ARTICLE doi: /mthe , available online at on IDEAL A B C FIG. 1. GFP expression of HIV-1 and FIV transduced h

Transcription

1 doi: /mthe , available online at on IDEAL ARTICLE Expression from Second-Generation Feline Immunodeficiency Virus Vectors Is Impaired in Human Hematopoietic Cells Mary A. Price, 1,* Scott S. Case, 1,* Denise A. Carbonaro, 1 Xiao-Jin Yu, 1 Denise Petersen, 1 Kathleen M. Sabo, 2 Michael A. Curran, 3 Barbara C. Engel, 1 Hovanes Margarian, 1 Janis L. Abkowitz, 2 Garry P. Nolan, 4 Donald B. Kohn, 1 and Gay M. Crooks 1, 1 Division of Research Immunology/Bone Marrow Transplantation, Childrens Hospital Los Angeles, Los Angeles, California 90027, USA 2 Division of Hematology, Department of Medicine, University of Washington, Seattle, Washington 98195, USA 3 Division of Immunology, University of California Berkeley, Berkeley, California 94720, USA 4 Department of Microbiology and Immunology, Stanford University, Stanford, California 94305, USA *These authors contributed equally to this work. To whom correspondence and reprint requests should be addressed. Fax: (323) gcrooks@chla.usc.edu. Vectors based on the feline immunodeficiency virus (FIV) have been developed as an alternative to those based on another lentivirus, human immunodeficiency virus-1 (HIV-1), because of theoretical safety advantages. We compared the efficiency of gene transfer and expression in human and feline hematopoietic progenitors using second-generation HIV-1 and FIV-based vectors. Vector pairs were tested using either human cytomegalovirus or murine phospho-glycerate kinase (PGK) internal promoters and were pseudotyped with the vesicular stomatitis virus G protein (VSV-G). Vector proviral copy numbers were similar in human and feline hematopoietic primary cells and cell lines transduced by HIV-1 or FIV vectors, demonstrating that both vectors are able to transfer genes efficiently to these cell types. HIV-1 vectors were well expressed in human primary hematopoietic cells and cell lines. However, transgene expression from FIV vectors was almost undetectable in human hematopoietic cells. In contrast, the FIV vector was expressed well in primary hematopoietic feline cells and human non-hematopoietic cells, demonstrating that low transgene expression from the FIV vector is a phenomenon specific to human hematopoietic cells. Northern blot analysis demonstrated decreased vector transcript levels in human CEM cells transduced with FIV relative to cells transduced with HIV-1, despite high vector copy numbers. No evidence of vector transcript instability was seen in studies of transduced CEM cells treated with actinomycin D. We conclude that FIV vectors can transfer genes into human hematopoietic cells as effectively as HIV-1 vectors, but that unknown elements in the current FIV backbone inhibit expression from FIV vectors in human hematopoietic cells. Key Words: gene therapy, gene transfer, lentiviral vectors, hematopoietic stem cells, feline immunodeficiency virus, HIV INTRODUCTION Lentiviral vectors have emerged as a more effective gene delivery system than oncoretroviral vectors for many cell types because of the ability of lentiviruses to transduce non-dividing cells and to deliver larger genomic elements without suffering vector instability [1]. However, a major concern about the use of lentiviral vectors is their biosafety, as the prototypic lentiviral vector has been based on human immunodeficiency virus-1 (HIV-1), a cause of a life-threatening progressive immune deficiency in humans. Although successive improvements in HIV-1- based lentiviral vectors have led to deletion of most of the HIV-1 genome, including multiple accessory genes which are essential to viral pathogenesis (for example, nef), concerns still remain about the potential for the vectors to recombine with genes for other HIV-1 proteins and the vesicular stomatitis virus G protein (VSV-G) used during vector packaging to yield a new pathogenic virus. While the probability of such a recombinant is exceedingly low due to lack of sequence homology, the consequences could be dire. Thus, in an effort to develop vectors with less potential for human pathogenesis, vectors have been derived from lentiviruses that are endemic to other species /02 $

2 ARTICLE doi: /mthe , available online at on IDEAL A B C FIG. 1. GFP expression of HIV-1 and FIV transduced human hematopoietic progenitors. (A) CD34+ cells transduced with vectors containing CMV promoters (n = 9). (B) CD34+CD38- cells transduced with vectors containing CMV promoters (n = 4). (C) CD34+ and CD34+CD38- cells transduced with vectors containing PGK promoters (n = 4). Parallel experiments with MND/GALV oncoretroviral vectors are also shown in (B). Filled black square, HIV-1; open circles, FIV; filled triangles, MND/GALV (in B only). and are not known to cause disease in humans, such as feline, simian and equine lentiviruses. Feline immunodeficiency virus (FIV) causes an immune deficiency syndrome in cats similar to that caused by HIV- 1 in humans; however, wild-type FIV is highly species-specific with no evidence of FIV disease or sero-conversion occurring in humans exposed to the virus. The restricted tropism of FIV results from a combination of the weak promoter activity of the FIV LTR [2] and relative inactivity of Rev-mediated nuclear export in non-feline cells [3]. Amino acid sequence analysis suggests