Mouse Model of Angiogenesis - Europe PMC

American Journal ofPathology, Vol. 152, No. 6, June 1998 Copyright ©) American Society for Investigative Pathology

Animal Model

Mouse Model of Angiogenesis

Thierry Couffinhal, Marcy Silver, Lu P. Zheng, Marianne Kearney, Bernhard Witzenbichler, and Jeffrey M. Isner From the Departments of Medicine (Cardiology) and Biomedical Research, St. Elizabeth's Medical Center, Tufts University School ofMedicine, Boston, Massachusetts

Neovascularization of ischemic muscle may be sufficient to preserve tissue integrity and/or function and may thus be considered to be therapeutic. The regulatory role of vascular endothelial growth factor (VEGF) in therapeutic angiogenesis was suggested by experiments in which exogenously aministered VEGF was shown to augment collateral blood flow in animals and patients with experimentally induced bindlimb or myocardial ischemia. To address the possible contribution of postnatal endogenous VEGF expression to collateral vessel development in ischemia tissues, we developed a mouse model of hindlimb ischemia. The femoral artery of one hindllmb was ligated and excised. Laser Doppler perfusion imaging (LDPI) was employed to document the consequent reduction in hindlimb blood flow, which typically persisted for up to 7 days. Serial in vivo examinations by LDPI disclosed that hindlimb blood flow was progressively augmented over the course of 14 days, ultimately reaching a plateau between 21 and 28 days. Morphometric analysis of capillary density performed at the same time points selected for in vivo analysis of blood flow by LDPI confirmed that the histological sequence of neovascularization corresponded temporally to blood flow recovery detected in vivo. Endothelial cell proliferation was documented by immunostaining for bromodeoxyuridine injected 24 hours before each of these time points, providing additional evidence that angiogenesis constitutes the basis for improved collateral-dependent flow in this animal model. Neovascularization was shown to develop in association with augmented expression of VEGF mRNA and protein from skeletal myocytes as well as endothelial cells in the ischemic hindlimb; that such reparative angiogenesis is indeed dependent upon VEGF up-regulation was confirmed by impaired neovascularization after administration of a neutralizing VEGF antibody. Se-

quential characterization of the in vivo, histological, and molecular findings in this novel animal model thus document the role of VEGF as endogenous regulator of angiogenesis in the setting of tissue ischemia. Moreover, this murine model represents a potential means for studying the effects of gene targeting on nutrient angiogenesis in vivo. (Am J Pathol 1998, 152:1667-1679)

Angiogenesis constitutes a physiological response to ischemia.1 Neovascularization of ischemic cardiac or skeletal muscle may be sufficient to preserve tissue integrity and/or function, and may thus be considered to be therapeutic.2'3 In the case of most neoplasms, angiogenesis is necessary for tumor growth and/or metastasis and has thus been considered to be pathological.4 Both therapeutic as well as pathological angiogenesis have been inferred to be dependent upon, if not regulated by, cytokines that are mitogenic for endothelial cells.5 Earlier studies of neoplasms established evidence that cytokine up-regulation was anatomically associated with foci of tumor necrosis,6 and similar observations were made in other naturally occurring models of pathological angiogenesis, such as proliferative retinopathy.7'6 These studies specifically implicated vascular endothelial growth factor (VEGF), subsequently shown to be upregulated in hypoxic myocytes9 and endothelial cells10 by transcriptional and post-transcriptional mechanisms.11-13 Observations regarding the contribution of VEGF to pathological angiogenesis were complemented by experiments in which soluble receptors14 or neutralizing antibodies15'16 were employed to block tumor and neovascularization of the iris after retinal ischemia. The regulatory role of VEGF in therapeutic angiogenesis was first suggested by experiments in which supplemental VEGF, administered as recombinant protein or Supported in part by grants (HL40518, HL02824, and HL57516 to J.M. Isner) from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, and the E.L. Weigand Foundation, Reno, NV. T. Couffinhal was supported by a grant from the University Hospital of Bordeaux and the Lavoisier grant from the French Ministry of Foreign

Affairs. Accepted for publication March 12, 1998. Address reprint requests to Dr. Jeffrey M. Isner, St. Elizabeth's Medical Center, 736 Cambridge Street, Boston, MA 02135.

1667

1668

Couffinhal et al

AJP June 1998, Vol. 152, No. 6

naked DNA, was shown to augment collateral blood flow in rabbits, rats, or pigs with experimentally induced hindlimb2'17 or myocardial3'18,19 ischemia. This notion was further supported by mid-gestational death of embryos in which targeted disruption of one or both VEGF alleles resulted in abnormal blood vessel development.2021 These studies, however, did not address or relate the possible contribution of postnatal endogenous VEGF expression to collateral vessel development in ischemia tissues. To more directly investigate the role of VEGF as well as other regulatory elements in postnatal angiogenesis, we developed a mouse model of hindlimb ischemia. The current series of investigations characterize and document the operative reduction and subsequent endogenous restoration of hindlimb blood flow in this animal model. Neovascularization, as evidenced by increased endothelial cell proliferation and capillary density, is shown to develop in association with augmented expression of VEGF mRNA and protein; that such reparative angiogenesis is indeed dependent upon VEGF up-regulation is confirmed by impaired neovascularization after administration of a neutralizing VEGF antibody. These experiments thus document the role of VEGF as endogenous regulator of angiogenesis in the setting of tissue ischemia and establish a novel animal model to exploit targeted mouse genetics in the study of therapeutic

angiogenesis.

Materials and Methods Animal Care C57BL/6 female mice (Jackson Laboratory, Bar Harbor, ME), 6 months of age and weighing 25 to 35 g, were used for all experiments. All protocols were approved by St. Elizabeth's Institutional Animal Care and Use Committee. The animals were anesthetized with pentobarbital, 160 mg/kg intraperitoneally (i.p.) for the surgical procedure as well as for laser Doppler measurements of limb perfusion (vide infra). Postoperatively, the animals were closely monitored, and analgesia (60 mg/kg levorphanol tartrate) was administered as required.

