Modulation of DNA damage repair activity could lead to new approaches to reduce cytotoxic side effects of chemotherapy. N,N′,N″-Triethylenethiophosphoramide (thioTEPA) induces the formation of amino-ethyl adducts of guanine, resulting in imidazole ring opening [formamidopyrimidine (Fapy)] and is associated with significant myelosuppression in dose-intensive therapies. In Escherichia coli, Fapy lesions are repaired by the Fapy-DNA glycosylase (Fpg) protein. We hypothesized that the expression of the Fpg could increase resistance of hematopoietic cells to thioTEPA-induced cytotoxicity. Expression of Fpg in bone marrow (BM) cells via a retrovirus vector was associated with demonstrable 8-oxodeoxyguanosine DNA glycosylase activity. BM cells were infected with a recombinant retrovirus, SF91, containing the Fpg gene and expressing the enhanced green fluorescence protein (EGFP) via an internal ribosomal entry site element. Control mice received BM transduced with the backbone containing IRES-EGFP alone. Fpg-transduced and GFP+ BM hematopoietic cells were resistant in vitro to thioTEPA at multiple concentrations. Mice transplanted with transduced cells were treated with four doses of thioTEPA (10 mg/kg) given over 7 weeks. Despite low transduction efficiency, peripheral blood leukocytes, hemoglobin, and platelet counts of thioTEPA-treated Fpg mice were significantly higher than treated control mice (P < 0.05). In addition, after treatment, the BM, spleen, and thymic cellularity as well as the number of GFP+ progenitor cells in the BM of treated mice were significantly higher than those of control group. Selection of Fpg-transduced cells in vivo was demonstrated by an increase in the mean fluorescence intensity of peripheral mononuclear cells of Fpg mice compared with pretreatment value. In addition, a significant increase in the EGFP-bright cells was demonstrated, suggesting preferential survival of high-expressing hematopoietic cells. Similar results were demonstrated in vitro with primary BM expressing the human functional counterpart of Fpg, OGG1. These results show that expression of the Fpg or hOGG1 protein protects hematopoietic cells from thioTEPA-induced DNA damage and suggest that a high level of expression of these repair proteins is required to establish resistance to this drug. Expression of Fpg and/or OGG1 may provide an novel approach to preventing thioTEPA-induced toxicity of primary hematopoietic cells.
Although dose-intensive anticancer drugs have been used in the treatment of advanced cancers, patients with relapsed or refractory disease still experience poor outcomes because of disease progression. Obstacles to effective therapy include cumulative and dose-limiting side effects of anticancer drugs, especially myelosuppression, and drug resistance of tumors acquired during chemotherapy. Over the last decade, high-dose chemotherapy in combination with autologous bone marrow transplantation has been used to reduce the toxicity to bone marrow that has led to improved treatment of resistant malignant cells (1, 2, 3). Several reports suggest that the response rate of various tumors, such as ovarian cancer, testicular germ cell tumors, and malignant lymphoma, can be enhanced by increased dose intensity of anticancer drugs (4). In high-dose chemotherapy protocols, alkylating agents such as cyclophosphamide, ifosfamide, and thioTEPA3 have major dose-limiting toxicity generally restricted to bone marrow. Therefore, a reduction of the bone marrow-related toxic side effects of these drugs could allow for further dose or time intensification, which could result in an improvement in the outcome of patients with advanced disease.
It is increasingly appreciated that cellular DNA repair systems are important in the sensitivity of cells to alkylating agents (5, 6, 7, 8). Overexpression of DNA repair proteins, such as O6-methylguanine DNA methyltransferase confer increased chemoresistance to chloroethylnitrosoureas, one class of alkylating agents (9, 10, 11, 12, 13, 14, 15, 16, 17, 18). The utility of this approach, however, is limited somewhat by the fact that chloroethylnitrosoureas are not widely used agents in human cancer therapies.
However, multiple pathways have been implicated in DNA repair, specifically repair in response to damage by DNA alkylating agents. One such pathway, the BER pathway, is initiated by glycosylase proteins followed by sequential removal and repair of single mutated DNA adducts. Among alkylating agents, thioTEPA has recently gained widespread usage in the treatment of a number of malignant conditions, with myeloid toxicity emerging as a common dose-limiting side effect (1, 2, 3, 19, 20). In the present study, we focused on the protection of thioTEPA-induced cytotoxicity using genetic manipulation of the BER pathway.
The cytotoxic activity of thioTEPA has been postulated to be attributable either to the formation of cross-links within cellular DNA or the liberation of aziridine (ethyleneamine; Refs. 21 and 22). Aziridine induces the formation of N7-aminoethyl adducts of guanine and adenine, leading to imidazole ring opening or Fapy lesions (23). These DNA lesions account for ∼80% of all thioTEPA-induced adducts (24), preventing the movement of the replication fork and blocking DNA synthesis one base 5′ to the lesion (25). Fapy lesions are repaired by the Fpg protein, which is encoded by the fpg gene (or MutM gene) in Escherichia coli (26). A mammalian functional analogue of fpg gene has been cloned and designated as hOGG1 (27, 28, 29, 30, 31, 32).
Recently, it has been shown by our laboratory and others that overexpression of Fpg protein in mammalian cells reduces the toxicity of aziridine and that these cells are resistant to thioTEPA treatment (33, 34, 35). On the basis of these findings, we hypothesized that the expression of the Fpg or hOGG1 would protect primary hematopoietic cells from thioTEPA-induced toxicity. To test this hypothesis, we used retrovirus-mediated gene transfer to express the fpg and hOGG1 genes in murine primary hematopoietic cells and examined the effects of glycosylase expression on protection of hematopoietic cells in vitro and in vivo from the cytotoxic effects of thioTEPA.
MATERIALS AND METHODS
Retroviral Vector Constructions.
