Jump to
Abstract
Abstract Blood supply through collateral arteries is of critical importance in occlusive arterial diseases such as coronary atherosclerosis. Induction of angiogenic growth factor within either the narrowing arteries or jeopardized myocardium may promote angiogenesis in vivo, leading to salvage of ischemic myocardium. We constructed a replication-defective adenovirus (AdCAsFGF-2) coding for human basic fibroblast growth factor (FGF)-2 that is modified, so that its secretion will be facilitated, by tagging a signal sequence derived from FGF-4. A large quantity of FGF-2 was detected in both the cell lysate and culture medium of COS cells infected with AdCAsFGF-2, indicating that FGF-2 was secreted at least partly from the infected cells. The conditioned medium from the infected COS cells stimulated DNA synthesis in and induced cellular proliferation of arterial smooth muscle cells. These effects were eliminated by adenovirus-mediated overexpression of a dominant-negative truncated FGF-receptor type 1. Implantation of a gel of basement membrane proteins containing fibroblasts infected with AdCAsFGF-2 into the ventral subcutaneous space of mice induced extensive cellular proliferation and the formation of functional arterioles. Cells surrounding the vessels were positively immunostained with antibodies recognizing either smooth muscle–specific α-actin or factor VIII antigen as a marker for endothelium. These results suggest that AdCAsFGF-2 may be useful for delivering functional FGF-2 into tissues and may lead to therapeutic angiogenesis in vivo.
- Received May 5, 1996.
- Accepted January 2, 1997.
Myocardial infarction is one of the major causes of death in industrial countries.1 2 It may be possible to save the jeopardized myocardium by increasing the blood supply either by dilating the narrowed coronary arteries responsible for the ischemia with a balloon catheter (percutaneous transluminal coronary angioplasty) or by surgically bypassing the narrowed section using an autologous graft of either arterial or venous origin. However, both procedures are effective for relatively large vessels only, and their benefits may be limited by restenosis.2 3 4 5 Another possible method would be to develop a collateral blood supply to take over for the diseased vessels. It is well known that occlusion of coronary arteries leads to the opening or development of collaterals in the human heart.6 Although the mechanisms and the responsible molecules underlying this phenomenon are not fully understood,6 7 angiogenic peptide growth factors such as acidic and basic FGF (FGF-1 and FGF-2, respectively) have been identified in the adult heart8 9 10 11 and in cultured cardiomyocytes.12
FGFs are potent and multifunctional growth factors mainly targeted at cells of mesoneurodermal origin, including arterial SMCs, ECs, fibroblasts, certain epithelial cells, and neural cells.7 13 14 15 FGF is a physiological mesodermal inducer in the early stages of frog development16 and also a physiological dominant growth factor for cardiomyocytes within 1 week of chicken embryogenesis.17 FGF can stimulate the growth of arterial endothelium; increase production of proteases such as plasmin18 and antiproteolysis proteins including plasminogen activator inhibitor-1, partially through activation of latent TGF-β by plasmin;19 and induce tube formation of ECs.7 Consequently, FGF alone can induce angiogenesis both in vitro and in vivo, and it is thought to be one of the major factors regulating angiogenesis in vivo.7 13 14 Thus, it seems a reasonable assumption that FGF, if efficiently introduced into the ischemic myocardium, may facilitate the effective formation of collaterals and thus induce beneficial effects. In fact, FGF-2 injected into dog coronary arteries20 or given periadventitially in the pig21 reduced the size of an infarction caused by subsequent ligation of the coronary artery. If such a growth factor could be introduced into the diseased area in a site-specific manner as an expression unit of the appropriate cDNA instead of as the protein itself and expressed for a sufficiently long period of time, more efficient and site-specific effects, with fewer systemic side effects, might well be attained.