Surgical Procedures Operative intervention was performed to create unilateral hindlimb ischemia in the mice. Exposure was obtained by performing an incision in the skin overlying the middle portion of the left hindlimb of each mouse. After ligating the proximal end of the femoral artery, the distal portion of the saphenous artery was ligated, and the artery and all side-branches were dissected free; after this, the femoral artery and attached side-branches were then excised. The overlying skin was then closed using a surgical stapler. The results of the surgery were assessed by tetrazolium dye staining and premortem angiography. At arbitrary time points postoperatively, 2% 2,3,5-triphenyltetrazolium chloride (tetrazolium red) dye (Sigma Chemical

Co., St. Louis, MO) in phosphate-buffered saline (PBS) was injected into the left ventricle (without opening the thoracic cavity) 30 minutes before sacrifice. This dye stains viable tissue red but does not penetrate ischemic or necrotic tissue. The extent of ischemia was subsequently evaluated at necropsy by gross inspection of cross sections of limb muscle obtained after removing the skin and cutting serial sections of the hindlimbs. Premortem angiography was performed immediately before sacrifice by opening the thoracic cavity and injecting radiopaque contrast material via a cannula placed in the thoracic aorta to confirm interruption of femoral arterial circulation in the operated hindlimb. To evaluate this surgical preparation as a murine model of nutrient angiogenesis, experiments were performed in untreated mice, PBS-treated mice, or mice treated with the angiogenesis inhibitor platelet factor-4 (PF-4, kindly provided by Dr. Ted Maione, Repligen, Dedham, MA) administered i.p., 1 mg daily for 10 days beginning at the time of surgery.

Hindlimb Blood Flow Laser Doppler Perfusion Imaging Laser Doppler perfusion imaging (LDPI) was used to record serial blood flow measurements over the course of 5 weeks postoperatively. The LDPI system (Lisca, North Brunswick and Hewitt, NJ)22,23 incorporates a 2-mW helium-neon laser to generate a beam of light that sequentially scans a 12 x 12 cm tissue surface to a depth of 600 p.m. During the scanning procedure, blood cells moving through the vasculature shift the frequency of incident light according to the Doppler principle. A photodiode collects the back-scattered light, and the original light intensity variations are transformed into voltage variations in the range of 0 to 10 V. A perfusion output value of 0 V was calibrated to 0% perfusion, whereas 10 V was calibrated to 100% perfusion. When the scanning procedure is terminated and the back-scattered light collected from all measurement sites, a color-coded image representing blood flow distribution is displayed on a television monitor. The perfusion signal is split into six different intervals, and each is displayed as a separate color. Low or no perfusion was displayed as dark blue, whereas maximal perfusion was displayed as red. Temporal variations in tissue perfusion were displayed as a conventional plot after conversion to ASCII code and exported to a spreadsheet software package. LDPI was used to record perfusion of both right and left limbs, preoperatively and at predetermined time points postoperatively. Excess hairs were removed by depilatory cream from the limb before imaging, and mice were placed on a heating plate at 370C to minimize temperature variation. Consecutive measurements were obtained over the same region of interest (leg and foot). Colorcoded images were recorded, and analyses were performed by calculating the average perfusion for each (ischemic and nonischemic) foot. To account for variables, including ambient light and temperature, calcu-

Mouse Model of Angiogenesis 1669 AJPJune 1998, Vol. 152, No. 6

lated perfusion was expressed as a ratio of left (ischemic) to right (normal) limb.

Tissue Preparation Animals were sacrificed at predetermined, arbitrary time points with an overdose of sodium pentobarbital. During the 24 hours before sacrifice, each animal received two i.p. injections (one every 12 hours) of bromodeoxyuridine (BrdU, 30 mg/kg; Amersham, Little Chalfont, UK). For immunohistochemistry, whole ischemic and nonischemic legs were immediately fixed in methanol for 1 hour, serially cut in a transverse fashion at 3-mm intervals, additionally fixed overnight in methanol, and then embedded in paraffin. For in situ hybridization, tissues were fixed for 24 hours in 4% paraformaldehyde in PBS before embedding in paraffin. After embedding, bones were carefully removed using a drill, and samples were embedded again so that the whole leg could be analyzed on each section. For immunohistochemistry and in situ hybridization, a minimum of three animals were examined for each time point. For protein extraction, tissue samples were rinsed in PBS to remove excess blood, flash frozen in liquid nitrogen, and stored at -80°C until used.

Immunohistochemistry Five-micron sections prepared from paraffin-embedded, transverse-cut tissue samples of whole leg were used for

immunohistochemistry. Immunohistochemical staining for endothelial cells was carried out using a rat monoclonal antibody (MAb) against mouse CD31 (PharMingen, San Diego, CA) or a rabbit polyclonal antibody (PAb) against factor VIII (Signet Laboratories, Dedham, MA). BrdU-positive nuclei were detected using a sheep PAb (Biodesign, Kennebunk, ME). Vascular endothelial growth factor (VEGF) was detected using a rabbit PAb prepared against human VEGF amino-terminal peptides 1 to 20 (catalog item sc-152, Santa Cruz Biotechnology, Santa Cruz, CA); the mouse VEGF amino-terminal peptides 1 to 20 are similar to the human sequence except for position 11 where a Thr replaces the Ser in the human sequence. Previous experiments showed that this antibody immunoprecipitates a 46-kd protein (nonreducing condition) or a 23-kd protein (reducing condition) from mouse uterus protein extract corresponding to the size of recombinant human VEGF165 used as a control. T lymphocytes were identified using an anti-CD3 antibody (Sigma), and macrophages were identified using the F4/80 antibody specific for mouse macrophages (Caltag Laboratories, Burlingame, CA). Immunoperoxidase staining was performed as previously described.24 Briefly, 5-gm sections were incubated in 3% hydrogen peroxide to block endogenous peroxidase activity. To prevent nonspecific antibody binding, the sections were preincubated for 20 minutes in a solution containing either 10% horse serum or 10% goat serum in PBS. An additional protein-blocking step was performed using an avidin-biotin blocking kit (Vector Laboratories, Burlingame, CA). For antibody staining, sec-

