|[Previous Story] [Next Story]
GENE DELIVERY--WITHOUT VIRUSES
Nonviral methods represent only a fraction of the gene delivery field, but they are catching up with viral vectors
CELIA M. HENRY, C&EN WASHINGTON
DR GOPAL MURTI/SCIENCE PHOTO LIBRARY
RUNNING RINGS Plasmids derived from bacteria carry DNA used for nonviral gene delivery.
For more than 20 years, researchers have been working to alleviate disease through gene therapy. In this type of treatment, a gene is delivered to cells, allowing them to produce their own therapeutic proteins.
For example, many researchers have focused on inserting genes for Factor VIII to treat hemophilia A and Factor IX for hemophilia B. Hemophilia is a popular model system because it is a single-gene defect, and replacing about 5% of the amount of protein present in normal individuals is enough to effect a therapeutic cure.
At the beginnings of gene therapy, genes were usually transferred by using viruses that could infect cells, deposit their DNA payloads, and take over the cells' machinery to produce the desirable proteins. Increasingly, however, researchers are ditching the viruses and using plasmids--small rings of DNA produced in bacteria--to get the genes into the target cells.
In the past, such nonviral methods of gene delivery--including "naked" DNA and DNA condensed with agents such as cationic lipids or polymers--were considered less efficient than viral vectors. (Naked DNA is simply plasmid DNA not complexed with anything else.) But industry and academic researchers working in the nonviral gene delivery field say that's not true anymore. Despite improvements, however, nonviral methods remain a small, though increasing, fraction of the research presented at gatherings such as the annual meetings of the American Society of Gene Therapy.
A number of the people working in nonviral gene therapy used to work with viral vectors. They had different reasons for turning to nonviral methods, but they do have one thing in common: They saw problems with viral gene therapy.
Jon A. Wolff, director and professor of pediatrics and genetics at the University of Wisconsin, Madison, Medical School, used to work with retroviruses, which carry their genetic material in RNA instead of DNA, as a way to deliver genes. But then he realized that viruses had some limitations. "I thought nonviral might be a simpler and easier way of [effecting delivery] and avoid the problems with viral" therapy, he says. The main problem is the immunogenicity of the viruses. "Viruses have evolved over a billion years to efficiently deliver DNA into our cells, but our immune system has also evolved over that time," he says.
Wolff also saw nonviral methods as an opportunity to use his undergraduate training in chemistry. "It seemed to be more of a molecular chemistry sort of challenge than working with [viral] vectors" was, he says.
Richard F. Selden, president and chief executive officer of Transkaryotic Therapies in Cambridge, Mass., has always worked with nonviral methods. He saw that there could be two basic approaches to developing gene therapy vectors: either taking what was then the state of the art in molecular biology and using that to jerry-build a system that would work for gene therapy or determining the ideal properties of a gene therapy system and then starting from scratch.
Selden took the "ideal properties" approach because he believed that the state of the art in molecular biology was moving so rapidly in the late 1970s that "it didn't make sense to say: 'Let's just stop where we are in 1978 and try to turn that into gene therapy.' "
In addition, Selden worried about the safety of retroviruses, at that time the main way of getting genes into cells. Perhaps the most notorious retrovirus is HIV, the virus that causes AIDS.
"The reason people thought retroviruses were okay was that they said that retroviruses were unassociated with any human disease," he says. "Even though there wasn't yet an association, there were clear data that should have made people worried. It was shown that a retrovirus, the mouse mammary tumor virus, was the causative agent of breast cancer in mice. Even though I never could have imagined how dangerous retroviruses really are, I don't think it took any great leap of faith to think that retroviruses have safety problems."
Selden has similar concerns today when he hears people say that the most popular viral vector for gene therapy--the adeno-associated virus--is unassociated with human disease. He thinks that viruses are inherently unsafe and should be avoided. "I thought that a nonviral method would allow a safer system and would also ultimately allow more control and would give better efficacy," he says.
"Viral vectors are always going to be handicapped by the regulatory, safety, and production issues," says Philip L. Felgner, chief scientific officer at Gene Therapy Systems in San Diego. With "nonviral methods, if they remain relatively simple carrier systems, it will be much easier to achieve practical pharmaceutical products. It's going to be a yin-yang or tug-of-war between high expression and practical issues."
