Conventional gene therapies still require improvement in various respects, such as transfection efficiency for targeting organs or tissues, safety (Li Q et al., 2001; Pfeifer A & Verma IM, 2001; Watson DJ et al., 2002), and so on. For example, transfection of naked plasmid DNA requires the use of a large quantity of DNA because of enzymatic degradation in the human body. Conventional non-viral vectors seem to be inferior to viral ones in efficiency, except for nucleofection (Nishikawa M & Hashida M, 2002; Schakowski F et al., 2004). On the other hand, gene therapy using viral vectors such as adenovirus, retrovirus etc., may induce tissue inflammation (Tomasoni S & Benigni A, 2004), though the efficiency of transfection is often relatively high. Also, the safety of viral transfection remains an issue, because of the risk of toxicity or mutation of viruses (Ferber D, 2001; Isner JM, 2002; Kay MA et al., 2001). Moreover, in vivo gene delivery to localized target tissues usually necessitates an invasive approach; for example, direct gene transfection to cardiomyocytes requires a surgical approach (Losordo DW et al.,1998) or catheterization (Kornowski R et al., 2000; Laitinen M et al., 2000). On the other hand, ex vivo gene transfection is less invasive. However, targeting of specific tissues by intravenous injection is usually inefficient.
Macrophages accumulate in ischemic tissues through chemotaxis (Ramsay SC et al., 1992). Therefore, we considered that intravenously administered macrophages might target ischemic tissue in vivo. Tabata, et al. reported that particles of gelatin are phagocytized by macrophages (Tabata Y & Ikada Y, 1987; Tabata Y & Ikada Y, 1988). Gelatin is a natural polymer derived from collagen, and is widely used in pharmaceutical or medical applications, i.e., it is safe and suitable for use in humans (Young S et al., 2005). Further, its isoelectric point (Ip) can be adjusted by modification of its residues, and positively charged gelatin can be impregnated with negatively charged substances (Ikada Y & Tabata Y, 1998), such as nucleic acid (Kasahara H et al., 2003). Therefore, gelatin may be suitable as a vector for transfecting phagocytes ex vivo.
Here, we describe a study aimed at examining the feasibility of a new concept for less-invasive, cell-based gene therapy by means of ex vivo gene transfection into isolated phagocytes (macrophages/monocytes) using a gelatin carrier, followed by intravenous injection of the transfected phagocytes. The present method has significant advantages over conventional cell-based gene delivery (Panetta CJ et al., 2002; Xie Y et al., 2001), in that the intravenously injected cells (phagocytes) not only produce protein from the transfected gene, but also have the potential for high tissue targeting ability. This method should be suitable for clinical gene therapy. Indeed, its clinical efficacy in patients with cardiovascular disease has recently been reported (Komeda M et al., 2009; Marui A et al., 2007; Tabata Y, 2009).
2. Main body
This study was performed in accordance with the Guideline of Tokai University School of Medicine on Animal Use, which conforms to the NIH Guide for the Care and Use of Laboratory Animals, (DHEW publication No. (NIH) 86-23, Revised 1985, Offices of Science and Health Reports, DRR/NIH, Bethesda, MD, 20205).
One hundred and twenty-one Fisher rats (male, 10 weeks old, Clea Japan Inc., Tokyo) and 61 nude SCID mice (male, 6 weeks old, Shizuoka Animal Center, Shizuoka, Japan) were used. Rats were anesthetized by inhalation of diethyl ether for harvesting of macrophages and with isofluran (1.5-3%) for thoracotomy, after which they were mechanically ventilated with a mixture of oxygen and nitrous oxide. Mice were anesthetized by intraperitoneal injection of sodium pentobarbital (40 mg/kg).
A model of myocardial ischemia-reperfusion injury was prepared in 41 rats. The remaining 80 rats were used for collecting activated macrophages. The heart was exposed via thoracotomy, and the proximal left anterior descending coronary artery was ligated (Gidh-Jain M et al.,1996) for 180 min, followed by reperfusion. A model of hindlimb ischemia was prepared in 61 mice. The left femoral artery was ligated and resected (Takeshita S et al., 1994).
