Author + information
- Received June 28, 2003
- Revision received December 10, 2003
- Accepted January 5, 2004
- Published online May 19, 2004.
- Yasuyo Hisaka, MS*,
- Masaki Ieda, MD†,
- Toshikazu Nakamura, PhD‡,
- Ken-ichiro Kosai, MD, PhD§,
- Satoshi Ogawa, MD, PhD† and
- Keiichi Fukuda, MD, PhD*,†,* ()
- ↵*Reprint requests and correspondence:
Dr. Keiichi Fukuda, Institute for Advanced Cardiac Therapeutics, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan.
Objectives This study investigated the possibility of achieving angiogenesis by using gene-modified cells as a vector.
Background Although gene therapy for peripheral circulation disorders has been studied intensively, the plasmid or viral vectors have been associated with several disadvantages, including unreliable transfection and uncontrollable gene expression.
Methods Human hepatocyte growth factor (hHGF) and thymidine kinase (TK) expression plasmids were serially transfected into NIH3T3 cells, and permanent transfectants were selected (NIH3T3 + hHGF + TK). Unilateral hindlimb ischemia was surgically induced in BALB/c nude mice, and cells were transplanted into the thigh muscles. All effects were assessed at four weeks.
Results The messenger ribonucleic acid expression and protein production of hHGF were confirmed. Assay of growth inhibition by ganciclovir revealed that the 50% (median) inhibitory concentration of50 at first mention (50% “I?” concentration)> NIH3T3 + hHGF + TK was 1,000 times lower than that of NIH3T3 + hHGF. The NIH3T3 + hHGF + TK group had a higher laser Doppler blood perfusion index, higher microvessel density, wider microvessel diameter, and lower rate of hindlimb necrosis, as compared with the plasmid- and adenovirus-mediated hHGF transfection groups or the NIH3T3 group. The newly developed microvessels were accompanied by smooth muscle cells, as well as endothelial cells, indicating that they were on the arteriolar or venular level. Laser Doppler monitoring showed that the rate of blood perfusion could be controlled by oral administration of ganciclovir. The transplanted cells completely disappeared in response to ganciclovir administration for four weeks.
Conclusions Gene-modified cell transplantation therapy induced strong angiogenesis and collateral vessel formation that could be controlled externally with ganciclovir.
Growth factors isolated recently, including vascular endothelial cell growth factor, fibroblast growth factor, angiopoietin, and hepatocyte growth factor (HGF), have been found to induce strong angiogenesis (1–5). A number of studies have reported induction of angiogenesis and collateral vessel formation by gene therapy with these factors in both animal experiments and clinical trials. Plasmid or viral vectors have been used in these therapies (2,6,7), but the adenovirus vector entails some serious problems, such as allergic reactions or difficulty with repeated treatment, despite sufficiently high transfection efficiency. Moreover, although plasmid vectors have recently been used in clinical settings, have not been associated with allergic reactions, and could be used repeatedly, their transfection efficiency has been low and has varied with the tissues injected or the patient. These gene delivery methods have the common drawbacks of not being able to choose the target cells and to selectively eliminate the transfected cells once they acquire the character of abnormal growth. Thus, new methods that would provide ideal gene delivery systems have long been awaited.
Regeneration therapy has recently been performed in many tissues and organs. Various types of cells regenerate from embryonic or adult stem cells, and these cells would be transplanted into patients. Rapid and sufficient establishment of angiogenesis and collateral vessel formation to promote the survival and function of the transplanted cells are especially important in terms of blood supply. We investigated regeneration of cardiomyocytes from adult stem cells and concluded that blood vessel formation into transplanted cells is crucial to their survival (8). Because angiogenic gene therapy with plasmid vectors has been insufficient to induce the rapid and powerful angiogenesis required for transplantation of the regenerated cells, a new method has been needed to address this problem.
In the present study, NIH3T3 cells were permanently transfected with a novel angiogenic human HGF (hHGF) and thymidine kinase (TK) of herpes simplex gene and then used as a gene therapy vector. Their effect on blood flow, angiogenesis, and collateral formation was investigated in a murine ischemic hindlimb model (9–11). In this paper, we report that gene-modified cells expressing hHGF and TK induced strong angiogenesis and collateral vessel formation, and that they were easily controlled externally with ganciclovir.
The NIH3T3 cells were cultured in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% fetal bovine serum and penicillin (100 μg/ml), streptomycin (250 ng/ml), and amphotericin B (85 μg/ml).
