Gene therapy progress and prospects: hydrodynamic gene delivery

Over the last few years, hydrodynamic tail vein delivery has established itself as a simple, yet very effective method for gene transfer into small rodents. Hydrodynamic delivery of plasmid DNA expression vectors or small interfering RNA allows for a broad range of in vivo experiments, including the testing of regulatory elements, antibody generation, evaluation of gene therapy approaches, basic biology and disease model creation (non-heritable transgenics). The recent development of the hydrodynamic limb vein procedure provides a safe nucleic acid delivery technique with equally high efficiency in small and large research animals and, importantly, the prospects for clinical translation.

Non-viral approaches for gene transfer

Gene therapy has great potential to treat or prevent a variety of both genetic and acquired conditions that include neuromuscular disorders, cardiovascular disease, cancer, and infectious diseases. For recessive genetic disorders such as Duchenne muscular dystrophy, delivery of the normal dystrophin gene to muscle should prevent the myofibers from dying. Despite the great promise and sound principles of gene therapy, its application to humans have been hampered by the inability to safely and effectively deliver genes to the target tissues. Among the several gene transfer methods under development, the use of non-viral delivery methods and specifically naked DNA is particularly attractive in that many of the concerns over the use of viral-mediated methods, such as immunogenicity of viral packaging proteins and cost of viral vector production can be avoided. Recently we used limb veins for efficient, repeatable, and safe delivery of nucleic acids to skeletal myofibers throughout the limb muscles of mammals in vivo. Promising results have been obtained in both rodents and larger animals including non-human primates. Studies in the mdx mouse model indicate that the approach should be of use for patients with Duchenne muscular dystrophy. Based upon these encouraging results, a human clinical trial to deliver the human dystrophin gene to patients with DMD is being planned. The initial objective is to preserve hand and forearm function to increase the quality of life.

Progress and prospects: naked DNA gene transfer and therapy

Increases in efficiency have made naked DNA gene transfer a viable method for gene therapy. Intravascular delivery results in effective gene delivery to liver and muscle, and provides in vivo transfection methods for basic and applied gene therapy and antisense strategies with oligonucleotides and small interfering RNA (siRNA). Delivery via the tail vein in rodents provides an especially simple and effective means for in vivo gene transfer. Electroporation methods significantly enhance direct injection of naked DNA for genetic immunization. The availability of plasmid DNA expression vectors that enable sustained high level expression, allows for the development of gene therapies based on the delivery of naked plasmid DNA.

Time course of gene expression after plasmid DNA gene transfer to the liver

BACKGROUND

High levels of expression in hepatocytes can be achieved after intraportal delivery of plasmid DNA vectors with up to 10% of all liver cells transfected. CMV promoter-driven expression is very high on Day 1 after injection, but is diminished strongly by Day 2. Expression slowly declines after 1 week. We describe experiments aimed at elucidating the reasons for this rapid decline in transgene expression.

METHODS

Histological methods were used to determine the presence and extent of liver damage and hepatocyte proliferation. Viral and liver-specific promoters were tested to study promoter shut-off, Southern blotting was performed to determine the loss of the pDNA vector over time, and several mouse models were used to study the host immunological response.

RESULTS

pDNA is lost rapidly early after injection, but remains at a relatively stable copy number after Day 4. Southern blotting experiments showed that plasmid DNA could be detected for at least 12 weeks after injection (0.2 copies per genome). The early rapid decline of expression is promoter dependent. A liver-specific albumin promoter resulted in similar levels of expression on Days 1 and 7, suggesting that promoter inactivation may be responsible for the instability of CMV promoter-driven expression. The slow decline in expression levels after 1 week appears to be the result of an immune response directed against the expressed transgene. Expression was much prolonged in immunosuppressed, immunodeficient, or antigen-tolerized mice.

CONCLUSION

The present data suggest that if promoter inactivation can be overcome, intravascular delivery of plasmid DNA could be a highly efficient, simple and non-toxic liver gene therapy approach. Intravascular delivery of pDNA allows for the rapid screening of novel expression vectors in vivo.

