Gene Therapy

Ectopic retroviral expression of LMO2, but not IL2Rγ, blocks human T-cell development from CD34+ cells: implications for leukemogenesis in gene therapy


The occurrence of leukemia in a gene therapy trial for SCID-X1 has highlighted insertional mutagenesis as an adverse effect. Although retroviral integration near the T-cell acute lymphoblastic leukemia (T-ALL) oncogene LIM-only protein 2 (LMO2) appears to be a common event, it is unclear why LMO2 was preferentially targeted. We show that of classical T-ALL oncogenes, LMO2 is most highly transcribed in CD34+ progenitor cells. Upon stimulation with growth factors typically used in gene therapy protocols transcription of LMO2, LYL1, TAL1 and TAN1 is most prominent. Therefore, these oncogenes may be susceptible to viral integration. The interleukin-2 receptor gamma chain (IL2Rγ), which is mutated in SCID-X1, has been proposed as a cooperating oncogene to LMO2. However, we found that overexpressing IL2Rγ had no effect on T-cell development. In contrast, retroviral overexpression of LMO2 in CD34+ cells caused severe abnormalities in T-cell development, but B-cell and myeloid development remained unaffected. Our data help explain why LMO2 was preferentially targeted over many of the other known T-ALL oncogenes. Furthermore, during T-cell development retrovirus-mediated expression of IL2Rγ may not be directly oncogenic. Instead, restoration of normal IL7-receptor signaling may allow progression of T-cell development to stages where ectopic LMO2 expression causes aberrant thymocyte growth.


Clinical gene therapy trials have shown marked progress during the last decade, in particular for treatment of hereditary diseases such as the primary immunodeficiency diseases.1, 2, 3, 4 Indeed, clinical benefit has been demonstrated in two types of severe combined immunodeficiencies, namely those caused by ADA and IL2RG gene defects.1, 2, 3, 4 The ADA gene has been considered a suitable target for many years, but only recently improvements in vector technology, stem cell mobilization and conditioning have led to clinical success.4 Similarly, two separate trials for SCID-X1 have shown the clinical feasibility of introducing a therapeutic gene into stem cells. The deficiency was restored and hematopoietic development was no longer blocked.1, 2, 3 The occurrence of leukemia in three patients in one of the gene therapy trials for SCID-X15, 6, 7 has emphasized insertional mutagenesis as an adverse effect. In the leukemic cells of these patients, retroviral integration occurred near the T-cell acute lymphoblastic leukemia (T-ALL) oncogene LIM-only protein 2 (LMO2)8 (third case reported at the European Society for Gene Therapy Meeting (ESGT) in Prague, 29 October to 1 November 2005).

The interleukin-2 receptor gamma chain (IL2Rγ), which is non-functional in SCID-X1, has been implicated as a cooperating oncogene to LMO2.9, 10 Two recent short reports using murine models have suggested that the IL2Rγ itself could possibly contribute to leukemic transformation.9, 10 In the first report, Copeland et al. show that insertions near Lmo2 and Il2rg were detected in a single-murine leukemia (tumor × 031) represented in the Mouse Retroviral Tagged Cancer Gene Database, which has led to the notion that these genes may act as cooperating oncogenes. However, several other oncogenes were targeted in the same leukaemic clone, a fairly common feature of replication competent marine leukaemia virus (MLV) used for these kinds of experiments. It is therefore not always clear which of the genes hit are functionally most important and whether cooperation between target genes functionally exists. For instance, the tumor clone with insertions in Lmo2 and Il2rg also contained insertions in Bmi1and Rap1gds1 (query date 21 July 2006). Both Bmi1, a repressor of Arf, and Rap1gds1 are known oncogenes and could therefore also have contributed to oncogenesis in these mice.11, 12, 13

In the second report, a murine model system was used in which stem cells were transduced with very high levels of human IL2RG using a lentiviral vector. These transplanted mice developed tumors with a high incidence, higher than reported for classical T-cell oncogenes such as LMO2 or LYL1 and comparable to the most active oncogene in transgenic mice, Tal1. This was a surprising finding because in human T-ALL, IL2RG has never been reported to act as oncogene. However, the very high levels of Il2RG in these mice compared to the slightly lower than normal levels obtained with retroviral MFG vectors in SCID-X1 trials may have contributed to the development of T-cell lymphomas in these mice. Similarly, insertional mutagenesis may have been a significant contributory factor.14 The report also did not indicate whether the tumors were clonal, expressed IL2RG, JAK3 was activated or the IL2RG gene might be mutated, for instance, owing to errors in lentiviral reverse transcription.

Genetic predisposition for tumor development has been proposed as an additional risk factor. Sorrentino et al.15 have recently developed a mouse model that is more prone to develop T-cell malignancies. In this model, both the Arf tumor-suppressor gene and the Il2rg gene were ablated. Retroviral transduction of IL2RG into the HSC of these mice emulates the high incidence of integration-dependent T-cell tumors than were found in the SCID-X1 trial. In parallel, the third patient who developed leukemia in the SCID-X1 trial had an insertion near BMI1, which represses ARF, whereas the family of the first leukemic patient had a history of juvenile tumor development.5 However, such a genetic predisposition does not explain the consistent latency of 30–34 weeks in the three cases of T-cell leukemia in SCID-X1 patients. All three patients had a retroviral insertion near LMO2. The consensus in the field is that the viral LTR caused aberrant expression of the LMO2 gene. Consequently, we have analyzed the mechanisms involved further by ectopic overexpression of LMO2 and IL2RG.

