Efficiency and Safety of O 6 -Methylguanine DNA Methyltransferase (MGMT P140K )-Mediated In Vivo Selection in a Humanized Mouse Model

Abstract

Efficient O⁶-methylguanine DNA methyltransferase (MGMT^P140K)-mediated myeloprotection and in vivo selection have been demonstrated in numerous animal models and most recently in a phase I clinical study in glioblastoma patients. However, this strategy may augment the genotoxic risk of integrating vectors because of chemotherapy-induced DNA damage and the proliferative stress exerted during the in vivo selection. Thus, to improve the safety of the procedure, we evaluated a self-inactivating lentiviral MGMT^P140K vector for transduction of human cord blood-derived CD34⁺ cells followed by transplantation of the cells into NOD/LtSz-scid/Il2rγ^−/− mice. These experiments demonstrated significant and stable enrichment of MGMT^P140K transgenic human cells in the murine peripheral blood and bone marrow. Clonal inventory analysis utilizing linear amplification-mediated polymerase chain reaction and high-throughput sequencing revealed a characteristic lentiviral integration profile. Among the bone marrow insertions retrieved, we observed considerable overlap to previous MGMT^P140K preclinical models or the clinical study. However, no significant differences between our chemotherapy-treated and nontreated cohorts were observed. This also hold true when specific cancer gene databases and a functional annotation of hit genes by the Panther Database with respect to molecular function, biological process, or cellular component were assessed. Thus, in summary, our data demonstrate efficient and long-term in vivo selection without overt hematological abnormalities using the lentiviral MGMT^P140K vector. Furthermore, the study introduces humanized mouse models as a novel tool for the pre-clinical assessment of human gene therapy related toxicity.

Introduction

Hematopoietic stem cell (HSC) gene therapy has been applied with considerable success to a number of congenital, monogenic disorders, including X-linked severe combined immunodeficiency disorder (SCID-X1), adenosine deaminase deficiency, chronic granulomatous disease (CGD), Wiskott–Aldrich syndrome, and various leukodystrophies (Rivat et al., 2012). However, this success has come at the expense of severe adverse events—most notably the occurrence of acute leukemia or myelodysplasias in several of these studies. Here in-depth analysis revealed retroviral insertions close to proto-oncogenes, leading to activation of these genes. As a consequence, major efforts have been undertaken to better define and predict the risks associated with HSC gene therapy. In this context, murine bone marrow (BM) transplantation models as well as surrogate in vitro assays were developed to assess genotoxicity and have significantly contributed to improve vector safety (Modlich et al., 2006; Montini et al., 2006, 2009). As a result, recent clinical trials were conducted with safety-improved, self-inactivating (SIN) gammaretroviral and lentiviral vectors (Zhang et al., 2013). Although this progress would not have been possible without meticulous studies in the various murine models, sole usage of mouse models for safety testing appears inappropriate given the inherent differences between murine and human biology (Hahn and Weinberg, 2002; Parekh and Crooks, 2012). Thus, relevant preclinical tools assessing genotoxicity directly in gene-modified human HSCs are needed. In this respect, “humanized” mouse models employing engraftment of human CD34⁺ cells in highly immunodeficient NOD/ShiLtSz-scid/Il2rγ^−/− (NSG) or NOD/ShiJic-scid/Il2rγ^−/− (NOG) mice are of great interest (Shultz et al., 2007; McDermott et al., 2010). In addition to the severe immunodeficiency that allows for the xeno-transplantation, it was recently demonstrated that the “scid” mutation also results in a stem cell defect in these mice even in the absence of selective pressure (Qing et al., 2012).

Extensive genotoxicity analysis appears particularly warranted for gene therapy strategies exerting additional genotoxic stress on transduced cells such as in vivo enrichment or exposure to mutagenic agents. This applies specifically to myeloprotective gene therapy approaches employing the (over)expression of selectable drug-resistant genes in hematopoietic stem and progenitor cells (Lachmann et al., 2013), as here the exposure to cytotoxic and mutagenic agents and the proliferative stress exerted during the in vivo selection process associated with this approach represent additional genotoxic risk factors. Besides mutants of the dehydrofolate reductase enzyme or the multidrug resistance 1 (MDR1) gene, mutants of the O⁶-methylguanine DNA methyltransferase (MGMT) gene have been investigated with considerable success in this context. MGMT encodes an evolutionarily conserved DNA repair protein that removes highly toxic O⁶-guanine adducts from the cellular DNA and confers resistance to chemotherapeutic drugs with a high O⁶-methylating or -chloroethylating potential such as temozolomide (TMZ) and other triazene derivatives or chloroethylnitrosoureas such as 1,3-bis(2-chlorethyl)-1-nitrosourea (BCNU), respectively (Milsom and Williams, 2007). Transgenic overexpression of the O⁶-benzylguanine (BG)-resistant MGMT^P140K point mutant has been demonstrated to allow for efficient enrichment of transduced hematopoietic cells by the combined application of the BG and BCNU or TMZ. These studies have been conducted in murine models (Reese et al., 1999; Ragg et al., 2000; Sawai et al., 2001; Jansen et al., 2002; Davis et al., 2003; Milsom et al., 2008; Giordano et al., 2011), including models of β-thalassemia (Persons et al., 2003; Zhao et al., 2009) and erythropoietic protoporphyria (Richard et al., 2004), humanized mice (Pollok et al., 2003; Zielske et al., 2003; Cai et al., 2006, 2008, 2010), as well as large animal models (Neff et al., 2003, 2005; Beard et al., 2009, 2010; Larochelle et al., 2009; Gori et al., 2012; Trobridge et al., 2012). MGMT^P140K (henceforth referred to as MGMT) also has been utilized to select for primary lung epithelium (Reese et al., 2008) and muscle cells (Lee et al., 2009) in vivo or to enrich for transgenic and protected lymphocytes in the context of novel gene therapy approaches for AIDS (Trobridge et al., 2009; Kiem et al., 2010). Moreover, a recent clinical phase I study has demonstrated efficient myeloprotection and in vivo enrichment of genetically modified hematopoietic cells following MGMT gene therapy in a cohort of glioblastoma patients with extended survival demonstrated in individual patients (Adair et al., 2012).

To screen for potential genotoxic side effects, the clonal repertoire of hematopoiesis after MGMT-mediated selection has been investigated for hematopoietic cells after exposure to alkylatying agents in vitro (Grund et al., 2010), in murine serial transplant (Giordano et al., 2011), canine and primate models (Neff et al., 2005; Beard et al., 2009, 2010; Larochelle et al., 2009), as well as in the recent clinical phase I study (Adair et al., 2012). Reassuringly, these studies so far have failed to demonstrate any adverse effects of MGMT-mediated chemoselection on the clonal repertoire of MGMT-transduced hematopoietic cells. Nevertheless, a number of questions remain, as some of these studies were compromised by relatively short observation periods (Pollok et al., 2003; Zielske et al., 2003; Cai et al., 2006) or insufficient depth of clonal analysis (Neff et al., 2003, 2005; Beard et al., 2009, 2010; Larochelle et al., 2009). Furthermore, the majority of these studies utilized gammaretroviral vectors (Pollok et al., 2003; Zielske et al., 2003; Cai et al., 2006, 2008, 2010; Adair et al., 2012), which inherently carry a relatively high risk for insertional mutagenesis (Modlich et al., 2009). Therefore, to improve the safety of MGMT-mediated myeloprotection and in vivo chemoselection strategies, we here analyzed MGMT expression from a third-generation SIN lentiviral vector employing the truncated elongation factor 1α (i.e., elongation factor-1α short version; EFS) promoter for transgene expression. This promoter was chosen as it was previously reported to direct robust and stable MGMT expression while avoiding the very high expression levels associated with viral promoters such as the spleen focus forming virus promoter (SFFV) (Milsom et al., 2008).

