Hematopoietic stem cell gene therapy: progress toward therapeutic targets

The concept of hematopoietic stem cell gene therapy is as exciting as that of stem cell transplantation itself. The past 20 years of research have led to improved techniques for transferring and expressing genes in hematopoietic stem cells and preclinical models now routinely indicate the ease with which new genes can be expressed in repopulating stem cells of multiple species. Both modified murine oncoretroviruses and lentiviruses transmit genes into the genome of hematopoietic stem cells and allow expression in the host following transplantation. Using oncoretroviruses, therapeutic genes for severe combined immunodeficiency, common variable gamma chain immunodeficiency, chronic granulomatous disease, Hurler's and Gaucher's Disease have all been used clinically with only modest success except for the patients with immunodeficiency in whom a partial T-cell chimerism has been dramatic. Since stem cell selection in vivo appears important to the therapeutic success of gene transfer, drug resistance selection, most recently using the MGMT gene, has been developed and appears to be safe. Future trials combining a drug resistance and therapeutic gene are planned, as are trials using safety-modified lentiviruses. The therapeutic potential of hematopoietic stem cell gene therapy, particularly given recent advances in stem cell plasticity, remains an exceptionally exciting area of clinical research.

Overexpression of human O6-alkylguanine DNA alkyltransferase (AGT) prevents MNU induced lymphomas in heterozygous p53 deficient mice

O6-alkylguanine DNA alkyltransferase (AGT) is a key mechanism in the prevention against MNU induced malignant transformation by removal of O6 methyl guanine (O6mG) adducts. We asked whether heterozygous p53 deficient mice (p53+/-) would be more susceptible to MNU induced lymphomas than wild type mice, and whether O6mG adducts were responsible for this susceptibility. To determine whether MGMT overexpression would be protective, p53+/- mice were bred to human MGMT transgenic mice (MGMT+) and treated with 50 mg/kg MNU. MNU increased the incidence of thymic lymphomas in non-transgenic p53+/- mice from 23% (n=13) to 68% (n=22) and decreased the mean latency from 433 to 106 days (P=0.01 compared to untreated mice). Wild type mice had an incidence of 30% (n=38) and a mean latency of 135 days after MNU. Overexpression of MGMT in the thymus of p53+/- mice significantly reduced the lymphoma incidence from 68 to 28% (n=17) and increased the latency from 106 to 167 days (P=0.003). Similarly, the lymphoma incidence in MGMT+/wild type mice decreased from 30 to 8% (n=12) and the latency increased to 297 days (P=0.2). Loss of the wild type allele was found in only 2/17 lymphomas occurring in p53+/- mice and there were no significant point mutations in exons 5-8 of p53. Furthermore, there was no loss of p53 function in these mice. These data demonstrate that unrepaired O6mG lesions act cooperatively with the reduced p53 dose and lead to lymphomagenesis in p53+/- mice, but AGT overexpression and rapid removal of O6mG adducts is protective.

MGMT expression in murine bone marrow is a major determinant of animal survival after alkylating agent exposure

Myelosuppression is commonly observed after alkylating agent chemotherapy due to low levels of O(6)-alkylguanine DNA alkyltransferase protein (AGT) in hematopoietic progenitors. Mice that lack AGT in all organs, O(6)-methylguanine-DNA methyltransferase gene knockout (MGMT(-/-)) mice are extremely hypersensitive to the methylating agent N-methyl-N-nitrosourea (MNU) and exhibit a 10-fold reduction in the LD(90). To determine whether bone marrow damage was the cause of the increased lethality, we transplanted 1 x 10(6) wild-type marrow into MGMT(-/-) mice and MGMT(-/-) marrow into wild-type mice and observed survival after MNU. Lethally irradiated MGMT(-/-) mice given > or = 25 mg/kg MNU 3 weeks after transplant of wild-type cells survived > 30 days (n = 11), whereas this dose was lethal to control MGMT(-/-) mice 9-12 days post treatment (n = 5). Conversely, lethally irradiated wild-type mice transplanted with MGMT(-/-) cells died after only 20-60 mg/kg MNU within 8-12 days (n = 6). No significant toxicities were found in other organs. Additionally, in an in vivo post transplant competition model, wild-type long-term repopulating cells had a > 200-fold competitive survival advantage over MGMT(-/-) cells, and after MNU treatment completely repopulated the mouse when transplanted at only one-tenth the cell number. We also observed a strong selection for transplanted marrow-derived wild-type stromal elements in the MGMT(-/-) background after drug treatment. These data indicate that alkylating agent hypersensitivity of MGMT(-/-) mice results from hematopoietic damage at the stem level. Thus, DNA repair involving AGT in hematopoietic cells is required for normal host survival following exposure to methylating and chloroethylating agents.