In developing nonviral vectors, the goal is to design a system that simultaneously achieves high efficiency, prolonged gene expression, and low toxicity, says Leaf Huang, professor of pharmaceutical sciences at the University of Pittsburgh. A number of vectors fulfill one or two of those criteria, but it is difficult to meet all three, he says. The most consistent problem is toxicity, he notes.
|HIGH UPTAKE Photomicrographs of monkey skeletal muscle sections histochemically stained for b-galactosidase expression. The cells on the left show a high level of expression, whereas the cells on the right show an average level of expression. Reprinted from Hum. Gene Ther., 12, 427 (2001).
NONVIRAL VECTORS can be divided into two broad categories--physical and chemical--according to Huang. Physical methods involve taking plasmids and forcing them into cells through such means as electroporation, sonoporation, or particle bombardment. Chemical methods use lipids, polymers, or proteins that will complex with DNA, condensing it into particles and directing it to the cells.
Nonviral systems for gene delivery have several potential advantages over viral vectors. Viruses cause an immune response that can make repeat administrations ineffective. Nonviral vectors also can usually carry more DNA than viruses, allowing the delivery of larger genes.
In addition, nonviral vectors are easier and less expensive to manufacture. The plasmids that are used in nonviral systems can be produced in bacteria such as Escherichia coli. The same production facilities can be used to manufacture a variety of plasmids incorporating different genes. And, as Robert C. Moen, president and CEO of Copernicus Therapeutics, Cleveland, points out, all the other components of nonviral systems are chemical, "so it's like a real pharmaceutical."
SYNTHETIC GENE delivery systems can be manufactured as pharmaceutical products, agrees Alain Rolland, senior vice president of preclinical R&D at Valentis in The Woodlands, Texas. "We have been able to manufacture gene-based medicines as single-vial, lyophilized, stable formulations," he says. Some of the products that Valentis is using in its clinical trials have been stable for more than two years at 4 °C and some, even at room temperature. In contrast, many viral vectors must be frozen to be stored for extended periods.
Synthetic delivery systems are also much easier to characterize than viral vectors. "Nonviral systems or gene-based medicines can be characterized as any pharmaceutical product" can, Rolland says. "Depending on the specific formulation, we can characterize the delivery components, the complex size, the DNA concentration, the synthetic gene delivery concentration, and so on. We can use a number of assays for specifications of our products."
PEELING AN ONION The multilamellar structure of cationic lipid-DNA complexes used in clinical nonviral gene delivery consists of DNA monolayers (purple chains) sandwiched between lipid bilayers with cationic (green) and neutral (white) head groups. New factors (red spheres) can be incorporated to facilitate DNA release inside cells. Ongoing research should clarify the relationship between the physical and chemical properties of the complex and its transfection efficiency. Adapted from Proc. Natl. Acad. Sci. USA, 97, 14046 (2000).
Mirus Corp., Madison, Wis., has been working on intravascular methods of delivering naked DNA to muscle in collaboration with Wolff, one of the company's founders. "We inject directly into the vascular system and then increase the permeability of the blood vessel so DNA can escape the blood vessel and target the myofibers of the muscle cells directly," says James E. Hagstrom, vice president for scientific operations at Mirus. The method relies on high volumes and rapid injections to increase the pressure in the blood vessels and allow DNA to "leak" out. In trials with primates, the company used a blood pressure cuff to increase the pressure.
Other researchers have suggested that such a method might not be "clinically relevant." That was also a concern for Wolff, but he thinks Mirus's studies are suggesting otherwise. "We did it in monkeys, and it worked really well and wasn't very toxic," he says. They have also used high-pressure techniques to deliver DNA to the liver in dogs and the heart in pigs. The method shows some toxicity to those organs. "By controlling how high the pressure is and the condition, one can balance the toxicity versus the benefit," Wolff says. "The actual data are looking more promising than I initially thought."