tions were first incubated with a primary antibody at appropriate dilutions overnight at 40C, rinsed 15 minutes with PBS, incubated with biotinylated secondary antibody for 30 minutes at room temperature, rinsed 15 minutes with PBS, and covered with streptavidin-horseradish-peroxidase (HRP) complex (Signet Laboratories). After 30 minutes at room temperature, the sections were rinsed with PBS and incubated with 0.05% (w/v) 3,3'-diaminobenzidine tetrahydrochloride dihydrate for VEGF staining, 3-amino-9-ethylcarbazole (for CD31, CD3, F4/80, or BrdU staining, or Vector SG (Vector Laboratories) for double immunostaining, including BrdU. A counterstain of 20% Gill hematoxylin was applied before coverslipping. For BrdU staining, slides were incubated before application of hydrogen peroxide in 2 N HCI at 370C for 30 minutes and then rinsed twice in borate buffer. Negative control slides were prepared by substituting preimmune rat serum for the primary monoclonal antibody or using preabsorption of VEGF Ab with the specific peptide (Santa Cruz). Capillary density, as an index of angiogenesis, was examined by counting the number of capillaries in light microscopic sections taken from the ischemic and nonischemic limbs. The entire leg below the knee from each animal was examined. Serial sections were cut at two different levels approximately 200 ,um apart. On each series VEGF, BrdU, and CD31 staining was performed. Thirty fields of CD31 staining were counted from the two levels for each of three animals per time point, for a total of 90 fields per time point. Initial pilot experiments were conducted to estimate the capillary density throughout the leg, and the counts were found to be approximately the same from one extremity of the leg to the other. Capillaries were counted under an 80x objective to determine the capillary density (mean number of capillaries/ mm2). Results of the time-course analysis of capillary density (vide infra) did not differ when the analysis was performed by determining the capillary/myocyte ratio. Pilot studies of immunostained serial adjacent sections had suggested that endothelial cells constituted the predominant proliferative cell type. To confirm this, double immunostaining was used to study BrdU incorporation into endothelial cells. Briefly, the first primary MAb (antiBrdU) was applied to the tissue sections, followed by biotinylated IgG and then the HRP complex. After incubation in Vector SG (Vector Laboratories) to yield a gray nuclear staining, sections were washed in PBS and incubated with MAb CD31, followed by biotinylated IgG and then the HRP complex. After washes in straight PBS, the antibody was detected by incubating the sections in 3-amino-9-ethylcarbazole to give a red reaction product. All sections for a time-course study were stained in the same run for each antibody tested. Proliferative activity, assessed by BrdU immunostaining, was expressed as the number of positive nuclei/mm2 and was recorded as described for the capillary density.

Immunoprecipitation Frozen tissue (limbs weighing approximately 0.4 g) was diced into small pieces in liquid nitrogen and briefly

1670

Couffinhal et al


homogenized with a polytron device in 3 ml of cold lysis buffer (PBS, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS) containing inhibitors. The mixture was further incubated for 30 minutes at 40C. Cellular debris was pelleted by two successive centrifugations (30 minutes at 14,000 x g) at 40C. The total cell lysate was precleared by incubating with 1 ,ug of rabbit IgG and 20 Al of protein A/G agarose conjugate for 2 to 3 minutes and pelleted by centrifugation at 6000 rpm at 40C for 5 minutes. Five milligrams of protein representing an equal contribution from two different animals was incubated with VEGF PAb (0.5 Ag/ml; Santa Cruz) overnight at 40C on a rotating device. Therefore, 40 ,ul of agarose beads were added for 1 hour at 40C, and the immunoprecipitate was collected by centrifugation (6000 rpm for 5 minutes) and washed four times with lysis buffer. Samples were analyzed by SDS-polyacrylamide gel electrophoresis on a 12% gel under reducing conditions and transferred to a nitrocellulose membrane. The membrane was blocked in 10% nonfat dry milk in Dulbecco-PBS (D-PBS) plus 0.1% Tween 20 for 1 hour at room temperature and subsequently incubated with VEGF antibody (0.2 ,ug/ml) for 90 minutes in D-PBS plus 5% milk. The membrane was washed three times for 20 minutes each in D-PBS plus 0.1% Tween and incubated with anti-rabbit IgG-peroxidase conjugate for 1 hour at room temperature. After a 1-hour wash in D-PBS plus 0.1% Tween, the membrane was stained with luminescence detection system (ECL, Amersham) according to the manufacturer's instructions. Positive controls included recombinant VEGF165 (kind gift of Bruce Keyt and Napoleon Ferrara, Genentech, South San Francisco, CA). Negative controls included a membrane probed with VEGF antibody preabsorbed with a 10-fold concentration of peptide for 1 hour (Santa Cruz).

Hybridization Probes Mouse VEGF165 cDNA was cloned using polymerase chain reaction (PCR) to serve as template for synthesis of specific cRNA. Briefly, total RNA from mouse uterus was reverse transcribed,2526 after which cDNA was amplified by PCR for 25 cycles using a VEGF primer set encompassing start and stop codons. The 573-bp PCR product corresponding to VEGF165 was purified and subcloned into pCRII (Invitrogen, San Diego, CA). The DNA sequence of the insert was determined using an ABI 373A automated DNA sequencer (Applied Biosystems 373A, Foster City, CA).

