Chemoprotective gene transfer II: multilineage in vivo ... - Nature

Gene Therapy (1998) 5, 842–847  1998 Stockton Press All rights reserved 0969-7128/98 $12.00 http://www.stockton-press.co.uk/gt

Chemoprotective gene transfer II: multilineage in vivo protection of haemopoiesis against the effects of an antitumour agent by expression of a mutant human O6-alkylguanine-DNA alkyltransferase N Chinnasamy1,2, JA Rafferty1,2, I Hickson1,2, LS Lashford3, SJ Longhurst1,2, N Thatcher4, GP Margison1, TM Dexter2 and LJ Fairbairn2 1

CRC Sections of Genome Damage and Repair and 2Haemopoietic Cell and Gene Therapeutics, Paterson Institute for Cancer Research, 3Academic Unit of Paediatric Oncology and 4CRC Department of Medical Oncology, Christie Hospital (NHS) Trust, Manchester, UK

Murine bone marrow cells were transduced ex vivo with a retrovirus encoding an O6-benzylguanine (O6-beG) insensitive, double mutant form of the human DNA repair protein O6-alkylguanine-DNA alkyltransferase (hATPA/GA). In animals reconstituted with the transduced bone marrow, about 50% of cells in the multipotent spleen colony-forming cells (CFU-S) and lineage restricted granulocyte–macrophage (GM-CFC) haemopoietic progenitor populations were found to be carrying the transgene and this correlated with the frequency of bone marrow cells and spleen colonies which stained positive for hATPA/GA by immunocytochemistry. Expression of hATPA/GA was associated with significant in vivo protection of both CFU-S (P = 0.001) and GM-CFC (P ⬍ 0.024) against the toxicity of the antitumour methylating agent, temozolomide, given in combination with O6-beG. Expression of hATPA/GA also led to a

reduction in the frequency of combined O6beG/temozolomide-induced micronuclei seen in polychromatic erythrocytes (P ⬍ 0.003). This study is the first to demonstrate in vivo protection of multipotent haemopoietic progenitors against the toxic and clastogenic effects of an O6-alkylating agent in the presence of O6-beG. It also represents the first report of reduced clastogenesis as a consequence of expression of an O6-beG-resistant ATase. In the accompanying article we report hATPA/GAmediated resistance of human CD34+ haemopoietic progenitors to combined O6-beG/O6-alkylating agent toxicity. Together these two reports suggest that a gene therapy strategy whereby protection of normal haemopoietic tissue may be combined with O6-beG-mediated tumour sensitisation may be efficacious in achieving an increase in therapeutic index.

Keywords: alkyltransferase; O6-benzylguanine; CFU-S; micronucleus; temozolomide

Introduction As discussed in detail in the accompanying article by Hickson et al, chemotherapy involving the O6-alkylating agents is often complicated by acute dose-limiting toxicity to bone marrow and other tissues such as lung and is also associated with a long-term risk of therapy-related secondary malignancy, particularly leukaemias. The relative sensitivity of bone marrow to these agents is probably due to the expression of low levels of O6-alkylguanine-DNA alkyltransferase (ATase; hAT).1 However, in order to overcome ATase-mediated drug resistance in tumours,2 administration of O6-benzylguanine (O6-beG), which inactivates ATase, is being used to achieve tumour sensitisation to O6-alkylating agents. Although able to potentiate the cytotoxicity of O6-alkylating agents in human tumour xenografts,3–6 O6-beG also increases the Correspondence: LJ Fairbairn, CRC Section of Haemopoietic Cell and Genome Therapeutics, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Wilmslow Road, Manchester M20 4BX, UK Received 28 January 1998; accepted 17 March 1998

toxic and clastogenic effects of temozolomide and BCNU in murine bone marrow progenitor cells in vivo.7 In the accompanying article by Hickson et al,8 we have shown that retrovirus-mediated transduction of an O6beG resistant human ATase (hATPA/GA) leads to in vitro protection of human haemopoietic cells, including CD34+ progenitors, against chemotherapeutic O6-alkylating agents in the presence of O6-beG. Whilst protection in in vitro colony-forming assays provides important proof of principle, this form of analysis is not informative with respect to the protection afforded to the repopulating haemopoietic stem cell compartment. Such data can only be obtained in vivo and for this reason we have developed a murine stem cell protection model. In this article we report multilineage protection of murine haemopoietic progenitors in vivo against the cytotoxic and clastogenic effects of combined O6-beG and temozolomide treatment, in mice reconstituted with bone marrow transduced with retrovirus encoding hATPA/GA.

