Insertional activation of cellular proto-oncogenes by replication-defective retroviral vectors can trigger clonal dominance and leukemogenesis in animal models and clinical trials. Here, we addressed the leukemogenic potential of vectors expressing interleukin-2 receptor common γ-chain (IL2RG), the coding sequence required for correction of X-linked severe combined immunodeficiency. Similar to conventional γ-retroviral vectors, self-inactivating (SIN) vectors with strong internal enhancers also triggered profound clonal imbalance, yet with a characteristic insertion preference for a window located downstream of the transcriptional start site. Controls including lentivirally transduced cells revealed that ectopic IL2RG expression was not sufficient to trigger leukemia. After serial bone marrow transplantation involving 106 C57Bl6/J mice monitored for up to 18 months, we observed leukemic progression of six distinct clones harboring γ-retroviral long terminal repeat (LTR) or SIN vector insertions in Evi1 or Prdm16, two functionally related genes. Three leukemic clones had single vector integrations, and identical clones manifested with a remarkably similar latency and phenotype in independent recipients. We conclude that upregulation of Evi1 or Prdm16 was sufficient to initiate a leukemogenic cascade with consistent intrinsic dynamics. Our study also shows that insertional mutagenesis is required for leukemia induction by IL2RG vectors, a risk to be addressed by improved vector design.
Gene transfer by retroviral vectors has demonstrated its potential for the correction of genetic disorders of the hematopoietic system.1, 2, 3, 4, 5 However, the semi-random vector insertions bear the risk of activating cellular genes by enhancer/promoter sequences contained in the vector. Transforming side effects of γ-retroviral gene transfer were observed in mice and nonhuman primates6, 7, 8 and also in gene therapy trials treating X-linked severe combined immunodeficiency (SCID-X1) patients.9, 10, 11 The insertional activation of cellular genes also influences the growth of hematopoietic cell clones, thus triggering clonal dominance.4, 12 Mapping of retroviral insertion sites in dominant clones could identify gene networks that support hematopoietic stem cell (HSC) growth in vivo.13 Clonal dominance was also observed in patients of the first chronic granulomatous disease (CGD) gene therapy trial, in association with insertions in genes encoding HSC-associated transcription factors and proto-oncogenes MDS-EVI1, PRDM16 and SETBP1,4 and similar observations were made in other clinical trials.14, 15, 16
Efforts are being made to develop ‘safer’ retroviral vectors that are less capable of activating neighboring cellular genes. Using a novel in-vitro immortalization assay, we showed that the deletion of the U3 region in the viral LTRs and the use of the same enhancer/promoter as internal promoter in a ‘self-inactivating’ (SIN) architecture rendered the vector less transforming.17, 18 However, immortal clones were also recovered after gene transfer with SIN retroviral vectors containing the spleen focus-forming virus (SFFV) U3 region as internal enhancer/promoter.17, 18
In-vivo assay systems were developed that can be used to analyze the leukemogenicity of retroviral gene transfer in normal C57Bl/6 mice8 and tumor-prone mouse models.19, 20 Further advancing nonclinical studies of insertional genotoxicity, gene knockout models may reveal disease-specific risk factors connected to a certain gene therapy approach. In a combined SCID-X1 and tumor suppressor knockout mouse model (IL2rg−/−/Arf−/−), an increased risk of tumor development after gene transfer of IL2RG (also addressed as common γ-chain, γc) was demonstrated when using a γ-retroviral vector (MFG.γc), containing a strong enhancer/promoter in the LTR.20 Similar to the situation of the first SCID-X1 clinical trials, selection of leukemic clones occurred, arguing for the contribution of insertional mutagenesis to the leukemic outcome. However, the oncogenic effect of IL2RG expression alone remained unclear. An active role of ectopic expression of common γc in leukemia induction was proposed in a study that used a lentiviral vector to overexpress murine IL2rg in wild-type bone marrow (BM) cells and found a high incidence of lymphoblastic tumors after prolonged follow-up.21 Unfortunately, information about the potential involvement of insertional mutagenesis and the number of insertions in the tumors was limited. Others challenged the concept that ectopic expression of IL2RG is sufficient for leukemogenesis,22, 23, 24 but conclusive results from animal models with long-term follow-up remained to be shown.
In the present study, we addressed the impact of vector design (γ-retroviral LTR, γ-retroviral SIN or lentiviral SIN) and transgene (IL2RG versus DsRED2 fluorescent protein) on leukemia induction after gene delivery into hematopoietic cells of C57Bl6/J mice. We chose wild-type C57Bl6/J mice to avoid a selective advantage introduced by restored expression of IL2RG. All vectors, except the clinically used vector MFG.γc, contain the same retroviral enhancer/promoter located either in the LTR or in an internal position of a vector with a SIN design.
