparable growth of Cord Formation Assay. Human endothelial cells were seeded at 2.0 ϫ 105 newly formed blood vessels at concentrations as low as 1 nmol/L (Fig. cells per well over a solid layer of Matrigel covering the bottom of a 24-well 1D). At this concentration (1 nmol/L), neither adrenomedullin (Fig. plate in the presence or absence of the test peptides, as described previously 1B) nor VEGF (Fig. 1C) caused any significant proliferation over the (18). After an overnight incubation, the tubular structures were photographed, control (Fig. 1A). The sprouting new vessels covered a surface of and the number of knots per photographic field were counted as a measure of 2,579,500 Ϯ 378,300 ␮m2 after treatment with PAMP, which is about lattice complexity. 10 times the surface covered by the primitive capillaries formed in the Real-Time Polymerase Chain Reaction Quantification of Gene Expres- untreated control (261,500 Ϯ 60,800 ␮m2; P Ͻ 0.001). This initial sion. Human endothelial cells were cultured in T-75 flasks until they reached a density of approximately 2.5 ϫ 106 cells per flask. Cells were treated with observation suggests that PAMP may be a more potent proangiogenic 10 nmol/L PAMP in serum-free medium for 24 hours. Total RNA was factor than previously described molecules. extracted using the RNeasy Mini kit from Qiagen (Valencia, CA) and reverse To further characterize this intriguing observation, we used an in transcribed using the SuperScript First-Strand Synthesis system (Invitrogen). vivo assay to more precisely quantitate the angiogenic properties of Quantification of gene expression was performed by real-time PCR as de- PAMP. This approach is called directed in vivo angiogenesis assay scribed previously (19). The PCR reaction was run in an Opticon cycler (MJ (10, 16). In this assay, small silicone capsules carrying the test Research, Waltham, MA) using Sybr Green PCR master mix (Applied Bio- substances are implanted under the skin of nude mice. After 11 days, systems, Foster City, CA). Thermocycling was performed in a final volume of 25 ␮L containing 2 ␮L of cDNA (1:10 dilution) and 400 nmol/L of primers (see below). All targets were amplified in triplicate in the same run as the house-keeping gene, using the following cycle scheme: After initial denatur- ation of the samples at 95°C for 2 minutes, 46 cycles of 95°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds were performed. Fluorescence was measured in every cycle, and mRNA levels were normalized by the 18S RNA values in all samples. A melting curve was run after PCR by increasing the temperature from 60°Cto96°C (0.5°C increments). A single peak was obtained for all amplicons, thus confirming the specificity of the reaction. Primers were as follows: adrenomedullin forward, ACA TGA AGG GTG CCT CTC GAA; adrenomedullin reverse, AGG CCC TGG AAG TTG TTC ATG; VEGF forward, TCA GAG CGG AGA AAG CAT TTG T; VEGF reverse, TCG GCT TGT CAC ATC TGC AA; bFGF forward, CGA CCC TCA CAT CAA GCT ACA AC; bFGF reverse, CCA GTT CGT TTC AGT GCC ACA T; platelet-derived growth factor (PDGF) A forward, TTC GGA GGA AGA GAA GCA TCG; PDGF A reverse, GCA CTT GAC ACT GCT CGT GTT G; PDGF B forward, AAC AAC CGC AAC GTG CAG T; PDGF B reverse, TCT CGA TCT TTC TCA CCT GGA C; PDGF C forward, TTG AGG AAC CCA GTG ATG GAA C; PDGF C reverse, CAG CTT CTG TGA ATT GTG GCA T; 18 S forward, ATG CTC TTA GCT GAG TGT CCC G; and 18 S reverse, ATT CCT AGC TGC GGT ATC CAG G. Xenograft Experiments. For the first xenograft experiment, 30 female athymic nude mice from the NIH colony in Frederick, Maryland, received s.c. Fig. 1. Comparative angiogenic potential of adrenomedullin, VEGF, and PAMP in the injections of 1.0 ϫ 107 A549 cells per mouse. Two weeks later, all of the mice chick embryo aortic ring assay. Aortic rings were embedded in Matrigel and exposed to had developed palpable tumors under the skin, and at this time, they were serum-free medium containing either PBS as a negative control (A) or 1 nmol/L of the peptides adrenomedullin (B), VEGF (C), or PAMP (D). Only in the case of PAMP is the randomly divided in three groups. Three times a week, each individual tumor crown of sprouting new vessels significantly larger than the control. Representative was measured (length, height, and thickness), and every mouse received an examples of four repeats. Bar ϭ 0.5 mm. 6490