Chemoprotective gene transfer II N Chinnasamy et al

843

Results Transduction of murine bone marrow PCR analysis of GM-CFC colonies derived from bone marrow plated immediately after co-cultivation with producer cells, indicated a 50–60% transduction frequency in this population (Figure 1). Similarly, PCR analysis indicated that 50–55% of day 12 spleen colonies from reconstituted mice arose from transduced marrow cells (data not shown). Furthermore, all GM-CFC and spleen colonies assayed showed the presence of the Y-chromosome indicating that they were derived from the transplanted marrow and that the mice had been fully reconstituted with donor cells. Expression of hATPA/GA in bone marrow cells and spleen colonies Immunohistochemical analysis of bone marrow cytospin preparations using a polyclonal antiserum specific for hAT was used to examine the expression of the transgene following reconstitution with the in vitro transduced bone marrow. When control, bone marrow cells that had been exposed to naive rather than retrovirus producing producer cells (mock-transduced) were incubated with the antibody, no signal was detected (Figure 2a), consistent with the lack of in situ cross-reactivity of this antibody with the murine ATase. In contrast, cells analysed 2 days following co-cultivation with hATPA/GA producer cells showed clear staining for hATPA/GA, which was localised to the nucleus (Figure 2b). The proportion of hATPA/GA positive cells (approximately 40%) detected by image analysis of nucleated bone marrow cells correlated with the frequency of transduction determined by PCR analysis of GM-CFC. Similarly, immunohistochemical analysis of spleen sections 12 days following reconstitution of mice with LhATPA/GA transduced (Figure 3b), but not mock-transduced (Figure 3a), bone marrow cells showed intense nuclear staining in around 50% of spleen colonies. Cytotoxic effects of temozolomide in mice reconstituted with mock or hATPA/GA-transduced bone marrow cells Four weeks following reconstitution with mock-transduced or LhATPA/GA-transduced bone marrow, mice

Figure 1 PCR analysis of GM-CFC colonies from hATPA/GA-transduced bone marrow cultures. Colonies analysed using primers specific for human ATase (transduced cells) (a) and mouse Y-chromosome (male, donor cells) (b). A ‘+’ symbol indicates a positive control reaction which for panel a was a plasmid containing an ATase cDNA sequence and for panel b was DNA prepared from a GM-CFC colony obtained from a male mouse. MW denotes molecular weight markers.

Figure 2 Immunohistochemical analysis of cytospin preparations taken from mock-transduced (a) and hATPA/GA-transduced (b) cultures 2 days following the transduction procedure, using a polyclonal antihuman ATase antibody and avidin-biotin peroxidase/diaminobenzidine as described in Materials and methods.

were challenged with temozolomide (60 mg/kg), either alone or following pretreatment with O6-beG at a dose (30 mg/kg) which has been previously shown to sensitise mouse bone marrow in vivo to killing by temozolomide,7 and the surviving fractions of CFU-S and GM-CFC were determined 24 h later. Treatment with temozolomide reduced CFU-S levels in mock-transduced mice both in the absence (32% survival) and presence (10% survival) of O6-beG. In mice transplanted with hATPA/GA transduced marrow, a statistically significant protection of bone marrow CFU-S in vivo following the exposure of mice to temozolomide alone (66% survival, P = 0.017) or in combination with O6-beG (38% survival, P = 0.001) was observed (Figure 4). Retroviral transduction had no effect on the formation of spleen colonies per se as the baseline value for the hATPA/GA transduced versus mock-transduced cells did not differ significantly. Similarly, in the same animals, GM-CFC survival was significantly higher in hATPA/GA transduced mice (33% survival) compared with control animals (13% survival, P = 0.024) following combined treatment.


