OUP user menu

Constitutively active HIF-1α improves perfusion and arterial remodeling in an endovascular model of limb ischemia

Tarak H. Patel, Hideo Kimura, Clifford R. Weiss, Gregg L. Semenza, Lawrence V. Hofmann
DOI: http://dx.doi.org/10.1016/j.cardiores.2005.05.002 144-154 First published online: 1 October 2005


Objective: Hypoxia-inducible factor 1 (HIF-1) regulates the expression of angiogenic growth factors. We analyzed the effect of intramuscular (i.m.) delivery of AdCA5, an adenovirus encoding a constitutively active form of the HIF-1α subunit, in a novel model of limb ischemia.

Methods: AdCa5 or AdLacZ (6 × 108 pfu) was injected into male New Zealand White rabbits that were untreated or subjected to occlusion of the left superficial femoral artery by endovascular coils. Expression of mRNAs was quantified 1, 3, and 7 days after adenovirus injection into rabbits without occlusion. Calf blood pressure (BP), angiography, and immunohistochemical analyses were performed 14 days after arterial occlusion and adenovirus injection.

Results: AdCA5 increased the expression of HIF-1α, monocyte chemotactic protein-1, placental growth factor, platelet-derived growth factor B, stromal-derived factor 1α, and vascular endothelial growth factor (VEGF) mRNA as well as HIF-1α and VEGF protein. On day 14, AdCA5-injected limbs showed improved calf BP ratios (0.89 ± 0.13 vs. 0.51 ± 0.05, p = 0.02), angiographic perfusion scores (3.50 ± 0.56 vs. 8.33 ± 1.31, p = 0.007), and distal deep femoral artery diameter ratio (1.84 ± 0.25 vs. 0.93 ± 0.22, p = 0.02) relative to those receiving AdLacZ. The capillary/myocyte ratio (0.93 ± 0.03 vs. 0.78 ± 0.06, p = 0.04) and arterial luminal area (0.32 ± 0.05 mm2 vs. 0.21 ± 0.03 mm2, p = 0.04) were significantly increased in the AdCA5 group.

Conclusion: In a model that resembles atherosclerotic obstruction of peripheral arteries in patients, the i.m. administration of AdCA5 promoted arteriogenic and angiogenic responses.

  • Angiogenesis
  • Collateral circulation
  • Gene therapy
  • Hypoxia-inducible factor

1. Introduction

Peripheral vascular disease (PVD) affects 7.5% of the population aged 60–64 years in the United States [1]. Surgical bypass and percutaneous revascularization (angioplasty and stenting) are the most common treatments for patients with critical limb ischemia. However, in many patients, the anatomical distribution and extent of vascular disease exclude them as candidates for these modes of therapy. As a result, the disease often follows a relentless and debilitating course leading to amputation. Annually, PVD is responsible for 200 lower limb amputations per million in the non-diabetic population and 3900 per million in diabetic patients [1]. Novel therapeutic strategies are needed for this patient population. Therapeutic angiogenesis aims to stimulate neovascularization of ischemic tissues by administration of angiogenic growth factors or DNA sequences encoding these proteins. Multiple angiogenic factors, including members of the vascular endothelial growth factor (VEGF), fibroblast growth factor (FGF), and platelet-derived growth factor (PDGF) families, have shown promising results in preclinical studies and safety in Phase I clinical trials, but they have shown either marginal or no efficacy in Phase II trials [2,3].

Hypoxia-inducible factor 1 (HIF-1) is a transcriptional activator that functions as a master regulator of oxygen homeostasis [4]. HIF-1 is a heterodimeric protein that is composed of a constitutively expressed HIF-1β subunit, and an O2-regulated HIF-1α subunit [5–7]. HIF-1α is modified at proline residues 402 and 564 via a hydroxylation reaction that utilizes O2 as a rate-limiting substrate, thus providing a mechanism by which changes in O2 concentration are transduced to HIF-1α. Prolyl hydroxylated HIF-1α is bound by the von Hippel–Lindau protein, which is the recognition component of an E3-ubiquitin ligase that targets HIF-1α for degradation by the 26S proteasome.

