Retroviral transduction of quiescent haematopoietic cells ... - Nature

Gene Therapy (1999) 6, 1084–1091  1999 Stockton Press All rights reserved 0969-7128/99 $12.00 http://www.stockton-press.co.uk/gt

Retroviral transduction of quiescent haematopoietic cells using a packaging cell line expressing the membrane-bound form of stem cell factor A Sehgal, N Weeratunge and C Casimir Department of Haematology, Imperial College School of Medicine, St Mary’s Campus, Norfolk Place, London W2 1PG, UK

Gene therapy vectors based on murine retroviruses are unable to transduce non-dividing cells. This has proven a particular problem in the haematopoietic system where the target cells of choice, the pluripotent stem cells are quiescent. In an attempt to circumvent this difficulty we have constructed a retroviral producer line that expresses the membrane bound form of human recombinant stem cell factor (SCF) on its cell surface. This should enable the retroviral producers to deliver a growth signal to the target cells simultaneous with their exposure to retrovirus. We tested the ability of these modified producers to transduce a growth factor-starved, SCF-dependent cell line (TF-1) and demon-

strated that these cells, though quiescent, can still be successfully transduced. This approach was extended to targeting of umbilical cord blood CD34+ cells, a predominantly quiescent population that normally require the addition of cytokines for efficient transduction. Using the SCFexpressing producer line in the absence of exogenously added cytokines, we observed a marked stimulation in transduction efficiency over that achieved using the parent producer line alone. Colonies derived from these cells arising in semi-solid media were also shown to be positive for expression of a retrovirally encoded reporter gene.

Keywords: retrovirus; ␤-galactosidase; stem cells; transduction; kit ligand

Introduction A major limitation affecting the use of retroviral vectors in gene therapy protocols is that target cells may often be quiescent. This makes them unsuitable targets for transduction with murine retroviral vectors as these vectors can only stably integrate into the genome of actively dividing cells.1,2 This situation has special relevance for pluripotent haematopoietic stem cells (PHSC) for example and attempts to transduce such cells have met with very limited success.3–6 Nevertheless, retroviruses currently represent the most highly developed of the available vector systems, having the ability to insert into the host chromosome with high efficiency and at a limited copy number. Moreover, retroviral vectors have been proven safe enough for some limited clinical application.5–9 It seems worthwhile, therefore, investigating methods to develop retroviruses further as a suitable vector system for gene delivery to quiescent cells such as PHSC. In an attempt to circumvent problems associated with gene delivery to PHSC we have genetically modified a retroviral producer cell line to express the membraneassociated form of human stem cell factor (SCF) on its cell surface. This has a number of potential benefits. Firstly, retroviruses should preferentially target the stem cells through cell surface SCF receptors. Secondly, the effective retroviral titre is raised through close association of the

Correspondence: C Casimir Received 12 August 1998; accepted 15 February 1999

producer and target cells. Finally, the producer cells can simultaneously deliver a specific growth signal along with the retrovirus. To construct a retroviral producer line expressing surface-bound SCF, we modified a retroviral producer line carrying the nlslacZ gene.10 This cell line, termed AM12lacZ-⌬SCF, was used to target a factor-dependent cell line, TF-1 and cord blood-derived CD34+ cells. We demonstrate here that quiescent TF-1 cells can be both induced to cycle and transduced by co-cultivation with these producer cells. In addition, efficient gene transfer to umbilical cord blood CD34+ cells and transfer to colonies derived from them was observed in the absence of exogenously added cytokines.

Results Construction of the AM12lacZ-⌬SCF cell lines In order to test whether retroviral vectors could be used to transduce quiescent cells we chose to modify an existing producer cell line MFG-nlslacZ10 (referred to here as AM12(lacZ)). The product of the nlslacZ gene, ␤-galactosidase, is easily detected histochemically and can be used to confirm retroviral transduction.10,11 Although scoring transduced cells by ␤-galactosidase expression is a less direct measure of retroviral transduction than detection of the retroviral genome by PCR, it is much more rapid and straightforward to perform. The membrane-bound SCF cDNA was delivered to the AM12(lacZ) cell line by calcium phosphate-mediated transfection with plasmid pREP8-⌬SCF.12 Plasmid