that FIV is more closely related to the non-primate lentiviruses (equine infectious anemia virus (EIAV) and visna virus) than to the primate lentiviruses (HIV-1 and simian immune virus (SIV)) [4]. Despite the fact that the native viruses of these species do not infect human cells, recombinants of vectors derived from these species with the VSV-G glycoprotein gene could also lead to novel, potentially pathogenic viruses capable of infecting human cells. Apart from the theoretical differences in the biosafety profiles of the lentiviral vector systems, their relative efficacy at gene transfer and expression in primary human cells will be an important determinant of their utility for clinical gene therapy. Although numerous studies have been published reporting effective gene transfer with each type of vector studied individually, there has only been one publication so far in which lentiviral vectors from different species were compared. O Rourke et al. [5] compared vectors derived from EIAV with vectors derived from HIV- 1. They found EIAV and HIV-1 vectors to be capable of similar gene transfer efficiency, but observed significantly lower expression levels from the EIAV vector, specifically in human hepatocytes. The lower expression was determined to result from greater instability of the EIAV-derived vector transcript in HepG2 cells. In this report, we have carried out an analogous comparison, examining second-generation lentiviral vectors derived from HIV-1 and FIV. Although FIV-based vectors have been shown to transduce and express in a variety of human cell lines and primary cells, including growth arrested and post-mitotic cells [6,7], the utility of FIV as a vector for human hematopoietic progenitors has not been reported. Thus, we have focused our studies on determining the efficiency of transduction and the level of expression of HIV-1- and FIV-based lentiviral vectors in human hematopoietic progenitors. We found that transduction with HIV-1-based vectors consistently resulted in significantly higher percentages of CD34+ and CD34+CD38- cells expressing the egfp reporter gene than transduction with FIV-based vectors, irrespective of viral titer or internal promoter. An analysis of copy number revealed that FIV was highly efficient at gene transfer into all cell types, including primary human hematopoietic progenitors. However, vector transcripts from FIV transgenes were decreased specifically in human (and not feline) hematopoietic cells. These data were supported by similar findings in human and feline hematopoietic cell lines. The identification of sequences within the FIV backbone that suppress vector expression in human hematopoietic cells will be an important step toward bringing FIV vectors to clinical use. RESULTS Transduction of Human Hematopoietic Progenitor Cells with FIV and HIV-1 We and others have demonstrated that HIV-1-based lentiviral vectors are able to efficiently transduce human hematopoietic progenitor cells after a relatively brief period of exposure to vector and with little or no cytokine 646

3 doi: /mthe , available online at on IDEAL ARTICLE A B FIG. 2. Transgene expression of HIV-1 and FIV vectors in human and feline cell lines. Identical cell numbers of each cell line were transduced with HIV-1 (black bars) or FIV (white bars) expressing GFP with either CMV (A) or PGK (B) internal promoters. Human (293, HEL, and CEM) and feline (FEA and 3201B) cell lines were transduced with supernatants also titered on murine 3T3 fibroblasts. The 3T3 relative units were calculated by dividing the titer on each cell line with the titer achieved on 3T3 cells. Numbers in the box below the graph show the ratio in each cell line of FIV titer:hiv-1 titer. stimulation [8]. To assess whether vectors based on FIV, another type of lentivirus, resulted in similar levels of transduction, we began by exposing cord blood CD34+ cells to FIV and HIV-1 vectors under identical experimental conditions. Second-generation, FIV-based (FELIX) vectors, pseudotyped with VSV envelope, were used as these had been shown previously to efficiently transduce other human cell types [6]. All vectors contained the marker gene GFP allowing transgene expression in transduced target cells to be measured by FACS analysis. To control for possible, unpredicted differences in internal promoter activity in each vector, two pairs of HIV-1 and FIV vectors were examined. In each case, GFP expression was driven by either the human CMV promoter or the murine PGK promoter. Vector supernatants were initially titered on the human kidney cell line 293, and then diluted as necessary to achieve equivalent titers of HIV-1 and FIV for each experiment. For each vector, a final titer of 1 to infectious units/ml was used at an MOI of Transduction of cord blood progenitors with HIV-1- based vectors consistently resulted in a significantly higher percentage of cells expressing GFP than those cells transduced with FIV-based vectors (Fig. 1). This result was seen irrespective of the internal promoter used in the vectors. The percentage of GFP-expressing cells was higher using HIV-1 vectors than FIV-1 vectors in every experiment whether the CMV promoter (Figs. 1A and 1B; P < , n = 13 combined data) or the PGK promoter (Fig. 1C; P = 0.04, n = 4) was used in the vectors. To determine if the low GFP expression with FIV could be overcome by increasing