HIF-1 regulates the expression of multiple angiogenic growth factors and cytokines, including angiopoietins, VEGF, PDGF-B, placental growth factor (PLGF), and stromal-derived growth factor 1 (SDF-1), as demonstrated by both gain-of-function and loss-of-function studies [8–10]. In addition to controlling the production of angiogenic factors in ischemic tissue, HIF-1 also controls cell-autonomous responses to hypoxia within human endothelial cells [11,12]. Intravitreous injection of AdCA5, an adenovirus encoding a constitutively active form of HIF-1α, stimulated neovascularization of retinal vessels that do not respond to VEGF alone, an effect that was due to its ability to activate the expression of both PLGF and VEGF [9], which act synergistically in stimulating retinal neovascularization [13]. Injection of plasmid DNA encoding a HIF-1α/VP16 fusion protein has also been shown to stimulate the recovery of blood flow in operative models of hindlimb ischemia [14] and myocardial infarction [15].

We have developed a non-invasive model of limb ischemia by endovascular occlusion, which eliminates the potentially confounding effects of a surgical wound on the vascular remodeling process [21]. Whereas early studies focused on effects of treatment on angiogenesis as measured by capillary density, it is now appreciated that the critical adaptive response to occlusion of a large artery is the remodeling of collateral vessels to accommodate increased flow, a process that has been termed arteriogenesis [22]. In this study, we demonstrate that AdCA5 gene therapy after endovascular occlusion of the superficial femoral artery (SFA) stimulates the recovery of limb perfusion through increased arterial remodeling.

2 Materials and methods

2.1 Animal care and use

Animal procedures, performed in accordance with protocols approved by the Johns Hopkins University Institutional Animal Care and Use Committee, conformed to the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes for Health (NIH Publication No. 85-23, revised 1996).

2.2 Adenovirus preparation

Replication-defective recombinant adenoviruses were constructed as described [9]. AdLacZ encodes Escherichia coli β-galactosidase. AdCA5 contains dual CMV promoters that drive expression of enhanced green fluorescent protein (EGFP), and a constitutively active form of human HIF-1α (CA5) that contains a deletion (residues 392–520) and two missense mutations (Pro567Thr and Pro658Gln). Large-scale adenovirus production was performed at the NHLBI PEGT Vector Core Facility (University of Pittsburgh).

2.3 Analysis of non-ischemic rabbits

2.3.1 Animal model

To assess the transduction efficiency of adenoviral vectors, 18 male New Zealand White rabbits (2.5–3.2 kg) were anesthetized by i.m. injection of acepromazine maleate (1 mg/kg) and ketamine hydrochloride (8 mg/kg). AdCA5 or AdLacZ (2 × 108 pfu/0.5 ml) was administered by i.m. injection at each of three sites in the left adductor muscle). Skin overlying injection sites was marked with a permanent black marker. Rabbits were euthanized and muscle samples from marked injection sites were obtained 1, 3, or 7 days after injection (n = 3 rabbits per treatment group per time point).