Materials and Methods

Generation of the SIN lentiviral MGMT vector construct

In our study a third-generation SIN lentiviral vector with an EFS as an internal promoter was used to coexpress MGMT^P140K mutant and enhanced green fluorescent (eGFP) protein coupled to internal ribosomal entry site. See Supplementary Materials and Methods (Supplementary Data are available online at www.liebertpub.com/hum).

Virus production

Viral supernatant was produced by transient transfection using calcium phosphate precipitation of 293T cells as previously described (Modlich et al., 2009). See Supplementary Materials and Methods for a brief description.

Evaluation of peripheral blood for human cell engraftment in NSG mice and end analysis

Starting at 6–8 weeks post-transplantation, peripheral blood (PB) was collected from NSG mice and analyzed at regular intervals for human engraftment by flow cytometry. Mice were euthanized 24–28 weeks after transplantation and analyzed for human engraftment and the levels of GFP⁺ cells in the PB, spleen, thymus, and bone marrow (BM). See Supplementary Materials and Methods.

Isolation and transduction of CD34⁺ cells

Human umbilical cord blood (CB) was collected from the donors after their informed written consent and after approval by the Hannover Medical School institutional ethics committee and the Norddeutsches Knochenmark- und Stammzellspender Register (Hannover, Germany). The CB was processed within 4 hr of collection, and human CD34⁺ cells were isolated according to the Miltenyi human CD34⁺ isolation kit protocol (Bergisch Gladbach, Germany). Isolated CD34⁺ cells were prestimulated for 48 hr in StemSpan medium (StemCell Technologies, Cologne, Germany) supplemented with 2 mM glutamine and 1% penicillin/streptomycin (PAA, Coelbe, Germany) and 100 ng/ml rhSCF, 100 ng/ml rhFLT3L, and 50 ng/ml rhTHPO (all from PeproTech GmBH, Hamburg, Germany). Subsequently, the cells were transduced (multiplicity of infection 20) with either EFS.MGMT.GFP or EFS.GFP viral supernatant on Retronectin-coated (10 μg/cm²; Takara, Otsu, Japan) dishes and using spinoculation at 2,000 rpm for 45 min at 4°C. Cells were cultured for a day further before transplantation into NSG mice.

Transplantation of NSG mice

A breeding colony of NSG was established and maintained at Hannover Medical School. Mice were kept in IVC racks (Allentown Inc., Moebris, Germany) in specific pathogen-free conditions. All animal experiments were approved by the local animal welfare committee and performed according to their guidelines. For the in vivo experiments 8–10-week-old NSG mice received a sublethal dose of 300 cGy total body irradiation by the PRIMUS linear accelerator (Siemens, Berlin, Germany) and were transplanted within 24 hr. The drinking water was supplemented with Ciprobay (Bayer, Leverkusen, Germany) for the first 3 weeks after irradiation to protect the mice from potential infections.

Administration of chemotherapy

Nine weeks after transplantation, the mice infused with the EFS.MGMT.GFP-transduced CD34⁺ cells were either treated with chemotherapy or left unexposed (non-treated group; NT). The treatment included weekly administration of 20 mg/kg BG (Sigma, Seelze bei Hannover, Germany) and 5–10 mg/kg of BCNU (Bristol Meyer Squibb GmBH & Co KGaA, Munich, Germany) 1 hr later for 2 consecutive weeks. BG and BCNU were prepared as described (Pollok et al., 2003). The mice transplanted with the EFS.GFP-transduced CD34⁺ cells (n=3), henceforth referred to as control, were not treated and monitored similar to the MGMT non-treated animals.

Linear amplification-mediated polymerase chain reaction

Vector insertion sites were amplified by linear amplification-mediated polymerase chain reaction (LAM-PCR) using 500 ng of DNA isolated from PB and BM samples using the QIAamp DNA mini Blood kit (Qiagen GmBH, Hilden, Germany). The LAM-PCR was performed as described previously (Schmidt et al., 2007) with the following modifications. After the double-strand synthesis step, each sample was equally divided into three parts for restriction digestion with different enzymes, namely, Tsp509I, HaeIII, and TaqαI (all from New England Biolabs, Ipswich, MA). Separate ligation reactions were carried out using enzyme-specific linkers (sequences in Supplementary Table S1) and were pooled for the denaturation step before the nested end-point PCRs. The PCR products were separated on a 2% agarose gel to visualize the bands and validate the PCR run. In total, we performed LAM-PCR on pretransplant (pre-TX; an aliquot of hCD34⁺ cells before transplantation), PB (collected before and after chemotherapy), and BM samples for mice from all groups.

High-throughput sequencing and integration-site analysis

For the 454-pyrosequencing run, LAM-PCR samples were digested with the SacI restriction enzyme (New England Biolabs) after the first exponential PCR at 37°C for 2 hr, followed by heat inactivation at 65°C for 20 min. This step removed most of the vector internal control bands. The digested DNA was diluted 1:30 with water, and 2 μl was used for the second exponential PCR along with the barcoded primers (Eurogentec, Seraing, Belgium). This final PCR product was purified via the QIAquick PCR Purification Kit (Qiagen GmBH), and the concentration was determined with a Nanodrop spectrophotometer (Thermo Scientific, Wilmington, Germany). Equal amounts of amplicons were used to create a library with a total of 500 ng DNA. The library was dispatched to GATC Biotech AG (Konstanz, Germany) for the 454 high-throughput sequencing with the Roche GS FLX+Titanium platform. The resulting sequences were sorted for their DNA barcode, using a custom Bioperl script. A sequence was considered valid only if it contained the part of the SIN lentiviral LTR along with the nested LTR primer and linker sequences with >95% identity. This step was performed using cross_match (Phil Green, University of Washington; www.phrap.org). Subsequent to these preprocessing steps, the sequences were uploaded to the freely available High Throughput Insertion Site Analysis Pipeline (HISAP) (Arens et al., 2012) for insertion-site analysis. The integration sites retrieved from HISAP analysis were ranked according to the read counts for each sample. The top integrations were selected for further validation using locus-specific PCR (ls-PCR). Insertions in various genes from our study were also compared with integration site databases retrieved from different MGMT studies (Beard et al., 2009, 2010; Grund et al., 2010; Giordano et al., 2011; Adair et al., 2012) and cancer databases [D'Antonio et al. (2012) and www.bushmanlab.org]. These post-HISAP processing steps were performed using in-house-developed scripts. The number of integrations for each BM sample along with the vector copy number (VCN) is given in Supplementary Table S2.