We used functional hematopoietic assays inducing differentiation of human CD34+ cells into myeloid, B, T and NK cell lineages to assess the effect of LMO2 and IL2RG overexpression. We demonstrate here that only the T-cell lineage is affected by LMO2 overexpression, whereas IL2Rγ overexpression has no detectable effects on T-cell development. As IL2RG was suggested to be oncogenic, we also assessed the expression of Il2RG, its associated cytokine receptors, LMO2 and all other classical T-ALL oncogenes in normal human HSC and thymocyte subpopulations in detail. We have presented a small part of the thymocyte data (without methods or figures) in a brief report elsewhere, but without the essential gene expression data nor B, NK and myeloid cell developmental assays.16 Our data on T-ALL oncogene expression shown here provide an explanation why LMO2 was preferentially targeted over other known T-ALL oncogenes. In addition, IL2Rγ may not act as an oncogene but by restoring IL7-receptor signaling, allows development of T-cells to stages where ectopic LMO2 expression hampers T-cell development in the thymus. This may create a pre-leukemic condition by accumulation of immature cells under intense proliferative pressure.

Materials and methods

Monoclonal antibodies

CD4-PE, CD16-PE, CD56-PE, CD8-APC, CD14-APC, CD19-APC and CD34-APC (BD Biosciences, Santa Clara, CA, USA), CD1a-RDI, CD13-RDI and CD33-RDI (Beckman Coulter, Fullerton, CA, USA).

Isolation and purification subsets from thymus and umbilical cord blood for gene expression profiling of normal human T-cell development

Material was handled and purified as described.17 Briefly, all stages of human T-cell development as well as CD34+umbilical cord blood (UCB) were isolated in duplicate. Thymi or cord blood cells from five individuals were pooled for each isolation. Purity of sorted population was determined on the FACS Calibur and shown to be >95% for all populations. When samples were less then 95% pure, they were excluded from further experiments.

Purification and stimulation of CD34+ umbilical cord blood and CD34+ peripheral blood stem cells for gene expression profiling

CD34+UCB cells were obtained from placentas of full-term pregnancies after informed consent according to legal regulations in the Netherlands. Mononuclear cells were isolated by Ficoll density gradient centrifugation (1.077 g/cm2; Nycomed Pharma AS, Oslo, Norway) and cryopreserved. After thawing, CD34+ cells from three to five individuals were isolated using magnetic beads (Miltenyi Biotech, Germany). Purity of the cells that were used was >90% as determined by flow cytometric analysis. CD34+ UCB cells were either lysed immediately or cultured for 2 and 4 days before isolating RNA. Cells were cultured in stem cell medium supplemented with 100 ng/ml rhSCF, 50 ng/ml rhFLT3-L and 10 ng/ml rhTPO as previously described.18, 19 Blood obtained from three individuals who were treated with cyclophosphamide and granulocyte-colony stimulating factor (G-CSF) was used as a source of peripheral blood stem cells (PBSC). Informed consent was obtained. The donors had the following characteristics: one healthy male donor, aged 35; one male myeloma patient in stable plateau phase, aged 53; one female myeloma patient in stable plateau phase, aged 48. CD34+ PBSC were isolated from using CliniMACS magnetic beads. Isolated CD34+ PBSC were cultured as was done for the two clinical trials for SCID-X1 and lysed for RNA isolation. Samples for all three individuals were stimulated and analyzed independently. Stimulation 1 was performed according to Cavazzana-Calvo et al.20 Briefly, cells were cultured at 0.5·106 cells/ml in ex vivo 10 medium containing 4% fetal calf serum, stem cell factor (SCF) (300 ng/ml), polyethylene glycol conjugated-megakaryocyte derived factor (PEC-MDF) (100 ng/ml), IL-3 (60 ng/ml) and FLT3-L (300 ng/ml) for 24 h. Then cells were cultured for an additional 3 days in the same media but with 50-ng/ml retronectin and 4-ng/ml protamine sulfate added. Stimulation 2 was performed according to Gaspar et al.1 Briefly, cells were cultured at 0.5 × 106 cells/ml in ex vivo 10 medium containing 1% human serum albumin, SCF (300 ng/ml), thrombopoietin (TPO) (100 ng/ml), IL-3 (20 ng/ml) and FLT3-L (300 ng/ml) for 40 h. After that cells were cultured for an additional 3 days in the same media but with 25-ng/ml retronectin added.

Gene expression profiling of CD34+ UCB cells and thymocytes

Microarray analysis was essentially performed as described previously17 and according to MIAME guidelines. For expression analysis of normal human T-cell development, the HG-U133A array from Affymetrix (22 283 probe sets) was used. For expression analysis of untreated (baseline) and stimulated CD34+ UCB cells as well as stimulated CD34+ PBSC, the HG-U133 Plus 2.0 Array from Affymetrix (54 675 probe sets) was used. For data analysis, probe intensity background was removed using robust multi-chip analysis.21 The intensity levels were quantile normalized.22, 23 Expression is considered absent when the arbitrary fluorescence value is below 80, corresponding to the cutoff given by MAS and GCOS software. The same criteria were used for making the absent–present calls for data extracted from UCB CD34+ cells, thymocytes as well as PBSC.

Production of MLV-based recombinant retrovirus

The retroviral plasmids LZRS-IRES-EGFP (control), LZRS-LMO2-IRES-EGFP and LZRS-IL2RG-IRES-EGFP were constructed and transfected into Phoenix amphotropic packaging cell lines using Fugene-6 transfection reagent. The LZRS vector has been described earlier.24 Stable high-titer producer clones were selected using puromycin (1 μg/ml). Recombinant virus-containing supernatant was harvested, filtered and stored at –80°C.