In Situ Hybridization In situ hybridization was carried out using the Genius labeling and detection kit (Boehringer Mannheim, Mannheim, Germany) as described elsewhere.27 After deparaffinizing, 10-,tm tissue sections were fixed in 4% paraformaldehyde for 30 minutes, treated with proteinase K (20 mg/ml) for 15 minutes, and then acetylated with triethanolamine and finally dehydrated and air dried. Labeled, single-strand RNA probes were prepared using digoxigenin-1 1-UTP according to the manufacturer's in-

structions. The mouse VEGF plasmid was either linearized with Notl and transcribed with SP6 RNA polymerase to generate an antisense probe or with BamHl and transcribed with T7 RNA polymerase to generate a sense probe. Hybridization was carried out at 500C for 16 hours with 500 ng/ml probe in a hybridization buffer containing 50% deionized formamide, 1 x Denhardt's solution, 4x saline sodium citrate (SSC), 10% dextran sulfate, and 0.4 mg/ml single-strand DNA. Slides were washed twice for 15 minutes each in 2x SSC at room temperature and then treated with RNAse A (100 gg/ml) in 2x SSC at 370C for 30 minutes. The slides were washed twice for 20 minutes each in 0.1 x SSC at 420C and then at room temperature. The alkaline-phosphatase-conjugated antidigoxigenin antibody (1/500 dilution) was incubated for 2 hours in 10% horse serum at ambient temperature, and the slides were washed overnight. The substrate (nitroblue tetrazolium salt, X-phosphate) was applied for 5 hours, after which slides were counterstained with methyl green and coverslipped. Controls included 1) hybridization with the sense probe and 2) use of an irrelevant antibody. To confirm endothelial cell expression of VEGF, evidence for VEGF mRNA expression was investigated using in situ hybridization; serially cut sections were probed for VEGF mRNA, and adjacent sections were immunostained for CD31 and VEGF protein.

In Vivo Application of Neutralizing VEGF Antibody In vivo injection of a neutralizing VEGF antibody24 was used to further evaluate the contribution of VEGF to angiogenesis in the ischemic hindlimb. Mice were injected i.p. with either a goat polyclonal VEGF blocking antibody (R&D Systems, Minneapolis, MN) or a control PAb of the same isotype in a volume of 100 ,lI in PBS. A dose of 100 Ag was administered on the day of the surgery, followed by 200 ,ug at days 3, 6, and 9 and 100 jig again at day 12; the animals were sacrificed at day 15.

Results Postoperative Course The mice recovered without difficulty from the surgery. Most did not use their ischemic hindlimb during the first postoperative week and then progressively began to use the limb again, and by the fifth week it was difficult to detect a difference between the ischemic and nonischemic limbs. Approximately 10% of mice developed signs of toe necrosis during the first postoperative week but then healed rapidly during the following weeks. Gross anatomic inspection after tetrazolium dye staining of the muscle in the ischemic limb at necropsy examination demonstrated a core of necrotic muscle surrounded by red-stained viable tissue.


1671

AJPJune 1998, Vol. 152, No. 6

IA

I

Pre

D1 4

Post

D3

D7

I.

D21

Figure 1. Hindlimb blood flow monitored serially in vivo by laser Doppler perfusion imaging (LDPI). In color-coded images, normal baseline (pre) perfusion in both hindlimbs is depicted in red. Immediately after operative excision of one femoral artery (post), marked reduction in blood flow of one hindlimb is depicted in blue. Perfusion remained severely impaired for 7 days, increased by day 14, and ultimately returned to near-normal levels by day 28.

Hindlimb Blood Flow Serial blood flow measurements were performed sequentially using LDPI in the same animal over a period of 35 days. Figure 1 illustrates the time course of flow recovery in the murine ischemic hindlimb, assessed by LDPI. After an initially precipitous reduction in flow for the ischemic versus normal limb, modest recovery was observed by day 7. By day 14, flow recovery in the ischemic limb was more substantial and by days 28 to 35 achieved a plateau. Treatment with PF-4 for 10 days beginning at the time of surgery led to a marked inhibition of flow recovery for 14 days; a statistically significant inhibition in blood flow recovery was maintained through day 35 (0.59 ± 0.06 at day 35 for PF-4-treated mice versus 0.90 ± 0.06 for untreated mice; P < 0.05; Figure 2).

Capillary Density and Proliferative Activity Revascularization of the ischemic hindlimb was evaluated by performing light microscopic examination of the ischemic and nonischemic limbs over the course of 35

days postoperatively. Capillary density was determined using a rat MAb against mouse CD31. Figure 3 illustrates the increase in capillary density observed in the ischemic hindlimb at 35 days postoperatively versus that measured in the normal limb. Time-course analysis of capillary density showed early destruction of capillaries in the ischemic limb. Thereafter, a statistically significant increase in capillary density was noted as early as day 7, was sustained at day 14, and increased further at days 21 and 35 (Figure 4A). In contrast, in the contralateral normal hindlimb, no significant alteration in capillary density was observed between days 0 and 35. Finally, the ischemic district was invaded by capillaries and small vessels (10 to 50 ,um in diameter). Thus, by day 35, capillary density in the ischemic limb was increased by fourfold compared with capillary density in the nonischemic limb. Similar results were obtained when capillaries were indexed by area; at day 35, capillaries/mm2 values for the ischemic versus normal limb were 1609 ± 245 versus 500 ± 167, respectively (P = 0.001). The serial measurements of capillary density measured by CD31 staining corresponded to the temporal evolution of improved perfusion

1672

Couffinhal et al

A/P Pjune 1998, Vol. 152, No. 6

1.2

0

9

1.0

t.v

0.8

Ischemic/Normal Leg Perfusion Ratio

density (268 + 195 versus 1053 + 371 capillaries/mm2; P < 0.01) and endothelial cell proliferation (16 + 29 versus 935 + 239 BrdU-positive cells/mm2; P < 0.01) were found in PF-4 versus PBS-injected mice, respectively (Figure 2).

o

C

0.6

T7 oo

0.4 0.2

1600 Capillary Density/mm2

1400 -

E

1200 -

T

1000P-~ a

800600-

400-

1*

2000-

1200 L

BrDU positive cells/mm2

1000 -

2

800-

T-1 600._

400-

200I

0-

C. o giR

*i* 11'1.

