Inhibitory Effects of Hypercholesterolemia and Ox-LDL on Angiogenesis-like Endothelial Growth in Rabbit Aortic Explants
Essential Role of Basic Fibroblast Growth Factor
Abstract Hypercholesterolemic (HC) rabbits exhibit suppressed compensatory vascular growth after restriction of arterial supply. However, neovascularization is commonly found in atheromas containing inflammatory cells. We used an in vitro model to determine the effects of hypercholesterolemia on angiogenesis in the absence or presence of inflammatory cells. HC rabbit aortic explants (1 mm2) with or without (n=90 each) lesion-forming inflammatory cells were cultured in a collagen matrix with serum-free medium. Explant-derived endothelial cell growth was organized into capillary-like microtubes (CLM) that could be videomicroscopically quantified. CLM growth from lesion-free HC explants was significantly reduced to 13±4% of the value in explants (n=90) from normocholesterolemic (NC, n=15) rabbits (P<.001). In contrast, in lesion-containing HC explants, the matrix was invaded by foam cells, and CLM growth was not inhibited. Immunoassayable basic fibroblast growth factor (bFGF, in pg/mL) in the culture medium was significantly lower in lesion-free HC (<5) than NC explants (11±2, P<.01) or HC explants with lesions (14±3). In addition, CLM growth was reduced in NC explants incubated with oxidized LDL (ox-LDL, 50-100 μg/mL). Exogenous bFGF (10 ng/mL) reversed the inhibitory effects of hypercholesterolemia and ox-LDL, whereas bFGF-neutralizing antibody (10 μg/mL) abolished CLM growth in all groups. In cultured rabbit aortic endothelial cells, ox-LDL reduced DNA synthesis, but this inhibition was reversed by bFGF. We conclude that hypercholesterolemia and ox-LDL inhibit angiogenesis-like endothelial growth because of a suppressed availability of endogenous bFGF. Retained responsiveness to exogenous bFGF suggests that inducing bFGF expression at targeted sites may improve collateral growth in hyperlipidemic arterial disease.
Presented in part in abstract form at the 41st American College of Cardiology Annual Scientific Sessions, April 12-16, 1992, Dallas, Tex; the 65th Scientific Sessions of the American Heart Association, November 16-19, 1992, New Orleans, La; and the 10th International Symposium on Atherosclerosis of the International Atherosclerosis Society, October 9-14, 1994, Montreal, Quebec, Canada.
- Received May 31, 1996.
- Accepted November 11, 1996.
Like luminal expansion, angiogenesis can represent compensatory vascular remodeling to help maintain organ perfusion in the presence of insufficient arterial supply. The response has been described in humans1 2 and in experiments with animals.3 4 The mechanisms of compensatory angiogenesis are a focus of research interest, in particular because of potential therapeutic implications.
Recent studies have demonstrated that administration of growth factors such as bFGF and VEGF can stimulate angiogenesis in NC animals.5 6 7 However, whether the release and effector functions of these growth factors remain intact or are impaired in the presence of dyslipidemic cell injury has not been fully evaluated. Previously, we demonstrated that adaptive growth of arteries and microvessels in response to surgical restriction of arterial supply was markedly suppressed in HC rabbits.8 In a more recent study, we showed that atherosclerotic impairment of angiogenesis in human coronary arterial explants can be reversed by exogenous bFGF.9
The present in vitro experiments with tissue explants from rabbits were designed to further characterize the role of bFGF in angiogenesis in a dyslipidemic environment. Neovascularization may occur in atheromatous lesions,10 but such localized angiogenesis is stimulated predominantly by the inflammatory process of lesion formation.11 12 Compensatory vascular growth usually originates in lesion-free areas. We quantified angiogenesis, release of endogenous bFGF, effects of exogenous bFGF, and effects of bFGF-neutralizing antibody in explants with and without atheromatous lesions from HC rabbits. Also, using explants from NC rabbits, we assessed these variables according to exposure or nonexposure to ox-LDL. Oxidatively modified lipoproteins suppress endothelial replication,13 14 15 a phenomenon that could limit vascular growth in vivo.
A rabbit aortic explant culture model was used so that interactions among dyslipidemia, mononuclear inflammatory cells (macrophage foam cells), and bFGF could be assessed in the absence of hemodynamic or systemic humoral factors. The in vitro model permitted us to compare the effects of growth factors and their neutralizing antibodies on organized endothelial growth in normal and atherosclerotic arterial tissue that was uninfluenced by the conventional cell culture conditions necessary to support endothelial cells. By use of an ELISA, explant-derived bFGF in each well was quantified and compared with endothelial response.
Male New Zealand White rabbits (2.5 to 3.0 kg) were randomized for 8 weeks to standard (n=15) or 1% cholesterol–supplemented (n=15) rabbit chow (Nutrition Biochem), yielding NC and HC groups, respectively (mean±SE plasma total cholesterol, 1.2±0.3 and 17.5±2.3 mmol/L, respectively; P<.001). The rabbits were killed after 8 weeks with pentobarbital (100 mg/kg body wt IV), and the descending thoracic aorta was rapidly excised.