the concentration of infectious vector particles, human CD34+ cells were transduced using vector concentrations ranging from to infectious units/ml. Less than 2% of cells transduced with FIV expressed GFP, irrespective of vector concentration. Thus, over a 30-fold range of concentration of FIV, GFP expression from FIV remained almost undetectable. In contrast, cells transduced in parallel with HIV-1 using the lowest concentration ( infectious units/ml) produced approximately 20% GFP marking (data not shown). Transduction of Feline Bone Marrow with FIV and HIV-1 To determine whether the FIV vectors were unable to efficiently transduce primary hematopoietic cells in general or whether the problem was species-specific, bone marrow cells from two cats were transduced with HIV-1 and FIV in parallel, again using identical transduction conditions and viral titers. In these experiments, the murine PGK promoter was used to drive GFP expression in the HIV-1 and FIV vectors, as we have previously noted that the CMV promoter functions poorly in feline cells transduced with lentiviral vectors (J.L.A., unpublished data). In contrast to the human bone marrow studies, high levels of GFP expression were achieved in feline cells transduced with both HIV-1 (29.5 ± 2.5%) and FIV (22.5 ± 0.5%; n = 2). Similar results were seen by analysis of CFU plated immediately after transduction (data not shown). Thus, FIV vectors were able to efficiently transduce and express the transgene in primary feline hematopoietic progenitors. Transduction of Human and Feline Cell Lines We next explored whether the differences in lentiviral vector transgene expression in feline and human cells existed only in primary hematopoietic cells or whether other, non-hematopoietic cell types showed a similar pattern. A variety of hematopoietic and non-hematopoietic human and feline cell lines were transduced in parallel with HIV-1 and FIV vectors. To standardize vector titer for HIV-1 and FIV supernatants, we determined the titer from the aliquots of each supernatant on the murine 647

4 ARTICLE doi: /mthe , available online at on IDEAL TABLE 1: Relationship of GFP expression to copy number Ratio Ratio Human %GFP %GFP Copy no./cell Copy no./cell %GFP:copy no. %GFP:copy no. cell type HIV FIV HIV FIV HIV FIV CD34+ n = ± ± ± ± ± ± 0.02 CD34 + CD38- n = ± ± ± ± ± ± :5000 dilution* :50,000 dilution* The %GFP was measured by FACS, copy number by real-time PCR. There was a significant difference in the ratio of %GFP:copy number between HIV and FIV transductions of primary human progenitors (CD34+ CD38- cells) (P = ). *Titers of undiluted vector were (HIV) and (FIV). fibroblast cell line, NIH-3T3. Similar titers were obtained on NIH-3T3 cells for HIV-1 and FIV vectors. Transduction of human and feline cell lines revealed a different pattern than the previous experiments conducted on primary human hematopoietic cells (Fig. 2). Transgene expression was similar for both HIV-1 and FIV in the nonhematopoietic human cell line 293, a human kidney epithelial line. However, in two human hematopoietic lines, HEL (an erythroleukemia line) and CEM (a T cell line), GFP expression from HIV-1 was higher than that seen with FIV. The two feline cell lines examined showed a reciprocal pattern of expression. FEA (a feline embryonic fibroblast line) expressed high levels of GFP with both HIV-1 and FIV. However, 3201B (a feline T lymphocyte line) expressed GFP well only when transduced with FIV. Thus, both HIV-1 and FIV vectors transduced non-hematopoietic cell lines from both species, but there was a trend suggesting that HIV-1 was optimal for transduction of human hematopoietic cell lines and FIV, for feline hematopoietic cell lines. Is the Problem Gene Transfer or Gene Expression in Primary Human Hematopoietic Cells? To study whether the low GFP expression in primary human hematopoietic progenitors was the result of low gene transfer or low gene expression, vector copy number was analyzed in cells grown from CD34+ and CD34+CD38- cells by real-time PCR. At all time points in long-term culture, copy number in cells transduced with FIV was equivalent to or higher than those transduced with HIV-1, despite the consistently low GFP expression from the FIV vector (Table 1). Thus, FIV vectors were able to transduce CD34+ and CD34+CD38- cells with at least equal efficiency to HIV-1 vectors, but were unable to express in transduced progenitors. Copy number was also assessed in 293 cells using aliquots of the same vector supernatants as was used for the hematopoietic progenitors. As previously shown, 293 cells showed similar levels of GFP expression with FIV and HIV-1 transduction, and expression correlated well with increasing titer (Table 1). Consistent with these results, copy number revealed high levels of gene transfer from both vectors that also increased with titer (Table 1). To test whether the expression problems seen with FIV occurred specifically with primary human hematopoietic cells or occurred with primary human cells in general, transgene expression in primary human bone marrow stroma was analyzed. Primary bone marrow stroma revealed a pattern of transduction and expression more similar to 293 cells than to primary CD34+ hematopoietic progenitors. High levels of GFP expression were seen in stroma with both HIV-1 and FIV vectors (mean 85% versus 75% respectively) and copy numbers