Transduction of quiescent haematopoietic cells A Sehgal et al

pREP8-⌬SCF is a derivative of the mammalian expression vector pREP8 (Invitrogen, De Schelp, The Netherlands), and contains SCF cDNA ⌬28.13 This cDNA is derived from a SCF splice-variant that lacks exon 6, the region containing the major proteolytic cleavage site, thus resulting solely in the synthesis of the membrane-bound form of SCF.14 This human isoform is also very inefficiently cleaved from the surface of mouse cells.15 The pREP8 plasmid also confers resistance to L-histidinol. Following transfection of AM12(lacZ), histidinol-resistant colonies were pooled by trypsinisation and expanded in Dulbecco’s MEM containing HXM (see Materials and methods), and 2.5 mm l-histidinol. Surface expression of SCF was confirmed by immunofluorescent staining with anti-SCF antibody (Figure 1). One of these polyclonal cell lines thus obtained, was termed AM12lacZ-⌬SCF and was used in all subsequent experiments. The titre of this cell line was tested on NIH3T3 cells and was found to be comparable with that of the parent producer cell line (see Materials and methods).

Proliferation assays on TF-1 cells To test the ability of the AM12lacZ-⌬SCF cell line to transduce quiescent cells, the growth factor-dependent human leukaemic cell line, TF-116 was chosen. TF-1 cells must be maintained in GM-CSF or IL-3 and die within several days when deprived of cytokine. Though not capable of long-term growth in SCF, TF-1 cells will survive and multiply for limited periods in this growth factor. We determined the optimal concentration of SCF required for proliferation of TF-1 cells and observed this to be 50 ng/ml (Figure 2a). The growth factordependence of TF-1 cells was confirmed by depriving them of SCF for various time periods and then re-challenging them with growth factor to see what proportion of the cells were still able to proliferate. After 24 h growth factor starvation, only about 50% of the cells were still capable of proliferation and total cell death occurred within 4 days (Figure 2b). We then compared the proliferative response of TF-1 cells co-cultured with irradiated AM12lacZ-⌬SCF cells (expressing membrane-bound SCF) to that obtained with soluble SCF and with TF-1 cells co-cultivated with irradiated AM12(lacZ) cells (lacking cell-surface SCF). TF cells exposed to soluble SCF alone gave a clear proliferative response but this was weaker than with GMCSF alone (Figure 3a). The AM12lacZ-⌬SCF cells stimu-

lated the cells to approximately the same degree as the soluble SCF alone (Figure 3a). Addition of soluble SCF to AM12lacZ-⌬SCF co-cultures was unable to increase the magnitude of this response (data not shown). Having established that the SCF on the surface of the AM12lacZ-⌬SCF cells was biologically active, we then compared the ability of AM12lacZ-⌬SCF and AM12(lacZ) to transduce TF-1 cells rendered quiescent by removal of growth factor support. Unexpectedly, no difference in the rates of transduction were observed (data not shown). This observation raised the possibility that the parent cell line, AM12(lacZ), may already be producing a growth factor which was capable of stimulating the TF1 cells to proliferate This indeed proved to be the case, as the proliferative response to co-culture with AM12(lacZ) cells was at least as good as that obtained with the AM12lacZ-⌬SCF producers (Figure 3a). Indeed, the AM12 cell line itself (from which AM12(lacZ) is derived) induced a potent proliferative response in TF-1 cells, though this was somewhat reduced in magnitude compared with the response to GM-CSF (Figure 3a). A large number of different cell types including fibroblasts and mononuclear cells in peripheral blood or bone marrow have been observed to produce IL-617 including retroviral packaging cell lines.18 Furthermore, Kitamura et al19 have also shown that IL-6 supports the growth of TF-1 cells. Based on this knowledge we hypothesised that the factor stimulating the TF-1 cells to proliferate was probably murine IL-6 produced by the AM12(lacZ) packaging cells. To test this hypothesis TF-1 cells were incubated with varying concentrations of antibody to murine IL-6 and then assayed for proliferation following co-cultivation with AM12(lacZ) producers. As shown in Figure 3b, pre-incubating AM12(lacZ) cells with increasing amounts of ␣-IL-6 antibody correspondingly decreased the proliferative response of the TF-1 cells. All subsequent assays on TF-1 cells involving co-cultivation with fibroblasts were therefore performed following pre-incubation of the irradiated fibroblasts with 5 ␮g/ml of ␣-IL-6 antibody. This step effectively reduced the background proliferative response of the TF-1 cells and made it possible to test the SCF-mediated retroviral transduction of quiescent TF-1 cells using our genetically modified retroviral producer line.