An improved MSCV-based bicistronic retroviral vector, MIEG3 (Fig. 1 A), expressing EGFP, was constructed as described previously (36) and contains the encephalomyocarditis virus IRES element in its original encephalomyocarditis virus viral configuration, resulting in more efficient translation of EGFP.
The pSF91-RE bicistronic retroviral vector (Fig. 1 A) was constructed by replacing the neoR cDNA in pSF91N (37) with the EcoRI-NotI IRES-EGFP fragment from pIRES-EGFP (Clontech, Palo Alto, CA). After transformation into DH5α competent cells (Life Technologies, Inc., Grand Island, NY), the colonies containing pSF91-RE were confirmed by minipreparation (Bio-Rad, Hercules, CA) and restriction endonuclease digestion. DNA sequencing was performed to confirm the integrity of the pSF91-RE sequence.
Cloning of Fpg cDNA was done using bacterial RNA as template as described previously (35). Briefly, Fpg oligo-primers (5′-primer, 5′-CCGGAATTCATGCCTGAATTACCCG-3′ and 3′-primer, 5′-GGCCGTCGACATTACTTCTGGCACTGCCGA-3′, containing added EcoRI and SalI sites, respectively) were used to perform a reverse-transcription-PCR. The PCR products were purified and digested with EcoRI and SalI. The digested fragment was gel purified and ligated into the EcoRI/SalI site in pGEX4T-1 (Life Technologies, Inc.). After transformation into HB101 competent cells (Life Technologies, Inc.), the colonies containing pGEX4T-1-Fpg were confirmed by PCR and restriction endonuclease digestion.
To construct pSF91-Fpg retrovirus vector, pSF91-RE retrovirus vector was double-digested with EcoRI and SalI (Fig. 1,A; New England Biolabs, Beverly, MA) and gel purified. The Fpg cDNA was amplified by PCR using Fpg oligo-primer (5′-CGAATTCACCATGCCTGAATTACCCGAAGTTG-3′, 5′-TCAGTCGACTTACTTCTGGCACTGCCGA-3′) to introduce a Kozak sequence (38, 39). The reaction included 10 mm KCl, 10 mm (NH4)2SO4, 20 mm Tris-Cl (pH 8.8), 2 mm MgSO4, 0.1% Triton X-100, 0.1 mg/ml BSA, 0.25 mm deoxynucleotide triphosphates, 10 pmol each of 5′ primer and 3′ primer, and 2.5 units of pfuTurbo DNA polymerase (Stratagene, La Jolla, CA) and was performed in a Thermocycler (MJ Research, Inc., Watertown, MA). The PCR products were purified and double-digested with EcoRI and SalI and ligated into the corresponding sites in pSF91-RE. After transformation into STBL2 competent cells (Life Technologies, Inc.), the colonies containing pSF91-Fpg were confirmed by PCR with a 5′-SF91 primer (AGTTAAGTAATAGTCCCTCTCTC) and a 3′-IRES primer (AGCGGCTTCGGCCAGTAACG; Fig. 1) and restriction endonuclease digestion. DNA sequencing was performed to confirm the integrity of the Fpg sequence.
hOGG1–6/pRSETB was a gift of Dr. Sankar Mitra (University of Texas Medical School, Galveston, TX). The hOGG1–6 cDNA was amplified by PCR using hOGG oligo primer (5′-ATCGAATTCCACCATGCCTGCCCGCGCGCTTCTGCCCA-3′, 5′-ATCGTCGACTTAGCCTTCCGGCCCTTTGGA-3′) to introduce a Kozak sequence. The PCR products were purified and double-digested with EcoRI and SalI and ligated into the corresponding sites in pSF91-RE. After transformation into DH5α competent cells (Life Technologies, Inc.), the colonies containing pSF91-hOGG1 were confirmed by PCR with 5′-SF91 primer and 3′-hOGG primer and restriction endonuclease digestion. DNA sequencing was performed to confirm the integrity of the hOGG1 sequence (Fig. 1). hOGG1–6 is identical to previously reported hOGG1 sequence and lacks six amino acids near the COOH terminus (amino acids 317–322; Refs. 35 and 40).
Generation of Retroviral Packaging Lines and Virus Supernatant.
Viral supernatant was produced from phoenix-AMPHO cells and phoenix-ECO cells (obtained from American Type Culture Collection, Manassas, VA) by transfection of 8 μg of purified plasmid DNA (Qiagen, Chatsworth, CA) in Lipofectamine transfection reagent (Life Technologies, Inc.), following the protocol provided by the manufacturer. Viral supernatants from phoenix-AMPHO were used to infect the GP+E86 packaging cell line (41). The GFP-expressing GP+E86 cells with high fluorescence intensity were established as stable producer populations.
The retroviral titer was determined by FACS analysis using a method reported previously (42). The titer of MIEG3, SF91-RE, or SF91-Fpg viral supernatants generated by individual GP+E86 clones were 1 × 105, 1 × 106, and 3 × 105 IU/ml, respectively, and the transient titer of SF91-hOGG1 generated by phoenix-ECO cells was 6 × 105 IU/ml. These titers were based on FACS analysis of GFP expression in infected cells and not on DNA analysis. The retroviral supernatant was used to infect to murine BM cells, and infected cells were analyzed for GFP expression using a FACSCalibur (Becton Dickinson, Mountain View, CA). A total of 3–5 ml of virus supernatant was used for each transduction.
Retroviral Infection of Hematopoietic Cells, Transplantation, and Treatment of Recipient Mice with thioTEPA.