2.3.2 RNA isolation and quantitative real-time RT-PCR (qRT-PCR)

Tissue samples were immersed in RNAlater (QIAGEN) and stored at −20 °C. The samples were homogenized in Trizol (Invitrogen). Total RNA was extracted, precipitated by addition of isopropanol, and purified using RNeasy Mini columns (QIAGEN) with on-column DNase I digestion. First-strand cDNA synthesis was performed with 5 μg of total RNA in 100-μl reactions using the iScript cDNA Synthesis Kit (BioRad). Real-time PCR was performed in an iCycler (BioRad) using iQ SYBR Green Mix (BioRad). Primer sequences were: 5′-CGGCGACGACCCATTCGAAC-3′ and 5′-GAATCGAACCCTGATTCCCCGTC-3′ (18S rRNA); 5′-CCACAGGACAGTACAGGATG-3′ and 5′-TCAAGTCGTGCTGAATAATACC-3′ (HIF-1α); 5′-TCACCTGCTGCTATACATTCAC-3′ and 5′-GGTTGGCAATGGCATCCTG-3′ (monocyte chemotactic protein-1; MCP-1); 5′-CAGCAGCACTCCGACAAG-3′ and 5′-ATGTAGCCGCCGTCACTC-3′ (PDGF-B); 5′-GCCATGAAGCTGTTCACTTG-3′ and 5′-CATTGAAAGGCACCACTTCC-3′ (PLGF); 5′-GAGAGCCACATCGCCAGAG-3′ and 5′-TTTCGGGTCAATGCACACTTG-3′(SDF-1); and 5′-CGAGACCTTGGTGGACATC-3′ and 5′-CTGCATGGTGACGTTGAAC-3′ (VEGF). For each set of primers, gradient PCR was performed for determination of the optimal annealing temperature. Serial dilutions of the cDNA samples were analyzed to determine the efficiency and dynamic range of the PCR. An assay requirement was that the standard deviation for the cycle threshold (CT) among three to five replicate samples was <0.3. CT was plotted vs. log (nanograms of input RNA) and the best-fit line was constructed. An assay requirement was that the correlation coefficient of the line was >0.99. The slope (m) of the line was used to determine PCR efficiency (E) based on the formula: E = (101/m)−1. In order for samples to be compared, the efficiencies were required to vary by no more than 5%. Thus, stringent criteria were imposed to assure linearity of all qRT-PCR assays. The expression (R) of each target gene mRNA relative to 18S rRNA was calculated based on the formula: R = 2Δ(ΔCT), where ΔCT=CT,targetCT,18S and Δ(ΔCT)=ΔCT,AdCA5−ΔCT,AdLacZ [9].

2.3.3 Immunoblot analysis

Tissue samples taken from the same adductor muscles as for RNA analysis were frozen in liquid nitrogen. The samples were homogenized in lysis buffer containing 20 mM HEPES, 1.5 mM MgCl2, 0.2 mM EDTA, 0.1 M NaCl, 0.2 mM DTT, and protease inhibitors, followed by addition of NaCl to a final concentration of 0.45 M. The supernatant was collected after centrifugation and glycerol was added to a final concentration of 20%. Aliquots (100 μg) were fractioned by SDS-PAGE and subjected to immunoblot analysis using a goat polyclonal anti-GFP antibody (ab6658; Abcam) at 1:1000 dilution, horseradish peroxidase-conjugated anti-goat Ig (Amersham) secondary antibody, and signal development with ECL reagents (Amersham).

2.4 Endovascular occlusion and gene transfer

2.4.1 Carotid arteriotomy

Animals were anesthetized with an i.m. injection of acepromazine maleate (1 mg/kg) and ketamine hydrochloride (8 mg/kg). Supplementary doses of thiopental sodium (12 mg/kg, i.v.) were given as needed. Animals were intubated and allowed to breathe room air. A sterile cutdown was performed and a 3-F sheath (Cook, Inc., Bloomington, IN) was inserted into the left common carotid artery.

2.4.2 Angiography

Angiography was performed on all animals immediately pre- and post-occlusion, and prior to euthanasia (day 14). Angiograms were performed using a 3-F pediatric pigtail catheter (Cook, Inc.) positioned in the distal abdominal aorta 3 cm above the aortic bifurcation. 300 μg of sodium nitroprusside was administered into the abdominal aorta followed by a 10-ml saline bolus. After 1 min, anteroposterior and lateral digital subtraction angiograms of the pelvis and hindlimbs were acquired using a Toshiba Infinix CC-I Interventional Angiography System, with a film rate of 15 frames/s (Toshiba America Medical Systems, Tustin, CA), providing temporal resolution of 66 ms (Fig. 1A). 16 ml of contrast (Omnipaque, Amersham Health, Buckinghamshire, UK) was power injected at 4 ml/s (Medrad, Inc., Indianola, PA).

Fig. 1

Intravascular occlusion of the superficial femoral artery and adenoviral gene therapy. Representative pelvic and hindlimb angiograms of an AdCA5-treated rabbit are shown. (A) Digital angiogram, day 0, before coiling. (B) Anterior–posterior digital spot image showing coils (arrows) in the left femoral artery and six needles (bracketed area) placed on the sites for adenoviral injection. (C) Digital subtraction angiogram, day 0, immediately after coiling. Arrowheads indicate location of coils. (D) Digital subtraction angiogram of an AdCA5-treated rabbit (day 14) showing collateral vessel development in the medial thigh. Arrowheads indicate location of the coils.