Retroviral gene transfer into quiescent TF-1 cells TF-1 cells were rendered quiescent by growth factor starvation, as described earlier. The cycling status of the cells

Figure 1 Immunofluorescent staining of modified retroviral producers. The retroviral producer cells AM12(lacZ) and AM12lacZ-⌬SCF were grown on cover slips and then stained with antibody to human SCF to detect cell surface expression: (a) AM12lacZ-⌬SCF and (b) AM12(lacZ).

1085


1086

Figure 2 Proliferation assays on TF1 cells. TF-1 cells were starved of growth factor overnight to establish quiescence. At various time-points following growth factor withdrawal, aliquots were re-challenged with growth factor (SCF) and the proliferative response was estimated by 3H-thymidine incorporation. (a) Proliferation of TF1 cells in response to stimulation with soluble recombinant human SCF. (b) Proliferative response of TF cells following growth factor starvation.

Figure 3 Proliferation assay on TF-1 cells co-cultured with retroviral producers. The proliferation of TF1 cells incubated alone with of growth factors, or following co-culture with retroviral producer cells was estimated as described above. (a) No cytokine, unstimulated cells; sSCF, cells stimulated with 25 ng/ml soluble recombinant human SCF; LacJP11, TF-1 cells co-cultured with the AM12lacZ-⌬SCF cell line; AM12-lacZ, TF1 cells co-cultured with the AM12(lacZ) cell line; AM12, TF1 cells co-cultured with the AM12(lacZ) cell line; GM-CSF, cells stimulated with 200 pg/ml GM-CSF. (b) Proliferation of TF-1 cells co-cultured with AM12(lacZ) in the presence of increasing amounts of anti-IL-6 monoclonal antibody.

was assessed by 3H-thymidine incorporation. As shown in Figure 4, quiescent cells showed many fewer labelled nuclei (Figure 4b) than logarithmic phase cells (Figure 4a). Quiescent TF-1 cells were co-cultivated with either irradiated AM12(lacZ) or the AM12lacZ-⌬SCF producers both of which had been pre-incubated with 5 ␮g/ml of ␣-IL-6 antibody. TF-1 cells co-cultured with AM12lacZ⌬SCF producers revealed a 3H-thymidine labelling pat-

tern indicative of SCF-dependent cycling, with increased numbers of labelled cells relative to quiescent cells and those co-cultured with AM12(lacZ) (Figure 4c and d); the latter cells resembled, more closely, the pattern seen with the quiescent starting population. Moreover, only TF-1 cells co-cultured for 16 h with the AM12lacZ-⌬SCF producer line stained positively for the presence of the nlslacZ gene An average of about 2–3%


1087

Figure 4 Transduction of TF-1 cells using the AM12lacZ-⌬SCF producer cell line. TF-1 cells rendered quiescent by growth factor starvation were transduced with retroviral vectors overnight, collected on to microscope slides and stained for incorporation of 3H-thymidine to detect cells that had divided during the co-culture period (a–d) or ␤-galactosidase activity to detect retroviral transduction (e-f). (a) Cycling TF1 cells alone; (b) quiescent TF1 cells alone; (c,e) quiescent TF1 cells following transduction with AM12lacZ-⌬SCF producers; (d,f) quiescent TF1 cells following transduction with AM12(lacZ) producers. The arrows in panels a–d denote examples of dividing cells and in panel e retrovirally transduced cells. Note that the same cells are arrowed in panels c and e.

of cells (2.6 ± 0.5) stained positive for ␤-galactosidase (Figure 4e). By contrast, following similar co-culture with the AM12(lacZ) producers, we observed no positively stained TF1 cells (across three independent experiments approximately 6000 cells were counted) (Figure 4f). The 3 H-thymidine labelling showed that transduced cells also stained positive for 3H-thymidine incorporation (arrowed cells in Figure 4c and e).