All animal experiments were performed in accordance with institutional guidelines approved by the Animal Care Committee of the Indiana University School of Medicine. BM cells were harvested from the hind limbs of 8–10-week-old female C57Bl/6J mice (The Jackson Laboratory, Bar Harbor, ME) 48 h after i.p. injection of 5-fluorouracil (150 mg/kg body weight; SoloPak Laboratories, Franklin Park, IL). Mononuclear cells were isolated by buoyant density centrifugation (400 × g for 30 min at 25°C) using Histopaque-1083 (Sigma Chemical Co., St. Louis, MO). Harvested mononuclear hematopoietic cells were prestimulated for 48 h at 37°C in 5% CO2 in the presence of 100 ng/ml of recombinant rat stem cell factor, 100 ng/ml of granulocyte-colony stimulating factor and 100 ng/ml of megakaryocyte growth and development factor (all from Amgen, Thousand Oaks, CA) in IMDM (Life Technologies, Inc.) supplemented with 20% FBS (HyClone, Logan, UT), 100 units/ml of penicillin, and 100 μg/ml of streptomycin. The cells were transduced as described previously (18) on FN CH-296-coated dishes using virus supernatants containing MIEG3 (empty vector), SF91-RE (control vector), SF91-Fpg (Fpg vector), or SF91-hOGG1 for 4 h in the presence of the cytokine mixture described above. A total of 10 ml of virus supernatant was used for each transduction. The supernatant was then replaced with fresh medium containing cytokines. BM cells were analyzed for GFP expression using FACSCalibur (Becton Dickinson, Mountain View, CA) 48 h after virus infection.
For analysis of cytotoxicity, BM cells were infected twice using the above protocol. Infected cells were then exposed to thioTEPA in multiple doses and assayed for surviving progenitor cells using standard methylcellulose colony assays (see below).
For in vivo studies, 2 × 106 transduced cells were infused into irradiated C57Bl/6J mice (11 Gy total body irradiation, 139Cesium split dose, with minimum of 3 h between doses; Nordion International, Kanata, Canada). Four weeks later, 10 mg/kg of thioTEPA were injected i.p. The same dose was repeated in 1 week. Three weeks after the second dose, the treatments were repeated. Some Fpg mice were given 0.9% NaCl as a nontreatment control group (NT group). Mice were sacrificed 3 weeks after the last thioTEPA treatment.
Total peripheral blood leukocyte (WBC) counts and Plt counts were analyzed on peripheral blood obtained from lateral tail veins with a Semi-Automated Microcell Coulter Model F820 (Sysmex, Kobe, Japan) using a size distribution curve of 30–300 fl for leukocyte determinations and 2–30 fl for Plt determinations. Hb concentrations were measured using lysing reagent, QuickLyser II (Sysmex) according to the manufacturer’s recommendations, based on a variation of the internationally standardized cyanmethoglobin method.
Flow Cytometric Analysis.
Flow cytometric analysis of peripheral blood mononuclear cells was performed every 3 weeks by using GFP expression. Lymphocytes harvested from the thymus and spleen of surviving control and SF91-Fpg-transduced recipient mice were analyzed phenotypically by three-color flow cytometry. Cells were also harvested and analyzed from the NT group mice as a second control. Single cell suspensions of thymus and spleen were prepared and filtered through a 40 μm nylon mesh (Becton Dickinson, Franklin Lakes, NJ). The cells were blocked in PBS (Life Technologies, Inc.) with 10% normal rat serum (PharMingen, San Diego, CA) for 10 min at 4°C. One million cells were incubated with 1 μg of CD4-PE, 0.5 μg of CD8-PerCP, 1 μg of CD3-PerCP, and 0.15 μg of B220-PE (PharMingen) for 30 min at 4°C and washed twice with PBS containing 0.1% BSA. The cells were analyzed by flow cytometric analysis using FACSCalibur (Becton Dickinson, Mountain View, CA).
Analysis of Clonogenic Cells.
BM cells were cultured in triplicate in IMDM, containing 30% FBS, 1% BSA, 100 units/ml of penicillin, 100 μg/ml of streptomycin, 2 mml-glutamine (Life Technologies, Inc.), 1 × 10−4m β-mercaptoethanol (Sigma Chemical Co.), 4 units/ml human erythropoietin (StemCell Technologies, Inc., Vancouver, British Columbia, Canada), 100 ng/ml recombinant rat stem cell factor, 100 units/ml of murine interleukin-3 (10 ng/ml; PeproTech, Rocky Hill, NJ), and 0.8% methylcellulose (Methocult M3134; StemCell Technologies, Inc.). Progenitor colonies were counted after 7 days of culture in a humidified environment at 37°C and 5% CO2. The number of GFP-expressing progenitor cells was determined by fluorescence microscopy.
PCR and Blot Analysis.
DNA was prepared from BM, spleen, and thymus using standard methods for amplification with PCR. 5′-SF91 primer and 3′-IRES primers were used (Fig. 1 A). The PCR products were then separated by electrophoresis in a 1% agarose gel, transferred to a nylon membrane (Osmonics, Westborough, MA), and hybridized to a 32P-labeled EcoRI-SalI double-gel-purified Fpg cDNA fragment excised from pGEX-4T-1-Fpg. Prehybridization, hybridization, and posthybridization washes were performed as described previously (17). Nylon filters were exposed to X-ray film at −70°C in the presence of calcium tungstate intensifying screens.
8-oxoG DNA Glycosylase Activity Assay.
