Transfection of Antisense p53 Tumor Suppressor Gene Oligodeoxynucleotides Into Rat Carotid Artery Results in Abnormal Growth of Vascular Smooth Muscle Cells
Background—Although loss of activity of an antioncogene, the p53 tumor suppressor gene product, has been postulated in the pathogenesis of human restenosis, little is known about the role of p53 in the regulation of vascular smooth muscle cell (VSMC) growth. In this study, to clarify the role of p53 in the pathogenesis of restenosis, we examined transfection of antisense p53 oligodeoxynucleotides (ODN) into VSMC in vitro and rat carotid artery in vivo.
Methods and Results—The specificity of antisense p53 ODN was confirmed by a significant decrease in p53 protein. Transfection of antisense p53 ODN into VSMC resulted in a significant increase in DNA synthesis and cell number as compared with sense and scrambled ODN (P<0.01). Importantly, transfection of antisense p53 ODN into rat intact carotid artery resulted in a significant increase in the ratio of neointima to medial area at 2 and 4 weeks after transfection, accompanied by a significant decrease in p53 protein (P<0.01). Moreover, cotransfection of wild-type p53 plasmid completely abolished neointimal formation induced by antisense p53 ODN. The sustained effect of a single antisense ODN administration was confirmed by the kinetics of ODN in the vessel wall with the use of FITC-labeled ODN.
Conclusions—Overall, the present study demonstrated that loss of p53 by antisense p53 ODN resulted in an abnormal VSMC growth in vitro and in vivo. These results demonstrated the potential contribution of p53 to the pathogenesis of restenosis.
Intimal hyperplasia is the pathological process that underlies restenosis after angioplasty, atherosclerosis, and vascular graft occlusion1 and develops in large part as a result of vascular smooth muscle cells (VSMC) proliferation and migration that are activated by vascular injury.1 The process of VSMC proliferation is dependent on the coordinated activation of a series of cell cycle–regulatory genes, which results in mitosis. Especially p53 (p53 tumor suppressor gene) has been postulated to negatively regulate the cell cycle in various cell types.2 3 4 However, little is known about the role of p53 in the regulation of VSMC. The presence of a functional p53 protein has been implicated as a critical determinant to regulate DNA replication, DNA repair, and programmed cell death.5 6 7 8 First, p53 has been shown to induce cell cycle arrest at the G1/S boundary border, related to its ability to induce expression of a cellular gene, WAF1/CIP1/SDI1, which encodes a 21-kDa inhibitor of G1 cyclin–dependent kinases.9 Second, p53 can induce apoptosis through bcl-2–dependent and bcl-2–independent pathways.10 11 Therefore the abnormal progression of the cell cycle seen in cancer is thought to result from the mutation of these negative cell cycle regulatory genes, especially p53.2 3
The importance of cell cycle regulation in VSMC is apparently great, since we and others have previously reported the successful prevention of restenosis after angioplasty by using antisense oligodeoxynucleotides (ODN) against cell cycle–regulatory genes, decoy cis element of E2F binding site, and gene transfer of nonphosphorylated Rb (retinoblastoma gene).12 13 14 Of importance, recent studies have suggested that loss of p53 activity by cytomegalovirus infection might be responsible for the pathogenesis of human restenosis.15 16 17 From this viewpoint, it is necessary to understand the negative regulation of the cell cycle by p53 in VSMC. This study examined the role of p53 in negative regulation of the cell cycle in the pathogenesis of restenosis, with the use of antisense strategy in vivo. The present study successfully demonstrated in vivo evidence of importance of p53 in the regulation of VSMC growth.
In Vitro Transfection of Antisense p53 ODN Into VSMC
Initially, we tested whether transfection of antisense p53 ODN into VSMC has an effect on VSMC growth. Transfection of antisense p53 ODN resulted in a significant increase in DNA synthesis assessed by 3H-labeled thymidine incorporation as compared with untreated and sense p53 ODN-treated VSMC (Figure 1a⇓, P<0.01). Of importance, transfection of antisense p53 ODN also resulted in a significant increase in the number of VSMC as compared with sense ODN-transfected VSMC (P<0.01, Figure 1b⇓). The specificity of antisense p53 ODN was confirmed by the observation that a marked decrease in p53 protein was only observed with transfection of antisense p53 ODN but not sense p53 ODN or HVJ complex alone (Figure 2⇓).
