Selective Ablation of Human Embryonic Stem Cells Expressing a “Suicide” Gene

Authors

  • Maya Schuldiner,

    1. Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem, Israel
    Search for more papers by this author
  • Joseph Itskovitz-Eldor,

    1. Department of Obstetrics and Gynecology, Rambam Medical Center, Faculty of Medicine, The Technion, Haifa, Israel
    Search for more papers by this author
  • Nissim Benvenisty M.D., Ph.D.

    Corresponding author
    1. Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem, Israel
    • Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem 91904, Israel. Telephone: 972-2-6586774; Fax: 972-2-6584972
    Search for more papers by this author

Abstract

Over the past few years, technological procedures have been developed for utilizing stem cells in transplantation medicine. Human embryonic stem (ES) cells can produce an unlimited number of differentiated cells and are, therefore, considered a potential source of cellular material for use in transplantation medicine. However, serious clinical problems can arise when uncontrolled cell proliferation occurs following transplantation. To avoid these potential problems, we genetically engineered human ES cell lines to express the herpes simplex virus thymidine kinase (HSV-tk) gene. Expression of the HSV-tk protein renders the ES cells sensitive to the U.S. Food and Drug Administration-approved drug ganciclovir, inducing destruction of HSV-tk+ cells at ganciclovir concentrations that are nonlethal to other cell types. The reversion rate of engineered cells was low even under prolonged selection with ganciclovir. The HSV-tk+ clones retained a normal karyotype and the ability to differentiate to cells from all three germ layers. Most importantly, tumors that arose in mice following subcutaneous injection of HSV-tk+ human ES cells could be ablated in vivo by administration of ganciclovir. By utilizing these cell lines, safety levels can be improved in transplantations involving tissues derived from human ES cells.

Introduction

The use of stem cells in transplantation therapy has recently been the focus of much research. The properties of stem cells, namely, self-renewal and vast differentiation potential, are of major importance in transplantation procedures as they provide a solution to the lack of tissue sources. Transplants of stem cells into tissues such as pancreas [1], brain [24], spinal cord [5], heart [6], and bone marrow [7, 8] of mice, rats, and nonhuman primates have shown some degree of penetration and normal differentiation in host tissue. In a few cases, phenotypic improvement in the animals' health was shown to occur subsequent to transplantation [5]. In a recent study, human fetal brain cells were injected into the nigrostriatum of patients suffering from Parkinson's disease. This resulted in a reduction of disease symptoms in most patients, though adverse reactions were found to occur in some patients. This was probably due to uncontrollable proliferation of the fetal cells resulting in neurotransmitter overproduction [9]. Cellular overproliferation and tumor formation are examples of clinical problems that must be addressed before using nonterminally differentiated cells. In order to control cell proliferation, a transplantable cell line expressing a negative selectable marker should be established.

Human embryonic stem (ES) cells are pluripotent cells derived from the inner cell mass of blastocyst-stage human embryos [1012]. It has been shown that these cells can differentiate in culture to cells of all three germ layers [13] and that their differentiation potential may be manipulated by growth factors [14]. Various cell types may be identified in differentiated human ES cells, for example, neurons [1518], pancreatic β cells [19], cardiomyocytes [20, 21], hematopoietic cells [22], and endothelial cells [23]. Recently, a method was established to genetically manipulate human ES cells [24]. This method allowed us to produce lines of genetically modified cells that expressed a negative selection marker. These cells may be eliminated when cellular overproliferation occurs in vivo.