Retroviral transduction of CD34+ cells BA number of reports have shown that the CD34+ cells obtained from umbilical cord blood are quiescent and transfect poorly in the absence of exogenous cytokines.20–23 We therefore extended our technique for SCF-mediated retroviral transduction of TF-1 cells, to transduce umbilical cord blood CD34+ cell populations to establish whether we could transfect these cells without the addition of cytokines. Approximately 2–5 × 105 CD34+

cells were co-cultivated with confluent, irradiated AM12(lacZ) and AM12lacZ-⌬SCF producers, as described above. Following exposure to the producers, the harvested CD34+ cells were stained for the presence of the nlslacZ gene. Approximately 20–30% of the CD34+ cells harvested from the AM12lacZ-⌬SCF producers stained positively for the presence of the nlslacZ gene (Figure 5b), whereas in TF-1 cells transduced using the AM12(lacZ) producers, ␤-galactosidase activity could be detected in 1% or fewer (Figure 5a). Although the level of transduction obtained with AM12(lacZ) producers did prove variable between cord blood samples, most likely dependent on the proportion of quiescent cells in the population, we consistently observed an approximately three-fold increase in transduction efficiency using the AM12lacZ-⌬SCF producers (Table 1). In some experiments, a similar analysis was performed on colonies derived from the transduced CD34+ cells by


1088

Figure 5 Histochemical staining of CD34+ cells for the presence of ␤-galactosidase gene following transduction. CD34+ umbilical cord blood cells were transduced by co-cultivation for 48 h with (a) AM12(lacZ) or (b) AM12lacZ-⌬SCF producer cell lines and stained for ␤-galactosidase activity.

Table 1 Transduction of CD34+ cord blood cells by SCFexpressing retroviral producers Experiment No.

AM12LacZ⌬SCF %ve cellsa

AM12(LacZ) %ve cellsa

AM12LacZ⌬SCF/ AM12(LacZ)

1 2 3 4 5 6

22.1 18.3 33.5 40.4 22.0 36.0

3.5 6.0 1.0 14.4 0.6 12.6

6.3 3.1 33.5 2.8 36.7 2.9

Average

28.72

6.35

4.5b

a

For each experiment 10 random views from two different cytospin preparations were scored for expression of ␤-galactosidase, approximately 2500–5000 cells. b P = ⬍0.001, Student’s paired t test.

plating in semi-solid media. Out of approximately 200 colonies from two independent experiments, none of those arising from co-cultivation with AM12(lacZ) stained positively for ␤-galactosidase (not shown). In contrast, small numbers (around 2%) of those co-cultured with the AM12lacZ-⌬SCF producers gave a positive staining reaction (Figure 6). Examples of erythroid (Figure 6a) and macrophage colonies (Figure 6b) are shown. It is worth noting that variegated expression was observed in a significant number of these colonies, suggesting that expression from the retroviral LTR may frequently become extinguished. The actual transduction frequency, therefore, may well be significantly higher. These data are very similar to those presented previously by Bagnis et al10 using an AM12(lacZ) vector to transduce CD34+ cells isolated from mobilised peripheral blood.

Discussion In this report we have sought to develop a method for the retroviral transduction of quiescent haematopoietic cells by employing a producer cell line that had been modified to express the membrane-bound form of SCF on its cell surface. The system was shown to be capable

of successfully transducing TF-1 cells, made quiescent by growth factor starvation and cord blood-derived CD34+ cells, a population known to be largely quiescent. The use of umbilical cord blood CD34+ cells as targets for retroviral transduction has become widespread. However, the ability to transduce these cells is very heavily dependent on the use of cytokines. For example Shi et al23 used a combination of SCF, IL-3 and IL-6 to produce a three-fold increase in transduction frequency and using the same growth factor combination Hanley et al22 described a 10-fold increase. This combination also gave improved transduction frequencies with a retroviral vector encoding the Fanconi’s anaemia group C protein24 and using a neoR reporter.25 Other successful combinations have included SCF, IL-3, GM-CSF, Epo26 and SCF, Flt3-ligand, GM-CSF, TNF-␣, TGF-␤.27 The system we have described in this report lends itself well to the transduction of populations of cells like these that are predominantly quiescent but without recourse to the use of exogenously added cytokines. This approach may also have value in other systems where target cells are normally quiescent, such as in liver or neural tissue and could obviate the necessity for ex vivo transduction, if packaging cells could be implanted. In this context, it could also be of use in targeting retrovirus to cells in close proximity to the producers, by localising growth factor spread or by limiting proliferation to specific subsets of cells bearing receptors. Similar approaches to ours, but using soluble cytokines have been used successfully for retroviral gene transfer to bone marrow progenitor cells28 and to support human haematopoiesis in long-term culture, using mouse stromal layers that were engineered to produce human cytokines.29 For the purpose, ultimately, of targeting haematopoietic stem cells, our decision to use SCF as the stimulating growth factor was dictated by a number of lines of evidence suggesting that SCF may be effective in increasing the efficiency of retroviral transfer to PHSC. For example, in bone marrow cell separations, cells expressing high levels of surface SCF receptor, c-kit, have been shown to contain the large majority of cells with long-term repopulating ability.30 In addition, PHSC selected by treating bone marrow cells with 5-fluorouracil in the presence of soluble SCF and IL-331 are positive for c-kit. Also, mice