Transduction of CD34+ UCB cells

CD34+ UCB cells from three to five individuals were pooled and pre-stimulated in Iscove-modified Dulbecco's medium supplemented with 10% heat inactivated fetal bovine serum, 50 ng/ml rhSCF, 50 ng/ml rhFLT3L and 10 ng/ml rhTPO for 24 h. Recombinant retrovirus-containing supernatant was incubated in Retronectin-coated 35 mm Petri dishes for 2–3 h at 37°C. Transduction was performed using 0.33 × 106–0.5 × 106 cells per dish. Cells were cultured for two additional days in the medium as described above. Extra recombinant retrovirus-containing supernatant was added after 24 h. The efficiency of retroviral transduction was evaluated after 48 h by flow cytometry and ranged between 25 and 39% (EGFP), 8 and 32% (LMO2-IRES-EGFP) and 7 and 12% (IL2RG-IRES-EGFP). We checked LMO2 and IL2RG expression in CD34+ UCB cells transduced with LMO2-IRES-EGFP and IL2Rγ-IRES-EGFP, respectively, by real-time quantitative–polymerase chain reaction (RQ–PCR). Sorted EGFP+ cells expressed 11-fold more LMO2 RNA than sorted EGFP cells, as determined by RQ–PCR (not shown), whereas for IL2RG the difference was 12 times (not shown).

Real-time quantitative–polymerase chain reaction

For validation of microarray experiments on UCB samples, RQ–PCR was performed for LMO2, LYL1 and TAL1 according to standard procedures. Primers and probes are as follows. LMO2: FW-LMO2-EMC: IndexTermCAAACTGGGCCGGAAGC, RV-LMO2-EMC: IndexTermACCCGCATTGTCATCTCAT, probe (FAM) Tr-LMO2-EMC: IndexTermCAAAAAGCCTGAGATAGTCTCTCCGGCAG. LYL1: FW-LYL1-EMC: IndexTermCCCCTTCCTCAACAGTGTCTACA, RV-LYL1-EMC: IndexTermCTCCCGGCTGTTGGTGAA, probe (FAM) Tr-LYL1-EMC: IndexTermCTCACAGTGGCTTGGTCTCCGCTTC. TAL1: FW-TAL1-EMC: IndexTermCCGGATGCCTTCCCTATGT, RV-TAL1-EMC: IndexTermTCCCGGCTGTTGGTGAAGA, probe (FAM) T-TAL1-EMC: IndexTermAGACCTTCCCCCTATGAGATGGAGATTACTGAT. For validation of LMO2 expression using the LZRS-IRES-EGFP construct we used the primers and probes described above. Validation of IL2RG expression using the LZRS-IRES-EGFP construct was performed with primers and probes obtained from Applied Biosytems (assay on demand No. Hs 00173950_m1). For each sample, ABL was amplified as a control target: fw: IndexTermTGGAGATAACACTCTAAGCATAACTAAAGGT, rev: IndexTermGATGTAGTTGCTTGGGACCCA, probe (FAM-TAMRA): IndexTermCCATTTTTGGTTTGGGCTTCACACCATT.

Co-culture assays

Mouse–human hybrid fetal thymus organ cultures (FTOC) were essentially performed as described.25 The co-culture assays were performed using confluent cultures of the murine bone marrow (BM) stromal cell line S17 plated in 12-well plates. The cell line was co-cultured with 10 000 CD34+ UCB cells per well. In some experiments, a part of the cells was taken from the co-culture after 1 week and tested for their ability to develop along the NK cell lineage in suspension culture in the presence of 100 units rhIL2.

Fetal thymus organ cultures

Fetal thymic lobes from littermates were obtained from C57Bl/6 mice on day 14 of gestation and endogenous thymocytes were depleted by γ irradiation (14.4. Gy). Hanging drops were created in Terasaki plates by adding one thymic lobe to a 25-μl cell suspension containing 30 000–40 000 transduced CD34+ UCB cells. Inverted Terasaki plates were incubated for 2 days in a humidified incubator (5% CO2 in air, 37°C) to facilitate the entry of UCB cells into the thymic lobes. Subsequently, lobes were transferred to Nuclepore filters, which were layered over 1 ml medium in 12-well plates. Starting at day 7, 10-ng/ml rhIL-7 was added to the culture medium twice a week. Cultures were incubated at 37°C in 5% CO2 for nine or 21 days after which lobes were mechanically disrupted into single-cell suspensions and flow cytometric analysis was performed.


We have previously reported global gene expression profiling of CD34+lin- UCB cells and eight consecutive stages of human thymic T-cell development.17 We mined these data for the transcription levels of several well-known T-ALL oncogenes in CD34+ progenitor cells. When normalizing expression against all probe sets on the array, in unstimulated CD34+ progenitor cells from UCB, LMO2 was expressed at the highest level, followed by LYL1, TAN1, HOX11 and TAL1. LCK was expressed at low, but detectable levels (Figure 1a). Other T-ALL oncogenes investigated were not transcribed at detectable levels. Expression was considered absent when the arbitrary fluorescence value decreased below 80, corresponding to the cutoff given by MAS and GCOS software.