Material and methods
Retroviral vectors and vector production
The retroviral vector SF91.DsRED pre was constructed as previously reported. In brief, SF91 is a LTR-driven retroviral vector harboring SFFV LTRs and a post-transcriptional regulatory element derived from woodchuck hepatitis virus (woodchuck post-transcriptional regulatory element, wPRE).25, 26 MFG.γc is a γ-retroviral LTR-driven retroviral vector used in the clinical trial for SCID-X1 in Paris.27 It harbors an MoMLV-based LTR, contains a leader region with splice donor and acceptor plus an extended packaging signal (portion of gag included), and no wPRE. In contrast, SRS.SF.IL2RG.pre and SRS.SF.DsRED.pre are γ-retroviral SIN vectors harboring an internal SFFV promoter, a safety-modified wPRE and a deletion in the U3 region to remove promoter/enhancer elements.28 The lentiviral SIN vector expressing the IL2RG cDNA under control of the SFFV promoter, pRRL.PPT.SF.IL2RG.pre, is a derivative of pRRL.PPT.PGK.GFP.pre.29
Bone marrow transduction and transplantation
Lineage-negative (Lin−) BM cells of untreated C57Bl6/J mice were transduced on 2 following days (days 3 and 4 after cell isolation) as previously described8 (Supplementary methods). The cumulative multiplicity of infection (MOI, number of infectious particles per exposed cell) was 4 in experiment I (exp. I), and 10 in exp. II and III. Transduced BM cells were transplanted into lethally irradiated (9–10 Gy) syngeneic C57Bl6/J recipient mice. The cell dose was 5 × 105 except for exp. I, in which we transplanted 5 × 106 cells after additional expansion for 5 days plus 1 × 105 fresh BM cells. For secondary BM transplantation (BMT) at least two secondary, lethally irradiated (10 Gy) recipients received at least 2 × 106 cells of one primary recipient. In exp. III, BM cells of two or three primary recipients were pooled and transplanted into three or four secondaries (see footnotes in Table 1).
Genomic DNA for Southern blot analysis was isolated from cells of the spleen or BM using the Blood DNA separation kit (Qiagen, Hilden, Germany). Genomic DNA (10 μg) was digested with appropriate enzymes and Southern blot performed according to standard protocols.
Ligation-mediated PCR (LMPCR) was performed as described.30 To amplify integration sites of γ-retroviral SIN vectors, a 3′ LMPCR was established using a primer that annealed in the wPRE sequence (Supplementary methods). Recovered sequences were screened using the NCBI mouse genome database (NCBI37, accessed August 2007).
Quantitative PCR for the PRE and IL2RG sequences was performed on a 7300 Real Time PCR System (Applied Biosystems, Foster City, CA), using QuantiTect SYBR Green (Qiagen). Information on primers is in Supplementary methods. For quantitative reverse transcription (RT)–PCR, RNA was extracted from BM or spleen cells using RNAzol (Wak Chemicals, Bad Homburg, Germany). RT was performed with 1–2 μg RNA using PowerScript MLV reverse transcriptase (Becton Dickinson, Heidelberg, Germany) and real-time PCR for Evi1 using Assay-on-Demand (Applied Biosystems) or QuantiTect Primer Assays (Qiagen) for Prdm16. To detect specific Evi1 and Mds/Evi1 transcripts, primers spanning exon boundaries were designed (see Supplementary methods).
Results and discussion
Aim of the study and experimental design
To address the impact of vector design and transgene sequence on the induction of clonal imbalance and leukemia, we performed BMT in C57Bl6 mice with long-term follow-up (Figure 1a). We conducted three independent experiments transplanting mice with BM cells transduced with LTR and SIN vectors expressing a fluorescent marker gene (DsRED) or a potentially therapeutic gene (IL2RG) from the SFFV enhancer/promoter (Table 1).
In the exp. I, we tested whether vector-mediated expression of IL2RG was sufficient to induce leukemia. We used three different vectors: a lentiviral SIN vector harboring the internal SFFV promoter (RRL.PPT.SF.IL2RG.pre), a γ-retroviral SIN vector with the same internal promoter (SRS.SF.IL2RG.pre) and the LTR-driven vector MFG.γc. The latter represents the vector used in the Paris clinical trial for the correction of SCID-X1,27 including the meanwhile four patients that developed lymphoproliferation/leukemia caused by insertional mutagenesis.9 Using a similar vector, a fifth case of leukemia was recently reported in the London trial.11 Our control vector lacking IL2RG was SF91.DsRED.pre, a γ-retroviral LTR vector expressing the DsRED fluorescent protein that we previously showed to induce leukemia in mice by insertional mutagenesis.8 As our previous studies showed that prolonged culture increased leukemia incidence,8 we thus expanded the cells for another 5 days after transduction. In exp. II, we used the same set of IL2RG vectors as in the first, and as a control a γ-retroviral SIN vector expressing DsRED from the internal SFFV promoter (SRS.SF.DsRED.pre). We applied a higher MOI than in exp. I (MOI 10) to obtain more efficient gene marking and to determine the maximal tolerated vector dose. Here, the cells were transplanted without additional expansion following transduction, because results of exp. I showed a low number of engrafting clones. A third experiment was conducted to directly compare clonal imbalance triggered by LTR vs SIN γ-retroviral vectors expressing DsRED (SF91.DsRED.pre vs SRS.SF.DsRED.pre Supplementary Figure 1). As in exp. II, the cells were infused without additional expansion post-transduction. All mice were prospectively observed for potential induction of clonal imbalance or leukemia.
