H U M A N G E N E T H E R A P Y 7:1431-1436 (August 1, 1996) Mary Ann Liebert, Inc. G r a n u l o c y t e C o l o n y - S t i m u l a t i n g Factor E x p r e s s i o n f r o m T r a n s d u c e d V a s c u l a r S m o o t h M u s c l e Cells P r o v i d e s Sustained N e u t r o p h i l Increases in R a t s DANIEL V. LEJNIEKS,' SANG W O N HAN,i-2 N. RAMESH,' STELLA LAU,' and WILLIAM R.A. OSBORNE' ABSTRACT Granulocyte colony-stimulating factor (G-CSF) regulates granulocyte precursor cell proliferation, neutrophil survival, and activation. Cyclic hematopoiesis, a disease that occurs both in humans and grey collie dogs is characterized by cyclical variations in blood neutrophils. Although the underlying molecular defect is not knovrn, long-term daily administration of recombinant G-CSF eliminates the severe recurrent neutropenia, indicating that expression of G-CSF by gene therapy would be beneficial. As a prelude to preclinical studies in affected collie dogs, w e monitored hematopoiesis in rats receiving vascular smooth muscle cells transduced to express G-CSF. Cells transduced with L r G S N , a retrovirus expressing rat G-CSF, were implanted in the carotid artery and control animals received cells transduced with L A S N , a retrovirus expressing h u m a n adeno­ sine deaminase ( A D A ) . Test animals showed significant increases in neutrophil counts for at least 7 weeks, with mean values of 3,670 ± 740 cellsl/il in comparison to 1,870 ± 460 cells//il in controls (p < O.OOI). Thus, in rats G - C S F gene transfer targeted at vascular smooth muscle cells initiated sustained production of 1,800 neutrophils//tl, a cell number that would provide clinical benefit to patients. Lymphocytes, red cells and platelets were not difTerent between control and test animals (p > 0.05). These studies indicate that retrovi­ rally transduced vascular smooth muscle cells can provide sustained clinically useful levels of neutrophils in vivo. O V E R V I E W S U M M A R Y 3,570 ± 740 cells//nl in comparison to 1,870 ± 460 cells//itl in controls (p < 0.001). This suggests that G-CSF gene Cyclic hematopoiesis is a disease that occurs both in hu- transfer targeted at vascular smooth muscle cells mediated mans and grey collie dogs and is characterized by cyclical production of 1,800 cellsVl, a neutrophil count that would variations in blood neutrophils. Although the underlying provide clinical benefit to patients. These studies show that molecular defect is unknown, long-term daily administra- retrovirally transduced vascular smooth muscle cells can tion of recombinant granulocyte colony-stimulating factor provide sustained clinically relevant levels of neutrophils in (G-CSF) eliminates the severe recurrent neutropenia, indi- y,-ŷ_ eating that expression of G-CSF by gene therapy would be beneficial. In a preclinical rat model, we monitored neu­ trophil production in rats receiving vascular smooth mus- I N T R O D U C T I O N cle cells transduced to express G-CSF. Cells transduced with LrGSN, a retrovirus expressing rat G-CSF, were implanted ^ ^ ranltlocyte colony-stimulating factor (G-CSF) is a in the carotid artery and control animals received cells V J cytokine that selectively stimulates the proliferation and transduced with LASN, a retrovirus encoding human differentiation of neutrophil precursors and accelerates neu- adenosine deaminase. Test animals showed a doubling of trophil matoration and release from the martow (Demefri and neutrophil counts for at least 7 weeks, with mean values of Griffin, 1991; Morstyn and Dexter, 1993). Recentiy, recombi- 'Department of Pediatrics, University ofWashington, Seattle, W A 98195. D̂epartment of Biochemistry-IB, Unesp Rio Claro-SP, 13500, Brazil. 