Optimization of Gene Transfer into Primitive Human Hematopoietic ...

2 downloads 0 Views 275KB Size Report
Primitive human hematopoietic cells in granulocyte-colony stimulating factor ... or a defined medium that had been supplemented with factors identified in ...
HUMAN GENE THERAPY 13:1317–1330 (July 20, 2002) © Mary Ann Liebert, Inc.

Optimization of Gene Transfer into Primitive Human Hematopoietic Cells of Granulocyte-Colony Stimulating Factor–Mobilized Peripheral Blood Using Low-Dose Cytokines and Comparison of a Gibbon Ape Leukemia Virus Versus an RD114-Pseudotyped Retroviral Vector JOHANNES C.M. VAN DER LOO,1 B.L. LIU,1 A.I. GOLDMAN, 2 S.M. BUCKLEY,1 and K.S. CHRUDIMSKY 1

ABSTRACT Primitive human hematopoietic cells in granulocyte-colony stimulating factor (G-CSF)–mobilized peripheral blood (MPB) are more difficult to transduce compared to cells from umbilical cord blood. Based on the hypothesis that MPB cells may require different stimulation for efficient retroviral infection, we compared several culture conditions known to induce cycling of primitive hematopoietic cells. MPB-derived CD341 cells were stimulated in the presence or absence of the murine fetal liver cell line AFT024 in trans-wells with GCSF, stem cell factor (SCF), and thrombopoietin (TPO) (G/S/T; 100 ng/ml) or Flt3-L, SCF, interleukin (IL)7, and TPO (F/S/7/T; 10–20 ng/ml), and transduced using a GaLV-pseudotyped retroviral vector expressing the enhanced green fluorescence protein (eGFP). Compared to cultures without stroma, the presence of AFT024 increased the number of transduced colony-forming cells (CFC) by 3.5-fold (with G/S/T), long-term culture-initiating cells (LTC-IC) by 4.6-fold (with F/S/7/T), and nonobese diabetic/severe immunodeficiency disease (NOD/SCID)-repopulating cells (SRC) by 6.8-fold (with F/S/7/T). Similar numbers of long-term culture-initiating cells (LTC-IC) and SRC could be transduced using AFT024-conditioned medium (AFT-CM) or a defined medium that had been supplemented with factors identified in AFT-CM. Finally, using our best condition based on transduction with the gibbon ape leukemia virus (GaLV)-pseudotyped vector, we demonstrate a 33-fold higher level of gene transfer (p , 0.001) in SRC using an RD114-pseudotyped vector. In summary, using an optimized protocol with low doses of cytokines, and transduction with an RD114 compared to a GaLV-pseudotyped retroviral vector, the overall number of transduced cells in NOD/SCID mice could be improved 144-fold, with a gene-transfer efficiency in SRC of 16.3% (13.3–19.9; n 5 6). OVERVIEW SUMMARY Primitive human hematopoietic cells in MPB are more difficult to transduce compared to their counterparts in umbilical cord blood. Previously, we have found that the level of gene transfer in these cells can be enhanced using extended prestimulation. Based on the observation by others that transplantable hematopoietic cells can be preserved in a low-dose cytokine environment in coculture with the cell line AFT024, or in a stroma-free medium formulated to mimic these conditions, we tested whether or not these con-

ditions could be used to support retroviral gene transfer in MPB. In a step-wise optimization, we demonstrate that these culture conditions significantly improve the survival of primitive hematopoietic cells in MPB and support oncoretroviral gene transfer. In addition, using an optimized stroma-free low-dose cytokine medium, we demonstrate that the number of gene modified cells in nonobese diabetic/severe immunodeficiency disease mice posttransplantation can be improved 33-fold using an RD114 compared to a gibbon ape leukemia virus (GaLV)-pseudotyped retrovirus vector.

1University of Minnesota Stem Cell Institute, Cancer Center, Division of Hematology, Oncology and Transplantation, Department of Medicine, 2 Department of Biostatistics and Cancer Center, University of Minnesota, Minneapolis, MN 55455.

1317

1318

VAN DER LOO ET AL.

INTRODUCTION

H

UMAN GRANULOC YTE -COLONY STIMULATIN G FACTOR

(GCSF)–mobilized peripheral blood (MPB) is an attractive target for gene transfer because of the ease with which the cells can be harvested and the accelerated rate of engraftment in allogeneic transplantation (Champlin et al., 2000; Bensinger et al., 2001). However, the clinical application of retroviral gene therapy using MPB has been hampered by relatively low transduction efficiencies. In the past 5 years, several modifications in gene-transfer technology have led to improved levels of gene transfer in primitive hematopoietic cells (Williams and Smith, 2000). These modifications include the use of alternate retroviral envelopes, such as the gibbon ape leukemia virus (GaLV) envelope (Kiem et al., 1997) or feline endogenous retrovirus (RD114) envelope (Kelly et al., 2000; Gatlin et al., 2001), the use of early acting cytokines that improve the survival of primitive hematopoietic cells (Petzer et al., 1996; Zandstra et al., 1997; Kiem et al., 1998), and the use of methods that improve contact between retroviral particles and target cell, such as infection on extracellular matrix protein fibronectin (Moritz et al., 1994; Hanenberg et al., 1996). Recently, using fibronectin-assisted gene transfer, Abonour et al. (2000) demonstrated transduction of human long-term repopulating hematopoietic stem cells derived from autologous MPB. At 1 year posttransplantation, 6 of 7 patients evaluated had evidence of gene transfer in bone marrow progenitor cells, demonstrating that the human long-term repopulating stem cell in MPB can be transduced using a murine oncoretrovirus. In a preclinical setting, several investigators have demonstrated efficient gene transfer into primitive hematopoietic cells from the umbilical cord blood (UCB) (Conneally et al., 1998; Marandin et al., 1998; van Hennik et al., 1998; Hennemann et al., 1999; Novelli et al., 1999; Sanyal and Schuening, 1999; Barquinero et al., 2000; Kelly et al., 2000). In these studies, the level of gene transfer was determined by the presence of gene-modified cells in nonobese diabetic/severe combined immunodeficiency disease (NOD/SCID) mice at several weeks posttransplantation. Compared to UCB, gene transfer into primitive hematopoietic cells of MPB in the same transplant model has been relatively inefficient (Pollok et al., 2001). Based on the quiescent nature of progenitor cells in MPB (Roberts and Metcalf, 1995), and the fact that mitosis of the target cell is a prerequisite for retroviral integration (Miller et al., 1990), we hypothesized that primitive cells in MPB compared to UCB may be more resistant to gene transfer as these cells may be more difficult to stimulate. This hypothesis is supported by a recent study that demonstrates uniformly inferior in vitro expansion of primitive hematopoietic cells in MPB compared to UCB (Lewis and Verfaillie, 2000). Also, this is consistent with our earlier observation that gene transfer into NOD/SCID-repopulating cells (SRC) in MPB, compared to a similar population derived from UCB, could be improved using a longer period of cytokine prestimulation (Pollok et al., 2001). Ultimately, to improve the level of gene transfer in MPB cells, we set out to evaluate culture conditions for prestimulation that have been found to preserve primitive hematopoietic cells in vitro as described below. The murine fetal liver cell line AFT024 has been described to have a positive effect on the maintenance of murine hema-

topoietic stem cells (Moore et al., 1997a). Moreover, in direct contact with human hematopoietic cells, or in noncontact transwells supplemented with early acting cytokines, AFT024 has been found to support the expansion of human long-term culture-initiating cells (LTC-IC), natural killer culture-initiating cells (NK-IC), and maintenance of cells capable of engrafting NOD/SCID mice and fetal sheep, in primary as well as secondary recipients (Thiemann et al., 1998; Lewis and Verfaillie, 2000; Lewis et al., 2001; Nolta et al., 2002). In AFT024contact and noncontact cultures, comparison of various combinations of early acting cytokines showed superior expansion of LTC-IC and NK-IC from MPB with low concentrations (10–20 ng/ml) of fetal liver tyrosine kinase 3-ligand (Flt3-L), stem cell factor (SCF), interleukin 7 (IL-7), and thrombopoietin (TPO) (Lewis and Verfaillie, 2000). As primitive hematopoietic cells were induced to cycle (Lewis and Verfaillie, 2000), and stem cell content was preserved when cultured under these conditions (Lewis et al., 2001), we hypothesized that similar conditions may be used to enhance the level of gene transfer in MPB. Here, we compared the effect of AFT024 and two combinations of cytokines on the gene-transfer efficiency in CFC, LTCIC and SRC derived from MPB. CD341 cells were cultured for 4 days in trans-wells in the presence or absence of AFT024 using medium supplemented with 100 ng/ml of G-CSF, SCF, and TPO (G/S/T), which we have used previously (Pollok et al., 2001). Alternatively, cells were cultured for 4 days in the presence of low doses (10–20 ng/ml) of Flt3-L, SCF, IL-7 and TPO (F/S/7/T), as described above. This was followed by infection with a GaLV-pseudotyped retroviral vector. Furthermore, we examined whether AFT024 could be replaced by a stroma-free medium without loss of transduced LTC-IC. In these experiments we used AFT-conditioned medium (AFT-CM) or a defined medium (MV8) that had been supplemented with factors identified in hematopoiesis-supporting stromal cultures, as previously described (Punzel et al., 1999; Lewis et al., 2001). Finally, using our optimal low-dose cytokine culture condition for prestimulation, we compared the efficiency of gene transfer into SRC using the GaLV-pseudotyped vector with a vector pseudotyped with the RD114-envelope protein. Addition of the RD114-pseudotyped vector to the study was based on recent reports of high levels of gene transfer in SRC from UCB (Kelly et al., 2000; Gatlin et al., 2001).