GFP-expressing BM cells were sorted by FACScan (Becton Dickinson, Mountain View, CA). Cells were gated based on forward and side light scatter to exclude debris. Postsort purity of GFP+ BM cells transduced with SF91-RE, SF91-Fpg, or SF91-hOGG-1 was 90.2, 96.6, or 94.5%, respectively. Whole cell extracts were prepared by sonicating the cells in PBS with 2 mm DTT at 45 W three times for 15 s on ice. Supernatants were collected after spinning the sonicated cells at 12,000 rpm for 10 min at 4°C. Protein concentration was determined by using Bio-Rad protein assay kit (Bio-Rad). 8-oxoG glycosylase assays were performed to determine Fpg or hOGG1 activity levels as described below. A 26-bp 8-oxoG-containing oligonucleotide (5′-AATTCACCGGTACC-8-oxo-GTCTAGAATTCG-3′) was 32P end-labeled as described previously (43, 44). Reaction mixtures (10 μl) containing 10 μg of protein cell lysate, 2.5 pmol of 32P end-labeled 8-oxoG 26-mer, 70 mm HEPES-KOH (pH 7.6), 100 mm KCl, and 10 mm EDTA were allowed to proceed for 60 min in a 37°C water bath. Reactions were halted by adding 10 μl of 96% formamide, 10 mm EDTA, xylene cyanol, and bromphenol blue. 8-oxoG assay products (10 μl) were separated on a 20% polyacrylamide gel containing 7 m urea. Gels were wrapped in saran wrap and exposed to film for visualization. Activity was determined as the ratio of the 14-mer (cut) versus 26-mer (uncut) bands using SigmaScan densitometry software (SPSS, Chicago, IL).
Both femurs of each primary recipient mouse were removed, and the marrow contents were flushed out using a syringe and a 26-gauge needle. The cells were resuspended in IMDM containing 10% FBS. Harvested cells of each mouse were used for DNA and protein preparation and progenitor assay as described above. Remaining cells were injected into secondary irradiated recipients at the cell concentration of 8 × 106/mouse.
Results are expressed as the mean ± SD. The probability of significant differences between groups in each figure was determined by Student’s t test.
Comparison of Expression in Primary Murine Bone Marrow Cells Transduced with MIEG3 versus SF91-RE Retroviruses.
High transgene expression in target cells may be critical to achieve optimal therapeutic intent with drug resistance genes. Expression of retrovirus-encoded transgenes depends in part on the enhancer/promoter elements contained in the U3 region of the viral LTR and on cis elements located in the UTRs on the RNA and varies significantly between backbones currently in use in preclinical and clinical gene therapy trials. Several reports have shown that the transcriptional activity of the U3 region contained in the MESV is relatively weak compared with that of the Friend mink cell focus-forming viruses, such as SFFV (45, 46, 47). Therefore, we initially compared the expression of transgenes in primary hematopoietic cells encoded after transduction with MIEG3, which contains a MESV LTR, and SF91-RE, which contains a SFFV LTR and an improved 5′ UTR (Fig. 1). We inserted EGFP with an IRES element into these backbones and analyzed GFP expression. BM cells were infected on FN-CH296 at the same MOI (<1) to ensure single copy integration. Mean fluorescence of GFP expression in primary murine BM cells transduced with SF91-RE was significantly higher than in cells transduced with MIEG3 (371.3 ± 53.2 versus 57.8 ± 24.6, respectively; P < 0.01; Fig. 1 B). Therefore, we used the pSF91-RE backbone for studies described below.
DNA Repair Activity of Cells Transduced with Fpg Vector.
To analyze the enzymatic activity of Fpg protein expressed in primary BM cells, these cells were infected at high (5–10) MOI by each virus supernatant containing either SF91-RE (GFP titer, 1 × 106 IU/ml) or SF91-Fpg (GFP titer, 3 × 105 IU/ml). Although these titers are based on GFP expression, we have demonstrated previously a close concordance of GFP expression and transduction based on DNA analysis using real-time PCR in transduced hematopoietic cells with this same vector backbone.4 The transduction efficiencies of SF91-RE or SF91-Fpg infected BM cells were 38.1 and 19.9%, respectively. GFP+ cells were sorted using FACS, and then the activity of the Fpg was determined using an oligonucleotide cleavage assay. A 26-bp oligonucleotide containing an 8-oxoG adduct at nucleotide 15 was incubated with cell lysate, as described in “Materials and Methods.” Equal amounts of protein were loaded in each assay comparing control cells (containing the empty vector) and SF91-Fpg-transduced BM cells. As seen in Fig. 2,A, cell extracts from SF91-Fpg-transduced BM cells contained significantly higher 8-oxoG repair activity compared with SF91-RE (control) cells (Fig. 2,A). Cell extracts from control cells did not contain any demonstrable repair activity, presumably because of the low level of expression of endogenous OGG1, a functional analogue of Fpg. On the basis of these results, we concluded that functional Fpg protein could be expressed in primary hematopoietic cells. Moreover, the level of GFP expression correlated with Fpg activity, when high- and low-GFP-expressing NIH/3T3 cells were isolated by GFP+ cell sorting and then subjected to the oligonucleotide cleavage assay (Fig. 2 B).
Expression of Fpg in BM of Mice Is Associated with Significantly Fewer Blood Cell Cytopenias after thioTEPA Treatment.
To determine the function of Fpg-expressing BM cells in vivo, we used a protocol that facilitates transduction of long-term repopulating BM cells with the SF91-Fpg retrovirus. Using this protocol, infection of BM cells for transplantation was accomplished at an MOI of 1–2. Transduction efficiencies of BM cells infected with SF91-RE (control) and SF91-Fpg virus prior to transplantation were 16.3 ± 0.8% and 3.9 ± 0.1%, respectively (Table 1). SF91-Fpg-transduced bone marrow demonstrated increased resistance to thioTEPA exposure in vitro, when analyzed by colony survival assays (50 μg/ml, 27 ± 3 versus 16 ± 1; mean ± SD colonies/103 preselected cells; n = 3; P < 0.01, significant differences also at 25 and 100 μg/ml thioTEPA). After reconstitution of transplanted animals, we analyzed the percentage of GFP-expressing peripheral blood leukocyte by flow cytometric analysis. GFP+ peripheral blood leukocytes were detected in SF91-RE (control), SF91-Fpg, and NT mice, and the frequency correlated with the transduction frequency of BM cells analyzed in vitro prior to infusion (Table 1). Only a low level (∼4–6%) of peripheral blood expressed GFP after infusion of SF91-Fpg-transduced BM cells, mimicking the low level of transduction seen in some human trials.