1431 1432 L E J N I E K S E T A L . nant G-CSF has been used to treat chronic neutropenias of var­ ious causes to decrease morbidity and mortality due to infection (Morstyn and Dexter, 1993). Regardless of cause, treataient of neutropenic individuals with recombinant G-CSF often results in at least a partial normalization of neutrophil number and func­ tion. However, due to the symptomatic rather than curative na­ tare of tiie treatment, these diseases often require lifelong daily injections with this honnone. In this category is cyclic hematopoiesis, a disease that occurs both in humans and grey collie dogs, and is characterized by cyclical variations in blood neutrophils, monocytes, lymphocytes, eosinophils, reticulocytes, and platelets due to periodic fluctoations in blood cell produc­ tion by the bone martow (Dale et al, 1972; lones and Lange, 1983; Dale and Hammond, 1988). The recurrent severe neu­ tropenia leads to bacterial infections and shortened life ex­ pectancy. The disorder can be cured by bone marrow transplan­ tation in grey collie dogs as well as in humans; and in both dogs and humans the disease can be transferted from an affected to a normal (Dale and Graw, 1974; Weiden et al, 1974; Krance et al, 1982). This transplantability strongly supports the concept that this is a disease of defective regulation of hematopoietic stem cells. Although the underlying molecular defect is not known, long-term daily administration of recombinant G-CSF eliminates the severe recurtent neutropenia (Lothrop et al, 1988; H ammond etal, 1989,1990). The constimtive expression ofG-CSF by gene therapy would provide clinical and probably economic benefits. The major colony-stimulating factors have been expressed in several cell types following gene transfer in vitro (Lang et al, 1985; Laker et al, 1987; W o n g et al, 1987; Browder et al, 1989), and the retrovims-mediated transfer and expression of interleukin-3 (IL-3) (Wong et al, 1989), granulocyte- macrophage colony-stimulating factor (GM-CSF) (Johnson et al, 1989), and G-CSF (Chang efaZ., 1989) in mouse hematopoi­ etic cells in vivo has been described. The retrovirally expressed IL-3 (Wong et al, 1989) and G M - C S F (Johnson et al, 1989) produced a fatal myeloproliferative syndrome in treated mice. Most noteworthy, however, long-term expression of G-CSF produced sustained neutrophilia that was not associated with disease (Chang et al, 1989). Similarly, a mouse mammary to­ mor constitatively expressing G-CSF produced sustained neu­ trophilia in mice without myeloproliferative disease (Lee and Lottsfeldt, 1984). The severe recurtent neutropenia in grey col­ lie dogs was not abrogated by in vivo IL-3 or G M - C S F treat­ ment (Hammond et al, 1990). IL-3 caused eosinophilia, whereas recombinant human G M - C S F caused neutrophilia and eosinophilia, but with both agents cycling of hematopoiesis per­ sisted (Hammond et al, 1990). In contrast, G-CSF prevented the recurtent neutropenia and obliterated periodic fluctuation of monocyte, eosinophil, reticulocyte, and platelet counts (Lothrop et al, 1988; H a m m o n d et al, 1990). W e have shown in rats long-term expression of human A D A (Lynch et al, 1992; Clowes et al, 1994) and erythropoietin (Osbome et al, 1995) from transduced smooth muscle cells seeded into carotid arteries. In vascular injury, where the ves­ sel is not completely re-endothelialized, vascular smooth mus­ cle cells form a pseudoendothelium that is in direct contact with the blood (Clowes et al, 1983). Vascular smooth muscle cells are readily obtained, cultured, transduced, and implanted, mak­ ing these cells a generally useful target tissue for gene therapy. W e