MATERIALS AND METHODS Animals A breeding colony of NOD.CB17-Prkdcscid (NOD/SCID) mice (Shultz et al., 1995) was established at the Research Animal Resources facility at the University of Minnesota using mice obtained from Jackson Laboratories (Bar Harbor, ME). Animals were kept under specific pathogen-free conditions in filter-top sterilized cages and maintained on autoclaved acidified water and sterilized Teklad mouse diet. Every other week, the drinking water was supplemented with 1 g/L sulfametoxazole and 200 mg/L trimethoprim (Barre-National Inc., Baltimore, MD). Animals were weaned at 4–5 weeks of age and used for transplantation at 6–8 weeks. Animal experiments had

1319

TRANSDUCTION OF HUMAN MPB CELLS been approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Minnesota.

Human cells MPBs were obtained from healthy adult volunteers. Donors received G-CSF (Neupogen, Amgen Inc. Thousand Oaks, CA) for 5 days at 10 mg/kg per day by subcutaneous injection. Nucleated cells were harvested by leukapheresis and CD341 cells were selected using the CliniMACS (Miltenyi Biotec, Auburn, CA) according to the manufacturer’s instructions at the Cell Therapy Core Facility of the University of Minnesota Cancer Center. The average purity of CD341 cells was 79.6% 6 15.9% (mean 6 standard deviation [SD], n 5 14). CD341 cells were stored until use in liquid N2 after controlled-rate freezing. The protocol was approved by the Institutional Review Board: Human Subjects Committee (IRB) of the University of Minnesota.

AFT024 stromal cells The murine fetal liver cell line AFT024 (kindly provided by Drs. I. Lemischka and K. Moore, Princeton University, Princeton, NJ) (Moore et al., 1997a) was maintained at 33°C in Dulbecco’s Modified Eagle Medium (DMEM; Gibco-BRL, Grand Island, NY) supplemented with 20% fetal bovine serum (FBS; Hyclone Laboratories, Logan, UT), 1024 mol/L 2-mercaptoethanol (2-ME; Sigma, St Louis, MO), 100 U/ml penicillin and 100 mg/ml streptomycin (Pen/Strep; stock solution 5000 U/ml penicillin and 5000 mg/ml streptomycin; Gibco-BRL). Cells were grown in tissue culture plates coated with 0.1% gelatin (Sigma) for 30 min at 37°C. When used as a feeder layer, AFT024 cells were grown to confluency and irradiated at 20 Gy using a Mark I 137Cesium irradiator (Shepard, Glendale, CA).

MFG-eGFP retroviral vectors The MFG-eGFP retroviral vector, which was kindly provided by Dr. G. Wagemaker, Erasmus University, Rotterdam, The Netherlands (Bierhuizen et al., 1997), had been pseudotyped with the gibbon-ape leukemia virus (GaLV) envelope (MFGeGFP clone 5) using PG13 packaging cells as previously described (Pollok et al., 1999). The MFG-eGFP retroviral vector had also been pseudotyped with the feline endogenous virus envelope protein RD114, a generous gift of Drs. K.E. Pollok (Indiana University, Indianapolis, IN) and P.F. Kelly (Children’s Hospital Research Center, Cincinnati, OH), using the FLYRD18 (HT1080) packaging cell line (Cosset et al., 1995). FLYRD18 cells had kindly been provided by Dr. Y. Takeuchi (University College London, London, England). Virus supernatants (SN) were collected and stored at 280°C until use.

Titering of supernatants Virus titers were determined based on eGFP expression in human erythroleukemia (HEL) 92.1.7 cells (ATCC No. TIB180). Briefly, HEL cells were seeded at 2 3 104 cells per well. Cells were transduced for 4 hr with 2 ml serial dilutions of virus SN in the presence of 8 mg/ml Polybrene (Aldrich Chemical Co., Milwaukee, WI), and analyzed by fluorescent-activated cell sorter (FACS) for eGFP-expression 48–96 hr later. Titers (number of infectious units per milliliter [IU/ml]) were calculated based on the number of cells at the time of infection, the dilution, and percent eGFP1 cells. GaLV and RD114-pseudotyped MFG-eGFP had a titer of 3 3 103 and 1 3 104 IU/ml, respectively. Real-time reverse transcriptase-polymerase chain reaction (RT-PCR) indicated a 4.4-fold difference in the relative level of retrovirus RNA in GaLV and RD114-SN that paralleled differences in viral titer.

AFT024-conditioned medium and MV8 AFT024 cells were grown to confluency in 75-cm2 flasks and irradiated as described above. After irradiation, the medium was replaced with 20 ml of RPMI 1640 (Gibco-BRL) per flask supplemented with 20% FBS, 2% Pen/Strep, 1025 mol/L 2-ME, and 10 ng/ml Flt3-L (Immunex, Seattle, WA), 10 ng/ml SCF (a generous gift from Amgen, Thousand Oaks, CA), 20 ng/ml IL-7 (R&D Systems, Minneapolis, MN) and 20 ng/ml TPO (a generous gift from the Kirin Brewery Co., Tokyo, Japan). After 48 hr at 33°C, AFT-CM was collected, filtered (0.45 mm, Sterivex-HV; Millipore, Bedford, MA), and stored at 280°C until use. MV8 medium consisted of Iscove’s Modified Dulbecco’s Medium (IMDM) (Gibco-BRL) supplemented with 20% FBS, 2% Pen/Strep, 2 mmol/L L -glutamine (Gibco-BRL), 1025 mol/L 2-ME, and 10 mg/ml Ndesulfated O-sulfated heparin (NDSNAc-Heparin; Seikagaku America, Falmouth, MA), as previously described (Lewis et al., 2001). In addition, MV8 contained 250 pg/ml G-CSF (Amgen), 200 pg/ml macrophage inflammatory protein-1a (Mip-1a; R&D Systems), 10 ng/ml monocyte chemoattractant protein-1 (MCP-1; R&D Systems), 10 ng/ml vascular endothelial cell growth factor (VEGF; R&D Systems), 2 ng/ml interleukin 8 (IL-8; R&D Systems), and 1 ng/ml IL-6 (R&D Systems). Finally, MV8 medium was supplemented with 20 ng/ml of Flt3-L, SCF and IL-7, and 10 ng/ml of TPO, aliquotted, and stored at 280°C until use.

Binding of virus to fibronectin To enhance the efficiency of gene transfer, the CD341 cells used in this study were infected in the presence of RetroNectin™, a recombinant fibronectin (FN) peptide (CH296) generously provided by Takara Shuzo Ltd., Otsu, Japan, as previously described (Hanenberg et al., 1996; Pollok et al., 2001). The number of infectious units present under these conditions was determined by endpoint dilution of virus SN on HEL cells in the presence of FN. The procedures used here were similar to the ones used for infection of CD341 cells described in this study. Briefly, for the GaLV-pseudotyped vector, HEL cells were transduced in FN-coated (4 mg/cm2) 24-well tissue culture plates (2 3 105 cells per well) with serial dilutions of virus SN (0.5 ml per well) for 4 hr on 2 subsequent days. For the RD114-pseudotyped vector, FN-coated plates were preincubated twice for 30 min with 0.5 ml of serial dilutions of RD114-SN, washed, and incubated with HEL cells overnight. The percent transduced cells was determined by FACScan 48–96 h later. GaLV and RD114-pseudotyped MFG-eGFP, when used in the presence of FN, had an effective titer of 6 3 104 and 5 3 105 IU/ml, respectively. The ratio in titer between RD114 and GaLV-SN as measured here (8.3-fold), was 2.5fold greater compared to the ratio in titer as measured without FN in the presence of polybrene (3.3-fold; see above), indicating that the RD114 vector bound FN with greater efficiency.

1320

CD341 cell pre-stimulation Before retroviral infection, CD341 cells were prestimulated for 4 days using one of several culture conditions. In the first series of experiments, cells were cultured in collagen-coated trans-wells (0.4 mm Transwell-COL; Corning, Acton, MA) inserted into 6-well plates, with or without a confluent layer of previously established and irradiated AFT024 in the bottom compartment. The culture medium consisted of IMDM, 10% FBS, 2% Pen/Strep, 2 mmol/L L -glutamine, supplemented with either G-CSF, SCF, and TPO (G/S/T; all at 100 ng/ml), as described (Pollok et al., 2001), or with F/S/7/T (20 ng/ml for F/S, and 10 ng/ml for 7/T). In subsequent experiments, cells were cultured for 4 days with or without AFT024 in the presence of F/S/7/T in trans-wells as described above, or in AFT-CM or MV8 supplemented with F/S/7/T in regular tissue culture plates.

Transduction of CD341 cells with GaLV and RD114-pseudotyped MFG-eGFP Cells were pre-stimulated as described above, and subjected to two 4-hour infections with GaLV-pseudotyped MFG-eGFP retrovirus SN on two subsequent days. Specifically, cells were collected, resuspended in retroviral SN, and transferred to transwells or tissue culture plates pre-coated (2 hours at room temperature) with FN (4 mg/cm2). In all experiments, virus SN was supplemented with the same cytokines used for pre-stimulation. Also, cells pre-stimulated with AFT024 were infected in the presence of AFT024. Between rounds of infections, cells were kept on FN, but virus SN was replaced by pre-stimulation medium. After the second infection, cells were collected, washed and used for assays. For transduction with the RD114pseudotyped vector, FN-coated tissue culture plates were preloaded with vector particles by a two-time incubation with virus SN for 30 min at room temperature as previously described (Kelly et al., 2000). Plates were then washed with tissue culture medium and CD341 cells were added for overnight incubation at 37°C in prestimulation medium with cytokines. The next day, cells were collected, washed, and used for assays.