After hematopoietic reconstitution, mice were administered 10 mg/kg of thioTEPA i.p. given once weekly for 2 weeks as one cycle with a second cycle of thioTEPA treatment administrated after a 3-week interval. Because repeated thioTEPA treatment leads to severe and cumulative pancytopenia in both mice and humans (48), leukocyte (WBC), Plt, and Hb determinations were used as a measure of bone marrow cytotoxicity attributable to repeated thioTEPA treatments. WBC, Plt counts, and Hb concentrations showed no significant difference between the SF91-RE (control) and SF91-Fpg mice prior to thioTEPA treatment (Fig. 3). Throughout 8 weeks of treatment with thioTEPA, the reduction in peripheral blood counts and Hb was less severe in SF91-Fpg mice compared with SF91-RE (control) mice (Fig. 3). The difference between the leukocyte counts in SF91-Fpg and SF91-RE (control) mice reached significance by the time of first postchemotherapy nadir and remained significant at 7 weeks (Fig. 3 A).
Plt counts were also higher in SF91-Fpg mice at the first nadir (238 ± 142 × 103/μl versus 100 ± 113 × 103/μl; P < 0.05; Fig. 3 B). Recovery of Plt counts from the nadir was faster in SF91-Fpg mice than that in the SF91-RE (control) mice, and the difference between the Plt counts of SF91-Fpg and SF91-RE (control) mice was significant at 4 and 5 weeks after thioTEPA treatment.
Hb concentrations of SF91-Fpg and SF91-RE (control) mice showed similar decreases in response to the first cycle of thioTEPA treatments. Recovery of Hb concentrations of SF91-Fpg mice was faster than that of the SF91-RE (control) group, and Hb concentrations of SF91-Fpg mice were significantly higher than controls at 4, 5, and 7 weeks after therapy.
During thioTEPA treatment, two of seven SF91-RE (control) mice died on weeks 2 and 3 after the beginning of the first cycle. The remainder, five of seven (71%) control animals, remained alive at 8 weeks after the initiation of the first cycle of thioTEPA treatment, whereas seven of seven animals transplanted with SF91-Fpg-transduced BM cells survived. As expected, survival of transplanted mice that did not receive thioTEPA was 100% (six of six). Autopsy of dead (n = 1) or moribund (n = 1) SF91-RE (control) animals demonstrated significant BM, splenic, and thymic hypoplasia. No other gross abnormalities were seen in the autopsied animals. The toxicity in other organs in surviving animals was also analyzed by histological examination. No histological changes were noted in multiple other tissues including, lung, heart, liver, small intestine, and kidney, and there were no differences noted between SF91-RE (control) or SF91-Fpg group mice.
BM, Spleen, and Thymus Cellularity, Progenitor Content, and Immunophenotyping.
The cellularity of the BM, spleen, and thymus provides another in vivo measure of thioTEPA-induced cytotoxicity. We examined bone marrow, spleen, and thymic cellularities after four thioTEPA doses and 12 weeks after transplantation. As seen in Fig. 4, expression of Fpg in bone marrow cells was associated with significantly higher marrow cellularity compared with SF91-RE (control; 40.0 ± 6.9 × 106versus 25.7 ± 7.8 × 106, Fpg versus Control; P < 0.01), and the marrow cellularity of SF91-Fpg mice treated with thioTEPA did not differ significantly from nontreated mice (Fig. 4,A). Similarly, increased cellularity was demonstrated in spleen (86.9 ± 18.9 × 106versus 31.1 ± 9.5 × 106, SF91-Fpg versus SF91-RE (control); P < 0.005; Fig. 4,B) and thymus (9.5 ± 5.9 × 106versus 1.5 ± 1.0 × 106, SF91-Fpg versus SF91-RE (control), respectively; P < 0.05; Fig. 4 C) of Fpg mice. In summary, expression of Fpg after retrovirus transduction nearly or completely reversed the hypocellularity associated with thioTEPA treatment in bone marrow, spleen, and thymus.
Expression of Fpg was also associated with significantly higher survival of BM progenitors after thioTEPA treatment compared with SF91-RE (control) mice (Fig. 4,D; 339 ± 91.8 versus 122 ± 86.8 × 105 cells, SF91-Fpg versus SF91-RE (Control); P < 0.005). The number of GFP+ progenitors was significantly higher in thioTEPA-treated SF91-Fpg mice compared with nontreated SF91-Fpg mice (3.3 ± 1.1 versus 0.6 ± 0.4%; mean ± SD; n = 6–7; P < 0.0001). Overall, mice transplanted with SF91-Fpg-transduced BM cells had nearly normal BM cellularity and progenitor cell content 3 weeks after the last thioTEPA treatment when compared with nontreatment SF91-Fpg animals (Fig. 4, A and D).
T- and B-cell numbers in spleen and thymus were also examined after thioTEPA treatment (Fig. 4, E and F). Both CD3+ and B220+ cells in the spleen were significantly protected in SF91-Fpg mice compared with SF91-RE (control) mice (CD3+ cells: 20.2 ± 6.6 × 106versus 9.3 ± 5.2 × 106; P < 0.05; B220+ cells: 32.4 ± 12.7 × 106versus 10.2 ± 9.5 × 106; P < 0.01). All T-cell subsets, including CD4+, CD8+ single positive, CD4+CD8+ double-positive and double-negative thymocytes were higher in SF91-Fpg mice compared with SF91-RE (control) mice (Fig. 4 F), indicating either protection of mature and immature thymocytes or protection of a significant numbers of lymphoid progenitors that were capable of giving rise to mature cells in the time frame of the experiment in Fpg mice.
Analysis of GFP Expression in Peripheral Blood after thioTEPA Treatment.