Locus-specific PCR and VCN determination

To validate individual vector integrations, reverse primers specific for the genomic region near the 3′ vector boundary were designed. The LTR primer (lvLTRIII) from the nested PCR step during the LAM-PCR served as the forward primer for validation of all integration sites. At first, conventional PCR using the Taq polymerase (PEQLAB Biotechnologie GmBH, Erlangen, Germany) was performed specifically on the sample from which the particular integration site was retrieved to confirm its presence and also check the functionality of the primers. Subsequently, real-time TaqMan PCR (Life Technologies GmBH, Frankfurt, Germany) was carried out on all the PB samples collected at different time points using a StepOnePlus (Applied Biosystems, Carlsbad, CA). Additionally, the VCN was determined using primers detecting the woodchuck posttranscriptional element (PRE) and polypyrimidine tract-binding protein 2 (PTBP2) as an internal reference. All quantitative PCRs (qPCRs) were performed in triplicates with standard deviations in cycles ranging from 0.02 to 1.23 (with signals obtained in a range of 24–36 cycles). A plasmid standard harboring sequences for the residual part of the LTR, PRE, and PTPB2 was used for the quantification. The list of all primers is provided in Supplementary Table S1.

Statistical tests

Statistical significance was determined by nonparametric t-test, Fisher's exact test, chi-square test, and two-way ANOVA test, and p-values≤0.05 were considered statistically significant.

Results

In our xenograft model, human CB-derived CD34⁺ were transduced with a third-generation SIN lentiviral construct expressing MGMT from an internal EFS promoter (EFS.MGMT.GFP; Fig. 1A) and transplanted into immunodeficient NSG mice. For control, NSG mice were transplanted with (EFS.GFP)-transduced cells. PB and BM samples from these animals were procured for integration-site analysis. In all our experiments, the progenies of 1–2×10⁵ CD34⁺ cells were transplanted per mouse with transduction efficiencies ranging between 10% and 35% (n=4 transplants). Subsequent to hematopoietic reconstitution, mice were subjected to combined BG/BCNU chemotherapy (Fig. 1B).

FIG. 1.

Schematic representation of the lentiviral vectors and the xenotransplant model. (A) Third-generation SIN lentiviral vectors expressing the mutant MGMT^P140K coupled to enhanced green fluorescent protein (eGFP; EFS.MGMT.GFP) or eGFP (EFS.GFP) alone. (B) NSG-based xenotransplant model. Chemotherapy (Chemo; two doses of 20 mg/kg BG and 5–10 mg/kg of BCNU injected 1 hr later, given 1 week apart), specified by black arrows. Peripheral blood (PB) samples were collected at the indicated time points to analyze for chimerism (solid white arrows) or for integration analysis (broken arrows). Δ, U3 region with a deletion; Δ, R, U5 comprise the long terminal repeat (LTR) of the vectors; ψ, packaging signal; BCNU, 1,3-bis(2-chlorethyl)-1-nitrosourea; BG, O⁶-benzylguanine; EFS, human elongation factor 1α short promoter (intron-deleted form); IRES, internal ribosomal entry site; NSG, NOD/ShiLtSz-scid/Il2rγ^−/−; PPT, polypurine tract; pre*, woodchuck posttranscriptional regulatory element; RRE, Rev responsive element; SA, splice acceptor; SD, splice donor; SIN, self-inactivating.

High-dose chemotherapy allows efficient enrichment of transgenic human cells but is toxic to NSG mice

In our initial experiment, four animals were treated with a BCNU dose of twice 10 mg/kg (a week apart). The pretransplant transduction efficacy was 18.8%. This regimen resulted in a profound granulocyto- and thrombocytopenia with nadir values of <0.1×10³/μl and 112×10³±28×10³/μl, respectively, measured 4 days after chemotherapy application (Fig. 2A and B). Within 8–21 days after treatment, three out of the four animals succumbed to severe pancytopenia. The only surviving animal was monitored for another 9 weeks (i.e., 17 weeks post-transplantation) with PB counts recovering to normal values 4–8 weeks after chemotherapy. Chemotherapy resulted in profound loss of human cells in the PB of all four NSG mice within 7 days of treatment (Fig. 2C). However, a significant enrichment of transduced human cells from 2.7%±0.8% before to 65%±8.8% one week after treatment (p=0.005; n=4) was observed in the human CD45⁺ PB population, indicating effective selection (Fig. 2D). Animals were analyzed for the engraftment of human cells in the BM immediately after death (two animals) and at the end of the experiment for the only long-term survivor. At these time points, the human cells in the BM mainly were composed of myeloid and B-cells (Fig. 2E). High levels of GFP⁺ human cells were detected in all hematopoietic compartments of the BM (Fig. 2F). Thus, the high-dose BG/BCNU treatment schedule clearly resulted in selection of MGMT-transduced cells, but at the expense of lethal myelotoxicity in the majority of animals. To circumvent this fatality and protect the murine hematopoietic system, we applied the same chemotherapy dose schedule but combined it with the transplantation of 1×10⁶ freshly isolated NSG whole BM cells 24 hr after the second dose of BCNU. While this approach substantially reduced the pancytopenia and led to survival of all animals until the end of the experiment, the percentage of human cells in the PB of these animals dropped to negligible levels (<0.01%) within 4–5 weeks after the murine cell transplant (Supplementary Fig. S1). However, when these animals were treated for a second time with a reduced chemotherapy regimen of 2×20 mg/kg BG/7.5 mg/kg BCNU given a week apart, transduced human cells were efficiently recovered and even enriched within the hCD45⁺ PB and BM populations.

FIG. 2.

High-dose chemotherapy allows effective chemoselection of transduced human cells but also lethal myelosuppression. Mice were administered two doses of 20 mg/kg 6-BG/10 mg/kg BCNU 8 weeks after transplant (large arrows). PB values for (A) granulocytes and (B) platelets are shown. Three out of four animals succumbed to pancytopenia (small arrows) within 5 weeks of chemotherapy administration. Gray areas indicate mean±SEM values of 16 nontransplanted wild-type NSG mice. PB was analyzed by flow cytometry for (C) human engraftment (%hCD45⁺ cells) and (D) the transduced population (GFP⁺) within the human cells before and after chemotherapy. BM was analyzed for the different hematopoietic sub-compartments either at the end of the experiment (open symbols) or when animals died (black symbols). (E) The contribution of B (hCD19⁺), T (hCD3⁺), immature myeloid (hCD33⁺), and mature myeloid (hCD11b⁺) cells to the human engraftment (hCD45⁺) and (F) the proportion of transduced (GFP⁺) cells in each lineage are depicted. Each animal is denoted by the same symbol throughout. BM, bone marrow; PB, peripheral blood.

Efficient in vivo selection of MGMT-transduced human cells with dose-reduced chemotherapy

On the basis of these results, we reduced the dose of BCNU to 7.5 mg/kg followed by 5.0 mg/kg (intermediate dose) or 5.0 mg/kg twice (low dose). Mice tolerated these chemotherapy regimens relatively well, although severe granulocytopenia (<0.1×10³/μl) was observed for both groups 4 days after chemotherapy application. Thrombocytopenia, however, was only moderate (652×10³±55×10³/μl for the intermediate and 580×10³±158×10³/μl for the low-dose regimen), and granulocyte as well as platelet values recovered to normal levels within 3 and 8 weeks of chemotherapy application, respectively (Fig. 3A and B). By the end of the experiment at 28 weeks, granulocyte counts dropped considerably in all animals (including non-treated controls; data not shown), which could be caused by long-term chemotherapy toxicity as well as the prolonged observation period. In both treatment groups, two out of five animals were lost 5–12 weeks post-transplant, although the cause of death appeared unrelated to the chemotherapy. Contribution of human cells in the PB is depicted in Fig. 3C, showing significant decrease in both the treated groups (p=0.008, medium dose, and p=0.0381, low dose, compared with the non-treated group, NT), 2 weeks after chemotherapy application. Engraftment levels in both the treated groups improved over time but remained well below the values of the NT cohort. For both the dose groups, chemotherapy resulted in considerable enrichment of the transduced human GFP⁺ population in the PB, peaking approximately 12 weeks after chemotherapy application (Fig. 3D). We detected an increase of GFP⁺ cells within hCD45⁺ PB cells from 2.3%±1.3% to 21.1%±7.2% (p=0.02; n=3) in the intermediate and from 0.89%±0.47% to 18.4%±10.4% (p=0.26; n=3) in the low-dose group. In the intermediate-dose group, these values were maintained until the end of the experiment (19.6%±2.3%), whereas selection was partially lost in the low-dose group (7.6%±4.7%; at week 26). In contrast, percentages of GFP⁺ cells remained low in non-chemotherapy-treated animals and even decreased throughout the experiment from 2.1%±0.42% to 0.5%±0.24% (p=0.02; n=6), a loss most likely reflecting the lower transduction rates that can be expected for primitive long-term SCID-repopulating cells (SRCs) in comparison to more short-term active stem cell clones.