Colony-forming cell assay Noncultured (day 0) and transduced CD341 cells were tested for the presence of colony-forming cells (CFC) in methylcellulose cultures (200–600 cells per well, three replicate wells per group). Medium consisted of IMDM containing 1.2% (wt/vol) methylcellulose (Fisher, Fair Lawn, NJ), 30% FBS, 2 mmol/L L -glutamine, 5 3 1025 mol/L 2-ME, 2% Pen/Strep, 4 U/ml human erythropoietin (Amgen), 20 ng/ml human granulocytemacrophage colony-stimulating factor (GM-CSF; Immunex), 20 ng/ml human SCF, and 20 ng/ml human IL-3 (R&D Systems). After 14 days of culture in a humidified atmosphere at 37°C and 5% CO2, plates were evaluated for the presence of transduced (eGFP1) or nontransduced (eGFP2) hematopoietic colonies of 50 cells or more using an inverted fluorescence microscope (Olympus, Melville, NY).

LTC-IC LTC-IC frequency in noncultured (day 0) and transduced CD341 cells was determined by limiting dilution analysis on a preestablished and 20 Gy-irradiated AFT024 feeder layer in 96well tissue culture plates, as previously described (Lewis et al.,

VAN DER LOO ET AL. 2001). Briefly, cells (90,000 at day 0, or the equivalent proportion of all cells posttransduction) were plated in 8 dilutions (2700 cells per well, highest concentration), threefold apart, using 22 wells per concentration in IMDM supplemented with 12.5% FBS, 12.5% horse serum (Stem Cell Technologies, Vancouver, BC, Canada), 2 mmol/L L -glutamine, 1 mmol/L hydrocortisone (Pharmacia & Upjohn, Kalamazoo, MI), and 2% Pen/Strep. Cultures were maintained at 37°C, 5% CO2 for 5 weeks using weekly half-medium changes. After 5 weeks, tissue culture medium in each well was replaced by methylcellulose medium (formulated as described above) and, after 14 days, wells were evaluated for the presence or absence of transduced (eGFP1) and nontransduced (eGFP2) colonies using an inverted fluorescence microscope.

Transplantation of human cells into NOD/SCID mice NOD/SCID mice (6–8 weeks old) received 300 cGy of total body irradiation (56 cGy/min), and were transplanted within 3–4 hours with CD341 cells in 0.3–0.4 ml Hanks’ balanced salt solution, 25 mmol/L HEPES, 10 U/ml heparin, and 0.01% bovine serum albumin (BSA, fraction V; Roche Diagnostics Corporation, Indianapolis, IN) by lateral tail vein injection. After 9–19 weeks, animals were killed, bone marrow was collected, and cells were analyzed for the presence of human cells and expression of eGFP by flow cytometry, as previously described (Pollok et al., 2001). Mouse anti-human CD45-PerCP, CD33-PE and mouse IgG1-PerCP were purchased from BD Biosciences (San Jose, CA), mouse IgG1-PE and CD19-PE were purchased from Caltag (Burlingame, CA). Cells were counterstained with propidium iodide (0.5 mg/ml) to allow for exclusion of dead cells. The absolute number of eGFP1 cells in the bone marrow was calculated based on a recovery of 25 3 106 cells per 2 femurs (averaged from n 5 47), which represents 11.8% of the marrow (Chervenick et al., 1968).

Statistics The relative levels of gene transfer, the number of CFC and LTC-IC, and the number of transduced CFC and LTC-IC in pooled experiments were analyzed by ANOVA using a randomized block design. Variation between individual MPB products was taken into account by defining each experiment as a block (Proc ANOVA/SAS version 6.12, SAS Institute Inc., Cary, NC). Within each experiment, values were analyzed relative to a control group, as indicated in the results. All analyses were performed after transformation of the data to a logarithmic scale. The upper and lower limits of error were computed from the mean differences 6 standard error (SE). Other analyses were done by two-sided paired or nonpaired Student’s t test, depending on the experimental design. Limiting dilution analyses were performed using maximum likelihood estimates from a Poisson distribution (Fazekas de StGroth, 1982). All inferences were made at a 5% level of significance.

RESULTS Effect of pre-stimulation on gene-transfer efficiency in CFU-C MPB-derived CD341 cells were prestimulated for 4 days in IMDM supplemented with either G/S/T (all at 100 ng/ml), or

TRANSDUCTION OF HUMAN MPB CELLS with F/S/7/T (F/S at 20 ng/ml, 7/T at 10 ng/ml), in the presence or absence of irradiated AFT024 in trans-wells. After prestimulation, cells from each group were transferred to FN-coated trans-wells and transduced with the GaLV-pseudotyped MFGeGFP retroviral vector. Statistical analysis indicated that the highest variability in gene transfer could be attributed to differences between MPB products (p , 0.0001, n 5 10, not shown). To control for this variability, data were expressed relative to a control reference group (G/S/T without AFT024) within each experiment. The other three groups were compared to the control using statistical modeling as described in the methods. The percentage of transduced CFC was lower when cells were prestimulated with F/S/7/T (Fig. 1A; F/S/7/T without AFT024 versus control; p , 0.05). When analyzed irrespective of the presence of AFT024, the effect of F/S/7/T was found to be highly significant (p , 0.01). Conversely, when analyzed independently of the choice of cytokines, AFT024 did not affect the percentage of transduced CFC (p 5 0.32). In contrast, AFT024 did affect the number of CFC (Fig. 1B; p 5 0.005). Compared to control, the number of CFC increased by 2.6-fold (1.8–3.6) in the presence of G/S/T (p , 0.05), and by 2.2-fold (1.5–3.1) in the presence of F/S/7/T (p , 0.05). Similarly, the number of CD341 cells increased in the presence of AFT024 irrespective of the cytokines (Table 1; p , 0.05). Finally, the absolute number of transduced CFC (Fig. 1C) was higher when cells were prestimulated and transduced with G/S/T in the presence of AFT024 (3.5-fold [2.1–5.8] compared to control, p , 0.05).

Effect of prestimulation on gene-transfer efficiency into LTC-IC CD341 cells, prestimulated and transduced as described above, were subsequently tested in the LTC-IC assay. Again, data were expressed relative to a control reference group within each experiment. Compared to control, the percentage of trans-

1321 duced LTC-IC was 2.5-fold higher (p , 0.05) in the presence of AFT024, but only in the group stimulated with the cytokines F/S/7/T (Fig. 2A; F/S/7/T with AFT024 versus F/S/7/T without AFT024). Furthermore, irrespective of the cytokine combination, AFT024 supported the survival of LTC-IC (Fig. 2B; p , 0.001). Compared to control, the number of LTC-IC was 2.2-fold (1.7–2.9) higher with G/S/T (p , 0.05) and 2.1-fold (1.6–2.7) higher with F/S/7/T (p , 0.05). Finally, with F/S/7/T in the presence of AFT024, the absolute number of transduced LTC-IC was 4.6-fold (3.2–6.5) higher compared to control (Fig. 2C; p , 0.01). In summary, the data show that the level of gene transfer in LTC-IC can be enhanced by low doses of F/S/7/T in the presence of AFT024. Statistical analysis confirmed that this effect could not be attributed to cytokines alone (p 5 0.56) or AFT024 alone (p 5 0.08), but was based on interaction between AFT024 cells and cytokines. As these data indicate an improved level of gene transfer in LTC-IC, we next set out to test whether gene transfer in transplantable hematopoietic cells could be improved using similar culture conditions.

Positive effect of AFT024 on the transduction of SRC To test the effect of AFT024 on the transduction of transplantable hematopoietic cells, CD341 cells were prestimulated with F/S/7/T in the presence or absence of AFT024, and transplanted into NOD/SCID mice (at 2.5 3 106 CD341 cells per mouse as counted before culture). After 13–19 weeks, bone marrow was analyzed for the presence of human cells (CD451 ) and for the absolute number of transduced cells. The level of engraftment was 2.5-fold (1.9–3.2) higher (p 5 0.03) in recipients of cells that had been cultured with AFT024 compared to cells cultured without AFT024 (Fig. 3A). In addition, the percent gene transfer among CD451 cells was 2.8-fold (2.1–3.7) higher (p 5 0.028) in the group cultured with AFT024 (2.4%, 1.8–3.2) compared to the group cultured without AFT024 (0.9%, 0.7–1.2; Fig. 3B). Finally, the absolute number of

FIG. 1. Effect of prestimulation on the transduction of MPB-derived CFC. MPB-derived CD341 cells (2 3 105 cells per group) were cultured for 4 days in the presence or absence of a confluent layer of AFT024 (20 Gy irradiated) in 0.4 mm trans-wells in medium supplemented with either G-CSF, SCF and TPO (G/S/T; all at 100 ng/ml) or Flt3-L, SCF, IL-7 and TPO (F/S/7/T; F/S at 20 ng/ml, at 7/T 10 ng/ml). Cells were then transferred to FN-coated trans-wells and transduced for 4 hr on 2 consecutive days with the (GaLV) MFG-eGFP vector (4 ml SN per well). Transduced and noncultured cells were tested in the CFC assay. All data have been expressed relative to a control reference group (G/S/T without AFT024). The reference level is indicated by a dotted line. A: Fold-gene transfer into CFC. B: Fold-number of CFC. C: Fold-number of transduced CFC. Bars indicate mean difference from control 6 SE of 10 independent experiments (each in triplicate). Data were analyzed using a randomized block design analysis of variance (ANOVA; see Methods). p values for overall comparison with control have been indicated in the left top corner of each graph. *p , 0.05 for individual groups compred to control. #p , 0.05 for comparison with group 3 (G/S/T with AFT024).