Transduction of human IL2RG into mouse BM cells was not leukemogenic in long-term primary recipients
Transgene expression determined 1 day after transduction was very efficient and higher in exp. II than in exp. I, in line with the MOI used (Figure 2a). In exp. II, 46% of the SRS.SF.IL2RG.pre transduced BM cells expressed IL2RG. This was the highest transduction efficiency achieved with the IL2RG containing vectors tested. The lentiviral construct gave 40% IL2RG+ cells. Expression of γc mediated by MFG.γc reached 31%.
Retroviral gene marking in the peripheral blood of recipient mice was analyzed by quantitative PCR 9 weeks after BMT (qPCR, Figure 2b and Table 1). In the SF91.DsRED.pre groups the mean copy numbers per cell ranged from 0.1 to 3.5 (mean 1.4). γ-Retroviral SIN vector-transduced BM (SRS.SF.IL2RG.pre or SRS.SF.DsRED.pre) reached a mean copy number of 0.54 and 0.36 (exp. II) and 0.71 (exp. III), respectively. Comparing lentiviral and γ-retroviral SIN vectors within the same experiment, the lentiviral marking was higher (exp. I and II). Despite a similar MOI, transduction efficiency with the MFG.γc vector was comparatively low, probably due to a higher proportion of competing virus-like particles in the supernatants.31 The mean marking in the peripheral blood of MFG.γc-treated mice thus reached 0.1 and 0.15 (Table 1).
With the exception of two recipient-type leukemias (due to irradiation, see supplements), all primary recipient mice stayed healthy (n=60), including the 32 mice transplanted with IL2RG-transduced cells. We conclude that γ-retroviral or lentiviral transduction with vectors expressing IL2RG itself had no oncogenic effects in wild-type recipient mice in an observation time of 11 months.
Leukemias after secondary transplantations of cells transduced with a γ-retroviral SIN vector encoding IL2RG, two of which harbor a single insertion
Serial transplantation stresses hematopoiesis and was shown to increase the risk of genotoxic side effects by retroviral vectors.6, 8 For secondary BMT, we used BM cells of selected mice with high gene marking from exp. I and II and of all mice from exp. III. Typically, two secondary recipients were transplanted with BM cells of one primary recipient. In exp. III, BM cells of up to three primary recipients were pooled and transplanted into three or four secondary recipients. Leukemias arose in seven secondary mice receiving cells transduced with the γ-retroviral SIN vector SRS.SF.IL2RG.pre, and in five mice receiving cells transduced with γ-retroviral LTR vectors (Table 2).
Four mice of the γ-retroviral SIN vector group developed a B220+ lymphoblastic leukemia (M32-1, M32-2, M35-1, M35-2; M32-2 shown in Figure 3a) that were based on the same leukemic clone, containing a single insertion site (Figure 3c). A single insertion only was also detected in an myeloproliferative disease-like myeloid leukemia (manifestation in two recipients, M31-1 and -2) and a third myeloid leukemic clone contained four insertions (M34-2, Figures 3a, c and Supplementary Figure 6). Strikingly, despite their amplification in different secondary mice, all clonally related leukemias manifested with similar latency and phenotype (Table 2), the lymphoblastic clone even in secondary recipients of different primary animals. The leukemic stem cell of this clone thus self-renewed during culture before primary BMT and progressed with a similar intrinsic program to conserve phenotype and latency of leukemogenesis while acquiring potential, yet unknown secondary mutations.32
Because it was difficult to directly detect the IL2RG transgene on the surface of freshly isolated leukemic cells, we cultured cells in medium free of interleukins that stimulate γc-associated receptors. After a few days, IL2RG was clearly detected by flow cytometry on the surface of leukemic cells; under these conditions, surface marker expression of B220 was largely lost while the expression of ckit and Sca-1 increased (Figure 3a, ‘cultured cells’).
The phenotype of three IL2RG-vector-associated leukemias was variable and not the classical cell background for IL2RG signaling.33 While the expression of Il2RG might have influenced the phenotype of the lymphatic leukemia, it can hardly be addressed as the initiating event. This conclusion is further supported by the large number of asymptomatic mice (32 primaries and 9 secondaries) that were repopulated with hematopoietic cells expressing IL2RG from γ-retroviral or lentiviral vectors. In agreement with our results, the overexpression of IL2RG by the CD2 promoter in a transgenic mouse line also did not lead to leukemogenesis.22
Also, five secondary recipients transplanted with SF91.DsRED.pre (LTR vector) transduced cells developed leukemia (Table 2). All cases had a myeloid phenotype (Supplementary Figure 3). Molecular analysis identified three distinct leukemic clones (Figure 3c). One contained only a single retroviral genome copy (M3-1 from exp. I), the other two had five copies (M(48–50)-1 and M(48–50)-3 from exp. III) or seven copies (M(48–50)-2 and M(48–50)-4 from exp. III). The latter two leukemic clones confirmed our previous finding of leukemogenesis triggered by multiple insertions of this vector.8 In contrast, the leukemia in association with a single insertion of this vector was unprecedented.
