Impaired Collateral Vessel Development Associated With Reduced Expression of Vascular Endothelial Growth Factor in ApoE−/− Mice
Background—The impact of disordered lipid metabolism on collateral vessel development was studied in apolipoprotein (apo)E−/− mice with unilateral hindlimb ischemia.
Methods and Results—Hindlimb blood flow and capillary density were markedly reduced in apoE−/− mice versus C57 controls. This was associated with reduced expression of vascular endothelial growth factor (VEGF) in the ischemic limbs of apoE−/− mice. Cell-specific immunostaining localized VEGF protein expression to skeletal myocytes and infiltrating T cells in the ischemic limbs of C57 mice; in contrast, T-cell infiltrates in ischemic limbs of apoE−/− mice were severely reduced. The critical contribution of T cells to VEGF expression and collateral vessel growth was reinforced by the finding of accelerated limb necrosis in athymic nude mice with operatively induced hindlimb ischemia. Adenoviral VEGF gene transfer to apoE−/− mice resulted in marked augmentation of hindlimb blood flow and capillary density.
Conclusions—These findings thus underscore the extent to which hyperlipidemia adversely affects native collateral development but does not preclude augmented collateral vessel growth in response to exogenous cytokines. Moreover, results obtained in the apoE−/− and athymic nude mice imply a critical role for infiltrating T cells as a source of VEGF in neovascularization of ischemic tissues.
We sought to test the hypothesis that disordered lipid metabolism retards collateral vessel growth to ischemic tissues. The apolipoprotein (apo)E-deficient mouse generated by gene targeting1 2 provides an appropriate animal model for such a study. ApoE is a 34-kDa glycoprotein that coats VLDLs, IDLs, and HDLs. It serves as a high-affinity ligand for LDL and putative chylomicron remnant receptors,1 2 thus facilitating the removal of apoE-coated lipoprotein particles from the circulation. In apoE-null (apoE−/−) mice, cholesterol is markedly elevated because of increased levels of VLDLs and IDLs.1 2
In the present series of experiments, we created unilateral hindlimb ischemia in apoE−/− mice and compared the development of collateral vessels in the ischemic limb with that observed in C57 wild-type mice that underwent similar surgery. We observed a marked retardation of collateral vessel development in apoE−/− mice, associated with reduced expression of vascular endothelial growth factor (VEGF) in the ischemic limb. The reduction in endogenous VEGF synthesis appears to result from reduced VEGF expression by skeletal myocytes, as well as a paucity of T-cell infiltrates expressing VEGF. Although hyperlipidemia adversely affected native collateral development, it did not preclude augmented collateral vessel growth in response to supplemental administration of VEGF achieved by adenoviral gene transfer.
Mouse Model of Unilateral Hindlimb Ischemia
Unilateral hindlimb ischemia was induced surgically in each inbred strain as previously described.3 This surgery typically results in oxygen saturations as low as 64% measured at day 0 and myonecrosis in the ischemic hindlimb.3
Laser Doppler Perfusion Imaging
Serial assessment of hindlimb blood flow was performed with a PIM 1.0 laser Doppler perfusion imager (Lisca) as previously described.3 4 Calculated perfusion was expressed as a ratio of left (ischemic) to right (normal) limb.
Previous studies3 established that laser Doppler flow velocity correlated with capillary density in the ischemic limb in C57 mice in which angiogenesis was retarded by administration of neutralizing antibodies prepared against VEGF as well as the angiogenesis inhibitor platelet factor 4 (PF-4). Double immunolabeling for bromodeoxyuridine (BrdU) and CD-31 established that endothelial cell (EC) proliferation peaks at 7 days and is subsequently reduced at days 14 and 21. In mice treated with PF-4 and euthanized 14 days after surgery, capillary density (268±195 versus 1053±371 capillaries/mm2, P<0.01) and EC proliferation (16±29 versus 935±239 BrdU-positive cells/mm2, P<0.01) were significantly reduced in PF-4– versus PBS-injected mice, respectively.
For immunohistochemistry, tissues were fixed in 100% methanol. A minimum of 3 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.
Five-micrometer, paraffin-embedded sections cut transversely from the mouse hindlimb were used for immunohistochemistry. ECs were identified with a rat monoclonal antibody (mAb) against mouse CD-31 (Pharmingen). Staining with CD-31 antibody,5 and in particular this anti-mouse specific CD31 antibody,6 7 has previously been shown to be limited to capillary endothelium.