To determine whether treatment of Fpg mice with thioTEPA was associated with selection of Fpg-expressing cells in vivo, we examined the number of GFP+ peripheral blood cells during treatment. Because of the bicistronic nature of the SF91-Fpg retrovirus vector, GFP expression is a useful marker not only of the number of transduced cells but also of the level of transgene expression. As shown in Fig. 5,A, GFP analysis demonstrated two interesting findings: (a) the number of very bright GFP+ cells was increased after thioTEPA treatment (MFI cutoff, 1150; %GFP+ pretreatment, 0.8%; %GFP+ posttreatment, 2.5%), suggesting in vivo selection for these cells; and (b) the number of low or medium GFP-expressing cells declined after treatment (%GFP+ pretreatment, 5.1%; %GFP+ posttreatment, 2.2%), suggesting a threshold level of Fpg expression was required to protect hematopoietic cells in vivo. These data suggest that increased protection of blood cells in vivo might be attained if higher transgene expression can be attained. As seen in Fig. 5,B, overall, MFI of peripheral blood mononuclear cells of SF91-Fpg mice was significantly increased after thioTEPA treatment compared with pretreatment MFI (944 ± 178 versus 523 ± 205; 8 weeks versus pretreatment; P < 0.05), whereas the MFI of SF91-Fpg mice that were not treated (NT) was not changed over this same period of analysis. MFI of treated SF91-RE (control) mice, as expected, did not change after thioTEPA treatment. Consistent with results of peripheral blood, high GFP-expressing BM cells in SF91-Fpg mice were observed after thioTEPA treatment, whereas GFP expression of BM cells in SF91-Fpg mice that were not treated with thioTEPA (NT) remained at a low level, and the number of GFP+ cells declined significantly (Fig. 5 C). These results suggest that BM cells expressing Fpg at high levels selectively survived in vivo after thioTEPA treatment.
Demonstration of Provirus Integration in Hematopoietic Tissue of Transplanted Mice.
Mice were analyzed by PCR for the presence of Fpg cDNA at 12 weeks after BM transplantation (Fig. 6,A). BM, spleen, and thymus DNA from transplanted mice was subjected to PCR amplification with vector-specific 5′-SF91 and 3′-IRES primers, and the products were analyzed after transfer with a labeled fragment of the bacterial Fpg cDNA. The 5′- and 3′-oligonucleotide primers were based on the SF91-RE vector specific-sequence (Fig. 1) and amplify a 960-bp fragment in the presence of the fpg gene in SF91-RE.
Twelve weeks after transplantation, DNA from BM, spleen, and thymus cells of SF91-Fpg thioTEPA-treated mice (Fig. 6,A, Lanes 6–12) and SF91-Fpg nontreated mice (Fig. 6 A, Lanes 13–18) transplanted with SF91-Fpg-transduced BM was positive for Fpg sequences, indicating the presence of transduced cells in both the myeloid and lymphoid lineages. As expected, BM, spleen, and thymus DNA transduced with SF91-RE (control) vector-infected BM (Lanes 1–5) were negative for Fpg sequences. BM harvested from these mice also demonstrated EGFP expression in progenitor colonies at a low level. Mice transplanted with SF91-Fpg-transduced BM and treated with thioTEPA demonstrated 3.3 ± 1.1 EGFP+ colonies per 105 cells versus 0.6 ± 0.4 EGFP+ colonies per 105 cells in BM from SF91-Fpg mice not treated with thioTEPA (mean ± SD, n = 6–7; P < 0.0001).
Retroviral Integration in BM Cells of Secondary Transplant Mice.
To determine whether hematopoietic stem cells were transduced and present in primary mouse recipients after thioTEPA treatment, serial transplants were performed. BM cells from either SF91-RE (control), SF91-Fpg-treated or SF91-Fpg primary mice that were not treated with thioTEPA were used as BM donors for secondary transplants. Six weeks after the infusion of cells, the secondary recipients were treated with thioTEPA in the same doses of primary recipient.
All secondary transplant recipients of donor stem cells derived from the SF91-Fpg-treated and nontreated mice survived four additional thioTEPA treatments, whereas only one-third of recipients of stem cells derived from SF91-RE (control) donors survived this treatment presumably, because the number or repopulating activity of stem cells was severely decreased by thioTEPA treatment in the primary (unprotected) animals.
BM DNA from surviving secondary transplant mice was subjected to PCR amplification with vector-specific 5′-SF91 and 3′-IRES primers, and the products were analyzed on blots probed with a labeled fragment of the bacterial Fpg cDNA. As expected, given the very low gene transfer efficiency in the initial BM, none (none of three) of secondary recipients of donor stem cells derived from SF91-Fpg-nontreated mice were positive for the Fpg sequence (Fig. 6,B). In contrast, BM from 50% (two of four) of secondary recipient of donor stem cells derived from SF91-Fpg treated mice were positive for the FPG sequence (Fig. 6 B), consistent with in vivo selection of transduced cells after thioTEPA treatment in primary SF91-Fpg mice. As expected, BM DNA from secondary recipient of donor stem cells derived from SF91-RE (control) mice was negative for Fpg sequences.
Transduction and Analysis of BM Progenitor Cells Transduced with SF91-hOGG1.