FIG. 3.

Reduced doses of BG/BCNU avoid myelosuppression and efficiently select MGMT^P140K-positive human cells. Mice were treated with two cycles of chemotherapy (black arrows) 8 weeks posttransplant. One group received 20 mg/kg BG/7.5 mg/kg BCNU and 20 mg/kg BG/5.0 mg/kg BCNU (intermediate dose; triangle), and the other cohort received two times 20 mg/kg BG/5.0 mg/kg BCNU (low dose; circle). The nontreated animals (NT; open square) served as controls. PB values for (A) granulocyte and (B) platelet are shown. Fluorescence-activated cell sorting analysis of PB for (C) human engraftment (% hCD45⁺ cells) and (D) transduced cells (GFP⁺) within the human cells before and after chemotherapy is depicted. At the end of the experiment, the BM was evaluated for the contribution of total GFP⁺ cells within the hCD45⁺ engraftment denoted as (E) percentages and (F) total BM absolute numbers. BM was analyzed for GFP⁺ cells within human hematopoietic lineages: B (hCD19⁺), T (hCD3⁺), myeloid (hCD11b⁺, hCD33⁺), progenitors (hCD34⁺ hCD38⁺), and primitive (hCD34⁺ hCD38⁻) cells for the NT (white column), 7.5–5.0 mg/kg (intermediate dose, checkered column), and 5.0–5.0 mg/kg (low dose, black column) animal cohorts. (G) Percentages and (H) absolute numbers for these GFP⁺ populations are shown. Values are plotted as mean±SEM, n=3–6 animals. *p≤0.05 by nonparametric test.

Furthermore, effective in vivo enrichment of EFS.MGMT.GFP-transduced cells was confirmed in our model when BM cells were isolated from animals euthanized at the end of the experiment 18 weeks after BG/BCNU application (Fig. 3E). Here, the percentage of transduced GFP⁺ cells within the human cell population was significantly increased in the intermediate- and the low-dose treatment groups compared with NT animals (54.5%±13.7%, p=0.02, and 52.9%±15.6%, p=0.02 [n=3], respectively, vs. 1.8%±1.1% [n=6]). Substantially increased counts in comparison with NT controls (2.3±1.2×10⁵) were also detected regarding absolute numbers of transgenic human cells in the murine BM, with 7.7±2.7×10⁵ cells (p=0.09; n=3) for the intermediate-dose group and 7.7±0.8×10⁵ cells (p=0.03; n=3) for the low-dose group (Fig. 3F). Significant enrichment of transgenic B cells (hCD19⁺), mature myeloid (hCD11b⁺) and myeloid progenitor cells (hCD33⁺), as well as more differentiated (hCD34⁺ hCD38⁺) and primitive (hCD34⁺ hCD38⁻) stem/progenitor cells after BG/BCNU application also was observed when the data were broken down to distinct hematopoietic BM subcompartments (Fig. 3G). The expansion of transgenic cells was achieved primarily in the hCD33⁺ myeloid progenitor compartment (Fig. 3H; representative flow cytometry plots for a treated and NT animal are depicted in Supplementary Fig. S2). Although T (hCD3⁺) cell engraftment was negligible in the BM, T cells comprised 50–70% of human cells in the spleen and 70–75% in the thymus (see Supplementary Fig. S3A), and we observed higher percentages of transduced T cells in the chemotherapy-treated groups (21.8%±18.1% and 13.9%±5.6%) versus the non-treated (6.9%±5.4% and 0.24%±0.08%) cohort in the thymus and the spleen, respectively (data not shown). Thus, these results demonstrate that moderate doses of chemotherapy regimen can significantly reduce hematotoxicity while still allowing for the efficient and stable enrichment of transgenic GFP/MGMT-positive human cells in the major hematopoietic compartments for the entire observation period of up to 18 weeks after chemotherapy.

Characteristic lentiviral integration profile observed in all groups

As a next step we sought to assess the effect of chemoselection on the clonal repertoire of transduced human cells engrafted in these NSG employing LAM-PCR followed by high-throughput pyrosequencing for integration-site analysis. After removal of the linker- and vector-specific sequences, 486 unique sequences were mapped to the human genome using the HISAP (Arens et al., 2012) and assigned to the following samples: 231 in 2 pretransplant samples (pre-TX; from reduced chemotherapy and GFP control experiments), 144 in treated (10 BM and 20 PB; samples from all dose groups), 54 in NT (4 BM and 8 PB), and 57 in control (3 BM and 6 PB) samples. Irrespective of the vector or treatment, all groups exerted similar insertion-site profiles (Fig. 4). In accordance with previous studies employing lentiviral vectors (Mitchell et al., 2004; Laufs et al., 2006; Grund et al., 2010), in our experiments we found most of the integrations within gene-coding regions; 76.5% in pre-TX samples, 67.4% in treated samples, 60.5% in NT samples, and 74% in control BM samples. Furthermore, among these insertions, 65.4%±4% occurred in introns. In line with a typical lentiviral insertion pattern, no preference for insertions in the vicinity of transcription start sites (TSS) was observed. Also, overall ∼20% of hits were found±10 kb around CpG regions with no significant differences between the groups. These results indicate a representative insertion profile of SIN lentiviral vectors that was not significantly altered by chemotherapy application or MGMT expression. An important technical limitation regarding PB samples after xenotransplantation is the low absolute number of human cells in these samples. Combined with a low chimerism in some animals, these are suboptimal conditions for LAM-PCR and 454-sequencing. Consequently, we mapped either no or only very few (≤5) insertions in the PB samples of our animals. Thus, we mainly focused on the BM-derived insertions for further clonality assessment (Supplementary Table S2). We compared insertion sites retrieved from chemotherapy-treated, NT, and GFP control vector-transduced BM samples to the pre-TX samples using the retroviral insertion sites published in previous MGMT studies (Beard et al., 2009, 2010; Grund et al., 2010; Giordano et al., 2011; Adair et al., 2012). Moreover, we matched the genes closest to lentiviral integrations found in our model with the available cancer databases [D'Antonio et al. (2012) and www.bushmanlab.org/links/genelists] summarized in Supplementary Table S3. However, no significant differences between the three in vivo groups and the pre-TX samples regarding overlaps with genes found in previous MGMT datasets or cancer datasets were detected. This argues against a selection for specific insertions as a result of chemotherapy treatment or expression of MGMT. We also compared integrations from these groups against annotated human genes in the Panther Database (Mi et al., 2012) for any enrichment of specific gene classes related to the biological process, molecular function, or cellular component function, but no enrichment in any of the experimental groups was found (data not shown).

