Bicistronic Vectors Containing Selectable Markers

Although coexpression of two proteins can be achieved through the use of separate promoters, the coexpression is frequently uncoupled due to promoter interference or shutoff of gene expression from one of the promoters, which causes the selected cells not to express the desired protein. To overcome this problem, the selectable marker may be expressed with the therapeutic gene as a translational or transcriptional fusion. A therapeutic protein can be directly linked to the carboxylterminus of the multidrug transporter P-glycoprotein (P-gp). The resulting fusion protein possesses functions of both P-gp and the target protein (125). Since P-gp is an integral membrane protein that functions on the cell plasma membrane, unless two proteins can be separated by a posttransla-tional proteolytic modification, the expressed target protein will be associated with the plasma membrane regardless of its normal cellular location. Thus, even though translational fusions guarantee protein coexpression, their potential is limited. On the other hand, transcriptional fusions, e.g., using bicistronic or polycistronic mRNA to encode more than one cDNA, may prove to be more generally applicable.

A. MDR1 Bicistronic Vectors Containing Internal Ribosome Entry Sites (IRES)

A DNA segment corresponding to one polypeptide chain plus the translational start and stop signals for protein synthesis can be loosely defined as a cistron. An mRNA encoding only a single polypeptide is called monocistronic mRNA; if it encodes two or more polypeptide chains, it may be called bicistronic or polycistronic mRNA. Almost all eukaryotic mRNA molecules are monocistronic. Initiation of translation of eukar-yotic mRNA is mediated by cap-binding protein that recognizes a methylated guanosine cap at the 5' terminus of mRNA. However, some viral mRNA molecules transcribed in eukaryotic cells are polycistronic. They can use a cap-independent mechanism to initiate translation in the middle of mRNA molecules. For picornavirus, this cap-independent internal initiation of translation is mediated through a unique internal ribosome entry site (IRES) within the mRNA molecule (126,127).

Identification of IRES sequences has led to the development of bicistronic vectors that allow coexpression of two different polypeptides from a single mRNA molecule in euk-aryotic cells (128,129). Using a bicistronic vector containing an IRES to coexpress a target gene and a selectable marker has several advantages. First, since two polypeptides are translated from the same mRNA molecule, the bicistronic vector guarantees coexpression of a selectable marker and a second protein. Secondly, bicistronic mRNA allows two polypeptides to be translated separately. Thus, this system does not compromise the correct intracellular trafficking of proteins directed to different subcellular compartments. In addition, using a bicistronic vector, expression of a target gene is proportionate to the expression of a selectable marker. Hence, expression of a target protein can be achieved quantitatively by applying selections of different stringencies.

To demonstrate coexpression of a dominant selectable marker with a therapeutic gene using a bicistronic vector, our laboratory has coexpressed P-gp with glucocerebrosidase (112,113), p-galactosidase (115), adenosine deaminase (118), a subunit of the NAPH oxidase complex (116,117), the shared gamma chain of the interleukin receptors (119), and a hammerhead ribozyme targeted to the U5 region of HIV-1 LTR (130). In those experiments, MDR1 served as a selectable marker linked to the target gene by an IRES from encephalo-myocarditis virus (EMCV) and constructed in a retroviral vector containing Harvey sarcoma virus LTR (131). Two configurations, in which MDR1 is placed either before or after the IRES, have been examined in some cases. As demonstrated in those experiments, P-gp and the target gene are coexpressed in the cells selected using cytotoxic P-gp substrates, such as colchicine or vincristine; the expressed target proteins are functional as detected using in vitro, or ex vivo analysis. In one case, using subcellular fractionation, we have demonstrated that P-gp and glucocerebrosidase are translocated separately to the cell plasma membrane and lysosomes, indicating correct intracellular protein trafficking (112). The demonstration that a noncoding RNA, such as a hammerhead ribozyme, can function even though tethered to an mRNA encoding a functional MDR1 provides an additional powerful way to use bicistronic vectors (130).

Another approach to the use of MDR 1-based bicistronic vectors is to develop ''suicide'' vectors for cancer gene therapy. Using MDR1 to protect bone marrow cells from cytotoxic drugs represents a promising approach to improve cancer chemotherapy. However, contaminating cancer cells may be inadvertently transduced with MDR 1, or transduced bone marrow cells may accidentally develop new tumors. In those cases, overexpression of P-gp could cause multidrug resistance in inadvertently transduced tumor cells that contaminate bone marrow, or in any transduced cells that later become malignant. A bicistronic ''suicide'' vector developed in this laboratory links P-gp expression with herpes simplex virus thymidine kinase (TK) expression (105,106). Thus the cells containing this vector can be eliminated through ganciclovir treatment.