Vical, based in San Diego, is another company that is working with naked DNA. Plasmids are mixed with carriers and injected into the muscle. These carriers don't condense DNA, but they allow the plasmids to get past the cell membrane and into the nucleus, according to Vijay B. Samant, Vical's CEO. In addition, the carriers help protect DNA from nucleases, which are enzymes that chew up nucleic acids. For vaccine applications, these so-called adjuvants help turn on the innate immune response, says David C. Kaslow, chief scientific officer at Vical.
Vical is using naked DNA to develop cancer vaccines. The plasmids express an antigen that is suppressed in the cancer cell. The antigen makes the tumor cells more visible to the immune system, which then attacks them. One product, called Allovectin-7, is in Phase III clinical trials for treatment of metastatic melanoma, a particularly deadly form of skin cancer. "We stand a chance of being the first gene therapy product approved in the U.S., assuming our results come out where they should be," Samant says.
|MOUSE LIVER Fluorescently labeled naked DNA following intravascular delivery to mouse liver. The images were taken at less than five minutes (left) and at one hour after injection. Nuclei are stained blue, actin filaments are stained green, and the plasmid DNA is stained red. Reprinted from J. Gene Med., 2, 76 (2000).
Genteric, Alameda, Calif., is working on a way to deliver naked DNA orally. Although the company calls the concept the "gene pill," President and CEO Martin D. Cleary points out that it doesn't yet "have a pill in the sense of a small white, brown, or pink object." The company is working to develop an oral administration vehicle for DNA. Right now, DNA is delivered to the target cells--the intestinal villi--by a catheter that delivers naked DNA directly to the intestine or by a process called oral gavage, which is the introduction of a fluid to the stomach through a tube.
The intestinal cells that are transfected have a short life: one to two days on average. Therefore, the gene expression is very short term. Cleary calls that the "beauty" of the gene pill. "You have the condition where, if you want a therapeutic effect, you administer a pill. When you want the therapeutic effect to stop, you stop administering the pill, and the protein manufacturing stops almost immediately," Cleary says. Insulin is one of the proteins whose gene they hope to deliver using the gene pill. Other delivery platforms that Genteric is working on include delivery to the salivary glands and the liver.
IN ANOTHER MAJOR area of nonviral gene delivery, chemicals such as cationic lipids or polymers are used to condense DNA into particles. These particles tend to range in size from 100 to 300 nm.
One of the challenges for DNA particles is releasing DNA in the cell. Mirus is working on particles as well as naked DNA. "We have the ability to get these particles into [liver cells] efficiently," Hagstrom says. "What we don't have is the ability to get high levels of gene expression from those particles."
The problem, Hagstrom continues, is that you need stable particles that are able to fall apart and release DNA once they're in the cells, so that DNA can get to the nucleus. "We think we've solved a lot of the assembly issues, but now we're working on the disassembly of the particles once they reach their target cell."
Although the mechanism of releasing DNA from the particles is not known, Cyrus R. Safinya, a professor of materials, physics, and biomolecular science and engineering at the University of California, Santa Barbara, proposes that the lipids are peeled away a layer at a time, like an onion, exposing DNA that can leave the complex [Science, 275, 810 (1997); Science, 281, 78 (1998); Proc. Natl. Acad. Sci. USA, 97, 14046 (2000)].
Moen claims that Copernicus is able to condense plasmid DNA to "its theoretical minimum size." Most of their work is being done with polylysine derivatives. They can compact an "average size" plasmid to about 20 or 25 nm. Moen believes that this small size allows the particles to get into the nucleus of nondividing cells, with efficiencies three to four orders of magnitude higher than those seen with noncompacted DNA, through the nuclear pores, which are 25 to 50 nm wide. It is usually thought that DNA enters the nucleus during cell division, because the nuclear membrane breaks down during mitosis. Poly(ethylene glycol) added to the particles helps keep them stable at high concentrations.
||FAKING IT Chromos' artificial chromosome (bottom) is very similar to a natural chromosome (top) but carries only the genetic information that has been engineered into it.
||IN THE PINC Valentis' PINC polymers, such as polyvinyl-pyrrolidone, interact with DNA without condensing it. The white and red shading depicts hydophobic and hydrophilic regions of the polymer. The white is the most hydrophobic and thus indicates that the polyvinyl backbone (hydrophobic) is coating the DNA and rendering its surface more hydrophobic.