VEGF was detected with a rabbit polyclonal antibody (pAb) prepared against human VEGF amino-terminal peptides 1 to 20 (Santa Cruz). The amino-terminal peptides 1 to 20 for mouse and human VEGF are similar except that a Thr (in 11) in the mouse replaces the Ser of the human sequence. This antibody immunoprecipitates a 46-kDa protein (nonreducing conditions) or a 23-kDa protein (reducing conditions) from mouse uterine protein extract corresponding to the size of recombinant human VEGF165 used as a control. Immunoperoxidase staining was performed as before.3 Negative control slides were prepared by substituting preimmune rat serum for the primary mAb or using preabsorption of VEGF Ab with the specific peptide (Santa Cruz).
For measurement of capillary density, tissues were prepared as described by Weidner et al.5 Two different sections were taken 3 mm apart between the knee and ankle. Capillaries were counted per 30 randomly chosen high-power fields (HPFs) on the 2 sections (3 animals per time point). The 30 HPFs (×60) constituted the complete muscle set of the hindlimb. Given the size (in mm2) of the HPFs and the magnification used, the results were calculated as capillaries per square millimeter.
Total leukocytes were identified with anti–CD-45 antibody (Gibco BRL), macrophages with F4/80 antibody (Caltag Labs), and T lymphocytes with anti–CD-3 antibody (Sigma Chemical Co). T cells and macrophages identified by cell-specific immunostains were manually counted in 90 randomly chosen HPFs of sections from the ischemic hindlimb of 3 animals per time point killed at postoperative days 2, 4, 7, and 14.
In Situ Hybridization
In situ hybridization was carried out with the nonradioactive Genius Labeling and Detection Kit (Boehringer Mannheim) as previously described.3 Controls included (1) hybridization with sense probe and (2) use of an irrelevant antibody. Sections were counterstained with methyl green nucleus–specific stain (Vector).
Five milligrams of protein pooled equally from diced frozen hindlimbs of 2 different animals was incubated with VEGF pAb (0.5 μg/mL) (Santa Cruz) or anti–Flk-1 pAb (0.5 μg/mL) (Santa Cruz) overnight as described previously.3 Positive controls included recombinant VEGF165 (Genentech). Negative controls included blotting with an irrelevant antibody. Experiments were performed in triplicate to ensure reproducibility.
Fluorescence-Activated Cell Sorting Analysis
Mononuclear cells were isolated from peripheral blood with the Ficoll Hypaque system and then washed in PBS. We used specific anti-mouse antibody (Pharmingen) conjugated with phycoerythrin (CD-3) or FITC (CD-4 and CD-8). Mononuclear cells were incubated with anti–CD-3 and CD-4 or anti–CD-3 and CD-8 in medium for 1 hour at 4°C, washed 3 times with cold medium to remove unbound antibody, and further fixed in 2% paraformaldehyde. After additional washes, the samples were analyzed with a fluorescence-activated cell sorting analyzer.
Intramuscular Adenoviral Transfection
Mice were transfected with E1-deleted recombinant adenovirus (4×109 pfu in 20% pluronic gel) expressing either LacZ containing a nuclear localization sequence nls-LacZ or murine VEGF cDNA. Transfection was performed by direct intramuscular injection into the ischemic hindlimb with a 27-gauge needle.
Recovery of Hindlimb Blood Flow Is Retarded in ApoE−/− Mice
In C57 controls (n=10), blood flow was precipitously reduced after surgery (Figure 1⇓), remained impaired for 7 days, increased to ≈55% of the nonischemic limb by day 14, and ultimately returned to near-normal levels by day 35. A similarly precipitous reduction in hindlimb blood flow occurred in apoE−/− mice, but in contrast to C57 mice, flow recovery was markedly attenuated (n=10). In apoE−/− mice, a statistically significant reduction in flow was evident by day 14, persisted at each of the subsequent time points (21, 28, and 35 days), and never achieved >55% of that measured for the contralateral normal limb.
Capillary Density Is Reduced in ApoE−/− Mice
A precipitous reduction in capillary density was observed in both strains of mice immediately after surgery (Figure 2⇓) as a result of ischemic necrosis associated with resection of the femoral artery. By day 4, capillary restitution in C57 ischemic limbs exceeded that in apoE−/− mice. A statistically significant reduction in capillary density per square millimeter of ischemic tissue was observed at days 4, 7, 14, 21, and 35 (Figure 2⇓). In either strain, vascularity was essentially limited to apparently viable muscle, as opposed to foci of fibrosis.