2.4.3 Endovascular occlusion

A 3-F straight flush catheter (Cook, Inc.) was advanced over a 0.018-in. guidewire (Transcend Micro-Guidewire, Meditech, Inc., Watertown, MA) into the SFA, just above its bifurcation into the saphenous and popliteal arteries [21]. Six fibered platinum endovascular coils (Vortex, Target-Boston Scientific, Newington, NH) were used to occlude the SFA (Fig. 1B and C). A single 2 × 3-mm coil was placed at the bifurcation of the saphenous and popliteal arteries. Four 2 × 5-mm coils were then inserted from distal to proximal. Finally, a single 2 × 3-mm coil was placed just distal to the lateral circumflex branch of the SFA. Care was taken not to occlude the lateral circumflex artery or the deep femoral artery (DFA) proximally.

2.4.4 Intramuscular delivery of adenovirus

Immediately after endovascular occlusion of the left SFA on day 0, animals were randomly assigned to receive injection of either AdCA5 or AdLacZ (n = 6 rabbits each). Each animal received six i.m. injections in the left adductor muscle (1 × 108 pfu in 0.3 ml at each of six sites). Sites were selected and marked with 25-gauge needles under fluoroscopy prior to injection (Fig. 1B).

2.5 Image analysis of collateral vessels

2.5.1 Angiographic analysis

A blinded observer reviewed images from all three angiographic studies. Quantitative analyses were performed to extract data relating to calf perfusion (angiographic perfusion score), the number of collateral vessels (angiographic vessel score), and the diameters of the vessels supplying the hindlimb (vessel diameter ratios). The angiograms were reviewed using the Toshiba DICOM Viewer for Infinix Celeve/i-series V1.14, which permitted “real-time” review of the entire angiographic run for each animal.

2.5.2 Angiographic perfusion score

A variation of the TIMI frame count [16] was utilized to determine how rapidly collateral vessels perfused vessels distal to the occlusion. The frame number on which opacification of the left SFA just distal to the occlusion occurred and the frame number on which contrast opacified the bifurcation of the right SFA were determined. The difference between the two values reflects the difference in angiographic perfusion between the left and right calves.

2.5.3 Angiographic vessel score

The frame demonstrating maximal arterial opacification of thigh vessels was imported to Adobe Photoshop 6.0 (Adobe Systems, Inc., San Jose, CA). The number of collateral vessels (Fig. 1D) that crossed the left mid-thigh was quantified by three observers blinded to the treatment groups, and the average was calculated. A collateral vessel was defined as having a stem, middle, and re-entry [17].

2.5.4 Vessel diameter ratio

The diameters of the distal hypogastric artery (DHA), distal DFA (DDFA), and lateral femoral circumflex artery (LFCA) were measured in both hindlimbs. Left/right ratios of vessel diameters were generated from the raw data.

2.6 Calf blood pressure (BP) ratio

Calf BP was measured in both hindlimbs pre-occlusion, immediately post-occlusion, and prior to euthanasia on day 14. Under anesthesia, the hindlimbs were shaved and cleaned, the pulse of the posterior tibial artery was identified using a Doppler ultrasound flow detector (811-B; Parks Medical Electronics, Inc., Aloha, OR), and the systolic BP in both limbs was determined using a sphygmomanometer. The mean of three separate measurements was determined for each hindlimb at each time point. The calf BP ratio was defined for each rabbit as the ratio of average systolic pressure in the left hindlimb to average systolic pressure in the right hindlimb.

2.7 Tissue preparation and histological analysis

All animals were euthanized by i.v. injection of thiopental (100 mg/kg). With a straight flush catheter positioned in the distal abdominal aorta, 500 ml of 10% formalin was injected. Both hip joints were disarticulated and each hindlimb was removed and immersed in 10% formalin for 24 h. Whole legs were then cut into 5-mm-thick transverse sections and embedded in paraffin, and 5-μm sections were prepared. Immunohistochemical staining was performed using mouse monoclonal antibody against human CD31 (Dako Cytomation, Inc., Carpinteria, CA) to detect vascular endothelial cells and mouse monoclonal antibody against human α-smooth muscle actin (αSMA; Research Diagnostics, Inc., Flanders, NJ) to detect vascular smooth muscle cells. Microwave irradiation was performed for antigen retrieval. Sections were incubated in 0.3% hydrogen peroxide to block endogenous peroxidase activity. Protein blocking, incubation with secondary biotinylated antibody, and avidin–biotin interaction were performed using a Vectastain ABC kit (Vector Laboratories, Burlingame, CA) and visualized using diaminobenzidine as a chromogen (DAB kit; Vector Laboratories).