Figure 1

Oncogene expression in human UCB and thymocytes. (a) Expression of T-ALL oncogene transcripts in unstimulated (day 0) and stimulated (day 2 and day 4) UCB CD34+ cells analyzed on Affymetrix HG- U133 Plus 2.0 Arrays. Stimulated: growth factor stimulation using Flt3-L, TPO and SCF. Data were normalized against all probe sets present on the array. Expression is considered absent when the arbitrary fluorescence value is below 80, corresponding to the cutoff given by MAS and GCOS software. (b) Expression of LMO2, TAL1 and LYL1 relative to ABL. The left panel represents data extracted from the microarray analysis described in Figure 1a. The right panel shows the average expression of LMO2, TAL1 and LYL1 relative to ABL as determined by RQ–PCR in two independent experiments. (c) Expression of T-ALL oncogene transcripts in CD34+ PBSC stimulated with Flt3-L, TPO, Il3 and SCF analyzed on Affymetrix HG- U133 Plus 2.0 Arrays. Data were normalized against all probe sets present on the array. Stimulation 1 was performed according to Cavazzana-Calvo et al.20 and stimulation 2 was performed according to Gaspar et al.1 (d) Expression of T-ALL oncogene transcripts during normal human T-cell development and progenitor cells analyzed on Affymetrix HG-U133A Arrays. Stages of T-cell development are plotted on the X-axis, from immature to mature thymocytes.

CD34+ cells used in gene therapy protocols are routinely stimulated with growth factors before and during the gene transfer procedure to facilitate transduction of progenitor cells.1, 4, 18, 19, 20 After stimulation with TPO, Flt3L and SCF, transcription of LMO2, LYL1, TAL1 and TAN1 was most pronounced, whereas the levels of LCK and HOX11 decreased (Figure 1a). The array data were validated by RQ–PCR for the LMO2, LYL1 and TAL1 genes (Figure 1b, right panel). Average values of two independent experiments are given. Results of the RQ–PCR were comparable to the array data, here normalized to ABL expression (Figure 1b, left panel). Recent studies by Bushman et al.26 indicate that transcriptional activity correlates with integration site selection. This notion was confirmed by other studies from this laboratory showing that retroviral integration closely parallels the transcriptional expression as measured by ESTs (i.e. transcriptional activation).27 Our studies in primary murine stem cells also demonstrated a strong correlation between the number of viral insertions and expression levels of immediately neighboring genes (Brugman et al., Mol Ther, 2006, 13, p S37, abstract). Although this correlation may not necessarily be valid for any given individual gene, it seems likely that genes that are actively transcribed (called present on the array) have a high probability of being located in euchromatin, making it more accessible for viral insertion. Therefore, these results predict that LMO2 is a likely oncogene to be targeted by retroviral integration, along with LYL1, TAL1 and TAN1.

CD34+ UCB cells are often used in research settings for gene therapy. However, in clinical SCID-X1 gene therapy, BM of very young infants is used. As it is difficult to obtain age-matched normal BM aspirates for research purposes, we have analyzed the expression of T-ALL oncogenes in G-CSF mobilized PBSC after in vitro stimulation to corroborate the UCB expression data. When CD34+ PBSC were stimulated according to transduction protocols used for the two published clinical SCID-X1 gene therapy trials,1, 20 the same pattern of T-ALL oncogene transcriptional activation was observed as was found in the primary as well as cytokine-stimulated CD34+ UCB cells (Figure 1c). The average expression values of CD34+ cells from three individuals are shown. These results may indicate that CD34+ cells display a similar pattern of proto-oncogene expression irrespective of their source. As the gene expression profiles of CD34+ cells from BM, PBSC and UCB are highly similar, with UCB being slightly more similar to BM,28 we would predict that the oncogenes LMO2, LYL1, TAL1 and TAN1 are also transcriptionally active in BM CD34+ cells.

As all three adverse events in the SCID-X1 trial presented as uncontrolled clonal T-cell proliferations, it was also of interest to study the expression of T-ALL oncogenes during normal human T-cell development. Compared with CD34+ UCB cells, the most immature cells in the thymus (CD34+ CD38−) showed a decreased expression of LMO2, similar to TAL1, but LYL1 expression was maintained until the CD4+CD8+ DP stage (Figure 1d). Thus, LMO2 expression is highest in CD34+ progenitor cells and rapidly decreases during T-cell differentiation, suggesting that ectopic expression of LMO2 at later stages of development may have adverse effects on T-cell development.

The effects of LMO2 overexpression have not been studied in human hematopoiesis. We utilized two culture systems to study ectopic LMO2 expression, schematically depicted in Figure 2. The murine BM stroma cell line (S17), supplemented with IL7 and SCF allows simultaneous development of human CD34+ UCB into B-lymphocytes and myeloid cells, and the mouse–human hybrid FTOC was used to study T-cell development. This assay allows for evaluation of human T-cell development ex vivo, but variability of individual human samples results in a relatively high inter-experiment variability, and quantitative assessments are most reliably performed by intra-experiment comparisons, which is the approach chosen here. The LMO2 transgene was expressed via an MFG-based retroviral vector, together with an IRES-EGFP cassette to allow easy detection of transduced cells.24

Figure 2

Schematic overview of differentiation assays. Development of human CD34+ UCB into B-lymphocytes and myeloid cells was studied using the murine BM stroma cell line (S17), supplemented with human recombinant IL7 and SCF. The mouse–human hybrid FTOC was used to study T-cell development. Transduced CD34+ UCB cells were used as input material for both assays. The LMO2 or IL2RG transgene were expressed via an MFG-based retroviral vector, together with an IRES-EGFP cassette allowing for easy detection of transduced cells.