Immunostaining for VEGF was performed on tissues harvested from ischemic and nonischemic hindlimbs at serial time points. No or very faint VEGF staining was detected in normal limbs (Figure 6A). By day 2, positive, albeit patchy and weak, VEGF immunostaining was detected in the ischemic limbs. By day 4, positive staining was more diffuse. Maximal VEGF immunostaining was detected at 7 and 14 days postoperatively. Immunostaining of adjacent tissue sections with cell-specific antibodies also localized positive VEGF immunostaining to inflammatory cells consisting of T cells and macrophages (Figure 6B). After 14 days, VEGF protein expression decreased significantly but was still detectable by day 35 postoperatively. Occasional tissue samples retrieved from the contralateral nonischemic limb disclosed faint VEGF immunostaining. Absent staining in sections treated with the peptide preabsorbed antibody confirmed the specificity of VEGF immunostaining. The results of VEGF immunostaining were confirmed by VEGF immunoprecipitation. Only a very faint VEGF band could be detected in the nonischemic limb at baseline (Figure 6C). Hindlimb ischemia induced an increase in VEGF expression within 4 days that reached a maximum between 7 and 14 days (Figure 6C). VEGF was barely detected in the nonischemic contralateral limb. The specificity of these results was confirmed by probing the immunoblot with VEGF antibody previously preabsorbed with the specific peptide.

4.0-

.

IN

VEGF Protein Expression

4.

1.1'

Figure 2. Quantitative evaluation of blood flos. capillary density, and cellolar proliferation in normal as wsell as ischemic (Day 15) hindliml) tissuLes treated wxith PBS (control), platelet factor-4 (PF-4), irrelevant antibody (control), or VEGF neutralizing antibody. *P < 0.05; **P < 0.01 versus contrcl.

in the ischemic/normal limbs described above and illustrated in Figure 1. BrdU staining was performed to evaluate proliferative activity in the ischemic versus normal hindlimbs. Proliferative activity peaked at 7 days (1235 -+- 254 versus 8 -+- 14 BrdU-positive cells/mm2 for the ischemic versus normal limbs, respectively (P < 0.001); proliferative activity was then subsequently reduced at days 14 and 21 (Figure 4B). Double immunolabeling for BrdU and CD31 demonstrated proliferating endothelial cells in the ischemic limb (Figure 5). Most proliferating endothelial cells localized to small capillaries, although endothelial cell proliferation was observed in small arteries as well. Capillary density and proliferative activity were also examined in mice treated with PF-4 and sacrificed 14 days after surgery. A significant decrease in capillary

VEGF mRNA Expression In situ hybridization was performed using a murine VEGF165 cRNA probe to identify VEGF mRNA. Before surgery, scarce hybridization was detected in the ischemic limb. Within 4 days postoperatively, however, VEGF mRNA was detected in skeletal myocytes and arterial smooth muscle cells as well as infiltrating cells (Figure 7). Tissue sections retrieved on day 7 disclosed a similar pattern, but by day 14, the hybridization signal was weaker. By day 21, VEGF mRNA was barely detectable. VEGF mRNA was also detected in endothelial cells located in small capillaries or venules. In some cases, endothelial cells of larger-caliber veins also displayed positive hybridization for VEGF (Figure 8); VEGF expression among endothelial cells of similar-caliber arteries, however, was less frequent.

Effect of VEGF Neutralizing Antibody Eight mice were injected every 3 days for the first 14 days postoperatively with VEGF neutralizing antibody (n = 4) or control antibody (n = 4). On the day of sacrifice, blood

Mouse Model of Angiogenesis 1673 AJPJune 1998, Vol. 152, No. 6

nonischeinic

ischemic d35

Figure 3. Capillary density in nonischemic versus ischemic tissues at day 35 after surgery. Capillaries were identified by a red-brown reaction product after CD31 immunostaining. These representative photomicrographs demonstrate increased capillary density in the ischemic versus the normal hindlimb at 35 days postoperatively.

flow in each limb was measured by LDPI, after which tissues were harvested to evaluate capillary density and proliferative activity. Mice receiving the control antibody demonstrated no statistically significant reduction in blood flow, capillary density, or cellular proliferation at 14

days (Figure 2). In contrast, recovery of blood flow in the ischemic limb of mice treated with neutralizing VEGF antibody was significantly retarded compared with control mice (Figure 2); these findings were similar to findings observed in mice receiving PF-4. Attenuated blood flow

1674

Couffinhal et al


A 2000.

IfthOldC L41g

--4b--

Norlad I.Ag

.1.. -

1800.

i

1400

1 i8-

1"00

1

1200.

600. 400-

200. gi. 0

7

21

14

28

35

Days post4rgery

B Double Immunostaining BrdU(gray) CD-31(red)

1600.

1400.

---o-

lscbmk Leg

-o-

Normal Leg

Figure 5. Double imnmunolabeling for BrdU (gray reaction product) and CD31 (red-brown reaction product) in ischemic hindlimb 7 days after surgery. Endothelial cells constitute the predominant proliferative cell type in the ischemic limb. Most proliferating endothelial cells localized to small capillaries, although endothelial cell proliferation was observed in small arteries as well. Arrows indicate proliferating endothelial cells.

1200. 1000.

80

800.

600. 400.