Stable transfection of hHGF and TK genes
The complementary deoxyribonucleic acid (cDNA) of the hHGF and TK genes was inserted into the pUC-SRα and pGK expression vector plasmids, respectively (10–13). pPUR and pcDNA3.1/Hygro(+) are selection plasmids that confer puromycin resistance and hygromycin resistance, respectively. After co-transfection of pUC-SRα/hHGF and pPUR into the NIH3T3 cells, using the Effectene Reagent (QIAGEN GmbH, Hilden, Germany), the puromycin-nonresistant cells were removed with puromycin (3 μg/ml), and the hHGF-producing cells were clonally selected (NIH3T3 + hHGF). pGK/TK and pcDNA3.1/Hygro(+) plasmids were then similarly co-transfected into the NIH3T3 + hHGF cells; the hygromycin-nonresistant cells were removed with hygromycin (200 μg/ml); and both hHGF- and TK-producing cells were clonally selected (NIH3T3 + hHGF + TK).
Reverse transcription-polymerase chain reaction (RT-PCR)
Expression of hHGF messenger ribonucleic acid was analyzed by RT-PCR using the primers that specifically detect human but not mouse HGF, as previously described (14).
ENzyme-linked immunosorbent assay (ELISA) for hHGF
Production of hHGF was determined by ELISA with anti-human–specific HGF monoclonal antibodies (Institute of Immunology, Tokyo, Japan) (6,15,16).
The adenoviral vector plasmid pAd.CA-hHGF, which is composed of a cytomegalovirus immediate early enhancer, a modified chicken beta-actin promoter, and hHGF cDNA, was constructed by the in vitro ligation method (17). The pAd.CA-hHGF plasmid was partially cut with PacI and then transfected into 293 cells, followed by culture with 0.5% overlaid agarose-α-minimal essential medium (MEM) containing 5% horse serum for 10 to 15 days. Viral plaques, which had been confirmed by restriction enzyme analysis and ELISA for hHGF, were propagated in 293 cells, purified by CsCl2gradient ultracentrifugation twice, and desalted with a desalting column (18). Viral particles were calculated by means260.> of optical density at 260 nm.
Murine model of hindlimb ischemia
All animal experiments were approved by the Animal Care and Use Committee of Keio University. After anesthetizing male BALB/c nude mice (eight weeks) with diethyl ether, the femoral artery was gently isolated, and the proximal portion was ligated with 7-0 silk ligatures (19,20).
Transplantation of continuously hHGF-producing NIH3T3 cells
The hindlimb ischemic mice (n = 192) were randomly classified into five groups. The control groups received 0.2 ml saline only (n = 14), 500 μg pUC-SRα/hHGF plasmids in 0.2 ml saline (n = 10), 109particles Ad.CA-hHGF in 0.2 ml phosphate-buffered saline (n = 10), or NIH3T3 in 0.2 ml DMEM (n = 14). The experimental group received NIH3T3 + hHGF + TK in 0.2 ml DMEM (n = 144). All injections were given via a 27-gauge needle (21). The numbers of cells transplanted ranged from 104to 107. They were injected into two different sites in the ischemic thigh (adductor) skeletal muscles on postoperative day 1. The direction of injection was parallel to the muscle fibers. Angiogenesis and collateral vessel formation were assessed at four weeks.
Laser Doppler blood perfusion analysis
The blood perfusion rate in the ischemic (left leg) and normal (right leg) hindlimb was measured with a laser Doppler perfusion image (LDPI) system (Moor Instruments), as described previously (20,22).
Frozen sections (4 μm) were cut from tissue specimens (23). Immunohistochemical staining for hHGF, endothelial cells, and alpha-smooth muscle actin (SMA) was carried out with anti-human HGF (R&D Systems Inc., Minneapolis, Minnesota), anti-human von Willebrand factor (vWF)/horseradish peroxidase (HRP), and anti-human SMA/HRP (Dakocytomation, Kyoto, Japan), respectively. Sections for staining and counterstaining were incubated with 3,3′-diaminobenzidine tetrahydrochloride and Mayer's hematoxylin solution, respectively. Elastica van Gieson staining was carried out by the standard method. Paraffin sections (3 μm) were cut from tissue specimens, and hematoxylin-eosin staining was carried out by the standard method.