CD34+ UCB cells were transduced with 8–32% efficiency. Expressing LMO2 using the MFG-based LZRS vector resulted in a 11-fold increased level of LMO2 mRNA as measured by RQ–PCR. When co-cultured with the stromal cell line S17, ectopic expression of LMO2 did not interfere with the development of CD34+ UCB cells into the B-cell lineage, as the percentage of CD19-positive cells did not differ from the control-transduced cells (4 vs 6%, respectively, Figure 3a). Similarly, the development of the myelo–monocytic cells was not affected, as illustrated by staining for CD13/CD33 and CD14. About one-third of the cells were positive for the myeloid markers CD13/CD33 (34 vs 33%) in both cultures, whereas 5% of the cells were positive for CD14. In parallel, part of the transduced cells was removed from co-culture after 1 week and cultured in suspension for an additional 14 days in the presence of IL2 to assay for the ability to develop into the NK lineage. Both LMO2 as well as control-transduced cells showed an equal percentage of CD16/56-positive cells indicating equal NK cell development. Thus, LMO2 overexpression does not affect the development into B, NK and myelo–monocytic lineages.

Figure 3

The effects of ectopic LMO2 expression on B, myelo–monocytic, NK and T lineage development. (a) Flow cytometric analysis of co-culture experiments after 12 days of culture. The FACS plots display CD45 and EGFP-positive cells in the lymphocyte gate. Numbers represent the percentage of cells within the quadrant. Representative results from 1 out of 3 co-culture experiments are shown. (bd) The FACS plots show cells in the lymphocyte gate. Numbers represent the percentage of cells within the quadrant. Representative results for 1 or 2 out of 5 experiments are shown. (b) Flow cytometric analysis of FTOC at day 9. (c) and (d) Flow cytometric analysis of FTOC at day 21.

A relatively high inter-experiment variability is seen in FTOC experiments due to variability of individual human samples. Quantitative assessments were, therefore, performed using intra-experiment comparisons. In FTOC experiments at day 9 of culture, LMO2-expressing cells showed a lower percentage of the early thymocyte marker CD1a29, 30 when compared with untransduced cells derived from the same culture (Figure 3b). In addition, in comparison with untransduced cells and control-transduced cells, a higher percentage of LMO2-transduced cells retained CD34 expression (12 vs 2–4%, representative of five different FTOC experiments conducted). These data indicate that progenitor cells ectopically expressing LMO2 are hampered in the initial stages of T-cell development.

Later in T-cell development (day 21 of culture), the total percentage of CD4+ cells was drastically reduced in LMO2-transduced cells compared with untransduced cells (Figure 3c), mostly because of the reduced percentage of CD4 single-positive cells (ISP), that is 57% in EGFP-transduced controls vs 23% in LMO2-IRES-EGFP transduced cells. Despite differences in efficiency of development, other experiments also demonstrated this reduction in ISP (Figure 3c, bottom panel). These results are representative of five different FTOC experiments conducted. Control-transduced and untransduced (EGFP-) cells displayed similar frequencies of CD4+ and CD8+ cells. Therefore, ectopic LMO2 expression led to an incomplete inhibition of T-cell development in the CD34+ DN stages and a severe block at the ISP stage. Apparently, as human T-cell development advances, differentiation is increasingly hampered by ectopic LMO2 expression.

It has been suggested that IL2RG, the gene involved in SCID-X1, could possibly contribute to leukemic transformation.9 We therefore investigated the effect of retrovirus-mediated expression of IL2RG on T-cell development. In parallel with the experiments described in Figure 3c, FTOC was performed using CD34+ cells that were transduced with the MFG-based virus containing the IL2RG gene together with IRES-EGFP, which resulted in 12-fold increased IL2RG mRNA as measured by RQ–PCR. Thymocytes derived from the cultures were analyzed for CD4 and CD8 expression (Figure 3d). The percentages of CD4 single positives (37 vs 38%) as well as CD4/CD8 double positives (6 vs 9%) were similar in untransduced and IL2RG-transduced cell populations. We therefore conclude that retroviral-mediated expression of IL2RG does not affect the development of T-cells. We also investigated whether cells expressing IL2RG under a retroviral promoter have a higher proliferation rate. When comparing percentages of transduced cells after culture with percentages of EGFP-positive cells at the start of the FTOC, a proliferative advantage was not apparent (data not shown).

The IL2Rγ chain pairs with other cytokine receptor chains (receptors for IL2, IL4, IL7, IL9, IL15, IL21) to form functional signal-transducing entities. We analyzed the transcription levels of IL2RG and its partner genes throughout T-cell development (Figure 4). The only chains expressed in developing T cells are IL2RG and IL7RA. Mature thymocytes also express IL2RB, but not IL2RA, the gene product that regulates responses via the full IL2R complex. Judging from the expression pattern, the only signaling events in which IL2Rγ could be involved during the course of T-cell development occurs through the IL7R, in accordance with the known role of both IL7 and IL7Rα in T-cell development.31, 32 Therefore, we conclude that the main effect of IL2RG gene therapy is restoration of IL7-driven signals during thymic development.

Figure 4

Expression of LMO2, IL2RG and associated IL receptor chains during normal human T-cell development. Stages of T-cell development are plotted on the X-axis, from immature to mature. Transcription levels are plotted on the Y-axis in arbitrary units (AU).


Recently, the oncogenic risk of retrovirus-mediated gene transfer has received major attention as a safety concern, and has been approached in various experimental model systems, including murine- and non-human primate studies.33, 34 In the SCID-X1 trial, in which severe adverse effects were reported, the T-ALL oncogene LMO2 was ectopically expressed due to insertional mutagenesis. In this study, we investigated the expression of LMO2 and other T-ALL oncogenes during hematopoiesis, in particular the various stages of human T-cell development. As retroviral insertions predominantly occur in actively transcribed genes,26, 27 high expression of LMO2 in immature CD34+ cells make this gene a likely target for deregulated expression through insertional mutagenesis, particularly since this high LMO2 expression is maintained by cytokines that are used in gene therapy protocols. We show here that overexpression of LMO2 is well tolerated in all major leukocyte lineages, with the exception of the T-cell lineage. Ectopic LMO2 expression hampers T-cell development in the thymus, creating a likely preleukemic condition by accumulation of immature cells under intense proliferative pressure. We speculate that such an accumulation at an early stage of T-cell development of the IL2RG-restored cells may have contributed to the apparently uniform latency time of 30–34 weeks.