200 0

--8-_ -__ -___ __ -_-__

0

14

-

21

-

-0

-

28

Days posumery

Figure 4. Quantitative evaluation of capillary density and cellular proliferation in tissues retrieved at necropsy from ischemic and normal hindlimbs. A: Time-course evaluation of capillary density demonstrating statistically significant increase in capillary density in the ischemic limb as early as day 7; this was sustained at day 14 and further increased at days 21 and 35. In contrast, histological examinations of the contralateral normal hindlimb disclosed no significant alteration in capillary density between 0 and 35 days. B: Timecourse evaluation of proliferative activity using BrdU immunostaining. Proliferative activity peaked at 7 days in the ischemic hindlimb and then was subsequentiy reduced at days 14 and 21. No cellular proliferation was detected in the contralateral normal hindlimb from 0 to 35 days. P < 0.05; "P < 0.01 ischemic versus nonischemic.

recovery was associated with reduced neovascularization during the first 14 days postoperatively as indicated by the reduction in capillary density (Figures 2 and 9), again similar to results observed for PF-4 treated mice. As expected, proliferative activity measured by incorporation of BrdU was significantly depressed compared with that observed among control mice (Figure 2).

Discussion These experiments establish the feasibility of serially monitoring in vivo neovascularization stimulated by operatively induced hindlimb ischemia in mice. This murine model thus represents a potential means for studying the effects of gene targeting on nutrient angiogenesis in vivo. Moreover, characterization of the sequence of molecular and histological features in this animal model provides insights into naturally occurring compensatory mecha-

nisms responsible for establishing collateral blood flow after the development of limb ischemia. Previous studies performed in a rat model of femoral artery ligation indicated that collateral vessel development is primarily responsible for reduction in minimal total resistance in the limb subserved by the ligated artery.28 In the mouse model, we excised the ligated femoral artery to obviate the possibility of recanalization or bridge-collateral formation. Consequently, we observed a period of profound ischemia, indicated by occasional toe necrosis, and compromised flow; the latter typically persisted for up to 7 days. Hindlimb blood flow was then progressively augmented over the course of 14 days, ultimately reaching a plateau between 21 and 28 days. Serial recordings made simultaneously in the normal, nonischemic limb were used to index alterations in flow detected in the ischemic limb. Morphometric analysis of capillary density performed at the same time points selected for in vivo analysis of blood flow by LDPI confirmed that the histological sequence of neovascularization corresponded temporally to blood flow recovery detected in vivo. Endothelial cell proliferation was documented by immunostaining for BrdU injected 24 hours before each of these time points, providing additional evidence that angiogenesis constitutes the basis for improved collateral-dependent flow in this animal model. The temporal pattern of VEGF expression documented at the RNA level by in situ hybridization and at the protein level by immunostaining and Western blot analysis corresponds to noninvasive and histological evidence of neovascularization in the mouse ischemic hindlimb. These findings suggest that VEGF, up-regulated in response to ischemia, functioned to endogenously promote neovascularization of the operated hindlimb and complement previous reports in which supplemental VEGF ad-


1675


A

.Q .e..

B

Figure 6. Time course of VEGF protein expression in mouse ischemic hindlimb. A: Representative

CDO

VEGF

D4

D7

D.15

D21

photomicrographs illustrating immunostaining for VEGF (DAB, brown reaction product) in hindlimbs of mice harvested at serial time points. A progressive increase in VEGF immunostaining is demonstrated up to 14 days postoperatively; the cellular sources of VEGF appear to include skeletal muscle cells and, as shown in B: T lymphocytes and macrophages (AEC, red-brown reaction product). A and B: Hematoxylin counterstain. C: Time course for VEGF expression was confirmed by VEGF immunoprecipitation. Day 0 animals were normal animals sacrificed without having undergone surgery.

1676

Couffinhal et al


AD'

t", 3

Aw

v:.-I-I..".....

AOFF:

Figure 7. Representative photomicrographs of VEGF mRNA in skeletal muscle cells and infiltrating cells detected by in situ hybridization using a digoxigeninlabeled probe; VEGF mRNA is seen as a purple-blue reaction product. Methyl-green counterstain identifies nuclei.

ministration has been shown to augment limb neovascuIarity.2'1 7,29-32 Tissues harvested from the mouse ischemic hindlimb were evaluated by immunohistochemistry as well as in situ hybridization to identify the cellular sites responsible for VEGF synthesis. Skeletal myocytes, previously shown to up-regulate VEGF expression in response to exercise33 and electrical stimulation,34 were observed to constitute the principal source of VEGF. Consistent with pathological examinations performed in other tissue types,35 VEGF expression was also noted in cellular infiltrates composed of macrophages and T cells. Necropsy examination also documented that endothelial cells contribute to VEGF expression in the ischemic hindlimb. Tissue samples from normal legs or contralat-

eral (nonischemic) limbs from an animal with unilateral limb ischemia did not display any hybridized endothelial cells. We have previously observed that hypoxia may induce VEGF expression (mRNA as well as protein) in cultures of human umbilical vein endothelial cells (HUVECs) as well as human microvascular endothelial cells.10 These in vitro studies documented that conditioned media from hypoxic HUVECs increased endothelial cell incorporation of tritiated thymidine and demonstrated evidence of receptor autophosphorylation in hypoxic but not normoxic HUVECs. Because endothelial cells also express the high-affinity VEGF receptors, Flk1/KDR and Flt-1, the current in vivo data confirm previous in vitro studies suggesting that endothelial cells include the requisite elements for an autocrine pathway. Activa-

f::w_.0'4.!i,aL: Mouse Model of Angiogenesis

1677


.......

.- _ DE>.': ._

F

:_:

fl-i

.t

|:

._

_ iv_. s_F

_-.

...

.

r,_

i

E'

_g.

E'

..

DF*.

w.

w

.j

i;

I ";;:

*:2 i * ,. | l

k IL s

1,

.

.g , .

1[

4

Figure 8. VEGF expression in endothelial cells of ischemic hindlimb tissue. Cells expressing VEGF mRNA were identified as endothelial cells by their specific membrane expression of CD31 antigen. Immunohistochemical staining of adjacent sections indicated that these cells expressed VEGF protein as well. Arrows indicate positively stained cells.