1089

Figure 6 Histochemical staining of colonies derived from transduced CD34+ cells for the presence of ␤-galactosidase gene. Cells transduced as for Figure 5 were plated into semi-solid media and incubated for a further 10 days. Colonies arising in the semi-solid media were stained in situ for ␤-galactosidase activity. (a) erythroid colony; (b) macrophage colonies.

carrying the steel mutation have defects in the gene encoding SCF and suffer from a variety of severe haematopoietic defects, are sterile, lack pigment (except in retina) and have reduced numbers of mast cells in their skin.14,32–34 Significantly, mice that harbour the ‘dickie’ allele of steel produce a, biologically active, soluble form of SCF but not the membrane-bound form,32,34 yet they display a phenotype very similar to those that make no SCF at all. Thus, the membrane-bound isoform of SCF may have greater physiological significance for early haematopoiesis than the soluble form. In addition, murine stromal cells transfected with human membrane-bound SCF cDNA have been shown to be more supportive of human haematopoietic progenitors in long-term bone marrow cultures, than cells transfected with a human soluble SCF cDNA.14 This line also directed increased levels of gene transfer to rhesus monkey long-term repopulating cells.35 We have also recently investigated whether a similarly modified retroviral producer line, which transduces a cDNA for the NADPH oxidase component p47phox, could improve gene transfer to quiescent human stem cells selected with 5-flurouracil31 with highly encouraging results.12

Materials and methods Cytokines and antisera Purified recombinant human SCF was obtained from Sigma (St Louis, MO, USA). Recombinant human GMCSF was kindly provided by Professor D Linch (UCLMS, London, UK). Anti-murine IL-6 neutralising antibody was obtained from R&D Systems, UK. Fluorescein isothiocyanate (FITC)-conjugated rabbit anti-goat IgG antibody was purchased from Zymed Laboratories (San Francisco, CA, USA). Cells and media The murine fibroblast cell line NIH3T3 and the amphotropic packaging cell line AM12,36 were routinely cultured in Dulbecco’s modified Eagle’s medium (DMEM, Life Technologies, Paisley, UK) containing 10% fetal calf serum, 2 mm l-glutamine, penicillin and streptomycin. The MFG-nlsLacZ producer cell line10 was kindly provided by C Porter (Institute for Cancer Research, London,

UK) and was termed AM12(lacZ) during this study. This retroviral producer line was maintained in standard DMEM supplemented with HXM (hypoxanthine (15 ␮g/ml), xanthine (250 ␮g/ml) and mycophenolic acid (25 ␮g/ml)), to select for the helper virus plasmids carrying the viral gag, pol and env genes. The titre of the AM12(lacZ) cell line was tested on NIH3T3 cells and was found to be 2.2 × 105 cfu/ml. In comparison, the titre of a AM12lacZ-⌬SCF producer line (carrying the membranebound SCF cDNA; see Results section) was found to be 1.6 × 105 cfu/ml. The human leukemic cell line, TF-1,16 was maintained in RPMI 1640 (Life Technologies) containing 10% fetal calf serum, and 2 mm l-glutamine, penicillin and streptomycin, and 200 pg/ml GM-CSF.

