Foamy virus–mediated gene transfer to canine repopulating cells

Recent stem-cell gene-therapy studies in children with severe combined immunodeficiency (SCID)–X1 and adenosine deaminase have demonstrated the enormous potential of stem-cell gene therapy and also the potential risks. Thus, it will be crucial to identify not only vector systems that allow efficient stem-cell transduction but also safer vector systems. Foamy virus vectors are derived from foamy or spuma retroviruses, which have many properties that distinguish them from γ-viruses or lentiviruses; an important characteristic for gene therapy being that they are nonpathogenic.^1-3 

Foamy viruses (FVs) have been isolated from a variety of mammalian species, including cows and cats, and are present in most captive primates used for research.^1,2  Efficient replication appears to be limited to the oral mucosa, allowing transmission by biting,⁴  and animals exposed to FVs become seropositive.¹  FVs are not found in humans despite the fact that the prototype FV was isolated from cultured human cells and originally named human foamy virus.⁵  This isolate is now believed to be a chimpanzee virus from a zoonotic infection or a culture contamination, and an extensive survey demonstrated that FVs are not endemic in human populations.⁶  There are rare cases in which humans have been infected with FVs via bites from captive primates,^1,7  but, like FV infection in natural hosts, no pathology has ever been associated with FV infection.

FV vectors have improved from early replication-competent vectors to third-generation vectors that are free of replication-competent retroviruses.⁸  FV vectors can transduce pluripotent murine⁹  and human^10-13  hematopoietic repopulating cells in murine models. Significant silencing of vector transgenes was not observed in these studies. FV vectors are able to form a stable transduction intermediate in quiescent cells¹⁴  which may explain how they efficiently transduce quiescent G₀-mobilized peripheral blood (PB) cells¹¹  that divide following transplantation. Reverse transcription occurs in the cell producing the virion rather than the infected target cell,¹⁵  so FV virions contain reverse-transcribed full-length double-stranded cDNA¹⁶  that might be stable in quiescent cells. FV vectors also have a unique integration profile relative to γ-retrovirus and lentiviral vectors, with less frequent integration near promoters than MLV vectors and less frequent integration in genes than HIV vectors.¹⁷ 

Here, we used the dog as a clinically relevant large animal model to study FV vector gene transfer into long-term repopulating hematopoietic stem cells.

Dogs were raised and housed at the Fred Hutchinson Cancer Research Center (FHCRC) under conditions approved by the American Association for Accreditation of Laboratory Animal Care. Animal experiments were reviewed and approved by the Fred Hutchinson Cancer Research Institutional Animal Care and Use Committee. All animals were provided with commercial chow and chlorinated tap water ad libitum. In preparation for the harvest of stem or progenitor cells, the dogs received canine granulocyte-colony stimulating factor (cG-CSF, 5 μg/kg body weight subcutaneously, twice daily) and canine stem-cell factor (cSCF, 25 μg/kg body weight subcutaneously, once daily) for 5 consecutive days. Leukapheresis was performed using the COBE BCT Spectra Apheresis System (Gambro BCT, Lakewood, CO). The machine was primed with autologous blood. A dual-lumen venous catheter was inserted and connected to the COBE machine. During the procedure, the dogs were constantly monitored for level of sedation or signs of distress, and a slow infusion of 10% calcium gluconate was given to prevent cramping.

As preparation for transplantation, all animals received a single myeloablative dose of 920 cGy total body irradiation administered from a linear accelerator at 7 cGy/minute. The animals received broad-spectrum antibiotics and recombinant cG-CSF after transplantation until absolute neutrophil count (ANC) was greater than 1000/μL. The animals also received cyclosporine to inhibit immune responses to the EGFP transgene from the day before transplantation to 35 days after the transplantation (animals G272, G306). For animal G264, cyclosporine was stopped 14 days after transplantation because of the development of an intussusception. Cyclosporine was also administered from day 40 to day 128 to animal G272 with a tapered dose from days 117 to 128.

FV vector plasmid pΔΦPF contains an EGFP reporter transgene expressed from a murine phosphoglycerate kinase (PGK) promoter and was constructed using standard molecular biology techniques by replacing the murine stem-cell virus promoter of pΔΦMscvF⁸  with a PGK promoter. The FV vector used in this study, ΔΦPF, was produced by calcium phosphate transfection as previously described,^8,18  except that 12 μg pΔΦPF, 12 μg pCiGSΔPsi,¹¹  1.5 μg pCiPS, and 0.75 μg pCiES in a total volume of 800 μL were used for each 10-cm tissue culture dish. Stocks were titered by determining the number of EGFP-transducing units on human HT-1080¹⁹  cells.