FIG. 4.

Retroviral integration-site analyses. Insertion sites from different samples of control, treated, non-treated, and pre-transplant (pre-TX) were analyzed for their location with respect to the CpG islands (±10 kb distance; black column), transcription start sites (TSS;±10 kb distance; checkered column), and hits within genes (white column). The chi-square test revealed no significant differences between the groups in any of the categories.

Oligoclonal repopulation in NSG recipients is not affected by chemotherapy application

The number of reads resulting from pyrosequencing for a particular insertion site is commonly used to estimate the abundance of a clone, thus providing an assessment of a sample's clonality (Biffi et al., 2011). The pre-TX samples consisted of various clones harboring integrations without any obvious overrepresentation, as 99.4%±0.5% of the detected insertions were represented by low (≤5) reads (Fig. 5). In contrast, in the treated, NT, and GFP control BM samples, 32.7%±6.6%, 43.0%±6.0%, and 25.1%±1.6% of insertions with low read counts were detected (p<0.01; two-way ANOVA). At the same time, high percentages of insertions with 101–1,000 (25.7%±6.3%, 12.9%±4.3%, and 35.4%±0.8%) and >1,000 (21.6%±9.4%, 13.2%±4.8%, and 11.7%±0.5%) reads were retrieved for treated, NT, and control BM samples, respectively. Statistical analysis using the chi-square test, however, revealed these differences between the three experimental cohorts as not significant. As the presence of insertion sites with high read counts (e.g., >1,000 reads) can be an indication of clonal imbalance, it is important to stress that we did not observe any overt clonal expansion in our animals (with the potential exception of animal RP10; see below). This and the fact that insertions with high read counts were detected in all three experimental cohorts (treated, non-treated, and GFP controls) seem to suggest that the chemotherapy and the selection process as such did not affect the clonality in the treated cohort. The overall reduced clonality in the BM samples of all groups suggests that only a few clones contributed to the in vivo repopulation repertoire over a prolonged period.

FIG. 5.

Clonal repertoire reduces in vivo after repopulation. Every integration site within a sample was ranked according to the pyrosequencing reads recovered. Percentages of insertions with ≤5 reads (black area), 6–100 reads (dark-gray area), 101–1,000 reads (light-gray area), and ≥1,000 reads (white area) are depicted for the pre-TX, GFP control, nontreated, and treated BM samples. The value above each column indicates the number of integration sites retrieved from the corresponding samples; n is the number of BM samples analyzed. Two-way ANOVA was used to compare pre-TX and BM samples (***p<0.01).

Verification of top clones reveals discrepancies in read counts and quantification by qPCR

Among the many insertions detected by the high-throughput sequencing for each animal (see Supplementary Table S2), the 10 most abundant integrations (top 10 clones) were designated based on their contribution to the total read counts of a specific sample. Similar to quantitative reverse transcription-PCR in validation of microarray experiments, ls-qPCR can be used to verify 454-read counts and to measure the contribution of a particular integration among all insertions in the PB and BM samples (Fig. 6A–E). As depicted, the proportion of a particular insertion varied at different time points, supporting the theory that different clones may contribute to the hematopoietic clonal repertoire at a given time in vivo. In our data, specific clones tended to increase over time, which in some cases correlated with the higher number of transduced human cells in PB samples. These data do not, however, allow for differentiation between single clones with several insertions or several clones with a single integration. Along this line, we could validate insertions close to the genes IFITM2, DNM2, JSRP1, FAM59A, and CCNB3, for the treated cohort and RAPGEF6, ARHGEF17, CAMSAP2, and TRAIP in the NT group (representative graphs in Fig. 6A–C). For one animal of the high-dose-treated cohort, the 454-results implied the presence of only one insertion close to TRGJP2 (Supplementary Table S2). As we observed >90% GFP expression in this animal, we cannot rule out that this represents a clonal amplification. However, we were unable to verify this insertion by ls-qPCR or to determine the VCN because of the very low amount of sample DNA available 7 days after chemotherapy when the animal died from myelosuppression. We here certainly cannot rule out clonal dominance, but given the discrepant results, the limited material available for analysis, the fact that the clone was not at all detected 3 weeks earlier (i.e., 1 week before chemotherapy), and also other explanations such as physiological clonal fluctuations in hematopoiesis have to be considered.

FIG. 6.

Validation of individual integration sites in the PB and BM. Locus-specific quantitative PCR (ls-qPCR) for the abundant integrations was performed on PB samples collected at different time points. Representative graphs indicate contribution of specific insertion to the total integrome (open symbols) along with the percentage of human cells (closed inverted triangle) and GFP⁺ human cells (asterisk) over time for individual insertion clones of (A, B) treated, (C) nontreated, and (D) control animals. Each graph represents one animal and with the gene symbol closest to the integration site detected in the BM of that animal. (E) Contribution of specific integrations to the total BM insertions retrieved by locus-specific qPCR (black column) and pyrosequencing (white column). (F) Engraftment levels of human cells (black column) and transgenic human cells in the BM (white column). RPXX codes indicate individual animals, and −X code denotes the rank of the particular insertion in this animal. Gene symbols are given over the respective bars.

Apart from animal RP10, we did not detect overt clonal imbalance in our mice. However, insertions around +65.2 kb of the IKAROS family zinc finger 2 (IKZF2) gene and +95.2 kb of the MDS1 and EVI1 gene complex (MECOM) were found in a control animal (RP49; Fig. 6D). IKZF2 is known to encode for a hematopoietic-specific transcription factor essential for lymphocyte development and is mutated in leukemias and lymphomas (Huret et al., 2000). MECOM, on the other hand, is a transcriptional regulator and is associated with esophageal and ovarian cancers, hematological malignancies, and so on (Metais and Dunbar, 2008). Furthermore, in the X-CGD clinical trial, insertions in this locus lead to clonal expansion and myelodysplasia in both the patients (Stein et al., 2010). As seen in Fig. 6D, the animal harboring these integrations showed an increased proportion of the IKZF2 insertion in the PB sample at 23 weeks, and at this point it also represented as the major insertion in the BM sample (Fig. 6E). In addition, in four of five methylcellulose colonies derived from BM cells of this animal, both insertions IKZF2 and MECOM were detected by ls-qPCR, suggesting that a single clone harbored both the insertions. This example clearly demonstrates that with the help of integration-site analysis and deep-sequencing technology, clinically relevant insertions can be detected in the humanized xenograft model. Furthermore, when comparing the clonal contribution in BM samples either quantified by ls-qPCR or 454-sequence read counts, we frequently observed substantial discrepancies in the representation by the two technologies (Fig. 6E). This incongruity was irrespective of the chimerism detected in the BM sample, arguing against insensitive detection as the basis of the observation (Fig. 6F). Furthermore, in few samples we observed discrepancies in the contribution of individual clones in BM and PB (see Supplementary Fig. S3B).

Discussion

We here report the use of a SIN lentiviral vector to express MGMT in human hematopoietic cells and successfully enriched these cells in vivo by the combined application of BG and BCNU. Recapitulating earlier studies by other groups in murine as well as xenotransplant models (Reese et al., 1999; Ragg et al., 2000; Sawai et al., 2001; Jansen et al., 2002; Pollok et al., 2003; Zielske et al., 2003; Cai et al., 2006, 2010; Milsom et al., 2008; Giordano et al., 2011), our study documents robust and stable chemoselection of transgenic human cells over extended periods in a model employing the xenotransplantation of MGMT-transduced CB-derived CD34⁺ cells. Insertion-site analysis by LAM-PCR coupled with high-throughput sequencing revealed a typical lentiviral integration pattern. An overall oligoclonal repertoire was detected in all experimental groups at the end of the 24–28-week observation period without enrichment of clones with integrations in or near-specific gene classes.