were similarly high with each vector (mean 7.5 and 3.5 copies per cell respectively). Thus, FIV vectors exhibited an inability to express specifically in human hematopoietic progenitor cells. The relationship between transduction and expression for each vector was demonstrated by expressing the ratio of percentage of GFP-positive cells to copy number in human hematopoietic progenitors (CD34+ and CD34+CD38- cells) and in 293 cells (Table 1). HIV-1 transduction resulted in a high ratio of expressing cells per copy number in 293 cells and in primary hematopoietic cells. In contrast, FIV transduction resulted in a high ratio of expressing cells per copy number in 293 cells; however, when analyzing human hematopoietic progenitors, the ratio of expressing cells per copy number was significantly lower with FIV than with HIV-1 (P = ). Northern blots performed on HIV-1 and FIV transduced CD34+ cells confirmed the findings obtained with FACS analysis of expression. Despite the higher copy number with the FIV than the HIV-1 vector, transcripts from the internal CMV promoter were barely detectable in cells transduced with FIV (Fig. 3). However, this was not the case for 293 cells transduced with FIV, which showed equivalent transcript levels from both vectors (data not shown), suggesting the low transgene expression occurs at an RNA level specifically in human hematopoietic cells. 648

5 doi: /mthe , available online at on IDEAL ARTICLE FIG. 3. Vector copy number and RNA expression in CD34+ cells. Shown is a northern blot of CD34+ cells analyzed 3 weeks after separate transductions (in duplicate or triplicate) with infectious units/ml of HIV-1 (lanes 1 and 2, HIV-1 transcript 950 bp) and FIV (lanes 3 5, FIV transcript 1500 bp) vectors (both using CMV promoters) and non-transduced NT cells (lanes 6 and 7). Copy number for each sample determined by real time PCR is shown under the corresponding lane. RNA Stability Low transgene expression can be due either to a transcriptional block or to decreased RNA stability. To determine which of these possibilities existed, RNA stability was assessed in the T cell line CEM, which had previously demonstrated higher expression levels with HIV-1 than FIV vectors (Fig. 2). In cells transduced with FIV, the internal CMV transcript was much less abundant than the CMV transcript in progenitor cells transduced with HIV (Figs. 3 and 4). When CEM cells were subjected to actinomycin D, the CMV transcript of FIV was just as stable (or more so) as the CMV transcript of HIV-1 (half-life 4 hours; Fig. 4). Thus it seems that the inability of FIV to express in human hematopoietic cells is not a function of transcript instability. DISCUSSION Lentiviral vectors offer a significant advance in the quest for efficient transduction of human hematopoietic stem cells because they can transduce nondividing cells (most primitive, pluripotent human hematopoietic stem cells are quiescent). Lentiviruses contain nuclear localization signals (NLS) in proteins of the pre-integration complex that allow entry of vector into non-dividing cells. This has been best studied in HIV-1, with evidence that NLS proteins are encoded by GAG matrix [9], VPR [10], and integrase [11] proteins. Poeschla et al. described transduction by FIV-based vectors of dividing and arrested human cell lines including 293 cells (a human kidney epithelial cell line) and primary cells [7]. High FIV expression was found in primary human macrophages and postmitotic human neurons. However, primary hematopoietic progenitor cells were not studied. A wide variation in titer for each cell line was noted: high levels in HeLa (human cervical adenocarcinoma), 293, RD (human rhabdomyosarcoma), and U87MG (human glioblastoma); medium levels in HOS (human osteosarcoma); and low levels in WI-38 (normal human lung fibroblasts). Curran et al. [6] demonstrated that third generation FIV vectors packaged without accessory proteins can transduce non-dividing primary human hepatocytes, aortic smooth muscle cells, dendritic cells, and cell lines. High-level transgene expression in murine liver was also accomplished using intravenous infusion of FIV vector [12]. Thus, effective transduction and expression in many non-hematopoietic cell types can be accomplished with FIV vectors. Data on FIV transduction in primary human hematopoietic progenitors, however, are lacking. Here we have described an expression defect from FIV vectors specific to human hematopoietic cells. The low GFP expression with FIV relative to HIV-1 vectors was seen in cells generated from both total CD34+ cells, which represent a heterogeneous population of progenitor cells, and CD34+CD38- cells, a population of cells greatly enriched for more primitive and quiescent progenitor cells. Low GFP expression in human CD34+ cells by FIV was not merely due to a lower concentration of infectious virus particles [13], because increasing the viral titer of the FIV supernatant more than 30-fold did not increase the percentage of GFP-expressing cells. In contrast to the human hematopoietic cells, feline bone marrow cells expressed equivalent levels of GFP with HIV-1 and FIV vectors. The suggestion of species-restricted efficacy was further confirmed in studies that compared transduction efficiency of feline and human hematopoietic cell lines. The HIV-1 vector was two to three times more effective than FIV in the human hematopoietic lines and the FIV vector was three to four times more effective than HIV-1 in the feline hematopoietic cell line. In human non-hematopoietic cells, the FIV vector performed equivalently to the HIV-1 vector. FIG 4. Analysis of RNA stability. Shown is a northern blot of CEM cells transduced with HIV-1 (H), FIV (F), or Mock (that is, not transduced) (M) and treated with actinomycin D to prevent further transcription. RNA was extracted at the times shown after actinomycin treatment. Shown is one of two identical experiments. 649