844

Figure 4 In vivo survival of CFU-S following treatment of control (open bars) or hATPA/GA-transduced (shaded bars) mice with temozolomide (60 mg/kg) either alone or in combination with O6-beG (30 mg/kg). Data expressed as percentage of the no treatment control ± s.e.m., P values as indicated by asterisks (Student’s t test). The baseline CFU-S number in transduced versus mock-transduced mice did not differ significantly.

Figure 3 Immunohistochemical analysis of spleen colonies from mice reconstituted with mock-transduced (a) or hATPA/GA-transduced (b) bone marrow cells, using a polyclonal antihuman ATase antibody and avidin-biotin peroxidase/diaminobenzidine as described in Materials and methods.

Clastogenic effects of temozolomide in mice reconstituted with mock or hATPA/GA transduced bone marrow cells Expression of hATPA/GA in the bone marrow of reconstituted mice had no effect on the background number of micronuclei observed in the absence of any treatment (Figure 5). When control mice were exposed to temozolomide at a dose (10 mg/kg) which generates maximum sublethal damage, the frequency of micronuclei observed was 18.7 ± 2.74 per 1000 polychromatic erythrocytes (PCE) in the absence of O6-beG and 27.5 ± 2.02 in mice pretreated with O6-beG. In hATPA/GA transduced mice, however, a lower frequency of micronucleus induction was seen, following treatment with temozolomide alone (12.0 ± 1.73 per 1000 PCE, P = 0.109 when compared with mock-transduced mice treated with temozolomide alone) or in combination with O6-beG (15.0 ± 0.00 per 1000 PCE, P = 0.003 when compared with mock transduced mice treated with temozolomide and O6-beG).

Figure 5 Bone marrow micronucleus frequency, expressed as number of micronuclei per 1000 polychromatic erythrocytes, following treatment of control (open bars) or hATPA/GA-transduced (shaded bars) mice with temozolomide (10 mg/kg) either alone or in combination with O6-beG (30 mg/kg).

Discussion The problem of tumour resistance to killing by chemotherapeutic O6-alkylating agents is often exacerbated by the extreme dose-limiting sensitivity of normal tissues such as bone marrow to these agents with subsequent


dose reductions jeopardising the effectiveness of therapy. Furthermore, DNA damage in surviving cells may lead to long-term consequences such as the induction of leukaemias in patients. The use of low molecular weight inactivators of ATase, such as O6-beG, has raised the prospect of increasing tumour sensitivity to these agents. However, we have previously shown that the sensitivity of human and murine bone marrow progenitors to the cytotoxic (acute) effects of agents such as temozolomide is increased by prior exposure to O6-beG, as is the frequency of clastogenic events (indicative of long-term effects) in murine bone marrow in vivo.7,9 In the accompanying study by Hickson et al we show O6-beGinsensitive protection of human haemopoietic cells against O6-alkylating agent toxicity in vitro. Here we report the production of mice expressing hATPA/GA in their bone marrow and the consequent effects of expression on the sensitivity of haemopoietic progenitors in vivo to the cytotoxicity and clastogenicity of temozolomide. A transduction efficiency of 50–60% was observed in the bipotent GM-CFC population plated immediately following co-cultivation with retroviral packaging cells. A similar frequency was seen in CFU-S and GM-CFC from mice 12 days and 4 weeks, respectively, after reconstitution with transduced marrow. The frequency of transduction determined by PCR analysis correlated well with the observation that approximately 40% of the cells from bone marrow cultures and 50% of colonies in spleens of recipient mice stained positive for hATPA/GA on immunohistochemical analysis of these tissues, indicating efficient transfer and expression of the hATPA/GA cDNA. Moreover, the hATPA/GA staining was predominantly nuclear in localisation, which is of obvious importance for achieving repair of genomic DNA. We next studied the effects of hATPA/GA expression in bone marrow on the in vivo sensitivity of haemopoietic progenitors to the cytotoxic effects of temozolomide, either alone or in combination with O6-beG. The dose of inactivator used had been previously shown to reduce the endogenous wild-type ATase activity substantially in mouse bone marrow and the dose of temozolomide was predicted to lead to decimation of the GM-CFC and CFUS compartments in unprotected mice when administered in combination with O6-beG.7 Exposure of hATPA/GAtransduced mice to temozolomide in the absence of O6beG, resulted in a clear and statistically significant protection of early bone marrow progenitors (CFU-S) compared with the controls (Figure 4). Most striking, however, was the effect on survival of progenitors in mice treated with the combination of O6-beG and temozolomide. In these animals, a significant protective effect of hATPA/GA expression was seen in both the CFU-S and GM-CFC compartments, with over a third of primitive progenitors in bone marrow surviving treatment. The significance of this protective effect is greater if one considers that a maximum of 50% of the progenitors exposed to temozolomide carried the transgene. This means that in this one cycle of treatment, at least two-thirds of the transduced CFU-S and GM-CFC cells expressed sufficient levels of hATPA/GA to ensure repair of DNA damage to below cytotoxic levels. In a clinical setting, however, cytotoxic antitumour treatment is not usually given as a single dose, but in several cycles of multiple doses. In such a case, it would seem likely that selection of transduced