1322

b n,

a Experiment

2AFT024 G/S/T (TW) 2AFT024 F/S/7/T (TW) 1AFT024 G/S/T (TW) 1AFT024 F/S/7/T (TW) AFT024 (TW) AFT-CM AFT-CM MV8

Condition 2.0 2.0 2.0 2.0 2.0 4.0 4.0 4.0 85 6 4 (n 5 7)

88 6 2 (n 5 5)

81 6 6 (n 5 6)b

Input % CD341 (n)

PRESTIMULATION CONDITIONS

3.2 6 0.5 2.4 6 0.4 5.0 6 0.7 3.5 6 0.8 3.2 6 0.7 5.5 6 1.2 8.1 6 1.5 8.1 6 0.9

AND

1.7 6 0.3 1.3 6 0.2 2.8 6 0.3 1.9 6 0.4 1.6 6 0.3 1.4 6 0.3 2.0 6 0.4 2.0 6 0.2

Fold-change in total cell number

TOTAL CELL NUMBERS

Output cells/well (3105)

ON

48 6 70 48 6 10 58 6 80 60 6 90 70 6 90 77 6 10 74 6 40 60 6 70

Output % CD341

CD341 CELLS

1.2 6 0.2 0.8 6 0.1 2.7 6 0.7 2.0 6 0.7 2.1 6 0.5 4.2 6 0.9 6.4 6 1.4 5.1 6 1.0

Output CD341 cells/well (3105)

0.8 6 0.2c 0.5 6 0.1c 1.8 6 0.5 1.4 6 0.6 1.2 6 0.2 1.2 6 0.2 1.8 6 0.3 1.5 6 0.3

Fold-change in number CD341 cells

1 contains cell counts for Figures 1 and 2, Experiments 2 and 3 contain cell counts for Table 2, respectively. number of independent experiments analyzed. c indicates p , 0.05 (paired t test, n 5 6), as compared to equivalent condition with (1) AFT024. Human G-CSF–mobilized peripheral blood CD341 cells were cultured for 4 days under various conditions and transduced with the GaLV-pseudotyped MFG-eGFP retroviral vector on day 5 and 6. In Experiment 1, cells were cultured in the presence or absence of (20 Gy-irradiated) AFT024 cells in noncontact 0.4 mm trans-wells (TW) in medium supplemented with either G-CSF, SCF, and TPO (G/S/T), or Flt3-L, SCF, IL-7 and TPO (F/S/7/T). In Experient 2, cells were cultured in trans-wells (TW) with AFT024 cells or in regular 6-well plates with AFT024-conditioned medium (AFT-CM), both in the presence of the cytokines F/S/7/T. In Experiment 3, cells were cultured in AFT-CM or MV8 with F/S/7/T. After transduction, the number of cells was enumerated and the percent CD341 determined by flow cytometric analysis. The recovery of CD341 cells cultured without AFT024 cells, compared to all other groups, was significantly lower.

3

2

1a

Experiment

OF

Input cells/well (3105)

TABLE 1. EFFECT

1323

TRANSDUCTION OF HUMAN MPB CELLS

FIG. 2. Effect of prestimulation on the transduction of MPB-derived LTC-IC. MPB CD341 cells had been prestimulated and transduced as described in Figure 1. Frequencies of transduced and nontransduced LTC-IC were determined in limiting dilution assays. All data have been expressed relative to a control reference group (G/S/T without AFT024). A: Fold-gene transfer into LTC-IC. B: Fold-number of LTC-IC. C: Fold-number of transduced LTC-IC. Data were analyzed using a randomized block design analysis of variance (ANOVA; see Methods). Bars indicate mean difference from control 6 SE of 6 independent experiments. p values for overall comparison with control have been indicated. *p , 0.05 for individual groups compared to control. # p , 0.05 for group 4 (F/S/7/T plus AFT024) compared to group 2 (F/S/7/T minus AFT024). eGFP1 cells recovered per animal in the group cultured with AFT024 was 6.8-fold (5.6–8.4) higher (p , 0.0001; 18.3 versus 2.7 per 50,000 cells) compared to the group cultured without AFT024 (Fig. 3C). When expressed per million CD341 cells injected (initial count before culture), 3.11 3 104 transduced cells could be detected in the (whole) marrow in recipients of cells prestimulated with AFT024, compared to 0.46 3 104 in recipients of cells prestimulated without AFT024. In conclusion, the data show that AFT024 positively affects the survival of transplantable primitive hematopoietic cells, as well as the level of gene transfer, when prestimulated and transduced in the presence of F/S/7/T.

AFT-CM and MV8 can replace AFT024 Next, we tested whether AFT024 could be replaced by AFTCM, or by MV8 that had been supplemented with factors identified in AFT-CM as previously described (Punzel et al., 1999; Lewis et al., 2001). CD341 cells were cultured and transduced in trans-wells in the presence of AFT024 or in 6-well plates with AFT-CM, each supplemented with F/S/7/T. No difference was found in the absolute number of transduced LTC-IC between these groups (p 5 0.44; Table 2). The percentage of transduced LTC-IC after prestimulation in AFT-CM (13.7%, 11.6–16.3) was lower (p , 0.001) compared to LTC-IC pre-

FIG. 3. Effect of AFT024 on engraftment and level of gene transfer in NOD/SCID mice. CD341 cells (10 3 106 CD341 cells per group) were prestimulated for 4 days and transduced with the (GaLV) MFG-eGFP retroviral vector in medium containing F/S/7/T, with (1AFT) or without (2AFT) irradiated AFT024 in trans-wells. Cells (13 3 106 and 19 3 106 in 2AFT and 1AFT groups, respectively, as counted after prestimulation) were transduced in 6 wells per group. Transduced cells were collected and injected into sublethally irradiated NOD/SCID mice (at 2.5 3 106 cells per mouse, cell count before culture). A: Percent human (CD451 cells) among nucleated cells in the bone marrow at 13–19 weeks posttransplantation (data pooled from 2 individual experiments, n 5 7 per group). For each animal, background staining with an isotype control antibody had been subtracted. B: Percent eGFP1 cells among human cells. C: Absolute number of transduced (eGFP1 ) cells among 50,000 nucleated cells analyzed. Horizontal lines indicate the average per group. p values for comparison between groups (derived from a two-sample t test) have been indicated.

1324

VAN DER LOO ET AL. TABLE 2. EFFECT

OF

PRESTIMULATION

G ENE -TRANSFER EFFICIENCY, EXPANSION , TRANSDUCED LTC-IC

ON THE OF

Group

na

Condition

% Gene Transferb in LTC-IC (range)

A

5

B

7

AFT024 (TW) AFT-CM AFT-CM MV8

28.1 13.7 7.1 12.2

(23.4–33.6) (11.6–16.3)** (5.4–9.3) (8.1–18.4)

AND

% Recoveryb of LTC-IC (range) 18.9 32.1 47.4 28.2

(17.1–20.9) (26.2–39.3)* (38.9–57.8) (20.4–39.2)#

RECOVERY % Recovery of b eGFP1 LTC-IC (range) 5.3 4.4 3.4 3.4

(4.5–6.3) (3.7–5.3) (2.5–4.5) (2.4–5.0)

an,

number of independent experiments. indicate mean (range). Range represents the mean 6 SE. Human G-CSF-mobilized peripheral blood CD341 cells were cultured for 4 days under various conditions and transduced with the GaLV-pseudotyped MFG-eGFP retroviral vector on day 5 and 6. At day 0, and after transduction, cells were cultured in the LTC-IC assay. Transduced and nontransduced LTC-IC were enumerated by fluorescence microscopy. In group A, cells were cultured with AFT024 cells in trans-wells (TW) or in regular 6-well plates with AFT024-conditioned medium (AFT-CM) in the presence of Flt3-L, SCF, IL-7 and TPO (4 3 104 cells/cm2, 5 experiments). In group B, cells were cultured in AFT-CM or MV8 in 6-well plates with Flt3-L, SCF, IL-7 and TPO (4 3 104 cells/cm2, 7 experiments). Averages were calculated and statistical analyses (paired t test) were performed after transformation of the data to a logarithmic scale, as described in the methods. In group A, *p , 0.05, **p , 0.001 for AFT024 (TW) versus AFT-CM. In group B, #p 5 0.052 for AFT-CM versus MV8. b Values

stimulated in the presence of AFT024 cells (28.1%, 23.4–33.6%). In contrast, the recovery of LTC-IC in cultures with AFT-CM (32.1%, 26.2–39.3) compared to AFT024 cells (18.9%, 17.1–20.9) was higher (p , 0.05; Table 2). Consequently, the number of transduced LTC-IC in these groups was similar. In a second set of experiments, AFT-CM was compared with MV8. In both groups, cells were cultured and transduced in the presence of F/S/7/T in 6-well plates. Again, there was no difference (p 5 0.95, Table 2) in the number of transduced LTC-IC between AFT-CM (3.4%, 2.5–4.5) and MV8 (3.4%, 2.4–5.0). In these experiments, the percentage of transduced LTC-IC was somewhat higher with MV8 compared to AFTCM (12.2% versus 7.1%; not significant [NS]), but the recovery of LTC-IC was lower (28.2% versus 47.4%, p 5 0.052, Table 2). Why the levels of gene transfer and recovery differed between AFT024, AFT-CM, and MV8 remains to be investigated. In summary, the data show that AFT-CM and MV8 can be used instead of AFT024 without a numeric loss of transduced LTC-IC.