6 ARTICLE doi: /mthe , available online at on IDEAL When the mechanism for the significantly different proportion of human hematopoietic cells that expressed GFP from the FIV and HIV-1 vectors was explored, we found that the levels of gene transfer were essentially identical (or even higher with FIV). Equivalent proviral copy numbers of the HIV-1 and FIV vectors were seen by realtime PCR analysis from multiple samples of transduced cells. Southern blot analysis of selected sets also showed similar findings as well as demonstrated vector stability (data not shown). These findings demonstrate that both vectors perform the steps between entry through proviral integration with similar efficiency. Thus, gene expression, rather than gene transfer, was implicated as the basis for the higher percentage of human cells that expressed GFP after transduction by HIV-1-based vectors compared with the FIV-based vector. The more effective gene expression by the HIV-1-based vector in human cells was clearly demonstrated by northern blot analysis. In samples which had higher proviral copy numbers of the FIV than the HIV-1 vector (based upon real-time PCR), the level of vector transcripts from the HIV-1-based vector consistently exceeded that from the FIV vector. The reasons for the poorer expression by FIV-based vectors in human cells remain to be fully elucidated. The LTR of both HIV-1 (in the absence of Tat) and FIV vectors would be expected to have minimal expression activity in cells without the viral-specific transcriptional activator proteins [14]. The U3 promoter of FIV has been reported to have low activity in human cells [6,7]. In accord with this prediction, northern blot analysis of human hematopoietic cells transduced by either the HIV-1 or the FIV vectors showed essentially no LTR-driven transcripts. The internal expression units were identical in the pairs of FIV and HIV-1 vectors, using either the human CMV or the murine PGK promoters to drive expression of the GFP reporter gene. The level of GFP expression and of transcripts from the FIV vector were significantly lower than those from the HIV-1 vector in primary human hematopoietic cells, as well as some human hematopoietic and T lymphoid cell lines. Poor expression from FIV vectors in human hematopoietic cells was seen irrespective of the internal promoter used. In contrast, essentially identical levels of GFP expression and vector transcripts were seen from the CMV promoter of HIV-1 and FIV vectors in 293T cells and human marrow stromal cells. Thus, the impaired expression by FIV vectors was restricted to human hematopoietic cells. The mechanism for the poorer expression from the FIV vector in human hematopoietic cells appears to occur at the transcriptional level. The northern blot analysis clearly showed significantly lower steady-state levels of transcripts from the internal CMV promoter of the FIV vector. The stability of the vector transcripts were similar in CEM cells treated with actinomycin D. These findings implicate a lowered transcriptional rate as the cause of the lower levels of expression by the FIV vector. Potentially, sequences of the FIV vector backbone may inhibit expression from the CMV and PGK promoters in hematopoietic cells. Inhibitory transcriptional proteins present in human hematopoietic cells may recognize these sequences and induce cis inhibition of the internal promoter. Presumably, the feline hematopoietic cells do not have a homologous protein that interacts with the FIV vector. Recent studies suggest that FIV-based vectors with enhancer/promoter deletions of the LTR (SIN vectors) express better in human cells, implicating the presence of expression suppressive elements within those sites (G.P.N. and M.A.C., unpublished data). Recently, O Rourke et al. have reported a similar phenomenon of impaired expression by a non-human lentiviral vector specifically in human cells [5]. A vector derived from EIAV was able to produce stable transduction of several different human cell lines at levels equivalent to an HIV-1-based lentiviral vector. However, transgene expression was significantly lower in the human cell lines transduced with EIAV than with the HIV-1 vector. Transduction of a horse fibroblast cell line with EIAV vectors resulted in efficient and stable expression. The mechanism by which the EIAV vectors were silenced in the human cells was not elucidated. However, in contrast to our own results with FIV, decreased stability of the EIAV vector-derived transcripts was noted. Thus, there may be multiple mechanisms by which transduction and expression by lentiviral vectors may be limited in species-specific ways, including cis inhibition of transcription and impaired polyadenylation. Because human hematopoietic stem cells represent a key target for clinical gene therapy applications, these