In the past, the majority of studies on MGMT-mediated in vivo selection, including the recent clinical trial, primarily used gammaretroviral vectors (Pollok et al., 2003; Zielske et al., 2003; Cai et al., 2006, 2008, 2010; Adair et al., 2012). As these vectors have a preference to integrate near the TSS of genes and carry a substantial risk of gene activation, we here evaluated a SIN lentiviral vector in order to reduce this risk. To direct MGMT expression, we specifically chose the EFS promoter, which in the context of the SIN gammaretroviral vector has been shown to allow for stable and functional MGMT expression while avoiding the toxicity associated with very high expression levels (Milsom et al., 2008). To assess the efficacy and safety of our lentiviral MGMT vector in a relevant in vivo model, we exploited the xenotransplantation of human CB-derived CD34⁺ cells into NSG mice (McDermott et al., 2010; Parekh and Crooks, 2012). Initial experiments with two cycles of 20 mg/kg BG/10 mg/kg BCNU resulted in effective selection of gene-marked human cells but also fatal myelotoxicity in the non-protected murine hematopoietic system. Infusion of 1×10⁶ fresh murine BM cells 24 hr after high-dose BCNU application prevented this toxicity, but also resulted in a loss of the vast majority of engrafted human cells. Similar observations have been described by other investigators and in part may be related to the excess of murine hematopoietic cells competing with their human counterparts for the murine BM niche (Cai et al., 2006). In contrast, reduced chemotherapy regimens employing 5–7.5 mg/kg BCNU allowed for effective selection of MGMT-transduced human cells without fatal myelotoxicity and resulted in significant enrichment of transgenic myeloid, B, and primitive CD34⁺ cells in the BM as well as selection of T cells in the spleen and thymus. These results are in accordance with previous findings stressing the importance of repetitive chemotherapy application to recruit dormant HSCs into an active cell cycle for effective MGMT-mediated chemoselection in murine transplant models (Jansen et al., 2002; Lin et al., 2011). In comparison to former xenotransplant studies (Cai et al., 2006), however, we now have considerably prolonged the observation period in the NSG recipients, enabling us to gauge the effectiveness of our selection strategy on the level of bona fide SRCs (McDermott et al., 2010). In this model, EFS-driven MGMT expression from our SIN lentiviral vector allowed for sustained in vivo selection over the entire observation period of up to 18 weeks after chemotherapy.

Insertion-site analysis by LAM-PCR demonstrated the well-known lentiviral integration pattern (Mitchell et al., 2004), characterized by a majority of hits (55–75%) in intronic sequences of genes and less hits in the vicinity of the TSS or CpG islands. This profile was detected in all experimental groups, ruling out a major influence of the experimental conditions associated with the in vivo selection process. However, when pre-TX and BM samples were compared, a substantial decrease in clonal diversity became obvious in all experimental groups. This is not surprising as only a minority of the originally transduced CD34⁺ cells have long-term engraftment capacity. Given that we transplanted progeny of 1–2×10⁵ hCD34⁺ cells and based on a frequency of 1–5 SRCs per 10⁴ CB CD34⁺ cells (Wang et al., 1997; McDermott et al., 2010) with a transduction efficiency of approximately 35–65%, we expect roughly 8–60 MGMT-positive SRCs in individual NSG mice. This correlates well with the results obtained from the BM samples of our BG/BCNU-treated animals (median 8 insertions, range 1–22).

In almost all animals, high read integrations were detected by end-point and quantitative PCR. When we analyzed the contribution of “top clones” to PB samples, we observed high variations with a strong trend for increased contributions to final samples (Fig. 6A–D). This observation clearly reflects substantial fluctuation in the contribution of individual insertion clones to the PB over time. The high-level contribution observed at the end of the experiment does not necessarily result from progressive clonal dominance but merely may reflect the fact that these clones initially were selected as “top clones” in the BM at even this time point. Furthermore, despite the successful validation of various insertions over time, we encountered substantial problems to quantify the contribution of a specific clone in a sample based on the pyrosequencing read counts. This was especially true for the PB samples and could not be overcome by reanalysis of the same sample. This technical limitation reflects the current concerns associated with clonality assessment based on sequential PCR-based techniques (Giordano et al., 2011; Brugman et al., 2013; Rittelmeyer et al., 2013).

We did not detect significant differences with regard to gene classes hit by insertions in the different cohorts, utilizing the Panther database as a reference (Mi et al., 2012). Importantly, this was also true for genes listed in cancer gene databases [D'Antonio et al. (2012) and www.bushmanlab.org/links/genelists]. Thus, in our humanized mouse model we could demonstrate successful MGMT-mediated in vivo selection without evidence for clonal imbalances generated by high-throughput sequencing. These results correlated well to the integration-site analysis reported for the in vitro BG/BCNU-selected human CD34⁺ cells transduced with an MGMT lentiviral vector employing the SFFV promoter (Grund et al., 2010). Similarly, insertional analysis in a number of large animal models employing gammaretro and lentiviral vectors revealed polyclonal profiles without enrichment of specific clones after chemotherapy. However, in these in vivo studies the integration sites were not monitored by pyrosequencing but by sub-cloning of LAM-PCR products (Neff et al., 2005; Beard et al., 2009, 2010; Larochelle et al., 2009). In a recent MGMT clinical trial, distinct prominent clones with integrations in PRDM16 and HMGA2 loci were described; however, no associated adverse events were observed (Adair et al., 2012). In contrast, enrichment of genes associated with cancer, DNA repair, or cell-fate decisions in response to chemotherapy exposure recently was reported from a high-dose chemotherapy murine serial transplant model employing MGMT-based chemoselection in up to tertiary recipients (Giordano et al., 2011). While some of these differences may arise from basic species disparity (Hahn and Weinberg, 2002), also profound differences among the chemotherapy regimen administered have to be taken into consideration.

Thus, in summary, our study convincingly demonstrates the potential of SIN lentiviral vectors in MGMT-based chemoprotection and in vivo selection strategies while at the same time avoiding HSC toxicity. Our xenotransplantation model combined with integration analysis allowed to detect and to monitor long-term the clonal fate of clinically relevant insertions. Nevertheless, the limited size of our experimental cohorts clearly calls for additional studies utilizing more animals and potentially for further evaluation of the construct specifically in the context of TMZ application, given the relevance of this agent in the treatment of glioblastomas.

Footnotes

Acknowledgments

We thank Axel Schambach for providing all vector backbones and Nico Lachmann for his help in generating MGMT vector constructs. We thank Doreen Lüttge for her excellent technical assistance and Rena Struβ for her help during mouse transplants. We appreciate Jannik Daudert for his help in developing scripts to evaluate integration analysis data. We are thankful to Sandra Zilz and the team of the medical school pharmacy for preparing cytotoxic drugs, and to Jörg Frühauf for his assistance during irradiation of mice (all from Hannover Medical School). This work was supported by grants from the Deutsche Forschungsgemeinschaft: Cluster of Excellence REBIRTH (Exc 62/1; T.M.) and MO 886/4–1 (T.M.), SPP1230 grant MO 886/3–1 (U.M. and T.M.) and European Union (Cell-PID, MR and UM). R.P. was partially supported by a stipend from Hannover Biomedical Research School.