Improved transduction of CFC and SRC with a RD114 compared to a GaLV-pseudotyped vector Using the two stroma-free conditions (AFT-CM and MV8) that showed optimal gene transfer with the GaLV-pseudotyped vector, we now compared transduction between this vector and a novel RD114-pseudotyped vector, which has shown to yield high levels of gene transfer in UCB-derived SRC (Kelly et al., 2000; Gatlin et al., 2001). In the first experiment, cells were transduced, tested for CFC, and transplanted into NOD/SCID mice. As no difference was found between groups pre-stimulated with AFT-CM or MV8 in the percent gene transfer or in the absolute number of transduced CFC and SRC (individual data not shown), groups were pooled. The percent gene transfer in CFC transduced with the RD114-pseudotyped vector (49.3%, 41.3–58.9), compared to the GaLV-pseudotyped vector (11.1%, 9.3–13.1), was significantly higher (Table 3, p ,

0.001). In addition, relatively more colony-forming unit granulocyte-macrophage (CFU-GM) and less burst-forming unit erythroid (BFU-E) were transduced with the RD114-pseudotyped vector (p , 0.0001 for comparison of BFU-E/CFU-GM ratio, Table 3). We speculate that this may be related to increased cell division among myeloid progenitor cells as SN from the RD114/MFG-eGFP packaging cell line has been shown to induce differentiation of human CD341 382 cells (Kelly et al., 2000). At 11 weeks posttransplantation, 33-fold more transduced cells (615 versus 18.4 per 50,000 cells; p , 0.001) were detected in recipients of RD114 compared to GaLV-transduced cells (Fig. 4C and Table 3). When expressed per million CD341 cells injected, 65.6 3 104 eGFP1 cells could be detected in the marrow of recipients of RD114-transduced cells, compared to 1.96 3 104 eGFP1 cells in recipients of GaLV-transduced cells. This difference was primarily because of a 21-fold difference in gene-transfer efficiency (0.8% versus 16.3%; p , 0.01, Fig. 4B and Table 3), as the level of engraftment between recipients of GaLV- or RD114-transduced cells was not significantly different (Fig. 4A and Table 3). Compared to our initial transplant data (shown in Fig. 3), the overall number of transduced cells in the marrow improved 144-fold (p , 0.0001), from 0.46 3 104 (F/S/7/T, without AFT024; GaLV) to 65.6 3 104 (F/S/7/T, AFT-CM or MV8; RD114) per million CD341 cells injected. As previously demonstrated by us and others (Hogan et al., 1997; van der Loo et al., 1998), the majority of human cells in the BM were of the B-cell (CD191 ) lineage (85% 6 14%, mean 6 SE, n 5 12). The ratio between myeloid and lymphoid cells, as well as the level of gene transfer in these subsets, was similar among all recipients in each group (not shown). The difference in gene transfer in CFC and SRC obtained with the GaLV and RD114-pseudotyped vector may have been partly caused by a difference titer; 3.3-fold based on infection of HEL cells, and 4.4-fold based on the relative level of retroviral RNA (see Methods). To control for this variable, MPB cells were transduced using GaLV- and RD114-pseudotyped

1325

15.2 (12.4–18.6) 33.7 (24.8–45.9) 2.2**

BFU-E

8.2 (6.1–10.9) 55.2 (44.5–68.3) 6.7***

CFU-GM

% eGFP expression (range)

11.1 (9.3–13.1) 49.3 (41.3–58.9) 4.4***

Total

GENE -TRANSFER EFFICIENCY OF

4.7 (3.1–7.3) 7.6 (5.0–11.6) 1.6 (NS)

CD451 0.7 (0.4–1.1) 0.9 (0.6–1.4) 1.4 (NS)

CD331

% Engraftment (range)

IN

2.3 (1.3–4.0) 6.8 (4.9–9.4) 2.9 (NS)

CD191

0.8 (0.4–1.5) 16.3 (13.3–19.9) 20.9*

% Gene Transfer in CD451 cells (range)a

NOD/SCID MICE

In vivo analysis

MOBILIZED PERIPHERAL BLOOD CD341 CELLS

18.4 (9.8–34.4) 615 (401–945) 33.4**

eGFP1 cells per 50,000 cells (range)a

a Relative percentage of gene transfer and absolute number of eGFP1 cells in recipients of GaLV and RD114 transduced cells were similar for the myeloid (CD331 ) and lymphoid (CD191 ) subpopulations (not shown). b Pooled data from AFT-CM and MV8 groups. G-CSF mobilized peripheral blood CD341 cells were cultured for 4 days in AFT-conditioned medium (AFT-CM) or MV8 stroma-free medium (MV8) supplemented with Flt3-L, SCF, IL-7 and TPO, and transduced with the MFG-eGFP retrovirus vector pseudotyped with either the gibbon ape leukemia virus (GaLV) or feline endogenous retrovirus (RD114) envelope. After infection, cells were cultured in standard methylcellulose cultures (300 to 1000 cells/dish, triplicate cultures) and injected into sublethally irradiated (3 Gy) NOD/SCID mice (4 3 106 CD341 cells/mouse input cells before culture, 3 mice/group). Colonies were scored after 14 days. NOD/SCID mice were analyzed at 11 weeks posttransplantation (raw data presented in Figure 4). As no difference was found in colony counts, gene-transfer efficiency, and engraftment between AFT-CM and MV8 (individual data not shown), groups were pooled. Data were analyzed after transformation to a logarithmic scale, as indicated in Methods. Range represents mean 6 SE. *p , 0.01; **p , 0.001, ***p , 0.0001, (NS) not significant (two-sample t test).

GaLV-transduced (n 5 6)b RD114-transduced (n 5 6)b Fold-difference

Group

AND

In vitro analysis

TABLE 3. ENGRAFTMENT

1326

VAN DER LOO ET AL.

FIG. 4. High-level gene transfer into SRC with RD114-pseudotyped vector. MPB CD341 cells (12 3 106 cells per group) were prestimulated for 4 days in AFT-CM or in MV8, both in the presence of F/S/7/T. Cells from each group were transduced with either a GaLV or an RD114-pseudotyped MFG-eGFP retroviral vector and transplanted into sublethally irradiated NOD/SCID mice (at 43 106 CD341 cells per mouse, cell count before culture). Bone marrow was analyzed at 11 weeks posttransplant by flow cytometry. AFT-CM and MV8 groups (not significantly different; individual data not shown) were pooled. A: Percent human (CD451 cells) among nucleated cells in the bone marrow. For each animal, background staining with an isotype control antibody had been subtracted. B: Percent gene transfer (percent eGFP1 cells among CD451 cells). C: Absolute number of transduced (eGFP1 ) cells among 50,000 nucleated cells from NOD/SCID bone marrow. Dead cells were excluded from all analyses based on staining with propidium iodide. Dots represent individual animals. Horizontal bars indicate the average per group. p values for comparison between groups (derived from a two-sample t test) have been indicated.

virus at a similar titer by diluting RD114-SN fourfold. Again, NOD/SCID mice transplanted with GaLV and RD114-transduced cells show similar levels of human cell engraftment (NS; Fig. 5A). Compared to recipients of GaLV-transduced cells, 7.7-fold more eGFP1 cells (827 versus 107 per 50,000 cells; p 5 0.003) were detected in recipients of RD114-transduced cells (Fig. 5C). Similar to the previous experiment, this could be attributed primarily to an increase in percent gene transfer (6.5-fold; 4.4% versus in 0.7 %; p , 0.001; Fig. 5B). As described in the methods, infection of HEL cells with both RD114- and GaLV-pseudotyped vectors in the presence and absence of FN showed that the RD114-pseudotyped vector bound FN with 2.5-fold greater efficiency. Therefore, the higher level of gene transfer in primitive human hematopoietic cells using the RD114- compared to the GaLV-pseudotyped vector can be attributed not only to a more efficient infection of the target cell, but also to a higher titer and increased efficiency in binding to FN.