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2004 American Association for Cancer Research. PAMP INDUCES ANGIOGENESIS the mice received injections of a specific amount of FITC-dextran, and the volume of blood circulating through the implant is quantified by measuring the fluorescence in the capsule. In addition, the new blood vessels growing into the silicone tube can be seen directly by transparency (Fig. 2, A–F). PAMP elicited a measurable angiogenic response at concentrations as low as 1 fmol/L (Fig. 2C), which was clearly distinct from the negative control (Fig. 2A). The angiogenic response elicited by PAMP was dose-dependent (Fig. 2, C–G). When compared with adrenomedullin and VEGF responses at equimolar concentrations, a clearly significant difference was observed over a wide concentration range. In this animal model, adrenomedullin and VEGF induce angiogenesis at nanomolar concentrations, whereas PAMP was already active in the femtomolar range (Fig. 2G). ϩ Proadrenomedullin NH -Terminal 20 Peptide Reduces Ca2؉ Fig. 3. PAMP reduces Ca2 flux in human microvascular endothelial cells. Addition 2 of 1 mmol/L ATP induces an increase in Ca2ϩ flux in endothelial cells (Ⅺ), whereas Flux in Endothelial Cells. Although the adrenomedullin receptor has addition of 10 nmol/L full-length PAMP greatly reduces this effect (छ). This blocking been well characterized at the molecular level (21), the structure of the effect was reversed by 100 nmol/L PAMP(12-20) (E). Representative example of five PAMP receptor is not yet available; but it has been shown that repeats. R.F.U., relative fluorescence units. exposure of adrenal medulla cells to PAMP results in a decrease of carbachol-induced Ca2ϩ influx (22). To demonstrate whether a similar biological response takes place in endothelial cells, we stimulated by the presence of 10 nmol/L PAMP in the medium (Fig. 3, dia- them with 1 mmol/L ATP to induce a Ca2ϩ influx (Fig. 3, squares)as monds). The peptide fragment PAMP(12-20) has been shown to have reported previously (17). This response was greatly reduced (Ͼ50%) opposite actions to full-length PAMP (23), suggesting its specific effect as a PAMP antagonist. To demonstrate the specificity of the PAMP-induced inhibition in Ca2ϩ influx, we added an excess of PAMP(12-20) and were able to recover the initial response (Fig. 3, circles). Taken together, these data show that there is a functional PAMP receptor in the membrane of the endothelial cells, and therefore this peptide may activate directly the angiogenic response described above.