Recent studies have shown that a recombinant adenoviral vector can achieve remarkably efficient gene transfer both in vitro and in vivo in a variety of tissues and cell types, including arterial wall cells22 23 24 25 26 27 28 29 30 31 32 33 and myocardium.12 34 35 36 37 Using adenoviral vectors, a high level of gene expression can be expected for weeks rather than days. In the present study, we constructed a replication-defective adenovirus expressing human FGF-2 under a potent constitutive promoter (AdCAsFGF-2) and investigated its angiogenic effects both in vitro and in vivo. In contrast to other growth factors, FGF-2 (and FGF-1) does not have conventional signal peptides, and the mechanisms underlying secretion from FGF-2–producing cells are still not fully understood.7 For secretion, some specialized mechanisms might be required. If so, only minimum biological effects of FGF-2 after gene transfer could be seen, since most of the effects of FGF-2 are mediated by binding to and activating of specific receptors on the cell membrane.38 For this reason, in constructing an adenoviral vector (AdCAsFGF-2), we used a modified cDNA of human FGF-2 in which a signal sequence derived from FGF-4 was ligated to the 5′ end of the native FGF-2 cDNA. In this way, we hoped to ensure that FGF-2 would be secreted from the infected cells. We have confirmed that COS cells, which usually do not produce endogenous FGF-2, produced and secreted to the medium a biologically active FGF-2 after infection with AdCAsFGF-2. Furthermore, we observed that application of AdCAsFGF-2 induced a marked cellular proliferation and angiogenesis in vivo. This material may well be useful for the efficient delivery of FGF-2 to tissues and may facilitate therapeutic angiogenesis in vivo.
Methods
Preparation of Adenoviral Vector
A recombinant cDNA for the secreted form of human FGF-2 was constructed by the addition of the signal sequence of FGF-4 encoding 22 amino acids to the 5′ end of the cDNA of human full-length FGF-2 coding 155 amino acids39 (provided by T. Maciag). Replication-defective E1 and E3 adenoviral vectors were prepared as described previously.33 40 41 42 Briefly, secreted FGF-2 (sFGF-2) cDNA was placed into a cassette cosmid vector under a CA promoter comprising a cytomegalovirus enhancer and chicken β-actin promoter43 (pAdCAsFGF-2). A recombinant adenovirus was constructed by in vitro homologous recombination in 293 cells using pAdCAsFGF-2 and the adenovirus DNA-terminal protein complex.40 The desired recombinant adenovirus, designated AdCAsFGF-2, was purified by ultracentrifugation through a CsCl2 gradient followed by extensive dialysis. Contamination by the wild-type adenovirus was excluded by the polymerase chain reaction designed for E1 amplification.44 The titer of the virus stock was assessed by a plaque-formation assay using the 293 cells and expressed in plaque formation units. An adenovirus expressing a truncated form of chicken FGF receptor type 145 that holds the entire extracellular domain and the transmembrane portion but only 26 amino acids of the cytoplasmic domain was also constructed (AdCAFGF-TR). This truncated receptor should disable the wild-type FGF receptor functionally as a dominant-negative mutation.16 17 45 The truncated chicken FGF receptor 1 cDNA was tagged with an influenza virus HA epitope (Ser-Tyr-Pro-Tyr-Asp-Val-Pro-Asp-Tyr-Ala-Ser-Leu) at its C terminal without making a frame shift and then inserted into the cosmid vector. We also used three control adenoviruses: AdCALacZ, expressing bacterial β-galactosidase; Ad1w, which did not express any exogenous gene32 33 37 41 42 ; and AdCATGFβ-TR, expressing a truncated TGF-β type II receptor that functions as a dominant-negative TGF-β receptor.42
Cell Culture
Arterial SMCs and ECs were prepared from the thoracic aortas of beef cattle by an explant method.33 COS-1 cells, 293 cells, and BALB/3T3 fibroblasts (clone A31) were obtained from the Japanese Cancer Research Resources Bank. Cells were cultured in DMEM (GIBCO-BRL) with 10% FBS supplemented with 2 mmol/L l-glutamine, 100 U/mL penicillin G, and 50 μg/mL streptomycin. SMCs and ECs from passages 3 to 8 were used in this study. In vitro gene transfer into cells was carried out by incubation with the adenoviral vector in serum-free medium (DMEM containing 0.05% BSA, 1 μg/mL insulin, 5 μg/mL transferrin, and 25 mmol/L HEPES [pH 7.4]) for 2 hours at room temperature under gentle agitation. After washing twice with PBS, cells were incubated in either growth medium or serum-free medium until assayed.