Ten fields from two serial sections (at the distal portion of the coiled area) of the left adductor muscle were randomly selected, and the number of capillaries was counted using CD31-stained slides. To avoid overestimation or underestimation of capillary density due to muscle atrophy or edema, the capillary/myocyte ratio was calculated. Arterial vessels with a diameter >100 μm were identified using the αSMA-stained slides under 40 × magnification, and the total number and lumenal area of the arteries in the adductor muscle were calculated using Image-J software (National Institutes of Health, Bethesda, MD). All histologic analyses were performed by a blinded observer.

For detection of EGFP and LacZ expression, adductor muscles were embedded in OCT and 10-μm frozen sections were examined by fluorescence microscopy (Olympus BX5ITF) or fixed in 2% formalin/2% glutaraldehyde, placed in β-galactosidase reporter gene staining solution (Sigma-Aldrich, St. Louis, MO) overnight at 37 °C, and counterstained with hematoxylin and eosin. For detection of HIF-1α and VEGF, frozen sections were fixed with 4% paraformaldehyde, and immunohistochemical staining was performed as described above using primary antibodies against human HIF-1α (1:50, Clone 54; BD Biosciences, Franklin Lakes, NJ) and human VEGF (1:500, C-1; Santa Cruz Biotechnology, Inc., Santa Cruz, CA).

2.8 Statistical analysis

Data are presented as mean ± standard error of the mean (S.E.M.). All p values were calculated with the use of MINITAB statistical software, version 13.1 (Minitab, State College, PA). Paired and unpaired t tests assuming unequal variances were used.

3 Results

3.1 Molecular analysis of adenoviral gene transfer

To demonstrate expression of AdCA5 in non-ischemic muscle, rabbits received an i.m. injection of AdCA5 or AdLacZ (as described in Section 2.3) and adductor muscle samples were harvested 1, 3, or 7 days later (n = 3 rabbits per treatment group per time point).Muscle lysates were subjected to immunoblot assay to detect EGFP, which is co-expressed with HIF-1αCA5 from dual CMV promoters in AdCA5. EGFP expression was present on day 1, maximal on day 3, and low but detectable 7 days after injection (Fig. 2A). EGFP was not detected in rabbits injected with AdLacZ, which only encodes β-galactosidase.

Fig. 2

Protein and gene expression after adenoviral injection into non-ischemic limbs. (A) Immunoblot assay of EGFP expression. Total protein was isolated from left adductor muscles on the indicated day after injection of AdLacZ or AdCA5. Lanes: (1–3) AdCA5, day 1; (4–6) AdLacZ, day 1; (7–9) AdCA5, day 3; (10–12) AdLacZ, day 3; (13–15) AdCA5, day 7; (16–18) AdLacZ, day 7; (19) positive control. Three tissue samples were analyzed from each rabbit. (B) Real-time RT-PCR assays of HIF-1α, MCP-1, PDGF-B, PLGF, VEGF, and SDF-1 mRNA expression. Total RNA was isolated from left adductor muscles at the indicated time after i.m. injection of AdCA5 and AdLacZ. The relative fold induction is defined as the ratio of mRNA expression in the AdCA5-injected muscle to expression in the AdLacZ-injected muscle, as determined by quantitative real-time RT-PCR. Mean and S.E.M. are shown (n = 3 rabbits per treatment group per time point).

Analysis of the same adductor muscles by qRT-PCR using primers that amplified both rabbit and human HIF-1α sequences showed that HIF-1α mRNA levels in AdCA5-treated limbs were maximal on day 3 and were declining but still elevated relative to AdLacZ-treated limbs on day 7 (Fig. 2B). In addition, mRNA levels were elevated on day 3 for MCP-1, PDGF-B, PLGF, and SDF-1, and on day 7 for VEGF. These results demonstrate that AdCA5 increased HIF-1α mRNA levels, leading to increased expression of genes encoding angiogenic growth factors/cytokines in non-ischemic muscle tissue.