1678 Couffinhal et al AJPJune 1998, Vol. 152, No. 6

CONTROL ANTIBODY

ANTIBODY ANTI-VEGF (D14)

Figure 9. Capillary density demonstrated by CD31 immunostaining in ischemic hindlimb of mice treated for 15 days with control antibody or neutralizing VEGF antibody. The reduction in neovascularization of the ischemic hindlimb in mice receiving the neutralizing VEGF antibody was similar to that seen in mice treated with PF-4.

tion of such an autocrine loop under hypoxic conditions might serve to amplify and/or protract the response of endothelial cells stimulated primarily by VEGF production from skeletal muscle cells via a paracrine pathway. Evidence that VEGF constitutes the principal regulatory mediator of endogenous neovascularization of ischemic tissues was established in the mouse model by two interventions. The first involved administration of recombinant PF-4, previously shown to inhibit angiogenesis in chick chorioallantoic membrane36 as well as solid tumors.37 Gengrinovitch et a138 have demonstrated that PF-4 may inhibit angiogenesis by disrupting VEGF receptor-mediated signal transduction and/or disrupting the binding of VEGF to cell surface heparan sulfates. To more specifically isolate the role of VEGF in modulating angiogenesis in this mouse model, we administered a VEGF neutralizing antibody. Similar to findings observed in mice receiving PF-4, recovery of blood flow, capillary density, and proliferative activity measured by incorporation of BrdU were all significantly depressed in the ischemic limb of mice treated with neutralizing VEGF antibody compared with control mice. Similar attenuation of spontaneous angiogenesis in freshly cut aortic rings cultured in a serum-free collagen gel and treated with a neutralizing VEGF antibody was recently described by Nicosia et al.39 In summary, this murine model of angiogenesis may be useful for defining the host response to the development of tissue ischemia. In addition, it is anticipated that this animal model will prove useful for studies designed to

assess the impact of gene targeting on neovascularization of ischemic tissues.

References 1. Folkman J: Clinical applications of research on angiogenesis. N Engl J Med 1995, 333:1757-1763 2. Takeshita S, Zheng LP, Brogi E, Kearney M, Pu LQ, Bunting S, Ferrara N, Symes JF, Isner JM: Therapeutic angiogenesis. A single intraarterial bolus of vascular endothelial growth factor augments revascularization in a rabbit ischemic hindlimb model. J Clin Invest 1994, 93:662-670 3. Pearlman JD, Hibberd MG, Chuang ML, Harada K, Lopez JJ, Gladston SR, Friedman M, Sellke FW, Simons M: Magnetic resonance mapping demonstrates benefits of VEGF-induced myocardial angiogenesis. Nature Med 1995, 1:1085-1089 4. Folkman J: Tumor angiogenesis: therapeutic implications. N Engl J Med 1971, 285:1182-1186 5. Folkman J, Klagsbrun M: Angiogenic factors. Science 1987, 235:442447 6. Plate KH, Breier G, Weich HA, Risau W: Vascular endothelial growth factor is a potential tumour angiogenesis factor in human gliomas in vivo. Nature 1992, 359:845-848 7. Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD, Shah ST, Pasquale LR, Theme H, Iwamoto MA, Parke JE, Nguyen MD, Aiello LM, Ferrara N, King GL: Vascular endothelial growth factor in ocular fluids of patients with diabetic retinopathy and other retinal disorders. N Engl J Med 1994, 331:1480-1487 8. Adamis AP, Miller JW, Bernal M-T, D'Amico DJ, Folkman J, Yeo T-K, Yeo K-T: Increased vascular endothelial growth factor levels in the vitreous of eyes with proliferative diabetic retinopathy. Am J Ophthalmol 1994, 118:445-450 9. Brogi E, Wu T, Namiki A, Isner JM: Indirect angiogenic cytokines upregulate VEGF and bFGF gene expression in vascular smooth


1679


10. 11.

12. 13. 14.

15.

16.

17.

18.

19.

20.

21.

22.

23. 24.

muscle cells, while hypoxia upregulates VEGF expression only. Circulation 1994, 90:649-652 Namiki A, Brogi E, Kearney M, Wu T, Couffinhal T, Varticovski L, Isner JM: Hypoxia induces vascular endothelial growth factor in cultured human endothelial cells. J Biol Chem 1995, 270:31189-31195 Mukhopadhyay D, Tsiokas L, Zhou X-M, Foster D, Brugge JS, Sukhatme VP: Hypoxic induction of human vascular endothelial growth factor expression through c-Src activation. Nature 1995, 375: 577-581 Levy AP, Levy NS, Wegner S, Goldberg MA: Transcriptional regulation of the rat vascular endothelial growth factor gene by hypoxia. J Biol Chem 1995, 270:13333-13340 Ikeda E, Achen MG, Breir G, Risau W: Hypoxia-induced transcriptional activation and increased mRNA stability of vascular endothelial growth factor in C6 glioma cells. J Biol Chem 1995, 270:19761-19766 Frank S, Hubner G, Breier G, Longaker MT, Greenhalgh DG, Werner S: Regulation of VEGF expression in cultured keratinocytes: implications for normal and impaired wound healing. J Biol Chem 1995, 270:12607-12613 Yuan F, Chen Y, Dellian M, Safabakhsh N, Ferrara N, Jian RK: Time-dependent vascular regression and permeability changes in established xenografts induced by an anti-vascular endothelial growth factor/vascular permeability factor antibody. Proc Natl Acad Sci USA 1996, 93:14765-14770 Adamis AP, Shima DT, Tolentino M, Gragoudas E, Ferrara N, Folkman J, D'Amore PA, Miller JW: Inhibition of VEGF prevents retinal ischemia associated iris neovascularization in a primate. Arch Ophthalmol 1996, 114:66-71 Takeshita S, Tsurumi Y, Couffinhal T, Asahara T, Bauters C, Symes JF, Ferrara N, Isner JM: Gene transfer of naked DNA encoding for three isoforms of vascular endothelial growth factor stimulates collateral development in vivo. Lab Invest 1996, 75:487-502 Hariawala M, Horowitz JR, Esakof D, Sheriff DD, Walter DH, Chaudhry GM, Desai V, Keyt B, Isner JM, Symes JF: VEGF improves myocardial blood flow but produces EDRF-mediated hypotension in porcine hearts. J Surg Res 1996, 63:77-82 Banai S, Jaklitsch MT, Shou M, Lazarous DF, Scheinowitz M, Biro S, Epstein SE, Unger EF: Angiogenic-induced enhancement of collateral blood flow to ischemic myocardium by vascular endothelial growth factor in dogs. Circulation 1994, 89:2183-2189 Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig M, Vandenhoeck A, Kendraprasad H, Eberhardt C, Declercq C, Pawling J, Moons L, Collen D, Risau W, Nagy A: Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 1996, 380:435-439 Ferrara N, Carver-Moore K, Chen H, Dowd M, Lu L, O'Shea KS, Powell-Braxton L, Hilan KJ, Moore MW: Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature 1996, 380:439-442 Linden M, Sirsjo A, Lindbom L, Nilsson G, Gidlof A: Laser-Doppler perfusion imaging of microvascular blood flow in rabbit tenuissimus muscle. Am J Physiol 1995, 269:H1496-H1500 Wardell K, Jakobsson A, Nilsson GE: Laser Doppler perfusion imaging by dynamic light scattering. IEEE Trans Biomed Eng 1993, 40: 309-316 Tsurumi Y, Murohara T, Krasinski K, Dongfen C, Witzenbichler B, Kearney M, Couffinhal T, Isner JM: Reciprocal relationship between