Macrophages were obtained from 80 rats. Thioglycolate (4%, 8 ml) was injected into the peritoneal cavity, and after 4 days, peritoneal macrophages were collected (Ribeiro RA et al.,1991). Monocytes were obtained from peripheral blood of healthy volunteers. Leukocyte-rich plasma was obtained by dextran 500 sedimentation and layered onto Nycoprep 1.068 (Nycomed, Birmingham, UK). The monocyte-containing layer was aspirated, washed twice and allowed to adhere to the dish for 90 minutes. Fibroblasts (NIH 3T3, Invitrogen Corporation, Carlsbad, CA) were also used. The cells were resuspended in RPMI 1640 medium (Sigma) containing 5% heat-inactivated fetal calf serum and cultured for 7-14 days. The cell viability and type were determined by trypan blue exclusion and by immunostaining using anti-macrophage antibody, up to 14 days.
2.1.3. Genes and vector
Complementary DNA (cDNA) of green fluorescent protein (GFP), Renilla luciferase or human hst1/fibroblast growth factor 4 (FGF4) (Kasahara H et al., 2003) was inserted into the expression vector pRC/CMV (Invitrogen Corporation, Carlsbad, CA) and the constructs were designated as pRC/CMV-GFP, pRC/CMV-luciferase and pRC/CMV-HST1-10, respectively. Preparation and purification of the plasmid from cultures of pRC/CMV-GFP-, pRC/CMV-luciferase-, or pRC/CMV-HST1-10-transformed Escherichia coli were performed by equilibrium centrifugation in cesium chloride-ethidium bromide gradients.
Gelatin was prepared from porcine skin (Tabata Y & Ikada Y, 1987). After swelling in water the gelatin particles used in this study were spheroids with a diameter of approximately 5-30 µm, water content of 95%, and pI of 11. Gelatin (2 mg) was incubated with 50 µg of the plasmid for 7 days at 4 °C to make a gelatin-DNA complex (Tabata Y & Ikada Y, 1987).
2.1.4. Ex vivo gene transfection
Macrophages, monocytes, and fibroblasts (1 x 106) were cultured with the gelatin-DNA complex (2 mg of gelatin plus 50 µg of DNA) for 14 days on a culture dish (100 mm in diameter). Gene expression of GFP was evaluated by fluorescence microscopy and fluorescence-activated cell sorting. Luciferase activity in the cell lysate was evaluated with a photon counter system after cell lysis (Fukuyama N et al., 1996).
2.1.5. Organ distribution of phagocytes injected intravenously and directly into ischemic muscle
To examine tissue-targeting by intravenous injection of transfected phagocytes, the distribution of the cells into organs was evaluated by immunohistochemistry. In the rat model of myocardial ischemia-reperfusion injury, the GFP-gene-transfected macrophages (1.0 x 106 each) were injected into the superficial dorsal vein of the penis at the initiation of reperfusion (n=7 and 5, respectively). In the mouse model of hindlimb ischemia, the GFP-gene-transfected monocytes (1.0 x 106) were injected into the caudal vein 14 days after induction of ischemia (n=5). To examine the tissue targeting by direct local injection of transfected phagocytes, the distribution of the cells into organs was also evaluated. In the rat model of myocardial ischemia-reperfusion injury (n=7) and the mouse model of hindlimb ischemia (n=5), the same numbers of transfected macrophages and monocytes were directly injected into ischemic myocardium and ischemic skeletal muscle, respectively. Tissue samples were obtained 24 hours after cell administration. Each tissue was homogenized and the homogenate was cytospun. Immunohistochemical analysis was done with anti-GFP antibody (GFP monoclonal antibody; Clontech, USA). GFP-positive macrophages were counted in each tissue and expressed as a percentage of total GFP-positive cells.