The difference in frequencies of leukemia induction between γ-retroviral and lentiviral vectors, or γ-retroviral LTR and SIN vectors, was not statistically significant (two-tailed Fisher's exact tests P-values: LTR DsRED versus SIN DsRED, P=0.48; γ-retroviral SIN DsRED versus lentiviral SIN IL2RG, P=0.142; γ-retroviral SIN IL2RG and lentiviral SIN IL2RG, P=0.142). However, the results clearly show that γ-retroviral SIN vectors with a strong viral enhancer/promoter are able to cause leukemia, as predicted by our cell culture model.17, 18
Clonal dominance by γ-retroviral SIN vector integration with a characteristic pattern in dominant clones
Because insertional mutagenesis by γ-retroviral SIN vectors bearing a strong internal enhancer/promoter initiated leukemias, we asked whether these vectors also induce clonal dominance.12 To test this we performed LMPCR on DNA from peripheral blood and BM in primary and secondary recipients transplanted with γ-retroviral SIN vector transduced cells expressing IL2RG or DsRED (samples taken at day of killing, Figure 4a). For γ-retroviral SIN vector insertions, the LMPCR initiates the first extension step from the wPRE element.17 We validated specificity, sensitivity and reproducibility of this protocol in defined cell clones (Supplementary Figure 2a, b). Further we confirmed that the LMPCR pattern obtained from DNA of blood and BM of a healthy recipient was almost identical and therefore the blood represents the clonal distribution (Supplementary Figure 2c).
We recovered an average of three dominant bands per mouse and a total of 63 unique insertions sites. In the SIN vector groups, the pattern of LMPCR products was similar to the LTR groups. Dominant clones became apparent by Southern blot if the mean copy number exceeded 0.5 per cell (Figure 4b). In exp. I, LMPCR showed selection for single clones already in the primary recipients (Supplementary Figure 4, M14 and M16), potentially triggered by the expansion of cells before BMT.
The location of the insertion sites with respect to the transcriptional start site (TSS) was similar for vectors encoding DsRED or IL2RG. This is concordant with our previous analysis of insertion sites of γ-retroviral vectors encoding various fluorescent proteins and cell-surface markers.13 We thus combined all results obtained with SIN vectors in the present study for statistical analyses of insertion pattern in dominant clones and leukemias (Figures 5a and b). We analyzed the number of genes common to this γ-retroviral SIN vector data set and the previously published retrovirus-tagged cancer gene database (RTCGD), which lists insertion sites of replication-competent γ-retroviruses recovered from murine tumor models.34 In total, 43% (27 of 63) of the genes that had their TSS nearest to the insertion sites were present in RTCGD, of which 17% (11 of 63) were listed as common insertion sites (CIS, genes that occur more than once in the RTCGD). By looking at all genes with their TSS within 100 kb of the vector integration, the number of integrations within 100 kb of an RTCGD member rose to 72% (41 of 57 sites, six sites were not in the 100 kb window), with a remarkable occurrence of nine sites that had more than one RTCGD member in this window. In total, 30% (17 of 57) genes had a RTCGD CIS member within 100 kb. Both the number of integrations close to RTCGD members and the number of RTCGD CIS is significantly higher than expected under a uniform random distribution, in which each gene has equal likelihood to be hit (Fisher's exact two-sided P<2.2 × 10−16 and P=7.6 × 10−15, respectively). In this respect, we found no difference (Fisher's exact two-sided test, P=0.4) between the present SIN data set and the formerly published Insertional Dominance Database (IDDb), which lists LTR vector insertion sites associated with dominant clones in similar BMT studies.13 The overrepresentation of RTCGD CIS genes close to γ-retroviral SIN vector insertions indicates clonal selection of ‘fitter’ clones (dominant clones), as previously shown in the context of LTR vectors.13
Most of the SIN vector insertions lie within introns/exons (28 out of 63 insertion sites listed in Supplementary Table 1, 44%). In concordance to the overrepresentation of insertions in introns/exons most of the SIN vector insertions occurred in a 5 kb window downstream of the TSS (Figure 5a). This pattern is distinct from that observed in unselected cells, which show preferential insertions at both sides of the TSS.35, 36 The SIN pattern identified in the present study was also significantly different from that found in the IDDb (Fisher's exact two-sided P<0.05) that shows a clustering of insertion sites in a window 5 kb upstream of the TSS (Figure 5b). A similar pattern as present in the IDDb was observed in clinical samples of a gene therapy trial.14 We also compared the orientation of LTR and SIN vector insertions with respect to their distance from the TSS. This analysis was based on the distance and orientation of all the genes within 150 kb of the 63 γ-retroviral SIN vector insertions retrieved in this study and all the vector insertions of the IDDb (Figure 5c and Supplementary Figure 7). In both the IDDb and the present SIN data set, vector sequences located closely upstream to the TSS preferably inserted in reverse orientation. With increased windows sizes, the forward and reverse oriented integrations occur in similar frequencies. In contrast, downstream of the TSS γ-retroviral SIN vectors inserts more frequently in forward orientation than do LTR vectors. As both our SIN data set reported here and the previously described IDDb reflect dominant clones, these differences in the orientation of the vectors in relation to the TSS of the neighboring genes might reflect selection effects and thus different mechanisms in which the two vector types can enhance gene transcription.