Reciprocal enhancement of gene transfer by combinatorial adenovirus transduction and plasmid DNA transfection in vitro and in vivo

The addition of replication-defective recombinant adenovirus to plasmid transfection (termed here "adenofection") has been shown to increase plasmid transgene expression in limited studies. Similarly, the addition of cationic liposomes to adenovirus increases adenovirus-mediated gene transduction (termed here "lipoduction"). Here we demonstrate that adenofection was effective at enhancing transgene expression when used in conjunction with a variety of different transfection reagents, including a monocationic liposome, a polycationic liposome, an activated dendrimer, a large multilamellar liposomal vesicle, and a protein/amphipathic polyamine complex. The effect was seen regardless of the cellular expression of the adenovirus receptor, CAR, in three different human cancer cell lines derived from rhabdomyosarcomas (Rh18 and RD, CAR-) and cervical carcinoma (HeLa, CAR+). The protein/amphipathic polyamine complex showed an adenofection effect but did not show a lipoduction effect, consistent with different mechanisms of action for adenofection and lipoduction. Using dual-color flow cytometric analysis of cells transfected with a plasmid expressing the enhanced blue fluorescent protein (pEBFP) and a recombinant adenovirus expressing the green fluorescent protein (Ad5-GFP), we demonstrate that adenofection works primarily by increasing gene expression within a cell, whereas lipoduction increases the percentage of cells expressing the transgene. In addition, these studies show that both adenofection and lipoduction can occur simultaneously, further increasing gene transfer. The combination of lipofection and adenovirus transduction also prolonged the duration of transient gene expression and was generally no more toxic than lipofection alone. The enhancement of gene transfer was also seen after injection of complexes directly into subcutaneous human xenograft tumors. Therefore, more effective gene transfer in vitro and in vivo of either plasmid DNA, adenovirus DNA, or both can be achieved by combining liposomal transfection with adenoviral transduction.

Combined radiation and p53 gene therapy of malignant glioma cells

More than half of malignant gliomas reportedly have alterations in the p53 tumor suppressor gene. Because p53 plays a key role in the cellular response to DNA-damaging agents, we investigated the role of p53 gene therapy before ionizing radiation in cultured human glioma cells containing normal or mutated p53. Three established human glioma cell lines expressing the wild-type (U87 MG, p53wt) or mutant (A172 and U373 MG, p53mut) p53 gene were transduced by recombinant adenoviral vectors bearing human p53 (Adp53) and Escherichia coli beta-galactosidase genes (AdLacZ, control virus) before radiation (0-20 Gy). Changes in p53, p21, and Bax expression were studied by Western immunoblotting, whereas cell cycle alterations and apoptosis were investigated by flow cytometry and nuclear staining. Survival was assessed by clonogenic assays. Within 48 hours of Adp53 exposure, all three cell lines demonstrated p53 expression at a viral multiplicity of infection of 100. p21, which is a p53-inducible downstream effector gene, was overexpressed, and cells were arrested in the G1 phase. Bax expression, which is thought to play a role in p53-induced apoptosis, did not change with either radiation or Adp53. Apoptosis and survival after p53 gene therapy varied. U87 MG (p53wt) cells showed minimal apoptosis after Adp53, irradiation, or combined treatments. U373 MG (p53mut) cells underwent massive apoptosis and died within 48 hours of Adp53 treatment, independent of irradiation. Surprisingly, A172 (p53mut) cells demonstrated minimal apoptosis after Adp53 exposure; however, unlike U373 MG cells, apoptosis increased with radiation dose. Survival of all three cell lines was reduced dramatically after >10 Gy. Although Adp53 transduction significantly reduced the survival of U373 MG cells and inhibited A172 growth, it had no effect on the U87 MG cell line. Transduction with AdLacZ did not affect apoptosis or cell cycle progression and only minimally affected survival in all cell lines. We conclude that responses to p53 gene therapy are variable among gliomas and most likely depend upon both cellular p53 status and as yet ill-defined downstream pathways involving activation of cell cycle regulatory and apoptotic genes.