2.1.6. Amelioration of ischemia by intravenous injection of angiogenic gene-transfected phagocytes
The angiogenic effect of intravenously injected FGF4-gene-transfected phagocytes on the ischemia models was evaluated. In the rat model of myocardial ischemia-reperfusion injury, FGF4-gene-transfected macrophages (n=5), non-transfected macrophages (1.0 x 106 each) (n=5), or saline (n=5) were injected into the superficial dorsal vein of the penis, or naked FGF4-DNA (50 µg) was injected directly into the ischemic myocardium (n=5), at the initiation of reperfusion. Fourteen days after the cell administration, blood flows in the ischemic and non-ischemic regions in the heart were evaluated with a noncontact laser Doppler flowmeter (FLO-N1, Omegawave Corporation). Then, tissue samples were obtained and histological analysis was performed. In a mouse model of hindlimb ischemia, just after induction of ischemia, FGF4-gene-transfected monocytes (n=15), non-transfected monocytes (n=8) (1.0 x 106 each), or saline (n=10) were injected into the caudal vein, or naked FGF4-DNA (50 µg) was injected directly into the ischemic muscle (n=12). Fourteen days after induction of ischemia, blood flows in the limbs were evaluated with the noncontact laser Doppler flowmeter (FLO-N1, Omegawave Corporation).
Ten-micrometer sections were cut from formalin-fixed, paraffin-embedded tissue. Two sections were used for H.E. staining and Azan staining, and eight sections were used for immunohistochemical staining. Immunohistochemical staining was performed by an indirect immunoperoxidase method. Anti-GFP antibody, anti-Mac1 antibody (BMA Biomedicals Ag, Switzerland), and anti-CD31 antibody (Serotec, UK) were used as primary antibodies. Mac1-antigen is specific to macrophages/monocytes. Anti-Ig, peroxidase-linked species-specific F(ab')2 fragments (Amersham Pharmacia Biotech UK Ltd., UK), were used as a secondary antibody. Double staining was performed with alkaline staining and peroxidase staining. The density of vessels stained with von Willebrand factor-antibody was calculated by morphometric assessment in 16 randomly selected fields for each heart and expressed as number / mm2.
2.1.8. Statistical analysis
Data are presented as mean values±SD. Differences were assessed by means of ANOVA (analysis of variance) with Scheffe's multiple comparisons test. A value of P<0.05 was considered statistically significant.
2.2.1. Ex vivo gene transfection
We studied whether genes could be transfected into isolated rat macrophages, human monocytes, and mouse fibroblasts ex vivo by using gelatin. Transfection of the GFP gene into isolated rat macrophages (Figs. 1A and B) and human monocytes (Figs. 1C and D), but not into mouse fibroblasts (data not shown), was achieved by culture with gelatin-DNA complex for 14 days. The gene transfection efficiency into rat macrophages was 68±11% (30 experiments, Fig. 2A) and that into human monocytes was 78±8% (30 experiments) as determined with a fluorescence-activated cell sorter. Sequential analysis after luciferase-gene transfection into rat macrophages revealed high expression after 14 days of culture (Fig. 2B).
2.2.2. Organ distribution of phagocytes injected intravenously or directly into ischemic muscle
We studied quantitatively whether intravenously injected luciferase-gene-transfected phagocytes could target ischemic tissues (Table 1). In non-ischemic rats, the injected macrophages were found almost exclusively in the spleen (98±4%) (n=7, Table 1). In non-ischemic mice, similar results were observed (n=7, data not shown). In a rat with myocardial ischemia-reperfusion injury, some of the intravenously injected macrophages were incorporated into the heart (Table 1).
The incorporation into the post-ischemic pericardium amounted to 13±6% (n=7) (non-ischemic rats 0±0%, n=7, Table 1). The incorporated cells expressed GFP (Fig. 3). Fibrosis with inflammatory infiltrates was recognized in the anterior wall of the left ventricle, extending to the interventricular septum (Figs. 3A and B). These infiltrates were mainly polymorphonuclear leukocytes and macrophages (Figs. 3C and D). Approximately 20 % of the macrophages showed GFP-positivity in this area (Figs. 3E and F). Similar tissue-targeting by intravenously injected monocytes was confirmed in a mouse model with hindlimb ischemia (13±7%, n=7, Table 1). Furthermore, we studied whether local intramuscular injection increased the degree of tissue targeting (Table 1). After direct injection of phagocytes into ischemic muscle, 86±10% and 88±6% of the cells remained in the target tissue in the two models. Thirteen and 11% of phagocytes injected into the cardiac or hindlimb muscle migrated to the spleen. In the other organs, accumulation of phagocytes was negligible.