In Vivo Transfection of Antisense p53 ODN Into Rat Intact Carotid Artery
As shown in Figure 3⇓, successful transfer of antisense p53 ODN by means of HVJ-liposome complex was confirmed by Western blotting of p53 protein. Five days after transfection, Western blots from untransfected intact vessels and intact vessels transfected with sense ODN demonstrated a single band corresponding to p53 protein. Administration of antisense p53 ODN decreased p53 protein but did not affect the level of tubulin protein (Figure 3⇓). Transfection of antisense p53 ODN into intact rat carotid artery resulted in significant decrease in p53 expression as compared with transfection of HVJ alone or sense ODN (P<0.01).
As shown in Figure 4a⇓, PCNA-stained nuclei were observed in the medial layer of blood vessels transfected with antisense p53 ODN at 4 days after transfection, whereas few PCNA-stained nuclei were found in vessels transfected with sense p53 ODN. In addition, untransfected intact carotid arteries also exhibited few PCNA-stained nuclei. Consistent with in vitro experiments, the ratio of PCNA-positive nuclei to total cells in the blood vessels transfected with antisense p53 ODN was significantly increased as compared with those transfected with sense p53 ODN (P<0.01, Figure 4b⇓). Next, we examined the effects of antisense p53 ODN on neointimal formation after transfection. As shown in Figure 5⇓, a and b, the vessels transfected with HVJ-liposome complex without ODN, sense p53 ODN, or scrambled p53 ODN exhibited little neointimal formation at 2 weeks after transfection. In contrast, a single administration of antisense p53 ODN resulted in a significant increase in the ratio of neointimal area to medial area (Figure 6⇓). There was no significant difference in medial area among HVJ-transfected, sense p53 ODN-transfected, scramble p53 ODN-transfected, and antisense p53 ODN-transfected vessels (Table 1⇓). Further evidence for the specificity of antisense p53 ODN was provided by the observation that treatment with neither antisense ACE ODN nor antisense thrombomodulin ODN affected neointimal formation (antisense ACE 124±95 μm2 vs antisense thrombomodulin 156±101 μm2, P>0.05). In addition, neointimal formation was due to the lack of p53 activity, as cotransfection of wild-type p53 plasmid and antisense p53 ODN completely abolished neointimal formation induced by antisense p53 ODN (Figure 7⇓). Surprisingly, neointimal formation was still observed 4 weeks after a single transfection in blood vessels transfected with antisense p53 ODN (P<0.01, Figure 8⇓). There was no significant difference in medial area between sense p53 ODN-transfected and antisense p53 ODN-transfected vessels (Table 2⇓).
To confirm the sustained effect of a single intraluminal administration of antisense p53 ODN, we performed in vivo transfer of FITC-labeled ODN into intact carotid arteries by using the HVJ-liposome method. Transfer of FITC-labeled phosphorothioate ODN by the HVJ method resulted in widespread distribution of fluorescence in medial vascular cells even 2 weeks after transfection (Figure 9⇓). The fluorescence was localized primarily in cell nuclei and persisted up to 2 weeks after transfection. Untreated or HVJ complex without ODN-treated vessels revealed no specific fluorescence in the elastic lamina. Intraluminal incubation of free FITC did not result in specific fluorescence in the vessel wall, demonstrating that this fluorescence was specific for FITC-labeled ODN (also see Figure 9⇓).
