Reciprocal Relationship between O-Methylguanine-DNA Methyltransferase P140K Expression Level and Chemoprotection of Hematopoietic Stem Cells

Retroviral-mediated delivery of the P140K mutant O-methylguanine-DNA methyltransferase (MGMT) into hematopoietic stem cells (HSC) has been proposed as a means to protect against dose-limiting myelosuppressive toxicity ensuing from chemotherapy combining O-alkylating agents (e.g., temozolomide) with pseudosubstrate inhibitors (such as O-benzylguanine) of endogenous MGMT. Because detoxification of O-alkylguanine adducts by MGMT is stoichiometric, it has been suggested that higher levels of MGMT will afford better protection to gene-modified HSC. However, accomplishing this goal would potentially be in conflict with current efforts in the gene therapy field, which aim to incorporate weaker enhancer elements to avoid insertional mutagenesis. Using a panel of self-inactivating gamma-retroviral vectors that express a range of MGMT activity, we show that MGMT expression by weaker cellular promoter/ enhancers is sufficient for in vivo protection/selection following treatment with O-benzylguanine/temozolomide. Conversely, the highest level of MGMT activity did not promote efficient in vivo protection despite mediating detoxification of O -alkylguanine adducts. Moreover, very high expression of MGMT was associated with a competitive repopulation defect in HSC. Mechanistically, we show a defect in cellular proliferation associated with elevated expression of MGMT, but not wild-type MGMT. This proliferation defect correlated with increased localization of MGMT to the nucleus/chromatin. These data show that very high expression of MGMT has a deleterious effect on cellular proliferation, engraftment, and chemoprotection. These studies have direct translational relevance to ongoing clinical gene therapy studies using MGMT, whereas the novel mechanistic findings are relevant to the basic understanding of DNA repair by MGMT. [Cancer Res 2008;68(15):6171–80] Introduction Retroviral-mediated transfer of drug resistance genes into hematopoietic stem cells (HSC) has been investigated for its potential to achieve the dual aim of preventing severe dose-limiting toxicity of cancer chemotherapy and enrichment of gene-modified cells in vivo (reviewed in ref. 1). Of several drug resistance genes proposed for this application, one of the best characterized is O-methylguanine-DNA methyltransferase (MGMT ; ref. 2). MGMT confers protection against clinically relevant O-alkylating agents, such as 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU; Carmustine) and temozolomide (Temodar), which are used in the treatment of a variety of tumors, predominantly brain tumors (3, 4). Hematotoxicity is dose limiting in many treatment protocols using O-alkylating agents (5, 6), likely due to low expression of MGMT in the HSC and progenitor (P) compartment (7). In addition, elevated expression of MGMT has been found to be a mechanism of drug resistance in tumor cells, particularly glial tumors, which have poor prognosis (8, 9). To circumvent tumor resistance, pseudosubstrate inhibitors of endogenous MGMT, such as O-benzylguanine and O-(4-bromothenyl)guanine (lomeguatrib), have been developed. These agents ablate MGMT activity and sensitize tumors to the cytotoxic effects of O-alkylating agents (10, 11). However, in human clinical trials the combination of O-benzylguanine or lomeguatrib with temozolomide or BCNU has been shown to exacerbate myelosuppression (12–14). This increased toxicity likely relates to the depletion of MGMT activity in bone marrow HSC/P. Several mutant versions of MGMT are resistant to such pseudosubstrate inhibitors, yet retain alkyltransferase activity. Of these, the P140K mutant (MGMT) seems to be optimal, combining resistance to pseudosubstrate inhibitors with efficient repair of cytotoxic O-alkylguanine adducts (15–18). Thus, one goal of current studies is to achieve transgenic expression of MGMT in bone marrow HSC/P. Coupled with O-alkylating agent treatment and O-benzylguanine depletion, tumor cells will be sensitized while simultaneously protecting the bone marrow compartment to effect a widened therapeutic window. Clinical studies are currently under way using gamma-retrovirus vectors expressing MGMT in patients with poor-prognosis brain tumors (19). In addition, we and others have shown that treatment of mice and dogs that have been transplanted with HSC transduced with O-benzylguanine–resistant MGMT vectors allows significant in vivo selection of transduced cells (16, 20–23). Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). Requests for reprints: David A. Williams, Children’s Hospital Boston, 300 Longwood Avenue, Karp 07212, Boston, MA 02115. Phone: 617-919-2697; Fax: 617730-0934; E-mail: [email protected]. I2008 American Association for Cancer Research. doi:10.1158/0008-5472.CAN-08-0320 www.aacrjournals.org 6171 Cancer Res 2008; 68: (15). August 1, 2008 Research Article Research. on September 7, 2017. © 2008 American Association for Cancer cancerres.aacrjournals.org Downloaded from The design of safer retroviral vectors has been an area of intense interest in the field of gene therapy since the occurrence of leukemia due to vector insertional activation of the LIM-only protein 2 (LMO2) proto-oncogene in the gene therapy trial for X-linked severe combined immunodeficiency (24). Further observations from clinical trials, as well as numerous in vitro and in vivo assays, have shown that the powerful enhancer elements present in the long terminal repeats (LTR) of gamma-retroviral vectors can readily alter cell fate by insertional up-regulation of genes that confer a growth advantage (reviewed in ref. 25). One potential solution to this side effect of retroviral insertion is to use internal promoter elements in the context of deletions of the 3¶ LTR viral enhancer, so-called ‘‘self-inactivating’’ or SIN vectors. These modified vectors have recently been shown to have attenuated ability to activate a variety of growth-promoting genes in an in vitro immortalization assay (26). In addition, by coupling the use of a SIN vector backbone with an internal promoter, which has reduced enhancer activity compared with the gamma-retroviral LTR, it is possible to further reduce the risk of transformation via insertional transactivation (27). The application of this approach to the design of vectors with which to express MGMT creates a conundrum because the detoxification of O-alkylguanine adducts by MGMT is a stoichiometric process. Thus, whereas using a weaker enhancer/promoter Figure 1. Retrovirus vectors and in vivo selection/protection of transduced bone marrow cells. A, schematic representation of SIN-MGMT retroviral vectors. Vectors were derived by modification of Sin.SF, Sin.EFS, and Sin.PGK (29). Abbreviations: SF, enhancer/promoter from the SFFV LTR; EFS, truncated form of the promoter/enhancer from the human elongation factor 1a gene; PGK, promoter/enhancer from the human phosphoglycerate kinase gene; MGMT, the human O-methylguanine-DNA methyltransferase cDNA (P140K mutant unless otherwise stated); IRES, the internal ribosome entry site from encephalomyocarditis virus; eGFP, enhanced green fluorescent protein; Venus, modified version of yellow fluorescent protein; Pre, truncated version of the woodchuck hepatitis virus posttranscriptional regulatory element that lacks any X protein coding sequence; DU3, deletion in the enhancer/ promoter region of the retroviral LTR; C, packaging signal. B, change in the percentage chimerism of gene-marked cells following treatment with O-benzylguanine/temozolomide. Lethally irradiated recipient mice were injected as described in Materials and Methods. Mice from each cohort were then either treated or not with a O-benzylguanine/temozolomide regimen at 7 wk posttransplant. The percentage of GFP cells in the peripheral blood at 15 wk posttreatment versus 1 wk pretreatment was used to determine the change in frequency of gene-marked cells with chemoselection. 