Preparation of Native and Ox-LDL
Native LDL (d=1.019 to d=1.063 g/mL) from pooled human plasma anticoagulated with EDTA (1 mg/mL) was isolated by sequential ultracentrifugation.16 17 To prepare ox-LDL, native LDL was exposed for 24 hours at 37°C to cultured RAECs (a gift of Dr Joseph Witztum, University of California, San Diego) or 5 μmol/L CuSO4 in PBS.16 17 After incubation, LDL was reisolated by ultracentrifugation (flotation at d=1.210 g/mL) with 1 mg/mL EDTA. Preparations were dialyzed against Krebs-Henseleit buffer formulated as previously described.17 Precautions were taken to prevent endotoxin contamination during lipoprotein isolation and oxidation, including the use of pyrogen-free sterile water for all reagents. All glassware was washed with 1% E-Toxa-Clean in sterile water, rinsed with sterile water, and baked at 185°C for 5 hours.18 Endotoxin contamination was monitored with the coagulation Limulus amebocyte lysate assay using Pyrotell sensitivity (λ=0.25 EU/mL) as the standard in serial dilutions (Associates of Cape Cod, Inc). The endotoxin concentration of the ox-LDL preparations was <0.1 EU·mL−1·mg−1 protein. Protein content in LDL preparations was estimated by the method of Lowry et al with bovine albumin as the standard. Thiobarbituric acid–reactive substances in LDL preparations were assayed as a measure of oxidative lipid modification.16 17
The organ culture technique used in this study was developed independently9 but has features in common with the aortic ring culture models reported by Nicosia, Kawasaki, and their associates.19 20 21 22 After removal of the adventitia under microscopic visualization, intima-media explants (1 mm2) were cut from the thoracic aorta with the use of a McIlwain tissue chopper (The Mickle Laboratory Engineering Co Ltd) set at 1 mm for both dimensions. Rat tail collagen type I solution (Collaborative Biomedical Products), 10× RPMI 1640 medium (GIBCO Laboratories), and 240 mmol/L NaHCO3 (8:1:1, vol/vol/vol) were mixed at 0°C. After the pH was adjusted to 7.4 with 1N NaOH, the reconstituted solution (300 μL per well, Corning 24-well culture plates) gelled in 10 minutes at 37°C. Single explants placed endothelial side down in the center of each well were covered with another 300 μL of collagen mixture, which entrapped the explants between two layers of collagen matrix. Serum-free RPMI 1640 medium (500 μL per well) containing penicillin (100 IU/mL) and streptomycin (100 μg/mL) was added. Cultures were incubated at 37°C in a 95% O2/5% CO2 atmosphere, and the medium was changed every other day. All explants were cultured for 14 days unless specified otherwise. After 2 or 3 days, sprouting CLM invaded the thin layer of collagen substrate and extended away from the edges in the plane of the fused collagen layers. CLM growth was visualized by serial video-microscopic imaging with a Nikon inverted microscope and a Cohu model 5172 video camera. Video images were analyzed by computer equipped with a Targa M8 video digitizing board and installed with a JAVA image analysis program (Jandel Scientific). The lengths of individual microtubes, including their branches, were determined by delineating their axes on the video screen with an electronic cursor. The sum of lengths of the axes was computed as an index of angiogenesis-like response, a procedure that combines criteria of CLM length and density that have been used previously in the assessment of CLM outgrowth from rat aortic rings.19 20 21 22 Because CLM arises exclusively from the cut edge of the explant,9 21 the sum of CLM lengths was initially normalized for explant circumference. However, the ratios of explant circumference to explant were, on average, very similar in different groups. Therefore, unless specified otherwise, in this study the index is expressed as the sum of CLM lengths per explant.
Histology and Immunochemistry
Collagen containing the explants after selected culture periods and tissue bordering the explants were fixed in 4% paraformaldehyde and embedded in paraffin. Deparaffinized sections were stained with hematoxylin and eosin for histology or incubated with specific antibodies for immunochemistry. To identify endothelial, smooth muscle, and rabbit macrophage cells,23 the sections were incubated for 60 minutes with antibodies to vWF (1:200, Incstar), α,γ-actin (HHF35 1:1, Enzo), and RAM11 antigen (1:100, Dako), respectively. An immunoperoxidase procedure (Vectastain, Vector Labs) performed in conjunction with species-specific biotinylated secondary antibodies was used to visualize bound primary antibody. Control stains were performed by using primary antibodies to irrelevant antigens or omitting the primary antibody. To visualize the antigen-antibody complex, sections were incubated with a diaminobenzidine solution, rinsed in distilled water, and counterstained with hematoxylin.22 To demonstrate lipid inclusion in macrophages, fresh tissues were stained with oil red O.
