Using gene transfer to circumvent off-target effects.

Many recombinant growth factors have failed in clinical trials due to off-target effects. We describe a method for circumventing off-target effects that involves equipping cells with a conditionally active signaling protein that can be specifically activated by an exogenously administered synthetic ligand. We believe that this approach will have many applications in gene and cell therapy.

Membrane localization is not required for Mpl function in normal hematopoietic cells.

Cellular trafficking of growth factor receptors, including cross-talk among receptors at the cell surface, may be important for signal transduction in normal hematopoietic cells. To test this idea, the signaling domain of Mpl (the thrombopoietin receptor) was targeted to the plasma membrane, or to the cytoplasm of murine marrow cells, and the ability of the cells to proliferate and differentiate in response to Mpl dimerized at the plasma membrane or free in the cytoplasm was assessed. Constructs encoding the signaling domain of Mpl linked to an FK506 binding protein domain (to permit dimerization by the membrane-permeable ligand AP20187) with or without a myristylation sequence (to target the receptor to the plasma membrane) and a hemagglutinin epitope tag were generated and introduced into murine marrow cells using a murine stem cell virus (MSCV)-based retroviral vector. Both populations of transduced marrow cells proliferated in Iscoves modified Dulbecco medium-10% FCS-100 nM AP20187 without exogenous growth factors for more than 100 days and achieved greater than a 10(7)-fold expansion of cells by day 50 (n = 4 transductions). Growth was dimerizer dependent, and myeloid, erythroid, and megakaryocytic progenitors were generated. Activation of Mpl either at the plasma membrane or in the cytoplasm allowed for the terminal maturation of transduced progenitor cells. Introduction of membrane-targeted or cytoplasmic Mpl into fetal liver cells from homozygous JAK2 knock-out mice or wild-type littermates demonstrated that both forms of Mpl require JAK2 for signaling. These data show that the activation of Mpl independent of its normal plasma membrane location can support production of the full range of normal hematopoietic progenitor cells in vitro.

Control of myoblast proliferation with a synthetic ligand.

Skeletal myoblast grafts can form contractile tissue to replace scar and repair injured myocardium. Although potentially therapeutic, generating reproducible and sufficiently large grafts remains a challenge. To control myoblast proliferation in situ, we created a chimeric receptor composed of a modified FK506-binding protein (F36V) fused with the fibroblast growth factor receptor-1 cytoplasmic domain. Mouse MM14 myoblasts were transfected with this construct and treated with AP20187, a dimeric F36V ligand, to induce receptor dimerization. Transfected myoblasts proliferated in response to dimerizer (comparable with basic fibroblast growth factor (bFGF) treatment), whereas the dimerizer had no effect on non-transfected cells. Similar to bFGF treatment, dimerizer treatment blocked myotube formation and myosin heavy chain expression and stimulated mitogen-activated protein (MAP) kinase phosphorylation in transfected cells. Non-transfected cells differentiated normally and showed no MAP kinase phosphorylation with dimerizer treatment. Furthermore, myoblasts treated with dimerizer for 30 days in culture reduced MAP kinase phosphorylation, withdrew from the cell cycle, and differentiated normally upon drug withdrawal, demonstrating reversibility of the effect. Thus, forced dimerization of the fibroblast growth factor receptor-1 cytoplasmic domain reproduces critical aspects of bFGF signaling in myoblasts. We hypothesize that in vivo administration of AP20187 following myoblast grafting may allow control over graft size and ultimately improve cardiac function.

Receptor specificity in the self-renewal and differentiation of primary multipotential hemopoietic cells.

To determine whether cytokine-induced signals generate unique responses in multipotential hemopoietic progenitor cells, the signaling domains of 3 different growth factor receptors (Mpl, granulocyte-colony-stimulating factor [G-CSF] receptor, and Flt-3) were inserted into mouse primary bone marrow cells. To circumvent the activation of endogenous receptors, each signaling domain was incorporated into an FK506 binding protein (FKBP) fusion to allow for its specific activation using synthetic FKBP ligands. Each signaling domain supported the growth of Ba/F3 cells; however, only Mpl supported the sustained growth of transduced marrow cells, with a dramatic expansion of multipotential progenitors and megakaryocytes. These findings demonstrate that the self-renewal and differentiation of multipotential progenitor cells can be influenced through distinct, receptor-initiated signaling pathways.

Cell proliferation through forced engagement of c-Kit and Flt-3.

To investigate the potential for functional interactions between heterologous receptors, the cytoplasmic domains of 2 different receptors (c-Kit and Flt-3) were coexpressed in the interleukin-3-dependent cell line Ba/F3. The receptor signaling domains were presented in the context of fusion proteins, with c-Kit linked to the FK506 binding protein (FKBP12) and Flt-3 linked to the FRB domain of the FKBP12-rapamycin-associated protein. The fusions were brought into apposition with the use of chemical inducers of dimerization (CIDs). Two classes of CID were employed. FK1012 and its synthetic analogue AP1510 bring together 2 copies of the FKBP12 domain, thereby inducing homodimerization of the c-Kit(FKBP12) fusion. A second type of CID, rapamycin, brings together one FKBP12 domain and one FRB domain, resulting in heterodimerization of the c-Kit(FKBP12) and Flt-3(FRB) fusions. Ba/F3 cell growth was promoted not only by FK1012- or AP1510-induced homodimerization of the c-Kit(FKBP12) fusion (as reported previously), but also by rapamycin-induced c-Kit(FKBP12)-Flt-3(FRB) heterodimerization. These findings demonstrate the potential for a direct functional interaction between c-Kit and Flt-3. (Blood. 2001;97:3662-3664)

In vivo selection using a cell-growth switch.

A major obstacle to stem-cell gene therapy rests in the inability to deliver a gene into a therapeutically relevant fraction of stem cells. One way to circumvent this obstacle is to use selection. Vectors containing two linked genes serve as the basis for selection, with one gene encoding a selectable product and the other, a therapeutic protein. Applying selection in vivo has the potential to bring a minor population of genetically corrected cells into the therapeutic range. But strategies for achieving in vivo selection have traditionally relied on genes that confer resistance to cytotoxic drugs and are encumbered by toxicity. Here we describe a new system for in vivo selection that uses a 'cell-growth switch', allowing a minor population of genetically corrected cells into the therapeutic range. But strategies for achieving in vivo selection have traditionally relied on genes that confer resistance to cytotoxic drugs and are encumbered by toxicity. Here we describe a new system for in vivo selection that uses a 'cell-growth switch', allowing a minor population of genetically modified cells to be inducibly amplified, thereby averting the risks associated with cytotoxic drugs. This system provides a general platform for conditionally expanding genetically modified cell populations in vivo, and may have widespread applications in gene and cell therapy.

Expansion of genetically modified primary human hemopoietic cells using chemical inducers of dimerization.

The inability to deliver a therapeutic gene to a sufficient percentage of hematopoietic stem cells is the major obstacle to using gene therapy to treat blood disorders. Providing genetically corrected stem cells with a reversible growth advantage could solve this problem. To this end we have employed small synthetic molecules that can reversibly dimerize and activate fusion proteins which contain a growth factor receptor signaling domain. We have shown that the thrombopoietin receptor (mpl) signaling domain can be used in this system to expand transduced multipotential progenitor cells from mouse bone marrow. In the present study we tested a similar retroviral vector in human CD34-selected cord blood cells. Following transduction, cells cultured in the presence of the dimerizing molecule AP1903 expanded 13.8- to 186-fold relative to cells cultured in the absence of AP1903. The cell type that emerged in suspension culture was erythroid. Contrary to our results in the murine system, cell expansion was transient. Activation of mpl caused the disappearance of BFU-E followed by a transient increase in CFU-E. In contrast, mpl activation had no discernable effect on transduced myeloid progenitor cells. AP1903-mediated expansion was restricted to transduced cells, as demonstrated by immunohistochemical staining. These findings indicate that synthetic dimerizing molecules can be used to expand primary human hematopoietic cells. (Blood. 2000;95:430-436)

Marrow sensitization to 5-fluorouracil using the ligands for Flt-3 and c-Kit.

In vivo administration of c-kit ligand (KL) expands early hemopoietic progenitors and stem cells and sensitizes clonogenic progenitors to 5-FU-mediated cell death. Studies were performed to determine whether the in vivo administration of Flk-2/Flt-3 ligand (FL) is also capable of sensitizing progenitors to 5-FU. Mice were treated with FL (100 micrograms/kg every 12 hours for a total of 3 doses), KL (50 micrograms/kg, same schedule) or both, either alone or in combination with 5-FU (a single 125 mg/kg injection 3 hours before the last dose of cytokine). Femurs and spleens were harvested 48 hours following the last dose of cytokine, and the total numbers of mononuclear cells and colony forming unit cells (CFU-C) per femur and spleen were determined. Statistically significant increases in the number of CFU-C per femur were observed in response to FL, KL and FL + KL. In the spleen, statistically significant increases in CFU-C were observed only with the FL + KL combination. 5-FU alone produced marked reductions in CFU-C both in the femur and in the spleen. In the femur, 5-FU-mediated reductions in CFU-C were enhanced 3- to 30-fold in the presence of concomitant KL, FL or KL + FL administration. Surprisingly, the combination of KL + FL was no more effective in sensitizing marrow CFU-C to 5-FU than was KL alone, suggesting that CFU-C that are capable of surviving the KL/5-FU combination cannot be driven into cell cycle by FL. The effects of concomitant cytokine/5-FU administration in the spleen contrasted sharply with those observed in the femur, as FL, KL and FL + KL all failed to enhance 5-FU-mediated reductions in CFU-C. The ability of FL + KL to stimulate CFU-C expansion in the spleen combined with the inability of this cytokine combination to augment 5-FU-mediated progenitor toxicity in the spleen supports the contention that cytokine-mobilized progenitors are not in cycle. FL's capacity to specifically sensitize marrow to the effects of cytotoxic drugs may have applications in bone marrow transplant conditioning regimens.

Multiple epsilon-promoter elements participate in the developmental control of epsilon-globin genes in transgenic mice.

To delineate the regulation of the human epsilon-globin gene, we investigated epsilon-gene expression during the development of transgenic mice carrying constructs with epsilon-promoter truncations linked to a micro-locus control region (microLCR). Expression levels were compared with those of microLCR epsilon mice carrying a 2 kilobase epsilon-promoter and betaYAC controls. epsilon mRNA in the embryonic cells of microLCR (-179)epsilon mice were as high as in microLCR epsilon mice suggesting that the proximal epsilon-promoter contains most elements required for epsilon-gene activation. epsilon mRNA in adult microLCR (-179) epsilon mice was significantly lower than in the embryonic cells indicating that elements involved in epsilon-gene silencing are contained in the proximal epsilon-promoter. Extension of the promoter sequence to -463 epsilon decreased epsilon-gene expression in the definitive erythroid cells, supporting previous evidence that the -179 to -463epsilon region contains an epsilon-gene silencer. However, the epsilon-gene of the microLCR(-463)epsilon mice was not silenced in the definitive cells of fetal and adult erythropoiesis indicating that additional silencing elements are located upstream of position -463epsilon. These results provide in vivo evidence that multiple elements of the distal as well as the proximal promoter contribute to epsilon-gene silencing.

Targeted expansion of genetically modified bone marrow cells.

The ability to specifically target a mitogenic signal to a population of genetically modified primary cells would have potential applications both for gene and cell therapy. Toward this end, a gene encoding a fusion protein containing the FK506-binding protein FKBP12, fused to the intracellular portion of the receptor for thrombopoietin (mpl), was introduced into primary murine bone marrow cells. Dimerization of this fusion protein through the addition of a dimeric form of the drug FK506, called FK1012, resulted in a marked proliferative expansion of marrow cells that was restricted to the genetically modified population. FK1012's proliferative effect was sustained and reversible. An apparent preference for differentiation along the megakaryocytic lineage was observed. This approach allows for the specific delivery of a mitogenic signal to a population of genetically modified primary cells and may have applications for studies in hematopoiesis and receptor biology, and for gene and cell therapy.

Stimulating cell proliferation through the pharmacologic activation of c-kit.

Previous studies have shown that expression of a membrane targeted chimeric protein containing the erythropoietin receptor (EpoR) cytoplasmic domain fused to the FK506-binding peptide FKBP12 allowed Ba/F3 cells to be rescued from interleukin-3 (IL-3) deprivation using a dimeric form of FK506, called FK1012. In this report, a similar approach is applied to the c-kit receptor. Expression of a membrane targeted fusion protein containing the c-kit receptor linked to one or more copies of FKBP12 allowed Ba/F3 cells to be switched from IL-3 dependence to FK1012-dependence. Similar results were obtained using an alternative dimerizer of FKBP12 domains called AP1510. Pharmacologic dimerization of chimeric proteins containing only a single FKBP12 domain confirmed that receptor dimerization is sufficient for proliferative signaling. Interestingly, while the proliferative effects of both FK1012 and AP1510 were reversible, FK1012-driven proliferation persisted for several days after drug withdrawal. Furthermore, much higher concentrations of FK506 were required to inhibit FK1012-mediated proliferation than were required to inhibit AP1510-mediated proliferation. The persistence of FK1012's effect appeared to be specific to clones expressing c-kit-containing fusion proteins. These results suggest that pharmacologically-responsive fusion proteins containing c-kit may be useful for specifically and reversibly expanding genetically modified hematopoietic cell populations.

Current status of stem cell therapy and prospects for gene therapy for the disorders of globin synthesis.

Sickle cell anaemia and beta-thalassaemia are today curable through the use of stem cell transplantation. Nevertheless, the disadvantages inherent in stem cell transplantation underscore the need for better therapies. A recent finding of potentially major importance is that complete eradication of host haematopoiesis is not an absolute requirement for achieving therapeutic effects in thalassaemia and sickle cell anaemia. Future stem cell transplantation protocols will use less toxic conditioning regimens in an effort to achieve a state of stable mixed chimerism between donor and host haematopoietic elements. An improved understanding of globin gene regulation and stem cell biology will allow for the first gene therapy trials for sickle cell anaemia and beta-thalassaemia in the relatively near future. Initial gene therapy protocols will emphasize safety, are likely to target progenitor cells, and will involve repeated cycles of mobilization, transduction and reinfusion, with little or no conditioning. These first generation gene therapy trials are unlikely to confer major therapeutic benefits, but will provide the foundation upon which subsequent, more effective protocols will be based.

A proliferation switch for genetically modified cells.

Receptor dimerization is the key signaling event for many cytokines, including erythropoietin. A system has been recently developed that permits intracellular protein dimerization to be reversibly activated in response to a lipid-soluble dimeric form of the drug FK506, called FK1012. FK1012 is used as a pharmacological mediator of dimerization to bring together FK506 binding domains, taken from the endogenous protein FKBP12. In experiments reported herein, FK1012-induced dimerization of a fusion protein containing the intracellular portion of the erythropoietin receptor allowed cells normally dependent on interleukin 3 to proliferate in its absence. FK506 competitively reversed the proliferative effect of FK1012 but had no influence on the proliferative effect of interleukin 3. Signaling pathways activated by FK1012 mimicked those activated by erythropoietin, because both JAK2 and STAT5 were phosphorylated in response to FK1012. This approach may provide a means to specifically and reversibly stimulate the proliferation of genetically modified cell populations in vitro or in vivo.

Cytokine prestimulation as a gene therapy strategy: implications for using the MDR1 gene as a dominant selectable marker.

A major obstacle to stem cell gene therapy is the extremely low efficiency of stem cell transduction. In vivo selection is a strategy for enriching a minor population of genetically modified bone marrow cells through the introduction of a drug resistance gene, followed by subsequent administration of the corresponding cytotoxic drug in vivo. Achieving persistent effects from in vivo selection is expected to require selection at the level of stem cells or, minimally, selection at the level of progenitors. Major limitations to in vivo selection are the nonhematologic toxicities of the cytotoxic drugs used and the resistance of stem cells and progenitors to killing by most cytotoxic agents. Experiments were performed in mice to evaluate whether the drugs used for selection in combination with multiple drug resistance gene 1 (MDR1) could have an enhanced effect on clonogenic progenitors if preceded by administration of the cytokine, stem cell factor (SCF). Single doses of taxol, navelbine, or vinblastine produced 10-fold reductions in the total number of mononuclear cells per femur, indicating a significant depletion of nonclonogenic precursor cells. However, for each of these agents, clonogenic progenitors, assayed as colony-forming unit cells and day-12 spleen colony-forming units, were relatively spared. Administration of SCF before taxol, navelbine, or vinblastine completely abrogated the progenitor-sparing phenomenon, because clonogenic progenitors were depleted as effectively as nonclonogenic precursor cells. Furthermore, the administration of SCF before drug administration allowed the dosages of taxol and vinblastine to be reduced by more than half, while retaining reductions in progenitor numbers that were unachievable using very high doses of the cytotoxic drug alone. Doxorubicin administration resulted in a 30- to 40-fold depletion in progenitors that was not significantly altered by preceding SCF administration. These results suggest that previous observations of in vivo selection using MDR1 gene transfer followed by taxol administration may have resulted from selection at the level of relatively mature, nonclonogenic precursor cells. Furthermore, these data suggest that cytokine prestimulation may be a useful strategy for improving the selection of drug-resistant clonogenic progenitors and, possibly, stem cells in vivo.