Physiologic Effects of Proadrenomedullin NH2-Terminal 20 Peptide on Endothelial Cells. For angiogenesis to occur, endothelial cells have to proliferate, migrate into new locations, and organize themselves into solid cords that eventually will develop into hollow tubes (2). These processes are promoted by proangiogenic substances, and all proangiogenic molecules must elicit at least one of these physiologic actions. To investigate which of these phenomena are induced by PAMP, we exposed human dermal microvascular endothelial cells to increasing concentrations of PAMP, adrenomedullin, and VEGF and compared their effects on growth (Fig. 4A), migration (Fig. 4B), and cord formation (Fig. 4C). Endothelial cell growth analysis showed that adrenomedullin and VEGF significantly increase proliferation over the control at a concentration of 10Ϫ8 mol/L (P Ͻ 0.001 for both). However, PAMP did not significantly enhance cell growth at the concentrations tested (Fig. 4A). We tested the migratory potential of endothelial cells over a range of peptide concentrations. Adrenomedullin did not modify cell migration significantly on the concentration range tested, whereas both VEGF and PAMP produced a 4-fold increase in migration at a concentration of 10Ϫ11 mol/L when compared with untreated controls (P Ͻ 0.001 for both). As previously reported for VEGF (24), both VEGF and PAMP showed a peak of migration stimulation at 10Ϫ11 mol/L. The shape of the peak was different for both peptides: more Fig. 2. Angiogenic potential in vivo of adrenomedullin (AM), VEGF, and PAMP as steep for PAMP and more gradual for VEGF (Fig. 4B). compared by the directed in vivo angiogenesis assay. Silicone capsules containing differ- We also tested the ability of adrenomedullin, VEGF, and PAMP to ent concentrations of the peptides in Matrigel were implanted under the skin of nude mice induce cord formation in a Matrigel assay. All three factors were able for 11 days. A to F, photographs of the angioreactors still attached to the skin at the end of the experiment. The new blood vessels can be seen growing from the tube opening. The to induce statistically higher cord formation in a dose-dependent capsules contain PBS as a negative control (A), 7 ␮mol/L bFGF as a positive control (B), manner when compared with untreated controls. VEGF was the most Ϫ15 Ϫ13 Ϫ11 Ϫ9 10 mol/L PAMP (C), 10 mol/L PAMP (D), 10 mol/L PAMP (E), and 10 efficient mediator, followed by adrenomedullin, whereas PAMP had a mol/L PAMP (F). Bar ϭ 2.0 mm. G. The Matrigel plugs were digested with dispase, and the FITC-dextran contents were measured to quantify the volume of blood circulating modest effect on cord formation (Fig. 4C). through the implants. Each bar represents the mean and SD of six independent values. We also studied the effects of exogenously added PAMP on the Statistically significant differences between implants treated with PAMP and VEGF at the -gene expression for other proangiogenic molecules (Fig. 4D). Real ,ءءء ;Ͼ P Ͼ 0.001 0.01 ,ءء ;P Ͻ 0.05 ,ء .ء same concentration are represented by P Ͻ 0.001. time PCR experiments showed that PAMP modestly induces the 6491

sponse (Fig. 5B) that was completely blocked by 100 nmol/L PAMP(12-20) (P Ͼ 0.05, comparing tumor cells plus PAMP(12-20) with Matrigel only control), indicating that PAMP signaling is some- how necessary for initiating angiogenesis. For further evaluation of in vivo effects of PAMP, we designed a xenograft experiment. The human lung cancer cell line A549 was injected under the skin of 30 athymic nude mice, and after 2 weeks, all animals developed palpable tumor masses at the injection site. These mice were divided into three groups, and each set received a different treatment three times a week. The control group was treated with the vehicle (PBS), and the tumor mass kept increasing until the mice had to be sacrificed 18 days after treatment began (Fig. 6, squares). The group of animals whose tumors received injections of 10 nmol/L PAMP(12-20) did not show any significant difference with the PBS-treated group (results not shown). In contrast, the mice that received 1 ␮mol/L PAMP(12-20) showed a slower rate of tumor growth (Fig. 6, diamonds). Statistical differences in tumor size between the groups were observed after 9 days of treatment and con- tinued for the rest of the experiment. We also investigated whether injection of full-length PAMP had any impact in xenograft tumor growth. At the concentrations tested (1 and 100 nmol/L PAMP), no difference was observed between tumors treated with peptide or vehicle (results not shown).