We also analyzed the insertion sites of lentiviral vectors in primary and secondary recipients. In the 12 primary mice, we identified only 17 independent sites (Supplementary Table 1 and Supplementary Figure 8). Six of these mice were transplanted with cells that were expanded prior to BMT, which is likely to reduce the number of engrafting clones. Because our analysis focuses on insertion sites of dominant clones we did not include minor LMPCR bands. In secondary mice, no additional insertion sites were found that were not already detected in the primary mice. This observation is similar to the pattern obtained with γ-retroviral LTR and SIN vectors. When looking at all genes within a window of 100 kb, 3 of the 17 lentiviral insertions showed RTCGD CIS genes. Due to the small sample size, the distribution around the TSS and a statistical comparison with the γ-retroviral SIN vector data set were not addressed.
Taken together, even though there was—because of the low number of cases—no measurable difference in the leukemogenicity of LTR and SIN vectors, the characteristic insertional behavior of γ-retroviral SIN vector containing the strong SFFV promoter/enhancer in dominant clones (downstream of TSS in forward orientation) suggests selection for a preferred mechanism of target gene activation.
Single hit leukemias are caused by insertional upregulation of Evi1 or Prdm16
We next asked whether the leukemias in mice developed due to the insertional activation of proto-oncogenes. All three leukemic clones that were induced by the γ-retroviral SIN vector encoding IL2RG and two of the three clones induced by the control LTR vector encoding DsRED contained insertions in Evi1, as detected in samples obtained from the major site of leukemia manifestation, that is, spleen or lymph nodes (Supplementary information and Supplementary Table 1). Four of the insertions were in the first intron of Evi1 and the fifth closely upstream of the ATG in exon 4 (Figure 6a). γ-Retroviral SIN and LTR vectors were found in similar positions. In all cases Evi1 transcripts were strongly upregulated (Figure 6b). It was now of interest to further characterize the Evi1 transcripts, as they may be expressed as a fusion mRNA of the Mds and Evi1 genes, then encoding the PR domain. Evi1 proteins lacking the PR domain are considered to be oncogenic while the PR domain containing Mds/Evi1 is associated with slower cell growth37 and the ratio of MDS/EVI to EVI1 message may predict the outcome of acute myeloid leukemia (AML) patients.38 In all leukemic clones, the Mds/Evi1 fusion transcript was expressed at much lower levels than in normal BM cells, whereas the Evi1 transcript initiating from the second exon was strongly upregulated (Figure 6c). We could not detect transcripts initiating from the first exon. γ-Retroviral vector insertions thus induced Evi1 transcripts that lacked the PR domain. Because four of the five insertions were located upstream of the second exon and a promoter in the first intron is postulated, they can well be described as upstream insertions. As the distance between the γ-retroviral insertion and the active TSS was up to 90 kbp, gene activation occurred over a long distance.
Ectopic expression of Evi1 by retroviral vectors in mice resulted in an myelodysplastic syndrome (MDS)-like disorder39, 40 or in B-ALL-like leukemia.41 In these studies, however, a potential confounding role of the vector insertion site was not addressed. Other studies speculate on necessary cooperation partners for leukemia induction by Evi1.40, 42, 43 The Evi1 locus is a frequent target in leukemias observed after infection of newborn mice with ecotropic replication-competent retrovirus (RCR).44 However, all leukemic clones induced by RCR contain further insertions. In the context of replication-defective vectors, insertion in this locus was previously found in association with development of monocytic leukemia6 or clonal dominance.12, 13, 45 Here, we confirm our previous observation of leukemia induction by a single insertion site that activated Evi16 and furthermore show that leukemic clones do not overexpress Mds/Evi1. We conclude that insertional activation of Evi1 may be sufficient to trigger leukemia with long latency and does not require coactivation of Mds exons encoding the PR domain.
The myeloid leukemia that contained only a single copy of the LTR vector encoding DsRED had the insertion site located 1376 bp upstream of the Evi1 homologue Prdm16. Prmd16 transcripts were strongly upregulated (500-fold). Similarly Prdm16 expression was upregulated in a dominant clone of a healthy recipient containing a vector insertion in the second intron (M16, 50-fold). PRDM16/MEL1 was identified as a fusion partner of RPN1 in AML and MDS with t(1;3)(p36;q21) translocations.46, 47 Similar to Evi1, Prdm16 generates two transcripts of different size, of which only the longer form contains the PR domain. Especially expression of the short form lacking the PR domain blocks granulocyte-colony stimulating factor (G-CSF)-induced differentiation and thus may be involved in leukemogenicity.48, 49 There are different chromosomal rearrangements generating fusion transcripts, for example, t(1;21)(p36;q22) inducing a RUNX1/PRDM16 fusion. Expression of PRDM16 alone without fusion is also linked to myeloid leukemias.50 Our study shows that insertional activation of Prdm16 may be sufficient to induce leukemia in mice. The role of the PR domain remains to be addressed.
Because only one single retroviral insertion was detected in three of the leukemic clones, we searched for further genomic alterations, particularly a loss of synthetic regions of human chromosome 7, as frequently observed in human leukemias with EVI1 activation.51 However, spectral karyotyping did not show major structural or numerical chromosomal alterations in the lymphoblastic or myeloid leukemic clones with upregulation of Evi1 or Prdm16, respectively (M32-2 and M3-1, Supplementary Figure 5).