Materials and Methods

Cell Culture and Growth Analysis

Human ES cells (H9 clone [10], passage 40–50) and embryoid bodies (EBs) were cultured as described [14], using a medium consisting of 80% KnockOut DMEM (an optimized Dulbecco's modified Eagle's medium for ES cells; GIBCO/BRL; Grand Island, NY; http://www.invitrogen.com), 20% KnockOut SR (a serum-free formulation; GIBCO/ BRL), 1 mM glutamine (GIBCO/BRL), 0.1 mM β-mercaptoethanol (Sigma; St. Louis, MO; http://www.sigmaaldrich.com), 1% nonessential amino acids stock (GIBCO/BRL), and 4 ng/ml basic fibroblast growth factor (bFGF) (GIBCO/BRL). During formation of EBs, bFGF was removed from the medium and no other growth factors were added, in order to allow spontaneous differentiation. Ganciclovir (Sigma) was administered 1 day after plating at the stated concentrations, and the medium was changed with fresh ganciclovir every 2 days. Cell densities were determined by fixating cells to tissue culture plates with 2.5% gluteraldehyde. The plates were stained with Methylene Blue (Sigma) dissolved in 0.1 M boric acid (pH = 8.5). Color was extracted by 0.1 M hydrochloric acid, and emission was read at 650 nM. As color density represents cell presence, growth charts based on color emission were made.

Transfection and Establishment of Transgenic Cell Lines

pPNT [25], a plasmid that contains two phosphoglycerate kinase (PGK) promoters driving either the neomycin resistance gene or the herpes simplex thymidine kinase (HSV-tk) gene and pPGK-enhanced green fluorescence protein (EGFP) [24] were introduced into human ES cells using ExGen 500 (Fermentas; St. Leon-Rot, Gemany; http://www.fermentas.de) as described [24]. More specifically, 107 cells were transfected with 12 μg of plasmid DNA, centrifuged, and incubated for 30 minutes with the transfection reagent. Twenty-four hours later, cells were replated and selected with G418. Following 10 days of selection, in the pPNT transfected cultures, nine colonies appeared, out of which six were indeed sensitive to ganciclovir, showing that the plasmid had integrated successfully and that HSV-tk was being expressed.

Fluorescence-Activated Cell Sorting (FACS) Analysis

FACS analyses of PGK-EGFP- and PNT-expressing cells were performed on a FACSCalibur system (Becton Dickinson; Franklin Lakes, NJ; http://www.bd.com), according to their green fluorescent emission. Cells were analyzed following trypsin digestion, centrifugation, and resuspension of cell pellets in phosphate-buffered saline (PBS). Prior to analysis, cells were filtered with a 40-μM MESH filter (Falcon; Becton Dickinson) and were kept on ice. Undifferentiated human ES cells were used to set the background level of fluorescence.

RNA and Reverse Transcription-Polymerase Chain Reaction (RT-PCR)

Total RNA was extracted as described [26] using extraction by guanidine thiocyanate followed by phenol-chloroform and isopropanol precipitation. cDNA was synthesized from 1 μg total RNA using random hexamer (pd(N)6) as a primer (Pharmacia Biotech; Upsala, Sweden; http://www.pnu.com) and Moloney murine leukemia virus reverse transcriptase (GIBCO/BRL). cDNA samples were subjected to PCR amplification with previously described DNA primers [14]. For each gene, the DNA primers were derived from different exons to ensure that the PCR product represented the specific mRNA species and not genomic DNA. PCR was performed using the Clontech Advantaq (Palo Alto, CA; http://www.clontech.com) RT-PCR kit and a two-step cycle at 68°C.

Cytogenetic Analysis

Cells prepared for cytogenetic analysis by G-banding were incubated in growth media with 0.1 μg/ml of Colcemid (Biological Industries, Bet Haemek, Israel) for 10 minutes. They were then trypsinized, resuspended in 0.075 M potassium chloride, and incubated for 10 minutes at room temperature. Fixation was performed using 3:1 methanol/acetic acid.

In Vivo Experiments in Mice

All animal experiments were conducted under the supervision of the Hebrew University Faculty of Sciences Animal Care and Use Committee (license: NS-01-05). Teratomas were formed by subcutaneous injection of 106 ES cells into nude or SCID/beige mice [27]. Tumors appeared 3–6 months following injection. Cell lines were formed by removing the tumors and plating the dissociated cells on 0.1% gelatin-coated plates with human ES cell medium. For the in vivo ganciclovir administration, 50 mg/kg in PBS was injected daily into the peritoneal cavity. Histological sections were made by paraffin embedding.