A fundamental pathological feature of vascular disease is the abnormal accumulation of cells within the intimal space, resulting in neointimal formation produced by alterations in the homeostatic balance between cell growth and cell death.18 Of importance, loss of p53 activity induced by cytomegaloviral infection might be related to the pathogenesis of human restenosis.15 16 17 However, there is no direct in vivo evidence that loss of p53 might induce restenosis. Thus it is important to understand the role of p53 in the regulation of VSMC. Our in vitro experiments demonstrated that transfection of antisense p53 ODN promoted DNA synthesis and growth of VSMC. Since regulation of the intimal cell population requires a delicate balance between cell influx, cell growth, and cell death,19 cells lacking p53 undergo cell cycle progression, resulting in abnormal cell growth. More importantly, in vivo transfection of antisense p53 ODN into intact carotid artery also resulted in a significant increase in neointimal area at 2 and 4 weeks after transfection. Several observations provide evidence that VSMC growth is induced by lacking p53: (1) PCNA-positive cells were only detected in blood vessels transfected with antisense p53 but not sense p53 ODN; (2) neointimal formation induced by antisense p53 ODN was observed both at 2 and 4 weeks after transfection; (3) neointimal formation was only observed in blood vessels transfected with antisense p53 ODN but not with HVJ-liposome alone, sense, or scrambled ODN; (4) neointimal formation was accompanied by a significant decrease in vascular p53 protein as compared with HVJ-liposome alone or sense ODN; (5) the increase in neointimal area was abolished by cotransfection of wild-type p53 plasmid. Because p53 plays an important role in the regulation of apoptosis,6 7 8 decrease in apoptosis by antisense p53 ODN also might be involved in the neointimal formation. The present study is consistent with the previous report that the atherosclerotic lesions in p53−/−/apoE−/− mice had a significant increase in cell proliferation rate compared with those in p53+/+/apoE−/− mice.20 p53 may play a pivotal role in restenotic as well as atherosclerotic lesion development associated with its function in cell replication control.
To confirm the sustained effect of a single administration of ODN, we examined whether ODN persisted in the vessel wall by using the HVJ-liposome method. In previous studies, the HVJ-liposome–mediated transfer method for antisense ODN, which uses the fusion system mediated by HVJ, results in intracellular ODN delivery bypassing endocytosis.21 22 Consequently, this method increases the effectiveness and prolongs the half-life of antisense ODN in vitro.22 The present study demonstrates that the HVJ-liposome method actually prolongs the in vivo half-life of ODN. In fact, there is evidence that FITC-labeled ODN persist in the nucleus for at least 2 weeks after a single transfection. This method is clearly efficient for inducing sustained action of antisense ODN in vivo. Indeed, a high transfection efficiency of the HVJ-liposome method into blood vessels has also been reported.23 Yonemitsu et al23 documented that HVJ-liposomes could achieve highly efficient gene transfection into the medial smooth muscle cells of intact arteries at 150 and 760 mm Hg of pressure (mean=85.3% and 93.5% of total VSMC) without any inflammatory reaction for ≥14 days.
Overall, the present study demonstrated that loss of p53 by antisense p53 ODN resulted in abnormal VSMC growth in vitro as well as in vivo. The observation in the present study is extremely important for understanding the molecular mechanisms of restenosis. In contrast, researchers have focused on the inhibitory effect of p53 on VSMC growth as a potential tool in molecular therapy to treat restenosis. Indeed, overexpression of the p53 gene in the injured arterial wall inhibits the proliferation of VSMC in vivo.24 Alternatively, a single intraluminal incubation of human p21, identified as a downstream mediator of p53 gene, resulted in significant inhibition of neointimal formation after vein grafting, associated with a phenotypic change of VSMC from neonatal to adult type in a rabbit model.25 Further studies of p53 in the pathological conditions in vascular disease may provide new insights into therapeutic strategies.
In Vitro Experiments
Human aortic VSMC (passage 5) were obtained from Clonetics Corp and cultured in modified MCDB131 medium supplemented with 5% fetal calf serum, 100 U/mL penicillin, 100 mg/mL streptomycin, 10 ng/mL epidermal growth factor, 2 ng/mL bFGF, and 1 mmol/L dexamethasone in the standard fashion.26 These cells showed the specific characteristics of VSMC by immunohistochemical examination and morphological observation. All the cells were used within passage 5 to 7.
Synthesis of ODN and Selection of Target Sequences
The sequences of phosphorothioate ODN against p53 were as previously reported, with the ATG initiation codon or its complement CAT in the sense and antisense, respectively19 : (p53-sense: 5′-ACTGCCATGGAGGAGCCG-3′, p53-antisense: 5′-CGGCTCCTCCATGGCAGT-3′). These antisense ODN inhibit p53 synthesis in human hematopoietic cells.19 We also used scrambled p53 ODN (5′-AGTGGCCTCCATCTCCTCCGG-3′) as a negative control.
