Problems And Directions For The Future

A. Low Titer

The low level of gene transfer into human and large animal HSC can be attributed to a combination of factors. One consistent problem with retrovirus-mediated gene transfer appears to be the relatively low titer of retrovirus preparations. The titer of even the best preparations of retrovirus rarely exceeds 1 million particles per milliliter of conditioned medium, and the movement of retrovirus particles in solution is limited by Brownian motion (222). Thus, most HSC do not come in contact with retrovirus particles under standard culture conditions. To increase interaction between retrovirus particles and HSC, investigators have used centrifugation to move HSC through a large volume of virus containing supernatant (223). Alternatively, the cells may be immobilized on a filter through which large volumes of retrovirus containing supernatant can be drawn (224,225). Both of these procedures have been shown to improve gene transfer into hematopoietic progenitor cells and may ultimately prove useful for the transduction of human HSC.

Another approach to increasing the contact between virus particles and HSC involves the use of fibronectin-coated gene transfer vesicles. It has been demonstrated that retrovirus particles and HSC bind specifically to a 35 kd fragment of fibro-nectin (226,227). The colocalization of the target cell and the retrovirus particle increases the probability that a cell will become transduced. The use of fibronectin has become standard in human gene transfer protocols. Titer can also be increased by concentration of virus particles in supernatant. A variety of protocols has been used to concentrate amphotropic, RD114, and VSV-G pseudotyped virus particles (228,229, 151-153).

B. Selection of Transduced HSC In Vivo

Recognizing the inefficient transduction of human HSC by oncoretroviral or lentiviral vectors, several groups have attempted to transfer drug resistance genes into HSC. These studies have a short-term goal of rendering bone marrow cells resistant to chemotherapeutic agents active against solid tumors, and a long-term goal of providing a means to select and amplify the rare transduced HSC. If the virus vector contained a therapeutic transgene along with the drug resistance gene, the result would be a high level of cells containing the therapeutic gene. This concept has been shown to be feasible in mouse models using the DHFR gene. When methotrexate selection is applied in the presence of nucleoside transport inhibitor nitrobenzylmercapt-purine riboside 5' monophosphate (NBMPR-P), DHFR-expressing HSC demonstrate preferential survival (230). Similar results were observed when the human ENT-2 gene was introduced into mouse HSC followed by a similar selection protocol (231). Other strategies have involved the introduction of the MGMT gene followed by treatment of the animals with temozolomide, O6-benzylgua-nine, or similar drugs. As in the case of DHFR, transduced cells expressing MGMT were resistant to the effects of the drug treatment (232-234).

Alternatively, other groups have attempted to induce HSC self-renewal by expressing either chimeric receptor molecules that can be dimerized by small molecules, thus transducing a signal, or by expressing growth-factor receptors that ordinarily are not expressed by HSC. Treatment with either the small molecule or the cytokine triggers a ''growth switch'' that selectively amplified transduced HSC. This approach has been demonstrated in both mouse and nonhuman primate models (235-238).

C. Low Levels of Retrovirus Receptors

Many groups have observed that while the transduction of progenitor cells with amphotropic oncoretroviral vectors is relatively efficient, the transduction of HSC in the same population of cells is inefficient. These observations conflict with data in the mouse models with ecotropic oncoretrovirus particles, which can transduce both progenitors and HSC efficiently. One hypothesis proposed that ecotropic receptor levels were relatively high on mouse HSC, but that amphotropic retrovirus receptor levels were low or absent on mouse, primate, and human HSC.

Using a semiquantitative RT-PCR assay, Orlic et al. showed that the mRNA encoding the mouse ecotropic retrovi-rus receptor (mCAT) was easily detectable in mouse HSC, progenitor, and bone marrow cells. The mRNA encoding the amphotropic receptor (amphoR) was easily detectable in human or mouse hematopoietic cell lines and primary hematopoietic progenitor cells. However, amphoR mRNA was nearly undetectable in mRNA isolated from mouse and human HSC (239). Further investigation identified and isolated a rare subpopulation of mouse HSC that had higher levels of amphoR mRNA. To correlate receptor mRNA levels with transduction, amphoR positive and negative HSC were simultaneously transduced with genetically distinguishable oncoretroviruses, one packaged by ecotropic packaging cells, the other by am-photropic packaging cells. Greater than 90% of the animals in these experiments contained ecotropic proviral sequences in their hematopoietic cells, indicating that there was sufficient cell division to allow the integration of oncoretroviral vectors into HSC. More than 50% of mice repopulated with amphoRpositive HSC also contained amphotropic proviral sequences, while less than 10% of mice repopulated with amphoR-nega-tive HSC contained amphotropic proviral sequences. In animals repopulated with amphoR-negative HSC, the relative number of ecotropic proviral sequences was 25-fold higher than the level of amphotropic proviral sequences, similar to the relative difference in the levels of ecoR and amphoR mRNA in HSC (239). Related work has demonstrated a low level of GALV receptor mRNA in human HSC as well (240). It would appear that a major obstacle to gene therapy for human hematopoietic diseases is the low level of retrovirus receptor on the surface of HSC (Fig. 5).