The hypothesis that upregulation of Evi1 or Prdm16 is sufficient to induce leukemia in mice may be predictive for humans: Recent findings suggest that MDS-related alterations were associated with expansion of clones that overexpressed EVI1 after retroviral vector insertions in the clinical trial to treat young adults with CGD (M Grez, personal communication, July 2007). Expression of EVI1 is closely connected to the development of MDS and AML of very poor prognosis in humans. Typically EVI1 overexpression is induced by inv(3)(q21q26) or by other chromosomal rearrangements, resulting in the generation of fusion genes (for example, ETV&/EVI1 or RUNX1/EVI1) but can also be overexpressed without further genomic rearrangements. However, 133 of 194 cases of human aAML with inv(3) documented in the Mitelman database (http://cgap.nci.nih.gov/Chromosomes/Mitelman) carried secondary chromosome aberrations, most frequently monosomy 7. As we did not observe loss of synthetic regions by spectral karyotyping, the murine leukemias established here will have to be examined in great detail for potential epigenetic changes or cryptic genetic lesions.
At the molecular level, multiple mechanisms of target gene regulation and protein–protein interactions with other crucial hematopoietic transcription factors may explain the remarkable leukemogenic potency of upregulated Evi1,52, 53 and we would therefore hypothesize that Prdm16 also encodes a multifunctional protein. Another remarkable finding of the present study was the consistent evolution of leukemias with the same insertion site in different primary and secondary recipients. This suggests the induction of a relatively fixed malignant program by upregulation of a critical proto-oncogene (such as Evi1 or Prdm16) in long-lived cells with a high replicative potential.
Perspectives for gene therapy
In our study, we analyzed the potential of different integrating retroviral vectors to initiate leukemia in murine repopulating hematopoietic cells. An interim analysis of the study reported here was communicated in response to the study of Woods et al.21, who had proposed a leukemogenic potency of IL2RG overexpression in mice based on studies conducted with lentiviral vectors.22 At the time our letter was communicated, our argumentation was based on the survival statistics with about 1-year follow-up. In the present study, an even longer observation time and detailed phenotypic and molecular analyses are reported. We observed vector related leukemias in 12 secondary recipients (n=46) but none in primary recipients (n=60). The leukemias were based on six independent clones, three of which developed in LTR vector-transduced cells and the remaining three in γ-retroviral SIN vector-transduced cells. The relatively small group sizes and the unequal gene marking between the groups complicate statistical evaluation of the difference in genotoxicity depending on vector design. This study, however, shows for the first time leukemia induction following insertional mutagenesis by γ-retroviral SIN vectors that harbor a strong internal enhancer/promoter, and thus confirm the predictive power of our previously published cell-based assay.17, 18
By comparing the insertion pattern of SIN vectors recovered from dominant clones with previously established databases, we obtained evidence that a vector-specific selection occurs in vivo for insertion sites that are more likely to activate cellular growth-promoting genes. In our cell-based immortalization assay, we recently obtained evidence for a reduced genotoxic potential of γ-retroviral and lentiviral vectors with more physiological cellular promoters (and Modlich U, unpublished data). The present study thus supports the emerging concept17, 19, 54 that the type of the vector backbone and the nature of its cis-regulatory sequences are crucial variables to improve the therapeutic index of integrating vectors. The use of retroviral vectors containing strong, uninsulated enhancers should rather be restricted for the discovery of cancer-initiating genes in model systems (as shown here for Evi1 and Prdm16), or to clinical situations that offer no alternative.
Cavazzana-Calvo M, Fischer A . Efficacy of gene therapy for SCID is being confirmed. Lancet 2004; 364: 2155–2156.
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 gammaretroviral vector. Lancet 2004; 364: 2181–2187.
Gaspar HB, Bjorkegren E, Parsley K, Gilmour KC, King D, Sinclair J et al. Successful reconstitution of immunity in ADA-SCID by stem cell gene therapy following cessation of PEG-ADA and use of mild preconditioning. Mol Ther 2006; 14: 505–513.
Ott MG, Schmidt M, Schwarzwaelder K, Stein S, Siler U, Koehl U et al. Correction of X-linked chronic granulomatous disease by gene therapy, augmented by insertional activation of MDS1-EVI1, PRDM16 or SETBP1. Nat Med 2006; 12: 401–409.
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.
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.
Seggewiss R, Pittaluga S, Adler RL, Guenaga FJ, Ferguson C, Pilz IH et al. Acute myeloid leukemia is associated with retroviral gene transfer to hematopoietic progenitor cells in a rhesus macaque. Blood 2006; 107: 3865–3867.
Modlich U, Kustikova OS, Schmidt M, Rudolph C, Meyer J, Li Z et al. Leukemias following retroviral transfer of multidrug resistance 1 (MDR1) are driven by combinatorial insertional mutagenesis. Blood 2005; 105: 4235–4246.
Cavazzana-Calvo M, Fischer A . Gene therapy for severe combined immunodeficiency: are we there yet? J Clin Invest 2007; 117: 1456–1465.
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.
Board of the European Society of Gene and Cell Therapy. Case of leukaemia associated with X-linked severe combined immunodeficiency gene therapy trial in london. Hum Gene Ther 2008; 19: 3–4.
Kustikova O, Fehse B, Modlich U, Yang M, Dullmann J, Kamino K et al. Clonal dominance of hematopoietic stem cells triggered by retroviral gene marking. Science 2005; 308: 1171–1174.