The hematological effects of folate analogs: implications for using the dihydrofolate reductase gene for in vivo selection.

Previous studies have shown that dihydrofolate reductase (DHFR) gene transfer protects marrow from methotrexate (MTX)-mediated toxicity; however, MTX treatment in vivo has not convincingly been shown to enrich DHFR-transduced progenitors or stem cells. Experiments were performed to better characterize the hematological effects of MTX, and maneuvers were tested with the aim of improving the utility of MTX as an agent for in vivo selection. Progenitors were assayed as colony forming unit cells in culture (CFU-C) and in the spleens of irradiated mice (day 11 CFU-S). A single injection of MTX at doses up to 250 mg/kg (more than three times the LD10) failed to reduce CFU-C numbers significantly in the femur or spleen assayed 1-3 days later. However, consistent declines in the number of mononuclear cells per femur reflected a significant depletion of nonclonogenic precursor cells. Preceding administration of pegylated stem cell factor (SCF), 100 micrograms/kg per day, increased CFU-C killing by a single dose of 5-fluorouracil (5-FU) 15- to 65-fold in the femur, and 5- to 15-fold in the spleen, consistent with previous reports. In contrast, despite preceding SCF administration there was no significant progenitor killing by MTX. Similar results were obtained using a second folate analog, trimetrexate. These results suggest that the mechanism by which folate analogs exert their hematological toxicity is through the depletion of relatively mature, nonclonogenic precursor cells, and not by killing progenitors. This information is relevant to the use of DHFR in gene therapy protocols, and suggests that folate analogs are poorly suited agents for selection at the level of clonogenic progenitor cells in vivo.

Forced expression of cytidine deaminase confers resistance to cytosine arabinoside and gemcitabine.

Genes that confer cellular resistance to cytotoxic agents have potential applications both for marrow protection following cancer chemotherapy and as dominant selectable markers for the selection of genetically modified hematopoietic cell populations in vivo. Cytidine analogues, including cytosine arabinoside (Ara-C), 2',2'-difluorodeoxycytidine (gemcitabine), and other drugs represent a clinically important class of chemotherapeutic agents for which no drug resistance gene has yet been described. Studies were performed to determine whether forced expression of a gene encoding the enzyme cytidine deaminase can confer resistance to cytosine arabinoside (Ara-C) and gemcitabine in vitro. A pooled population of NIH3T3 cells overexpressing cytidine deaminase from a retroviral construct based on the LXSN-vector (LCDSN) demonstrated a 4.5-fold increased resistance to Ara-C based on the 50% inhibitory concentration (ID50) and a 3.7-fold increased resistance to Ara-C based on the 80% inhibitory concentration (ID80) relative to cells transduced with a control vector. In the hemopoietic cell line CCRF-CEM, the same retroviral construct conferred a 2.1-fold increased resistance to Ara-C by ID50 and a 3.0-fold increased resistance to Ara-C by ID80 relative to cells transduced with a control vector. CEM cells transduced with LCDSN were also resistant to gemcitabine (2.4-fold by ID50 and 2.5-fold by ID80). Furthermore, Ara-C-resistance could be completely reversed in LCDSN-transduced CEM cells by a specific inhibitor of cytidine deaminase, tetrahydrouridine (THU). Expression of the transgene was confirmed by RNase-protection assay and by an enzyme activity assay. These results provide the first direct evidence that forced expression of cytidine deaminase confers cellular resistance to Ara-C and gemcitabine.