Transmission Electron Microscopy
To evaluate the ultrastructure of cell outgrowth from the explants, fragments of collagen containing the explants were fixed in 3% glutaraldehyde buffered with 0.1 mol/L PIPES overnight. Specimens were then rinsed in PIPES buffer, treated with 2% OsO4 for 1 hour, and embedded in Spurr’s resin after serial dehydration. Serial 1-mm sections stained with toluidine blue/basic fuchsin were examined by light microscopy. Sections containing endothelial microtubes were cut into 60-nm sections, stained with uranyl acetate/lead citrate, and examined in a JEOL 100C transmission electron microscope.
NC Versus HC Explants
The explants were cut from predetermined sites that corresponded to lesion-resistant areas (ventral descending thoracic aorta maximally remote from intercostal artery ostia and designated site 1) or lesion-prone areas (aortic arch designated site 2; Fig 1⇓). For each treatment group, 6 explants from each rabbit were cultured. In NC explants, CLM growth was not different between sites 1 and 2 (see “Results”). In the absence of serum or exogenous growth factors, time-dependent CLM growth of NC explants was compared with that of HC explants from site 1 or 2. In some experiments, site 1 HC explants were cocultured with site 2 HC explants to ascertain whether diffusible factors from site 2 explants might stimulate growth of site 1 explants. We use the term “lesion-free” to designate exclusively those explants obtained from lesion-resistant areas (site 1) of HC aortas.
Treatment With Native and Ox-LDL
Because oxidatively modified LDL exerts inhibitory effects on the growth of endothelial cells in culture,14 15 we incubated NC explants with native or oxidatively modified LDL preparations. LDL preparations were added to the medium on the first and third days of culture. Simultaneous experiments with NC explants cultured without or with graded concentrations (10, 50, or 100 μg/mL) of ox-LDL or native LDL (100 μg/mL) were conducted. To determine whether the LDL preparations exerted direct cytotoxic effects, cultured RAECs were incubated with 100 μg/mL native LDL or 100 μg/mL ox-LDL for 24 or 48 hours. DNA synthesis was evaluated by including in the medium 3 μCi/mL [3H]thymidine (Moravek Biomedicals) during the last 6 hours of a 24-hour incubation period.24 Cells were counted with a hemacytometer, and the percentage of dead cells was determined by trypan blue positivity.
Treatment With Growth Factors and Growth Factor–Neutralizing Antibodies
To characterize responsiveness to established angiogenesis-modulating factors, explants were treated every other day with 10 ng/mL TGF-β1, 10 ng/mL TNF-α, 10 ng/mL VEGF, or 5 to 20 ng/mL bFGF (all from R&D Systems). To determine whether endogenous bFGF was important for CLM growth, NC and lesion-containing HC explants were also treated every other day with 10 μg/mL bFGF-neutralizing antibody (goat IgG, R&D Systems). For comparison, some explants were treated with 10 μg/mL TGF-β1–neutralizing antibody (R&D Systems) or 10 μg/mL nonimmune goat IgG (Sigma Chemical Co). To evaluate the reversibility of ox-LDL–induced effects on DNA synthesis, cultured RAECs with or without exposure to ox-LDL were treated with 2 ng/mL bFGF.
Measurement of bFGF Concentrations
Minute bFGF concentrations can be accurately detected by sensitive ELISA in the culture medium during the early hours of in vitro angiogenesis.25 To determine whether the levels of endogenously released bFGF were related to CLM growth, bFGF concentrations in the culture medium were measured by sandwich ELISA with the use of the human FGF basic Quantikine kit (R&D Systems, catalog No. DFB00). bFGF standards (provided in the kit) and medium samples were incubated in wells of the microtiter plate coated with a monoclonal antibody specific for bFGF at room temperature for 2 hours. After the plates were washed, a conjugate containing a polyclonal antibody against bFGF was added and the complex was incubated for another 2 hours. After the substrate solution was added, the optical density of each well was assayed in a spectrophotometer with the wavelength set to 450 nm. bFGF concentrations of the medium samples were determined with the use of standard curves. bFGF concentrations in the plasma from NC or HC rabbits were also measured. Although the manufacturer indicated that the minimum detectable concentration was 1 pg/mL, values <5 pg/mL determined with the standard curves, as recommended, lacked precision and therefore have been reported as <5 pg/mL.
Differences between group means were compared using either paired t tests for single comparisons or modified t statistics (Bonferroni test) for multiple comparisons. An ANOVA followed by Scheffé’s test for significance was used to compare time-dependent CLM growth between groups. A value of P<.05 was considered significant. Values are reported as mean±SE.