The data presented above show that expression of the bacterial fpg gene in BM cells increases the resistance of these cells to cytotoxic doses of thioTEPA. A potential obstacle to successful application of gene transfer of Fpg in human gene therapy trials might be an immunological reaction against expressed recombinant protein, in this case a bacterial-derived protein. Therefore, we next investigated the capacity of the hOGG1 gene, a functional analogue of bacterial fpg, to protect BM cells from thioTEPA exposure. BM cells were infected with virus supernatant containing SF91-hOGG1 retrovirus (Fig. 1,A and Table 1). Infection was carried out at an MOI of 5–10. Transduction efficiency of BM cells infected with the SF91-OGG1 virus was 69.3%, as determined by GFP expression. GFP+ cells were purified using FACS, and the OGG1 protein activity was determined using an oligonucleotide cleavage assay. As seen in Fig. 7,A, cell extracts from SF91-OGG1-transduced BM cells contained significantly higher 8-oxoguanine repair activity compared with SF91-RE (control) cells. Subsequently, transduced BM cells were exposed to increasing concentrations of thioTEPA in vitro. Surviving cells were analyzed by standard colony-forming assays or incubated in liquid culture for 7 days in the presence of cytokines and analyzed by flow cytometric analysis. Transduction efficiencies of BM infected at an MOI of 2–3 with SF91-RE (control) and SF91-hOGG1 virus prior to exposure of thioTEPA were 32.1 ± 8.1% and 27.0 ± 5.4%, respectively. A significant increase in the survival of progenitor cells transduced with SF91-hOGG1 compared with SF91-RE (control) cells was demonstrated after exposure to thioTEPA (Fig. 7 B). Protection was more pronounced in middle and lower concentrations of thioTEPA, whereas no increase in survival of progenitor cells was observed in the highest concentration. These data suggest that in a manner similar to Fpg, higher levels of hOGG1 expression are required to protect cells from the cytotoxic effects of thioTEPA.
We also examined whether SF91-hOGG1-transduced BM cells are selected in vitro after exposure to thioTEPA. After transduction with hOGG1, the proportion of GFP+ BM cells was increased after exposure to 25 μg/ml of thioTEPA (Fig. 7 C), whereas the proportion of GFP+ SF91-RE (control) BM cells was not changed after thioTEPA treatment. These results indicate that SF91-hOGG1-transduced BM cells are protected from the toxicity of thioTEPA in vitro.
Current approaches using gene transfer technology in human therapies have focused on correction of genetic diseases and modulation of tumor immunogenicity, tumor sensitization, or tumor vaccination strategies in cancer treatment protocols (49). Another approach for cancer therapy using gene transfer technology focuses on reduction of therapy-induced side effects, particularly myelosuppression, which is a common limiting side effect of dose-intensified chemotherapy regimens [reviewed in Moritz and Williams (50)]. To date, transfer of several genes encoding proteins associated with chemotherapy resistance has been investigated in preclinical models and early phase human trials, including dihydrofolate reductase, MGMT, multidrug resistance protein 1, and glutathione S-transferase proteins. In the present study, we have investigated the expression of genes encoding glycosylases of the base excision repair pathway, fpg and hOGG1, and demonstrate increased resistance of target cells to thioTEPA, a commonly used alkylating agent associated with myelosuppression. Expression of either Fpg or OGG1 in primary murine hematopoietic cells, which normally express very low endogenous levels of OGG1 (26, 31), was associated with an increase the oxoguanine DNA glycosylase activity. The mechanism of the differences in selective phenotypes in progenitor versus differentiated blood cells is currently unclear. Two possible explanations, that overexpression of the repair protein is toxic to more differentiated cells in vivo or that differentiated cells require higher levels of protein for protection, are currently being investigated. An additional explanation is that the lower activity of APE seen in more differentiated cells5 may reduce the activity of Fpg/OGG1 as reported recently by Hill et al. (51).
Fpg and hOGG1 proteins are functionally very similar enzymes, and cellular repair activity for 8-oxoG and methyl-Fapy primarily relies on these proteins. Despite functional similarity, the primary amino acid sequences of these proteins differ significantly. Moreover, OGG1, but not Fpg, is a member of the endonuclease III superfamily that contains a hairpin-helix-hairpin/GPD motif involved in the glycosylase/AP (apurinic/apyrimidinic)-lyase activity (52). In addition, the two proteins, despite the functional similarities, differ with respect to their substrate specificity. hOGG1 excises 8-oxodG paired with dC but has little or no effect when the lesion is paired with other DNA bases, whereas Fpg excises 8-oxodG paired with dC, dG, and dT (53, 54). Asagoshi et al. (55) have compared the repair activity of methyl-Fapy lesions between Fpg and OGG1 in detail and have reported that Fpg did not show obvious consensus sequences for efficient repair, hOGG1 showed a preference for 5′-(C/G)-Fapy-C-3′ sequences. Therefore, it is possible that overexpression of Fpg protein compensates the repair activity of endogenous hOGG1 in murine and presumably human hematopoietic cells. On the other hand, potential immunological reaction against the bacterial Fpg protein may be an obstacle in future human clinical trials.
Few previous studies have investigated the use of overexpressing BER enzymes in affording protection against DNA-damaging agents, and those have met with limited success (56). For example, the first enzyme in the human DNA BER pathway, MPG, removes the damaged base after alkylating agent treatment. Although the overexpression of MPG might be expected to enhance repair, particularly if it was rate limiting, the overexpression of MPG leads to increased sensitivity to alkylating agents and chromosomal aberrations, suggesting that this first step in BER is not rate limiting and that the accumulation of unrepaired AP sites is cytotoxic to cells (57).
Overexpression of another BER protein, Ape1/ref-1, has produced mixed results with some studies showing cellular protection, whereas others either demonstrated no protection or a deleterious effect, depending on the drugs used (58). For example, overexpression of the yeast AP endonuclease, Apn1, but not the human Ape1/ref-1 in Chinese hamster ovary cells resulted in cells that were resistant to the alkylating agent methyl methanesulfonate and the oxidizing agent, hydrogen peroxide (59). Additionally, a chimeric BER repair protein made up of MGMT fused to Ape1/ref-1 showed significant protection against 1,3-bis(2-chloroethyl)-1-nitrosourea and methyl methanesulfonate in mammalian cells (44). This protection has also been observed in a MGMT-Apn1 fusion.6 Finally, when the polymerase involved in BER, DNA polymerase β, is overexpressed in mammalian cells, the cells acquired increased resistance to chemotherapeutic drugs (cisplatin, melphalan, and mechlorethamine); however, this resistance was achieved at the expense of an increase in a mutator phenotype (60, 61). However, recent studies have shown that the targeting of nuclear OGG1 protein to the mitochondria results in enhanced protection of mammalian cells against oxidizing DNA-damaging agents (62). Other members of the DNA BER pathway, such as DNA ligase I, human endonucleases III (NTH1), and Ogg1-2a (mitochondrial Ogg), are currently being evaluated for their protective ability against DNA-damaging agents (56, 63). Additionally, any chemotherapy agent leading to Fapy lesions, including chloroethyl alkylating agents and agents leading to oxidant DNA damage, could potentially be mitigated by expressing Fpg or OGG-1.