Protein extracts from ischemic hindlimb muscles were obtained at days 4, 7, 14, 21, and 35 after surgery and hybridized with an antibody against mouse CD-31. Western blot analysis confirmed that CD-31 (130-kDa) protein was reduced at early time points in apoE−/− compared with C57 mice (Figure 2⇑).
Flk-1 Receptor Expression Is Not Reduced Out of Proportion to the Reduction in CD31-Positive ECs in ApoE−/− Mice
Immunoprecipitation studies indicated that Flk-1 expression was reduced in tissue specimens retrieved from the ischemic limb of apoE−/− versus control C57 mice, but only in proportion to the reduced EC population indicated by Western blot analysis of CD31 in apoE−/− versus control C57 mice (Figure 3⇓); specifically, temporal expression of EC Flk-1 protein in ischemic limbs of C57 and apoE−/− mice corresponded to sequential recovery of capillary density. Antiphosphotyrosine blot of the Flk-1 immunoprecipitation confirmed that the Flk-1 immunoprecipitate was functional as early as day 4 in both C57 and apoE−/− mice, with an increase in phosphorylation at later time points (data not shown).
VEGF Protein and mRNA Are Reduced in Tissues Retrieved From Ischemic Limbs of ApoE−/− Mice
We considered that impaired angiogenesis in apoE−/− mice might be related to reduced endogenous VEGF. Indeed, development of peak VEGF protein levels in ischemic limb tissue was both delayed and reduced in apoE−/− versus control mice (Figure 3⇑).
Tissue specimens harvested from ischemic limbs of apoE−/− and C57 mice were also immunostained for VEGF. Consistent with the results of immunoprecipitation/Western analysis, VEGF was less abundant in tissue sections from apoE−/− mice compared with controls. In C57 controls, VEGF immunostaining could be localized to skeletal myocytes in the ischemic limb and foci of inflammatory cells. In apoE−/− mice, VEGF immunostaining of skeletal myocytes was less abundant (Figure 4⇓). Moreover, we observed a striking reduction in VEGF immunostaining contributed by infiltrating inflammatory cells (Figure 4⇓). Immunostaining of adjacent, serially cut sections for total leukocytes (anti–CD-45 antibody), macrophages (F4/80 antibody), and T lymphocytes (anti–CD-3 antibody) indicated that the inflammatory infiltrates in the C57 controls included a large T-cell population, beginning within 2 days after surgery (1.02±0.02 cells/HPF) and diminishing to background levels by day 14 (0.14±0.02 cells/HPF). In contrast, ischemic tissues harvested at day 2 from apoE−/− mice displayed a marked reduction in infiltrating T cells (0.20±0.02 cells/HPF, P=0.0021) (Table⇓, Figure 5⇓).
In situ hybridization using murine VEGF165-specific digoxigenin-labeled probes was performed 7 days after surgery. By use of VEGF165-antisense probe in C57 mice, VEGF mRNA was detectable in ischemic muscle fibers (Figure 6⇓), which corresponded to immunostaining for VEGF protein. VEGF mRNA was barely detectable in legs of control animals that had not undergone surgery (not shown). In contrast, both VEGF protein and mRNA expression were markedly reduced in the skeletal muscle sections from apoE−/− mice. Use of VEGF sense control probe did not result in an identifiable signal (Figure 6⇓).
Absolute T-Cell Count Is Not Reduced in ApoE−/− Mice
Because histopathological examination disclosed a paucity of infiltrating T lymphocytes in the ischemic limbs of apoE−/− mice, flow cytometry was performed on blood obtained from C57 and apoE−/− mice (n=16) to determine expression of CD-3, CD-4, and CD-8.
The percentage of circulating CD-3–positive cells was reduced by 26.8% in apoE−/− mice compared with nontransgenic mice of similar (C57) background (26.0±1.8% versus 35.5±4.1%, P=0.06). The mean number of total leukocytes among apoE−/− mice, however, was more than that measured for the C57 mice (7990±850 versus 3930±280, P=0.001). Thus, the total number of T cells in the apoE mice was not reduced but in fact exceeded the number measured in the C57 wild-type mice. With regard to T-cell subtypes, the CD-4/CD-8 ratio was 1.43±0.2 for apoE−/− mice versus 1.40±0.1 in C57 controls.