A third approach is to link two drug resistance genes together using a bicistronic vector system to extend the ability of the vector to confer drug resistance. Examples include the use of MDR1 with dihydrofolate reductase that confers methotrexate resistance (132), MDR1 plus methylguanine methyl-transferase (MGMT) that confers resistance to certain alkylating agents (102,103,133), and MRP1 plus gamma-glutamylcysteine synthetase that confers resistance to alkylat-ing agents as well (83).

Finally, bicistronic vectors can be used to introduce marker genes into selected cells. For example, MDR1 vectors containing green fluorescent protein or p-galactosidase have been constructed to determine the efficiency of expression of the target gene in transduced and MDR1 selected cells (124).

B. Efficiency of IRES-dependent Translation

Using an IRES to generate a bicistronic mRNA ensures coex-pression of two different proteins. However, IRES-dependent mRNA translation (or cap-independent translation) is less efficient than cap-dependent translation, so that the two proteins are not expressed in equal amounts. It has been shown that in a monocistronic vector, insertion of an IRES upstream from an open reading frame of either P-gp or dihydrofolate re-ductase (DHFR) reduces the translation efficiency by 2- to 10-fold (129,134). Using a bicistronic vector, expression of neo in the position downstream from the IRES is 25% to 50% of that observed when neo is in the upstream position (128). The asymmetric expression pattern of the bicistronic vector results in a significant difference in MDR1 transducing titer between a configuration with P-gp placed before the IRES and a configuration in which P-gp is placed after the IRES. We have found that the apparent titer of a bicistronic vector containing ADA-IRES-MDR1 was only 7% of the titer of a bicistronic vector containing MDR1-IRES-ADA (118). Similar reductions in MDR1 transducing titer and in expression of the nonselected downstream gene was seen with MDR1-fi-galactosidase bicistronic vectors too (115). The apparent MDR1 transducing titer of the retrovirus is based on the drug resistance conferred by expression of P-gp as the result of retroviral infection; thus the viral titer is proportional to the P-gp expression level. Insufficient expression of P-gp is unable to protect the cells from cytotoxic drug selection. To achieve P-gp expression at the same level, the lower efficiency of translation would have to be compensated for by a higher level of transcription, which can occur only in a minority of the cells in the transduced population. This may account for the apparent lower MDR 1-transducing titer of bicistronic vectors with a configuration of P-gp placed after the IRES. On the other hand, when cells express P-gp at the same level (i.e., the cells survived vincristine or colchicine selection at the same concentration), ADA expressed fromADA-IRES-MDR1 is 15-fold higher than the ADA expressed from MDR1-IRES-ADA. This difference is probably due to a combination of the lower translation efficiency of ADA located downstream from the IRES and the high transcription level of ADA-IRES-MDR1 as the result of vincristine selection. A similar asymmetric expression of P-gp and human p-galactosidase A is also observed in NIH3T3 cells, where the difference is about 8-fold.

IRES-dependent translation is a complex process, in which mRNA containing IRES interacts with various cellular proteins, including IRES transacting factors [reviewed in Hellen and Sarnow (135)]. The efficiency of IRES-dependent translation can be affected by the cell type (136), IRES origin (137,138), and the size and structure of a particular mRNA molecule. We have found that the titer of retrovirus containing pHa-MDR1 was higher than pHa-MDR1-IRES-ADA, even though P-gp translation was cap-dependent in both cases. Pgp expressed from pHa-MDR1 was also at a higher level in a vincristine resistant cell population than the P-gp expressed from pHa-MDR1 -IRES-ADA. A possible explanation for the relatively low retroviral titers observed is RNA instability or alternative splicing, since no DNA rearrangement was detected by Southern blot analysis of the transduced cells using an MDR1 probe.