We have argued elsewhere that murine and human T-cell development is highly similar at the genome-wide level.17, 35 However, at the level of individual genes there are differences in expression and function. For instance, defects in the two genes involved in T-negative SCID, ZAP70 and IL2RG, have species-specific phenotypes when deficient. The Il2rg-deficient mice lack T, B and NK cells, whereas the human SCID-X1 patients usually have normal or increased numbers of B cells. Thus, murine studies should be interpreted with care when extrapolating to human diseases.

In the introduction of this paper we referred to two recent murine studies that have suggested that Il2rg may be a direct oncogene. We argued that both studies are hampered by an experimental design that makes extrapolation to the leukemias that arose in the SCID-X1 trial problematic. It is important to note that the phenotype of the murine tumors found by Woods et al.10 is very different (B220+CD3+) from that of the T-ALL like tumors that occurred in the patients enroled in the SCID-X1 trial. In our view, the retroviral-mediated expression of IL2RG in hematopoietic precursors does not represent a preleukemic or otherwise leukemogenic event because this chain is also highly expressed in normal hematopoietic CD34+ precursors as well as in all subsets of T-cell development (Figure 4). In addition, signaling through IL2Rγ can only occur when the chain is in close proximity to a partner chain (or multiple chains) bound to its specific ligand, resulting in transphosphorylation of JAK proteins. Therefore, it is unclear how the extremely high expression of IL2Rγ alone could lead to tumor development in the study by Woods et al. Furthermore, IL2Rγ expression in treated SCID patients was within the normal range, although this was only measured in peripheral blood T cells.8, 20 Persistent activation of JAK3 was also not detected in these patients indicating normal function of IL2Rγ in the setting of retroviral expression.8 In addition, a brief report described Il2RG transgenic mice that overexpress human Il2RG using the CD2 promoter. In 54 out of 54 mice analyzed, no T-cell lymphomas were found.36 Taken together, these observations argue against IL2Rγ acting as a cooperative oncogene in the human gene therapy setting.

Most T-negative SCID have an early developmental block like the one found in IL2Rγ deficiency. Thus, we would predict that gene therapy aimed at other T-negative SCID, such as deficiencies in the RAG1, RAG2, IL7RA or JAK3 genes, would also present with the risk of leukemia development as long as retroviral vectors with high incidence of insertional mutagenesis are used.

Since LMO2 expression is normally completely downregulated at the transition of the CD34+CD1a− to CD34+CD1a+ stage (Figure 1b), ectopic LMO2 expression is apparently tolerated to some extent in the early stages of T-cell development, although it hampers further development resulting in an incomplete developmental block. The block in early human T-cell development is in line with data from LMO2-transgenic mice.37, 38 In mice, the initial block is at the DN stage, similar to the first abnormalities in human development we report here, at the CD34+CD1a+ stage, the equivalent of murine DN3.35

As the most immature thymocytes undergo extensive proliferation (up to 10 000 fold),39 any restoration of proliferative signals will have a tremendous selective advantage. The LMO2-induced arrest at a stage of differentiation in which proliferative pressure is high may generate a pre-leukemic state in which additional genetic lesions could lead to full-blown T-ALL, for instance through mutations in the Notch1 gene.40 Ironically, by restoring T-cell development, IL2RG retroviral gene therapy allows LMO2 to act as an oncogene for T-ALL development. Of course, this scenario will only occur in those cases in which normal regulation of LMO2 expression is disrupted by nearby retroviral insertion.

We have used DNA microarrays and quantitative PCR to survey expression of a large number of genes in various subpopulations and their quantitative expression levels are useful in predicting which genes are likely targets for insertional mutagenesis. Therefore, a corollary of our findings is that it may be useful to include gene expression profiling of the target cells in gene therapy to assess which oncogenes are transcribed and thereby represent potential targets for insertional oncogenesis. Current efforts to generate self-inactivating (SIN) vectors may overcome some of the problems associated with insertional mutagenesis41, 42 as these vectors may not transactivate nearby cellular genes as strongly. Finally, we find no evidence for involvement of overexpression of the therapeutic IL2RG gene in abnormal T-cell development. Our results also predict that in other gamma retrovirus-mediated gene therapy efforts directed at restoring T-cell development (ADA, RAG1, RAG2), cells with insertions near the LMO2 gene might result in similar aberrant LMO2 expression.


  1. 1

    Gaspar HB, Parsley KL, Howe S, King D, Gilmour KC, Sinclair J et al. Gene therapy of X-linked severe combined immunodeficiency by use of a pseudotyped gamma retroviral vector. Lancet 2004; 364: 2181–2187.

    CAS  Article  Google Scholar 

  2. 2

    Hacein-Bey-Abina S, Fischer A, Cavazzana-Calvo M . Gene therapy of X-linked severe combined immunodeficiency. Int J Hematol 2002; 76: 295–298.

    CAS  Article  Google Scholar 

  3. 3

    Hacein-Bey-Abina S, Le Deist F, Carlier F, Bouneaud C, Hue C, De Villartay JP et al. Sustained correction of X-linked severe combined immunodeficiency by ex vivo gene therapy. N Engl J Med 2002; 346: 1185–1193.