recentiy cloned rat G-CSF (Han et al, 1996) and as a pre­ lude to the treatment of collie dogs with cyclic hematopoiesis, we investigated the ability of transduced vascular smooth mus­ cle cells to provide therapeutic levels of G-CSF in rats. MATERIALS AND METHODS Retroviral vectors The G-CSF expression vector was made by digesting rat G-CSF c D N A (Han et al, 1996) with Eco RI and Dra I and lig­ ating the isolated D N A fragment (700 bp) into the viral plasmid L X S N (Miller and Rosman 1989), previously digested with Eco RI and Hpa I to provide tiie expression vector LrGSN. In LiGSN, rat G-CSF c D N A expression is driven by tiie strong viral L T R promoter and the neo resistance gene is expressed ffom SV-40 early region promoter/enhancer (Hock et al, 1989). From PA317 packaging cell lines (Miller and Buttimore, 1986), LiGSN had a viral titer of 8 X 10'̂ cfu/ml. The retroviral vector L A S N , which encodes nonsecreted human adenosine deaininase (ADA) (Hock et al, 1989), was chosen as a control vector. Cell culture and transduction Ecotropic PE501 and amphotropic PA317 retrovims pack­ aging cell lines (Miller and Buttimore, 1986; Miller and Rosman, 1989), NIH-3T3 thymidine kinase-negative cells (MUler and Buttimore, 1986), and primary cultures of rat vas­ cular smooth muscle cells were grown in Dulbecco/Vogt-mod- ified Eagle's medium ( D M E M ) with high glucose (4.5 gram/liter) supplemented with 1 0 % fetal bovine seram in hu­ midified 5 % C02/95% air at 37°C. Rat smooth muscle cell cultores were prepared by enzymatic digestion of the aorta from male Fisher 344 rats. These cells were characterized by positive staining for muscle cell-specific actins with H H F 3 5 antibody (Geary et al, 1994) while stain­ ing negative for von Willebrand factor (Geary et al, 1994), an endothelial cell-specific marker. Early-passage smooth muscle cells were exposed to 16-hr viras harvests from PA317-LrGSN and PA317-LASN amphotropic viras-producing cell lines for a period of 24 hr in the presence of Polybrene (4 figlml), and se­ lected in G418 antibiotic (1 mg/ml). G-CSF bioassay Cytokine secretion ffom LrGSN-transduced cells was moni­ tored using a murine cell line, NFS-60, that proliferates in re­ sponse to G-CSF (Dale et al, 1992). Recombinant canine G-CSF (kindly supplied by Amgen, Thousand Oaks, C A ) was used to constract a proliferation-response curve with murine NFS-60 cells (Dale et al, 1992). In brief, short-tenn prolifera­ tion was determined by measuring tritiated thymidine incorpo­ ration by cells seeded at a concentration of 10' cells/well in 96- well microtiter plates. Proliferative response to conditioned medium was measured after 24 hr at 37°C, 5 % C O 2 and har­ vesting the cells 4 hr later on fiberglass filters using an automated cell harvester (Cambridge Technology, Cambridge, M A ) . Cell implantation Male Fisher 344 rats (275-325 grams) were premedicated witti 0.04 mg/kg atropine subcutaneously and 2.5 mg/kg enrofloxacin G-CSF EXPRESSION IN R A T S 1433 i.m. and were anesthetized using 44 mg/kg ketamine, 5 mg/kg xylazine, and 0.5 mg/kg acepromazine i.p. Animals were placed in dorsal recumbency and an incision was made along the ven­ tral midline of ttie neck from the angle of the mandible to the ttioracic inlet (Clowes et al, 1994). The left common carotid and its intemal and extemal branches were exposed using blunt dis­ section. Temporary ligatures were placed around the caudal com­ mon carotid artery and the cranial extent of the intemal carotid. The extemal carotid was permanentiy ligated and an arterotomy made between this ligature and the bifiircation of the extemal and intemal branches. A Fogarty 2F arterial embolectomy cattieter was