observations suggest that HIV-1 based vectors would be preferable to FIV-based vectors in terms of gene expression capabilities. Further studies are needed to define the specific basis for the deficiency of expression from FIV-based vectors. Additional studies are also needed to increase understanding of the relative biosafety of these two systems. Ultimately, safe and effective lentiviral vectors may lead to therapeutic applications of gene transfer for a wide range of genetic diseases of neurological, musculoskeletal, cardiovascular, and pulmonary systems, as well as the hematopoietic system. METHODS Cell sources and isolation. Human umbilical cord blood was collected into centrifuge tubes containing citrate-phosphate-dextrose (Sigma, St. Louis, MO) according to a protocol approved by the Institutional Review Board (the Committee on Clinical Investigations, CCI) at Childrens Hospital Los Angeles. Mononuclear cells were isolated by density centrifugation using Ficoll-Hypaque (Amersham Pharmacia Biotech, Piscataway, NJ) and incubated with antibodies for CD34+ cell enrichment using the MiniMACS magnetic separation system (Miltenyi Biotec, Auburn, CA). CD34+CD38- cells were identified in the enriched CD34+ populations by incubating with anti-cd34-fitc (HPCA2; Becton Dickinson Immunocytometry Systems (BDIS), San Jose, CA) and anti-cd38-pe (leu 17, BDIS), and isolated using a FACSVantage flow cytometer (BDIS) equipped 650

7 doi: /mthe , available online at on IDEAL ARTICLE with an argon laser using regions as described [8]. Isotype control antibodies were used as follows: for 10 6 cells in 100 l of phosphate buffered saline (PBS, Bio-Whittaker, Walkersville, MD), 20 l IgG1-phycoerythrin (PE) (BDIS), 20 l IgG1-Fluorescein isothiocyanate (FITC) (BDIS). Quadrants were set to include at least 97% of the isotype-negative cells. Human bone marrow stromal cells were cultured by plating whole human bone marrow cells into a T75 flask (Fisher Scientific, Pittsburgh, PA) in human stromal cell medium (Iscove s Modified Dulbecco s Medium (IMDM, Bio-Whittaker), 15% fetal calf serum (FCS), 15% horse serum (HS), 2-mercaptoethanol (2 M-E, Sigma), 10 6 mol/l hydrocortisone (Sigma), penicillin/streptomycin, and glutamine). Non-adherent cells were removed the following day. The remaining adherent stromal cells were passaged at least four times before they were used in long-term bone marrow cultures to support transduction. Feline bone marrow was obtained from the humerus or femur of domestic cats following ketamine anesthesia, using the protocol approved by the IACUC at the University of Washington. Cells were collected in preservative-free heparin (American Pharmaceutical Partners, Los Angeles, CA). Mononuclear cells were then isolated by density (< g/ml) centrifugation using Percoll (Sigma). Very low density (< g/ml) mononuclear cells, known to contain feline progenitor and stem cells, were obtained by a second density gradient centrifugation to enrich hematopoietic stem cells approximately 50- to 100-fold as defined in limiting dilution autologous transplantation studies [16]. This enrichment is comparable to the CD34 selection of human mononuclear bone marrow cells. Human cell lines were obtained from ATCC and were as follows: 293 cells (a human kidney epithelial cell line transformed with Adenovirus 5), HEL (an erythroleukemia line), and CEM (a T cell line). Feline cell lines, FEA (a feline embryonic fibroblast line) and 3201B (a feline T lymphocyte line) were provided by the laboratory of J.L.A. Lentiviral constructs. The HIV-1-based vector used was the phr vector, which contains the intact HIV-1 LTR and has been previously described [17]. The FIV-based vector was the second-generation vector pflx- CPLksFELIX [18]. Expression of the egfp reporter gene (BD Biosciences/Clontech, Palo Alto, CA) was driven from internal promoters from either the early immediate enhancer/promoter of human cytomegalovirus (CMV) or the murine phosphoglycerokinase (PGK) gene. Lentiviral vectors were packaged by co-transfection of 293T cells, as described [8]. In brief, for each T cells per 100 mm 3 dish, 10 g of either HIV-1 or FIV vector plasmid was transfected with the appropriate packaging plasmid (10 g of pcmv_r8.91 was used to express HIV-1 gag, pol, tat, and rev proteins without the accessory genes vif, vpu, vpr, and nef [8,17]; 6 g of pcpr_env for FIV [6]) and envelope plasmid (both FIV and HIV-1 vectors were pseudotyped using 2 g of the VSV-G expression plasmid, pmd.g [17]). The viral supernatant was collected and virus particles were concentrated by ultracentrifugation as described [8]. Vector titers were determined by measuring the frequency of target cells expressing egfp as measured by FACS analysis after transduction with serial dilutions of the vector samples [13]. In some experiments vector titers were also determined on other cell lines (feline and murine fibroblasts). Transduction conditions. Cord blood CD34+ and CD34+CD38- cells were isolated as above, washed, and resuspended in cytokine containing X-Vivo 15 serum-free medium (Bio-Whittaker, Walkersville, MD; cytokines at 2 concentration). CD34+CD38- cells (0.2 to ) or CD34+ cells (0.25 to ) were plated into either 96-well or 24-well plates (Falcon, BDIS) previously coated with fibronectin (retronectin, fraction CH296, Takara Shuzo Co., Otsu Shiga, Japan) for 1 hour at room