"We use the polylysine because, when it's complexed with DNA, it's not toxic. When you use a lipid-based system, often the lipids themselves are associated with some toxicity," Moen says. "Polylysine by itself in large quantities can have toxicity. But when polylysine, which is very positively charged, is complexed with DNA, which is very negatively charged, they combine, and all the toxicities seem to go away for both agents."
Huang's lab has worked to solve the problem of toxicity in using cationic lipids. Rather than premixing DNA and lipid, he injects them sequentially, lipid first. A few minutes later he injects DNA. "We found that the transfection activity in the lung is just as high, if not more, than with the complex," Huang says. "More importantly, the inflammatory toxicity is greatly reduced." His suspicion is that the reduced inflammation results from less DNA being delivered to macrophages, which produce cytokines that cause inflammation. He doesn't believe that such an approach would work for local delivery to tumors.
Targeted Genetics is working on both viral and nonviral methods of gene delivery, with the nonviral methods being focused on shorter term expression. "We like to use [nonviral vectors] in the case of cancer, where you actually want the cell that receives the gene to die," says Ralph W. Paul, director of technology discovery. The company has two different nonviral systems. One is a lipid-DNA complex that is currently in clinical trials for head and neck cancer as well as ovarian cancer. Targeted Genetics' other nonviral system uses cations to condense DNA, which is then placed in a liposome.
Insert Therapeutics is also working with condensing agents. However, the company is working with cyclodextrin-containing polymers rather than lipids or the more traditional polymers. The company's work is based on research done in the laboratory of Mark E. Davis, a professor of chemical engineering at California Institute of Technology.
"At Caltech, we have been working on the structure-property relationships between the molecular structure of polymers that bind and condense DNA and their cellular gene delivery properties," Davis says. "Our work is showing that very subtle structural variations in the polymer can have large effects in the gene delivery properties. The mechanistic origins of these relationships are under investigation as well."
Davis says one of the important things about using cyclodextrin-containing polymers as a gene delivery vector is the intellectual property (IP) considerations. "These delivery products have their own intellectual property and do not require additional IP for application," Davis says. "This will be an important feature in the future, as the IP for fully formulated products that are based on either lipids or polycations for nonviral gene delivery is becoming very messy."
Gene Therapy Systems is developing "peptide nucleic acid-dependent gene chemistry" to improve gene expression. PNA is a DNA analog that has a peptide backbone rather than a sugar and phosphate backbone. PNA can bind to its target sequence on DNA and actually displace the nontarget strand, Felgner says.
"We've made a molecule called a peptide nucleic acid clamp," Felgner says. "It's called a clamp because it has two places that it binds to DNA, two ways that it binds to the same sequence." First, PNA forms a Watson-Crick duplex with DNA. Then, another strand binds to the duplex, so two PNA strands are bound to the same target. The researchers attach various targeting ligands to PNA that possibly help DNA cross the nuclear membrane.
For delivery to skeletal muscle, Valentis is using polymers that interact with DNA without condensing it. Skeletal muscle is dense tissue, without a lot of space for particles to diffuse through, Rolland notes. "We found that, with this noncondensing system, we can retain the flexibility of DNA. The DNA is better able to diffuse through the muscle and get access to a larger number of cells, which translates to higher levels of expression," he says.
These PINC (protective, interactive, noncondensing) polymers have been designed to protect DNA from nucleases. Unlike condensing polymers, PINC polymers don't interact with the phosphate groups in the DNA backbone. Instead, they interact with DNA through hydrogen bonding or hydrophobic interactions. They fit in the major groove and coat the DNA, modifying its surface properties.
"We've been looking at this system mainly for solid tissues, such as skeletal muscle, cardiac muscle, and solid tumors. In solid tissues, you need to keep the DNA in a flexible way so that it can diffuse better through the extracellular matrix and get access to a larger number of cells," Rolland says. "If you condense the DNA in some of these tissues, you don't see very good diffusion and mainly see expression at the site of injection or along the needle track."
THE FIRST-GENERATION PINC polymers were polyvinyl derivatives. They are being used in two products that are currently in clinical trials, interleukin-12 and interferon-a for head and neck cancer and melanoma. Other polymers include a poly(oxyethylene)-poly(oxypropylene) block copolymer called poloxamer, which has entered clinical trials with a gene-based medicine for cardiovascular disease.