Western Blot Analysis
COS cells in 90-mm dishes were infected with AdCAsFGF-2 at various multiplicity of infection (moi) or left uninfected. Two days after incubation in the growth medium, the medium was replaced with serum-free medium (7 mL/90-mm dish), and cells were incubated for a further 22 hours until lysed in RIPA buffer (50 mmol/L NaCl, 30 mmol/L sodium pyrophosphate, 50 mmol/L NaF, 5 mmol/L EDTA, 10 mmol/L Tris, pH 7.4, 1% Triton X-100, 1 mmol/L PMSF, 0.2 U/mL aprotinin, 10 mmol/L pepstatin A, and 25 mmol/L leupeptin). The culture medium was also collected. The cell lysates and culture media were treated with heparin-Sepharose beads (Pharmacia) 4 hours at 4°C and washed three times with 1.0 mol/L NaCl buffered with 100 mmol/L Tris-HCl (pH 7.4). The precipitates were subjected to SDS-PAGE (15%) and transferred onto polyvinylidene difluoride membranes (Millipore). A recombinant human FGF-2 protein (Intergen) was also used as a positive control and molecular marker (18 kD). The membrane was probed with a mouse monoclonal antibody against human FGF-2 (Ab 98, a gift from Takeda Research Laboratory, Oosaka) and then visualized with an alkaline phosphatase–conjugated anti-mouse IgG and chromogenic reagents (Promega).
COS cells were infected with AdCAFGF-TR at various moi or left uninfected, and lysed in RIPA buffer after incubation for 3 days. The lysates were analyzed by immunoblotting with either a monoclonal anti-HA epitope antibody (12CA5) or an extracellular FGF receptor antibody45 after partial purification with wheat germ agglutinin-Sepharose (Pharmacia) and separation by SDS-PAGE (8%). The reactive proteins were visualized as described above.
Measurement of DNA Synthesis
The serum-free conditioned medium (7 mL/90-mm dish) was prepared as described in the previous section from COS cells infected with either AdCAsFGF-2, AdCALacZ, or Ad1w at various moi. Confluent SMCs, ECs, or 3T3 fibroblasts in 24-well plates were incubated in serum-free medium for 50 hours and then challenged for 20 hours with one of the conditioned media (200 μL per well) plus fresh serum-free medium (200 μL per well). Cells were then pulsed for 4 hours with 1 μCi/mL [3H]thymidine (DuPont-NEN). The incorporation of [3H]thymidine into the trichloroacetic acid–insoluble material was measured using a scintillation counter. Some cells had been infected with AdCAFGF-TR at various moi 2 days before the conditioned media were applied.
Cell Proliferation Assay
SMCs were infected with either AdCAsFGF-2 (10 moi), Ad1w (10 moi), AdCAsFGF-2 (10 moi) plus AdCAFGF-TR (100 moi), AdCAsFGF-2 (10 moi) plus Ad1w (100 moi), or AdCAsFGF-2 (10 moi) plus AdCATGFβ-TR (100 moi). The next day, cells were harvested and plated sparsely in serum-free medium. The number of cells in fixed fields that were randomly selected (four fields per dish; two dishes for each group) was counted daily under a microscope.