VSMC were seeded onto 10-cm dishes (Corning). VSMC were grown to confluence and made quiescent by incubation in defined serum-free medium, which was supplemented with insulin (5×10−7 mol/L), transferrin (5 mg/mL), and ascorbate (0.2 mmol/L)27 before transfection. Seventy-two hours after transfection, the cells were fixed with 10% trichloroacetic acid (TCA) in saline, followed by extraction of total protein with urea-TX (9 mol/L urea, 2% Triton-X and 5% 2-mercaptoethanol). Samples containing 100 μg protein were run on 12.5% SDS polyacrylamide gels. Proteins were incubated with a monoclonal antibody to p53 (1:20, Calbiochem) at 4°C overnight. To quantify and compare levels of proteins, the density of each band was measured by densitometry. Staining with Coomassie Brilliant Blue R (Sigma) was used to confirm that identical protein amounts were loaded in all samples for Western blotting. Western blotting of tubulin with antitubulin antibody (anti-human mouse IgG; 1:100; Oncogene) was also performed to confirm equal amounts of loaded proteins.
Effect of Antisense ODN on DNA Synthesis and VSMC Growth
VSMC were seeded onto 96-well tissue culture plates. At 80% confluence, VSMC were rendered quiescent by incubation for 48 hours in defined serum-free medium. Then, 10 μL HVJ-liposomes (containing 15 μmol/L ODN) was added to the wells. The cells were incubated at 4°C for 10 minutes and then at 37°C for 30 minutes. The cells were washed 3 times with balanced salt solution (BSS) containing 2 mmol/L CaCl2. Relative rates of DNA synthesis were assessed by determination of 3H-thymidine incorporation into TCA-precipitable material over the next 24 hours after 16 hours of incubation. After 3 days, an index of cell proliferation was determined with the use of a WST cell counting kit (Wako).
Preparation of HVJ-Cationic Liposomes In Vitro
In this study, we used HJV-cationic liposome method as the in vitro transfection method because this method demonstrated highly effective transfection efficiency, especially in in vitro cultured cells.28 In brief, a lipid mixture containing 6 mg phosphatidylcholine, 3 mg cholesterol, and 0.75 mg 3β-[N-(N′ N-dimethylaminoethane)carbamoyl]cholesterol was dissolved in chloroform and evaporated with the use of a rotary evaporator. The dried mixture was hydrated with 200 μL of BSS (137 mmol/L NaCl, 5.4 mmol/L KCl, 10 mmol/L Tris-HCl, pH 7.6) containing ODN. Purified HVJ (Z strain) was inactivated by UV irradiation (110 erg/mm2 per second) for 3 minutes just before use. The liposome suspension was mixed with HVJ in a total volume of 1 mL BSS. Free HVJ was removed from the HVJ liposomes by sucrose density gradient centrifugation. The top layer of the sucrose gradient was collected for use.
In Vivo Experiments
Synthesis of ODN and Selection of Target Sequences
The sequences of phosphorothioate ODN against p53 used in in vivo study were changed to the identical sequence of rat p53: (p53-sense: 5′-ACTGTCATGGAGGATTCA-3′, p53-antisense: 5′-TGAATC-CTCCATGACAGT-3′). We also used scrambled p53 ODN (5′-AGTGGCCTCCATCTTATCAAG-3′) as a negative control. Antisense thrombomodulin (5′-ACCCAGAAAGAAAATCCC-3′; 12) and antisense ACE (5′-GCCCCCATGGCGCGGT-3′)(position −8 to +8 of the rat sequence) were used as additional negative controls.
In Vivo Transfer Into Intact Rat Carotid Artery
Male Sprague-Dawley rats (weight 250 to 300 g; Charles River) were anesthetized with ketamine, and the left common carotid artery was surgically exposed.12 13 A cannula was introduced into the common carotid artery through the external carotid artery. Then, 200 μL of HVJ-liposome complex was infused into the segment and incubated for 10 minutes at room temperature. After 10-minute incubation, the infusion cannula was removed. After transfection, blood flow to the common carotid artery was restored by release of the ligatures, and the wound was closed. No adverse neurological or vascular effects were observed in any animal undergoing this procedure.
Conventional HVJ-Liposome Preparation In Vivo
For in vivo experiments, we used the conventional anionic HVJ-liposome method.12 13 29 Briefly, phosphatidylserine, phosphatidylcholine, and cholesterol were mixed in a weight ratio of 1:4.8:2. Dried lipid was hydrated in BSS containing sense or antisense ODN. Liposomes were prepared by shaking and sonication.