25. 26.

27.

28. 29.

30.

31.

32.

33.

34.

35.

36. 37. 38.

39.

VEGF and NO in the regulation of endothelial integrity. Nature Med 1997, 3:879-886 Sambrook J, Fritsch EF, Maniatis T: Molecular Cloning: A Laboratory Manual, ed 2. Cold Spring Harbor, NY, Cold Spring Harbor Laboratory Press, 1989, pp 8.60-8.63 Duplaa C, Couffinhal T, Dufourcq P, Llanas B, Moreau C, Bonnet J: The integrin very late antigen-4 is expressed in human smooth muscle cell: involvement of alpha4 and vascular cell adhesion molecule-1 during smooth muscle cell differentiation. Circ Res 1997, 80:159-169 Couffinhal T, Kearney M, Witzenbichler B, Chen D, Murohara T, Losordo DW, Symes JF, Isner JM: Vascular endothelial growth factor/ vascular permeability factor (VEGFNPF) in normal and atherosclerotic human arteries. Am J Pathol 1997, 150:1673-1685 Unthank JL, Nixon J, Lash JM: Early adaptations in collateral and microvascular resistances after ligation of the rat femoral artery. J Appl Physiol 1995, 79:73-82 Tsurumi Y, Takeshita S, Chen D, Kearney M, Rossow ST, Passeri J, Horowitz JR, Symes JF: Direct intramuscular gene transfer of naked DNA encoding vascular endothelial growth factor augments collateral development and tissue perfusion. Circulation 1996, 94:3281-3290 Bauters C, Asahara T, Zheng LP, Takeshita S, Bunting S, Ferrara N, Symes JF, Isner JM: Physiologic assessment of augmented vascularity induced by VEGF in ischemic rabbit hindlimb. Am J Physiol 1994, 267:H1263-H1271 Isner JM, Pieczek A, Schainfeld R, Blair R, Haley L, Asahara T, Rosenfield K, Razvi S, Walsh K, Symes J: Clinical evidence of angiogenesis following arterial gene transfer of phVEGF165. Lancet 1996, 348:370-374 Baumgartner I, Pieczek A, Manor 0, Blair R, Kearney M, Walsh K, Isner JM: Constitutive expression of phVEGF165 following intramuscular gene transfer promotes collateral vessel development in patients with critical limb ischemia. Circulation 1998, 97:1114-1123 Hang J, Kong L, Gu J-W, Adair TH: VEGF gene expression is upregulated in electrically stimulated rat skeletal muscle. Am J Physiol 1995, 269:H1827-H1831 Breen EC, Johnson EC, Wagner H, Tseng H-M, Sung LA, Wagner PD: Angiogenic growth factor mRNA responses in muscle to a single bout of exercise. J AppI Physiol 1996, 81:355-361 Freeman MR, Schneck FX, Gagnon ML, Corless C, Soker S, Niknejad K, Peoples GE, Klagsbrun M: Peripheral blood T lymphocytes and lymphocytes infiltrating human cancers express vascular endothelial growth factor: a potential role for T cells in angiogenesis. Cancer Res 1995, 55:4140-4145 Maione TE, Gray GS, Petro J, Hunt AJ, Donner AL, Bauer SI, Carson HF, Sharpe RJ: Inhibition of angiogenesis by recombinant human platelet factor-4 and related peptides. Science 1990, 247:77-79 Sharpe RJ, Byers HR, Scott CF, Bauer SI, Maione TE: Growth inhibition of murine melanoma and human colon carcinoma by recombinant human platelet factor 4. J Natl Cancer Inst 1990, 82:848-853 Gengrinovitch S, Greenberg SM, Cohen T, Gitay-Goren H, Rockwell P, Maione T, Levi B-Z, Neufeld G: Platelet factor-4 inhibits the mitogenic activity of VEGF121 and VEGF165 using several concurrent mechanisms. J Biol Chem 1995, 270:1-7 Nicosia RF, Lin YJ, Hazelton D, Qian X: Endogenous regulation of angiogenesis in the rat aorta model. Am J Pathol 1997, 151:13791386