Author Disclosure Statement

No competing financial interests exist.

References

Adair

J.E.

, Beard

B.C.

, Trobridge

G.D.

, et al. (2012). Extended survival of glioblastoma patients after chemoprotective HSC gene therapy. Sci. Transl. Med., 4, 133ra57.

Arens

, Appelt

J.U.

, Bartholomae

, et al. (2012). Bioinformatical clonality analysis of next generation sequencing derived viral vector integration sites. Hum. Gene Ther. Methods, 23, 111–118.

Beard

B.C.

, Sud

, Keyser

K.A.

, et al. (2009). Long-term polyclonal and multilineage engraftment of methylguanine methyltransferase P140K gene-modified dog hematopoietic cells in primary and secondary recipients. Blood, 113, 5094–5103.

Beard

B.C.

, Trobridge

G.D.

, Ironside

, et al. (2010). Efficient and stable MGMT-mediated selection of long-term repopulating stem cells in nonhuman primates. J. Clin. Invest., 120, 2345–2354.

Biffi

, Bartolomae

C.C.

, Cesana

, et al. (2011). Lentiviral vector common integration sites in preclinical models and a clinical trial reflect a benign integration bias and not oncogenic selection. Blood, 117, 5332–5339.

Brugman

M.H.

, Suerth

J.D.

, Rothe

, et al. (2013). Evaluating a ligation-mediated PCR and pyrosequencing method for the detection of clonal contribution in polyclonal retrovirally transduced samples. Hum. Gene Ther. Methods, 24, 68–79.

Cai

, Hartwell

J.R.

, Cooper

R.J.

, et al. (2006). In vivo effects of myeloablative alkylator therapy on survival and differentiation of MGMTP140K-transduced human G-CSF-mobilized peripheral blood cells. Mol. Ther., 13, 1016–1026.

Cai

, Ernstberger

, Wang

, et al. (2008). In vivo selection of hematopoietic stem cells transduced at a low multiplicity-of-infection with a foamy viral MGMT(P140K) vector. Exp. Hematol., 36, 283–292.

Cai

, Wang

, Bailey

, et al. (2010). Differential secondary reconstitution of in vivo-selected human SCID-repopulating cells in NOD/SCID versus NOD/SCID/gamma chain mice. Bone Marrow Res., 2011, 252953.

10.

D'Antonio

, Pendino

, Sinha

, and Ciccarelli

F.D.

(2012). Network of Cancer Genes (NCG 3.0): integration and analysis of genetic and network properties of cancer genes. Nucleic Acids Res., 40, D978–D983.

11.

Davis

B.M.

, Reese

J.S.

, Lingas

, and Gerson

S.L.

(2003). Drug selection of mutant methylguanine methyltransferase from different oncoretroviral backbones results in multilineage hematopoietic transgene expression in primary and secondary recipients. J. Hematother. Stem Cell Res., 12, 375–387.

12.

Giordano

F.A.

, Sorg

U.R.

, Appelt

J.U.

, et al. (2011). Clonal inventory screens uncover monoclonality following serial transplantation of MGMT P140K-transduced stem cells and dose-intense chemotherapy. Hum. Gene Ther., 22, 697–710.

13.

Gori

J.L.

, Beard

B.C.

, Ironside

, et al. (2012). In vivo selection of autologous MGMT gene-modified cells following reduced-intensity conditioning with BCNU and temozolomide in the dog model. Cancer Gene Ther., 19, 523–529.

14.

Grund

, Maier

, Giordano

F.A.

, et al. (2010). Analysis of self-inactivating lentiviral vector integration sites and flanking gene expression in human peripheral blood progenitor cells after alkylator chemotherapy. Hum. Gene Ther., 21, 943–956.

15.

Hahn

W.C.

, and Weinberg

R.A.

(2002). Rules for making human tumor cells. N. Engl. J. Med., 347, 1593–1603.

16.

Huret

J.L.

, Minor

S.L.

, Dorkeld

, et al. (2000). Atlas of genetics and cytogenetics in oncology and haematology, an interactive database. Nucleic Acids Res., 28, 349–351.

17.

Jansen

, Sorg

U.R.

, Ragg

, et al. (2002). Hematoprotection and enrichment of transduced cells in vivo after gene transfer of MGMT(P140K) into hematopoietic stem cells. Cancer Gene Ther., 9, 737–746.

18.

Kiem

H.P.

, Wu

, Sun

, et al. (2010). Foamy combinatorial anti-HIV vectors with MGMTP140K potently inhibit HIV-1 and SHIV replication and mediate selection in vivo . Gene Ther., 17, 37–49.

19.

Lachmann

, Brennig

, Phaltane

, et al. (2013). Myeloprotection by cytidine deaminase gene transfer in antileukemic therapy. Neoplasia, 15, 239–248.

20.

Larochelle

, Choi

, Shou

, et al. (2009). In vivo selection of hematopoietic progenitor cells and temozolomide dose intensification in rhesus macaques through lentiviral transduction with a drug resistance gene. J. Clin. Invest., 119, 1952–1963.

21.

Laufs

, Guenechea

, Gonzalez-Murillo

, et al. (2006). Lentiviral vector integration sites in human NOD/SCID repopulating cells. J. Gene Med., 8, 1197–1207.

22.

Lee

A.S.

, Kahatapitiya

, Kramer

, et al. (2009). Methylguanine DNA methyltransferase-mediated drug resistance-based selective enrichment and engraftment of transplanted stem cells in skeletal muscle. Stem Cells, 27, 1098–1108.

23.

Lin

, Cheung

, Roth

J.C.

, et al. (2011). Imaging stem cell-derived persistent foci after in vivo selection of lentiviral MGMT-P140K transduced murine bone marrow cells. Mol. Ther., 19, 1342–1352.

24.

McDermott

S.P.

, Eppert

, Lechman

E.R.

, et al. (2010). Comparison of human cord blood engraftment between immunocompromised mouse strains. Blood, 116, 193–200.

25.

Metais

J.Y.

, and Dunbar

C.E.

(2008). The MDS1-EVI1 gene complex as a retrovirus integration site: impact on behavior of hematopoietic cells and implications for gene therapy. Mol. Ther., 16, 439–449.

26.

, Muruganujan

, and Thomas

P.D.

(2012). PANTHER in 2013: modeling the evolution of gene function, and other gene attributes, in the context of phylogenetic trees. Nucleic Acids Res., 41, D377–D386.

27.

Milsom

M.D.

, and Williams

D.A.

(2007). Live and let die: in vivo selection of gene-modified hematopoietic stem cells via MGMT-mediated chemoprotection. DNA Repair (Amst.), 6, 1210–1221.

28.

Milsom

M.D.

, Jerabek-Willemsen

, Harris

C.E.

, et al. (2008). Reciprocal relationship between O6-methylguanine-DNA methyltransferase P140K expression level and chemoprotection of hematopoietic stem cells. Cancer Res., 68, 6171–6180.

29.

Mitchell

R.S.

, Beitzel

B.F.

, Schroder

A.R.

, et al. (2004). Retroviral DNA integration: ASLV, HIV, and MLV show distinct target site preferences. PLoS Biol., 2, E234.

30.