    CAS  Article  Google Scholar 

  4. 4

    Aiuti A, Slavin S, Aker M, Ficara F, Deola S, Mortellaro A et al. Correction of ADA-SCID by stem cell gene therapy combined with nonmyeloablative conditioning. Science 2002; 296: 2410–2413.

    CAS  Article  Google Scholar 

  5. 5

    Hacein-Bey-Abina S, von Kalle C, Schmidt M, Le Deist F, Wulffraat N, McIntyre E et al. A serious adverse event after successful gene therapy for X-linked severe combined immunodeficiency. N Engl J Med 2003; 348: 255–256.

    Article  Google Scholar 

  6. 6

    Marshall E . Gene therapy. Second child in French trial is found to have leukemia. Science 2003; 299: 320.

    CAS  Article  Google Scholar 

  7. 7

    Check E . Gene therapy put on hold as third child develops cancer. Nature 2005; 433: 561.

    Google Scholar 

  8. 8

    Hacein-Bey-Abina S, Von Kalle C, Schmidt M, McCormack MP, Wulffraat N, Leboulch P et al. LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science 2003; 302: 415–419.

    CAS  Article  Google Scholar 

  9. 9

    Dave UP, Jenkins NA, Copeland NG . Gene therapy insertional mutagenesis insights. Science 2004; 303: 333.

    Article  Google Scholar 

  10. 10

    Woods NB, Bottero V, Schmidt M, von Kalle C, Verma IM . Gene therapy: therapeutic gene causing lymphoma. Nature 2006; 440: 1123.

    CAS  Article  Google Scholar 

  11. 11

    Foot AB, Oakhill A, Kitchen C . Acute monoblastic leukemia of infancy in Klinefelter's syndrome. Cancer Genet Cytogenet 1992; 61: 99–100.

    CAS  Article  Google Scholar 

  12. 12

    Lessard J, Sauvageau G . Bmi-1 determines the proliferative capacity of normal and leukaemic stem cells. Nature 2003; 423: 255–260.

    CAS  Article  Google Scholar 

  13. 13

    Hussey DJ, Nicola M, Moore S, Peters GB, Dobrovic A . The (4;11)(q21;p15) translocation fuses the NUP98 and RAP1GDS1 genes and is recurrent in T-cell acute lymphocytic leukemia. Blood 1999; 94: 2072–2079.

    CAS  Google Scholar 

  14. 14

    Woods NB, Bottero V, Schmidt M, von Kalle C, Verma IM . Gene therapy: is IL2RG oncogenic in T-cell development?: X-SCID transgene leukaemogenicity (reply). Nature 2006; 443: E6–E7.

    CAS  Article  Google Scholar 

  15. 15

    Shou Y, Ma Z, Lu T, Sorrentino BP . Unique risk factors for insertional mutagenesis in a mouse model of XSCID gene therapy. Proc Natl Acad Sci USA 2006; 103: 11730–11735.

    CAS  Article  Google Scholar 

  16. 16

    Pike-Overzet K, de Ridder D, Weerkamp F, Baert MR, Verstegen MM, Brugman MH et al. Gene therapy: is IL2RG oncogenic in T-cell development? Nature 2006; 443: E6–E7, (E5 discussion).

    Article  Google Scholar 

  17. 17

    Dik WA, Pike-Overzet K, Weerkamp F, de Ridder D, de Haas EF, Baert MR et al. New insights on human T cell development by quantitative T cell receptor gene rearrangement studies and gene expression profiling. J Exp Med 2005; 201: 1715–1723.

    CAS  Article  Google Scholar 

  18. 18

    van Hennik PB, Verstegen MM, Bierhuizen MF, Limon A, Wognum AW, Cancelas JA et al. Highly efficient transduction of the green fluorescent protein gene in human umbilical cord blood stem cells capable of cobblestone formation in long-term cultures and multilineage engraftment of immunodeficient mice. Blood 1998; 92: 4013–4022.

    CAS  Google Scholar 

  19. 19

    Wognum AW, Visser TP, Peters K, Bierhuizen MF, Wagemaker G . Stimulation of mouse bone marrow cells with kit ligand, FLT3 ligand, and thrombopoietin leads to efficient retrovirus-mediated gene transfer to stem cells, whereas interleukin 3 and interleukin 11 reduce transduction of short- and long-term repopulating cells. Hum Gene Ther 2000; 11: 2129–2141.

    CAS  Article  Google Scholar 

  20. 20

    Cavazzana-Calvo M, Hacein-Bey S, de Saint Basile G, Gross F, Yvon E, Nusbaum P et al. Gene therapy of human severe combined immunodeficiency (SCID)-X1 disease. Science 2000; 288: 669–672.

    CAS  Article  Google Scholar 

  21. 21

    Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003; 4: 249–264.

    Article  Google Scholar 

  22. 22

    Bolstad BM, Irizarry RA, Astrand M, Speed TP . A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 2003; 19: 185–193.

    CAS  Article  Google Scholar 

  23. 23

    de Ridder D, Staal FJ, van Dongen JJ, Reinders MJ . Maximum significance clustering of oligonucleotide microarrays. Bioinformatics 2006; 22: 326–331.

    CAS  Article  Google Scholar 

  24. 24

    Kinsella TM, Nolan GP . Episomal vectors rapidly and stably produce high-titer recombinant retrovirus. Hum Gene Ther 1996; 7: 1405–1413.

    CAS  Article  Google Scholar 

  25. 25

    Weerkamp F, Baert MR, Brugman MH, Dik WA, de Haas EF, Visser TP et al. Human thymus contains multipotent progenitors with T/B lymphoid, myeloid, and erythroid lineage potential. Blood 2006; 107: 3131–3137.