Proliferation assays GM-CSF: The protocol employed in this study was a modification of the one reported by Callard et al.37 Actively cycling TF-1 cells (1 × 106) were washed twice with PBS to remove all traces of cytokine. Washed cells (5 × 105) were then maintained in media without cytokines; the remaining cells were cultured in media containing GM-CSF and served as a positive control. After 24 h of growth-factor deprivation, 1 × 104 (100 ␮l) quiescent cells were dispensed into sterile flat-bottomed microtitre wells. Varying concentrations of GM-CSF were divided into aliquots to triplicate wells. The positive control cells were treated in the same manner. The plates were incubated for 48 h at 37°C in 5% CO2 in air, after which 1 ␮Ci 3H-thymidine (specific activity: 5 Ci/mmol; Amersham International, Bucks, UK) was added to each well. Following a further incubation period of 12–18 h, the labelled cells were harvested on to glass-fibre discs using Dynatech Automash Cell Harvester (LabTech, Sussex, UK). The incorporated 3H-thymidine was counted on a liquid scintillation counter (Wallac 1209 Rack Beta, Wallac, Milton Keynes, UK) and expressed as disintegrations per minute (d.p.m.) after correction for quenching and efficiency of counting. SCF: Actively cycling TF-1 cells were washed twice with PBS and transferred to media containing 50 ng/ml SCF.


1090

After culturing TF-1 cells in this media for up to 5 days the cells were starved of growth factor and the proliferation assay was performed as detailed above. 3

H-thymidine autoradiography This was carried out essentially as reported by Wilkinson and Green.38 To label transduced cells, 12–16 h before harvesting the TF-1 cells, 1 ␮Ci 3H-thymidine was added to the growth medium. Cells were stained for ␤-galactosidase activity and then dipped in photographic emulsion, air dried and exposed for 1 week at −70°C. Detection of cell surface SCF by immunofluorescence AM12(lacZ) and LacJP cells were washed in PBS, deposited on to slides using Cytospin 3 (Shandon, Runcorn, UK) and incubated with goat anti-human SCF antibody (R&D Systems, Oxford, UK) for 1 h at room temperature. After gentle washing, the cells were incubated with FITC-conjugated rabbit ␣-goat Ig antibody (Zymed Laboratories) for 1 h at room temperature, and viewed under a fluorescence microscope. ␤-Galactosidase staining Cells, either grown in microtitre plates (24- or six-well), or deposited on microscope slides (54 g, 7 min) using a Cytospin (Shandon), were washed twice with PBS and fixed with 0.5% glutaraldehyde washed in PBS to remove the fixative and stained for 2–4 h at 37°C with 1 mg/ml X-gal (5,bromo-4,chloro-3,indolyl-␤-d-galactoside) in PBS buffer supplemented with sodium deoxycholate (0.01%), NP40 (0.2%), 20 mm MgCl2, 5 mm potassium ferricyanide, 5 mm potassium ferrocyanide.

Isolation of CD34+ cells from cord blood Anti-coagulated cord blood was diluted 1:4 with PBS containing 2 mm EDTA. The diluted cell suspension (35 ml) was layered over 15 ml of Ficoll–Paque (Pharmacia, Herts, UK). This was centrifuged for 35 min at 400 g at 20°C in a swinging-bucket rotor (without brake). The upper layer was then removed, leaving the mononuclear cell layer undisturbed at the interphase. The interphase cells were collected, washed twice in PBS containing 2 mm EDTA and resuspended in this buffer at a final concentration of 108 cells per 300 ␮l buffer. CD34+ cells were isolated from cord blood using the MiniMACS CD34 Progenitor Cell Isolation Kit (Miltenyi Biotec, Surrey, UK) which positively selects CD34-expressing cells by indirect magnetic labelling. These cells (approximately 5 × 106–1 × 107 mononuclear cells) were then magnetically labelled and CD34+ cells isolated as described by the manufacturers. The yield of CD34+ cells by the MiniMACS separation method averaged 0.8% and purity was between 80 and 95%. Viral titre and transduction Retroviral producers were grown to confluence, washed twice with PBS and incubated overnight with fresh medium. Supernatant was harvested, filtered through a 0.22-␮m filter and stored in aliquots at −70°C. Dilutions of supernatant were added to sub-confluent NIH3T3 cells in medium supplemented with 4 ␮g/ml polybrene and incubated overnight. The medium was then removed, the cells washed twice with PBS and incubated in fresh medium for a further 48 h before staining for ␤-galactosidase.

Retroviral transductions were performed by co-culture of target and retroviral producer cells for 16–48 h at 37°C in IMDM (CD34+ cells) or RPMI (TF-1 cells) media. The producer cells were grown to confluence and irradiated (25 Gy) before use. Quiescence was induced in TF-1 cells as described for proliferation assay.