Angiogenesis Is Retarded in Nude Mice
If T cells constitute a critical source of VEGF, we reasoned that angiogenesis should be similarly retarded in nude mice. Indeed, in congenitally athymic C57BL/6J-nu mice (Jackson Laboratory), routine femoral arterial resection led to early (<10 days) and extensive limb necrosis (Figure 7⇓). To permit limb survival beyond 10 days, the femoral artery was excised slightly more distally; this extended limb survival to 14 days after surgery. Recovery of blood flow, however, remained severely impaired, resulting in toe necrosis by day 15, at which time the experiments were terminated.
Necropsy examination 14 days after surgery disclosed a marked reduction in capillary density in nude compared with C57 mice: CD-31 immunostaining showed extensive territory lacking capillaries with only rare foci of neovascularization (Figure 8⇓). In the absence of T cells in the ischemic limbs of nude mice, the macrophage component of the inflammatory cell infiltrate was increased in comparison to C57 mice (Table⇑, Figure 5⇑). Immunostaining of sections from these same tissues disclosed a marked decrease in VEGF expression within 14 days of follow-up compared with control strains (Figure 8⇓).
Intramuscular Adenoviral Murine VEGF Gene Transfer Augments Angiogenesis in ApoE−/− Mice
If angiogenesis is retarded in apoE−/− mice as a result of reduced VEGF expression, then VEGF replacement should augment angiogenesis. Ischemic hindlimb skeletal myocytes of apoE−/− mice transfected with adenoviral murine VEGF in fact displayed augmented VEGF expression as early as day 3 (Figure 9⇓) compared with apoE−/− mice transfected with nls-LacZ. Within 21 days, hindlimb blood flow was improved in apoE−/− mice transfected with VEGF compared with nls-LacZ. Over 35 days, blood flow recovery in VEGF-transfected apoE−/− mice was equivalent to that in C57 mice (Figure 10⇓).
These findings provide inferential evidence that collateral vessel development is impaired in the setting of hypercholesterolemia. Hypercholesterolemia due to an inherited deficiency of apoE lipoprotein was associated with retarded restoration of blood flow to an ischemic limb. Although capillary density was progressively augmented in C57 control mice, capillary density in the ischemic limb of apoE−/− mice was, by comparison, persistently reduced to a statistically significant degree.
The apparent reduction in collateral vessel development in the apoE−/− mice was associated with a relative reduction in the levels of VEGF protein in the ischemic limb compared with C57 mice. This finding is consistent with previous studies in which targeted disruption of the gene encoding for VEGF leads to impaired angiogenesis.8 9 The apparent restoration of collateral vessel development by constitutive overexpression of VEGF in apoE−/− mice confirms the critical role of VEGF in directing hindlimb collateral vessel development. Moreover, these experiments suggest that even in the face of disordered lipid metabolism, VEGF receptor expression and function are sufficiently preserved to permit collateral vessel augmentation in response to exogenously administered ligand.
Immunostaining of tissue sections harvested from ischemic limbs of the C57 mice clearly demonstrated that skeletal myocytes constitute an important source of endogenous VEGF synthesis. Smooth muscle cells,10 ischemic myocardium,11 and circulating monocytes12 have been shown to express VEGF. In addition, however, cell-specific immunostaining demonstrated large numbers of T cells among inflammatory cell infiltrates in the ischemic limb of the C57 mice; in contrast, few T cells were observed in tissue sections from the ischemic hindlimbs of apoE−/− mice.
The finding that leukocytes may synthesize and export angiogenic cytokines has been reported previously.13 14 15 16 Blotnick et al13 showed that cultured T lymphocytes isolated from normal human peripheral blood could synthesize heparin-binding epidermal-like growth factor (primarily from CD4+ cells) and basic fibroblast growth factor (bFGF, from CD4+ and CD8+ cells). Freeman et al15 subsequently showed that CD4+ and CD8+ cells could export bioactive concentrations of VEGF as well; moreover, tumor-infiltrating lymphocytes in tissue sections cut from bladder cancers were shown to stain positively for VEGF protein. These results were confirmed at the mRNA level by Iijima et al.16
To specifically investigate the relative magnitude of VEGF contributed by T cells to restoration of blood flow in the ischemic limb, we performed the identical surgical procedure used to establish hindlimb ischemia in the C57 and apoE−/− mice on athymic nude mice. None of the latter survived beyond 10 days without experiencing severe limb necrosis. Even when the surgical procedure was modified by excision of a shorter segment of the femoral artery, toe necrosis routinely developed within 15 days. These experiments thus suggest that VEGF contributed by T cells in the setting of tissue ischemia is critical for timely development of collateral vessel growth and restoration of blood flow to the ischemic district.