The method has been described previously.^20,21  Briefly, cells were labeled with biotinylated monoclonal antibody 1H6 (IgG1 anti–canine CD34) at 4°C for 30 minutes. The cells were washed twice and then incubated with streptavidin-conjugated microbeads for 30 minutes at 4°C, washed, and then separated using an immunomagnetic column technique (Miltenyi Biotec, Auburn, CA) according to the manufacturer's instructions.

CD34-enriched cells from PB were exposed directly (without prior cryopreservation) to FV vectors at an MOI of 6 to 8 transducing units/cell for 18 hours in 75-cm² canted-neck flasks (Corning, Corning, NY) coated with CH-296 (RetroNectin; Takara Shuzo, Otsu, Japan) at a concentration of 2 μg/cm² in Iscoves modified Dulbecco medium supplemented with 10% FBS (GIBCO BRL), 1% sodium pyruvate, 1% L-glutamine, 1% penicillin/streptomycin (Gibco BRL, Gaithersburg, MD) in the presence of fms-like tyrosine kinase 3 ligand, cSCF, and cG-CSF at a concentration of 50 ng/mL each. After transduction, nonadherent and adherent cells were pooled, counted, and infused intravenously into the animal.

CD34-enriched cells were cultured in a double-layer agar culture system. Isolated cells were cultured in alpha minimal essential medium supplemented with FBS (Hyclone, Logan, UT), bovine serum albumin (fraction V; Sigma, St Louis, MO), 0.5% (wt/vol) agar (Difco, Detroit, MI), overlaid on medium with 0.3% agar (wt/vol) containing 100 ng/mL cSCF, cG-CSF, canine granulocyte-macrophage colony-stimulating factor and 4 U/mL erythropoietin. Cultures were incubated at 37°C in 5% CO₂ and 95% air in a humidified incubator. After infection, CD34⁺ cells were plated at a density of 2000 cell/plate (based on cell numbers prior to infection). Nontransduced control cells were plated at the same time. All cultures were performed at least in triplicate. The total number as well as the number of EGFP-positive colonies were enumerated at day 14 of culture by fluorescence microscopy.

EGFP-expressing white blood cells were quantitated by flow cytometric analysis of at least 250 000 events (propidium iodide [1 μg/mL]–excluding, forward and right-angle light scatter-gated) on a fluorescence activated cell sorting (FACS) Vantage (Becton Dickinson, San Jose, CA). For analysis of red blood cells and platelets, a FACS Calibur was used (Becton Dickinson). Flow cytometric data were analyzed by CELLQuest v3.1f software (Becton Dickinson) with gating to exclude fewer than 0.1% control cells in the relevant region. The results were then plotted over time in an Excel chart (Microsoft, Redmond, WA). Murine anti–human monoclonal antibodies conjugated to phycoerythrin (PE) and shown to bind to canine CD antigens were used to detect CD21 (clone CA2.1D6; SeroTec, Raleigh, NC) for B cells and CD14 (clone TÜK4; DAKO, Carpinteria, CA) for monocytes. The monoclonal antibody DM5 used to detect granulocytes and the anti-CD3 (clone 17.6B3) used for T cells were kindly provided by Drs Peter Moore (University of California, Davis, CA) and Brenda Sandmaier (Fred Hutchinson Cancer Research Center, Seattle, WA).