2.2.3. Amelioration of ischemia by intravenously injected angiogenic-gene-transfected phagocytes
In the rat model with myocardial ischemia-reperfusion injury, we studied the angiogenic
effect of intravenously injected macrophages transfected with FGF4 gene by using gelatin. Intravenous injection of these macrophages (1.0 x 106) significantly increased the regional blood flow in the ischemic myocardium (78±7.1%, n=8, in terms of ischemic/non-ischemic myocardium flow ratio) compared with the other three treatments (P<0.05, ANOVA), that is, intravenous administration of saline (35±10%, n=8), intramuscular administration of naked DNA encoding FGF4 (50 µg, direct intramyocardial injection after thoracotomy) (58±5.3%, n=8), and intravenous administration of the same number of non-transfected macrophages (42±12%, n=8) (Fig. 4A). Histological analyses revealed angiogenesis in the ischemic tissue after the administration of transfected cells (Figs. 4B and C). Similar results were observed in the mouse model with hindlimb ischemia. Intravenous injection of FGF4-gene-transfected monocytes (1.0 x 106) enhanced regional blood flow in the ischemic leg (Fig. 4D). The increase of blood flow in the mice with transfected monocytes (93±22% in terms of flow ratio in the ischemic/non-ischemic leg) was significantly larger than those obtained with the other three treatments described above (38±12, 55±12, and 39±15%, P<0.05, ANOVA). Neither lymph node swelling in any part of the body nor pathologic change in the spleen or lung, such as angioma or abnormal immune response, was found in any of the animals.
The advantages of the present method are as follows. First, genes can easily be transfected into phagocytes (macrophages/monocytes). In preliminary experiments, we found that genes can also be transfected into endothelial progenitor cells (Nagaya N et al., 2003). Compared with other transfection methods, the transfection efficiency was high (68±11%) and it is not necessary to use a potentially hazardous viral vector (Kobinger GP et al., 2004; Lei Y et al., 2004; Watson DJ et al., 2002). Second, the phagocytes can target pathologic tissues by chemotaxis even after intravenous injection, and targeting should be considerably improved if they are administered locally. Moreover, the injection is repeatable. We confirmed that the angiogenic gene-transfected phagocytes enhanced angiogenesis after ischemia-reperfusion injury in rat heart and ameliorated ischemia in a mouse hindlimb model. Gelatin prepared as described here is not rapidly degraded in the human body, and should be effective for prolonged transfection gene therapy, as well as for drug delivery systems, etc.. This method appears to be suitable for safe and effective cell-based gene therapy.
The injected phagocytes migrated into pathologic tissues, presumably in response to the release of cytokines such as monocyte chemoattractant protein 1 by injured endothelial cells (Yoshimura T & Leonard EJ, 1990). Adhesion molecules such as P-selectin (Ikada Y et al., 2002) are probably involved in the recruitment of phagocytes to the vessel wall. The injected phagocytes also migrated to the spleen, but no pathologic change was found in the spleen.
The present method has several advantages over conventional methods of cell-based
gene therapy, such as fibroblast-based and smooth muscle cell-based approaches (Koch K-C et al., 2006; Ott HC et al., 2005; Xie Y et al., 2001). For example, monocytes do not aggregate in vessels, whereas fibroblasts or smooth muscle cells cannot be injected intravenously because of aggregation. The transfected phagocytes not only synthesize protein from the transfected gene, but also are partially targeted to the impaired tissue. In addition, the transfection rate was better than those of methods such as lipofection, viral vectors and electroporation (Kobinger GP et al., 2004; Veit K et al., 1999). The newly developed technique of nucleofection has a transfection efficiency of 40-70% (Maasho K et al., 2004), which is similar to that of our method, but our procedure is easier to use (Maasho K et al., 2004; Mertz KD et al., 2002).
The therapeutic effect obtained here (intravenous administration of phagocytes transfected ex vivo with FGF4 DNA/biodegradable gelatin complex) was superior to that of conventional gene therapy, which we reported previously, i.e., intramuscular injection of naked DNA, in heart and leg ischemia models. The major disadvantage of our method is the cell preparation time of 2 weeks before therapy can be started, and further work is needed to speed up this process. However, our results suggest that this method may an effective therapeutic option from the viewpoint of regenerative medicine.
The authors wish to thank Jobu Itoh, Yoshiko Shinozaki, and Takayuki Hasegawa for their technical work.