Safinya is trying to figure out just how DNA-lipid complexes work. "If you look at nonviral gene delivery, it involves using synthetic particles that have been designed based on chemical and physical concepts and then looking at how they interact with a cell," he says. "A lot of those events are physically and chemically dependent on the nature of the assembly."
Safinya's goal is to identify chemical and physical parameters and determine how they correlate with transfection. Such parameters could include the lipid structure, its charge distribution, or the mechanical properties of the membrane. He notes that scientists have virtually total control over the properties of the synthetic complex. "You can begin to design specific molecules that will counteract host mechanisms that are set to destroy anything that comes in," he says.
One of the goals should be to design particles that can target particular cells, much as viruses do, Safinya says. "Viruses have coevolved with cells, so that they will come in and arrest certain natural processes. Then they'll take over the machinery and make thousands of copies of themselves. That's sort of what we're trying to do."
One of the ways to target DNA is to use ligands on the particle surface. Many people are developing such targeting moieties. For example, Copernicus is targeting the polymeric immunoglobulin receptor, which is found in cells lining the airway. Selective Genetics is using fibroblast growth factor as a ligand.
Some companies are working on what is called cell-based or ex vivo gene therapy. For example, Transkaryotic Therapies places therapeutic genes, using an electric field to open pores in the cellular membrane, in cells harvested from a patient. The cells are then reimplanted.
Transkaryotic Therapies primarily uses fibroblasts, which are connective tissue cells. These cells are particularly good at dividing in culture but don't grow much once they're back in the body, Selden says. "Another big advantage of fibroblasts is that when you put them in the body, they stay where you put them," he says. "They lay down a nice collagen matrix." To treat hemophilia, the fibroblasts are implanted in abdominal fat, he says, because the Factor VIII protein is large and is not taken up efficiently when injected subcutaneously.
Transkaryotic Therapies published the results from its Phase I study for hemophilia A this past summer [New Engl. J. Med., 344, 1735 (2001)]. "We have shown that four of the first six patients with hemophilia A had clinical benefit, with two of those patients not bleeding spontaneously for almost a year," Selden says.
Immune Response Corp. is also working on cell-based gene therapy. They are engineering adult stem cells to produce Factor VIII. They haven't yet figured out how to get the cells to populate the liver, but they're working on it, according to C. Richard Ill, senior director for molecular biology.
Immune Response's shift to stem cells comes after eight years of working on methods to deliver DNA as particles using condensing agents. Although they had little luck with condensing agents, their concurrent efforts to increase the potency of the Factor VIII plasmid led them to create a novel gene for the protein.
MUCH OF GENE THERAPY is done with complementary DNA (cDNA), which is reverse transcribed from messenger RNA, because the actual genomic sequences are too big. For example, the Factor VIII gene has about 185,000 base pairs. The cDNA has only 9,000 base pairs, because all the noncoding regions, or "introns," have been removed.
The problem with using cDNA is that cells have complexes called spliceosomes that take RNA being produced from DNA and splice out the introns at what are called splice donor and splice acceptor sites. Even though the cDNA is pure coding sequence, the spliceosome still looks for recognition sequences on the RNA and splices part of the RNA.
"We've gone through using computer algorithms and searched for these splice donor and splice acceptor and branch sequences," Ill says. "We essentially said we don't want those sequences in our coding sequence." The DNA sequence is changed so that a different set of nucleotides is used to code for the same amino acid. RNA produced from these altered sequences is more stable, with less likelihood of producing aberrant proteins, he says. This technology could have applications anywhere that cDNA is used, not just gene therapy.
ANOTHER TECHNOLOGY that may find application in cell-based gene therapy is the human artificial chromosome being developed by the Canadian company Chromos Molecular Systems, Burnaby, British Columbia. These chromosomes have all the features of natural chromosomes, including a centromere, telomeres, and structural, noncoding DNA. However, they contain only the genetic information that Chromos scientists engineer into them. The artificial chromosomes reside in the nucleus beside the natural chromosomes without integrating into the natural chromosomes. When the cell divides, the artificial chromosomes replicate just like natural chromosomes.