All of the subsequent studies involved the analysis of rabbits that were subjected to both endovascular occlusion of the left SFA and injection of AdLacZ (n = 6) or AdCA5 (n = 6) into the adductor muscle along the course of the occluded vessel (as described in Section 2.4). For each of these animals, BP measurements, angiography, and immunohistochemistry were performed as described below.

3.2 Calf BP ratios

The left/right calf BP ratio showed no significant difference before occlusion and fell to zero immediately after occlusion in all rabbits. On day 14, the calf BP ratio was significantly higher in AdCA5–treated as compared to AdLacZ-treated animals (0.89 ± 0.13 vs. 0.51 ± 0.05, p = 0.02; Fig. 3A) due to increased BP in the left calf. There was no significant difference in the right calf systolic BP between groups (data not shown).

Fig. 3

Effect of AdCA5 on blood pressure (BP) and angiographic parameters. Rabbits were studied 14 days after unilateral SFA occlusion and injection of AdCA5 (closed bars) or AdLacZ (open bars). (A) Calf BP ratio performed on days 0 and 14 (*p = 0.02). (B) Angiographic vessel score on day 14. (C) Angiographic perfusion score on day 14 (*p = 0.007). (D) Vessel diameter ratio of the left/right distal deep femoral artery on day 14 (*p = 0.02). Mean and S.E.M. are shown (n = 6 rabbits each).

3.3 Angiographic analysis of collateral blood vessels

Angiographic vessel scores showed no statistical difference in the number of collateral vessels present in the thighs of AdCA5- and AdLacZ-treated animals (12.67 ± 1.50 vs. 12.67 ± 2.03) (Fig. 3B). However, the angiographic perfusion score was significantly lower in the AdCA5-treated group relative to the AdLacZ group (3.50 ± 0.56 vs. 8.33 ± 1.31, p = 0.007) (Fig. 3C), indicating increased perfusion of AdCA5-treated limbs. Left/right ratios of vessel diameters revealed no significant difference in the DHA (1.15 ± 0.07 vs. 1.07 ± 0.05, p = 0.42) or LCFA (1.02 ± 0.07 vs. 1.03 ± 0.12, p = 0.95) between groups, but showed a significantly increased diameter of the DDFA, the vessel adjacent to the site of adenovirus administration, in the AdCA5-treated group (1.84 ± 0.25 vs. 0.93 ± 0.22, p = 0.02) at day 14 (Fig. 3D).

3.4 Histological determination of capillary density and arterial luminal area

Adductor muscle sections from rabbits on day 14 were examined by immunohistochemistry using an anti-CD31 antibody to detect vascular endothelial cells (Fig. 4A and B). Capillary density, defined as the capillary/myocyte ratio, was significantly greater in the AdCA5-treated group than in the AdLacZ-treated group (0.93 ± 0.03 vs. 0.78 ± 0.06, p = 0.04) (Fig. 4C).

Fig. 4

Analysis of angiogenesis. Capillary morphometry was performed by CD31 immunohistochemical analysis of sections from left adductor muscles harvested on day 14 after SFA occlusion and injection of AdCA5 (A) and AdLacZ (B). Magnification, 400 ×; scale bar, 100 μm. (C) Capillary/myocyte ratio on day 14 (*p = 0.04). Mean and S.E.M. are shown (n = 6 rabbits each).

Immunohistochemistry was also performed using anti-αSMA antibodies to detect vascular smooth muscle cells (Fig. 5A and B). The number and total luminal area of collateral arteries with a diameter >100 μm were calculated. While the number of collateral arteries was not significantly different between the two groups (12.4 ± 2.4 vs. 12.5 ± 1.3), the luminal area of arteries in the AdCA5-treated limbs was significantly increased relative to the AdLacZ-treated limbs (0.32 ± 0.05 vs. 0.21 ± 0.03 mm2, p = 0.04) (Fig. 5C and D). These data are consistent with angiographic analysis, which demonstrated no difference in the number of collateral vessels but a significant difference in the diameter of the DDFA and in the perfusion of the ischemic limb.