Provirus copy numbers were determined by measuring EGFP-gene levels with the TaqMan 5′ nuclease quantitative real-time polymerase chain reaction (PCR) assay.²²  Genomic peripheral blood leukocyte DNA (300 ng) was amplified at least in duplicate with a EGFP-specific primer/probe combination (5′-CTG CAC CAC CGG CAA-3′ and 5′-GTA GCG GCT GAA GCA CTG-3′; probe, 5′-FAM-CCA CCC TGA CCT ACG GCG TG-TAMRA-3′; Synthegen, Houston, TX). A canine IL-3–specific primer/probe combination (5′-ATG AGC AGC TTC CCC ATC C-3′, 5′-GTC GAA AAA GGC CTC CCC-3′; probe, 5′-FAM-TCC TGC TTG GAT GCC AAG TCC CAC-TAMRA-3′) was used to adjust for equal loading of genomic DNA. Standards consisted of dilutions of DNA extracted from cell lines containing a single-copy EGFP vector. Negative controls consisted of DNA extracted from peripheral blood mononuclear cells obtained before transplantation, from control animals, or from water. Reactions were run using the ABI master mix (Applied Biosystems, Branchburg, NJ) on the ABI Prism 7700 sequence detection system (Applied Biosystems) using the following thermal cycling conditions: 50°C for 2 minutes and 95°C for 10 minutes, then 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. Southern blots were performed on genomic DNA from peripheral blood leukocytes with an FV-specific probe and EcoNI digestion as described²³  and compared with dilutions of genomic DNA from a cell line with a single-copy FV vector provirus.

Integration site analysis by linear amplification-mediated–polymerase chain reaction (LAM-PCR) was performed on canine DNA isolated from peripheral blood leukocytes. One hundred nanograms of DNA served as template for LAM-PCR that was performed as described previously²⁴  with the following modifications. Briefly, 0.25 pmol vector-specific 5′-biotinylated primer f3LTR1 (5′-GT GAT TGC AAT GCT TTG TGC-3′) was used to anneal and extend linear fragments containing the LTR with 5 U ThermalAce DNA polymerase (Invitrogen, Carlsbad, CA). MspI or both BspHI and PciI restriction enzymes (NEB, Beverly, MA) were used to digest DNA after creating double-stranded DNA. Enzyme-specific linker cassettes were ligated onto the overhangs created by the restriction enzymes with 4 U Fast-Link DNA Ligase (Epicentre, Madison, WI) for 15 minutes at room temperature. Two additional rounds of nested PCRs with 25 pmol LTR-specific primers f3LTR2 (5′-ACC GAC TTG ATT CGA GAA CC-3′) and f3LTR3 5′-GCT AAG GGA GAC ATC TAG TG-3′) amplified the virus long terminal repeat (LTR) and genomic flanking regions using 4% of the first nested PCR as template for the second round. PCR products were visualized on Spreadex gels (Elcrom Scientific, Cham, Switzerland).

We transplanted 3 myeloablated animals with autologous peripheral blood CD34⁺ cells transduced by a EGFP-expressing FV vector. In all 3 animals a stable ANC greater than 500/μL was reached within 9 to 15 days. One dog (G264) developed a transplantation-related intussusception at day 17 and did not survive surgery to repair the condition. Intussusception in our canine colony is associated with irradiation and treatment with cyclosporine and occurs in both autologous and allogeneic transplantations. For the 2 animals that survived long term, a platelet count greater than 50 000/μL was reached at an average of 49 days (Table 1) Figure 1 displays ANC and platelet counts after transplantation for all 3 animals in this study. Engraftment was similar to our results using lentiviral vectors²⁵  and significantly faster than historic controls that received hematopoietic stem cells transduced by γ-retroviral vectors in a 3-day transduction protocol.²⁰  Long-term engraftment was stable in both of the long-term surviving dogs, and complete blood counts were within normal values over 1 year after transplantation; for G272 the ANC was 6.6 × 10³/μL and the platelet count was 2.9 × 10⁵/μL at day 400 after transplantation, for G306 the ANC was 9.8 × 10³/μL and the platelet count was 4.0 × 10⁵/μL at day 447 after transplantation.

Transduction efficiency prior to transplantation was assessed by flow cytometric determination of EGFP-positive CD34-enriched cells or by scoring EGFP-positive colony-forming cells by fluorescence microscopy on day 14. Transduction frequencies were high, especially given the relatively low MOI (8-10) with 11% to 24% EGFP-expressing colonies. Table 1 summarizes the results of the pretransplantation analysis of hematopoietic progenitor cells.