Assay of growth inhibition by ganciclovir in vitro
After seeding cells on six-well plates (105cells/well) and culturing for 24 h, they were exposed to ganciclovir in concentrations ranging from 0 to 10−3g/ml for 72 h (24,25).
Detection of ganciclovir-induced apoptosis with annexin V
Annexin V is an early apoptotic marker. The NIH3T3 + hHGF + TK group was exposed to 10−7g/ml ganciclovir for 48 h, and the apoptotic cells were detected with an annexin V-enhanced green fluorescent protein (EGFP) apoptosis detection kit (Medical & Biological Labs Co. Ltd., Nagaya, Japan) (26).
Regulation of transplanted cell growth with ganciclovir in vivo
We investigated the dose-response relationship of growth inhibition by ganciclovir by transplanting NIH3T3 + hHGF + TK (107cells) and administering ganciclovir two weeks later. The transplanted mice received different doses (0, 1, 10, 50, or 80 mg/kg per day) of ganciclovir orally once a day for four weeks.
The data were processed using StatView J-4.5 software. Results are reported as the mean value ± SE. Comparisons of values among all groups were performed by one-way analysis of variance. The Scheffe's Ftest was used to determine the level of significance. The probability level accepted for significance was p < 0.05.
Permanently hHGF-transfected NIH3T3 cells produced hHGF protein
The NIH3T3 + hHGF cells were obtained after two weeks of exposure to puromycin, and NIH3T3 + hHGF + TK cells were obtained after two more weeks of exposure to hygromycin. We confirmed that both the NIH3T3 + hHGF and NIH3T3 + hHGF + TK groups expressed hHGF mRNA and then produced hHGF protein at a rate of 17.3 ± 1.4 and 19.1 ± 2.0 pg/106cells per 24 h, respectively (Fig. 1).
Ganciclovir-inhibited cell growth and induced apoptotic cell death
It is well known that HGF regulates cell growth. To determine whether transfection of hHGF affects the growth of NIH3T3 cells, we counted the numbers of cells in vitro (Fig. 2A). The growth rate of the hHGF-transfected NIH3T3 cells seemed to increase slightly, but the increase was not significant on day 3. Transfection of the TK gene had no effect on their growth rate.
Next, we investigated the growth-inhibitory effect of ganciclovir on these cells (Fig. 2B). The IC50of ganciclovir for the NIH3T3 + hHGF + TK group was ∼1,000 times lower than that for the NIH3T3 + hHGF group. These findings confirmed that the TK plasmid genes had been effectively transfected, and that hardly any of the cells that expressed the TK gene survived exposure to ganciclovir at a concentration of 10−6g/ml, which did not affect the control cells.
Enhanced green fluorescent protein fluorescence was detected at the membranes of NIH3T3 + hHGF + TK cells after ganciclovir exposure (Fig. 2C), indicating that cell death was attributable to apoptosis.
Human HGF-producing cell therapy augmented angiogenesis and collateral vessel formation
To evaluate whether transplantation of hHGF-producing cells improves the perfusion of ischemic hindlimbs, we first determined the rate of necrosis of the ischemic hindlimb. Necrosis was rated on a three-grade scale. The rate of necrosis of the foot and toes in the saline group was 35.7% and 42.9%, respectively. The rates in the pUC-SRα/hHGF group were 20% and 40%, respectively, and in the Ad.CA-hHGF group 10% and 40%, respectively. These therapeutic approaches were effective in comparison with the saline group, but they were not sufficient to fully prevent the necrosis. To further ameliorate limb necrosis, we examined angiogenic gene-modified cell transplantation therapy. The NIH3T3 (107cells) group had rates of 14.3% and 35.7%, respectively, suggesting that the vector cell transplantation itself might improve perfusion of the ischemic limb to some extent. In contrast, the rates in the NIH3T3 + hHGF + TK (107cells) group were 5.8% and 14.5%, respectively (Fig. 3). The rate of necrosis was surprisingly reduced in the NIH3T3 + hHGF + TK group, indicating that transplantation of hHGF-producing cells might be one of the most effective methods of improving limb ischemia.
Vessel density and size
Immunostaining clearly revealed the presence of numerous vessels in the NIH3T3 + hHGF + TK group (Fig. 4A, panel c) and a lower number of vessels in the saline (Fig. 4A, panel a) and NIH3T3 (Fig. 4A, panel b) groups. Quantitative analysis revealed that the vessel density in the ischemic region was significantly higher (Fig. 4B), and the minimum diameter of the vWF-positive vessels was significantly greater (Figs. 4A, panel d, and 4C) in the NIH3T3 + hHGF + TK group.