Modlich

, Bohne

, Schmidt

, et al. (2006). Cell-culture assays reveal the importance of retroviral vector design for insertional genotoxicity. Blood, 108, 2545–2553.

31.

Modlich

, Navarro

, Zychlinski

, et al. (2009). Insertional transformation of hematopoietic cells by self-inactivating lentiviral and gammaretroviral vectors. Mol. Ther., 17, 1919–1928.

32.

Montini

, Cesana

, Schmidt

, et al. (2006). Hematopoietic stem cell gene transfer in a tumor-prone mouse model uncovers low genotoxicity of lentiviral vector integration. Nat. Biotechnol., 24, 687–696.

33.

Montini

, Cesana

, Schmidt

, et al. (2009). The genotoxic potential of retroviral vectors is strongly modulated by vector design and integration site selection in a mouse model of HSC gene therapy. J. Clin. Invest., 119, 964–975.

34.

Neff

, Horn

, Peterson

, et al. (2003). Methylguanine methyltransferase-mediated in vivo selection and chemoprotection of allogeneic stem cells in a large-animal model. J. Clin. Invest., 112, 1581–1588.

35.

Neff

, Beard

B.C.

, Peterson

L.J.

, et al. (2005). Polyclonal chemoprotection against temozolomide in a large-animal model of drug resistance gene therapy. Blood, 105, 997–1002.

36.

Parekh

, and Crooks

G.M.

(2012). Critical differences in hematopoiesis and lymphoid development between humans and mice. J. Clin. Immunol., 33, 711–715.

37.

Persons

D.A.

, Allay

E.R.

, Sawai

, et al. (2003). Successful treatment of murine beta-thalassemia using in vivo selection of genetically modified, drug-resistant hematopoietic stem cells. Blood, 102, 506–513.

38.

Pollok

K.E.

, Hartwell

J.R.

, Braber

, et al. (2003). In vivo selection of human hematopoietic cells in a xenograft model using combined pharmacologic and genetic manipulations. Hum. Gene Ther., 14, 1703–1714.

39.

Qing

, Lin

, and Gerson

S.L.

(2012). An intrinsic BM hematopoietic niche occupancy defect of HSC in scid mice facilitates exogenous HSC engraftment. Blood, 119, 1768–1771.

40.

Ragg

, Xu-Welliver

, Bailey

, et al. (2000). Direct reversal of DNA damage by mutant methyltransferase protein protects mice against dose-intensified chemotherapy and leads to in vivo selection of hematopoietic stem cells. Cancer Res., 60, 5187–5195.

41.

Reese

J.S.

, Davis

B.M.

, Liu

, and Gerson

S.L.

(1999). Simultaneous protection of G156A methylguanine DNA methyltransferase gene-transduced hematopoietic progenitors and sensitization of tumor cells using O6-benzylguanine and temozolomide. Clin. Cancer Res., 5, 163–169.

42.

Reese

J.S.

, Roth

J.C.

, and Gerson

S.L.

(2008). Bone marrow-derived cells exhibiting lung epithelial cell characteristics are enriched in vivo using methylguanine DNA methyltransferase-mediated drug resistance. Stem Cells, 26, 675–681.

43.

Richard

, Robert

, Cario-Andre

, et al. (2004). Hematopoietic stem cell gene therapy of murine protoporphyria by methylguanine-DNA-methyltransferase-mediated in vivo drug selection. Gene Ther., 11, 1638–1647.

44.

Rittelmeyer

, Rothe

, Brugman

M.H.

, et al. (2013). Hepatic lentiviral gene transfer is associated with clonal selection, but not with tumor formation in serially transplanted rodents. Hepatology, 58, 397–408.

45.

Rivat

, Santilli

, Gaspar

H.B.

, and Thrasher

A.J.

(2012). Gene therapy for primary immunodeficiencies. Hum. Gene Ther., 23, 668–675.

46.

Sawai

, Zhou

, Vanin

E.F.

, et al. (2001). Protection and in vivo selection of hematopoietic stem cells using temozolomide, O6-benzylguanine, and an alkyltransferase-expressing retroviral vector. Mol. Ther., 3, 78–87.

47.

Schmidt

, Schwarzwaelder

, Bartholomae

, et al. (2007). High-resolution insertion-site analysis by linear amplification-mediated PCR (LAM-PCR). Nat. Methods, 4, 1051–1057.

48.

Shultz

L.D.

, Ishikawa

, and Greiner

D.L.

(2007). Humanized mice in translational biomedical research. Nat. Rev. Immunol., 7, 118–130.

49.

Stein

, Ott

M.G.

, Schultze-Strasser

, et al. (2010). Genomic instability and myelodysplasia with monosomy 7 consequent to EVI1 activation after gene therapy for chronic granulomatous disease. Nat. Med., 16, 198–204.

50.

Trobridge

G.D.

, Wu

R.A.

, Beard

B.C.

, et al. (2009). Protection of stem cell-derived lymphocytes in a primate AIDS gene therapy model after in vivo selection. PLoS One, 4, e7693.

51.

Trobridge

G.D.

, Beard

B.C.

, Wu

R.A.

, et al. (2012). Stem cell selection in vivo using foamy vectors cures canine pyruvate kinase deficiency. PLoS One, 7, e45173.

52.

Wang

J.C.

, Doedens

, and Dick

J.E.

(1997). Primitive human hematopoietic cells are enriched in cord blood compared with adult bone marrow or mobilized peripheral blood as measured by the quantitative in vivo SCID-repopulating cell assay. Blood, 89, 3919–3924.

53.

Zhang

, Thrasher

A.J.

, and Gaspar

H.B.

(2013). Current progress on gene therapy for primary immunodeficiencies. Gene Ther., 20, 963–969.

54.

Zhao

, Pestina

T.I.

, Nasimuzzaman

, et al. (2009). Amelioration of murine beta-thalassemia through drug selection of hematopoietic stem cells transduced with a lentiviral vector encoding both gamma-globin and the MGMT drug-resistance gene. Blood, 113, 5747–5756.

55.

Zielske

S.P.

, Reese

J.S.

, Lingas

K.T.

, et al. (2003). In vivo selection of MGMT(P140K) lentivirus-transduced human NOD/SCID repopulating cells without pretransplant irradiation conditioning. J. Clin. Invest., 112, 1561–1570.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB

0.14 MB

0.26 MB

0.10 MB

0.09 MB

0.05 MB

Efficiency and Safety of O 6 -Methylguanine DNA Methyltransferase (MGMT P140K )-Mediated In Vivo Selection in a Humanized Mouse Model

Abstract

Introduction

Materials and Methods

Generation of the SIN lentiviral MGMT vector construct

Virus production

Evaluation of peripheral blood for human cell engraftment in NSG mice and end analysis

Isolation and transduction of CD34+ cells

Transplantation of NSG mice

Administration of chemotherapy

Linear amplification-mediated polymerase chain reaction

High-throughput sequencing and integration-site analysis

Locus-specific PCR and VCN determination

Statistical tests

Results

High-dose chemotherapy allows efficient enrichment of transgenic human cells but is toxic to NSG mice

Efficient in vivo selection of MGMT-transduced human cells with dose-reduced chemotherapy

Characteristic lentiviral integration profile observed in all groups

Oligoclonal repopulation in NSG recipients is not affected by chemotherapy application

Verification of top clones reveals discrepancies in read counts and quantification by qPCR

Discussion

Footnotes

Acknowledgments

Author Disclosure Statement

References

Supplementary Material

Isolation and transduction of CD34⁺ cells