In Vivo Angiogenesis
BALB/3T3 cells were infected with either AdCAsFGF-2 or AdCALacZ at moi 30 or left uninfected, and incubated in growth medium. Three days later, cells (5×106) were harvested using trypsin, washed twice with growth medium, resuspended in 100 μL of ice-cold DMEM containing 10% serum, and then mixed with 500 μL of an ice-cold gel of basement membrane proteins (growth factor–reduced Matrigel, Becton Dickinson) on ice. BALB/c male mice (5 weeks of age) were anesthetized by inhalation of diethyl ether, and the ventral skin was shaved. Infected cells in Matrigel (total volume, 0.6 mL) were injected into a subcutaneous space. In some experiments, adenoviral vectors (2×108 plaque formation units) alone were mixed in Matrigel and injected into mice. Seven days later, the mice were killed and the gel plugs fixed in formaldehyde. The fixed gel was embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin-eosin. Some sections were subjected to immunohistostaining with primary antibodies recognizing either SMC-specific α-actin (from Boehringer) or von Willibrand factor (factor VIII antigen, from DAKO), using a biotinylated rabbit anti-mouse IgG antibody (Nitirei) as a secondary antibody, peroxidase-labeled streptavidin, and diaminobenzidine. The cells were lightly counterstained with hematoxylin. Representative sections were photographed by a technician blinded to treatment regimen. All animals were treated under protocols approved by Kyushu University animal care committees.
Results
Expression and Secretion of FGF-2 From Cells Infected With AdCAsFGF-2
We constructed a replication-defective adenovirus (AdCAsFGF-2) expressing a modified form of human FGF-2 that should be secreted efficiently from infected cells. COS cells, which do not normally produce FGF-2 at a detectable level, were infected with AdCAsFGF-2 at various moi and incubated in growth medium for 48 hours and then in serum-free medium for a further 22 hours. The cell lysates and culture media were separately analyzed by immunoblotting with an antibody against human FGF-2 after a partial purification with heparin-Sepharose followed by SDS-PAGE. FGF-2 of 18 kD and of various sizes (18, 20, and 24 kD) was detected in both the lysates and the culture media (Fig 1⇓). The results showed that the infected COS cells produced FGF-2 and that at least some FGF-2 was indeed secreted from the cells into the medium. We hypothesize that different degrees of glycosylation may result in production of FGF-2 of various weights in the medium, although we do not know the exact mechanisms. The morphology of the AdCAsFGF-2–infected cells under light microscopy was no different from that of cells infected with AdCALacZ. Moreover, COS cells infected with AdCAsFGF-2 excluded trypan blue, indicating that the cells were alive.
Expression and secretion of FGF-2 protein from cells infected with AdCAsFGF-2. Confluent COS cells were infected with AdCAsFGF-2 at the indicated moi or left uninfected, and incubated for 48 hours. Then the medium was changed to serum-free DMEM and incubated for a further 22 hours. The cell lysates and supernatants were examined for the presence of FGF-2 by immunoblotting after partial purification with heparin-Sepharose (see “Methods”). The recombinant human FGF-2 protein was also analyzed. Molecular markers are in kilodaltons.
Cells Infected With AdCAFGF-TR Become Refractory to FGF-2
We also constructed AdCAFGF-TR expressing a truncated FGF receptor 1 that should function as a dominant-negative FGF receptor.16 17 45 BALB/3T3 cells were infected with AdCAFGF-TR at various moi, and cell lysates were analyzed by immunoblotting with either an antibody recognizing the HA epitope (Fig 2A⇓) or an antibody against the extracellular domain of FGF receptor45 (Fig 2B⇓). The truncated FGF receptor was expressed in a moi-dependent manner.
Expression of the truncated FGF receptor 1 in cells infected with AdCAFGF-TR. COS cells were infected with AdCAFGF-TR at indicated moi or left uninfected (control, 2 lanes). Three days later, cell lysates were analyzed by immunoblotting with either an anti-HA antibody (A) or an anti-FGF receptor antibody (B) after partial purification with wheat germ agglutinin-Sepharose followed by SDS-PAGE (8%). Molecular markers are in kilodaltons.