Fig. 5

Analysis of arterial remodeling. α-Smooth muscle actin immunohistochemistry was performed on sections of adductor muscles injected with AdCA5 (A) and AdLacZ (B) Magnification, 40 ×; scale bar, 1 mm. (C) Mean number of arteries with diameter >100 μm. (D) Total luminal area of the same arteries was determined (*p = 0.04). Mean and S.E.M. are shown (n = 6 rabbits each).

Histochemical assays were also performed to demonstrate a correlation between AdCA5 infection and expression of HIF-1α protein. Adjacent sections of adductor muscle analyzed 3 days after SFA occlusion and AdCA5 injection demonstrated a colocalization of fluorescence, resulting from EGFP expression, and increased HIF-1α expression (Fig. 6A and C). In adjacent sections of adductor muscle from AdLacZ-treated limbs, X-gal histochemistry revealed β-galactosidase activity resulting from LacZ gene expression (Fig. 6B) but no increased expression of HIF-1α (Fig. 6D).

Fig. 6

Expression of adenovirus-encoded gene products. Representative sections of left adductor muscles 3 days after intravascular occlusion and injection of AdCA5 (A and C) or AdLacZ (B and D). EGFP expression was identified by fluorescence microscopy (A). HIF-1α expression was detected by immunohistochemistry (C and D). β-Galactosidase expression was detected by X-gal histochemistry (B). Magnification, 100 ×; scale bar, 400 μm.

In the same tissue samples, immunohistochemistry demonstrated the presence of VEGF protein (Fig. 7C) in EGFP-positive regions (Fig. 7A) of adductor muscle. In contrast, VEGF protein was not detected in adductor muscle from AdLacZ-treated rabbits that expressed β-galactosidase (Fig. 7B and D). It is noteworthy that increased VEGF protein expression was induced 3 days after SFA occlusion and AdCA5 injection (Fig. 7C), whereas VEGF mRNA expression was induced 7 days after AdCA5 injection in the absence of SFA occlusion (Fig. 2B). Taken together, the molecular, histological, and functional data indicate that increased HIF-1α expression in AdCA5-treated adductor muscle leads to increased expression of angiogenic growth factors/cytokines that significantly augment vascular remodeling in response to SFA occlusion.

Fig. 7

Expression of VEGF. Sections of left adductor muscles 3 days after intravascular occlusion and injection of AdCA5 (A and C) or AdLacZ (B and D). EGFP and β-galactosidase expression were demonstrated by fluorescence microscopy (A) and X-gal histochemistry (B), respectively. VEGF protein was detected by immunohistochemistry (C and D). Magnification, 100 ×; scale bar, 400 μm.

4 Discussion

In this paper, we have reported two novel and important results. First, we have utilized a novel non-invasive model of limb ischemia resulting from endovascular occlusion of the SFA as an animal model of PVD. Second, we have demonstrated that i.m. injection of AdCA5, an adenovirus encoding a constitutively active form of HIF-1α, significantly improves the recovery of BP and tissue perfusion and significantly promotes arteriogenesis following arterial occlusion.

4.1 Non-invasive endovascular occlusion

The endovascular model of hindlimb ischemia that we have utilized [21] differs from standard surgical models [18,19,20] in four important respects. First, as endovascular coils are used to occlude the artery from within, this model obviates the need for surgical procedures, which disrupt pre-existing collateral vessels around the operated area. Second, the wound healing response in surgical models leads to the production of numerous cytokines, which may confound results. Third, the animal does not suffer pain or impaired motility of the affected limb after endovascular procedures. Fourth, our model results in less ischemia in the affected limb. Prior models have rendered both the thigh and the calf ischemic. In our model only, the SFA is occluded, whereas flow through both the DFA and LFCA is maintained so that perfusion of the thigh is not compromised. Arteriogenesis, remodeling of arteries to increase regional perfusion, can occur in the absence of ischemia under the condition that the molding forces (e.g., circumferential wall stress, fluid shear stress) exist [22]. If the medial thigh muscles are severely ischemic, these target tissues of adenovirus injection can produce a variety of cytokines and modify the expression of genes encoding angiogenic and arteriogenic factors in the absence of gene transfer and thus confounding the analysis. Other investigators have recognized the limitation of rendering the thigh ischemic and have used surgical methods that limit ischemia to the calf only [23,24].