Maturation of the vessels was investigated by staining three consecutive frozen sections of ischemic skeletal muscle. Amazingly, most of the vessels in the NIH3T3 + hHGF + TK group were vWF/α-SMA–double positive (Figs. 5A, panels a and b, and 5B). However, there was no increase in elastic fiber-positive cells, as compared with the saline and NIH3T3 groups (Figs. 5A, panel c, and 5B). These findings showed that NIH3T3 + hHGF + TK cell transplantation strongly induced angiogenesis not only at the capillary level but also at the microvessel (arteriole) level, and it caused angiogenesis at the large blood vessel level.
Laser Doppler blood perfusion
The LDPI analysis was performed to study subcutaneous blood perfusion. Representative images are shown in Figure 6A, and quantitative analysis of blood perfusion is shown in Figure 6B. No blood perfusion was observed in the hindlimb immediately after femoral artery ligation (Fig. 6A, panel a). Perfusion of the proximal part of the thigh had recovered at four weeks in the saline and NIH3T3 groups, but perfusion distal to the heel joint had markedly decreased (Fig. 6A, panels b and c). In the NIH3T3 + hHGF + TK (104cells) group, perfusion of the ischemic limb almost recovered to the control (nonischemic) level, but perfusion distal to the heel was slightly decreased compared with the control level (Fig. 6A, panel d). In the NIH3T3 + hHGF + TK (107cells) group, perfusion of the ischemic limb was 118.1% (i.e., much greater than that in the control hindlimb) (Figs. 6A and 6B, panel e). To adjust the recovery of blood perfusion in the ischemic limb to the appropriate level, we transplanted NIH3T3 + hHGF + TK (107cells), monitored the LDPI level, and began giving ganciclovir when blood perfusion reached the control level (two weeks). This method enabled us to adjust the blood perfusion rate in the ischemic limb to the same level as in the control limb (Figs. 6A and 6B, panel f).
When the NIH3T3 + hHGF + TK cells were transplanted into the normal nonischemic limb, the blood perfusion increased more than that in the control limb. Up to six weeks after transplantation, no evidence of angiosarcoma or hypervascular tumor was observed in the transplanted limb or other parts of the body (data not shown).
In vivo production of HGF protein
Immunohistochemical staining demonstrated the production of hHGF protein in transplanted NIH3T3 + hHGF + TK cells, but not in transplanted NIH3T3 cells (Fig. 7A).
Cell regulation with ganciclovir and TK
Figure 7Bshows a quantitative analysis of the inhibitory effect of ganciclovir on blood perfusion. At a concentration of 50 mg/kg/day of ganciclovir, the blood perfusion was adjusted in the ischemic limb to the same level as in the control limb, and no significant side effects were produced. Histologic examination revealed the natural history of the transplanted cells (Fig. 7C, panels a to c). The transplanted cells formed a mass between the skeletal muscles, which gradually increased in size but did not infiltrate into the skeletal muscle. Two weeks after transplantation of the NIH3T3 + hHGF + TK cells, we began giving ganciclovir orally every day for two to four weeks and then examined tissue samples (Fig. 7C, panels d to f). The NIH3T3 + hHGF + TK cells gradually underwent apoptosis, and by four weeks, no transplanted cells could be detected. The surrounding muscle cells and the generated vessels were unaffected by ganciclovir.
In this study, we assessed angiogenic gene-modified cell transplantation therapy with fibroblasts permanently transfected with hHGF and TK genes in a murine hindlimb ischemia model. This therapy had the following merits: 1) it induced angiogenesis and collateral vessel formation more effectively than with plasmid and viral vectors. 2) The combination of TK and ganciclovir allowed the angiogenesis to be adjusted by monitoring LDPI. 3) This therapy could be stopped at any time desired for any reason. 4) There was no possibility of the hHGF gene being expressed in nontarget organs or nontarget cells as a result of leakage or dispersion of the vectors. If the plasmid vector was integrated into the genome and neoplastic transformation occurred, it would be difficult to control cell growth. 5) The angiogenic effect can be easily predicted, because the transfection efficiency of the gene is always 100%. 6) The cell vector will be much more effective in patients who require rapid angiogenesis, because plasmid or viral vectors require a week for maximal expression, and the duration of maximal expression is short.