Fig. 4. Influence of adrenomedullin (AM), VEGF, and PAMP in the physiology of DISCUSSION endothelial cells. A. Human microvascular endothelial cells were seeded in 96-well plates in the presence of increasing concentrations of the angiogenic peptides in serum-free medium. After 3 days in culture, the number of viable cells was estimated by an Here, we have demonstrated that the adrenomedullin gene-related 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay and expressed as per- peptide PAMP is a potent angiogenic factor that is active at concen- cent growth over untreated controls. Each point represents the mean and SD of eight independent measurements. B. The same cells were seeded in the top chamber of a ChemoTx microplate coated with fibronectin in which different concentrations of the peptides had been placed in the bottom chambers. Each point represents the mean and SD of four independent measurements. C. Endothelial cells were seeded over a layer of solidified Matrigel in the presence of different concentrations of angiogenic peptides. The complexity of the cord network was estimated by the number of knots per microscopic statistically ,ء .field. Each bar represents the mean and SD of three independent wells ,ءءء ;Ͼ P Ͼ 0.001 0.01 ,ءء ;P Ͻ 0.05 ,ء .significant differences with the untreated cells P Ͻ 0.001. D. The same cells were exposed to 10 nmol/L PAMP overnight, and their contents in mRNA of angiogenic molecules were measured by real-time PCR. Each bar represents the ratio between the gene of interest in the treated cells and the contents in the untreated cells. The horizontal line indicates no change over untreated conditions. expression of its own gene (adrenomedullin/PAMP), about 50% over basal levels. PAMP was also capable of elevating the expression of VEGF, bFGF, and PDGF C by 80, 300, and 300%, respectively. Conversely, no significant modification in the expression of PDGF A or PDGF B was observed (Fig. 4D). The addition of 10 nmol/L adrenomedullin or VEGF to endothelial cells had no effect on the expression of the adrenomedullin/PAMP gene (results not shown).

A Proadrenomedullin NH2-Terminal 20 Peptide Antagonist In- hibits Angiogenesis In vivo. To further evaluate the role of PAMP in angiogenesis, we evaluated its influence in tumor growth, using PAMP(12-20) as a PAMP-receptor antagonist. First, we studied the competition between synthetic full-length PAMP at 1 nmol/L concentration and increasing doses of PAMP(12-20) in the directed in vivo angiogenesis assay. We observed a dose-dependent inhibition of the angiogenic response elicited by PAMP (Fig. 5A). A 100-fold excess of the peptide fragment (100 nmol/L) inhibited angiogenesis to the basal levels (P Ͼ 0.05, comparing 100 nmol/L of the peptide fragment with Fig. 5. Characterization of PAMP(12-20) as an inhibitor of PAMP-induced angiogen- untreated control). The peptide fragment by itself did not modify basal esis. A. Different concentrations of full-length PAMP and PAMP(12-20) were added to angiogenesis (last bar in Fig. 5A). angioreactors and inserted under the skin of nude mice following the directed in vivo Because tumor cells produce many angiogenic factors (25), we angiogenesis assay protocol as indicated. Each bar represents the mean and SD of six independent measurements. Statistical differences with PAMP alone (first bar) are rep- P Ͻ 0.01). B. PAMP(12-20) was also able to inhibit the) ءء P Ͻ 0.05) or) ء evaluated the contribution of PAMP to the total angiogenic response. resented as Two human non-small cell lung cancer cell lines (A549 and H1299) angiogenesis induced by two human lung cancer tumor cells embedded in the Matrigel capsules. PAMP(12-20) was added at 100 nmol/L. Each bar represents the mean and SD were embedded in Matrigel and placed in the directed in vivo angio- of six independent measurements. Statistical differences with untreated cells are repre- .(P Ͻ 0.05) ء genesis assay. Both cell lines induced a significant angiogenic resented as 6492