High expression of naked plasmid DNA in muscles of young rodents

There is a time window at 2 weeks of age for achieving very high levels of foreign gene expression from the intramuscular injection of naked plasmid DNA in mice and rats. The highest expression, over 1 microg of luciferase protein/muscle, was obtained in Balb/C mice using constructs containing the CMV promoter, a chimeric intron and the luc+ luciferase gene. Approximately 50% of the myofibers were intensely blue following the intramuscular injection of a beta-galactosidase expression vector in 2 week old Balb/C mice. The effects of age, mouse strain and construct were multiplicative, resulting in >1000-fold greater luciferase and approximately 20-fold more beta-galactosidase-positive cells. These high levels of expression were unstable and were not observed in larger animals (dog, rhesus monkey). These results indicate that enormous levels of foreign gene expression can be obtained in muscle with naked DNA in vivo and will enable the temporary effects of gene function and expression in rodent muscle to be expeditiously studied.

Gene transfer into mammalian cells using histone-condensed plasmid DNA

A recombinant histone (NLS-H1) containing both the SV40 large T antigen nuclear localization signal and the carboxy-terminal domain of human histone H1(0) was produced in bacteria. NLS-H1-plasmid DNA complexes, in the presence of chloroquine, mediated reporter gene transfer into cultured cells with similar efficiencies as plasmid DNA-cationic lipid (lipofectin) complexes. NIH-3T3 or COS-7 cells transfected with NLS-H1-plasmid DNA-lipofectin complexes expressed at least 20 times more luciferase or had at least 2.5 times more beta-galactosidase-positive cells than those transfected with plasmid DNA-lipofectin complexes. Foreign gene expression was also improved by other DNA-binding proteins and cationic lipid formulations, yet the greatest enhancement was obtained with complexes containing either NLS-H1 or calf thymus histone H1. Histone H1-plasmid DNA-lipofectin complexes were internalized by a greater number of cells than plasmid DNA-lipofectin complexes.

A plasmid-based self-amplifying Sindbis virus vector

Sindbis virus was used as a self-amplifying eukaryotic expression vector. A recombinant cDNA genome of this (+)-strand RNA virus was placed under the transcriptional control of a Rous sarcoma virus LTR (RSV) promoter. Transfection of this plasmid construct into mammalian cell lines (3T3, HepG2, and 293 cells) resulted in expression of the luciferase reporter gene. High-expression levels were also measured after transfection into primary rat myoblasts. In differentiated myotubes, expression levels generated by the Sindbis virus vector were up to 200 times higher than those obtained with a conventional RSV expression vector. In vivo expression was detected after injection of plasmid DNA into mouse quadriceps. In vivo expression was transient and undetectable by day 16. This self-amplifying expression vector can be used for generating high-level expression of transgenes in vitro and in vivo. Its transient nature in vivo could allow for safe, short-term delivery of gene products in gene therapy protocols. It should facilitate the study of Sindbis and other RNA viruses.

Dystrophin expression improves myofiber survival in mdx muscle following intramuscular plasmid DNA injection

Expression of Becker-like and full-length human dystrophins was stable for at least 6 months in mdx mouse muscle following intramuscular plasmid DNA injection. Intramuscular injection of a single plasmid DNA encoding both luciferase and dystrophin resulted in stable luciferase expression for at least 2 months in mdx muscle, whereas injection of plasmid DNA encoding only luciferase did not result in stable luciferase expression. These results suggest that expression of either full-length or Becker-like dystrophins protects mdx mouse myofibers from degeneration.