In addition to IRESes derived from viruses, several IRES elements have been identified in human genes. Those IRESes play important roles in cell cycle-dependent or stress-response translation regulation [reviewed in Sachs (139)]. In contrast to viral IRESes, human IRESes are shorter and are complementary to 18s rRNA [reviewed in Mauro and Edelman (140)]. It has been found that a 9-nt sequence from the 5'-UTR of the mRNA encoding the Gtx homeodomain protein can function as an IRES. Ten linked copies of the 9-nt sequence are 3- to 63-fold more active than the classical EMCV IRES in all 11 cell lines tested (141). Similarly, an IRES isolated from the human EIF 4G gene also exhibits 100-fold more IRES activity than EMCV IRES in 4 different cell lines (142). In addition to higher efficiency and smaller size, translation from a human IRES can be regulated by cellular events (142), which may be advantageous for certain cancer gene therapies.

C. Flexibility Using Bicistronic Vectors in Coordinating Expression of Selectable Markers and a Therapeutic Gene

Selectable bicistronic vectors provide great flexibility in coordinating expression of a selectable marker, such as P-gp, and a therapeutic gene. The low translation efficiency of the IRES results in asymmetric expression of genes positioned before and after the IRES. This asymmetric expression pattern makes it possible to alter the relative expression level of a therapeutic gene and P-gp to achieve maximum therapeutic effects while applying minimal selective pressure using a cytotoxic drug. By choosing different configurations, i.e., placing MDR1 before or after the IRES, we can select cells expressing a therapeutic gene at either a low level (MDR1 before the IRES) or a high level (MDR1 after IRES).

In addition, expression of a therapeutic gene can also be achieved at a desired level by altering the selection conditions. The degree of multidrug resistance conferred by P-gp corresponds to the amount of P-gp expressed on the plasma membrane. Using a bicistronic vector, the expression of a target gene is proportional to the expression of P-gp, which is directly linked to the selection conditions. In a highly stringent selection, instead of increasing the concentration of cytotoxic drug, P-gp reversing agents can also be applied in combination with low concentrations of cytotoxic drugs (143). P-gp reversing agents, also known as chemosensitizers, are noncytotoxic hydrophobic compounds that interact with P-gp and cause a direct inhibition of P-gp function. In the presence of a P-gp reversing agent, most P-gp-expressing cells are killed by the cytotoxic drug unless they express a large amount of P-gp to overcome the inhibitory effects. Using a combination of cytotoxic drug and chemosensitizer allows selection of cells expressing the therapeutic gene at a high level without need for a high concentration of cytotoxic drug. This strategy is especially desirable for an in vivo selection in which avoiding systemic toxicity is essential.

High expression of the target gene can be selected using cytotoxic drugs, cytotoxic drugs combined with chemosensi-tizers, or the vector configured to place the target gene placed before the IRES. However, those approaches also reduce the overall number of cells that can survive the selection. Nevertheless, using a minimum concentration of drug, the selectable bicistronic vector provides options for selecting a large population of cells with low expression of the target gene, or a small population of cells with high expression of the target gene. Both options may be useful for gene therapy. For instance, ADA levels in normal individuals occur over a very broad range. Heterozygous carriers can be immunologically normal even with as little as 10% of the normal amount of ADA [reviewed by Blaese (144)]. Expression of ADA at a low level in a large number of cells may prove sufficient to treat SCID. On the other hand, high ADA-expressing lymphoid cells, even through present as a small percentage of total cells, are also able to correct the SCID syndrome due to a beneficial by-stander effect (145). In gene therapy applications, the choice of the approach depends on the thera peutic strategy for a specific disease. Experiments on animal models are essential to prove the concepts that underlie gene therapy using selectable markers such as MDR1.


Efficient delivery of a therapeutic gene to the appropriate target cells and its subsequent maintenance and expression are important steps for successful gene therapy. Genes introduced into cells are rapidly lost unless there is a mechanism to retain these genes within the nucleus and to ensure that the genes are also replicated and partitioned into daughter cells during cell division. Long-term expression of the transgene within cells can be achieved either via the integration of the transferred DNA into the host genome or maintenance of the introduced DNA as an autonomously replicating extrachromosomal element or episome. In either case, inclusion of a drug-selectable marker, like the MDR1 gene, in the construct would ensure that rapidly dividing cells containing the transgene are given a selective growth advantage.

Delivery modalities can be viral or nonviral. Retroviral gene transfer, one of the most exploited systems for gene transfer into actively dividing cells, has been discussed earlier in this chapter while liposomal gene delivery will be discussed later in the chapter. In this section, nonretroviral and/or epi-somal vectors expressing selectable markers will be described.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook

Post a comment