The transduction frequency of hematopoietic repopulating cells was measured by flow cytometric detection of EGFP in PB granulocytes and lymphocytes after transplantation (Figure 2). We observed long-term (> 500 days) EGFP expression in repopulating lymphocytes as high as 19%, and greater than 15% in both animals, and granulocyte EGFP expression as high as 19%, and greater than 13% in both animals. Transduction efficiency in CFCs from G306 and G272 at 1 year after transplantation was 13% and 19%, respectively. The level of transgene-expressing cells was stable over time, in contrast to marking with γ-retroviral vectors, which typically declines over time.^20,26  These transduction levels were slightly higher than what we have observed with lentiviral vectors (up to 12%),²⁵  and they were achieved with approximately 14-fold lower MOIs. Additionally, marking levels with FV vectors were higher in lymphocytes as compared with lentiviral vectors. FV vector provirus copy numbers were also determined by real-time PCR at several time points in DNA from PB (Figure 2C). There was approximately 2-fold difference between flow cytometry and real-time PCR results (compare Figure 2, panels A, B, and C), suggesting that on average, repopulating cells contained no more than 2 vector copies. At later time points both real-time PCR and flow cytometry results stabilized, indicating that gene expression from the FV vector was stable over time. Similar provirus levels were observed when 2 blood samples from dog G306 were analyzed by Southern blot, and the vector genome appeared to be intact (Figure 2D).

To assess gene expression in different hematopoietic lineages, PB cells were labeled with antibodies against granulocytes (DM-5), T lymphocytes (CD3), and monocytes (CD14) and analyzed by flow cytometry at different time points. We also analyzed BM-derived CD34⁺ cells. Figure 3A shows representative results obtained in 2 of the dogs approximately 1 year after transplantation. Sustained EGFP expression could be detected in all PB-cell subsets examined and the percentage of EGFP⁺ cells in bone marrow CD34⁺ cells was similar to that in PB. In both dogs we were also able to detect EGFP⁺ platelets and erythrocytes (Figure 3B). Because EGFP fluorescence intensity in these cell populations is significantly lower than in white blood cells, these data likely underestimate the true transduction rates in these lineages.

We have followed these animals by performing complete blood counts and LAM-PCR to monitor the potential development of myeloproliferation, lymphoproliferation, or leukemia. To date there has been no evidence of malignancy, and the 2 animals that survived the transplantation remain healthy. LAM-PCR performed using PB DNA showed that engrafted, repopulating cells were highly polyclonal (Figure 4). LAM-PCR analysis also demonstrated the transduction of multipotential repopulating clones. For 2 clones identified by LAM-PCR, we designed genome-specific primers and were able to find these sequences in FACS-purified CD3, CD14, CD21, and DM5 (granulocytes)–positive subpopulations. To rule out the possibility that clonal multilineage marking may have been due to contamination of the purified subsets with cells of the opposite lineage, we performed integrant-specific SYBR green quantitative PCR²⁷  on the purified DM5 (myeloid) and CD3 (lymphoid) subsets (data not shown). For the clone analyzed, marking was 6.9-fold higher in the lymphoid subset relative to the myeloid subset, but marking in the purified myeloid cells was at least 9.7-fold higher than would be expected from contamination of CD3 cells in the myeloid subset. These studies indicate that FV-mediated gene transfer results in polyclonal repopulation with multipotential clones.

In the current study we evaluated the ability of FV vectors to transduce hematopoietic stem cells in a clinically relevant large animal model after a short, 18-hour transduction protocol. Stable gene transfer into long-term hematopoietic repopulating cells was observed with up to 19% of PB cells expressing the transgene. This is a remarkable transduction frequency considering the short transduction period and the low MOI used. This is also the first report of efficient HSC transduction with FV in a clinically relevant large animal model. Transgene expression by FV vectors was stable over time and did not decline as is commonly observed using γ-retroviral vectors.^20,26  A rigorous, preclinical assay for stem-cell transduction is to measure marking rates in large animal repopulating cells that differentiate into all blood lineages and persist for the lifetime of the recipient. We were able to follow 2 dogs that received FV vector–transduced mobilized PB cells for 16 and 23 months and observed stable transgene expression in the PB of both dogs. EGFP-expressing cells were detected in all hematopoietic lineages, including red blood cells and platelets. In addition, LAM-PCR analysis demonstrated polyclonal repopulation and the transduction of multipotential repopulating cells. These data strongly suggest that long-term, multipotent repopulating cells were transduced by FV vectors.

Although FV vectors require mitosis for transduction, their persistence as stable transduction intermediates in quiescent G₀ cells²⁸  may in part explain their ability to efficiently transduce long-term repopulating cells. Unlike γ-retroviruses or lentiviruses, FVs undergo reverse transcription in the cell producing the virion rather than the target cell,^15,16  which may contribute to their stability, and thus their capacity for transducing stem cells after a short exposure to vector.