DISCUSSION Using murine oncoretrovirus vectors, several investigators have demonstrated efficient gene transfer into primitive human hematopoietic cells derived from UCB (Conneally et al., 1998; Marandin et al., 1998; van Hennik et al., 1998; Hennemann et al., 1999; Novelli et al., 1999; Sanyal and Schuening, 1999; Barquinero et al., 2000; Kelly et al., 2000). In direct comparison, gene transfer into MPB-derived hematopoietic cells has been relatively inefficient. In a recent paper we demonstrated that primitive cells in MPB, compared to UCB, require a longer culture period for more efficient gene transfer (Pollok et al.,

2001). In the study presented here, we investigated the effect of an alternate combination of cytokines, with or without stromal cells that had been described to allow maintenance or even expansion of primitive hematopoietic cells (Moore et al., 1997a; Thiemann et al., 1998; Lewis et al., 2001). In line with the documented effect of the stromal cell line AFT024 on the maintenance of engraftable hematopoietic cells (Moore et al., 1997a; Lewis et al., 2001), we report here that prestimulation in the presence of AFT024 improved the absolute number of transduced hematopoietic cells in the NOD/SCID mouse. This was caused in part to an increase in gene-transfer efficiency, and in part to the positive effect of AFT024 on survival of the target cell. Furthermore, this effect was maintained under stroma-free conditions, using either AFT-CM or a defined medium, designated MV8, which had been supplemented with factors previously identified in AFT-CM (Punzel et al., 1999; Lewis et al., 2001). Finally, using the conditions that were optimized based on the level of gene transfer with a GaLV-pseudotyped vector, we demonstrated an additional increase in the number of transduced SRC using an RD114-pseudotyped retroviral vector. This finding is consistent with the high level of gene transfer with an RD114- compared to a GaLV-pseudotyped vector in human peripheral blood cells (Porter et al., 1996; Onodera et al., 1998) as well as in UCB-derived SRC (Kelly et al., 2000; Gatlin et al., 2001). The lack of difference in gene transfer between transduction with GaLV- and RD114-pseudotyped vectors, as demonstrated in a canine model (Goerner et al., 2001), may be related to the differentiating effect of the FLYRD18 packaging cell line as suggested by others (Kelly et al., 2000). In the current study, differentiation of SRC was avoided by washing the FN-coated plates preloaded with virus prior to infection, as previously described (Kelly et al., 2000). Although not all pre-

TRANSDUCTION OF HUMAN MPB CELLS

1327

FIG. 5. Comparison of gene transfer with virus supernatants of identical titer. MPB CD341 cells (20 3 106 cells per group) were prestimulated for 4 days in MV8 in the presence of F/S/7/T. Cells from each group were transduced with a GaLV-pseudotyped retroviral vector or with a 4-fold–diluted RD114-pseudotyped vector, and injected into sublethally irradiated NOD/SCID mice (at 4 3 106 CD341 cells per mouse, cell count before culture). Bone marrow was analyzed at 9 weeks posttransplant by flow cytometry. A: Percent human (CD451 ) cells among nucleated cells in the bone marrow. B: Percent gene transfer. (C) Absolute number of transduced (eGFP1 ) cells among 50,000 nucleated cells. Dead cells were excluded from all analyses based on staining with propidium iodide. Dots represent individual animals. Horizontal bars indicate the average per group. p values for comparison between groups (derived from a two-sample t test) have been indicated.

stimulation conditions tested here with the GaLV-pseudotyped vector had been repeated for transduction with the RD114pseudotyped vector, the higher level of gene transfer in SRC with this vector clearly warrants future optimization. Specifically, future studies will be aimed at testing whether the improved yield of transduced cells using RD114 negates the requirement for AFT-CM or MV8 when cultured in the presence of low doses of cytokines, or whether these conditions actually enhanced the effect of RD114. When comparing differently cultured populations of MPB CD341 cells, there were differences in the percent gene transfer in both CFC and LTC-IC between groups that showed a similar level of cell expansion. This indicates that the rate of proliferation may not have been the only variable limiting the gene-transfer efficiency in these cells. One of the alternative mechanisms by which cytokines enhance gene transfer into hematopoietic cells is upregulation of viral receptor expression. In both CD341 and CD341 382 cells, cytokines have been shown to increase expression of the receptor specific for the amphotropic virus envelope PiT-2 (Crooks and Kohn, 1993; Orlic et al., 1996; Kaubisch et al., 1999). Similar results have been obtained for the GaLV-receptor PiT-1 in hematopoietic cell lines (Sabatino et al., 1997). There was a positive correlation between the percent gene transfer with an amphotropic or GaLV-pseudotyped vector and the level of PiT-2 or PiT-1 mRNA, respectively (Sabatino et al., 1997). Furthermore, this study showed that receptor expression was independent of the cell cycle status of the cells. Therefore, we speculate that the increase in gene transfer in CFC and LTC-IC, as observed in the current study with the GaLV-pseudotyped vector, may have been partly caused by changes in expression of PiT-1. Alternatively, the level of cell expansion measured may not have accurately reflected the rate of proliferation, as different prestimulation conditions may have differently affected the rate of cell

death. Future investigations on the relationship between cytokine stimulation and receptor expression for both GaLV and RD114-pseudotyped vectors may provide additional insight in how to further improve the gene-transfer efficiency in MPB cells. Various coculture systems have been described that allow for the expansion of primitive hematopoietic cells (Gan et al., 1997; Moore et al., 1997a; Brandt et al., 1998; Thiemann et al., 1998; Reese et al., 1999; Shih et al., 1999; Tsuji et al., 1999). Of the cells tested, AFT024 has been found to support the survival of primitive human and mouse hematopoietic cells for up to several weeks, as demonstrated in vitro, and by transplantation in vivo (Moore et al., 1997a; Thiemann et al., 1998; Lewis and Verfaillie, 2000; Lewis et al., 2001). AFT024 is one of 225 stromal cell lines that were generated from a day-14 murine fetal liver (Moore et al., 1997a). Surprisingly, analysis of the cytokine expression profiles in supportive and nonsupportive clones showed that stem cell support did not correlate with patterns of cytokine transcription (Wineman et al., 1996). Comparison of AFT024 with lines that failed to support stem cells ultimately led to the identification of dlk, a trans-membrane molecule with epidermal growth factor (EGF)-like repeat motifs that, in soluble form, could transfer hematopoietic supportive activity to a non-supportive line (Moore et al., 1997b). The presence of a soluble factor, or factors, is consistent with our current data and with the observation that primitive hematopoietic cells can be maintained or expanded in AFT024 noncontact cultures or with AFT-CM (Punzel et al., 1999; Lewis and Verfaillie, 2000; Lewis et al., 2001). Another class of molecules that is important for maintenance of LTC-IC are large 6-O-sulfated heparan sulphate proteoglycans (HSPG) (Gupta et al., 1996). This type of HSPG has been identified in supernatant of hematopoiesis-supportive but not non-supportive cell lines (Gupta et al., 1996, 1998). In the ab-

1328

VAN DER LOO ET AL.

sence of stroma, but in medium supplemented with 6-O-sulfated HSPG and low doses of growth factors, LTC-IC could be maintained for 5 weeks in vitro (Gupta et al., 2000). Because HSPG can bind growth-promoting and growth-inhibitory factors (Bruno et al., 1995; Gupta et al., 2000), it has been speculated that HSPG are involved in the presentation of factors to hematopoietic cells (Roberts et al., 1988; Gupta et al., 1998, 2000). Here, we demonstrate that AFT024, or AFT-CM, can be replaced by MV8 with similar recovery of transduced LTC-IC, supporting this hypothesis. The latter medium had been supplemented with O-sulfated heparin as well as with factors previously identified in cultures that support LTC-IC (Punzel et al., 1999; Lewis et al., 2001). We speculate that the variation in gene transfer in LTC-IC, when cultured in the presence of AFT024 or with AFT-CM or MV8 (Table 2), may be caused by an inhibitory effect of high doses of HSPG on the transduction. Our data show that the reduction in gene-transfer efficiency in these experiments was accompanied by a proportional increase in the number of cells, and vice versa, generating equal numbers of transduced cells under both culture conditions. At present it is not known whether and how these parameters are linked. More study will be required to explain this effect and to define the optimal concentration of HSPG in MV8 medium for gene transfer. The level of gene transfer in MPB-derived SRC reported here is comparable to the levels reported in the literature (Schiedlmeier et al., 2000; Schilz et al., 2000). However, it may be difficult to compare the absolute levels of gene transfer between these studies and our current study directly because of differences in the source of cells (cells from cancer patients that had received irradiation or chemotherapy versus cells from healthy volunteers) and gene transfer methodology (length of culture, choice of cytokines, centrifugation, and virus titer). In our hands, the gene transfer efficiency in SRC ranged from 13% to 20% despite a relatively low virus titer of 1 3 104 IU/ml as determined on HEL cells (and 293 cells; not shown) in the absence of FN, and of 2 3 105 IU/ml, as determined on HEL cells in the presence of FN. The multiplicity of infection (MOI) in our transplant experiments, based on the effective titer of 2 3 105 IU/ml, was estimated to be only 0.3. We speculate that the absolute level of gene transfer in MPB-derived primitive hematopoietic cells may be improved even further by infecting the cells at a higher MOI and, based on our recently published data (Pollok et al., 2001), by prestimulating the cells for a more extended period of time using conditions that allow maintenance of primitive human hematopoietic cells, such as the conditions presented in this study.

ACKNOWLEDGMENTS We thank Sue Fautsch at the Cell Therapy Core Facility of the University of Minnesota Cancer Center for preparing the CD341 cells, Scott Wissink for his technical help, Karen E. Pollok (Wells Center for Pediatric Research, Indianapolis, IN) for performing real-time RT-PCR on viral SN, and Drs. Catherine M. Verfaillie (Stem Cell Institute, University of Minnesota, Minneapolis, MN), David A. Williams (Experimental Hematology, Children’s Hospital Research Center, Cincinnati, OH) and Karen E. Pollok for critically reviewing the manuscript.

Supported in part by the American Cancer Society Institutional Research Grant (IRG-58-001-40-IRG-22), the Children’s Cancer Research Fund (University of Minnesota), the Grant-inAid of Research, Artistry and Scholarship (University of Minnesota), and the National Institutes of Health (5 P01-CA6549305).