Phenotypic alterations of transduced cells, particularly in the setting of chemoresistance as shown here, may require high-level expression of vector-encoded protein. Issues related to vector-directed transgene expression include positional variegation of the level of expression, extinction of expression over time, and vector silencing (64). To address these issues, multiple oncoretroviral backbones have been tested, and newer, often chimeric backbones or backbones containing specific sequence modifications have been developed to improve transgene expression in primary cells, particularly hematopoietic stem cells. Our data demonstrate considerable differences in expression in primary hematopoietic cells between an MSCV-based and SF91-based vector systems confirming previous observations (37). Of note, the comparison based on the expression of the IRES-controlled EGFP may not reveal the complete gain of expression that can be achieved with SF91, as modifications of the 5′UTR designed to improve the efficiency of translation of a cap-dependent transgene (37) are not reflected. Consistent with the hypothesis that the level of expression of Fpg is likely critical, only transduced cells expressing high levels of transgene from the SF vector backbone survived thioTEPA treatment in vitro or in vivo, and expression of Fpg in the MSCV backbone led to no demonstrable thioTEPA resistance in primary cells.7 Thus, the level of expression required to protect hematopoietic cells from thioTEPA-induced cytotoxicity appears to be very high, and successful application of gene transfer to generating a chemoresistance phenotype in human trials will likely require vectors that encode even higher levels of protein in these cells. In this regard, the use of modules for posttranscriptional enhancement of gene expression, such as constitutive RNA transport elements, posttranscriptional regulatory elements affecting RNA translation, and appropriate use of splicing signals may be critical and are currently being tested (65). Expression of Fpg may also profit from optimizing the codon usage.
Another potential problem in this approach concerns the low level retrovirus transduction of hematopoietic stem cells in large animals and humans, which may be an obstacle to the successful application of gene transfer technology to therapeutic applications. This may be attributable to the quiescent nature of human stem cells and poor virus receptor expression in the target cells (66, 67). Typically, current protocols result in modification of <5% of peripheral blood cells (68, 69), although two recent studies incorporating a variety of transduction modifications present more encouraging results (70, 71). Low levels of gene transfer may reduce the effective generation of the chemotherapy resistance phenotype and thus hinder clinical effectiveness of this approach. However, one potential advantage of the use of chemotherapy-resistance genes in the setting of cancer therapy is in vivo selection of transduced stem and progenitor cells, as demonstrated previously with mutants of dihydrofolate reductase and MGMT (18, 72).
In conclusion, the results presented here demonstrate a potential application of using DNA BER proteins for BM protection from DNA alkylating agents. A clear implication of these data is the necessity of correct pairing of the specific BER enzyme with a vector backbone that encodes appropriate expression in the target cell population and matching of BER protein with specific chemotherapy agents.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by NIH Grant P01 CA75426.
The abbreviations used are: thioTepa, N′,N′,N″-triethylenethiophosphoramide; BER, base excision repair; Fapy, formamidopyrimidine; Fpg, Fapy-DNA glycosylase; hOGG1, human oxoguanine DNA glycosylase; MSCV, murine stem cell virus; EGFP, enhanced green fluorescence protein; FACS, fluorescence-activated cell sorter; BM, bone marrow; IMDM, Iscove’s modified Dulbecco’s medium; Plt, platelet; Hb, hemoglobin; 8-oxoG, 8-oxoguanosine; LTR, long-terminal repeat; UTR, untranslated region; MESV, murine embryonic stem cell virus; SFFV, spleen focus-forming virus; MOI, multiplicities of infection; FACS, fluorescence-activated cell sorting; MFI, mean fluorescence intensity; MGMT, O6-methylguanine DNA methyltransferase; MPG, methylpurine DNA glycosylase.
K. Pollok and D. A. Williams. Differential transduction efficiency of SCID-repopulating cells derived from umbilical cord blood and G-CSF mobilized peripheral blood, manuscript in preparation.
M. R. Kelley and D. A. Williams, unpublished data.
M. R. Kelley and J. C. Roth. Human yeast chimeric repair protein protects alkylating cells against agent, manuscript in preparation.
M. Kobune and D. A. Williams, unpublished data.
|Experimental group .||% EGFP expression of infused cells .||% EGFP-expressing peripheral leukocytes after transplantationa .|
|Control||16.3 ± 0.8||25.9 ± 15.1|
|Fpg||3.9 ± 0.1||5.9 ± 2.5|
|NTb||3.9 ± 0.1||3.9 ± 2.2|
|Experimental group .||% EGFP expression of infused cells .||% EGFP-expressing peripheral leukocytes after transplantationa .|
|Control||16.3 ± 0.8||25.9 ± 15.1|
|Fpg||3.9 ± 0.1||5.9 ± 2.5|
|NTb||3.9 ± 0.1||3.9 ± 2.2|
Analyzed 4 weeks after transplantation (n = 6–7).
NT, nontreatment group transduced with SF91-Fpg.
We thank Susanne Ragg, Wen Tao, Melissa Limp Foster, and members of our laboratory for helpful discussion, as well as Jeffery Bailey for assistance in the preparation of all reagents.