For the measurement of vascular p53 content, rats were killed 5 days after transfection. After infusion of PBS, the carotid arteries were removed and dissected free of periadventitial tissues and immediately frozen in liquid nitrogen. Western blotting was performed for analysis of p53 protein, as described above. Carotid arteries frozen in liquid nitrogen were physically crushed with a Cryo-Press CP-100 (Microtech Nichion Corp), followed by extraction of total protein with lysis buffer (50 mmol/L Tris-HCl, pH 8.0, 20 mmol/L EDTA, 1% SDS, 100 mmol/L NaCl). Samples containing 30 μg protein were run on 12.5% SDS polyacrylamide gels. Proteins were incubated with a monoclonal antibody to p53 (1:100; Calbiochem) at 4°C overnight.
Quantification of Cell Proliferation in Medial Lesions
Paraffin-embedded sections of carotid artery harvested 4 days after transfection were used for quantification of cell proliferation.12 13 Four sections of each vessel spaced at 0.4-mm intervals were measured by a computerized image analyzer system (Image Command 5098, Olympus). Monoclonal antibody against proliferating cell nuclear antigen (PCNA) (PC-10; 1:500, Neo Markers) was used as specific markers for proliferating cells in paraffin-embedded sections. The number of PCNA-positive nuclei was counted in the media by the image analyzer system at ×400 magnification. Frequency of cell proliferation was expressed as PCNA index, defined as the ratio of the number of PCNA-positive nuclei to that of all the nuclei in the media. Samples were coded so that the analysis was performed without knowledge of which treatment each individual vessel had received. Intraobserver variability was determined from triplicate measurements performed by 1 observer for all sections (2.4±0.3%).
Measurement of Neointimal Formation
At 2 and 4 weeks after transfection, each carotid artery was processed for morphological study.12 13 For histological analyses, a segment of each artery was perfusion-fixed with 4% paraformaldehyde at physiological pressure (110 mm Hg) and subsequently processed. Medial and luminal areas were measured on a digitizing tablet after staining with hematoxylin. The medial area of the vessel was readily demarcated as the area between the internal and external elastic laminae. At least 3 individual sections from the middle of the transfected arterial segments were analyzed. Animals were coded so that the analysis was performed without knowledge of which treatment each individual animal had received.
Cotransfection of wild-type p53 plasmid was also examined. The p53 expression vector is driven by cytomegalovirus promoter.24 As the control, we constructed a control expression vector driven by the same promoter without p53 plasmid.
In Vivo Transfection of FITC-Labeled ODN
FITC-labeled phosphorothioate ODN on the 3′ and 5′ ends were provided by Nihon Seifun Inc. Transfer of phosphorothioate FITC-labeled ODN was performed under the following protocol; HVJ complex with FITC-labeled phosphorothioate ODN (3 μL) was incubated for 10 minutes. The vessels were harvested 2 weeks after transfection and perfusion-fixed with 4% paraformaldehyde. Sections were examined by fluorescence microscopy after staining in Erichrome black T solution. Elastic fibers stained dark red and were readily distinguishable from the specific FITC ODN by the treatment of Erichrome black T solution.30
All values are expressed as mean±SEM. ANOVA with subsequent Bonferroni’s test was used to determine the significance of differences in multiple comparisons. A value of P<0.05 was considered significant.
This work was partially supported by grants from the Uehara Memorial Foundation, the Hoan-sya Foundation, the Japan Cardiovascular Research Foundation, a Japan Heart Foundation Research Grant, and a grant-in-aid from The Tokyo Biochemical Research Foundation.
The Methods section of this article can be found at http://www.circulationaha.org
- Received July 19, 1999.
- Revision received September 24, 1999.
- Accepted October 6, 1999.
- Copyright © 2000 by American Heart Association
Libby PD, Schwartz E, Brogi H, Tanaka S, Clinton K. A cascade model for restenosis: a special case of atherosclerosis progression. Circulation. 1992;86(suppl III):III-47–III-52.
Hollstein M, Sidransky D, Vogelstein B, Harris CC. P53 mutations in human cancers. Science. 1991;253:49–53.
Shaw P, Bovey R, Tardy S, Sahli R, Sordat B, Costa J. Induction of apoptosis by wild-type p53 in a human colon tumor-derived cell line. Proc Natl Acad Sci U S A. 1992;89:4495–4499.