5, nontreated age-matched control mice; n, O-benzylguanine/temozolomide treatment. **, P < 0.01 (Wilcoxon’s rank sum test); ns, not significant. Data represent the sum of three independent experiments with 11 to 23 mice per experimental group. C, Kaplan-Meier plot of survival of O-benzylguanine/temozolomide–treated mice from each experimental group with time posttreatment. Vertical arrows, timing of O-benzylguanine/temozolomide (6BG/TMZ ) treatment. **, P < 0.01, versus other groups (log-rank test). Data represent the sum of three independent experiments with 13 to 23 mice per experimental group. Cancer Research Cancer Res 2008; 68: (15). August 1, 2008 6172 www.aacrjournals.org Research. on September 7, 2017. © 2008 American Association for Cancer cancerres.aacrjournals.org Downloaded from element to drive expression of MGMT may decrease the risk of insertional mutagenesis, the success of such vectors in chemoprotection and selection may be diminished. In addition, theoretically the risk of cell transformation may be increased due to increased frequency of unresolved O-alkyl adducts related to the lower levels of MGMT protein. In this study, we used a panel of SIN gamma-retroviral vectors that express a range of MGMT activities to establish the level of MGMT expression that is optimal for selection/protection of HSC/P. Importantly, the promoter/enhancer elements used in these vector constructs have differing potential to promote insertional mutagenesis (27). We show that the lower-level expression of MGMT driven by weaker cellular promoters is sufficient for selection of HSC following O-benzylguanine/ temozolomide treatment and detoxifies O-methyguanine lesions to the level of untreated controls. Surprisingly, we also found a previously undescribed deleterious effect of very high overexpression of MGMT on HSC reconstitution in vivo and cell proliferation in vitro. These studies may further elucidate the function(s) of MGMT protein and have important implications for the use of MGMT vectors in clinical trials. Materials and Methods Additional details of methods are provided in Supplementary data 1. Mice. All animals were maintained in a specific pathogen-free environment and all experiments were approved by the Institutional Animal Care and Use Committee of Cincinnati Children’s Research Foundation. Both C57BL/6J (referred to as C57B6 hereafter) and B6.SJLPtprc Pepc/BoyJ mice were obtained from The Jackson Laboratory. C57BL/6J pUR288(lacZ)-transgenic mice (referred to as lacZ) have previously been described (28). Generation of viral vectors. SIN gamma-retroviral vectors used in this study have previously been described (29). A cassette consisting of the encephalomyocarditis virus internal ribosome entry site, followed by either enhanced green fluorescent protein (eGFP) or Venus, was inserted directly 3¶ of the MGMT cDNA (Fig. 1A). Viral supernatants were generated as previously described (30). Vector titers were in the range of 8.0 10 to 8.4 10 IU/mL. Transduction of 32D cells and primary murine bone marrow cells. Transduction of 5-fluorouracil–pretreated primary murine bone marrow cells was done as described in ref. 30. The volume of viral supernatant used for transduction was adjusted relative to the established titer to achieve a transduction frequency that was similar across all transductions (typically in the range of 19–34%). Thirty hours after the final exposure to viral particles, transduced bone marrow cells were isolated by flow sorting (FACS Vantage, Becton Dickinson). The transduction of 32D cells was done essentially as for bone marrow cells with the exception that cells received a single round of transduction and were cultured in Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 10% FCS and 10 ng/mL murine interleukin-3 (mIL3; Peprotech). Transplant and analysis of chimerism. Bone marrow was injected into the tail vein of lethally irradiated recipient mice [11.75 Gy, 56 cGy/min, Cs source Mark I Model 68A Irradiator (J.L. Shepherd and Associates), split dose]. Peripheral blood cell counts were obtained using a Hemavet 850 FS Hematology Analyzer (Drew Scientific). Following red cell lysis (Pharm Lyse, Becton Dickinson), cells were either directly assessed for fluorescence (FACS Canto, Becton Dickinson) or stained with fluorescently labeled monoclonal antibodies [antimouse CD3q (145-2C11), B220 (RA3-6B2), TER-119, GR1 (RB6-8C5), MAC1 (M1/70), CD5 (53-7.3), and CD8a (53-6.7); all from E-Bioscience] following the manufacturer’s instructions, before flow analysis. Viable cells were identified by exclusion