Although athymic nude mice are by definition T cell–deficient, this was not the case in the apoE−/− mice. The absolute T-cell number in apoE−/− mice was in fact 92.8% higher than in C57 mice. Absent reduced T-cell numbers, the marked reduction in T-cell infiltrates seen in tissue sections from the apoE−/− ischemic limbs implies an associated defect in T-cell function. Although not previously associated with disordered lipid metabolism, T-cell dysfunction acquired with age17 constitutes a potential precedent for the present findings. Stohlawetz et al18 in fact showed that transendothelial migration of human T cells may be compromised in elderly versus young subjects, although the basis for this defect in transmigration was not disclosed. More recently, several groups19 20 21 have shown that T-cell recruitment is in part determined by preserved affinity for specific cell adhesion molecules. It is important to underscore the fact that VEGF expression in T cells is upregulated by hypoxia15 ; thus, failure of T cells to exit the circulation and infiltrate the ischemic tissue precludes what may be an important compensatory mechanism for achieving site-specific increase in VEGF synthesis.
The principle that disordered lipid metabolism may impair angiogenesis is further supported by other recent experimental findings. Murugesan and Fox22 established that oxidized LDL inhibits EC motility. Chen et al23 24 showed that bFGF reverses atherosclerotic impairment of capillary-like microtubes, which develop in vitro from explants of human coronary arteries. Thus, these preliminary reports, together with the present findings, provide compelling evidence that lipid disorders may retard angiogenesis but do not preclude a favorable response to cytokine therapy.
This study was supported in part by NIH grants HL-02824, HL-53354, and HL-57516 to Dr Isner and grants from the University Hospital of Bordeaux and the French Ministry of Foreign Affairs (La Voisier grant) to Dr Couffinhal.
- Received December 31, 1998.
- Revision received February 19, 1999.
- Accepted March 16, 1999.
- Copyright © 1999 by American Heart Association
Zhang SH, Reddick RL, Piedrahita JA, Maeda N. Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science. 1992;258:468–471.
Vecci A, Garlanda C, Lampugnani MG, Resnati C, Matteucci C, Stoppacciaro A, Schnurch H, Risau W, Ruco L, Mantovani A, Dejana E. Monoclonal antibodies specific for endothelial cells of mouse blood vessels: their application in the identification of adult and embryonic endothelium. Eur J Cell Biol. 1994;63:247–254.
Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig M, Vandenhoeck A, Harpal K, 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.
Blotnick S, Peoples GE, Freeman R, Eberlein TJ, Klagsbrun M. T lymphocytes synthesize and export heparin-binding epidermal growth factor-like growth factor and basic fibroblast growth factor, mitogens for vascular cells and fibroblasts: differential production and release by CD4+ and CD8+ T cells. Proc Natl Acad Sci U S A. 1994;91:2890–2894.
Fromer CH, Klintworth GK. An evaluation of the role of leukocytes in the pathogenesis of experimentally induced corneal vascularization. Am J Pathol. 1976;82:157–170.
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.
Miller RA. The aging immune system: primer and prospectus. Science. 1996;273:70–74.
Gerszten RE, Luscinskas FW, Ding HT, Dichek DA, Stoolman LM, Gimbrone MA, Rosenzweig A. Adhesion of memory lymphocytes to vascular cell adhesion molecule-1–transduced human vascular endothelial cells under simulated physiological flow conditions in vitro. Circ Res. 1996;79:1205–1215.
Stemme S, Holm J, Hansson GK. T lymphocytes in human atherosclerotic plaques are memory cells expressing CD45RO and the integrin VLA-1. Arterioscler Thromb. 1992;12:206–211.
Chen C-H, Cartwright J Jr, Li Z, Lou S, Nguyen HH, Gotto AM Jr, Henry PD. Inhibitory effects of hypercholesterolemia and ox-LDL on angiogenesis-like endothelial growth in rabbit aortic explants: essential role of basic fibroblast growth factor. Arterioscler Thromb Vasc Biol. 1997;17:1303–1312.