4.2 AdCA5 administration improves the recovery of blood flow by stimulating both angiogenesis and arteriogenesis

We demonstrate for the first time that adenoviral vector-mediated overexpression of a constitutively active form of HIF-1α can improve the blood flow of ischemic limbs following endovascular occlusion by inducing both angiogenesis and arteriogenesis. Although it has been reported that a fusion protein containing HIF-1α and herpes virus VP-16 sequences can enhance angiogenesis in an operative model of rabbit hindlimb ischemia [14], no assessment of arteriogenesis was performed. In our model, functional evidence of increased arterial remodeling in response to AdCA5 injection included significantly increased calf BP and decreased angiographic perfusion time 14 days after occlusion in the AdCA5-treated group. Angiographic and histological evidence of arterial remodeling were also observed. There was no significant difference in the number of collateral vessels detected by angiography in the AdCA5- and AdLacZ-treated groups, but in the AdCA5-treated group, the DDFA, the vessel adjacent to the site of vector administration, was significantly larger in diameter. Furthermore, histological examination revealed that the luminal area of arteries was significantly increased in sections from the AdCA5-treated animals.

Capillary density was also greater in the AdCA5-treated group, indicating that HIF-1 can promote both angiogenesis and arteriogenesis. However, the difference in angiogenesis between groups, although statistically significant, was modest and probably not physiologically meaningful since the defect in blood supply was due to the occlusion of a large conduit vessel. The term arteriogenesis refers to the process by which the luminal diameter of preexisting arterioles is increased to provide collateral sources of blood flow in response to critical narrowing of a major artery [1,21]. This process should be clearly distinguished from angiogenesis, which refers to the formation of capillary branches by sprouting from preexisting capillaries. Angiogenesis is important for the improvement of local tissue perfusion in order to match O2 supply and demand, which occurs during cell proliferation or hypertrophy. However, an increased number of capillaries in the distal vascular bed cannot substitute for obstruction of a proximal conduit vessel [1,22,25].

Several angiogenic genes/proteins have been evaluated for therapeutic angiogenesis in hindlimb and cardiac ischemia models, including FGF-2, hepatocyte growth factor, MCP-1, PDGF-B, PLGF, SDF-1, and VEGF [26]. VEGF has been most extensively studied. Animal studies have indicated that administration of VEGF may augment vascularity and perfusion in ischemic tissues. However, overexpression of VEGF alone resulted in increased numbers of leaky vessels with tissue edema and inflammation [27,28]. Clinical trials of VEGF therapy have demonstrated safety, but Phase II trials have failed to show its efficacy [29]. Overexpression of both VEGF and angiopoietin-1 augments vessel formation without excessive permeability [27], and administration of PDGF-B and FGF-2 synergistically promotes functional and stable vascular networks in animal models [30]. Both MCP-1 and PLGF have been shown to promote arteriogenesis in preclinical studies [31,32].

These results suggest that the establishment of a stable and functional vessel network requires the orchestrated expression of multiple angiogenic factors. We previously demonstrated that intravitreous injection of AdCA5 activates the expression of genes encoding the angiogenic growth factors VEGF, PLGF, PDGF-B, ANGPT1, and ANGPT2 [9]. In this study, we have demonstrated that AdCA5 can induce increased expression of PDGF-B, PLGF, MCP-1, SDF-1, and VEGF mRNA in non-ischemic adductor muscle. The concerted expression of these factors provides a molecular basis for the observed arteriogenic and angiogenic responses that were induced by AdCA5, resulting in improved perfusion in our endovascular occlusion model. Given these findings in a model that resembles atherosclerotic obstruction of peripheral arteries in patients, further studies are warranted to determine whether AdCA5 administration may represent a novel treatment option for patients with extensive PVD who are not candidates for conventional therapies.


We thank Andrea Gambotto and Susan Schoonover for large-scale adenovirus preparations; William Baldwin III and Rene Rodriguez for advice on histology; and Brian Kelly, Carolyn Magee, and Lori Piptone for technical assistance.

This work was supported by grants R01-HL55338 and P01-HL65608 to G.L.S. from the National Institutes of Health.


  • 1 T.H.P. and H.K. contributed equally to this study.

  • Time for primary review 36 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
View Abstract