Results

Creation of Genetically Modified Clones

In order to create cell lines that carried a negative selection marker, we transfected human ES cells with a plasmid encoding the HSV-tk gene under the control of a housekeeping gene (PGK) promoter. The introduced plasmid vector also contained the neomycin resistance gene that facilitated selection of cells harboring the foreign DNA. Expression of HSV-tk induces conversion of the prodrug nucleoside ganciclovir to its drug form as a phosphorylated base analogue. The phosphorylated ganciclovir is incorporated into the DNA of replicating cells causing irreversible arrest at the G2/M check point followed by apoptosis [28, 29]. All isolated lines of human ES cells transfected with HSV-tk were eliminated when ganciclovir was administered (Fig. 1A). The transfected clones died when exposed to a range of ganciclovir concentrations varying over three orders of magnitude (from 2 × 10−8 M to 1 × 10−5 M), while the control human ES cells remained unaffected (Fig. 1B). We chose to continue working with an intermediate concentration of 2 × 10−6 M ganciclovir, which induced cell death in all HSV-tk+ cell lines. At that concentration, when ganciclovir was removed from the medium after 9 days of selection, no renewed cell growth occurred as could be seen by screening the plates under a microscope in addition to staining for colony formation. As human ES cells proliferate rapidly, lack of growth over a time frame of 3 weeks indicated a terminal effect of the drug (Fig. 1C).

Figure Figure 1..

Effects of ganciclovir on human ES cells expressing the HSV-tk gene.A) Human ES cell clones expressing the HSV-tk gene show sensitivity to the presence of the prodrug ganciclovir (Ganc). B) Dose response of HSV-tk+clones to ganciclovir treatment. Six HSV-tk+clones (TK+) and human ES cells (as a control) were grown in the presence of varying concentrations of ganciclovir. The bars represent standard error values. C) Time course of the effects of ganciclovir on HSV-tk+cells. Six HSV-tk+clones and control human ES cells were treated with 2 × 10−6M ganciclovir for 9 days. After 9 days, ganciclovir was removed from the medium and cells were grown for an additional 10 days. The bars represent standard error values.

Susceptibility of HSV-tk+ Cells to Ganciclovir

Transplantation procedures could expose human ES cells to conditions that are not encountered in culture. First, in vivo, the cells would be free from the selective pressure exerted by neomycin to retain the foreign DNA, and in the case of treatment with ganciclovir, the cells could be under heavy selection in favor of those cells that had lost the HSV-tk gene. Moreover, the cells will differentiate over time and will also be surrounded by cells that are resistant to ganciclovir, which could change their overall sensitivity to the treatment. These problems may cause resistance to the ganciclovir treatment. In order to examine whether HSV-tk+ cells could be selectively eliminated, we grew the cells together with HSV-tk cells expressing the green fluorescent protein (GFP+) and treated them with ganciclovir. When the cultures were analyzed using FACS, it was clear that selective killing of the HSV-tk+ cells occurred within the same time frame and at the same dosage as cultures of HSV-tk+ cells only (Fig. 2). As no sensitivity of GFP+ cells to ganciclovir was documented, it seems that no major cell fusion event occurred.

Figure Figure 2..

Selective ablation of HSV-tk+cells.Analysis of cultures of HSV-tkcells expressing GFP (GFP+) with HSV-tk+(TK+) cells. Following treatment with ganciclovir (Ganc), selective elimination of the HSV-tk+cells was shown through FACS analysis of GFP expression. Peaks were colored differentially in order to help discriminate between signals. Notice that the TK+cells had no fluorescence, whereas the GFP+cells formed two distinct peaks, one with low fluorescence and one with high fluorescence. (This distribution is a property of the PGK-GFP cell line used).