renomedullin/PAMP gene is one of the many genes that is induced by low oxygen tension through a mechanism involving the transcription factor HIF-1 (32). Under hypoxic conditions, the alternative splicing mechanism of the gene favors the production of the longer mRNA species that produces PAMP but not adrenomedullin, thus increasing the relative amount of PAMP that is secreted from the cell (27). This regulatory mechanism is consistent with the angiogenic activity that we have described for PAMP. How Many Angiogenic Factors Are Necessary to Induce Neo- vascularization? Although tumor cells and their environment produce more than one angiogenic factor at a time, it is surprising that the inhibition of a single factor can result in complete abrogation of angiogenesis (33, 34). Angiogenesis may be composed of an interde- pendent chain of events that when one of those individual steps is blocked, the whole process has to stop. The profound effects of PAMP(12-20) on human lung tumor cells in the directed in vivo angiogenesis assay and the xenograft experiments is consistent with Fig. 6. Antitumoral effect of PAMP(12-20) in a xenograft model. Athymic nude mice this proposed hypothesis. Our finding that PAMP induces the expres- Ⅺ ␮ छ carrying A549 tumors were treated with either PBS ( )or1 mol/L PAMP(12-20) ( ), sion of several proangiogenic genes in endothelial cells is also in statistical ,ء .three times a week. Each point represents the mean and SD of 10 mice P Ͻ 0.01. agreement with that hypothesis. This would suggest that PAMP has ,ءء ;P Ͻ 0.05 ,ء .differences of tumor volume between treatments both a direct effect on endothelial cells by binding to its receptor and lowering Ca2ϩ flux and an indirect effect by inducing production of trations 6 orders of magnitude lower than previously recognized other angiogenic promoters. angiogenic molecules such as adrenomedullin and VEGF. In addition, In summary, our experimental findings have revealed PAMP to be we have shown that a receptor for PAMP is present in human an extremely potent angiogenic factor with activity several orders of endothelial cells and that these cells react to the presence of PAMP by magnitude higher than the previously established “gold standard,” increasing their migration and cord formation potential, at the same VEGF. We have shown that the peptide fragment PAMP(12-20) time that expression for other angiogenic factors is boosted. We also functions as an antagonist that can suppress in vivo angiogenesis found that the peptide fragment PAMP(12-20) acts as an angiogenesis induced by the intact ligand or by tumor cell–derived peptide. Finally, antagonist for either neovascularization induced by synthetic PAMP PAMP(12-20) treatment partially blocked the in vivo growth of hu- or by tumor cells and reduces tumor growth in a xenograft model. man tumors as assessed in a nude mouse xenograft model. In light of