of 7-amino-actinomycin D. Drug treatment. O-Benzylguanine (Sigma) was dissolved at 7.5 mg/mL, using sonication for 1 h at 4jC in polyethylene glycol 400 (Electron Microscopy Sciences). The stock was further diluted to 2.5 mg/mL using icecold PBS and then injected i.p. into mice at 30 mg/kg. Temozolomide (Chemodex) was dissolved at 40 mg/mL in DMSO and was diluted to 8 mg/mL in PBS and placed on ice and immediately injected i.p. into mice at 80 mg/kg. Temozolomide was administered 2 h after O-benzylguanine, and this schedule was repeated on 2 consecutive days. Analysis of apoptosis and cell cycle. Cell cycle analysis was done using the Becton Dickinson APC bromodeoxyuridine (BrdUrd) Flow Labeling Kit according to the manufacturer’s instructions. 32D cells were starved for 16 h in IMDM, 1% FCS and were then incubated in media with 10% FCS, 10 ng/mL mIL3 at 37jC for the time periods indicated. 32D cells were then labeled with BrdUrd for 20 min at 37jC before fixation. Quantification of active MGMT. Quantification of MGMT activity was done as described by Watson and Margison (31). Immunoblots. The lysate from 5 10 cells was subjected to immunoblot as described by Gu and colleagues (32). The following primary antibodies were incubated with the membranes at the indicated dilutions: rabbit antip21 (C-19; Santa Cruz; 1:1,000); rabbit polyclonal anti-p27 (Cell Signaling Technology; 1:1,000); mouse anti–cyclin D1 (DCS6; Cell Signaling Technology; 1:200); mouse anti–cyclin E (E-4; Santa Cruz; 1:500); mouse anti–h-actin (AC-15; Sigma; 1:5,000); mouse anti-MGMT (MT3.1; Millipore; 1:1,000); and rabbit polyclonal anti–phospho-RB (Ser) (Cell Signalling Technology; 1:1,000). The membranes were then washed thrice and incubated with the appropriate secondary antibody [antirabbit IgG-horseradish peroxidase (HRP) conjugate or antimouse IgG-HRP conjugate, both from Cell Signaling Technologies, diluted 1:2,500]. HRP was visualized using either Lumiglo (Cell Signaling Technologies) or Supersignal West Femto (Pierce). The separation of soluble proteins from chromatinand nuclear matrix–bound proteins before immunoblot was done as described (33). Immunocytologic assay for DNA adducts and Comet assay. Lineagenegative (lin ) bone marrow cells were isolated from mice 2 h after treatment with O -benzylguanine and temozolomide as described above. Immunocytologic assay for DNA alkylation products by adduct-specific antibodies was done as described (34) using either mouse anti–(O-methylguanine) monoclonal antibody EM2-3 (35) or rabbit anti–imidazole-ring opened (N-methylguanine) serum (kindly provided by Dr. A. Povey, University of Manchester, Manchester, United Kingdom). The alkaline comet assay was done on sorted lin bone marrow cells as described (36). Quantification of fluorescence signals relating to DNA adducts and DNA strand breaks was also as described (36). Immunofluorescence intracellular staining for MGMT. 3T3 cells were grown on glass coverslips in 12-well non–tissue-culture-coated dishes (Becton Dickinson). Fixation and permeabilization were done using the Becton Dickinson Cytofix/Cytoperm Fixation/Permeabilization Solution Kit according to the manufacturer’s instructions. The samples were then incubated with mouse anti-MGMT (MT3.1, Millipore) diluted 1:100 in Permeabilization/Wash solution for 1 hour at room temperature. After washing, samples were incubated in Rhodamine-conjugated donkey anti-mouse IgG secondary antibody (Jackson Immunoresearch Laboratories) at 37jC for 45 min. After washing with Permeabilization/Wash solution, samples were mounted using Vectashield containing 4¶,6-diamidino-2-phenylindole (DAPI). Images were obtained on a Zeiss Axiovert 200M microscope (Carl Zeiss Microimaging, Inc.) and images were collected with a Hamamatsu camera using Openlab software (Improvision). Mutation frequency analysis. Mutation frequency was determined in lacZ bone marrow as described (37). Statistical analysis. All quoted errors are SEs of the mean unless otherwise stated. The Student t test or Wilcoxon rank-sum test was used to determine statistical significance dependent on whether all data conformed to a normal distribution as defined by the Shapiro-Wilk test.