Colony assays Following transduction cells were collected by centrifugation and respended in 500 ␮l of IMDM. This was added to 2.5 ml StemGem (obtained from Professor J Hatzfeld, Laboratoire de Biologie Cellulaire et Moleculaire des Facteurs de Croissance, Institut de Recherches Scientifique sur le Cancer, Villejuif, Paris, France; a ready-to-use semisolid media, which contains the necessary cytokines for multilineage colony growth) and plated into two 2-ml Petri dishes. Plates were incubated for 10 days at 37°C, 5% CO2 and saturated humidity for 10 days before scoring colonies under the inverted microscope. Staining for ␤-galactosidase activity was as described above. Approximately 1 ml of the staining buffer was carefully layered over the surface of the semi-solid media, to avoid disturbing the colonies and allowed to diffuse overnight.

Acknowledgements We would like to thank Dr Colin Porter (ICR, London) for providing the NIH3T3 and AM12(lacZ) cell lines and Joanna Povey for the gift of plasmid pREP8-⌬SCF. Thanks are also due to Professor David Linch (UCLMS, London) and Immunex, Seattle, WA, USA for their kind gifts of GM-CSF and SCF cDNA, respectively. This work was supported by grants from the MRC and the CGD Research Trust.

References 1 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. 2 Roe T et al. Integration of murine leukemia virus DNA depends on mitosis. EMBO J 1993; 12: 2099–2108. 3 Dunbar CE et al. Retrovirally marked CD34-enriched peripheral blood and bone marrow cells contribute to long-term engraftment after autologous transplantation. Blood 1995; 85: 3048–3057. 4 Brenner MK et al. Gene transfer into human hemopoietic progenitor cells. Br Med Bull 1995; 51: 167–191. 5 Fairbanks LD et al. Biochemical and immunological status following gene therapy and PEG-ADA therapy for adenosine deaminase (ADA) deficiency. Adv Exp Med Biol 1994; 370: 391–394. 6 Kohn DB et al. Engraftment of gene-modified umbilical cord blood cells in neonates with adenosine deaminase deficiency. Nature Med 1995; 1: 1017–1023. 7 Blaese RM et al. T lymphocyte-directed gene therapy for ADASCID: initial trial results after 4 years. Science 1995; 270: 475–480. 8 Bordignon C et al. Gene therapy in peripheral blood lymphocytes and bone marrow for ADA-immunodeficient patients. Science 1995; 270: 470–475. 9 Brenner MK et al. Gene marking to determine whether autologous marrow infusion restores long-term haemopoiesis in cancer patients. Lancet 1993; 342: 1134–1137. 10 Bagnis C et al. Retroviral transfer of the nlsLacZ gene into human CD34+ cell populations and into TF-1 cells: future prospects in gene therapy. Hum Gene Ther 1994; 5: 1325–1333. 11 Clapp DW et al. Myeloproliferative sarcoma virus directed expression of beta-galactosidase following retroviral transduction of murine hematopoietic cells. Exp Hematol 1995; 23: 630–638.