CLM Emerging From Normal Aortic Endothelium
As shown for human coronary and rat aortic ring cultures,9 19 20 21 22 endothelial cells formed microtubes that sprouted from explant edges and underwent progressive elongation (Fig 2A⇓). The growing microtubes had a capillary-like ultrastructure, with single layers of endothelial cells surrounding the lumen (mean luminal diameter, 10±4 mm) containing no collagen (Fig 2B⇓). The cell lining of the lumen stained with antibody against vWF, an endothelial marker; this staining did not occur with omission of the primary antibody (Fig 2C⇓ and 2D⇓). There were no vWF-positive cells that were not associated with the microtubes. In contrast, isolated spindle-shape cells emerging from explant edges stained for muscle actin (HHF35 antibody) (data not shown). These cells probably represented smooth muscle cells or myofibroblasts and were not associated with the microtubes.21 Although explant-derived CLM lacked well developed basement membranes and pericytes, several features of CLM growth resemble angiogenesis in vivo.3 26 The angiogenesis-like features include sprouting, progressive elongation, and penetration of the CLM through the interstitium-like collagen substrate.
Microvascular structures originating in the planar explants were located within a thin layer of collage, which resulted in a predominantly two-dimensional growth distribution. Serial video microscopic images showed that after an initial lag phase of 2 days, during which time “primordial” endothelial buds developed at the edges of the explants, there was an elongation of microtubes that progressed nearly linearly for 3 to 5 days, after which growth halted. After 10 days, the microtubes began to degenerate (Fig 3⇓). In pilot studies with NC explants, the day-7 CLM growth index for lesion-resistant site 1 explants (2.45±0.28 μm per explant, n=24; 6×4) was not different from that for lesion-prone site 2 explants (2.37±0.31 μm per explant, n=24). Explants from site 1 were used to represent the NC explants, and the mean 7-day angiogenesis index was 2.40±0.21 μm per explant (n=90, 6×15). Expressed with respect to explant circumference, the index was 0.61±0.07 μm/mm.
Effect of Dietary Hypercholesterolemia on CLM Growth
Light microscopic examination of tissue surrounding explants revealed no intimal thickening and a complete absence of mononuclear leukocytes or lipid-laden macrophages at HC lesion-resistant site 1 (Fig 1⇑). Compared with NC explant controls, these explants invariably exhibited very poor growth, with a day-7 index of 0.31±0.12 μm per explant (0.07±0.03 μm/mm, P<.001), or 13±4% of controls. The index, expressed with respect to the circumference of individual explants, was (P<.001). The poorly developed CLM degenerated quickly after day 7 and disappeared almost completely after 14 days (Fig 3⇑). There were no foam cells surrounding lesion-free HC explants (Fig 4A⇓).
Tissue surrounding HC lesion-prone site 2 explants contained abundant subendothelial macrophage foam cells that stained positively with RAM11 antibody and oil red O. These cells invaded the matrix, and some of them were later observed in areas distant from the explant edges (Figs 1⇑ and 5⇓). Intraintimal neovascularization could not be demonstrated during examination for vWF-positive cells (data not shown). In contrast to lesion-free site 1 HC explants, lesion-containing site 2 explants showed no growth inhibition (day-7 index, 3.05±0.44 μm per explant, or 0.73±0.13 μm/mm, P>.1 compared with NC controls; Figs 3⇑ and 4B⇑). Another distinguishing feature was the invariable association of RAM11-positive macrophage foam cells that were apposed to the proximal and distal segments of the CLM (Figs 4B⇑ and 5⇓). Electron photomicrographs revealed intimate contact between the abluminal surface of the endothelial cells and the macrophages (Fig 6⇓). However, the presence of macrophages did not appear to be associated with changes in the ultrastructure of the CLM (Figs 2⇑ and 6⇓). In contrast to the CLM growth pattern in NC explants, the CLM continued to grow in these macrophage-rich explants until day 10 and showed only a slight decline after 14 days (Fig 3⇑).
To demonstrate that lesion-containing explants released proangiogenic factors that were missing from lesion-free explants, coculture experiments were performed. Lesion-free HC explants exposed to cocultured lesion-containing explants from the same aorta exhibited a threefold increase in CLM growth (1.20±0.26, n=24, P<.05), with growth most marked in the vicinity of the lesion explants.
Effects of Ox-LDL
Native LDL (100 μg LDL protein per milliliter) prepared without antioxidants (other than EDTA) exerted a nonsignificant effect on CLM growth in NC explants. In contrast, ox-LDL dose-dependently (10 to 100 μg/mL) inhibited CLM growth (Table 1⇓). In cultured RAECs, 100 μg/mL ox-LDL decreased thymidine incorporation into DNA at 24 hours without increasing the percentage of dead cells, whereas 100 μg/mL native LDL had no effect. The cell counts were not different from those for untreated controls (Table 2⇓).
Effect of Exogenous bFGF on CLM Growth and DNA Synthesis
In NC explants treated with 10 ng/mL TGF-β1 or 10 ng/mL TNF-α, CLM growth was reduced to <50%; CLM growth in explants treated with 10 ng/mL VEGF was not different from that of controls (Table 3⇓). In contrast, bFGF dose-dependently (5 to 20 ng/mL) increased CLM growth in NC explants (Table 3⇓ and Fig 7⇓). Moreover, microtube growth was sustained for 14 days, after which degradation became noticeable (Fig 7⇓). bFGF partially corrected growth suppression in lesion-free HC explants and NC explants incubated with ox-LDL (Table 4⇓).
In cultured RAECs not exposed to LDL preparations, bFGF increased DNA synthesis twofold and cell count by 20%. In cells treated with ox-LDL, bFGF corrected the replication deficit and produced increases in DNA synthesis that exceeded those observed in untreated cells (Table 2⇑).
Role of Endogenous bFGF on CLM Growth
Treatment with bFGF-neutralizing antibody at the beginning of the culture nearly abolished CLM growth in NC explants. In lesion-containing HC explants, bFGF-neutralizing antibody did not block the emigration of foam cells but did abolish CLM growth to the same extent as in the NC group (Figs 4C⇑ and 8⇓). When bFGF-neutralizing antibody was given on day 10 to HC explants with lesions, the well developed CLM underwent rapid fragmentation and degeneration within 48 hours. This change suggested that the growth factor contributed not only to the growth but also to the preservation of the CLM. In NC explants or HC explants containing foam cell lesions, treatment with 10 ng/mL TGF-β1–neutralizing antibody or 10 ng/mL nonimmune IgG did not affect CLM growth (Figs 4D⇑ and 8⇓).
Correlation Between Endogenously Released bFGF and CLM Growth
As shown in Table 5⇓, bFGF concentration (in picograms per milliliter) in the media of day-5 cultures of NC explants (11±2) was similar to that of HC explants (14±3) containing lesions. In contrast, bFGF concentration in the media of lesion-free HC cultures was significantly lower (<5) than that of the other groups. Treatment of NC explants or lesion-containing HC explants with bFGF-neutralizing antibody significantly reduced measurable bFGF. Treatment with TGF-β1–neutralizing antibody or nonimmune IgG had no effect. The plasma bFGF concentration of HC rabbits (6±1, n=6, P<.05) was lower than that of NC rabbits (10±2).
In this study, we demonstrated that rabbit aortic explants in organ culture gave rise to an organized, capillary-like growth of endothelial cells. In HC rabbits, lesion-free explants containing no immunostainable macrophages exhibited markedly suppressed CLM growth. When explants were occupied by lesion-forming inflammatory cells, however, endothelial microtubes were surrounded by emigrating macrophages, and organized endothelial growth was not suppressed or even enhanced. Ample growth in the NC group and in the lesion-containing HC group was sensitive to the inhibitory effects of a bFGF-neutralizing antibody. Exogenous bFGF, on the other hand, substantially stimulated growth in the lesion-free HC group. In combination, these results suggest that suppressed growth in the lesion-free HC group reflected a reduced availability or possibly production of bFGF and that the enhanced growth in the presence of emigrating macrophages was mediated by a bFGF-dependent mechanism.
Restriction of the arterial blood supply determines an increased flow through residual arterial pathways. Adult arteries chronically exposed to high flow are known to undergo adaptive growth and remodeling with modulation in bFGF expression.27 28 Because hypercholesterolemia may impair flow-dependent regulatory mechanisms,29 30 it is possible that atherosclerotic arteries become insensitive to the growth-stimulatory effects of flow loads. In this study, we demonstrated dyslipidemic effects on angiogenesis in the absence of flow alterations.
Although new capillaries usually arise from parent capillaries,31 32 endothelial cells derived from a large artery (eg, human coronary or rat aorta) are capable of forming CLM in a serum-free organ culture system.9 19 20 21 22 In our model, the matrix provided a three-dimensional, interstitium-like environment that promoted organized endothelial growth. CLM originating in the edges of the explant were located in the plane of the endothelial monolayer of the explant. The thinness of the collagen coating the wells and possibly the sequential layering of two collagen coats facilitated the predominantly planar growth of the microtubes. Serial observations showed that elongation of the CLM occurred slowly over a period of days. As in other in vitro angiogenesis models,31 32 microtubes growing from the explants lacked the organization of mature capillaries. However, the formation of sprouts and subsequent progressive elongation and branching are features in common with the early stages of capillary angiogenesis in vivo.3 4 26 The slow, progressive pattern in the absence of endothelial cells unassociated with microtube assembly differs from the rapid (within 1 to 2 hours) tube-forming rearrangements of endothelial cells plated on a Matrigel matrix. Microtube formation under the latter conditions occurs with little or no endothelial cell proliferation.33 In contrast, in rat aortic ring cultures, endothelial cells forming capillaries incorporate BrdU into their nuclei,20 a response suggesting the importance of cell replication. The formation of microtubes in our experiments indicates that macrovascular aortic endothelium can undergo phenotypic changes promoting microvessel-like cellular assembly.21 Elongation of microtubes has been suggested to reproduce in vitro the processes of angiogenesis in vivo, including proteolytic basement membrane breakdown, migration, proliferation, and tube formation with intraluminal collagenolysis.26 34 The major strength of the model is that the organ culture system is not dependent on exogenous mitogens (eg, serum, growth factors, or cytokines). Therefore, the model appears suitable for the detection of missing factors (or excess inhibitors) regulating cell growth.
For quantification of microtube growth, we used a method similar to that employed in studies of rat aortic ring explants.19 20 21 22 The spatial distribution of microtubes in a single plane facilitated their morphometric quantification. Our conclusions are not critically dependent on the precision of the quantitative morphometric method, because semiquantitative assessment would have sufficed to demonstrate the striking effect of hypercholesterolemia in the absence of inflammatory cells (<15% of control). Previous experiments in rabbits suggested that compensatory growth of large and small arteries induced by surgical restriction of the arterial blood supply is markedly suppressed.8 Results of the present study indicate that an atherogenic dyslipidemia influences intrinsic growth properties of artery cells that do not depend on flow-sensing mechanisms or systemic humoral effects. The observed growth impairment cannot therefore be ascribed to a defective shear stress–dependent gene activation mechanism.
Of interest is the observation that ox-LDL inhibited CLM growth in explants from NC rabbits. This finding suggests that the uptake or intracellular generation of ox-LDL by endothelial cells in vivo might play a role in impairing vascular growth.35 36 At 24 hours, ox-LDL reduced premitotic DNA synthesis in cultured RAECs in the absence of a change in cell number. These results suggest that the ox-LDL–induced inhibition of CLM growth did not reflect lethal cytotoxic effects. Reversal of inhibition in response to bFGF or explants in coculture supports the view that hypercholesterolemia and ox-LDL acted without causing irreversible cell injury. We have previously reported that LPC, a phospholipid metabolite accumulating in atherosclerotic arterial walls and ox-LDL, inhibits DNA synthesis of endothelial cells.9 Recently, Murugesan and Fox37 demonstrated that LPC exerts antimigratory effects on endothelial cells. Because migration of endothelial cells plays a pivotal role in angiogenesis, LPC is one factor that might contribute to inhibited angiogenesis mediated by ox-LDL.
Effects of TGF-β1 and TNF-α on endothelial cell growth are complex and appear to depend on growth factor concentration and substrate conditions. Although these agents may inhibit endothelial cell growth in two-dimensional cultures, low concentrations may actually stimulate tube formation in three-dimensional matrices.12 In our model, both growth factors in high concentrations exerted inhibitory effects on CLM growth. VEGF did not increase or decrease CLM growth in this model. In contrast, bFGF, an angiogenic factor both in vivo and in vitro, increased CLM growth in NC explants. Growth suppression in these explants by bFGF-neutralizing antibody indicated the importance of endogenous bFGF in the angiogenic response. The importance of endogenous bFGF in angiogenesis has also been demonstrated in cultures of aortic rings from NC rats.38 In preliminary studies, we confirmed that bFGF can induce expression of its own gene,39 40 and we found that ox-LDL decreased bFGF gene expression in vascular endothelial cells.41 Recent experiments in our laboratory indicate that steady-state bFGF mRNA levels can be dose-dependently reduced by ox-LDL in cultured endothelial cells.41 These findings suggest that decreased production of bFGF is responsible for suppressed angiogenesis in the presence of dyslipidemia.
When explants contained inflammatory cell–rich lesions (ie, fatty streaks), CLM growth was activated and the growth of CLM was invariably accompanied by macrophage foam cells in close apposition to the outer surface of the microtubes. This finding suggests that endothelial cells exposed to a dyslipidemic environment, although possessing an intrinsically suppressed capacity for replication and organized growth, can still respond to external or paracrine angiogenic stimuli. It also indicates that neovascularization of atheromatous lesions reflects predominantly a local angiogenesis stimulated by the inflammatory processes of lesion formation.10 11 Because atheromatous neovascularization is a local phenomenon, its presence does not imply that adaptive growth of mature vessels is preserved in the presence of hypercholesterolemia.
Numerous angiogenic factors have been demonstrated to be released from macrophages undergoing inflammatory activation.12 The close approximation between macrophages and microtube-forming endothelial cells is likely to facilitate paracrine interactions between these two cell types. Although contributions by other growth factors or cytokines such as VEGF,7 42 acidic FGF,11 43 or interleukin-844 expressed by macrophages cannot be ruled out, the degree of suppression of CLM growth by bFGF-neutralizing antibodies suggests a strong dependence on bFGF. The specificity of the bFGF contribution was confirmed by the absence of inhibitory effects of TGF-β1–neutralizing antibody and nonimmune IgG.