Analysis of the effects of differentiation was conducted by taking differentiated teratoma cells that had formed following injection of the ES cell clones into immune-deficient mice. The teratoma cells were removed from the mice after 8 weeks, during which they were not subject to selection. The dissociated cells that were grown in culture formed lines of teratoma cells that had a differentiated morphology but proliferated rapidly. These cells retained their susceptibility to ganciclovir at 2 × 10−6 M (Fig. 3). This analysis also confirms that differentiated cells that are not grown continuously on neomycin selection still retain HSV-tk expression. Furthermore, we monitored the reversion rate of the ganciclovir susceptibility in order to get an estimation of the possible number of drug-resistant cells found subsequent to transplantation. The cells were grown for two passages in the absence of the positive-selection drug neomycin to mimic lack of selection following transplantation. They were then replated in the presence of ganciclovir and grown for a further 10 days, after which resistant colonies were counted. Reversion was noted at frequencies between 10−6 and 10−7. Using several “suicide” genes in each transplanted cell line may reduce the problem of reversion.

Figure Figure 3..

Sensitivity of teratoma cells to ganciclovir.A) Differentiated HSV-tk+teratoma cells showed sensitivity to the presence of the prodrug ganciclovir (Ganc). B) Time course of differentiated HSV-tk+teratoma cells in response to 2 × 10−6M ganciclovir for 8 days. After 8 days, ganciclovir was removed from the media and cells were grown for an additional 8 days. The bars represent standard error values.

Retention of Normal Karyotype and Differentiation Potential of Genetically Modified Clones

Our aim was to ascertain that the genetically modified clones still retained their normal, pluripotent characteristics. Cytogenetic analysis of the transfected clones revealed a normal karyotype (Fig. 4A). Moreover, the undifferentiated cells could easily aggregate to form normal-looking EBs [13] in vitro and pluripotent teratomas in vivo (Fig. 5B). RT-PCR analyses of cDNA from both HSV-tk+ EBs and HSV-tk+ teratoma cells revealed the cells' capacity to differentiate to all three germ layers as well as to a wide variety of cell types. This was made clear in the expression of the endodermal tissue-specific markers α-fetoprotein (αFP, primitive endoderm), albumin (liver), and amylase (pancreas); the mesodermal markers β-globin (blood), c-actin (c-act, cardiac muscle), and enolase (muscle); and the ectodermal markers keratin-1 (skin), glial fibrillary acidic protein (GFAP, glia), and neurofilament heavy chain (NF-H, mature neurons) (Fig. 4B).

Figure Figure 4..

Retention of normal karyotype and differentiation potential by the HSV-tk+cell lines.A) Karyotype of an HSV-tk+clone using G-banding. All cell lines had a normal, XX chromosome set. B) RT-PCR analysis of differentiation markers from the three germ layers on human ES cells, HSV-tk+EBs, and HSV-tk+teratoma (Ter) cells. The markers used were: αFP, albumin, amylase, β-globin, cardiac actin (c-act), enolase, keratin-1, GFAP, and NF-H. As a control for the presence of cDNA, the housekeeping genes β-actin and glyceraldehyde-3-phosphate dehydrogenase were used.

Figure Figure 5..

Elimination of HSV-tk+cells in vivo after teratoma formation in mice.A) Growth curve of three HSV-tk+teratomas prior to and following ganciclovir administration. Bars indicate standard error. Slashed line indicates the predicted rise in tumor mass without treatment, as calculated from the rate of growth prior to ganciclovir administration. Note the reduction in tumor mass following treatment. B) HSV-tk+tumor appearance in the presence and absence of ganciclovir. Treatment with ganciclovir (Ganc) caused a dramatic reduction in vascularization, as can be seen in the photos of a whole tumor mass. In addition, hematoxilin and eosin (H&E) staining showed large fibrotic areas in the treated tissues relative to the control, and DAPI nuclear staining stressed the lack of nuclei in those areas. Scale bars = 10 μm.