Yonish-Rouach E, Resnitzky D, Lotem J, Sachs L, Kimchi A, Oren M. Wild-type p53 induces apoptosis of myeloid leukaemic cells that is inhibited by interleukin-6. Nature. 1991;352.
Miyashita T, Harigai M, Hanada M, Reed JC. Identification of a p53-dependent negative response element in the bcl-2 gene. Cancer Res. 1994;54:3131–3135.
Morishita R, Gibbons GH, Ellison KE, Nakajima M, Zhang L, Kaneda Y, Ogihara T, Dzau VJ. Single intraluminal delivery of antisense cdc 2 kinase and PCNA oligonucleotides results in chronic inhibition of neointimal hyperplasia. Proc Natl Acad Sci U S A. 1993;90:8474–8478.
Morishita R, Gibbons GH, Horiuchi M, Ellison KE, Nakajima M, Zhang L, Kaneda Y, Ogihara T, Dzau VJ. A novel molecular strategy using cis element “decoy” of E2F binding site inhibits smooth muscle proliferation in vivo. Proc Natl Acad Sci U S A. 1995;92:5855–5859.
Chang MW, Barr E, Seltzer J, Jiang YQ, Nabel GJ, Nabel EG, Parmacek MS, Leiden JM. Cytostatic gene therapy for vascular proliferative disorders with a constitutively active form of the retinoblastoma gene product. Science. 1995;287:518–522.
Speir E, Modali R, Huang ES, Leon MB, Shawl F, Finkel T, Epstein SE. Potential role of human cytomegalovirus and p53 interaction in coronary restenosis. Science. 1994;265:391–394.
Tsai HL, Kou GH, Chen SC, Wu CW, Lin YS. Human cytomegalovirus immediate-early protein IE2 tethers a transcriptional repression domain to p53. J Biol Chem. 1996;271:3534–3540.
Gibbons GH, Dzau VJ. The emerging concept of vascular remodeling. N Engl J Med. 1994;30:1431–1438.
Bi S, Lanza F, Goldman JM. The involvement of “tumor suppressor” p53 in normal and chronic myelogenous leukemia hemopoiesis. Cancer Res. 1994;54:582–586.
Morishita R, Gibbons GH, Ellison KE, Nakajima M, Leyen HVL, Zhang L, Kaneda Y, Ogihara T, Dzau VJ. Intimal hyperplasia after vascular injury is inhibited by antisense cdk 2 kinase oligonucleotides. J Clin Invest. 1994;93:1458–1464.
Morishita R, Gibbons GH, Horiuchi M, Nakajima M, Ellison KE, Lee W, Kaneda Y, Ogihara T, Dzau VJ. Molecular delivery system for antisense oligonucleotides: enhanced effectiveness of antisense oligonucleotides by HVJ-liposome mediated transfer. J Cardiovasc Pharmacol Ther. 1996;2:213–222.
Yonemitsu Y, Kaneda Y, Morishita R, Nakagawa K, Nakashima Y, Sueishi K. Characterization of in vivo gene transfer into the arterial wall mediated by the Sendai virus (hemagglutinating virus of Japan) liposomes: an effective tool for the in vivo study of arterial diseases. Lab Invest. 1996;75:313–323.
Yonemitsu Y, Kaneda Y, Tanaka S, Nakashima Y, Komori K, Sugimachi K, Sueishi K. Transfer of wild-type p53 gene effectively inhibits vascular smooth muscle cell proliferation in vitro and in vivo. Circ Res. 1998;82:147–156.
Bai HZ, Morishita R, Kida I, Yamakawa T, Zhang WD, Aoki M, Matsushita H, Noda A, Nagai R, Kaneda Y, Higaki J, Ogihara T, Sawa Y, Matsuda H. Inhibition of intimal hyperplasia after vein grafting by in vivo transfer of human senescent cell-derived inhibitor-1 gene. Gene Ther. 1998;5:761–769.
Bonin PD, Leadley RJ, Erickson LA. Growth factor-induced modulation of endothelial-1 binding to human smooth muscle cells. J Cardiovasc Pharmacol. 1993;22:S125–S127.
Kaneda Y, Iwai K, Uchida T. Increased expression of DNA cointroduced with nuclear protein in adult rat liver. Science. 1989;243:375–378.