Kustikova OS, Geiger H, Li Z, Brugman MH, Chambers SM, Shaw CA et al. Retroviral vector insertion sites associated with dominant hematopoietic clones mark ‘stemness’ pathways. Blood 2007; 109: 1897–1907.
Deichmann A, Hacein-Bey-Abina S, Schmidt M, Garrigue A, Brugman MH, Hu J et al. Vector integration is nonrandom and clustered and influences the fate of lymphopoiesis in SCID-X1 gene therapy. J Clin Invest 2007; 117: 2225–2232.
Schwarzwaelder K, Howe SJ, Schmidt M, Brugman MH, Deichmann A, Glimm H et al. Gammaretrovirus-mediated correction of SCID-X1 is associated with skewed vector integration site distribution in vivo. J Clin Invest 2007; 117: 2241–2249.
Aiuti A, Cassani B, Andolfi G, Mirolo M, Biasco L, Recchia A et al. Multilineage hematopoietic reconstitution without clonal selection in ADA-SCID patients treated with stem cell gene therapy. J Clin Invest 2007; 117: 2233–2240.
Modlich U, Bohne J, Schmidt M, von Kalle C, Knoss S, Schambach A et al. Cell-culture assays reveal the importance of retroviral vector design for insertional genotoxicity. Blood 2006; 108: 2545–2553.
Zychlinski D, Schambach A, Modlich U, Maetzig T, Meyer J, Grassman E et al. Physiological promoters reduce the genotoxic risk of integrating gene vectors. Mol Ther 2008; 16: 718–725.
Montini E, Cesana D, Schmidt M, Sanvito F, Ponzoni M, Bartholomae C et al. Hematopoietic stem cell gene transfer in a tumor-prone mouse model uncovers low genotoxicity of lentiviral vector integration. Nat Biotechnol 2006; 24: 687–696.
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.
Woods NB, Bottero V, Schmidt M, von Kalle C, Verma IM . Gene therapy: therapeutic gene causing lymphoma. Nature 2006; 440: 1123.
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 E6–7.
Pike-Overzet K, de Ridder D, Weerkamp F, Baert MR, Verstegen MM, Brugman MH et al. Ectopic retroviral expression of LMO2, but not IL2Rgamma, blocks human T-cell development from CD34+ cells: implications for leukemogenesis in gene therapy. Leukemia 2007; 21: 754–763.
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: E5; discussion E6–7.
Hildinger M, Abel KL, Ostertag W, Baum C . Design of 5′ untranslated sequences in retroviral vectors developed for medical use. J Virol 1999; 73: 4083–4089.
Schambach A, Wodrich H, Hildinger M, Bohne J, Krausslich HG, Baum C . Context dependence of different modules for posttranscriptional enhancement of gene expression from retroviral vectors. Mol Ther 2000; 2: 435–445.
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.
Schambach A, Mueller D, Galla M, Verstegen MM, Wagemaker G, Loew R et al. Overcoming promoter competition in packaging cells improves production of self-inactivating retroviral vectors. Gene Therapy 2006; 13: 1524–1533.
Dull T, Zufferey R, Kelly M, Mandel RJ, Nguyen M, Trono D et al. A third-generation lentivirus vector with a conditional packaging system. J Virol 1998; 72: 8463–8471.
Schmidt M, Hoffmann G, Wissler M, Lemke N, Mussig A, Glimm H et al. Detection and direct genomic sequencing of multiple rare unknown flanking DNA in highly complex samples. Hum Gene Ther 2001; 12: 743–749.
Relander T, Brun A, Hawley RG, Karlsson S, Richter J . Retroviral transduction of human CD34+ cells on fibronectin fragment CH-296 is inhibited by high concentrations of vector containing medium. J Gene Med 2001; 3: 207–218.
Huntly BJ, Gilliland DG . Cancer biology: summing up cancer stem cells. Nature 2005; 435: 1169–1170.
Pike-Overzet K, van der Burg M, Wagemaker G, van Dongen JJ, Staal FJ . New insights and unresolved issues regarding insertional mutagenesis in X-linked SCID gene therapy. Mol Ther 2007; 15: 1910–1916.
Akagi K, Suzuki T, Stephens RM, Jenkins NA, Copeland NG . RTCGD: retroviral tagged cancer gene database. Nucleic Acids Res 2004; 32 (Database issue): D523–D527.
Cattoglio C, Facchini G, Sartori D, Antonelli A, Miccio A, Cassani B et al. Hot spots of retroviral integration in human CD34+ hematopoietic cells. Blood 2007; 110: 1770–1778.
Wu X, Li Y, Crise B, Burgess SM . Transcription start regions in the human genome are favored targets for MLV integration. Science 2003; 300: 1749–1751.
Nucifora G, Laricchia-Robbio L, Senyuk V . EVI1 and hematopoietic disorders: history and perspectives. Gene 2006; 368: 1–11.
Barjesteh van Waalwijk van Doorn-Khosrovani S, Erpelinck C, van Putten WL, Valk PJ, van der Poel-van de Luytgaarde S, Hack R et al. High EVI1 expression predicts poor survival in acute myeloid leukemia: a study of 319 de novo AML patients. Blood 2003; 101: 837–845.