High expression of the mdr3 multidrug-resistance gene in advanced-stage chronic lymphocytic leukemia

Chronic lymphocytic leukemia (CLL) is characterized by an often indolent course with a poor therapeutic response at advanced stage. We investigated the expression of the human multidrug resistance genes mdr1 and mdr3 in 31 patients with CLL. Using specific probes for mdr1 and mdr3 mRNA, respectively, expression of both genes could be found in 29 of 31 patients. Of those, nine had high expression of mdr1 and 13 of mdr3. Although 29 of 31 patients showed coexpression of mdr1 and mdr3, the mRNA levels were not interrelated. Prior treatment did not significantly influence the level of mdr1 or mdr3 expression. In patients with advanced CLL (Rai stage 3 + 4) the mdr3 expression was significantly higher than in early-stage CLL (Rai stage 0 to 2) (mean +/- SEM, 25.4 +/- 4.2 U v 4.2 +/- 1.1 U; P less than .0001). Such a difference was not present for mdr1 expression (21.5 +/- 4.3 U v 10.7 +/- 3.1 U; P = .09). These data indicate that advanced-stage CLL is associated with an increased mdr3 expression, which may concur with a decreased sensitivity to chemotherapy.

A phase II study of epidoxorubicin in colorectal cancer and the use of cyclosporin-A in an attempt to reverse multidrug resistance

We determined the ability of the multidrug resistance (MDR) reversal agent cyclosporin-A to increase anthracycline drug accumulation in colorectal tumour cells in vitro, using the technique of on-line flow cytometry. Data of four previously untreated patients showed that cyclosporin-A can increase intracellular net-uptake of daunorubicin. A phase II study was initiated in 24 colorectal cancer patients. They received cyclosporin-A at a dose of 3 mg kg-1 over 1 h as i.v. infusion, at 7 h and at 1 h preceding cytotoxic drug administration. At the end of the second cyclosporin-A administration epidoxorubicin 90 mg m-2 was administered as i.v. bolus. Cycles were repeated every 3 weeks. Median cyclosporin-A peak blood levels and levels at 18 h after cytotoxic drug administration appeared to be 6248 ng ml-1 and 1012 ng ml-1 respectively. Only one partial response was observed, despite these high cyclosporin-A levels. Cyclosporin-A did not cause major toxicity, only a 29% incidence of hot flushes was observed. Epidoxorubicin toxicities were as expected but the frequency of severe leucocytopenia was striking. This treatment schedule can not be considered active in colorectal cancer.

The use of E. coli exonuclease III to generate single stranded DNA in BrdUrd cell-cycle analysis permits simultaneous detection of cell surface antigens

An immunocytochemical method for the simultaneous flow cytometric quantitation of total cellular DNA, incorporated 5-bromo-2'-deoxyuridine (BrdUrd) and one or more cell surface antigens has been developed. Biotin labeling of cell surface antigens, critically tuned fixation techniques and an enzymatic denaturation of cellular DNA are the essential features of this method. Enzymatic denaturation of cellular DNA was shown to prevent loss of cell surface antigen-bound biotin moieties, and thus to preserve cell surface immunofluorescence distribution. After a mild protein extraction and the introduction of breaks into the chromatin using restriction endonucleases, E. coli exonuclease III was used to generate stretches of single stranded DNA. This approach permits detection of the incorporated BrdUrd using anti-BrdUrd monoclonal antibodies. The enzymatic denaturation protocol was optimized using in vitro BrdUrd-labeled L1210 murine leukemia cells, and applied to both in vivo and ex vivo BrdUrd-labeled murine bone marrow cells. With this new method it is possible to study DNA content, cell cycle kinetics and cell surface antigen expression simultaneously, and hence functional relationships between these parameters can be investigated.