temperature, then blocked with 2% bovine serum albumin (BSA) for 30 minutes at room temperature. Viral supernatants were thawed at room temperature for 2 3 hours vortexing every 30 minutes, and added 1:1 to the medium containing the progenitor cells. The final concentration of cytokines for each well was 5 ng/ml interleukin-3 (IL- 3, R&D, Minneapolis, MN), 10 ng/ml interleukin-6 (IL-6, Sandoz Pharmaceuticals Corp, East Hanover, NJ), and 25 ng/ml c-kit ligand (KL, R&D, Minneapolis, MN). Based on titers performed on 293 cells before freezing, a final concentration of infectious particles/ml of FIV or HIV supernatant was used unless otherwise stated. After overnight incubation with virus, the cells were washed three times with PBS. Mock (non-transduced) cells were cultured in identical conditions, but without virus addition. A similar method was used to transduce feline bone marrow progenitor cells. Very low density (< g/ml) marrow cells ( ) were suspended in IMDM medium (Gibco/BRL) with 20% heat inactivated FCS (Summit Biotechnology, Fort Collins, CO), 1% BSA, 2 ng/ml recombinant canine stem cell factor (a gift from Amgen), 10 4 M -mercaptoethanol, 1 U/ml recombinant human erythropoietin, and penicillin/streptomycin/amphotericin and placed into fibronectin-coated plates. Viral supernatant was added 1:1 to medium and cells to make a final volume of 750 l. After 12 hours, cells were washed and resuspended in 750 l fresh medium and cultured for 7 days in a 5% CO 2 humidified atmosphere at 37 C, then subjected to flow cytometry analysis to assess GFP expression. Before culture, some cells were removed and placed in triplicate methylcellulose culture to assay progenitors [16]. After days, individual colonies were counted and scored for GFP positivity. Additionally, colonies were plucked and analyzed for GFP fluorescence by flow cytometry. Long-term bone marrow cultures. After transduction, progenitors were cultured on irradiated (2000 cgy) allogeneic human bone marrow stroma in long-term bone marrow culture (LTBMC) medium (IMDM, 20% FCS, 0.5% bovine serum albumin (BSA; Sigma), 2-mercaptoethanol, 10 6 mol/l hydrocortisone (Sigma), penicillin/streptomycin, and glutamine) in 5 ng/ml IL-3, 10 ng/ml IL-6, and 25 ng/ml KL. Cultures were fed by half medium replacements twice weekly. Analysis of GFP expression. Non-adherent cells were harvested from longterm bone marrow cultures every 1 3 weeks, and analyzed on a FACSCalibur (BDIS) using an argon laser. The non-transduced population was used as a negative control to set gates for negative and positive GFP cells. Molecular analysis of proviral copy number and gene expression. Genomic DNA extraction for real time PCR was performed using the Puregene DNA Isolation Kit (Gentra Systems, Minneapolis, MN). The standard for real-time PCR and Southern blotting was a 293 clone transduced with HIV-1 vector and demonstrated by Southern analysis to contain one copy per cell of egfp. The standard curve was made by serially diluting the one copy per cell standard DNA into non-transduced 293 cells. Equal amounts of DNA ( ng) were amplified in triplicate for both the standard curve and samples. The probe concentration was 200 nm and primer concentration was 400 nm. All samples were amplified in Universal Master Mix (Applied Biosystems) at a final volume of 50 l. The primer sequences for detection of GFP were as follows: 5 -CTGCTGCCCGACAACCA-3 and 5 -GAACTCCAGCAGGACCATGTG-3, the probe sequence was 5-6FAM- CCCTGAGCAAAGACCCCAACGAGA-TAMRA-3. The reaction conditions wereas follows: stage 1, 50 C for 2 minutes for 1 cycle; stage 2, 95 C for 10 minutes for 1 cycle; stage 3, 95 C for 15 seconds, then 60 C for 1 minute for 40 cycles. -Actin was used as a loading control. In lieu of multiplexing the -actin primers and probe in the same reaction, the template DNA was added to each well for a final volume of 19 l and from which 1.5 l was transferred to a separate well on the same plate. This well was amplified using undiluted primers and probe for -actin (both from Applied Biosystems, Foster City, CA). All samples were run on the ABI Prism 7700 Sequence Detection System, either immediately or after storage at 20 C. Northern analysis. Total RNA was extracted from transduced (mock, HIV, or FIV) 293 (human embryonic kidney) cells or CD34+ (cord blood derived) cells using the RNAqueous RNA isolation kit by Ambion (San Antonio, TX). The procedure for northern blot analysis has been described [14]. For each sample, 10 g of RNA was electrophoresed under denaturing conditions (1.2% agarose formaldehyde gel) for 20 hours (35 volts), denatured, neutralized, and transferred by capillary transfer to a nylon membrane (Osmonics, Westborough, MA). The RNA was cross-linked to the membrane, which was hybridized overnight with a [ 32 P]dCTP radiolabeled (random primed; Prime-It Kit, Stratagene, La Jolla, CA) full-length probe to egfp, and exposed to autoradiography film (X-OMAT film, Kodak). The membrane was stripped and rehybridized with a radiolabeled human actin cdna probe. The expected size of the internally derived HIV transcript is 950 bp and the FIV transcript is 1500 bp. RNA extraction and message instability. To evaluate message stability, total RNA was analyzed by northern blot from transduced cells treated 651