REFERENCES ABONOUR, R., WILLIAMS, D.A., EINHORN, L., HALL, K.M., CHEN, J., COFFMAN, J., TRAYCOFF, C.M., BANK, A., KATO, I., WARD, M., WILLIAMS, S.D., HROMAS, R., ROBERTSON, M.J., SMITH, F.O., WOO, D., MILLS, B., SROUR, E.F., and CORNETTA, K. (2000). Efficient retrovirus-mediated transfer of the multidrug resistance 1 gene into autologous human long-term repopulating hematopoietic stem cells. Nat. Med. 6, 652–658. BARQUINERO, J., SEGOVIA, J.C., RAMIREZ, M., LIMON, A., GUENECHEA, G., PUIG, T., BRIONES, J., GARCIA, J., and BUEREN, J.A. (2000). Efficient transduction of human hematopoietic repopulating cells generating stable engraftment of transgeneexpressing cells in NOD/SCID mice. Blood 95, 3085–3093. BENSINGER, W.I., MARTIN, P.J., STORER, B., CLIFT, R., FORMAN, S.J., NEGRIN, R., KASHYAP, A., FLOWERS, M.E., LILLEBY, K., CHAUNCEY, T.R., STORB, R., and APPELBAUM, F.R. (2001). Transplantation of bone marrow as compared with peripheral-blood cells from HLA-identical relatives in patients with hematologic cancers. N. Engl. J. Med. 344, 175–181. BIERHUIZEN, M.F.A., WESTERMAN, Y., VISSER, T.P., DIMJATI, W., WOGNUM, A.W., and WAGEMAKER, G. (1997). Enhanced green fluorescent protein as selectable marker of retroviral-mediated gene transfer in immature hematopoietic bone marrow cells. Blood 90, 3304–3315. BRANDT, J.E., GALY, A.H., LUENS, K.M., TRAVIS, M., YOUNG, J., TONG, J., CHEN, S., DAVIS, T.A., LEE, K.P., CHEN, B.P., TUSHINSKI, R., and HOFFMAN, R. (1998). Bone marrow repopulation by human marrow stem cells after long-term expansion culture on a porcine endothelial cell line. Exp. Hematol. 26, 950–961. BRUNO, E., LUIKART, S.D., LONG, M.W., and HOFFMAN, R. (1995). Marrow-derived heparan sulfate proteoglycan mediates the adhesion of hematopoietic progenitor cells to cytokines. Exp. Hematol. 23, 1212–1217. CHAMPLIN, R.E., SCHMITZ, N., HOROWITZ, M.M., CHAPUIS, B., CHOPRA, R., CORNELISSEN, J.J., GALE, R.P., GOLDMAN, J.M., LOBERIZA, F.R., Jr., HERTENSTEIN, B., KLEIN, J.P., MONTSERRAT, E., ZHANG, M.J., RINGDEN, O., TOMANY, S.C., ROWLINGS, P.A., VAN HOEF, M.E., and GRATWOHL, A. (2000). Blood stem cells compared with bone marrow as a source of hematopoietic cells for allogeneic transplantation. IBMTR Histocompatibility and Stem Cell Sources Working Committee and the European Group for Blood and Marrow Transplantation (EBMT). Blood 95, 3702–3709. CHERVENICK, P.A., BOGGS, D.R., MARSH, J.C., CARTWRIGHT, G.E., and WINTROBE, M.M. (1968). Quantitative studies of blood and bone marrow neutrophils in normal mice. J. Physiol. 215, 353– 359. CONNEALLY, E., EAVES, C.J., and HUMPHRIES, R.K. (1998). Efficient retroviral-mediated gene transfer to human cord blood stem cells with in vivo repopulating potential. Blood 91, 3487–3493. COSSET, F.L., TAKEUCHI, Y., BATTINI, J.L., WEISS, R.A., and COLLINS, M.K. (1995). High-titer packaging cells producing recombinant retroviruses resistant to human serum. J. Virol. 69, 7430– 7436. CROOKS, G.M., and KOHN, D.B. (1993). Growth factors increase am-

TRANSDUCTION OF HUMAN MPB CELLS photropic retrovirus binding to human CD341 bone marrow progenitor cells. Blood 82, 3290–3297. FAZEKAS DE STGROTH, S. (1982). The evaluation of limiting dilution assays. J. Immunol. Methods 49, R11–R23. GAN, O.I., MURDOCH, B., LAROCHELLE, A., and DICK, J.E. (1997). Differential maintenance of primitive human SCID-repopulating cells, clonogenic progenitors, and long-term culture-initiating cells after incubation on human bone marrow stromal cells. Blood 90, 641–650. GATLIN, J., MELKUS, M.W., PADGETT, A., KELLY, P.F., and GARCIA, J.V. (2001). Engraftment of NOD/SCID mice with human CD34(1) cells transduced by concentrated oncoretroviral vector particles pseudotyped with the feline endogenous retrovirus (RD114) envelope protein. J. Virol. 75, 9995–9999. GOERNER, M., HORN, P.A., PETERSON, L., KURRE, P., STORB, R., RASKO, J.E., and KIEM, H.P. (2001). Sustained multilineage gene persistence and expression in dogs transplanted with CD34(1) marrow cells transduced by RD114-pseudotype oncoretrovirus vectors. Blood 98, 2065–2070. GUPTA, P., MCCARTHY, J.B., and VERFAILLIE, C.M. (1996). Stromal fibroblast heparan sulfate is required for cytokine-mediated ex vivo maintenance of human long-term culture-initiating cells. Blood 87, 3229–3236. GUPTA, P., OEGEMA, T.R., BRAZIL, J.J., DUDEK, A.Z., SLUNGAARD, A., and VERFAILLIE, C.M. (1998). Structurally specific heparan sulfates support primitive human hematopoiesis by formation of a multimolecular stem cell niche. Blood 92, 4641–4651. GUPTA, P., OEGEMA, T.R., BRAZIL, J.J., DUDEK, A.Z., SLUNGAARD, A., and VERFAILLIE, C.M. (2000). Human LTC-IC can be maintained for at least 5 weeks in vitro when interleukin-3 and a single chemokine are combined with O-sulfated heparan sulfates: requirement for optimal binding interactions of heparan sulfate with early-acting cytokines and matrix proteins. Blood 95, 147–155. HANENBERG, H., XIAO, X.L., DILLOO, D., HASHINO, K., KATO, I., and WILLIAMS, D.A. (1996). Colocalization of retrovirus and target cells on specific fibronectin fragments increases genetic transduction of mammalian cells. Nat. Med. 2, 876–882. HENNEMANN, B., CONNEALLY, E., PAWLIUK, R., LEBOULCH, P., ROSE-JOHN, S., REID, D., CHUO, J.Y., HUMPHRIES, R.K., and EAVES, C.J. (1999). Optimization of retroviral-mediated gene transfer to human NOD/SCID mouse repopulating cord blood cells through a systematic analysis of protocol variables. Exp. Hematol. 27, 817–825. HOGAN, C.J., SHPALL, E.J., MCNIECE, I., and KELLER, G. (1997). Multilineage engraftment in NOD/LtSz-scid/scid mice from mobilized human CD341 peripheral blood progenitor cells. Biol. Blood Marrow Transplant. 3, 236–246. KAUBISCH, A., WARD, M., SCHOETZ, S., HESDORFFER, C., and BANK, A. (1999). Up-regulation of amphotropic retroviral receptor expression in human peripheral blood CD341cells. Am. J. Hematol. 61, 243–253. KELLY, P.F., VANDERGRIFF, J., NATHWANI, A., NIENHUIS, A.W., and VANIN, E.F. (2000). Highly efficient gene transfer into cord blood nonobese diabetic/severe combined immunodeficiency repopulating cells by oncoretroviral vector particles pseudotyped with the feline endogenous retrovirus (RD114) envelope protein. Blood 96, 1206–1214. KIEM, H.-P., ANDREWS, R.G., MORRIS, J., PETERSON, L., HEYWARD, S., ALLEN, J.M., RASKO, J.E.J., POTTER, J., and MILLER, D.A. (1998). Improved gene transfer into baboon marrow repopulating cells using recombinant human fibronectin fragment CH-296 in combination with interleukin-6, stem cell factor, FLT-3 ligand, and megakaryocyte growth and development factor. Blood 92, 1878–1886. KIEM, H.P., HEYWARD, S., WINKLER, A., POTTER, J., ALLEN,