BALB/3T3 cells were infected with AdCAFGF-TR at various moi, AdCALacZ (30 moi), or left uninfected. Two days later, submaximal concentrations of either FGF-2 (10 ng/mL) or PDGF-BB (30 ng/mL) were applied and DNA synthesis was measured (Fig 3⇓). DNA synthesis in response to FGF-2 decreased in a moi-dependent manner in cells infected with AdCAFGF-TR. Cells infected with AdCAFGF-TR at 30 moi became completely unresponsive to FGF-2, whereas the infected cells remained fully responsive to PDGF-BB. These data indicate that AdCAFGF-TR directs cells to express a dominant-negative FGF receptor and that signaling by FGF-2 is specifically abolished in the infected cells, as previously reported.45
DNA synthesis stimulated by either FGF-2 or PDGF-BB in cells infected with AdCAFGF-TR. Confluent BALB/3T3 cells were infected with either AdCAFGF-TR (AdFTR) at indicated moi, AdCALacZ (AdLacZ) at moi 30, or left uninfected (−). Two days later, cells were stimulated with either FGF-2 (FGF, 10 ng/mL) or PDGF-BB (30 ng/mL), and [3H]thymidine incorporation was measured. Data are shown as mean±SD (n=4). SF denotes serum-free medium.
Secreted FGF-2 From Cells Infected With AdCAsFGF-2 Is Biologically Active
To confirm that the secreted FGF-2 was biologically active, the conditioned medium from COS cells infected with AdCAsFGF-2 was added to bovine SMCs, and DNA synthesis was measured. As shown in Fig 4⇓, [3H]thymidine incorporation was enhanced in the presence of the conditioned medium derived from COS cells infected with AdCAsFGF-2 (3 moi) but not with AdCALacZ. When cells had been infected with AdCAFGF-TR, the conditioned medium prepared from AdCAsFGF-2–infected COS cells failed to elicit DNA synthesis (Fig 4⇓). This finding is further support that FGF-2 in the supernatant indeed induced DNA synthesis in cells. Similar effects of the conditioned medium were also seen in both ECs and BALB/3T3 fibroblasts (data not shown).
Stimulation of DNA synthesis in SMCs by the conditioned medium. Arterial SMCs were challenged with the conditioned medium prepared from COS cells infected with either AdCAsFGF-2 (AdFGF) at 3 moi or AdCALacZ (AdLacZ) at 10 moi, or left uninfected (uninfect). [3H]Thymidine incorporation in SMCs was measured. Some cells had been infected with AdCAFGF-TR expressing a dominant-negative FGF receptor type 1 (indicated as AdFTR) at 30 moi, and 2 days later they were stimulated with the conditioned medium from COS cells infected with AdCAsFGF-2 (3 moi). Data are shown as mean±SD (n=5).
When SMCs infected with either AdCAsFGF-2 or Ad1w at 10 moi or left uninfected were plated sparsely in serum-free medium, only SMCs infected with AdCAsFGF-2 proliferated thereafter (Fig 5⇓). Both uninfected cells and cells infected with Ad1w exhibited a gradual decrease in cell number. The proliferative response observed in cells infected with AdCAsFGF-2 (10 moi) was completely eliminated when cells were coinfected with AdCAFGF-TR (100 moi) (Fig 5⇓). With AdCAsFGF-2 (10 moi) plus Ad1w (100 moi) or with AdCAsFGF (10 moi) plus AdCATGFβ-TR (100 moi), cellular growth was similar to that seen with AdCAsFGF-2 (10 moi) alone (Fig 5⇓).