Tissue Regulation of Proadrenomedullin NH2-Terminal 20 significant recent clinical benefits of antiangiogenic therapies based Peptide Levels. An obvious question raised by our observations is on VEGF blockade (7), PAMP antagonists provide a conceptually how a peptide with such extraordinary angiogenic effects can be attractive tool to further explore this strategy. produced normally without causing an excessive amount of blood vessel formation. The expression of the adrenomedullin/PAMP gene ACKNOWLEDGMENTS during development is tightly regulated in a spatio-temporal manner (26), underscoring the need for a controlled liberation of these mol- We gratefully acknowledge the expert technical assistance of Vickie Keck ecules. In addition, PAMP is rapidly degraded by neutral endopepti- with the xenograft studies. dases (27, 28), resulting in extremely low levels of circulating PAMP (29). The silicone implant used in the directed in vivo angiogenesis REFERENCES assay may provide a secluded environment protecting PAMP from protease degradation, thus contributing to the angiogenic response 1. Carmeliet P. Angiogenesis in health and disease. Nat Med 2003;9:653–60. 2. Distler J, Hirth A, Kurowska-Stolarska M, Gay RE, Gay S, Distler O. Angiogenic and observed at very low concentrations of PAMP. These data suggest a angiostatic factors in the molecular control of angiogenesis. Q J Nucl Med 2003;47: scenario in which PAMP will be secreted at a low concentration in the 149–61. 3. Shannon AM, Bouchier-Hayes DJ, Condron CM, Toomey D. Tumor hypoxia, che- proximity of the target cells (endothelial cells), and after effecting its motherapeutic resistance and hypoxia-related therapies. Cancer Treat Rev 2003;29: paracrine action, it will be rapidly cleaved by neutral endopeptidases 297–307. to avoid excessive proliferation of blood vessels. In a way, this 4. Folkman J, Watson K, Ingber D, Hanahan D. Induction of angiogenesis during the transition from hyperplasia to neoplasia. Nature 1989;339:58–61. biology will not be too different from the mechanisms exhibited by 5. Cavallaro U, Christofori G. Molecular mechanisms of tumor angiogenesis and tumor other potent biological substances that exert their function locally, progression. J Neurooncol 2000;50:63–70. such as neurotransmitters. 6. Scappaticci FA. The therapeutic potential of novel antiangiogenic therapies. Expert Opin Investig Drugs 2003;12:923–32. The rapid and efficient cleavage of PAMP by resident endopepti- 7. Yang JC, Haworth L, Sherry RM, et al. A randomized trial of bevacizumab, an dases may also explain the lack of effect observed in the xenograft anti-vascular endothelial growth factor antibody, for metastatic renal cancer. N Engl J Med 2003;349:427–34. experiment when we added full-length peptide. Alternatively, we can 8. Lo´pez J, Martı´nez A. Cell and molecular biology of the multifunctional peptide, speculate that the tumor cells may be producing enough PAMP to adrenomedullin. Int Rev Cytol 2002;221:1–92. saturate the receptors present in the endothelial cell surface, thus any 9. Zhao Y, Hague S, Manek S, Zhang L, Bicknell R, Rees MC. PCR display identifies tamoxifen induction of the novel angiogenic factor adrenomedullin by a non estro- additional peptide would not modify the angiogenic response. Similar genic mechanism in the human endometrium. Oncogene 1998;16:409–15. results have been reported for adrenomedullin in breast cancer cells 10. Martı´nez A, Vos M, Gue´dez L, et al. The effects of adrenomedullin overexpression (30). in breast tumor cells. J Natl Cancer Inst (Bethesda) 2002;94:1226–37. 11. Oehler MK, Hague S, Rees MC, Bicknell R. Adrenomedullin promotes formation of There are many substances that regulate the expression of the xenografted endometrial tumors by stimulation of autocrine growth and angiogenesis. adrenomedullin/PAMP gene, including cytokines, other hormones, Oncogene 2002;21:2815–21. 12. Ishikawa T, Chen J, Wang J, et al. Adrenomedullin antagonist suppresses in vivo growth factors, interferons, lipopolysaccharide, etc. (8). Hypoxia is a growth of human pancreatic cancer cells in SCID mice by suppressing angiogenesis. common inducer of angiogenesis in tumors (31), and the ad- Oncogene 2003;22:1238–42. 6493