Expression of mdr1 and mdr3 multidrug-resistance genes in human acute and chronic leukemias and association with stimulation of drug accumulation by cyclosporine

We determined the expression levels of the mdr1 and mdr3 multidrug-resistance genes (also known as PGY1 and PGY3, respectively) in peripheral blood cells from 69 adult patients with acute and chronic leukemias, using an RNase protection assay. Expression of mdr1 was found in samples from patients with acute nonlymphocytic leukemia (13 of 17), chronic myelocytic leukemia (CML, chronic phase, 10 of 10; blast crisis, three of four), acute lymphocytic leukemia (ALL, eight of 11), B-cell chronic lymphocytic leukemia (B-CLL, 17 of 17), hairy cell leukemia (HCL, one of two), and T-cell prolymphocytic leukemia (one of one), but not in B-cell prolymphocytic leukemia (B-PLL, 0 of seven). Expression of mdr3 was only detected in samples from B-cell lymphocytic leukemias: CML, lymphoid blast crisis (one of one), B-cell ALL (two of two), B-CLL (17 of 17), B-PLL (seven of seven), and HCL (two of two). In vitro drug uptake studies by on-line flow cytometry showed that in leukemia cells expressing either mdr1 or mdr3, the steady-state accumulation of daunorubicin could be significantly increased by addition of cyclosporine and, to a lesser extent, by verapamil. Because cyclosporine and verapamil are known as inhibitors of the mdr1-encoded P-glycoprotein drug-efflux pump, and because the mdr1 and mdr3 genes are highly homologous, our data suggest that the mdr3 gene encodes a functional drug pump in B-cell lymphocytic leukemias. The results of this study may have implications for clinical therapy for acute or chronic leukemias expressing the mdr1 or mdr3 gene, in particular, treatment with combinations of cytotoxic drugs plus agents that reverse multidrug resistance. Since mdr1 and mdr3 are frequently expressed in untreated as well as treated leukemia, such combination therapy should be considered for untreated patients as well as treated patients.

Expression of the mdr3 gene in prolymphocytic leukemia: association with cyclosporin-A-induced increase in drug accumulation

Typical multidrug resistance in human and animal cell lines is caused by overactivity of an unidirectional transmembrane drug efflux pump, encoded by the MDR genes, called mdr genes in mice and humans and pgp genes in hamsters. In humans, two mdr genes, mdr1 and mdr3, with approximately 80% nucleotide homology, have been identified. There is increasing evidence that overexpression of the mdr1 gene plays a role in resistance to anticancer agents in specific tumor types. However, currently no data are available on a possible role for mdr3 in drug resistance. Here we report high levels of expression of mdr3 gene sequences in leukemic cells from 6 out of 6 patients with prolymphocytic leukemia (PLL). No mdr1 expression was detected in 5 out of 6 of these samples, whereas a low level of mdr1 expression was found in a sample from one PLL patient in the course of transformation to non-Hodgkin's lymphoma. Except for this patient, all other PLL cases studied had not received prior chemotherapy. In vitro drug uptake studies showed that daunorubicin accumulation in PLL cells was increased by cyclosporin A. Since cyclosporin A is an inhibitor of the mdr1-encoded P-glycoprotein drug pump, these data suggest that in PLL cells mdr3 also codes for a drug efflux pump. Our findings could partly explain the primary refractoriness of PLL to chemotherapy.

Overexpression of the mdr1 gene in blast cells from patients with acute myelocytic leukemia is associated with decreased anthracycline accumulation that can be restored by cyclosporin-A

Typical multi-drug resistance (MDR) in human and animal cell lines is caused by overactivity of a unidirectional drug efflux pump. This pump is composed of a 170-kDa transmembrane glycoprotein (P-glycoprotein) that is encoded by the so-called mdr1 gene. The functionally relevant characteristic of MDR cells is a defect in drug accumulation that can be restored by agents which inhibit the P-glycoprotein pump. The purpose of our study was to find out whether P-glycoprotein inhibitors could increase the daunorubicin (DNR) accumulation in acute myelocytic leukemia (AML) cells, overexpressing the mdr1 gene. Using dot blot analysis with an mdr1-specific cDNA probe, we identified leukemic cell samples, obtained from chemotherapy-resistant AML patients, that had relatively high levels of mdr1 expression. These leukemic cells showed a reduced ability to accumulate DNR in vitro, as quantitated by flow cytometry. Addition of cyclosporin-A (Cy-A), a drug known to inhibit the P-glycoprotein pump, to the incubation medium resulted in an increase (up to 60%) in steady-state drug uptake by the leukemic cells. The degree of Cy-A-induced increase in drug accumulation in the leukemic cells correlated approximately with the level of overexpression of the mdr1 gene. Our data indicate that Cy-A is a good candidate for combination chemotherapy with cytotoxic drugs in clinical trials, aimed at the treatment of drug resistance in AML.