Human mesenchymal stem cells provide stromal support for efficient CD34+ transduction

Human mesenchymal stem cells (hMSC)-nonhematopoietic cells within the bone marrow microenvironment that can be culture expanded to a uniform population of fibroblastic cells-have been shown to support long-term hematopoiesis of CD34+ cells. Because direct contact between stromal elements and CD34+ cells enhances long-term engraftment, we postulated that hMSC would be a good alternative to the more heterogeneous stroma currently used in gene transfer studies. We used hMSC to support retroviral gene transfer of the G156A MGMT (deltaMGMT) gene encoding an alkyltransferase (AGT), which confers drug resistance to a combination of O6-benzylguanine (BG) plus the alkylating agents BCNU and temozolomide (TMZ) in human hematopoietic progenitors. In the presence of IL-3, IL-6, SCF, or leukemia inhibitory factor (LIF) and Flt-3 ligand, hMSC facilitated expansion and retroviral transduction of human peripheral blood-mobilized CD34+ cells. Furthermore, the transduced cells expressed AGT in 29% of hematopoietic cells and were 5-fold more resistant to BCNU and TMZ than were untransduced cells. Unirradiated hMSC present as support cells were simultaneously transduced and expressed AGT in 26% of the cells. Thus, the homogeneous nature of hMSC, and their ability to support gene transfer and be transduced themselves suggest they may be useful in clinical gene transfer protocols and have broad therapeutic applications.

Human long-term culture initiating cells are sensitive to benzylguanine and 1,3-bis(2-chloroethyl)-1-nitrosourea and protected after mutant (G156A) methylguanine methyltransferase gene transfer

Human hematopoietic progenitors express low levels of O6-alkylguanine-DNA alkyltransferase and are sensitive to 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU), particularly following O6-benzylguanine (BG)-mediated O6-alkylguanine-DNA alkyltransferase inhibition. Expression of the BG-resistant mutant (G156A) methylguanine methyltransferase (deltaMGMT) gene in hematopoietic cells confers resistance to BG and BCNU. Because BCNU targets both early and late human hematopoietic cells and results in prolonged and cumulative myelosuppression, we attempted to protect early hematopoietic progenitors (long-term culture initiating cells (LTC-ICs)) by retroviral-mediated transfer of the deltaMGMTgene. A total of 33-56% of LTC-ICs were transduced with MFG-deltaMGMT retrovirus as determined by evidence of provirus in secondary colony-forming units at 5 weeks of culture under conditions optimal for the survival and proliferation of early hematopoietic progenitors. The addition of flt-3 ligand to cultures increased the transduction rate of LTC-ICs. Furthermore, 17.8 +/- 8.1% of deltaMGMT-transduced LTC-ICs survived doses of BG and BCNU; these doses allowed the survival of only 0-1% of untransduced LTC-ICs. This finding compares favorably with the 8-12% of CD34+ cell-derived colony-forming units that we previously showed became resistant to BG and BCNU after deltaMGMTgene transfer. Thus, deltaMGMT transduction of human early hematopoietic progenitor LTC-ICs confers resistance to BG and BCNU and may allow transduced LTC-ICs selective survival and enrichment over untransduced cells in patients undergoing BG and BCNU chemotherapy.

DeltaMGMT-transduced bone marrow infusion increases tolerance to O6-benzylguanine and 1,3-bis(2-chloroethyl)-1-nitrosourea and allows intensive therapy of 1,3-bis(2-chloroethyl)-1-nitrosourea-resistant human colon cancer xenografts

O6-Benzylguanine (BG) is a potent inhibitor of the DNA repair protein 06-alkylguanine DNA alkyltransferase (AGT), and sensitizes tumors to BCNU in vitro and in xenografts. The combination of BG and BCNU is now undergoing phase I clinical testing. The maximally tolerated dose of BCNU given after BG is expected to be lower then the doses tolerated as a single agent owing to BG sensitization of hematopoietic progenitors. We have previously shown that retroviral expression of G156A mutant MGMT (deltaMGMT) in mouse and human marrow cells results in significant BG and BCNU resistance. In this study we evaluated the effect of deltaMGMT-transduced marrow infusion on the therapeutic index of multiple BG and BCNU treatments in tumor-bearing nude (nu/nu athymic) mice. Prior to subcutaneous implantation of BCNU-resistant SW480 human colon cancer cells, cohorts of mice were given intraperitoneal injections of nonablative doses of BG (30 mg/kg) and BCNU (10 mg/kg, one-half of the LD10) and then infused with 1-2 x 10(6) isogeneic deltaMGMT (n = 29 mice) or lacZ-transduced (n = 20 mice) marrow cells. The xenograft-bearing mice were treated with multiple cycles of BG (30 mg/kg) and BCNU (10-25 mg/kg). After three cycles, deltaMGMT mouse bone marrow was repopulated with CFU containing the provirus, and demonstrated a 2.7-fold increase in AGT activity and a 5.5-fold increase in BCNU IC90 compared with LacZ mice. After five cycles, the BCNU IC90 of CFU cells increased nine-fold over control cells, indicating selective enrichment of CFU precursor cells expressing high levels of deltaMGMT. Starting with the third cycle of therapy, tolerance to BG and BCNU was significantly improved in deltaMGMT mice compared with LacZ mice, as evidenced by preserved peripheral blood counts, bone marrow cellularity, and CFU content 1 and 2 weeks posttreatment and a significantly higher survival rate. Xenograft growth was significantly delayed in mice tolerating multiple cycles and higher dose intensity of BG and BCNU as compared with mice receiving less intensive therapy. We conclude that deltaMGMT-transduced marrow cells can improve the therapeutic index of BG and BCNU by selectively repopulating the marrow and providing significant marrow tolerance to this combination, allowing intensive therapy of a BCNU-resistant tumor.

Simultaneous protection of G156A methylguanine DNA methyltransferase gene-transduced hematopoietic progenitors and sensitization of tumor cells using O6-benzylguanine and temozolomide

O6-Benzylguanine (BG) potentiates temozolomide (TMZ) cytotoxicity in tumors by inactivating O6-alkylguanine DNA alkyltransferase but also increases toxicity in hematopoietic cells. To improve the hematopoietic cell tolerance to alkylating agents, we retrovirally transduced the BG-resistant mutant G156A methylguanine DNA methyltransferase gene (deltaMGMT) into hematopoietic progenitors and evaluated whether deltaMGMT expression in hematopoietic colony-forming units would result in greater drug resistance to TMZ. DeltaMGMT expression in human and mouse colony-forming units followed by BG treatment resulted in a >7.7-fold increase in the TMZ 90% inhibitory concentration (IC90) and a 5.6-fold increase in the 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) IC90 relative to untransduced cells. This degree of protection enabled deltaMGMT-transduced CD34 cells to become much more resistant to BG and TMZ than SW480 cells, which express high O6-alkylguanine DNA alkyltransferase and are normally resistant to TMZ or BCNU alone. DeltaMGMT-transduced long-term culture initiating cells were also resistant to the BG and TMZ combination, as were untransduced long-term culture initiating cells, suggesting that noncycling early progenitors may be partially protected from TMZ. These data indicate that retroviral transduction of deltaMGMT into hematopoietic progenitors followed by BG and TMZ treatment may selectively protect hematopoietic cells more efficiently than BCNU, allowing dose-intensive and repetitive therapy without the risk of cumulative myelosuppression.

Selection for G156A O6-methylguanine DNA methyltransferase gene-transduced hematopoietic progenitors and protection from lethality in mice treated with O6-benzylguanine and 1,3-bis(2-chloroethyl)-1-nitrosourea

A retroviral gene therapy approach was developed to protect early hematopoietic progenitors from 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU), a stem cell toxin, and O6-benzylguanine (BG), an inhibitor of a key BCNU resistance protein, O6-alkylguanine DNA alkyltransferase (AGT). The retroviral vector MFG was used to transfer the G156A MGMT (deltaMGMT) cDNA, encoding a mutant AGT that is resistant to inhibition by BG, into murine bone marrow-derived hematopoietic progenitors. Following transplantation into lethally irradiated mice, the transduced cells were subjected to in vivo BG and BCNU treatment to examine the ability to enrich for transduced cells expressing deltaAGT. Transplantation of deltaMGMT-transduced cells resulted in deltaAGT expression in 30% of bone marrow nucleated cells 13 weeks after transplantation. After one cycle of BG and BCNU, deltaAGT expression was observed in 60% of bone marrow cells, and the percentage of colony-forming units (culture; CFU-C) containing proviral sequence increased from 67 to 100%. CFU-C obtained from BG and BCNU-treated deltaMGMT animals up to 23 weeks after transplantation were more resistant to combination BG and BCNU than CFU-C from mice transplanted with lacZ-transduced cells and treated with BG and BCNU or from mice transplanted with deltaMGMT-transduced cells and left untreated. The degree of drug resistance in deltaMGMT-transduced hematopoietic progenitors to BG and BCNU was much greater than we observed previously with wild-type MGMT gene transfer and treatment with BCNU alone. Furthermore, whereas 21 of 22 mice transplanted with deltaMGMT-transduced cells survived in vivo BG and BCNU administration, only 3 of 13 mice transplanted with lacZ-transduced progenitors survived similar drug treatment. Thus, deltaMGMT-transduced murine bone marrow cells selectively survive in vivo BG and BCNU exposure, resulting in prolonged enrichment for the transduced cells and protection from mortality induced by this drug combination.

Axon-target interactions maintain synaptic gene expression in retinae transplanted to intracranial regions of the rat

In the present study we examined the effects of optic axon-CNS target interactions on gene expression in the rat retina. These studies took advantage of a transplantation paradigm that allowed us to assay gene expression in retinae transplanted to different intracranial locations in the neonatal rat that either promoted (dorsal midbrain) or precluded (cerebral cortex) the formation of retino-collicular connections. Using in situ hybridization experiments, we observed that transplantation to the dorsal midbrain resulted in a relatively normal pattern of nicotinic acetylcholine receptor (nAChR) beta-3 subunit and glutamate receptor 3 (GluR3) gene expression. In contrast, retinae transplanted to the cerebral cortex (which did not result in normal retino-collicular interactions) showed a dramatic reduction in nAChR beta-3 subunit and GluR3 gene expression. These results agree with those obtained in the adult goldfish retina, where it has been demonstrated that an optic nerve-optic tectum interaction is responsible for the re-induction nAChR and NMDA receptor gene expression during optic nerve regeneration. Taken together, these results support the hypothesis that proper axon-target interactions are required for maintenance of nAChR and glutamate receptor gene expression in the mature vertebrate retina.

Potentiation of lymphomagenesis by methylnitrosourea in mice transgenic for LMO1 is blocked by O6-alkylguanine DNA-alkyltransferase

We evaluated induction of lymphomas by the methylating carcinogen, N-methylnitrosourea [MNU], in transgenic mice expressing both LMO1 and the DNA repair gene, MGMT, in the thymus. The goal was to determine whether environmental mutagens shorten the latency or increase the incidence of LMO1 + lymphomas and whether mice transgenic for both LMO1 and MGMT, and thereby able to repair O6-methylguanine DNA adducts induced by MNU, would be protected. Mice heterozygous for LMO1 or MGMT were crossed and offspring treated with MNU at 6 weeks of age. MNU induced lymphoma incidence was highest in the LMO1 mice, 91% and lowest in the hMGMT + mice, 15%. MNU induced K-ras mutations in codon 12 in non-MGMT transgenics resulted in a shorter latency of tumors and accounting for half of the early lymphomas in LMO1 mice. The effect of MNU was abrogated in the LMO1/hMGMT transgenic mice, indicating the ability of MGMT expression to block the carcinogenic effect of MNU even in cancer prone mice. Thus, methylating agents potentiate lymphomagenesis of LMO1, in part through activation of K-ras and the MAPK pathway, a process which appear to synergize with LMO1 mediated transcription activation. O6-alkylguanine DNA-alkyltransferase mediated DNA repair effectively blocks chemical carcinogenesis in mice carrying the LMO1 oncogene.

Retroviral transduction of a mutant methylguanine DNA methyltransferase gene into human CD34 cells confers resistance to O6-benzylguanine plus 1,3-bis(2-chloroethyl)-1-nitrosourea

Human CD34 cells express low levels of the DNA repair protein O6-alkylguanine-DNA alkyltransferase (AGT) and are sensitive to 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU). Gene transfer of the AGT gene, methylguanine DNA methyltransferase (MGMT), results in only modest BCNU resistance. Recently, an AGT inhibitor, O6-benzylguanine (BG), entered clinical trials. In preclinical studies, BG potentiated the cytotoxic effect of BCNU in tumors but increased toxicity to normal CD34 cells. We transferred a mutant MGMT containing a glycine-to-alanine mutation at position 156, resulting in marked resistance to BG, into Chinese hamster cells; the K562 cell line and human CD34 cells used the retroviral backbone MFG. In each instance, cells expressed increased AGT and were much more resistant to the combination of BG and BCNU than the parental cells or cells transduced with wild-type MGMT. Furthermore, the transduction efficiency in human CD34 cells was in excess of 70%, and the proportion of CD34 transduced cells resistant to the combination was > 30%. Thus, retroviral-mediated transduction of a mutant MGMT into CD34 cells appears to be an effective way to induce selective resistance to a drug combination designed to overcome a significant resistance mechanism to nitrosoureas in tumors.

Opsin gene expression and regulation in retinal transplants

In the present study we investigated the expression and regulation of the opsin gene in retinal transplants. Embryonic retinae were transplanted to intracranial locations in neonatal rodents in which they either reliably projected to the superior colliculus, or in locations (such as the cerebral cortex) in which they did not project to subcortical visual nuclei. Our results show that, regardless of the graft location, the developmental schedule of opsin gene expression in the outer nuclear layer was similar to normal, and that it was maintained in transplants for at least 6 months. To test if ambient light affected opsin gene expression, we dark-reared rats containing a retinal transplant for up to 26 days before assaying for opsin transcripts. In situ hybridization experiments showed that opsin gene expression in the transplants of these dark-reared recipients was not different either from transplants in animals reared in cyclic light conditions, or from the retina in situ. These observations support the hypothesis that the opsin gene is activated and maintained by molecular mechanisms intrinsic to the photoreceptor.

Transfer of drug resistance genes into hematopoietic progenitors to improve chemotherapy tolerance

A number of drug resistance genes have been identified that may be useful in gene therapy approaches to ameliorate chemotherapy toxicity. Hematopoietic tissue is the most suitable target for drug resistance gene therapy because myelosuppression is the dose-limiting toxicity of the many chemotherapeutic agents. Recent studies have shown that murine and human hematopoietic progenitors can be transduced ex vivo using retroviral vectors to overexpress P-glycoprotein, dihydrofolate reductase, and O6-alkylguanine DNA alkyltransferase. In all instances, gene transfer results in significant drug resistance in hematopoietic progenitors both in vitro and in vivo. Clinical trials are underway to evaluate the role of MDR-1 gene therapy in amelioration of chemotherapy induced myelosuppression. Other genes being examined for their potential to transfer drug resistance to hematopoietic cells include genes encoding aldehyde dehydrogenase, nucleotide excision repair proteins, multidrug resistant protein, and superoxide dismutase. As a group these proteins could confer significant levels of chemotherapy drug resistance to bone marrow cells. When compared with other somatic gene therapy approaches, drug resistance gene therapy has the aim of protecting normal cells and preventing toxicity. In addition many of these genes could be used to select for cells carrying the drug resistance gene as well as cotransduced therapeutic gene. Thus, gene transfer of drug resistance genes will have broad applications in the field of gene therapy as well as in protecting hematopoietic cells from chemotherapy toxicity.