    CAS  Article  Google Scholar 

  26. 26

    Schroder AR, Shinn P, Chen H, Berry C, Ecker JR, Bushman F . HIV-1 integration in the human genome favors active genes and local hotspots. Cell 2002; 110: 521–529.

    CAS  Article  Google Scholar 

  27. 27

    Mitchell RS, Beitzel BF, Schroder AR, Shinn P, Chen H, Berry CC et al. Retroviral DNA integration: ASLV, HIV, and MLV show distinct target site preferences. PLoS Biol 2004; 2: E234.

    Article  Google Scholar 

  28. 28

    Ng YY, van Kessel B, Lakhorst HM, Baert MR, van den Burg CM, Bloem AC et al. Gene expression profiling of CD34+ cells from various hematopoietic stem-cell sources reveals functional differences in stem-cell activity. J Leukoc Biol 2004; 75: 314–323.

    CAS  Article  Google Scholar 

  29. 29

    Spits H, Blom B, Jaleco AC, Weijer K, Verschuren MC, van Dongen JJ et al. Early stages in the development of human T, natural killer and thymic dendritic cells. Immunol Rev 1998; 165: 75–86.

    CAS  Article  Google Scholar 

  30. 30

    Weerkamp F, de Haas EF, Naber BA, Comans-Bitter WM, Bogers AJ, van Dongen JJ et al. Age-related changes in the cellular composition of the thymus in children. J Allergy Clin Immunol 2005; 115: 834–840.

    Article  Google Scholar 

  31. 31

    Puel A, Ziegler SF, Buckley RH, Leonard WJ . Defective IL7R expression in T(-)B(+)NK(+) severe combined immunodeficiency. Nat Genet 1998; 20: 394–397.

    CAS  Article  Google Scholar 

  32. 32

    Roifman CM, Zhang J, Chitayat D, Sharfe N . A partial deficiency of interleukin-7R alpha is sufficient to abrogate T-cell development and cause severe combined immunodeficiency. Blood 2000; 96: 2803–2807.

    CAS  Google Scholar 

  33. 33

    Li Z, Dullmann J, Schiedlmeier B, Schmidt M, von Kalle C, Meyer J et al. Murine leukemia induced by retroviral gene marking. Science 2002; 296: 497.

    CAS  Article  Google Scholar 

  34. 34

    Calmels B, Ferguson C, Laukkanen MO, Adler R, Faulhaber M, Kim HJ et al. Recurrent retroviral vector integration at the Mds1/Evi1 locus in nonhuman primate hematopoietic cells. Blood 2005; 106: 2530–2533.

    CAS  Article  Google Scholar 

  35. 35

    Weerkamp F, Pike-Overzet K, Staal FJ . T-sing progenitors to commit. Trends Immunol 2006; 27: 125–131.

    CAS  Article  Google Scholar 

  36. 36

    Thrasher AJ, Gaspar HB, Baum C, Modlich U, Schambach A, Candotti F et al. Gene therapy: X-SCID transgene leukaemogenicity. Nature 2006; 443: E5–E6, (discussion E6–E7).

    CAS  Article  Google Scholar 

  37. 37

    Larson RC, Osada H, Larson TA, Lavenir I, Rabbitts TH . The oncogenic LIM protein Rbtn2 causes thymic developmental aberrations that precede malignancy in transgenic mice. Oncogene 1995; 11: 853–862.

    CAS  Google Scholar 

  38. 38

    Neale GA, Rehg JE, Goorha RM . Ectopic expression of rhombotin-2 causes selective expansion of CD4-CD8- lymphocytes in the thymus and T-cell tumors in transgenic mice. Blood 1995; 86: 3060–3071.

    CAS  Google Scholar 

  39. 39

    Allman D, Sambandam A, Kim S, Miller JP, Pagan A, Well D et al. Thymopoiesis independent of common lymphoid progenitors. Nat Immunol 2003; 4: 168–174.

    CAS  Article  Google Scholar 

  40. 40

    Weng AP, Ferrando AA, Lee W, Morris JPt, Silverman LB, Sanchez-Irizarry C et al. Activating mutations of NOTCH1 in human T cell acute lymphoblastic leukemia. Science 2004; 306: 269–271.

    CAS  Article  Google Scholar 

  41. 41

    Trono D . Virology. Picking the right spot. Science 2003; 300: 1670–1671.

    CAS  Article  Google Scholar 

  42. 42

    Baum C, von Kalle C, Staal FJ, Li Z, Fehse B, Schmidt M et al. Chance or necessity? Insertional mutagenesis in gene therapy and its consequences. Mol Ther 2004; 9: 5–13.

    CAS  Article  Google Scholar 

Download references


We thank Dr Langerak, Dr van Velden and Dr Dik for critically reading the manuscript. In addition, we are grateful for the primers and probes made available by Dr Dik. We also thank E de Haas for performing the sorting and Dr Kwee Yong, Department of Haematology, University College London, for supplying the PBSC. This work was supported in part by the 5th and 6th EU Framework program (Contract Nos. QLK3-CT-2001-0427 (INHERINET) and LSHB-CT-2004-005242 (CONSERT), as well as by the Translational Gene Therapy Research Programme of ZonMw – the Netherlands Organization for Health Research and Development. AJT is supported by the Wellcome Trust.

Author information



Corresponding author

Correspondence to F J T Staal.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Pike-Overzet, K., de Ridder, D., Weerkamp, F. et al. Ectopic retroviral expression of LMO2, but not IL2Rγ, blocks human T-cell development from CD34+ cells: implications for leukemogenesis in gene therapy. Leukemia 21, 754–763 (2007).

Download citation


  • gene therapy
  • IL2Rγ
  • LMO2

Further reading