inserted through the arteriotomy site and the inte­ rior of the common carotid was balloon injured. The balloon cattieter was withdrawn and a 24-gauge angiocatheter (Becton Dickinson) was inserted. A total of 2 X 10* transduced vascular smooth muscle cells were introduced into the lumen of the com­ mon carotid artery and allowed to seed the artery wall for 15 min. The angiocatiieter was withdrawn and the arteriotomy site was closed by placing a second permanent ligatare around the exter­ nal carotid artery just caudal to the site. Blood flow was re-es­ tablished through the common carotid artery and the intemal carotid artery and the incision was closed. Blood counts Anticoagulated blood samples (100 fil) were obtained from the tail vein with the animals under light ether anesthesia. Samples were obtained from 19 days before surgery and from 3 to 7 days post surgery for up to 7 weeks. Total white blood cells (WBC) platelets, and hematocrit values were determined using a Coulter T-540 counter and differential W B C counts were obtained manually. RESULTS We monitored NFS-60 cell proliferation in the presence of conditioned medium from PA 317-LrGSN amphotropic pack­ aging cells and LrGSN-transduced Fisher rat vascular smooth 5000 4000 ^ 3000 O O z < 2000 1000 FIG. 1. Effect of seeding transduced vascular smooth muscle cells on absolute neutrophil count (ANC). Open symbols, ani­ mals implanted with LrGSN-transduced cells; solid symbols, animals receiving LASN-transduced cells. Surgery was on day zero. 1434 LEJNIEKS E T AL. muscle cells, using PA 317-LASN packaging cells and LASN- transduced smooth muscle cells, respectively, to provide con­ trol medium. G-CSF expression was 2 ng/24 hr per iC cells from packaging cells and 8 ng/24 hr per IO'' cells from trans­ duced smooth muscle cells. These assays indicate expression of a bioactive gene product from our retroviral vector, but are probably an underestimate because purified rat G-CSF was not available and recombinant canine G-CSF was used to generate a standard curve. The absolute neutrophil counts of animals receiving LrGSN- transduced cells increased rapidly after cell implantation, and by day 10 a relatively constant elevated plateau had been achieved which was sustained for at least 7 weeks (Fig. 1). In contrast, neutrophti counts obtained before and after surgery from animals seeded with LASN-transduced cells did not show changes (Fig. 1). This suggests that the consequences of surgery were not responsible for the increased number of neutrophils observed in the test rats. The surgical procedure was well tol­ erated by the animals, witii no evidence of fever or other malaise, and this may be reflected in the lack of neutrophil in­ creases in animals receiving LASN-transduced cells. Pooled hematopoietic cell data from 6 control and 8 test rats are shown in Table 1. Neutrophil counts recorded from rats after LrGSN- cell implantation were significantiy elevated over control rats receiving LASN-transduced cells (p < 0.001). A simtiar sta­ tistical comparison of platelets, lymphocytes, and red cells, as measured by hematocrit, showed no significant differences (p > 0.05). This is noteworthy as previous stadies of G-CSF administration to mice have shown reduction in red cell num­ bers (Molineux et al, 1990; Pojda et al, 1990). D I S C U S S I O N We have shown that retrovirally transduced vascular smooth muscle cells allow sustained expression of G-CSF that stimulated significantiy elevated neutrophil production for pe­ riods of up to 7 weeks. In treated rats, we documented mean increases of 1,800 neutrophils//Al, which would be a thera­ peutic cell number for patients with severe chronic neutrope­ nia or cyclic