8 ARTICLE doi: /mthe , available online at on IDEAL with actinomycin D. CEM cells ( cells/ml) were transduced overnight with either HIV-1-based vector or FIV-based vector, or in R10 medium alone (RPMI (Irvine Scientific, Santa Ana, CA), 10% FCS, penicillin/streptomycin, and glutamine); HIV-1 and FIV vectors were present at a concentration of i.u/ml. The following morning, the cells were washed three times and plated in R10 medium. After 2 weeks, the transduced CEM cells were plated into six flasks at a concentration of 10 6 cells/ml in 15 ml of R10 with 10 g/ml actinomycin D (Sigma-Aldrich, St Louis, MO). Over the next 12 hours, the cells were harvested at 2-hour intervals, pelleted, treated with RNAstat 60 (Tel-TEST, Inc., Friendswood, TX), and frozen at 80 C for use the following day in northern blot analysis of GFP transcripts. ACKNOWLEDGMENTS We thank Lora W. Barsky, Dennis Haas, and Jason Zhang (Childrens Hospital Los Angeles) for technical assistance and advice. We are very grateful to the Labor and Delivery staff of Kaiser Permanente Sunset for their generous and vital assistance in the collection of cord blood for these studies. Supported by NIH grants RO1DK54567and P50HL54850 (G.M.C.), P01CA59318 (G.M.C. and D.B.K.), HL31823 (J.L.A.), American Cancer Society PF CCE (M.A.C.), and Martell Foundation (G.M.C and D.B.K.). G.M.C. is a Scholar of the Leukemia and Lymphoma Society. D.B.K. is an Elizabeth Glaser Scientist from the Pediatric AIDS Foundation, and is the recipient of a Distinguished Clinical Scientist Award from the Doris Duke Charitable Foundation. G.P.N. was supported in this work as a Scholar of the Leukemia Society of America, a Howard Hughes Young Faculty Award, and a Burrough s Wellcome New Investigator in Pharmacology Award, and was supported by National Institutes of Health grants P01-AI39646, AI35304, N01-AR , and A1/GF , and the Juvenile Diabetes Foundation. RECEIVED FOR PUBLICATION AUGUST 12; ACCEPTED SEPTEMBER 25, REFERENCES 1. Kohn, D. B., Carter, B., and Glorioso, J. (2002). Gene therapy and viruses. In: Clinical Virology (D. D. Richman, R. J. Whitley, and F. G. Hayden, Eds). A.S.M. Press, Washington, DC. 2. Miyazawa, T., et al. (1992). Production of feline immunodeficiency virus in feline and non-feline non-lymphoid cell lines transfection of an infectious molecular clone. J. Gen. Virol. 73: Tomonaga, K., et al. (1994). Comparison of the Rev transactivation of feline immunodeficiency virus in feline and non-feline cell lines. J. Vet. Med. Sci. 56: Olmsted, R. A., Hirsch, V. M., Purcell, R. H., and Johnson, P. R. (1989). Nucleotide sequence analysis of feline immunodeficiency virus: Genome organization and relationship to other lentiviruses. Proc. Natl. Acad. Sci. USA 86: O Rourke, J. P., Newbound, G. C., Kohn, D. B., Olsen, J. C., and Bunnell, B. A. (2002). Comparison of gene transfer efficiencies and gene expression levels achieved with equine infectious anemia virus- and Human Immunodeficiency Virus Type 1-derived lentivirus vectors. J. Virol. 76: Curran, M. A., Kaiser, S. M., Achacoso, P. L., and Nolan, G. P. (2000). Efficient transduction of nondividing cells by optimized feline immunodeficiency virus vectors. Mol. Ther. 1: Poeschla, E. M., Wong-Staal, F., and Looney, D. J. (1998). Efficient transduction of nondividing human cells by feline immunodeficiency virus lentiviral vectors. Nat. Med. 4: Case, S. S., et al. (1999). Stable transduction of quiescent CD34+CD38- human hematopoietic cells by HIV-1 lentiviral vectors. Proc. Natl. Acad. Sci. USA 96: Bukrinsky, M. I., et al. (1993). A nuclear localization signal within HIV-1 matrix protein that governs infection of non-dividing cells. Nature 365: Heinzinger, N. K., et al. (1994). The Vpr protein of human immunodeficiency virus type 1 influences nuclear localization of viral nucleic acids in nondividing host cells. Proc. Natl. Acad. Sci. USA 91: Gallay, P., Hope, T., Chin, D., and Trono, D. (1997). HIV-1 infection of nondividing cells through the recognition of integrase by the importin/karyopherin pathway. Proc. Natl. Acad. Sci. USA 94: Stein, C. S., et al. (2001). In vivo treatment of hemophilia A and mucopolysaccharidosis type VII using nonprimate lentiviral vectors. Mol. Ther. 3: Haas, D. L., Case, S. S., Crooks, G. M., and Kohn, D. B. (2000). Critical factors influencing stable transduction of human CD34 + cells with HIV-1- derived lentiviral vectors. Mol. Ther. 2: Miyazawa, T., Kawaguchi, Y., Kohmoto, M., Tomonaga, K., and Mikami, T. (1994). Comparative functional analysis of the various lentivirus long terminal repeats in human colon carcinoma cell line (SW480 cells) and feline renal cell line (CRFK cells). J. Vet. Med. Sci. 56: Hao, Q.-L., Shah, A. J., Thiemann, F. T., Smogorzewska, E. M., and Crooks, G. M. (1995). A functional comparison of CD34+CD38- cells in cord blood and bone marrow. Blood 86: Abkowitz, J. L., Taboada, M. R., Sabo, K. M., and Shelton, G. H. (1998). The ex vivo expansion of feline marrow cells leads to increased numbers of BFU-E and CFU-GM but a loss of reconstituting ability. Stem Cells 16: Naldini, L., et al. (1996). In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector. Science 272: Curran, M. A., and Nolan, G. P. (2002). Recombinant feline immunodeficiency virus vectors. Preparation and use. Methods Mol. Med. 69: Robbins, P. B., et al. (1997). Increased probability of expression from modified retroviral vectors in embryonal stem cells and embryonal carcinoma cells. J. Virol. 71: