Genetically modified CD34 cells as cellular vehicles for gene ... - Nature

Gene Therapy (2000) 7, 43–52  2000 Macmillan Publishers Ltd All rights reserved 0969-7128/00 $15.00 www.nature.com/gt

CELL-BASED THERAPY

RESEARCH ARTICLE

Genetically modified CD34+ cells as cellular vehicles for gene delivery into areas of angiogenesis in a rhesus model J Go´mez-Navarro1, JL Contreras2, W Arafat1, XL Jiang2, D Krisky3, T Oligino3, P Marconi3, B Hubbard2, JC Glorioso3, DT Curiel1 and JM Thomas2 1

Gene Therapy Center, Birmingham; 2Transplant Immunobiology, Department of Surgery, Birmingham, AL; and 3Department of Molecular Genetics and Biochemistry, E1240 Biomedical Science Tower, Pittsburgh, PA 15261, USA

To develop a cellular vehicle able to reach systemically disseminated areas of angiogenesis, we sought to exploit the natural tropism of circulating endothelial progenitor cells (EPCs). Primate CD34+ EPCs were genetically modified with high efficiency and minimal toxicity using a non-replicative herpes virus vector. These EPCs localized in a skin autograft model of angiogenesis in rhesus monkeys, and sustained the expression of a reporter gene for several weeks while circulating in the blood. In animals infused with autologous CD34+ EPCs transduced with a thymidine kinase-encoding herpes virus, skin autografts and subcutaneous Matrigel pel-

lets impregnated with vascular growth factors underwent necrosis or accelerated regression after administration of ganciclovir. Importantly, the whole intervention was perfectly well tolerated. The accessibility, easy manipulation, lack of immunogenicity of the autologous CD34+ cell vehicles, and tropism for areas of angiogenesis render autologous CD34+ circulating endothelial progenitors as ideal candidates for exploration of their use as cellular vehicles when systemic gene delivery to those areas is required. Gene Therapy (2000) 7, 43–52.

Keywords: rhesus monkey; CD34 cell; gene transfer; herpes simplex virus; herpes simplex virus thymidine kinase; anti-angiogenesis

Introduction A variety of strategies has been developed to accomplish gene therapy for cancer and other systemic diseases and conditions characterized by the development of angiogenesis, including inflammatory diseases and organ engrafting.1–4 Fundamental to the clinical implementation of these strategies is the development of vector systems that allow direct in vivo gene transfer to disseminated loci of target tissues. To this end, both viral and non-viral vectors have been employed.5–8 In general, despite their utility in selected contexts, currently available viral vectors and liposomes have been quite limited in their ability to accomplish highly efficient gene delivery to relevant target cells in vivo. Most importantly, the lack of vectors with the capacity to achieve cell-specific targeting has broadly restricted the implementation of gene therapy in many disease contexts, and notably in disseminated malignant tumors. The innate and adaptive immune responses against viral vectors and accompanying inflammation have also presented formidable barriers, even in the neoplastic disease context, for significant gene delivery and expression, and thus for successful clinical translation.5,9–11 Thus, shortcomings in the current generCorrespondence: JM Thomas, Transplant Immunobiology, Department of Surgery, 1808 Seventh Avenue South, BDB 563, Birmingham, AL 35294, USA Received 19 July 1999; accepted 9 August 1999

ation of vectors have limited the overall efficacy of gene therapy approaches. As an alternative, cells have also been employed as vectors. In this approach, therapeutic genes are transferred to isolated cells ex vivo and the transduced cells are reimplanted in vivo. In this manner, the genetically modified cell becomes itself the ultimate vector for gene delivery.12–14 A variety of genes can conceivably be introduced in candidate cell vehicles, and expression of these gene payloads can be obtained in the ultimate environment where these cells finally localize. A loco-regional effect subsequent to the expression of the therapeutic gene can thus be achieved in areas otherwise hardly accessible to gene transfer. For example, we have previously shown that human umbilical vein endothelial cells (HUVECs) genetically modified with an adenovirus encoding HSV thymidine kinase (tk) may be used in vivo as cellular vehicles for the induction of a cytotoxic bystander effect on tumor cells.15 Cellular vehicles offer numerous advantages as vectors and have had therefore historical precedence. However, stringent criteria, largely unmet, are required for developing a clinically feasible cellular vehicle vector system: (1) simple and non-invasive harvest from the patient; (2) non-deleterious manipulation and/or expansion in culture; (3) high levels of cell transduction with available vectors; (4) low risk and easy reimplantation; and (5) adequate cellular tropism. In the context of a paucity of candidate cells that could meet the stringent

Endothelial progenitors for systemic gene delivery J Gomez-Navarro et al

44

requirements mentioned above for a clinically useful cellular vehicle, it has recently been reported that the CD34+ leukocyte fraction of human peripheral blood contains putative endothelial progenitors which originate in the bone marrow and localize in areas of post-ischemic angiogenesis.16 Additional evidence from a canine model confirmed that a subset of donor bone marrow cells can be mobilized to the peripheral circulation and colonize endothelial flow surfaces.17 In both reports, purified CD34+ cells were shown to differentiate in vitro into mature endothelium.16 In addition to these studies based on animal models of transplantation of CD34+ cells, a variety of other experimental observations support the intimate role of CD34+ cells in angiogenesis, including studies on implanted vascular grafts18–24 and molecularly based histological25 and embryological studies.26–36 Several threads of evidence converge therefore to strongly support the existence in the adult of circulating progenitor cells with endothelial potential.17 Transduction of human and non-human primate CD34+ cells, though, has been hampered by its consistently low efficiency.37–44 Unfortunately, maneuvers for improving transduction in vitro, such as prolonging exposure to viral vectors, stimulating cell proliferation, or culturing on stromal cells, affect the phenotypic features of progenitor cells in undesirable ways.45,46 To explore the concept of systemic gene delivery into areas of angiogenesis, we sought to use cellular vehicles naturally endowed with angiogiogenic tropism. Fundamental to this goal, we have shown that recombinant, nonreplicative herpes virus vectors (HSV) provide a novel means to modify with high efficiency and low toxicity human and nonhuman primate CD34+ cells genetically. The remarkable transducing ability of the vector facilitates short-term incubations for genetic manipulation of CD34+ angioblasts, avoiding undesired cellular phenotypic changes in culture. Furthermore, we have observed, by infusing genetically modified CD34+ cells in rhesus macaques bearing skin autografts, the expression of reporter genes and the presence of vector genome in areas of angiogenesis. Importantly, the accessibility, easy manipulation, lack of immunogenicity of the autologous CD34+ cell vehicles, and tropism for areas of angiogenesis of such a cell render autologous CD34+ endothelial progenitors ideal candidates for use as cellular vehicles targeted against areas of angiogenesis distributed systemically.

Results Human CD34+ cells are transduced with high efficiency by a replication-deficient herpes virus To explore novel means of efficiently transducing human and non-human primate CD34+ cells, we evaluated a second-generation herpes virus vector. Its main features are its being non-replicative by virtue of several deletions in genes essential for replication and its reduced cytotoxicity obtained by additional early gene deletions (Figure 1a). After an exposure for 12 h of CD34+ cells obtained from a mobilized normal human donor to the TOZ.1 herpes vector at 0.3 and 3 p.f.u. per cell, 11.7% and 99.9% of cells expressed the reporter gene lacZ, respectively, according to a fluorescence-activated cell sorter (FACS) analysis using FDG (Figure 1b, population 4). In contrast, Gene Therapy

less than 1% of cells transduced with the control T3 vector, non-expressing the reporter gene, were positive in this analysis. Similarly high efficiency was obtained with CD34+ cells isolated from bone marrow of a non-human primate (data not shown). Thus, an exceptionally high efficiency of transduction can be achieved in primate CD34+ cells with a non-replicative herpes virus after a short incubation in serum free and stromal cell free conditions.

Transduction of CD34+ cells with replication-deficient herpes virus has low toxicity in vitro To determine the toxicity of multiple-gene deleted HSV vectors in CD34+ cells, we continuously exposed human CD34+ cells in vitro to the TOZ.1 herpes virus vector for up to 7 days, and measured viable cells at 2, 4 and 7 days using a quantitative colorimetric assay. There was no significant decrease in the number of cells except at day 2 at an MOI of 30, a dose of virus one order of magnitude higher than needed for an optimal level of gene transfer (Figure 2). On the contrary, a virus-related cell proliferation was observed mostly at day 2. Analogous experiments in Vero cells showed a similar lack of deleterious effects on cell survival (data not shown). Cytotoxicity studies revealed therefore minimal or null deleterious effect of TOZ.1 on cell survival at the optimal MOI of 3 or less. Autologous transplanted CD34+ cells home into areas of angiogenesis in an in vivo model To determine the capacity of genetically modified CD34+ cells to localize into areas of angiogenesis, we transduced them ex vivo with the highly efficient HSV-based vector TOZ.1, which encodes the reporting gene lacZ. The modified CD34+ cells were autologously transplanted in three rhesus macaques bearing 2-day-old skin autografts. Approximately 1–2 × 106 CD34 cells/kg were infused in each animal. Considering the purity of the cell preparation (70–85%), the efficiency of transduction (90–95%), and the estimated blood volume (65 ml/kg), the estimated concentration of transduced CD34+ cells in these 3–4 kg animals was between 9600 and 11 000 cells/ml. This is in contrast to an estimated concentration in nonmobilized animals of 50 CD34+ cells/ml (assuming a percentage of 0.5% of CD34+ cells and 10 000 mononuclear cells/ml). Additionally, as an alternative model of angiogenesis, subcutaneous pellets of either Matrigel or Gelfoam were impregnated with the potent angiogenic growth factors VEGF and bFGF. Biopsies of both skin grafts and pellets, obtained at several time points, were stained with X-gal to reveal activity of the lacZ gene-encoded beta-galactosidase. Small, round, intensely blue cells were identified in neovascular structures (Figure 3a, b and c), both in grafts and pellets. Staining with anti-factor VIII antibodies confirmed the vascular character of these tube-like structures. Thus, the possibility of circulating CD34+ cells to home into areas of angiogenesis was confirmed in this model. A key aspect of our hypothesis is the localization of endothelial progenitors in areas of angiogenesis. In situ hybridization was performed as a confirmatory technique to obtain evidence for the presence of the genetically modified endothelial progenitors in the target skin autografts. In situ hybridization with a probe specific for a HSV vector sequence demonstrated nuclear staining of


45

Figure 1 Genetic modification of endothelial progenitors by herpes virus vectors (a) For constructing recombinant herpes virus vectors, a typical HSV1 150 000 kb viral genome has been deleted in genes essential for replication (ICP4 and ICP27) and genes additionally related to viral cytotoxicity (ICP22, UL41), rendering them substantially less toxic and allowing longer transgene expression. Further, the reporter gene lacZ and the protoxin gene thymidine kinase were inserted under the control of HSV immediate – early gene promoters ICPO and ICP4, respectively. The structure of TOZ.1 is ICP4−, ICP22−, ICP27−, UL24−: ICP4TK, UL41−: ICP0 lacZ. T3 was used as negative control for lacZ staining, and its structure is ICP4−, ICP22−, ICP27−, UL24−: ICP4TK The diagram shows the vectors backbone, deleted genes, and transgenes. Methods for constructing HSV-derived vectors have been described elsewhere.60 Open squared boxes represent HSV sequence deletions. Black boxes represent expression cassettes. (b) Expression of lacZ as determined by production of the fluorescent metabolite of FDG following infection by TOZ.1. Frozen human CD34+ cells were cultured for 1 h and incubated for 12 h (1) in serum-free media alone; (2) in the presence of the non-lacZ expressing herpesvirus T3; and either (3) with 0.3 p.f.u. or with (4) 3 p.f.u. of the lacZ expressing herpes virus TOZ.1. Next, they were stained with FDG for flow cytometry analysis.

graft cells (Figure 3f). In situ hybridization on HSVinfected CD34+ cells with an irrelevant probe (specific for plasmid DNA) was negative (Figure 3d), as was in situ hybridization with the lacZ gene probe on grafts from animals which did not receive genetically modified CD34+ cells (data not shown). Evidence for the presence of the HSV vector and for the expression of the lacZ reporter gene can thus be observed in skin autografts, while angiogenesis was taking place therein.

Sustained low levels of transgene expression in peripheral blood and bone marrow The use of genetically modified endothelial progenitors in vivo requires that expression of encoded transgenes be long enough to assure their effect. Therefore, we sought to detect transgene expression in blood and bone marrow in animals receiving infusions of HSV-infected CD34+ cells. To examine the duration of expression of the reporter gene lacZ, the ex vivo-modified autologous monGene Therapy


46

In vivo administration of CD34+ cells modified by a replication-deficient herpes virus appears to be safe During the first 12 months after infusion of CD34+ cells, no symptom or sign of toxicity has been observed in any of the three different primates treated with TOZ.1-transduced cells. Specifically, there was no change in behavior, weight or vital signs. Furthermore, there was no alteration in any standard hematocytometric or biochemical study, including liver and renal function biochemical profiles, with the exception of a reversible and minor increase in alkaline phosphatase during ganciclovir treatment (data not shown). However, to fully prove the safety of the replication-deficient herpes virus vector further studies are necessary.

Discussion

Figure 2 HSV exhibits minimal cytotoxicity on human CD34+ cells. To evaluate the cytotoxicity of the TOZ.1 vector, a cell survival assay was performed. Frozen human CD34+ cells were cultured for 1 h and then infected with the herpes virus TOZ.1 at MOIs of 0, 0.3, 3 and 30. Next, cells were plated in a 96-well plate, in quadruplicate wells, in media containing the growth factors IL-3, IL-6 and SCF. Cell survival was determined after 2, 4 and 7 days of continuous exposure in culture by means of a colorimetric cell proliferation assay. Two independent experiments were performed. Bars, mean percentage of surviving cells in a representative experiment, ± s.d. of the mean. There was no significant decrease in the number of cells except at day 2 at an MOI of 30, a dose of virus one order of magnitude higher than needed for an optimal level of gene transfer. On the contrary, a virus-related cell proliferation was observed mostly at day 2. Analogous experiments in Vero cells showed a similar lack of deleterious effects on cell survival (data not shown). Cytotoxicity studies revealed therefore minimal or null deleterious effect of TOZ.1 on cell survival at the optimal MOI of 3 or less.

key cells were introduced intravenously and the animals were monitored by flow cytometry of peripheral blood MNC to detect the proportion of FDG-positive cells in the circulating CD34+ population. We were able to detect the gene in both peripheral blood and bone marrow aspirates obtained serially for a period of weeks (3 weeks in two animals, 6 weeks in an additional animal) (Figure 4). In contrast, analysis of an identically gated cell population in mobilized but non-transplanted animals did not reveal any FDG+ cells.

Skin autografts in macaques infused with endothelial progenitors genetically modified with a tk-encoding HSV are rejected after treatment with ganciclovir As an additional way to verify the expression of genes contained in the transplanted CD34+ cells in the target angiogenesis areas, and to determine if we could exploit such expression with therapeutic intent, we infected CD34+ cells with the TOZ.1 vector, which encodes the gene thymidine kinase. Treatment with ganciclovir in animals that received the TOZ.1 transduced CD34+ cells induced the necrosis of skin autografts (Figure 5a) and the disappearance of the subcutaneous pellets (Figure 5b). In contrast, in animals treated with ganciclovir but transplanted with non-transduced cells, skin autografts were unaffected and pellets were evident for longer periods of time. Gene Therapy

We have examined, in a rhesus primate model, the possibility of genetically modifying CD34+ endothelial progenitor cells and have used them as cellular vehicles for systemic delivery into areas of angiogenesis. We found that human and non-human primate CD34+ cells can indeed be transduced with high efficiency and minimal toxicity in vitro with a novel, non-replicative herpes virus vector. Furthermore, we have observed, by infusing genetically modified CD34+ cells in rhesus macaques bearing skin autografts, the expression of reporter and toxin genes and the presence of vector genome in areas of angiogenesis. Importantly, the autologous transplantation of transduced CD34+ cells is well tolerated by the monkeys, and the expression of the reporter gene is sustained for several weeks. We have not yet analyzed or optimized, however, the quantitative relation between the number of infused cells, the quantity of transgene, and its selective expression in target areas of angiogenesis. The question of whether HSV-transduced CD34+ cells home specifically to sites of angiogenesis requires further resolution. This important practical issue is the subject of current studies in a tumor mouse model, more amenable to this quantitative and invasive study involving numerous animals. This uncertainty however does not alter the essence of the observations delineated herein, particularly since gene expression in CD34+ cells can be regulated by endothelial cell promoters. Efficient transduction of normal and malignant human hematopoietic cells, including CD34+ cells, has been previously reported.47 Levels of 70–100% cell transduction were achieved with a disabled infectious herpes virus that is restricted to a single cycle of replication (DISCHSV). Although toxicity appeared to be low, the impact of viral replication in hematopoietic cells and the effects of non-attenuated expression of viral genes have not been completely characterized. TOZ.1, the herpes virus vector we used, is non-replicative by virtue of the deletion of two genes essential for replication, and has been deleted in two other genes additionally responsible for herpesrelated cytotoxicity. As expected, our studies confirmed a remarkable absence of cytotoxicity of TOZ.1 in CD34+ and Vero cells in vitro. Thus, recombinant herpes virus TOZ. 1 offers some theoretical safety advantages with respect to other non-deleted herpes virus vectors. Recombinant adenovirus vectors have also been used for obtaining transient gene expression in CD34+ cells. Efficiency, however, has been generally much lower48 than with herpes virus vectors,49–54 with a maximum,


47

a

d

c

b

e

f

Figure 3 Localization of genetically modified cells in areas of angiogenesis. Photomicrographs of skin autograft biopsies obtained 5 days after autologous transplantation of CD34+ cells transduced with the lacZ-encoding herpes virus TOZ.1. Sections were stained with X-gal and eosin (a and b). Staining with anti-factor VIII antibody confirms the vascular character of the displayed structure (c). In situ hybridization was performed as a confirmatory technique to obtain evidence for the presence of the genetically modified endothelial progenitors in the target skin autografts. Hybridization with a probe specific for a HSV sequence demonstrated nuclear staining of graft cells (f). In the inset, two cells positive for HSV genome are shown. Both seem to be located in the wall of capillary tubes (f). As controls, we performed in situ hybridization on slides containing HSV-infected CD34+ cells using an irrelevant probe (specific for plasmid DNA) (d), or the lacZ gene probe (e). Hybridization with the lacZ gene probe on grafts from animals, which did not receive genetically modified CD34+ cells, was also negative (not shown).

somewhat exceptional compared with most other published attempts, of 45% modified CD34+ cells from the total number of cells infected.49 Retroviral vectors, while offering the prospect for long-term expression, are not only inefficient in transduction of human CD34+ cells, but they are further restricted to cycling cells for their integration.55,56 Unfortunately, conditions found to be required for improving retrovirus gene transfer, such as stimulation with growth factors for inducing active cell cycling and prolonged incubation with viral vectors, may critically alter the phenotype and function of target cells. In fact, differentiation and loss of desired properties of stem cells occurs frequently, consequently minimizing the relevance of the discreet improvements obtainable in gene transfer.46 Thus, recombinant, replication-deficient herpes virus offers advantages over currently available alternative vectors, which have limitations for achieving efficient gene transfer of CD34+ cells in a way that maintains the phenotypic and functional integrity of freshly isolated CD34+ cells. The duration of expression of genes encoded by herpes virus recombinant vectors in dividing cells should be limited, given that the virus life cycle does not include

chromosomal integration of viral DNA. In non-dividing cells, however, transgene expression has been observed to persist from days (neurons) up to several weeks (muscle).57 Using flow cytometry, we could observe expression of the reporter gene lacZ in (non-cycling) mononuclear cells of the peripheral blood of transplanted animals for at least 3 weeks. This analysis, which was less invasive than sampling the autograft or subcutaneous pellets, was complemented by the indirect observation of HSV-thymidine kinase expression at the target angiogenesis areas, as suggested by disappearance or necrosis of pellets or skin autograft after ganciclovir treatment. Of note, the prodrug treatment was begun 2 weeks after transplantation of transduced CD34+ cells. Thus, herpes virus can provide sufficiently sustained expression of genes of interest in CD34+ cellular vehicles for applications that require relatively short-term interventions. It should be noted that we have not yet directly examined the immunogenicity of HSV-transduced CD34+ cells by cytotoxic T lymphocyte analysis. Repeated administration of CD34+ cells is a practical possibility, the implementation of which will require studying the immune response associated with transplantation of Gene Therapy


48

Figure 4 Sustained low levels of transgene expression in peripheral blood and bone marrow CD34+ cells from rhesus monkeys were transduced with the lacZ-encoding herpes virus TOZ.1 and autologously transplanted. Mononuclear cells from the peripheral blood were obtained at different time points. Cells were stained with an anti-CD34 antibody and exposed to the galactosidase reagent FDG, and double positives were measured by flow cytometry. We were able to detect gene expression (as measured with the fluorescent substrate of FDG) in both peripheral blood and bone marrow biopsies obtained serially for 3 weeks in two animals, and 6 weeks in an additional animal. A representative curve is shown.

genetically modified autologous cells. Interestingly in this regard, CD34+ cells have been associated with induction of immune tolerance to highly immunogenic alloantigens.58 At this point, further studies are certainly needed to define the broad biodistribution of endothelial progenitors after autologous CD34+ cell transplantation. Ultimately, the biodistribution of infused cellular vehicles will determine the therapeutic index of delivering genes with cytotoxic or anti-vascular effect by using our proposed approach. In this regard, the expression of a toxin gene in the endothelial progenitors infused in the skin graft-bearing monkeys was remarkably well tolerated but, at the same time, efficacious enough for inducing graft necrosis. Of note, the parallel distribution of lacZ positive cells between blood and bone marrow in our studies is probably biased by the fact that it was difficult to obtain bone marrow biopsies of enough volume without the inclusion of contaminating blood, simply from the trauma of the procedure. Nonetheless, the observed lack of toxicity confirms that the selectivity of tropism achievable by the endothelial progenitors, yet to be determined, does not compromise the feasibility of our gene delivery approach. Furthermore, it is conceivable that the therapeutic genes used for modifying the endothelial progenitors can be put under control of regulatory sequences specifically activated in proliferating endothelium, thus limiting the expression of the therapeutic genes only to those endothelial progenitors that have localized into areas of active angiogenesis. Three transplanted macaques tolerated the entire procedure without complications, and more than 18 months later do not present any evidence of hematological, immunological, or solid organ toxicity. With regard to the function of the genetically modified CD34+ cells, we have not yet performed in vitro or in vivo assays to determine their capacity to proliferate and differentiate into Gene Therapy

Figure 5 Effects of the expression of a toxic gene by CD34+ endothelial progenitors. Rhesus monkeys were autologously transplanted with CD34+ cells transduced with the thymidine kinase-encoding herpes virus TOZ.1. After 15 days, treatment with ganciclovir, 10 mg/kg/day for 14 days, was administered. One animal did not receive CD34+ cells and served as control. (a) A shortening of the period during which pellets were palpable was observed in the transplanted group. (b) Sequence of changes that occurred during the treatment. The autograft was successfully accepted and vascularized by 4–5 days after transplant. Subsequently, the treatment with ganciclovir at 14–28 days after transplant induced a progressive necrotic process that was complete 5 days after the treatment ended.

hematopoietic progenitors and differentiated mature cells. Although destiny of the progeny of transduced CD34+ cells may be relevant in certain contexts, however, we planned ultimately to kill transduced cells by treatment with ganciclovir, and therefore did not studied this issue further at this point. Whereas a large number of attempts have sought to overcome the limitations of currently available viral and non-viral vectors, particularly using ex vivo gene transfer into parenchymal cells, little work has been proposed to employ ‘cellular vehicles’ in an autologous context.


Further, although the homing capacity of CD34+ endothelial progenitors has been recently described in adult animals,16,17 there is no previous experience with cellular vectors that display the natural ability both freely to circulate systemically and localize into areas of angiogensis. Our experiments involved skin autografts merely as models of angiogenesis that could be readily performed in non-human primates. We believe that the physiology involved, however, has allowed establishment of proofof-concept. Our results suggest that CD34+ cells can indeed localize into areas of angiogenesis and effect the expression of an incorporated transgene, and allow the evaluation of their potential as cellular vehicles with capacity for recruitment into areas of angiogenesis. The timing of destruction of the autografts and subcutaneous pellets suggested a local toxic effect induced by the prodrug ganciclovir that was efficacious even despite a probably low number of genetically modified CD34+ cells implanted therein. This destruction likely occurred by transformation of the prodrug into toxic metabolites in cells expressing thymidine kinase, with subsequent cell death. This mechanism would be consistent with the local replication of CD34+-derived endothelial cells.16 In addition, toxic bystander effects from ganciclovir metabolites, ischemia from vascular damage and occlusion, and inflammatory reactions could all contribute to the necrosis observed. Beyond the dependence on ganciclovir treatment, we have not yet precisely discriminated a primary mechanism underlying the necrosis. Support for a potential bystander effect, however, is provided by our recent in vitro and in vivo observations demonstrating a direct toxic bystander effect of human CD34+ HSVtktransduced cells on human endothelium and tumor cells (J Go´mez-Navarro and W Arafat, unpublished data). These facts establish a compelling rationale for our current attempts at trying to define the potential role of CD34+-based cellular vehicles in a mouse tumor model. The fact that our previous work was performed using both human and non-human primate CD34+ cells, a most stringent fact in terms of levels of gene transfer achievable by available vectors, emphasizes the potential feasibility for clinical application. Thus, the observations derived from our studies, we believe, have the potential to offer a method for systemic gene delivery into disseminated areas of angiogenesis, potentially including growing tumor foci. The process for obtaining autologous CD34+ cells is currently widely available, all of which makes the ultimate goal of applying this gene delivery and anti-vascular strategy to human gene therapy eminently attractive.

Materials and methods Recombinant herpes virus vectors To study the transduction efficiency of herpes virus vectors in CD34+ cells, we used TOZ.1, a non-replicative vector encoding the reporter gene lacZ and deleted in the genes ICP4, ICP22, ICP27 and UL41, which are related to viral cytotoxicity (Figure 1). As a negative control we used T3, a similar vector except for lacking the lacZ gene. The construction and preparation of these vectors, derived from a KOS strain mutant, has been described elsewhere.61 Viral stocks were prepared and titered using 7B cells, which express ICP4 and ICP27, as previously

described. Of note, purification of viral stocks was done by centrifugation in OptiPrep (Gibco, Gaithersburg, MD, USA).

49

Human and primate normal CD34+ cells We sought to develop a vector system that could have potential for rapid translation to clinical use. To avoid in advance the aforementioned problems with interspecies variation in transduction efficiency, human peripheral blood mononuclear cells (MNC) were obtained by leukapheresis of a normal donor mobilized by treatment with 15 ␮g/kg/day of G-CSF for 5 consecutive days. Human CD34+ cells were purified by immunoadsorption to 95% purity (Systemix, Palo Alto, CA, USA). Cells were then frozen at −70°C until further use. Alternatively, we used non-human primate cells obtained in this case from aspirates of normal bone marrow of similarly mobilized Macacca mulatta. These target cells were chosen having in mind the rhesus monkey as a relevant animal species to be used in eventual preclinical experiments. MNC were isolated by density centrifugation over Ficoll–Paque (Pharmacia, Piscataway, NJ, USA) and CD34+ cells were isolated on a Ceprate (CellPro Inc, Bothell, WA, USA) immunoaffinity column according to the manufacturer’s protocol. Transduction of CD34+ cells Highly enriched CD34+ cells were plated in 1 ml of serum-free Stem-Pro culture media (Gibco) supplemented with Stem-Pro Nutritional Supplement (Gibco), interleukin-3 (IL-3, 5 ng/ml), IL-6 (10 ng/ml) and stem cell factor (SCF, 25 ng/ml). All growth factors were purchased from R&D Systems (Minneapolis, MN, USA). Cells were incubated at 1 × 106 cells per well in a 12-well flat bottom plate. Herpes virus stock was directly added to the cultures at 0, 0.3, 3 and 30 multiplicities of infection (MOI). Infections were initiated 1 h after starting the culture. In the case of primate CD34+ cells, culture was started approximately 8 h after BM extraction. Rigorous aseptic technique was observed during the entire procedure. At 12 h after infection, flow cytometric analysis or autologous transplantation of primate cells was performed. Cytotoxicity studies in vitro To evaluate the cytotoxicity of the TOZ.1 vector, a cell survival assay was performed. Human CD34+ cells were plated in quadruplicate wells on 96 well plates, at 10 000 cells per well, and infected with 0, 0.3, 3 and 30 infectious particles per cell in serum-free Stem-Pro media. In addition, Vero cells (African green monkey kidney cells) were also assayed. Cell viability was determined after 2, 4 and 7 days using a colorimetric cell proliferation assay that measures conversion of a tetrazolium salt to formazan by viable cells, as described by the manufacturer (Cell Titer 96 Aqueous Non-radioactive Cell Proliferation Assay, Promega, Madison, WI, USA). Briefly, 20 ␮l assay reagent mixture was added to each well containing 100 ␮l of cell suspension. The plates were incubated for 3 h at 37°C before absorbance was measured at 490 nm in a 96 well plate reader (Molecular Devices, Menlo Park, CA, USA). To quantify cell numbers for the experimental results, a standard curve was generated by plating CD34+ cells in triplicate wells on the day of the assay at the following concentrations: 50 000, 20 000, 10 000, 5000, 2000, Gene Therapy


50

and 0 cells per well. From the resulting standard curve, cell numbers were calculated for the experimental groups using SOFTmax computer software (Molecular Devices).

Determination of transduction efficiency by flow cytometry analysis of ␤-galactosidase activity To determine transduction efficiency, either aliquots of transduced cells or peripheral blood mononuclear cells were stained with anti-human CD34-biotin antibody (CellPro) or an irrelevant antibody conjugated with streptavidin-PerCP (Becton-Dickinson, San Jose, CA, USA). After 20 min of incubation in the dark, and then being washed, the cells were stained with fluorescein di-betad-galactopyranoside (FDG; Sigma, St Louis, MO, USA) by the FACS-gal technique of Fiering et al.59 Briefly, after incubation in a 37°C water bath for 5 min, 50 ␮l of distilled water containing 2 mm FDG was mixed thoroughly with an equal volume of cells suspended in staining media (SM) (10 mm HEPES and 4% FCS in 1 × phosphatebuffered saline (PBS), pH 7.3). The cells were then returned to the water bath for exactly 1 min. FDG cell loading was stopped by adding 500 ␮l of ice-cold SM. Cells were maintained in ice until the FACS analysis was performed approximately 30 min later. Cells not incubated with FDG served as the negative control for FDG fluorescence and were used to set the upper boundary of the negative cell autofluorescence. Cells emitting signal over this threshold were counted as FDG(+) cells. Flow cytometry data acquisition was performed in the first 30 min after staining in a FACScalibur flow cytometer (Becton Dickinson), and analysis was performed with CellQuest software. In vivo model of angiogenesis and general management of primates To evaluate the homing capacity of CD34+ cells into areas of angiogenesis, we established an in vivo model based on implantation in primates of skin autografts, which should undergo a robust angiogenic reaction. In addition, as a versatile alternative source of study material, we implanted subcutaneously, in several separate areas of the abdominal wall, pellets of either Matrigel (500 ␮l) (Becton Dickinson) or clinical grade Gelfoam (0.5 cm2) (Upjohn, Kalamazoo, MI, USA). These materials were impregnated with 1 ␮g of the potent angiogenic factors vascular endothelial growth factor (VEGF; a kind gift from Genentech, South San Francisco, CA, USA) and basic fibroblast growth factor (bFGF; Boehringer Mannheim, Indianapolis, IN, USA). Two days after skin graft and pellet implantation, 1– 2 × 106 CD34+ enriched transduced cells/kg were infused into the saphenous vein. Normal, 3–4 kg rhesus macaques (Macacca mulatta), 2–3 years old, were obtained from breeding colonies at LABS (Yemassee, SC, USA), the University of Wisconsin (Madison, WI, USA), and Hazleton (Alice, TX, USA). They were conditioned in a quarantine facility for 60 days. Grafts, biopsies and venous blood sampling were performed under ketamine anesthesia. The antibiotic cephazolin was administered prophylactically (three doses at 12.5 mg/kg) immediately after transplantation. Surgical and non-surgical procedures were carried out in accordance with the provisions of the NIH ‘Guide for the Care and Use of Laboratory Animals’ under the supervision of the University of Alabama at Birmingham Animal Care and Use CommitGene Therapy

tee. The animals were maintained in a restricted-access facility and received NIH monkey chow and water through an automatic water system containing an in-line 0.22 ␮m filter.

Tropism for areas of angiogenesis of gene-modified CD34+ cells and transgene expression After autologous transplantation, transduced cells, which should express the TOZ.1-encoded lacZ reporter gene, were sought in sites of angiogenesis by X-gal staining, as follows. Wedge biopsies were obtained from skin grafts and subcutaneous Matrigel and Gelfoam pellets of primates autologously transplanted 5 days before with genetically modified CD34+ cells. The samples were snap frozen (−70°C) in tissue embedding media (Histo Prep; Fisher Scientific, Pittsburgh, PA, USA). Frozen blocks were sectioned, rinsed twice with PBS, and then fixed at 4°C for 2 h in 0.5% glutaraldehyde. Next, sections were rinsed twice at 37°C for 15 min in PBS containing 1 mm MgCl2 and incubated with X-gal solution (PBS pH 7.4, 5 mm potassium ferrocyanide, 5 mm potassium ferricyanide, 1 mm MgCl2, and 1 mg/ml X-gal) at room temperature. After development of color, tissues were rinsed twice thoroughly in PBS with 3% DMSO, then three times in 70% ethanol, and cleared with a substitute of xylene (Hemo-DE; Fisher). Sections were then counterstained with eosin or antibodies against endothelial-specific Factor VIII (Dako, Carpinteria, CA, USA), and analyzed by light microscopy for the appearance of the characteristic blue staining representing ␤-galactosidase activity. To confirm the presence in areas of angiogenesis of cells transduced by the TOZ.1 vector, in situ hybridization was performed for transgene (lacZ) vector sequences. Serial paraffin sections were mounted on glass slides, deparaffinized, rinsed in 2 × standard saline citrate (SSC), and immersed in 3% hydrogen peroxide at 37°C for 10 min to quench endogenous peroxidase activity. After proteinase K digestion for 15 min at 37°C, slides were then dehydrated and air-dried, and a cover slip was applied and then heated at 95°C for 10 min to denature viral DNA. Sections were incubated for 2 h at 50°C with a biotinylated probe in 50% formamide, 3 × SSC, Denhardt’s solution, dextrose sulfate, and 10 mm DTT. The probe was obtained using the following primers against the lacZ gene sequence: 5⬘-TTG CTG ATT CGA GGG GTT AAC CGT CAC GAG and 3⬘-ACC AGA TGA TCA CAC TGC GGT GAT TAC GAT. Slides were then washed in 50% formamide, 2 × SSC, and 25 mm DTT at 50°C for 20 min and incubated with a horseradish peroxidase-conjugated streptavidin (Zymed Laboratories, South San Francisco, CA, USA). The reaction product was visualized by means of the AEC chromogen. Figure 3a and b show in situ hybridization of CD34 cytospin slides processed without and with the biotinylated lacZ probe, respectively. Figure 3c shows the skin graft biopsy obtained in a rhesus monkey 2 weeks after infusion of HSV-infected CD34+ cells and similarly processed. In the inset, two cells positive for HSV genome are shown. Both seem to be located in the wall of capillary tubes. As a further means to show the functional consequences of expression of genes delivered by CD34+ cells at the angiogenesis areas, we exploited the presence in the TOZ. 1 herpes virus vector of the thymidine kinase gene. To this end, treatment with ganciclovir (Cytovene; Syntex Laboratories, Palo Alto, CA, USA) was initiated


15 days after CD34+ cell infusion with 5 mg/kg i.v. for 14 days. A control animal received non-transduced CD34+ cells and the same ganciclovir treatment. All macaques underwent periodic blood sampling for monitoring of hematocytometric and biochemical standard parameters.

Acknowledgements This work was supported by NIH grant AI-94012 to Judith M Thomas. We thank Clement Asiedu and Zhi Huang for their help with flow cytometry, and Arabella Tilden for her assistance with CD34+ processing and storage. Last but not least, Nat Borden took care of the rhesus animals with utmost ability.

23

24

25

26

27

References

28

1 Kong HL, Crystal RG. Gene therapy strategies for tumor antiangiogenesis. J Natl Cancer Inst 1998; 90: 273–286. 2 Folkman J. Antiangiogenic gene therapy. Proc Natl Acad Sci USA 1998; 95: 9064–9066. 3 Kluth D. Inflammation, genes and gene therapy. JR Coll Phys Lond 1998; 32: 167–170. 4 Knechtle SJ, Zhai Y, Fechner J. Gene therapy in transplantation. Transplant Immunol 1996; 4: 257–264. 5 Anderson WF. Human gene therapy. Nature 1998; 392: 25–30. 6 Bilbao G, Contreras JL, Gomez-Navarro J, Curiel DT. Improving adenoviral vectors for cancer gene therapy. Tumor Targeting 1998; 3: 59–79. 7 Thurston G et al. Cationic liposomes target angiogenic endothelial cells in tumors and chronic inflammation in mice. J Clin Invest 1998; 101: 1401–1413. 8 Kay MA, Liu D, Hoogerbrugge PM. Gene therapy. Proc Natl Acad Sci USA 1997; 94: 12744–12746. 9 Vrancken PM, Perkins AL, Kay MA. Method for multiple portal vein infusions in mice: quantitation of adenovirus-mediated hepatic gene transfer. Biotechniques 1996; 20: 278–285. 10 Worgall S, Wolff G, Falck-Pedersen E, Crystal RG. Innate immune mechanisms dominate elimination of adenoviral vectors following in vivo administration. Hum Gene Ther 1997; 8: 37–44. 11 Kafri T et al. Cellular immune response to adenoviral vector infected cells does not require de novo viral gene expression: implications for gene therapy. Proc Natl Acad Sci USA 1998; 95: 11377–11382. 12 Paillard F. Mesothelial cells: the panacea for ex vivo gene therapy? Hum Gene Ther 1997; 8: 1839–1840. 13 Gage FH. Cell therapy. Nature 1998; 392: 18–24. 14 Smith A. Cell therapy: in search of pluripotency. Curr Biol 1998; 8: R802–R804. 15 Rancourt C et al. Endothelial cell vehicles for delivery of cytotoxic genes as a gene therapy approach for carcinoma of the ovary. Clin Cancer Res 1998; 4: 265–270. 16 Asahara T et al. Isolation of putative progenitor endothelial cells for angiogenesis. Science 1997; 275: 964–967. 17 Shi Q et al. Evidence for circulating bone marrow-derived endothelial cells. Blood 1998; 92: 362–367. 18 Stump MM, Jordan GLJ, DeBakey ME, Halpert B. Endothelium grown from circulating blood on isolated intravascular Dacron hub. Am J Pathol 1963; 43: 361–363. 19 Gonzalez IE, Vermeulen F, Ehrenfeld WK. Relations between circulating blood and pathogenesis of atherosclerosis. Isr J Med Sci 1969; 5: 648–651. 20 Kennedy LJJ, Weissman IL. Dual origin of intimal cells in cardiac-allograft arteriosclerosis. New Engl J Med 1971; 285: 884–887. 21 Frazier OH, Baldwin RT, Eskin SG, Duncan JM. Immunochemical identification of human endothelial cells on the lining of a ventricular assist device. Tex Heart Inst J 1993; 20: 78–82. 22 Scott SM, Barth MG, Gaddy LR, Ahl ET Jr. The role of circulat-

29

30 31

32

33 34

35

36 37 38

39 40

41

42

43

44

45

46

ing cells in the healing of vascular prostheses. J Vasc Surg 1994; 19: 585–593. Shi Q et al. Proof of fallout endothelialization of impervious Dacron grafts in the aorta and inferior vena cava of the dog. J Vasc Surg 1994; 20: 546–556. Rafii S et al. Characterization of hematopoietic cells arising on the textured surface of left ventricular assist devices. Ann Thorac Surg 1995; 60: 1627–1632. Young PE, Baumhueter S, Lasky LA. The sialomucin CD34 is expressed on hematopoietic cells and blood vessels during murine development. Blood 1995; 85: 96–105. Kabrun N et al. Flk-1 expression defines a population of early embryonic hematopoietic precursors. Development 1997; 124: 2039–2048. Eichmann A et al. Ligand-dependent development of the endothelial and hemopoietic lineages from embryonic mesodermal cells expressing vascular endothelial growth factor receptor 2. Proc Natl Acad Sci USA 1997; 94: 5141–5146. Shalaby F et al. A requirement for Flk1 in primitive and definitive hematopoiesis and vasculogenesis. Cell 1997; 89: 981–990. Caprioli A et al. Blood-borne seeding by hematopoietic and endothelial precursors from the allantois. Proc Natl Acad Sci USA 1998; 95: 1641–1646. Choi K et al. A common precursor for hematopoietic and endothelial cells. Development 1998; 125: 725–732. Hatzopoulos AK et al. Isolation and characterization of endothelial progenitor cells from mouse embryos. Development 1998; 125: 1457–1468. Nishikawa SI et al. Progressive lineage analysis by cell sorting and culture identifies FLK1 + VE-cadherin+ cells at a diverging point of endothelial and hemopoietic lineages. Development 1998; 125: 1747–1757. Labastie MC et al. Molecular identity of hematopoietic precursor cells emerging in the human embryo. Blood 1998; 92: 3624–3635. Hirashima M, Kataoka H, Nishikawa S, Matsuyoshi N. Maturation of embryonic stem cells into endothelial cells in an in vitro model of vasculogenesis. Blood 1999; 93: 1253–1263. Cleaver O, Krieg PA. VEGF mediates angioblast migration during development of the dorsal aorta in Xenopus. Development 1998; 125: 3905–3914. Shalaby F et al. Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice. Nature 1995; 376: 62–66. Medin JA, Karlsson S. Viral vectors for gene therapy of hematopoietic cells. Immunotechnology 1997; 3: 3–19. Dunbar CE et al. Transduction of hematopoietic stem cells in humans and in nonhuman primates. Stem Cells 1997; 15 (Suppl 1): 135–139; discussion 139–140. Havenga M, Hoogerbrugge P, Valerio D, van Es HH. Retroviral stem cell gene therapy. Stem Cells 1997; 15: 162–179. Hennemann B et al. Optimization of retroviral-mediated gene transfer to human NOD/SCID mouse repopulating cord blood cells through a systematic analysis of protocol variables. Exp Hematol 1999; 27: 817–825. Douglas JT et al. A system for the propagation of adenoviral vectors with genetically modified receptor specificities. Nat Biotechnol 1999; 17: 470–475. Case SS et al. Stable transduction of quiescent CD34 (+)CD38(−) human hematopoietic cells by HIV-1-based lentiviral vectors. Proc Natl Acad Sci USA 1999; 96: 2988–2993. Chatterjee S et al. Transduction of primitive human marrow and cord blood-derived hematopoietic progenitor cells with adenoassociated virus vectors. Blood 1999; 93: 1882–1894. Miyoshi H et al. Transduction of human CD34+ cells that mediate long-term engraftment of NOD/SCID mice by HIV vectors. Science 1999; 283: 682–686. Briones J et al. Retroviral gene transfer into human hematopoietic cells: an in vitro kinetic study. Haematologica 1999; 84: 483–488. Tisdale JF et al. Ex vivo expansion of genetically marked rhesus peripheral blood progenitor cells results in diminished longterm repopulating ability. Blood 1998; 92: 1131–1141.

51

Gene Therapy


52

47 Dilloo D et al. A novel herpes vector for the high-efficiency transduction of normal and malignant human hematopoietic cells. Blood 1997; 89: 119–127. 48 Watanabe T et al. Gene transfer into human bone marrow hematopoietic cells mediated by adenovirus vectors. Blood 1996; 87: 5032–5039. 49 Neering SJ et al. Transduction of primitive human hematopoietic cells with recombinant adenovirus vectors. Blood 1996; 88: 1147–1155. 50 Seth P et al. Adenovirus-mediated gene transfer to human breast tumor cells: an approach for cancer gene therapy and bone marrow purging. Cancer Res 1996; 56: 1346–1351. 51 Marini FC, Zhang X, Trapnell B, Andreeff M. Inefficient infection of primitive hematopoietic precursors cells (CD34+) using CMV driven adenoviral vectors (ad-v). Blood 1996; 88 (Suppl 1): 292b. 52 Garcia Sanchez F et al. Use of a safety recombinant adenovirus vector for selective bone marrow purging carrying the chemosensitization gene cytosine deaminase. International Society of Hematology, Charlottesville, VA, Carden Jennings, 1996, Abstr. 619: 1138. 53 Fu SQ, Garcia Sanchez F, Chung I, Deisseroth AB. Adenoviral vectors and hematopoietic cells. Blood 1997; 89: 1460–1467.

Gene Therapy

54 Frey BM et al. High-efficiency gene transfer into ex vivo expanded human hematopoietic progenitors and precursor cells by adenovirus vectors. Blood 1998; 91: 2781–2792. 55 Miller DG, Adam MA, Miller AD. Gene transfer by retrovirus vectors occurs only in cells that are actively replicating at the time of infection. Mol Cell Biol 1990; 10: 4239–4242. 56 Springett GM et al. Infection efficiency of T lymphocytes with amphotropic retroviral vectors is cell cycle dependent. J Virol 1989; 63: 3865–3869. 57 Huard J et al. Gene transfer to muscle using herpes simplex virus-based vectors. Neuromusc Disord 1997; 7: 299–313. 58 Rachamim N et al. Potential role of CD34 stem cells in tolerance induction. Transplant Proc 1997; 29: 1935–1936. 59 Fiering SN et al. Improved FACS-Gal: flow cytometric analysis and sorting of viable eukaryotic cells expressing reporter gene constructs. Cytometry 1991; 12: 291–301. 60 Krisky, DM et al. Rapid method for construction of recombinant HSV gene transfer vectors. Gene Therapy 1997; 4: 1120–1125.

Reference added in proof 61 Krisky DM et al. Deletion of multiple immediate–early genes from herpes simplex virus reduces cytotoxicity and permits long-term gene expression in neurons. Gene Therapy 1998; 5: 1593–1603.