The use of VSV-G pseudotyped retroviruses may eventually circumvent problems with low titer and low numbers of receptors. As noted previously, VSV-G particles can be efficiently concentrated to high titers without loss of activity (151,152). Since the VSV-G receptor is a membrane phosholi-pid (150), it would appear that there should be fewer problems with low receptor numbers. Problems associated with the toxicity of VSV-G env have slowed the development of VSV-G mediated gene therapy (149), but recent work with concentrated and purified VSV-G vectors has shown great promise in model systems (152,177,178). Other oncoretroviruses have been isolated from cats and sheep with novel envelopes that may recognize more abundant receptors on human HSC. For example, oncoretroviruses pseudotyped with the feline RD114 virus envelope have been shown to give high levels of gene transfer in canine models (241).

D. Quiescent HSC

Because Moloney Murine Leukemia Virus (MMuLV)-based retroviral vectors require cell division to integrate into the host cell genome, another obstacle to retrovirus-mediated gene transfer is the fact that most HSC are quiescent. Although relatively efficient transduction of mouse HSC can be achieved with ecotropic MMuLV vectors, it is not clear that human HSC are as prone to enter the cell cycle during the time they are exposed to the virus. One approach to this prob lem has been to expose human HSC to retroviral vectors under culture conditions that would lower the levels of cell cycle inhibitors. A recent study showed that the level of the cell cycle inhibitor p15(INK)4B could be decreased in human CD34+ progenitor cells by culture in serum-free medium containing antibodies against TGF-p. In addition, antisense oligonucleotides inhibited the expression of a second cell cycle inhibitor P27(KIP-1) in CD34 + cells and promoted cell cycling (242). Human CD34+ CD38- HSC were cultured in serum-free medium containing SCF, Il-6, IL-3, FL, anti TGF-p antibodies and antisense p27(KIP-1) oligonucleotides for 12 h and exposed to retrovirus vectors containing a neo gene for 12 h. Parallel cultures of cells were cultured in the same cytokines minus the anti TGF-p antibodies and antisense p27(KIP-1) oligonucleotides for 12 h and then exposed to the same virus. Cells from both conditions were equally capable of engrafting and proliferating in the BNX mouse model. One year after transplantation, no neo transgenes were observed in the human hematopoietic cells derived from cells not exposed to the anti TGF-p antibodies and antisense p27(KIP-1) oligonucleotides. In contrast the neo transgene was detected in about 10% of cells derived from CD34+ CD38- HSC exposed to anti TGF-p antibodies and antisense p27(KIP-1) oligonucleotides (242). These results clearly show that cell-cycle induction can promote retrovirus-mediated gene transfer into HSC without damaging the ability of the HSC to repopu-late BNX mice.

The gene transfer problems relating to cell cycle may ultimately be overcome by the use of lentivirus-based vectors,

Figure 5 Comparison of transduction by ecotropic and amphotropic retrovirus particles. In the mouse (top), the high level of ecotropic retrovirus receptor allows many cells to bind a virus. Subsequent cell division allows these viruses to integrate into the genome. Amphotropic transduction of mouse primate or human HSC (bottom) is limited by low levels of amphotropic receptor expression on HSC. There are fewer opportunities for virus to bind to cells and therefore any cell division that occurs is unlikely to result in a viral integration.

Figure 5 Comparison of transduction by ecotropic and amphotropic retrovirus particles. In the mouse (top), the high level of ecotropic retrovirus receptor allows many cells to bind a virus. Subsequent cell division allows these viruses to integrate into the genome. Amphotropic transduction of mouse primate or human HSC (bottom) is limited by low levels of amphotropic receptor expression on HSC. There are fewer opportunities for virus to bind to cells and therefore any cell division that occurs is unlikely to result in a viral integration.

which do not require cell division for integration. VSV-G pseudotyped lentivirus vectors have been shown to be greater than 20-fold more efficient at transducing freshly isolated human Lin- CD34+ CD38- HSC than MMuLV vectors (243). In addition, the expression level of the GFP reporter gene in lentivirus-transduced cells persisted for 5 weeks in culture, while expression of GFP from the MMuLV vector was lower and rapidly was silenced (243). Although the initial transductions of nonhuman primate and mouse HSC with lentivirus vectors have been similar to results with oncoretrovirus vectors, lentivirus mediated gene transfer still has theoretical advantages over oncoretrovirus-mediated gene transfer.

E. Summary

Improving the efficiency of gene transfer and the expression of the transferred genes are the critical basic research goals that will extend gene therapy from possibility to reality. Basic research into the biology of the PHSC has demonstrated low levels of retrovirus receptors and cell cycling as well as solutions to these problems. Basic research into retrovirus envelope proteins and backbones have lead to new recombinant retrovirus vectors that are capable of transferring and expressing genes at higher rates. As clinical experience is acquired in the use of retrovirus vectors for marking human HSC, it is clear that therapeutic gene transfer to human HSC could soon be a practical method for the treatment of ADA deficiency and other immune disorders. I predict that success in treating these diseases by gene transfer to HSC would be rapidly expanded to the treatment of a wide variety of other inherited and acquired hematological diseases.

Was this article helpful?

0 0

Post a comment