Effect of AdCAsFGF-2 infection on cellular proliferation of SMCs. SMCs were infected with either AdCAsFGF-2 (AdFGF, ▪) or Ad1w (○) at 10 moi, or left uninfected (control, □). The next day, cells were plated sparsely in serum-free medium (day 0). Cell numbers in the fixed fields (four fields per dish, two dishes for each group) were counted daily under microscopy. Some cells were coinfected with AdCAsFGF-2 (10 moi) and either AdCAFGF-TR (AdFTR, 100 moi) expressing a dominant-negative FGF receptor 1 (▴), Ad1w (100 moi ▵), or AdCATGFβ-TR (AdTβ-TR, 100 moi) expressing a truncated TGF-β type II receptor (•). Ratio of cell number to cell number 24 hours after plating (day 1) is shown as mean±SD (n=8). Another set of experiment showed a similar result.
These results (Figs 4⇑ and 5⇑) together indicate that biologically active FGF-2 was indeed secreted from the cells infected with AdCAsFGF-2 and that AdCAsFGF-2 can stimulate cellular proliferation through autocrine and/or paracrine loops.
In Vivo Angiogenesis Induced by AdCAsFGF-2
We investigated whether FGF-2 produced by AdCAsFGF-2 could induce in vivo angiogenesis. BALB/3T3 fibroblasts infected with either AdCAsFGF-2 or AdCALacZ were mixed with growth factor–reduced Matrigel (a gel of basement membrane proteins) and injected into the ventral subcutaneous space of BALB/c mice. Uninfected fibroblasts were used as a control. Seven days later, the gel plugs were histologically examined. An enormous cell proliferation and multiple vessel formation, most of which contained red blood cells inside the lumen, were observed in the gel plug containing fibroblasts infected with AdCAsFGF-2 (Fig 6E⇓ through 6G). Substantially fewer cells and virtually no vessel formation were found in the gels containing cells either infected with AdCALacZ or left uninfected (Fig 6A⇓ through 6D). Immunohistostaining with antibodies recognizing either SMC-specific α-actin or factor VIII antigen as a marker for ECs demonstrated that SMCs and ECs were indeed present around the vessels (Fig 6F⇓ and 6G⇓). These results demonstrate that FGF-2 secreted from the AdCAsFGF-2–infected cells can induce cellular proliferation and angiogenesis in vivo.
Cellular proliferation and angiogenesis in vivo induced by AdCAsFGF-2–infected cells in Matrigel. BALB/3T3 fibroblasts infected with either AdCAsFGF-2 (E, F, and G) or AdCALacZ (C and D), or left uninfected (A and B) were mixed with Matrigel and injected subcutaneously into BALB/c mice. Seven days later, the gel plugs were histologically evaluated after hematoxylin-eosin staining (A, C, and E). The gel containing cells infected with AdCAsFGF-2 showed a marked cellular proliferation and numerous vascular formations with red blood cells inside (indicated by arrows, C). Immunohistostaining was performed using antibodies against SMC-specific α-actin (B, D and F) or von Willibrand factor (factor VIII antigen) as a marker for ECs (G). These immunohistostained samples were counterstained with hematoxylin. Original magnification ×200 (A through F) or ×400 (G).
When Matrigel containing AdCAsFGF-2 alone was injected, vessel formation was also observed, but only at the marginal area of the gel adjacent to the surrounding tissues. Neovascularization was not observed in the center of the gel plug (Fig 7⇓).