bFGF, a ubiquitous growth factor expressed by many cells including endothelial cells and macrophages,12 43 exerts its multiple effects by means of autocrine and paracrine pathways.39 45 Although it lacks a typical hydrophobic signal-peptide sequence, bFGF synthesized by endothelial cells has been demonstrated to be deposited in the subendothelial extracellular matrix.46 This indicates that bFGF can be secreted by a nonclassic pathway that is apparently independent of the endoplasmic reticulum–Golgi complex.47 With recent improvements in immunoassay techniques, elevated bFGF levels have been detected during the early hours of in vitro angiogenesis.25 In this study, bFGF was detected in the serum-free media of explant cultures. The concentrations of secreted bFGF in the media correlated well with the degrees of CLM growth. In the poorly growing lesion-free HC group, bFGF concentration was also extremely low. Increased bFGF release from lesion-containing HC explants promoted CLM growth. Because activated macrophages are known to secrete numerous growth factors, including bFGF,12 one plausible explanation is that lesion-derived macrophages were a source of bFGF that could correct the defective growth of endothelial cells exposed to hypercholesterolemia. The paracrine effect of bFGF may contribute to localized neovascularization of atheromas. Although bFGF may be secreted by inflammatory and activated endothelial cells in vivo, it is interesting to note that plasma bFGF concentrations were lower in HC than in NC rabbits. Whether these lower concentrations play a role in limiting compensatory macrovascular and microvascular growth after surgical restriction of the arterial blood supply in vivo is uncertain, however. Elucidation of underlying biochemical or genetic mechanisms accounting for the reduced availability of bFGF will require detailed investigation. In addition, it will be important to determine in monoculture systems the comparative effects of ox-LDL on bFGF expression by endothelial cells and macrophages. Maintained responsiveness of HC explants with suppressed growth to exogenous bFGF suggests that receptor and transduction functions for the factor are partly preserved, but effector pathways of bFGF in HC states will require further characterization.
In conclusion, endothelial cells of a large artery exhibit a bFGF-dependent capillary-like growth in a three-dimensional, interstitium-like matrix. Exposure of the cells to elevated cholesterol concentrations in vivo or to ox-LDL in vitro produces growth impairment that can be partly reversed by exogenous bFGF. Stimulation of the endogenous production of angiogenic mitogens or administration of orally active factor analogues is an intriguing option for the treatment of hyperlipidemic occlusive arterial disease.
Selected Abbreviations and Acronyms
|bFGF||=||basic fibroblast growth factor|
|ELISA||=||enzyme-linked immnosorbent assay|
|RAEC||=||rabbit aortic endothelial cell|
|TGF-β||=||transforming growth factor-β|
|TNF-α||=||tissue necrosis factor-α|
|VEGF||=||vascular endothelial growth factor|
|vWF||=||von Willebrand factor|
This study was supported in part by grants HL-40884 and HL-36894 from the National Institutes of Health, Bethesda, Md; grant and award 38520 from The Methodist Hospital Foundation, Houston, Tex; and Grant-in-Aid 95G-239 from the American Heart Association, Texas Affiliate, Austin, Tex. The authors are grateful to Suzanne Simpson for her excellent editorial assistance.
Sasayama S, Fujita M. Recent insights into coronary collateral circulation. Circulation. 1992;85:1197-1204.
Habib GB, Heibig J, Forman SA, Brown BG, Roberts R, Terrin ML, Bolli R, and the TIMI Investigators. Influence of coronary collateral vessels on myocardial infarct size in humans. Circulation. 1991;83:739-746.
Gregg DE. The natural history of coronary ‘collateral’ development. Circ Res. 1974;35:335-344.
Schaper W. Coronary collateral development: concepts and hypothesis. In: Schaper W, Schaper J, eds. Collateral Circulation. Norwell, Mass: Kluwer Academic Publishers; 1993:41-64.
Baffour R, Berman J, Garb JL, Rhee SW, Kaufman J, Friedman P. Enhanced angiogenesis and growth of collaterals by in vivo administration of recombinant basic fibroblast growth factor in a rabbit model of acute lower limb ischemia: dose-response effect of basic fibroblast growth factor. J Vasc Surg. 1992;16:181-191.
Lazarous DF, Scheinowitz M, Shou M, Hodge E, Rajanayagam S, Hunsberger S, Robison WG Jr, Stiber JA, Correa R, Epstein SE, Unger EF. Effects of chronic systemic administration of basic fibroblast growth factor on collateral development in the canine heart. Circulation. 1995;91:145-153.
Takeshita S, Kearney M, Loushin C, Brogi E, Zheng LP, Horowitz J, Ferrara N, Symes JF, Isner JM. In vivo evidence that vascular endothelial growth factor stimulates collateral formation by inducing arterial cell proliferation in a rabbit ischemic hindlimb. J Clin Invest. 1994;93:662-670.
Henry PD. Hypercholesterolemia and angiogenesis. Am J Cardiol. 1993;72(suppl):61C-64C.
Brogi E, Winkles JA, Underwood R, Clinton SK, Alberts GF, Libby P. Distinct patterns of expression of fibroblast growth factors and their receptors in human atheroma and nonatherosclerotic arteries. J Clin Invest. 1993;92:2408-2418.