12 Povey J et al. Enhanced retroviral transduction of 5FU-resistant human bone marrow (stem) cells using a genetically modified packaging cell line. Blood 1998; 92: 4080–4089. 13 Anderson DM et al. Alternate splicing of mRNAs encoding human mast cell growth factor and localization of the gene to chromosome 12q22-q24. Cell Growth Diff 1991; 2: 373–378. 14 Toksoz D et al. Support of human hematopoiesis in long-term bone marrow cultures by murine stromal cells selectively expressing the membrane-bound and secreted forms of the human homolog of the steel gene product, stem cell factor. Proc Natl Acad Sci USA 1992; 89: 7350–7354. 15 Majumdar MK et al. Identification and mutation of primary and secondary proteolytic cleavage sites in murine stem cell factor cDNA yields biologically active, cell-associated protein. J Biol Chem 1994; 269: 1237–1242. 16 Kitamura T et al. Establishment and characterization of a unique human cell line that proliferates dependently on GM-CSF, IL-3, or erythropoietin. J Cell Physiol 1989; 140: 323–334. 17 Van Snick J. Interleukin-6: an overview. Ann Rev Immunol 1990; 8: 253–278. 18 Einerhand MP, Bakx TA, Valerio D. IL-6 production by retrovirus packaging cells and cultured bone marrow cells. Hum GeneTher 1991; 2: 301–306. 19 Kitamura T, Takaku F, Miyajima A. IL-1 up-regulates the expression of cytokine receptors on a factor-dependent human hemopoietic cell line, TF-1. Int Immunol 1991; 3: 571–577. 20 Leitner A et al. Lack of DNA synthesis among CD34+ cells in cord blood and in cytokine-mobilized blood. Br J Haematol 1996; 92: 255–262. 21 Hao QL et al. A functional comparison of CD34+ CD38− cells in cord blood and bone marrow. Blood 1995; 86: 3745–3753. 22 Hanley ME et al. Umbilical cord blood cell transduction by retroviral vectors: preclinical studies to optimize gene transfer. Blood Cells 1994; 20: 539–543. 23 Shi YJ et al. Comparative analysis of retrovirus-mediated gene transduction into CD34+ cord blood hematopoietic progenitors in the presence and absence of growth factors. Blood Cells 1994; 20: 517–524. 24 Walsh CE et al. Transduction of CD34-enriched human peripheral and umbilical cord blood progenitors using a retroviral vector with the Fanconi anemia group C gene. J Invest Med 1995; 43: 379–385. 25 Abe T et al. Transduction of retrovirus-mediated NeoR gene into CD34+ cells purified from granulocyte colony-stimulating factor (G-CSF)-mobilized infant and cord blood. Exp Hematol 1997; 25: 966–971.

26 Lu L et al. High efficiency retrovirus-mediated gene transduction into single isolated immature and replatable CD34(3+) hematopoietic stem/progenitor cells from human umbilical cord blood. J Exp Med 1993; 178: 2089–2096. 27 Bello-Fernandez C et al. Efficient retrovirus-mediated gene transfer of dendritic cells generated from CD34+ cord blood cells under serum-free conditions. Hum Gene Ther 1997; 8: 1651–1658. 28 Bodine DM et al. Development of a high-titre retrovirus producer cell line capable of gene transfer into rhesus monkey haematopoietic stem cells. Proc Natl Acad Sci USA 1990; 87: 3738–3742. 29 Hogge DE et al. Enhanced detection, maintenance, and differentiation of primitive human hematopoietic cells in cultures containing murine fibroblasts engineered to produce human steel factor, interleukin-3, and granulocyte colony-stimulating factor. Blood 1996; 88: 3765–3773. 30 Orlic D et al. Purification and characterization of heterogeneous pluripotent hematopoietic stem cell populations expressing high levels of c-kit receptor. Blood 1993; 82: 762–770. 31 Berardi AC et al. Functional isolation and characterization of human hematopoietic stem cells. Science 1995; 267: 104–108. 32 Flanagan JG, Chan DC, Leder P. Transmembrane form of the kit ligand growth factor is determined by alternative splicing and is missing in the Sld mutant. Cell 1991; 64: 1025–1035. 33 Huang EJ et al. Differential expression and processing of two cell associated forms of the kit-ligand: KL-1 and KL-2. Mol Biol Cell 1992; 3: 349–362. 34 Brannan CI et al. Steel-Dickie mutation encodes a c-kit ligand lacking transmembrane and cytoplasmic domains. Proc Natl Acad Sci USA 1991; 88: 4671–4674. 35 Bodine DM et al. Long-term in vivo expression of a murine adenosine deaminase gene in rhesus monkey hematopoietic cells of multiple lineages after retrovirus-mediated gene transfer into CD34+ bone marrow cells. Blood 1993; 82: 1975–1980. 36 Markowitz D, Goff S, Bank A. Construction and use of a safe and efficient amphotropic packaging cell line. Virology 1988; 167: 400–406. 37 Callard RE, Kotowicz KT, Smith SH. Assays for human B cell responses to cytokines. In: Balkwill FR (ed). Cytokines: A Practical Approach. IRL Press: Oxford, 1995, pp 143–177. 38 Wilkinson D, Green J. In situ hybridization and three-dimensional reconstruction of serial sections. In: Rickwood D, Cockcroft DL (eds). Postimplantation Mouse Embryos: A Practical Approach. IRL Press: Oxford, 1990, pp 155–171.

1091