Elimination of Tumors In Vivo in a Murine Model System

Before routine use of these cell lines is advisable, it is important to ascertain that cell death and elimination can occur in vivo following tumor formation in a whole animal. For this purpose, we injected immune-deficient mice subcutaneously with the HSV-tk+ ES cells. After 2–6 months, four mice developed teratoma tumors. Tumor volumes ranged from ∼100–2,000 mm3. During the first week following tumor observation, measurements were made of their size to ascertain continuation of tumor development. Following this time period, they were subjected to ganciclovir treatment. One mouse was used as a control and was injected daily with saline, whereas the three remaining mice were injected daily with 50 mg/kg ganciclovir for 10 days. During this period, tumor mass was reduced (Fig. 5A), with one of the tumors disappearing completely. The mouse carrying the tumor that disappeared following treatment was grown for an additional 3 months in which no tumor recurrence was observed. Two mice still showed tumor mass following 10 days of treatment. Yet, the tumors no longer grew during treatment duration and even shrank to about 20% of their expected size without treatment. In order to analyze the effects of the treatment on these tumors, they were removed and compared with the control tumor. Interestingly, the tumors treated with ganciclovir had a milky white appearance, marking the absence of a blood supply. This is in contrast to the vascularized appearance of the control tumor (Fig. 5B). When a treated tumor was dissociated in an attempt to make a cell line, no live cells could be detected, and no cell line could be formed although untreated tumors form cell lines with ease (results not shown). Histological sections were made from the tumors and stained with hematoxilin and eosin. This staining shows that the tumors were indeed pluripotent teratomas harboring many cell types (Fig. 5B). However, in the treated tumors, large sections were fibrotic in appearance, showing mostly connective tissue. This was seen more easily following 4′6-diamidino-2-phenylindole (DAPI) staining of nuclei, which demonstrated very few cells relative to similar sections in control tissue (Fig. 5B). Terminal deoxytransferase-mediated deoxy uridine nick end-labeling (assay) staining showed few apoptotic cells, but this analysis was problematic due to the vast areas of connective tissue with no cells in the treated tumors (results not shown).

Discussion

Human ES cells are a powerful tool for transplantation medicine. These cells are capable of proliferating indefinitely in culture while remaining pluripotent. Potentially, ES cells may be genetically manipulated so as to evade the host immune system. As such, they constitute both an unlimited and a widely applicable cell source [3032]. One of the risks involved in using stem cells is the possibility of massive proliferation. This can be a problem if the stem cells present in the grafted sample take on inappropriate fates [9]. To address these problems, we engineered several human ES cell lines that constitutively expressed the HSV-tk gene. As expression of this gene causes sensitivity to the U.S. Food and Drug Administration-approved prodrug ganciclovir, transplanted cells can be targeted specifically, allowing for nonintrusive removal of grafts in cases of undesirable side effects. Our results show that HSV-tk+ ES and teratoma cells were ablated when exposed to a wide range of ganciclovir concentrations (10−5-10−7), while normal cells remained unaffected. The recommended i.v. dose of ganciclovir is from 5–50 mg/kg [33], which produces steady-state plasma levels higher than 10−6 M [34]. This concentration is within the range that effectively killed all the HSV-tk+ human ES cells. As the reversion rate is low, it appears that, in the event of malignant transformation, treatment with ganciclovir may reduce tumor mass by more than six orders of magnitude, allowing dramatic elimination of the cells. Since our reversion rate was measured in vitro, it may be that in vivo more ganciclovir-resistant cells may be accumulated. Thus, in order to reduce the reversion rate, it may be necessary in the future to produce lines with two unrelated suicide genes in separate genomic locations. As these cell lines express the HSV-tk gene under a “housekeeping” promoter, treatment with ganciclovir will cause cell death of all dividing graft tissue. This may cause elimination of the much-needed graft in addition to the tumor tissue. Though problematic in cases of bone marrow transplantation, this may not prove to be a difficulty in cases where the graft constitutes nondividing cells, such as neural tissue. However, if graft elimination does occur, patients may have to undergo additional transplantation procedures. An additional option is to construct much more sophisticated cell lines in which the HSV-tk gene can be silenced in the appropriately grafted tissue. This may be performed, for example, by flanking the HSV-tk gene with lox sites and expressing cre recombinase under a promoter specific for the required tissue [35]. Possibly, much simpler constructs driving expression of the suicide gene under a stem cell-specific promoter would negatively select only rapidly dividing, undifferentiated cells.