Chromos grows the artificial chromosomes from naturally occurring acrocentric chromosomes, which have their centromeres very close to one end, producing short and long arms. "We've been able to identify a sequence that we refer to as the megareplicator," says Alistair Duncan, president and CEO. "This means that when we target into it, it triggers a big amplification. The short arm begins to grow. A second centromere ultimately appears. Now you have a structure that has a long arm of the chromosome, then a centromere, then it has the new chromosome. It does not have any genetic information in it other than what we've targeted in there as part of the triggering process." This dicentric chromosome is unstable, and when the cell divides, it breaks, leaving the original chromosome and the new artificial one.
The artificial chromosomes are large structures: between 40 million and 60 million base pairs long. Chromos has demonstrated information-carrying capacities of between 1 million and 1.5 million base pairs, Duncan says. "Our system allows you to start contemplating working with gene arrays, multiple copies of a gene, or geno-mic sequences. It allows you to start contemplating putting on genes and coupling them to appropriate regulators. It allows you to really start thinking about being able to focus in on novel and complex proteins that might not otherwise be able to be produced in existing systems," he says.
Selective Genetics is carrying out local gene delivery using a technology called gene-activated matrix, which is composed of a biocompatible matrix, according to Barbara A. Sosnowski, vice president for technology development. It can be made of materials such as collagen, a hydrogel, carboxymethylcellulose--any biocompatible material. The matrix is mixed with a gene therapy vector, viral or nonviral.
The mixture can be made to assume any desired shape for use at a wound site. If it is lyophilized, it can form an implantable sponge. It can also be used as a gel that is injected via syringe. As the body tries to heal itself, it sends wound-repair cells to the site. Those cells invade the matrix, take up the gene, and express the protein during the entire healing process. If a gene-encoding platelet-derived growth factor b is used, the expressed protein acts as a chemoattractant for other cells to come to the site, Sosnowski says.
The matrix can be used to accelerate the wound-healing process in different types of tissues by changing the gene. Selective Genetics is using the gene-activated matrix to stimulate diabetic ulcer wound healing, bone formation, and blood vessel formation, Sosnowski says.
"What's really nice about the technology," Sosnowski explains, "is that it remains localized at the desired site. Not only is the gene contained within the injured site, but if genes encoding proteins with cell retention signals are used, the protein also remains localized. The protein doesn't wander outside of that local environment."
Some of the work in nonviral gene delivery can be thought of as trying to create an artificial virus, without the viral components. DNA-lipid complexes already look morphologically like viruses, Huang points out.
"Most of the time in viral systems, we worry about taking things out, taking away, or trying to ameliorate any of the deficiencies the virus has, masking immunogenicity," Paul says. "We spend a lot of time trying to take away natural properties of the virus. In the nonviral systems, what we're essentially doing is creating synthetic viruses piece by piece, by adding in one element at a time to get the transfection more efficient, more targeted, able to escape any sort of nonspecific inflammation or immune response." Perhaps the vectors of the future will be hybrids, as researchers from the viral and nonviral sides meet in the middle.
Despite promising results in the clinic, DNA delivery without viruses still represents only a fraction of the research in gene therapy. But that may be changing.
"I already see people shifting from viral to maybe something in between viral and nonviral," Rolland says. "I would say that it will take some success in the clinic first to make people convinced that synthetic gene delivery systems are going to be the way to go. Some people are very strong supporters of nonviral systems and are encouraged by some recent tantalizing clinical results. However, until unequivocal success of pivotal trials in the clinic, people are going to be doubtful."
CHEMOATTRACTANT Selective Genetics' gene-activated matrix attracts repair cells to a wound site. The schematic shows a collagen matrix holding the gene therapy vector (tiny dark spots). A gene-activated matrix placed in a wound on a rabbit's back glows pink to show high levels of platelet-derived growth factor b mRNA expression (top). In the image at bottom left, the mRNA (tiny black granules) and its resulting protein (brown) remain localized in the cells in the matrix.
[Previous Story] [Next Story]
Chemical & Engineering News
Copyright © 2001 American Chemical Society