1329 J.M., MILLER, A.D., and ANDREWS, R.G. (1997). Gene transfer into marrow repopulating cells: comparison between amphotropic and gibbon ape leukemia virus pseudotyped retroviral vectors in a competitive repopulation assay in baboons. Blood 90, 4638–4645. LEWIS, I.D., ALMEIDA-PORADA, G., DU, J., LEMISCHKA, I.R., MOORE, K.A., ZANJANI, E.D., and VERFAILLIE, C.M. (2001). Umbilical cord blood cells capable of engrafting in primary, secondary, and tertiary xenogeneic hosts are preserved after ex vivo culture in a noncontact system. Blood 97, 3441–3449. LEWIS, I.D., and VERFAILLIE, C.M. (2000). Multi-lineage expansion potential of primitive hematopoietic progenitors: superiority of umbilical cord blood compared to mobilized peripheral blood. Exp. Hematol. 28, 1087–1095. MARANDIN, A., DUBART, A., PFLUMIO, F., COSSET, F.L., CORDETTE, V., CHAPEL-FERNANDES, S., COULOMBEL, L., VAINCHENKER, W., and LOUACHE, F. (1998). Retrovirus-mediated gene transfer into human CD34138low primitive cells capable of reconstituting long-term cultures in vitro and nonobese diabetic-severe combined immunodeficiency mice in vivo. Hum. Gene Ther. 9, 1497–1511. MILLER, D.G., ADAM, M.A., and MILLER, A.D. (1990). Gene transfer by retrovirus vectors occurs only in cells that are actively replicating at the time of infection. Mol. Cell. Biol. 10, 4239–4242. MOORE, K.A., EMA, H., and LEMISCHKA, I.R. (1997a). In vitro maintenance of highly purified, transplantable hematopoietic stem cells. Blood 89, 4337–4347. MOORE, K.A., PYTOWSKI, B., WITTE, L., HICKLIN, D., and LEMISCHKA, I.R. (1997b). Hematopoietic activity of a stromal cell transmembrane protein containing epidermal growth factor-like repeat motifs. Proc. Natl. Acad. Sci. U.S.A. 94, 4011–4016. MORITZ, T., PATEL, V.P., and WILLIAMS, D.A. (1994). Bone marrow extracellular matrix molecules improve gene transfer into human hematopoietic cells via retroviral vectors. J. Clin. Invest. 93, 1451–1457. NOLTA, J.A., THIEMANN, F.T., ARAKAWA-HOYT, J., DAO, M.A., BARSKY, L.W., MOORE, K.A., LEMISCHKA, I.R., and CROOKS, G.M. (2002). The AFT024 stromal cell line supports longterm ex vivo maintenance of engrafting multipotent human hematopoietic progenitors. Leukemia 16, 352–361. NOVELLI, E.M., CHENG, L.Z., YANG, Y.D., LEUNG, W., RAMIREZ, M., TANAVDE, V., ENGER, C., and CIVIN, C.I. (1999). Ex vivo culture of cord blood CD34(1) cells expands progenitor cell numbers, preserves engraftment capacity in nonobese diabetic/severe combined immunodeficient mice, and enhances retroviral transduction efficiency. Hum. Gene Ther. 10, 2927–2940. ONODERA, M., NELSON, D.M., YACHIE, A., JAGADEESH, G.J., BUNNELL, B.A., MORGAN, R.A., and BLAESE, R.M. (1998). Development of improved adenosine deaminase retroviral vectors. J. Virol. 72, 1769–1774. ORLIC, D., GIRARD, L.J., JORDAN, C.T., ANDERSON, S.M., CLINE, A.P., and BODINE, D.M. (1996). The level of mRNA encoding the amphotropic retrovirus receptor in mouse and human hematopoietic stem cells is low and correlates with the efficiency of retrovirus transduction. Proc. Natl. Acad. Sci. U.S.A. 93, 11097– 11102. PETZER, A.L., ZANDSTRA, P.W., PIRET, J.M., and EAVES, C.J. (1996). Differential cytokine effects on primitive (CD341CD38-) human hematopoietic cells: novel responses to Flt3-ligand and thrombopoietin. J. Exp. Med. 183, 2551–2558. POLLOK, K.E., VAN DER LOO, J.C.M., COOPER, R.J., HARTWELL, J.R., MILES, K.R., BREESE, R., WILLIAMS, E.P., MONTEL, A., SESHADRI, R., HANENBERG, H., and WILLIAMS, D.A. (2001). Differential transduction efficiency of SCID-repopulating cells derived from umbilical cord blood and granulocyte colony-stimulating factor-mobilized peripheral blood. Hum. Gene Ther. 12, 2095–2108.

1330 POLLOK, K.E., VAN DER LOO, J.C.M., COOPER, R.J., KENNEDY, L., and WILLIAMS, D.A. (1999). Costimulation of transduced T lymphocytes via T-cell receptor-CD3 complex and CD28 leads to increased transcription of integrated retrovirus. Hum. Gene Ther. 10, 2221–2236. PORTER, C.D., COLLINS, M.K., TAILOR, C.S., PARKAR, M.H., COSSET, F.L., WEISS, R.A., and TAKEUCHI, Y. (1996). Comparison of efficiency of infection of human gene therapy target cells via four different retroviral receptors. Hum. Gene Ther. 7, 913–919. PUNZEL, M., GUPTA, P., ROODELL, M., MORTARI, F., and VERFAILLIE, C.M. (1999). Factor(s) secreted by AFT024 fetal liver cells following stimulation with human cytokines are important for human LTC-IC growth. Leukemia 13, 1079–1084. REESE, J.S., KOC, O.N., and GERSON, S.L. (1999). Human mesenchymal stem cells provide stromal support for efficient CD34(1) transduction. J. Hematother. Stem Cell Res. 8, 515–523. ROBERTS, A.W., and METCALF, D. (1995). Noncycling state of peripheral blood progenitor cells mobilized by granulocyte colonystimulating factor and other cytokines. Blood 86, 1600–1605. ROBERTS, R., GALLAGHER, J., SPOONCER, E., ALLEN, T.D., BLOOMFIELD, F., and DEXTER, T.M. (1988). Heparan sulphate bound growth factors: A mechanism for stromal cell mediated haemopoiesis. Nature 332, 376–378. SABATINO, D.E., DO, B.Q., PYLE, L.C., SEIDEL, N.E., GIRARD, L.J., SPRATT, S.K., ORLIC, D., and BODINE, D.M. (1997). Amphotropic or gibbon ape leukemia virus retrovirus binding and transduction correlates with the level of receptor mRNA in human hematopoietic cell lines. Blood Cells Mol. Dis. 23, 422–433. SANYAL, A., and SCHUENING, F.G. (1999). Increased gene transfer into human cord blood cells by centrifugation-enhanced transduction in fibronectin fragment-coated tubes. Hum. Gene Ther. 10, 2859–2868. SCHIEDLMEIER, B., KUHLCKE, K., ECKERT, H.G., BAUM, C., ZELLER, W.J., and FRUEHAUF, S. (2000). Quantitative assessment of retroviral transfer of the human multidrug resistance 1 gene to human mobilized peripheral blood progenitor cells engrafted in nonobese diabetic/severe combined immunodeficient mice. Blood 95, 1237–1248. SCHILZ, A.J., SCHIEDLMEIER, B., KUHLCKE, K., FRUEHAUF, S., LINDEMANN, C., ZELLER, W.J., GREZ, M., FAUSER, A.A., BAUM, C., and ECKERT, H.G. (2000). MDR1 gene expression in NOD/SCID repopulating cells after retroviral gene transfer under clinically relevant conditions. Mol. Ther. 2, 609–618. SHIH, C.C., HU, M.C., HU, J., MEDEIROS, J., and FORMAN, S.J. (1999). Long-term ex vivo maintenance and expansion of transplantable human hematopoietic stem cells. Blood 94, 1623–1636. SHULTZ, L.D., SCHWEITZER, P.A., CHRISTIANSON, S.W., GOTT, B., SCHWEITZER, I.B., TENNENT, B., MCKENNA, S., MOBRAATEN, L., RAJAN, T.V., GREINER, D.L., and LEITER, E.H. (1995). Multiple defects in innate and adaptive immunologic function in NOD/LtSz-scid mice. J. Immunol. 154, 180–191.

VAN DER LOO ET AL. THIEMANN, F.T., MOORE, K.A., SMOGORZEWSKA, E.M., LEMISCHKA, I.R., and CROOKS, G.M. (1998). The murine stromal cell line AFT024 acts specifically on human CD341CD382 progenitors to maintain primitive function and immunophenotype in vitro. Exp. Hematol. 26, 612–619. TSUJI, T., ITOH, K., NISHIMURA-MORITA, Y., WATANABE, Y., HIRANO, D., MORI, K.J., and YATSUNAMI, K. (1999). CD34(high1)CD38(low/2) cells generated in a xenogeneic coculture system are capable of both long-term hematopoiesis and multiple differentiation. Leukemia 13, 1409–1419. VAN DER LOO, J.C.M., HANENBERG, H., COOPER, R.J., LUO, F.Y., LAZARIDIS, E.N., and WILLIAMS, D.A. (1998). Nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mouse as a model system to study the engraftment and mobilization of human peripheral blood stem cells. Blood 92, 2556–2570. VAN HENNIK, P.B., VERSTEGEN, M.M.A., BIERHUIZEN, M.F.A., LIMON, A., WOGNUM, A.W., CANCELAS, J.A., BARQUINERO, J., PLOEMACHER, R.E., and WAGEMAKER, G. (1998). Highly efficient transduction of the green fluorescent protein gene in human umbilical cord blood stem cells capable of cobblestone formation in long-term cultures and multilineage engraftment of immunodeficient mice. Blood 92, 4013–4022. WILLIAMS, D.A., and SMITH, F.O. (2000). Progress in the use of gene transfer methods to treat genetic blood diseases. Hum. Gene Ther. 11, 2059–2066. WINEMAN, J., MOORE, K., LEMISCHKA, I., and MULLERSIEBURG, C. (1996). Functional heterogeneity of the hematopoietic microenvironment: rare stromal elements maintain long-term repopulating stem cells. Blood 87, 4082–4090. ZANDSTRA, P.W., CONNEALLY, E., PETZER, A.L., PIRET, J.M., and EAVES, C.J. (1997). Cytokine manipulation of primitive human hematopoietic cell self- renewal. Proc. Natl. Acad. Sci. U.S.A. 94, 4698–4703.

Address reprint requests to: Johannes C.M. van der Loo Cincinnati Children’s Hopsital Medical Center Divison of Experimental Hematology 3333 Burnet Avenue, ML 7013 TCHRF 2059 Cincinnati, OH 45229-3090 E-mail: [email protected] Received for publication February 7, 2002; accepted after revision June 20, 2002. Published online: July 8, 2002.