To assess gene transfer into functionally defined precursor cells we determined the percentage of transduced CFCs after transplantation. The overall percentage of transduced CFCs correlated well with the transgene-expression levels in PB and bone marrow leukocytes 1 year after transplantation. gene-transfer levels detected in PB were very close to those determined in bone marrow CD34⁺ cells (Figure 3A). These data suggest that there is no block in the differentiation of transduced cells and no selective elimination of mature gene-modified cells. Comparison of EGFP expression and provirus copy numbers determined by real-time PCR showed that significant silencing did not occur over time. We noted higher transgene expression in lymphocytes relative to granulocytes in both long-term animals when compared with animals that received CD34⁺ cells transduced with lentiviral vectors²⁵  (unpublished observations, H.-P.K., May 2004). The numbers of animals were too small in these 2 studies to observe a statistical difference between FV and lentiviral vectors, but it is interesting to speculate that there may be differences between the stem-cell pools transduced by lentivirus and FV vectors. Additional studies will be needed to determine whether this is true.

Immune responses against the transfer vector or the transgene itself are a concern in gene therapy. Both humoral and cytotoxic lymphocyte responses to gene-modified cells have been reported after transplantation of EGFP-expressing CD34⁺ cells following a non–myeloablative-conditioning regimen.²⁹  We have also encountered immune responses against EGFP/EYFP in baboons after a fully myeloablative-conditioning regimen.²⁹  Additionally, we have observed a decrease in transgene-expressing cells in some dogs that received a transplant with cells transduced by γ-retroviral or lentiviral vectors, suggestive of an immune response against genetically marked cells. In this study we therefore included cyclosporine as an immunosuppressive drug after transplantation to prevent potential immune responses to the gene-modified cells.

The risk of malignant transformation in stably transduced hematopoietic cells by insertional mutagenesis is of great concern because of the development of leukemia in a gene-therapy trial for X-linked SCID.³⁰  The number of proviral copies per cell may be an important consideration in this regard.^31,32  In the current study we found that provirus copy numbers measured by real-time PCR were approximately 2-fold higher than the percentage of transgene-expressing cells measured by flow cytometry, suggesting an average proviral copy number of 2, which is similar to our results with lentivirus and γ-retrovirus vectors in dogs. This is also similar to the copy numbers observed after the transduction of human nonobese diabetic (NOD)/SCID-repopulating cells by FV vectors (1.6-2.6) at a similar MOI (5),¹⁰  demonstrating the efficiency with which FV vectors transduce hematopoietic repopulating cells. For lentiviral vectors, much higher MOIs (approximately 100) were used to obtain similar gene-marking levels in canine²⁵  and NOD/SCID-repopulating cells.³³  Interestingly, for FV vectors we observed similar transduction rates in pretransplantation progenitor cells and in long-term repopulating cells; however, for lentiviral vectors, transduction rates in progenitor cells (49%-81%) were much higher than in long-term repopulating cells (1%-12%).²⁵  Thus, FV vectors may be advantageous in that efficient transduction of true repopulating cells relative to progenitor and mature hematopoietic cells reduces the MOIs necessary for gene transfer and thus may reduce the total number of vector integrations in a transplanted cell population.

In conclusion, we report efficient and reproducible transduction of long-term, multipotent canine repopulating cells using a short overnight transduction protocol with FV vectors. The short transduction protocol should be particularly important when treating diseases in which maintenance of stem cells in culture is an obstacle to successful gene therapy and possibly for transplantations using less-toxic non–myeloablative preparative regimens in which stem cell potential in the graft will be critical.

We thank Michele Spector, DVM, the technicians in the canine facilities of the Fred Hutchinson Cancer Research Center, and the investigators of the Program in Transplantation Biology who participated in the weekend treatments. We thank Drs Rainer Storb, Peter Moore, and Brenda Sandmaier for providing antibodies for subset analyses; Amgen Inc. for providing canine-specific growth factors; and the technicians of the hematology and pathology laboratories of the Fred Hutchinson Cancer Research Center. We also acknowledge the assistance of Bonnie Larson and Helen Crawford in preparing the manuscript.

This work was supported by the National Institutes of Health, Bethesda, MD (grants HL36444, DK47754, HL074162, DK56465, and HL53750).