cells would occur. Such an effect is not without precedence in murine models10,11 and were this to translate into the human situation, cells expressing hATPA/GA would become more prevalent in patients following the first cycle of chemotherapy. One of the problems encountered with repeat dosing of temozolomide during chemotherapy is a progressive delay in marrow recovery.12 The selection of a drug resistant population in the bone marrow may abrogate this progressive toxicity and permit the maintenance of an intensive dose schedule or provide the potential for dose escalation. A more insidious side-effect of chemotherapy with genotoxic agents is the occurrence of secondary malignancies, predominantly leukaemias,13 as a result of the survival of sublethally damaged cells. Whilst strategies such as the use of growth factors or peripheral blood stem cell rescue can sometimes ameliorate short-term toxicities, they are likely to exacerbate the longer term problem of iatrogenic tumours as patient survival increases. It will become of increasing importance to address this problem and to find ways to reduce the frequency of induction of secondary leukaemia. Therefore we assessed the effects of hATPA/GA expression on the frequency of preleukaemic lesions in order to determine whether the gene therapy approach might also give rise to protection against iatrogenic tumours. As a direct assay for mutation frequency in human bone marrow progenitors does not yet exist, we used the mouse bone marrow micronucleus test, which serves as one of the ‘industry standard’ tests for the detection of carcinogens, including leukaemogens,14 as a surrogate marker for mutation in the stem cell and progenitor compartments. Our observation that expression of hATPA/GA in around 50% of mouse bone marrow cells leads to significant protection against micronucleus formation in PCE whether O6-beG pretreatment is administered or not, strongly suggests that it may prove possible to provide protection against the mutagenic, and thus potential leukaemogenic, effects of O6-alkylating agents as well as the cytotoxicity of these drugs. This is the first demonstration of in vivo protection of bone marrow against the combined toxic and clastogenic effect of an O6-alkylating agent via gene transfer of a human DNA repair function which is refractory to inactivation by O6-beG. The ability to protect the haemopoietic system as outlined, under conditions where considerable tumour sensitisation to killing by temozolomide will be achieved, holds out real prospects for increasing the therapeutic index of temozolomide/O6-beG combination therapy. Given the compelling nature of our data in both murine and human primary haemopoietic progenitors, we are moving towards evaluating this strategy in a clinical setting.

Materials and methods Reagents for treatment of animals Temozolomide was obtained from the Cancer Research Campaign Formulation Unit, Department of Pharmaceutical Sciences, University of Strathclyde, Glasgow, UK. It was dissolved in DMSO at 30 mg/ml and diluted in phosphate-buffered saline for injection into mice. O6-benzylguanine (a generous gift from Dr RS McElhinney and Prof B McMurry, Trinity College, Dublin, Republic of

845


846

Ireland) was suspended homogeneously in corn oil using a TEFLON glass homogeniser (Scientific Laboratory Supplies, Nottingham, UK). Both solutions were freshly prepared immediately before use.

Retroviral vector construction and generation of producer cell line The hATPA/GA cDNA was cloned into a BamHI site engineered into the retroviral vector pLX derived from the LN series of vectors,15 to produce pLhATPA/GA which expressed the mutant hAT protein under the control of the MoMuLV LTR. Producer cell lines were routinely grown in Dulbecco’s minimum essential medium (DMEM; Life Technologies, Paisley, UK) supplemented with 10% newborn calf serum (Life Technologies) 2 mm glutamine. The pL-hATPA/GA vector construct and pcDNA3neo (10:1) were cotransfected into the GP+E86 cells using Lipofectin (Life Technologies). G418 resistant, L-hATPA/GA-positive GP+E86 cells (ecotropic) were cocultivated for 7 days with GP+envAm12 (amphotropic) cells in a ‘ping-pong’ method to increase the titre.16 Following selection with 200 ␮g/ml of hygromycin for 7 days to kill ecotropic producers, a population of amphotropic producer cells were obtained which produced around 5 × 105 infectious particles per millilitre. Retroviral infection of bone marrow, transplantation and temozolomide treatment of mice Femoral bone marrow cells were harvested from 9–12week-old male B6D2F1 mice 48 h after the administration of a single dose of 5-fluorouracil (150 mg/kg, i.p.; David Bull Laboratories, Warwick, UK), and cells were immediately prestimulated for 48 h at 37°C in DMEM supplemented with 20% foetal calf serum, 0.1% BSA, 2 mm glutamine, 200 U/ml recombinant human IL-6 (Sandoz, Basle, Switzerland), 10 ng/ml recombinant murine IL-3 (R&D Systems, Abingdon, UK) and 100 ng/ml recombinant rat stem cell factor (Amgen, Thousand Oaks, CA, USA). The bone marrow cells were then cocultured with either LhATPA/GA retroviral producers or GP+envAm12 cells (mock infection) for 48 h in the same medium with 4 ␮g/ml polybrene, after which supernatant cells were collected and used for both cytospin preparation and injection into 8–10-week-old, lethally irradiated (15.2 Gy, 60Co source at 0.95 Gy per hour) female B6D2F1 mice (6 × 106 cells per mouse). Four weeks after transplantation, groups of four mice were treated with temozolomide (60 mg/kg, i.p.) either alone or 2 h after receiving a single dose of O6-beG (30 mg/kg, i.p.). Mice were killed 24 h later by cervical dislocation and femoral bone marrow cells were harvested for colonyforming unit–spleen (CFU-S) and granulocyte–macrophage colony-forming cell (GM-CFC) analysis as described previously.7 Individual GM-CFC colonies plated in methylcellulose were taken directly for PCR analysis whilst spleen colonies were dissected out carefully and DNA isolated for PCR analysis. PCR analysis PCR analyses were carried out on GM-CFC colonies and DNA isolated from day 12 spleen colonies using hATPA/GA specific primers. The hATPA/GA sense and antisense primers (shown below) gave an expected PCR product of 624 bp.

Sense primer 5⬘ATGGACAAGGATTGTGAAATGAAACG3⬘ Antisense primer 5⬘TCAGTTTCGGCCAGCAGGCGG3⬘

Y chromosome specific primers were used as a control to detect male bone marrow cells as described.17

Immunohistochemistry Cytospin preparations of bone marrow cells and sections of day 12 spleen colonies were stained with a rabbit polyclonal antibody specific for human ATase using an avidin biotin peroxidase complex (Vector Laboratories, Peterborough, UK) in a modification of the procedure previously described.18 Micronucleus induction Three each of control and hATPA/GA mice were treated with temozolomide (10 mg/kg, i.p.) either alone or 2 h after a single dose of O6-beG (30 mg/kg, i.p.) in corn oil. Mice were killed by cervical dislocation 24 h later and femoral marrow cells were flushed into foetal calf serum. Bone marrow smears were fixed in methanol for 15 min, then stained with acridine orange as described by Tinwell and Ashby19 and scored blindly by two separate investigators for micronucleated PCE among 2000 PCE. The results were expressed per 1000 PCE. Statistical analysis The unpaired Student’s t test was used to compare the survival of CFU-S and GM-CFC, and the frequency of bone marrow micronuclei in the control versus hATPA/GA expressing mice exposed to temozolomide alone or in combination with O6-BeG.

Acknowledgements We thank Mrs Lorna Woolford, Mr Mark Willington, Mr John Bailey and Ms Dorothy Gagen for expert technical assistance, Mr Richard Swindell for statistical advice and Dr Brian Lord for helpful discussion. This work was supported by the Medical Research Council (NC, IH), Leukaemia Research Fund (SJL) and Cancer Research Campaign, UK. TMD is a Gibb Research Fellow of the Cancer Research Campaign

References 1 Gerson SL et al. Human CD34+ hematopoietic progenitors have low, cytokine-unresponsive O6-alkylguanine-DNA alkyltransferase and are sensitive to O6-benzylguanine plus BCNU. Blood 1996; 88: 1649–1655. 2 Brent TP et al. Identification of nitrosourea-resistant human rhabdomyosarcomas by in situ immunostaining of O6-methylguanine-DNA-methyltransferase. Oncol Res 1993; 5: 83–86. 3 Mitchell RB, Moschel RC, Dolan ME. Effect of O6-benzylguanine on the sensitivity of human tumor xenografts to 1,3-bis(2chloroethyl)-1-nitrosourea and on DNA interstrand cross-link formation. Cancer Res 1992; 52: 1171–1175. 4 Gerson SL et al. Synergistic efficacy of O6-benzylguanine and 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) in a human colon cancer xenograft completely resistant to BCNU alone. Biochem Pharmacol 1993; 45: 483–491. 5 Dolan ME, Pegg AE. O6-benzylguanine and its role in chemotherapy. Clin Cancer Res 1977; 3: 837–847. 6 Wedge SR, Porteous JK, Newlands ES. Effect of single and multiple administration of an O6-benzylguanine/temozolomide combination: an evaluation in human melanoma xenograft model. Cancer Chemother Pharmacol 1997; 40: 266–272.


7 Chinnasamy N et al. O6-benzylguanine potentiates the in vivo toxicity and clastogenicity of temozolomide and BCNU in mouse bone marrow. Blood 1997; 89: 1566–1573. 8 Hickson I et al. Chemoprotective gene transfer I: transduction of human haemopoietic progenitors with O6-benzylguanineresistant O6-alkylguanine-DNA alkyltransferase attenuates the toxic effects of O6-alkylating agents in vitro. Gene Therapy 1998; 5: 835–841. 9 Fairbairn LJ et al. O6-benzylguanine increases the sensitivity of human primary bone marrow cells to the cytotoxic effects of temozolomide. Exp Hematol 1995; 23: 112–116. 10 Moritz T et al. Retrovirus-mediated expression of a cDNA repair protein in bone marrow protects haemopoietic cells from nitrosourea-induced toxicity in vitro and in vivo. Cancer Res 1995; 55: 2608–2614. 11 Allay JA, Davis BM, Gerson SL. Human alkyltransferase-transduced murine myeloid progenitors are enriched in vivo by BCNU treatment of transplanted mice. Exp Hematol 1997; 25: 1069–1076. 12 Newlands ES et al. Phase I trial of temozolomide (CCRG 81045: M&B 39831: NSC 362856). Br J Cancer 1992; 65: 287–291. 13 Smith MA, McCaffrey RP, Karp JE. The secondary leukemias:

14

15 16

17

18

19

challenges and research directions. J Natl Cancer Inst 1996; 88: 407–418. Tinwell H, Ashby J. Comparative activity of human carcinogens and NTP rodent carcinogens in the mouse bone marrow micronucleus assay: an integrative approach to genetic toxicity data assessment. Environ Health Perspect 1994; 102: 958–962. Miller AD, Rosman GJ. Improved retroviral vectors for gene transfer and expression. BioTechniques 1989; 7: 980–990. Fairbairn LJ, Spooncer E. Retroviral gene transfer into haemopoietic cells. In: Testa NG, Molineux G (eds). Haemopoiesis, A Practical Approach. IRL Press: Oxford, 1993, pp 175–188. Han YM, Yoo OJ, Lee KK. Sex determination in single mouse blastomeres by polymerase chain reaction. J Assist Reprod Genet 1993; 10: 151–156. Lee SM et al. Intercellular and intracellular heterogeneity of O6-alkylguanine DNA alkyltransferase expression in human brain tumours – possible significance in nitrosourea therapy. Carcinogenesis 1996; 17: 637–641. Tinwell H, Ashby J. Comparison of acridine orange and Giemsa stains in several mouse bone marrow micronucleus assays – including a triple dose study. Mutagenesis 1989; 4: 476–481.

847