Effects of cyclosporin A and verapamil on the intracellular daunorubicin accumulation in Chinese hamster ovary cells with increasing levels of drug-resistance

Multidrug-resistance (MDR) is characterized by the presence of a 170 kDa glycoprotein (P-glycoprotein) in the plasma membrane. P-glycoprotein is thought to act as an efflux pump, leading to reduced drug accumulation in MDR cells. This defect in drug accumulation can be overcome by membrane transport modulating agents (MTMAs). We determined the concentration of MTMA needed for maximal restoration of daunorubicin content in 4 Chinese hamster ovary cell lines with increasing levels of drug-resistance using flow cytometry. Stimulation of daunorubicin accumulation occurred in a dose-dependent manner. The required level of MTMA needed for maximal drug accumulation increased with the level of drug-resistance. CHrA3 cells, which have a level of resistance comparable to clinical samples, needed relatively low concentrations of MTMA for maximal restoration of drug accumulation. This indicates that, in trial combining drugs and MTMAs, low dosages of MTMAs could be sufficient for optimal potentiation of cytotoxicity.

A rapid and sensitive flow cytometric method for the detection of multidrug-resistant cells

Multidrug-resistant (MDR) cells are characterized by a defect in drug accumulation caused by activity of an energy-dependent rapid drug efflux pump. The action of this drug pump can be inhibited by specific agents, referred to as membrane transport modulating agents (MTMAs), resulting in a restoration of the intracellular drug accumulation. This paper presents a flow cytometric assay for the detection of MDR cells, which is based on the ability of these cells to respond to MTMAs. Daunorubicin net-uptake kinetics were measured of anthracycline-sensitive (A2780/S) and -resistant (A2780/R) human ovarian carcinoma cells in vitro. A2780/R cells accumulated significantly less (about a factor of 5) daunorubicin as compared to A2780/S cells. Addition of verapamil or cyclosporin A to A2780/R cells at steady-state daunorubicin uptake led to a dose-dependent increase in cellular daunorubicin accumulation. The sensitivity of the assay was determined by testing mixtures of A2780/S and A2780/R cells. Analysis of A2780/S cells contaminated with A2780/R cells showed that as few as 2.5% MDR cells could readily be detected in the mixture. In conclusion, this functional assay enables the detection of MDR cells in a heterogeneous cell suspension and is ideally suited for the study of the occurrence of typical MDR in human cancer.

High-speed photodamage cell selection using bromodeoxyuridine/Hoechst 33342 photosensitized cell killing

One of the major drawbacks of droplet sorting in a flow cytometer is the relatively low sorting speed. Thus, we have developed an alternative, faster sorting technique: photodamage cell sorting. In a photodamage cell sorter all unwanted cells, as detected with the first, measuring laser, are killed with the second, damaging laser. Thus, the cells need to be photosensitive to the second laser. In addition, a mechanism is needed to switch this laser on and off based on the sorting criteria. In our photodamage cell sorter, the ZAPPER, we use an acousto-optic crystal to switch the laser beam. Cells are made photosensitive by vital staining with photosensitizers. With cells grown in the presence of 5-bromo-2'-deoxyuridine (BrdUrd) and stained with Hoechst 33342 (H42) at least a 5-decade cell reduction is accomplished after irradiation with 400 mW UV light. With this system, sorting rates have been achieved of 30,000 cells per second. Due to the selection based on photodynamic killing, this sorting technique is restricted to the selection of viable cells. Photodamage cell sorting seems well suited for isolating viable cells occurring in low percentages or for the sorting of large numbers of cells. Another application can be the sorting of large or fragile cells.