hematopoiesis. In these patients and cyclic neu­ tropenic dogs, provision of neutrophil counts in excess of 5001 fll prevent severe recurtent infection and would be ther­ apeutic (Lothrop et al, 1988; Hammond et al, 1989; Morstyn and Dexter, 1993). W e observed no significant differences in lymphocyte and platelet numbers and hematocrit between animals treated with G-CSF-expressing cells and controls, indicating vector-encoded G-CSF production stimulated neutrophil production without other hematological effects. This is of interest because previ­ ous stodies in mice have shown that G-CSF administration caused reduced erythropoiesis and anemia (Molineux et al, 1990; Pojda et al, 1990). Because long-term administi-ation of recombinant G-CSF to humans (Morstyn et al, 1988; Morstyn and Dexter, 1993) and dogs (Hammond et al 1989, 1990) specifically stimulates neutrophil production, our data suggest that rats provide a more appropriate human model for the phys­ iological study of G-CSF administration than mice. G-CSF ex­ pression, unlike GM-CSF (Johnson et al, 1989) and n..-3 (Wong et al, 1989), did not cause pathological changes in hematopoiesis. In previous studies, these latter cytokines pro­ duced a fatal myeloproliferative syndrome in mice (Johnson et al, 1989; Wong et al, 1989). The implantation of genetically modified vascular smootii muscle cells to patients will requke an approach ottier than ar­ terial seeding. Recentiy, we proposed tiie seeding of synthetic PTFE (polytetrafluoroethylene) grafts as a method to retum transduced vascular smooth muscle cells to their donor (Geary et al, 1994) and this technique can be readily stodied in collie dogs with cyclic hematopoiesis, an appropriate clinical model for affected patients (Dale et al, 1988; Lothrop et al, 1988). W e have shown that the rat carotid artery seeding procedure re­ sulted in approximately 10' cells being retained within the artery (Osbome et al, 1995). On the basis of this fmding, we estimate that the potentially therapeutic level of neutrophils we observed in rats was derived from 10' transduced cells. W e pre­ viously estimated that 10* transduced cells can be seeded into a 10-cm X 4-mm PTFE graft (Osbome et al, 1993; Geary et al, 1994) and from the curtent study, this should provide a ther­ apeutic level of G-CSF to a dog. It is also probable that neu­ tropenic dogs and patients are more sensitive to G-CSF than normals. Our data show that transduced vascular smooth muscle cells do not inactivate retroviral vector sequences, in agreement with previous studies of retrovirally mediated gene expression in these target cells (Lynch et al, 1992; Clowes et al, 1994; Osbome et al, 1995). This is in contrast to skin fibroblasts where vector inactivation has been documented in both rats (Palmer et al, 1991) and dogs (Ramesh et al, 1993). Thus, data is accumulating to show that vascular smooth muscle cells pro­ vide an ideal target tissue for gene therapy. These cells are read­ ily obtained, cultored, transduced, and retumed to their donor. Implantation of these cells in the blood circulation suggests that their use for the secretion of not only hormones but also clot- Table 1. Control and Treated Rat Blood Counts LASN (« = 6) LrGSN (n = 8) '^< 0.001. bp > 0.05. Pre-surgery Post-surgery Pre-surgery Post-surgery Neutrophils per [ll 1,850 ± 390 1,870 ± 460" 1,880 ± 750 3,670 ± 740" Lymphocytes per p,/ 6,870 ±1,180 6,180 ±970^ 6,470 ±1,140 6,240 ± 1,300'' HCT 43.7 ± 2.84 45.0 ± 2.22'' 44.3 ± 2.38 43.9 ± 1.55'' Platelets per p./ (XIO^) 673 ± 167 716 ± 96'' 678 ± 122 704 ± 94'' G-CSF E X P R E S S I O N IN R A T S 1435 ting factors for the treatment of patients with hemophilia and enzymes for