The potential use of HSV-tk+ cells in transplantation medicine is enhanced by the fact that these cells retained the characteristic properties of ES cells. This was demonstrated by the fact that no karyotype modification occurred during the manipulation process. Perhaps even more importantly, the genetically modified clones also retained their capacity to differentiate into a wide variety of tissues, as can be seen both by their ability to form differentiated teratoma tumors in vivo, and by their ability to aggregate into EBs harboring cells from the three embryonic germ layers. The expression of markers for neuronal, cardiac, and pancreatic cells indicates that they might possibly be used in the treatment of pathologies such as neurodegenerative diseases, cardiomyopathy, and diabetes mellitus.

We ascertained the safety of selective elimination of human ES cells in vivo using cellular transplantation experiments in a murine animal model. Those experiments showed that even tumors that arose after a lengthy period of up to 6 months retained their sensitivity to ganciclovir. More experiments that follow up on tumor formation in various organs for extended periods of time are needed before clinical applications are considered. As the model system consisted of solid tumors, remaining fibrotic tissue could be observed even following treatment. Such cases of residual scar tissue have been described routinely both after standard chemotherapy for solid tumors [3638] and following gene therapy procedures that introduce suicide genes through viral infection of the host [39]. Though interfering with follow-up of tumor progression, these masses of connective tissue no longer endanger the patient in most cases.

Previous experiments with suicide genes have focused on using gene therapy to treat malignancies. Though there are other suicide genes, including a tetracycline-inducible form of the diphtheria toxin [40] and bacterial cytosine deaminase [41], most gene therapy research has focused on the introduction of the HSV-tk gene. This focus is due to the limited side effects of treatment with ganciclovir. Successful advanced-stage clinical trials with ganciclovir are presently being conducted with the objective of treating malignancies such as melanomas and glioblastomas [34, 42]. One of the main problems in gene therapy is the delivery system. No such problem is encountered with human ES cells as these cells can be grown indefinitely in culture, thus enabling production of premade genetically modified clones as required and bypassing the need for difficult methods of gene transfer. The genetic manipulation of stem cells to express a suicide gene, such as the HSV-tk gene or Fas intracellular domain, prior to transplantation has recently been utilized in reducing graft-versus-host disease symptoms of transplanted lymphocytes [43, 44]. This procedure, though immensely important for bone marrow transplantation, modifies lymphocytes that are restricted in their biomedical uses. In contrast, human ES cells can differentiate into any cell type, allowing such cell lines to be used in virtually any sort of transplantation. Additionally, their immortality allows for the addition of any other needed genetic modification in culture forming a customized cell line, a situation not possible for other stem cells due to their low proliferative potential in culture. Our work demonstrates, for the first time, the possible biotechnological applications of these cells once differentiation procedures can be combined with genetic modification. Such a combination will allow the formation of an unlimited cell source for potentially all cell types of the body, custom made to fit the unique requirements of each transplantation procedure or patient.

This manuscript describes the formation of cell lines expressing the HSV-tk negative selection marker. The formation of many cell lines, each with this transgene at a unique integration site, may be used as a tool for studying basic processes in early human embryogenesis. Through selection of revertants, showing no sensitivity to ganciclovir, it is possible to obtain deletion mutants of whole genomic areas. The effects of these deletions on differentiation processes, X inactivation, and self-renewal of stem cells will allow the discovery of new genes and gene functions. This has previously been performed in mouse ES cells [45], but to date, no such library exists in human ES cells. Our results also reveal the potential for making human ES cells a safer reagent in transplantation medicine without sacrificing their pluripotent characteristics.

Acknowledgements

This study was partially supported by the Juvenile Diabetes Foundation and by the Herbert Cohn Chair (N.B.). S.M. is a Clore fellow. Cytogenetic analysis of the clones was carried out at the Genetics Department of Hadassah Medical School.

Ancillary