13. Cuttitta F, Pıó R, Garayoa M, et al. Adrenomedullin functions as an important tumor 24. Yahata Y, Shirakata Y, Tokumaru S, et al. Nuclear translocation of phosphorylated survival factor in human carcinogenesis. Microsc Res Tech 2002;57:110–9. STAT3 is essential for VEGF-induced human dermal microvascular endothelial cell 14. Isaacs JS, Jung Y, Mimnaugh EG, Martıńez A, Cuttitta F, Neckers LM. Hsp90 migration and tube formation. J Biol Chem 2003;278:40026–31. regulates a VHL-independent HIF-1␣ degradative pathway. J Biol Chem 2002;277: 25. Chlenski A, Liu S, Cohn SL. The regulation of angiogenesis in neuroblastoma. 29936–44. Cancer Lett 2003;197:47–52. 15. Auerbach R, Lewis R, Shinners B, Kubai L, Akhtar N. Angiogenesis assays: a critical 26. Montuenga LM, Martıńez A, Miller MJ, Unsworth EJ, Cuttitta F. Expression of overview. Clin Chem 2003;49:32–40. adrenomedullin and its receptor during embryogenesis suggests autocrine or paracrine 16. Gue´dez L, Rivera AM, Salloum R, et al. Quantitative assessment of angiogenic modes of action. Endocrinology 1997;138:440–51. responses by the directed in vivo angiogenesis assay. Am J Pathol 2003;162: 27. Martıńez A, Hodge DL, Garayoa M, Young HA, Cuttitta F. Alternative splicing of the 1431–9. proadrenomedullin gene results in differential expression of gene products. J Mol 17. Lau KL, Kong SK, Ko WH, Kwan HY, Huang Y, Yao X. cGMP stimulates Endocrinol 2001;27:31–41. endoplasmic reticulum Ca(2ϩ)-ATPase in vascular endothelial cells. Life Sci 2003; 28. Nagatomo Y, Kitamura K, Kangawa K, Fujimoto Y, Eto T. Proadrenomedullin 73:2019–28. N-terminal 20 peptide is rapidly cleaved by neutral endopeptidase. Biochem Biophys Res Commun 1996;223:539–43. 18. Macpherson GR, Ng SS, Forbes SL, et al. Anti-angiogenic activity of human 29. Washimine H, Kitamura K, Ichiki Y, et al. Immunoreactive proadrenomedullin endostatin is HIF-1-independent in vitro and sensitive to timing of treatment in a N-terminal 20 peptide in human tissue, plasma and urine. Biochem Biophys Res human saphenous vein assay. Mol Cancer Ther 2003;2:845–54. Commun 1994;202:1081–7. 19. Martıńez A, Saldise L, Ramı´rez MJ, et al. Adrenomedullin expression and function in 30. Miller MJ, Martıńez A, Unsworth EJ, et al. Adrenomedullin expression in human the rat carotid body. J Endocrinol 2003;176:95–102. tumor cell lines: its potential role as an autocrine growth factor. J Biol Chem 20. Ouafik L, Sauze S, Boudouresque F, et al. Neutralization of adrenomedullin inhibits 1996;271:23345–51. the growth of human glioblastoma cell lines in vitro and suppresses tumor xenograft 31. Pugh CW, Ratcliffe PJ. Regulation of angiogenesis by hypoxia: role of the HIF growth in vivo. Am J Pathol 2002;160:1279–92. system. Nat Med 2003;9:677–84. 21. McLatchie LM, Fraser NJ, Main MJ, et al. RAMPs regulate the transport and ligand 32. Garayoa M, Martıńez A, Lee S, et al. Hypoxia-inducible factor-1 (HIF-1) up- specificity of the calcitonin-receptor-like receptor. Nature 1998;393:333–9. regulates adrenomedullin expression in human tumor cell lines during oxygen 22. Katoh F, Kitamura K, Niina H, et al. Proadrenomedullin N-terminal 20 peptide deprivation: a possible promotion mechanism of carcinogenesis. Mol Endocrinol (PAMP), an endogenous anticholinergic peptide: its exocytotic secretion and 2000;14:848– 62. inhibition of catecholamine secretion in adrenal medulla. J Neurochem 1995;64: 33. Giavazzi R, Sennino B, Coltrini D, et al. Distinct role of fibroblast growth factor-2 459– 61. and vascular endothelial growth factor on tumor growth and angiogenesis. Am J 23. Belloni AS, Rossi GP, Andreis PG, et al. Proadrenomedullin N-terminal 20 Pathol 2003;162:1913–26. peptide (PAMP), acting through PAMP(12-20)-sensitive receptors, inhibits 34. Takano S, Tsuboi K, Matsumura A, Nose T. Anti-vascular endothelial growth factor Ca2ϩ-dependent, agonist-stimulated secretion of human adrenal glands. Hyper- antibody and nimustine as combined therapy: effects on tumour growth and angio- tension 1999;33:1185–9. genesis in human glioblastoma xenografts. Neuro-oncol 2003;5:1–7.

6494

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2004 American Association for Cancer Research. Proadrenomedullin NH2-Terminal 20 Peptide Is a Potent Angiogenic Factor, and Its Inhibition Results in Reduction of Tumor Growth

Alfredo Martínez, Enrique Zudaire, Sergio Portal-Núñez, et al.

Cancer Res 2004;64:6489-6494.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/64/18/6489

Cited articles This article cites 34 articles, 8 of which you can access for free at: http://cancerres.aacrjournals.org/content/64/18/6489.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/64/18/6489.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/64/18/6489. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.