treatment of lysosomal storage disorders. ACKNOWLEDGMENTS We thank Dr. David Dale for many helpful discussions. This work was supported by grants D K 43727 and D K 47754 from the National Institotes of Health. REFERENCES BROWDER, R.M., ABRAMS, J.S., WONG, P.M.C, and NIENHUIS, A.W. (1989). Mechanism of autocrine stimulation in hematopoietic cells producing interleukin-3 after retrovirus-mediated gene transfer. Mol. Cell. Biol. 9, 204-213. CHANG, J.M., METCALF, D., GOUDA, T.J., and JOHNSON, G.R. (1989). Long-term exposure to retrovirally expressed granulocyte- colony-stimulating factor induces a non-neoplastic granulocytic and progemtor cell hyperplasia without tissue damage in mice. J. Clin. Invest. 84, 1488-1496. CLOWES, A.W., REIDY, M.A., and CLOWES, M.M. (1983). Kinetics of cellular proliferation after arterial injury: 1. Smooth muscle growth in the absence of endothelium. Lab Investigation 49, 327-333. CLOWES, M.M., LYNCH, CM., MILLER, A.D., MILLER, D.G., OS­ BORNE, W.R.A., and CLOWES, A.W. (1994). Long-tenn biologi­ cal response of injured rat carotid artery seeded with smooth muscle cells expressing retrovirally introduced human genes. J. Clin. Invest. 93, 644-651. DALE, D.C, and GRAW, R. (1974). Transplantation of allogeneic bone manow in canine cyclic neutropenia. Science 183, 83-84. DALE, D.C, and H A M M O N D , W.P. (1988). Cyclic neutropenia: A Clinical review. Blood Rev. 2, 178-185. DALE, D.C, W A R D , S.B., KIMBALL, H.R., and WOLFF, S.M. (1972). Studies on neutrophil production and tumover in grey collie dogs with cyclic neutropenia. J. Clin. Invest. 51, 2190-2196. DALE, D.C, LAU, S., NASH, R., BOONE, T., and OSBORNE, W.R. (1992). Effect of endotoxin on semm granulocyte and granulocyte- macrophage colony-stimulating factor levels in dogs. J. Infect. Dis. 165, 689-694. DEMETRI, G.D., and GRIFFIN, J.D. (1991). Granulocyte colony-stim­ ulating factor and its receptor. Blood 78, 2791-2808. GEARY, R.L., CLOWES, A.W., LAU, S., VERGEL, S., DALE, D.C, and OSBORNE, W.R.A. (1994). Gene transfer in baboons using prosthetic vascular grafts seeded with retrovirally-transduced smooth muscle cells: A model for local and systemic gene therapy. Hum. Gene Ther. 5, 1213-1218. H A M M O N D , W.P., PRICE, T.H., SOUZA, L.M., and DALE, D.C. (1989). Treatment of cyclic neutropenia with granulocyte colony- stimulating factor. N. Engl. J. Med. 320, 1306-1311. H A M M O N D , W.P., BOONE, T.C, DONAHUE, R.E., SOUZA, L.M., and DALE, D.C. (1990). A comparison of treatment of canine cyclic hematopoiesis with recombinant human G-CSF, GM-CSF, and IL-3 and canine G-CSF. Blood 76, 523-532. HAN, W.S., RAMESH, N., and OSBORNE, W.R.A. (1996). Cloning and expression of the cDNA encoding rat granulocyte colony-stim­ ulating factor. Gene (in press). HOCK, R.A., MILLER, A.D., and OSBORNE, W.R.A. (1989). Expression of human adenosine deaminase from various strong pro­ moters after gene transfer into human hematopoietic cell lines. Blood 74, 876-881. JOHNSON, G.R., GONDA, R.J., METCALF, D., HARIHARAN, I.K., and CORY, S. (1989). A letiial myeloproliferative syndrome in mice transplanted with bone manow cells infected with a retrovirus ex­ pressing granulocyte-macrophage colony-stimulating factor. E M B O J. 8, 441^48. JONES, J.B., and LANGE, R.D. (1983). Cyclic hematopoiesis: Animal models. Exp. Hematol. II, 571-580. KRANCE, R.A., SPRUCE, W.E., FORMAN, S.J., ROSEN, R.B., HECHT, T., H A M M O N D , W.P., and BLUME, K.G. (1982). Human cyclic neutropenia transfened by allogeneic bone manow grafting. Blood 60, 1263-1266. LAKER, C, STOCKING, C, BERGHOLZ, U., HESS, N., DELA- MARTER, J.F., and OSTERTAG, W. (1987). Autocrine stimula­ tion after transfer of the granulocyte-macrophage