Buonamici S, Li D, Chi Y, Zhao R, Wang X, Brace L et al. EVI1 induces myelodysplastic syndrome in mice. J Clin Invest 2004; 114: 713–719.
Jin G, Yamazaki Y, Takuwa M, Takahara T, Kaneko K, Kuwata T et al. Trib1 and Evi1 cooperate with Hoxa and Meis1 in myeloid leukemogenesis. Blood 2007; 109: 3998–4005.
Cuenco GM, Ren R . Both AML1 and EVI1 oncogenic components are required for the cooperation of AML1/MDS1/EVI1 with BCR/ABL in the induction of acute myelogenous leukemia in mice. Oncogene 2004; 23: 569–579.
Du Y, Spence SE, Jenkins NA, Copeland NG . Cooperating cancer-gene identification through oncogenic-retrovirus-induced insertional mutagenesis. Blood 2005; 106: 2498–2505.
Louz D, van den Broek M, Verbakel S, Vankan Y, van Lom K, Joosten M et al. Erythroid defects and increased retrovirally-induced tumor formation in Evi1 transgenic mice. Leukemia 2000; 14: 1876–1884.
Mucenski ML, Taylor BA, Ihle JN, Hartley JW, Morse III HC, Jenkins NA et al. Identification of a common ecotropic viral integration site, Evi-1, in the DNA of AKXD murine myeloid tumors. Mol Cell Biol 1988; 8: 301–308.
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.
Xinh PT, Tri NK, Nagao H, Nakazato H, Taketazu F, Fujisawa S et al. Breakpoints at 1p36.3 in three MDS/AML(M4) patients with t(1;3)(p36;q21) occur in the first intron and in the 5′ region of MEL1. Genes Chromosomes Cancer 2003; 36: 313–316.
Mochizuki N, Shimizu S, Nagasawa T, Tanaka H, Taniwaki M, Yokota J et al. A novel gene, MEL1, mapped to 1p36.3 is highly homologous to the MDS1/EVI1 gene and is transcriptionally activated in t(1;3)(p36;q21)-positive leukemia cells. Blood 2000; 96: 3209–3214.
Nishikata I, Sasaki H, Iga M, Tateno Y, Imayoshi S, Asou N et al. A novel EVI1 gene family, MEL1, lacking a PR domain (MEL1S) is expressed mainly in t(1;3)(p36;q21)-positive AML and blocks G-CSF-induced myeloid differentiation. Blood 2003; 102: 3323–3332.
Xiao Z, Zhang M, Liu X, Zhang Y, Yang L, Hao Y . MEL1S, not MEL1, is overexpressed in myelodysplastic syndromes patients with t(1;3)(p36;q21). Leuk Res 2006; 30: 332–334.
Hazourli S, Chagnon P, Sauvageau M, Fetni R, Busque L, Hebert J . Overexpression of PRDM16 in the presence and absence of the RUNX1/PRDM16 fusion gene in myeloid leukemias. Genes Chromosomes Cancer 2006; 45: 1072–1076.
Russell M, List A, Greenberg P, Woodward S, Glinsmann B, Parganas E et al. Expression of EVI1 in myelodysplastic syndromes and other hematologic malignancies without 3q26 translocations. Blood 1994; 84: 1243–1248.
Wieser R . The oncogene and developmental regulator EVI1: expression, biochemical properties, and biological functions. Gene 2007; 396: 346–357.
Morishita K . Leukemogenesis of the EVI1/MEL1 gene family. Int J Hematol 2007; 85: 279–286.
Evans-Galea MV, Wielgosz MM, Hanawa H, Srivastava DK, Nienhuis AW . Suppression of clonal dominance in cultured human lymphoid cells by addition of the cHS4 insulator to a lentiviral vector. Mol Ther 2007; 15: 801–809.
We thank Marina Cavazzana-Calvo and Alain Fischer (Hopital Necker, Paris) for the MFG.γc plasmid. This study was supported by grants from the Deutsche Forschungsgemeinschaft (DFG SPP1230 to ZL and CB), the European Union (CONSERT to CB), the Bundesministerium für Bildung und Forschung (BMBF TreatID to CB) and the National Cancer Institute (R01-CA107492-01A2, to CB). AS was supported by the Else-Kröner Foundation.
About this article
Cite this article
Modlich, U., Schambach, A., Brugman, M. et al. Leukemia induction after a single retroviral vector insertion in Evi1 or Prdm16. Leukemia 22, 1519–1528 (2008). https://doi.org/10.1038/leu.2008.118
- single hit leukemia
- γ-retroviral self-inactivating vector
- insertional mutagenesis
Human Lentiviral Gene Therapy Restores the Cellular Phenotype of Autosomal Recessive Complete IFN-γR1 Deficiency
Molecular Therapy - Methods & Clinical Development (2020)
Current Protocols in Mouse Biology (2018)
PRDM16 isoforms differentially regulate normal and leukemic hematopoiesis and inflammatory gene signature
Journal of Clinical Investigation (2018)
Preclinical Efficacy and Safety Evaluation of Hematopoietic Stem Cell Gene Therapy in a Mouse Model of MNGIE
Molecular Therapy - Methods & Clinical Development (2018)
Human Gene Therapy (2017)