Angiogenesis in vivo induced by AdCAsFGF-2 in Matrigel. Matrigel containing AdCAsFGF-2 was injected subcutaneously into BALB/c mice. Seven days later, the gel plugs were histologically evaluated after hematoxylin-eosin staining. Cellular proliferation and arterial formation (indicated by arrows in B) were observed in the marginal area adjacent to the gel. Original magnification ×40 (A) or ×200 (B)
Discussion
In this study, we constructed an adenoviral vector expressing a secreted form of human FGF-2 with the aim of inducing an efficient therapeutic angiogenesis to a specific site as an alternative to the systemic administration of a large quantity of angiogenic growth factor protein. We confirmed that FGF-2 was indeed secreted to the medium from the infected cells (Fig 1⇑). The secreted FGF-2 induced DNA synthesis in SMCs (Fig 4⇑). Infected SMCs proliferated in serum-free medium by autocrine and/or paracrine loops (Fig 5⇑). These effects were abolished (Figs 4⇑ and 5⇑) in cells that had been infected with an adenovirus expressing a dominant-negative truncated FGF receptor 1 (Figs 2⇑ and 3⇑), further supporting the possibility that the increased DNA synthesis and cellular proliferation were due to FGF-2 secreted from the AdCAsFGF-2–infected cells. Subcutaneous injection into mice with a gel containing fibroblasts infected with AdCAsFGF-2 resulted in extensive cellular proliferation and neovascularization (Fig 6⇑). These vessels contained red blood cells and were surrounded with SMCs and ECs (Fig 6⇑), indicating that mature arterioles had been formed in vivo. These SMCs and ECs probably migrated from neighboring tissues attracted to FGF-2 produced by fibroblasts in the gel plug. However, the possibility may remain that some of the fibroblasts become differentiated to ECs and/or SMCs, since fibroblasts have been documented to be progenitors of ECs in vivo.46
We confirmed that biologically active FGF-2 was indeed secreted to the medium from AdCAsFGF-2–infected COS cells, which normally do not produce endogenous FGF-2 (Figs 1⇑ and 4⇑). The amount of FGF-2 in the medium may be somewhat underestimated, because some fraction of FGF-2 may be tightly bound with extracellular matrices rather than floating in the medium. In our study, we could not evaluate precisely how effective the modification of tagging a signal sequence on a native FGF-2 would be. There has been a report comparing the biological properties of adenoviral vectors that express either a secreted form of FGF-1 (acidic FGF) or a native FGF-1,47 which is the closest member to FGF-2 among the FGF family.13 14 In that study,47 the FGF-1 was modified in the same way as FGF-2 was in the present study. The modified FGF-1 was secreted to the medium to a much greater extent than the native FGF-1 and was more effective than native FGF-1 in the induction of both EC proliferation and the formation of a capillary-like network in vitro, although not much difference was observed in in vivo angiogenesis.47 Considering that the two forms of FGF were modified in the same way in that47 and the present study, the modification of the signal sequence on FGF-2 would be expected to induce similar effects.
The effectiveness of angiogenic growth factors such as FGFs and another potent angiogenic cytokine, VEGF, has been reported in ischemic limb48 49 50 51 and ischemic myocardium20 21 52 53 54 models. For example, Lazarous et al54 investigated the effects of a long-term systemic administration of a large quantity of FGF-2 protein (1.74 mg/d for 4 weeks) on collateral development in a relatively large mammal, the dog, and found that the FGF-2 treatment induced a marked improvement in collateral flow. Furthermore, it has been reported that FGF-155 and VEGF56 stimulated endothelial proliferation, leading to a relining of injured arteries with regenerated endothelium, and also reduced intimal thickening in rat carotid arteries.56 Periadventitial administration of FGF-2 in a chronic ischemic region of the porcine coronary artery restored the ability of the endothelium to induce relaxation in response to various vasoactive substances.57 However, in most studies, large quantities of growth factor proteins were administered daily: This protocol could impose both financial and practical burdens if such a method were to be used therapeutically. In contrast, adenovirus-mediated gene transfer performed only once may be able to produce a sufficient quantity of protein for an extended period of time in a relatively site-specific manner. Gene expression by an adenoviral vector is limited to 2 to 3 weeks in the myocardium34 35 36 37 and 4 to 6 weeks in arterial walls;33 however, this may not be a serious problem, since FGF-2 treatment for 1 week during the period of maximal ischemia was sufficient to induce a major effect on collateral flow.54 This limitation in gene expression may even be advantageous in avoiding unnecessary angiogenesis and minimizing systemic side effects.