colony-stimulat­ ing factor gene and autonomous growth are distinct but interdependent steps in the oncogenic pathway. Proc. Natl. Acad. Sci. USA 84, 8458-8462. LANG, R.A., METCALF, D., GOUGH, N.M., DUNN, A.R., and GOUDA, R.J. (1985). Expression of a hematopoietic growth factor cDNA in a factor-dependent cell line results in autonomous growth and tamorigenicity. Cell 43, 531-542. LEE, M.Y., and LOTTSFELDT, J.L. (1984). Augmentation of neu­ trophilic granulocyte progenitors in the bone manow of mice with tamor-induced neutrophilia: Cytochemical stady of in vitro colonies. Blood 65, 499-506. LOTHROP, CD., WARREN, D.J., SOUZA, L.M., JONES, J.B., and MOORE, M.A.S. (1988). Conection of cyclic hematopoiesis with recombinant human granulocyte colony-stimulating factor. Blood 72, 1324-1328. LYNCH, CM., CLOWES, M.M., OSBORNE, W.R.A., CLOWES, A.W., and MILLER, A.D. (1992). Long-term expression of human adenosine deaminase in vascular smooth muscle cells of rats: A model for gene ttierapy. Proc. Nati. Acad. Sci. USA 89, 1138-1142. MILLER, A.D., and BUTTIMORE, C (1986). Redesign of retroviras packaging cell lines to avoid recombination to helper virus produc­ tion. Mol. Cell. Biol. 6, 2895-2902. MILLER, A.D., and ROSMAN, G.J. (1989). Improved retroviral vec­ tors for gene transfer and expression. BioTechniques 7, 980-990. MOLINEUX, G., POJDA, Z., and DEXTER, T.M. (1990). A compar­ ison of hematopoiesis in normal and splenectomized mice treated with granulocyte colony-stimulating factor. Blood 75, 563-569. MORSTYN, G., and DEXTER, T.M. (1993). Filgrastim (r-metHuG- CSF) in clinical practice. In Basic and Clinical Oncology. (Marcel Dekker, New York) p. 351. MORSTYN, G., SOUZA, L.M., KEECH, J., SHERIDAN, W., CAMPBELL, L., ALTON, N.K., GREEN, M., METCALF, D., and FOX, R. (1988). Effect of granulocyte colony stimulating factor on neutropenia induced by cytotoxic chemotherapy. Lancet 1, 667-672. OSBORNE, W.R.A., GEARY, R., LAU, S., DALE, D.C.D., and CLOWES, A.W. (1993). Transduced vascular smooth muscle cells in a canine model of gene therapy. Clin. Res. 41, 194A. OSBORNE, W.R.A., RAMESH, N., LAU, S., CLOWES, M.M., DALE, D.C, and CLOWES, A.W. (1995). Gene tiierapy for long- term expression of erythropoietin in rats. Proc. Nati. Acad. Sci. USA 92, 8055-8058. PALMER, T.D., ROSMAN, G.J., OSBORNE, W.R.A., and MILLER, A.D. (1991). Genetically-modified skin fibroblasts persist long after ttansplantation but gradually inactivate inttoduced genes. Proc. Nati. Acad. Sci. USA 88, 1330-1334. POJDA, Z., MOLINEUX, G., and DEXTER, T.M. (1990). Hematopoietic effects of short-term in vivo tteatment of mice witii various doses of rhG-CSF. Exp. Hematol. 18, 27-31. RAMESH, N., LAU, S., PALMER, T.D., STORB, R., and OSBORNE, W.R.A. (1993). High-level human adenosine deaminase expression in dog skin fibroblasts is not sustained following ttansplantation. Hum. Gene Ther. 4, 3-7. WEIDEN, P.L., ROBEMETT, B., GRAHAM, T.C, ADAMSON, J., and 1436 LEJNIEKS ET AL. STORB, R. (1974). Canine cyclic neutropenia. A stem cell defect. J. Clin. Invest. 35, 950-953. W O N G , P.M.C, C H U N G , S.-W., and NIENHUIS, A.W. (1987). Rettoviral transfer and expression of the interleukin-3 gene in he­ mopoietic cells. Genes & Dev. 1, 358-365. W O N G , P.M.C., CHUNG, S.-W., DUNBAR, C.E., BODINE, D.M., RUSCETTI, S., and NIENHUIS, A.W. (1989). Rettovirus-mediated transfer and expression of the interleukin-3 gene in mouse hematopoi­ etic cells results in a myeloprohferative disorder. Mol. Cell. Biol. 9, 798-808. Address reprint requests to: Dr. W.R.A. Osborne Department of Pediatrics, M S 356320 University of Washington Seattle, W A 98195 Received for publication Feb. 5, April 17, 1996. 1996; accepted after revision