Sunderkoetter C, Steinbrink K, Goebeler M, Bhardwaj R, Sorg C. Macrophages and angiogenesis. J Leukoc Biol. 1994;55:410-422.
Yla-Herttuala S, Palinski W, Butler SW, Picard S, Steinberg D, Witztum JL. Rabbit and human atherosclerotic lesions contain IgG that recognizes epitopes of oxidized LDL. Arterioscler Thromb. 1994;14:32-40.
Mangin E Jr, Kugiyama K, Nguy J, Kerns S, Henry PD. Effects of lysolipids and oxidatively modified low density lipoprotein on endothelium-dependent relaxation of rabbit aorta. Circ Res. 1993;72:161-166.
Brand K, Banka CL, Mackman N, Terkeltaub RA, Fan ST, Curtiss LK. Oxidized LDL enhances lipopolysaccharide-induced tissue factor expression in human adherent monocytes. Arterioscler Thromb. 1994:14:790-797.
Nicosia RF, Tuszynski GP. Matrix-bound thrombospondin promotes angiogenesis in vitro. Cell Biol. 1994;124:183-193.
Tsukada T, Rosenfeld M, Ross R, Gown A. Immunocytochemical analysis of cellular components in atherosclerotic lesions. Arteriosclerosis. 1986;6:601-613.
Zhan X, Hu X, Friesel R, Maciag T. Long term growth factor exposure and differential tyrosine phosphorylation are required for DNA synthesis in BALB/c 3T3 cells. J Biol Chem. 1993;268:9611-9620.
Langille BL. Remodeling of developing and mature arteries: endothelium, smooth muscle, and matrix. J Cardiovasc Pharmacol. 1993;21(suppl):S11-S17.
Malek AM, Gibbons GH, Dzau VJ, Izumo S. Fluid shear stress differentially modulates expression of genes encoding basic fibroblast growth factor and platelet-derived growth factor B chain in vascular endothelium. J Clin Invest. 1993;92:2013-2021.
McLenachan JM, Williams JK, Fish D, Ganz P, Selwyn AP. Loss of flow-mediated endothelium-dependent dilation occurs early in the development of atherosclerosis. Circulation. 1991;84:1273-1278.
Sorensen KE, Celermajer DS, Georgakopoulos D, Hatcher G, Betteridge DJ, Deanfield JE. Impairment of endothelium-dependent dilation is an early event in children with familial hypercholesterolemia and is related to the lipoprotein(a) level. J Clin Invest. 1994;93:50-55.
Folkman J, Shing Y. Angiogenesis. J Biol Chem. 1992;267:10931.
Kubota Y, Kleinman HK, Martin GR, Lawley TJ. Role of laminin and basement membrane in the morphological differentiation of human endothelial cells into capillary-like structures. J Cell Biol. 1988;107:1589-1598.
Leek RD, Harris AL, Lewis CE. Cytokine networks in solid human tumors: regulation of angiogenesis. J Leukoc Biol. 1994;56:423-435.
Maggi E, Chiesa R, Melissano G, Castellano R, Astore D, Grossi A, Finardi G, Bellomo G. LDL oxidation in patients with severe carotid atherosclerosis: a study of in vitro and in vivo oxidation markers. Arterioscler Thromb. 1994;14:1892-1899.
Chen CH, Nguyen HH, Henry PD, Gotto AM Jr. Inhibition of atherosclerotic human coronary smooth muscle cell proliferation by blocking endogenous basic fibroblast growth factor. Circulation. 1994;90(suppl I):I-511. Abstract.
Chen CH, Jiang W, Luo S, Via D, Henry PD, Gotto AM Jr. Oxidized LDL inhibits expression of basic fibroblast growth factor in vascular endothelial cells. J Invest Med. 1996;44:276A. Abstract.
Leung DW, Cachianes G, Kuang WJ, Goeddel DV, Ferrara N. Vascular endothelial growth factor is a secreted angiogenic mitogen. Science. 1989;246:1306-1309.
Folkman J, Klagsbrun M. Angiogenic factors. Science. 1987;235:442-447.
Koch AE, Polverini PJ, Kunkel SL, Harlow LA, DiPietro LA, Elner VM, Elner SG, Strieter RM. Interleukin-8 as a macrophage-derived mediator of angiogenesis. Science. 1992;258:1798-1801.
Sato Y, Rifkin DB. Autocrine activities of basic fibroblast growth factor: regulation of endothelial cell movement, plasminogen activator synthesis, and DNA synthesis. J Cell Biol. 1988;107:1199-1205.
Vlodavsky I, Folkman J, Sullivan R, Fridman R, Ishai-Michaeli R, Sasse J, Klagsbrun M. Endothelial cell-derived basic fibroblast growth factor: synthesis and deposition into subendothelial extracellular matrix. Proc Natl Acad Sci U S A. 1987;84:2292-2296.