Before the clinical value of an adenoviral vector can be fully assessed, a direct cytopathic effect and inflammatory response potentially associated with the current E1 and E3 adenoviral vector should be tested in vivo using large mammals. The potential systemic side effects observed with a large quantity of FGF protein should also be tested in the cases of adenoviral vectors in large mammals. These undesirable side effects include (1) hypotension due to vasodilation,54 57 58 as well as anemia and a decrease in circulating platelets54 ; (2) a potential shift of phenotype of the myocardium from an adult differentiated type to a fetal dedifferentiated type, as has been shown in cultured neonatal rat cardiomyocytes59 ; and (3) fibroproliferative changes in tissues and/or intimal hyperplasia in injured arteries.60 61 62 63 These side effects are also implied by the enormous cellular proliferation seen in the present study (Fig 6⇑). Finally, it will be important to exclude the possibility of tumorigenesis, since NIH-3T3 fibroblasts transfected to express constitutively a secreted form of FGF-1 became tumorigenic.39 In terms of systemic side effects, it would be interesting to examine whether use of a native FGF-2 might reduce the risk.
A potential use has been reported for adenoviral vectors that express either FGF-1 (both a native form and a secreted form)47 or VEGF165.64 In terms of their properties, FGF-1 and FGF-2 are mitogenic for not only ECs but also for SMCs and fibroblasts, whereas VEGF acts more like an endothelium-specific growth factor. VEGF has signal peptides for secretion. At present, it is not known which angiogenic growth factor might be most suited for therapeutic angiogenesis. Moreover, it has not been established which, if any, combination might achieve more beneficial effects, although synergistic effect between FGF-2 and VEGF on endothelial proliferation has been reported both in vitro65 and in vivo.66 Furthermore, a combination of proliferative growth factors, such as FGF and VEGF, and a growth-inhibitory factor, TGF-β, may be interesting, since TGF-β potentiated the effect of FGF-2–and VEGF-induced angiogenesis in vitro, depending on its concentration.67 68 On the other hand, the potential fibrosis in response to FGF60 61 62 63 might be attenuated by a coadministration with an adenoviral vector expressing a dominant-negative TGF-β receptor that can abolish the diverse signaling by TGF-β, including the transcriptional activation of extracellular matrix proteins.42 Some of these interesting issues are under investigation in our laboratory.
Selected Abbreviations and Acronyms
| DMEM | = | Dulbecco’s modified Eagle’s medium |
| EC | = | endothelial cell |
| FGF | = | fibroblast growth factor |
| HA | = | hemagglutinin |
| moi | = | multiplicity of infection |
| PDGF | = | platelet-derived growth factor |
| SDS-PAGE | = | SDS–polyacrylamide gel electrophoresis |
| SMC | = | smooth muscle cell |
| TGF | = | transforming growth factor |
| VEGF | = | vascular endothelial growth factor |
Acknowledgments
This study was supported in part by a grant-in-aid for scientific research from the Ministry of Education, Science, and Culture of Japan and by grants from Sandoz Foundation for Gerontological Research (Switzerland), Mochida Memorial Foundation for Medical Science (Tokyo, Japan), and Takeda Research Foundation for Metabolic Diseases (Oosaka, Japan) to H. Ueno. We thank Dr T. Maciag (American Red Cross Hospital), Dr I. Saito (University of Tokyo), and Takeda Research Laboratory for cDNA, a cosmid vector, and an antibody against FGF-2, respectively. We acknowledge the excellent technical support of S. Nishio.
References
This Issue
Jump to
Article Tools
Share this Article
- Adenovirus-Mediated Expression of the Secreted Form of Basic Fibroblast Growth Factor (FGF-2) Induces Cellular Proliferation and Angiogenesis In Vivo