Angiogenic gene-modified cell transplantation therapy has several drawbacks. One is that once the cells are transplanted into patients, their growth cannot be controlled. To solve this problem, we double-transfected the cells with the TK gene, and the results confirmed that permanently transfected cells could be killed with ganciclovir after the establishment of angiogenesis and collateral vessel formation. The finding that the IC50of ganciclovir for the TK-transfected cells was 1,000 times lower than that in the nontransfected cells indicated that this system might be capable of being used in clinical settings.
We used NIH3T3, a fibroblast line derived from fetal NIH/Swiss mice, for the following reasons: 1) the transfection efficiency of the plasmid is high; and 2) their growth rate is relatively high in vitro, making it easy to expand the cells. However, their growth rate in vivo is not as high as that of carcinoma cell lines, probably because NIH3T3 cells have a mechanism of growth inhibition by cell-cell contact. To apply this method in clinical medicine, the selection of a human cell line will be requried. Considering the time and cost for preparation of the cells, an autograft might require a long time and be expensive. It took at least two months to prepare the hHGF– and TK–double-transfected cells, and a number of additional experiments were needed to confirm their effectiveness and safety. We think that allograft cells should be used to prepare gene-modified cells. In view of the time, cost, effectiveness, and safety of the cells, allografts would be much better than autografts.
Regenerative medicine has recently been the subject of investigations in many fields, and a number of regenerative cells have been established. The authors have reported that regenerative cardiomyocytes can be generated from marrow mesenchymal stem cells, and transplantation of the regenerated cells will be examined in various organs. One of the reasons why we are considering angiogenic gene-modified cell transplantation therapy is the need for a rapid blood supply to the transplanted cells. To achieve that goal, we can co-transplant target organs with these gene-modified cells in combination with the regenerated cells. Once the blood supply has become established, the angiogenic cells are no longer needed, and they can be eliminated by ganciclovir.
Bone marrow mononuclear cells have recently been used to induce angiogenesis as a means of treating arteriosclerosis obliterans (27). Although bone marrow mononuclear cells contain endothelial cells, the population of endothelial progenitor cells is <1%. The effectiveness of this therapy may be explained not only by the presence of endothelial progenitor cells but also by the fact that bone marrow mononuclear cells produce various cytokines and angiogenic growth factors. The advantage of angiogenic therapy with bone marrow mononuclear cell autografts is that the cells do not undergo immunorejection. The drawback of this therapy is that the cells may contain a variety of types of cells, such as osteogenic or chondrogenic stem cells, or induce inflammation by secreting cytokines. Using angiogenic gene-modified cells avoids the problem of transplanting different types of cells; however, the efficiency and safety of this procedure needs to be fully investigated before clinical application can become a reality.
The authors gratefully acknowledge Kensuke Kimura, MD, Isao Shibuya, PhD, and Haruko Kawaguchi, MS, for their kind assistance and helpful discussions.
☆ This study was supported in part by research grants from the Ministry of Education, Science and Culture, Japan, and by Health Science Research Grants for Advanced Medical Technology from the Ministry of Welfare, Japan.
- Dulbecco's modified Eagle's medium
- enhanced green fluorescent protein
- enzyme-linked immunosorbent assay
- human hepatocyte growth factor
- 50% (median) inhibitory concentration50.>
- laser Doppler perfusion image
- reverse transcription-polymerase chain reaction
- smooth muscle actin
- thymidine kinase
- von Willebrand factor
- Received June 28, 2003.
- Revision received December 10, 2003.
- Accepted January 5, 2004.
- American College of Cardiology Foundation
- Suri C,
- McClain J,
- Thurston G,
- et al.
- Jainchill J.L,
- Aaronson S.A,
- Todaro G.J
- McKnight S.L
- Yamada A,
- Matsumoto K,
- Iwanari H,
- et al.
- Chen S.H,
- Chen X.H,
- Wang Y,
- et al.
- Morishita R,
- Sakaki M,
- Yamamoto K,
- et al.
- Shintani S,
- Murohara T,
- Ikeda H,
- et al.
- Smee D.F,
- Martin J.C,
- Verheyden J.P,
- Matthews T.R
- Koopman G,
- Reutelingsperger C.P,
- Kuijten G.A,
- Keehnen R.M,
- Pals S.T,
- van Oers M.H
- Asahara T,
- Masuda H,
- Takahashi T,
- et al.