© Christopher Zacharow/Images.com

Gene therapy's heady days of introducing the gene for adenosine deaminase into immune cells to treat a life-threatening congenital defect, as was done in the first gene therapy trial in 1990, have given way to an atmosphere of caution. In 1999, Jesse Gelsinger died of multiple organ failure four days after receiving adenovirus-based therapy for a rare liver disorder. In 2002, a child developed leukemia after receiving retroviral therapy for X-linked severe combined immunodeficiency (X-SCID), and another child in the same trial was similarly diagnosed the following year (with the possibility of a third recently reported). In June it was reported that a promising trial for hemophilia was halted because of apparent adverse immunological reactions to both the vector and the transgene.

The cause of Gelsinger's death, while not definitively established, may have been the high numbers of viral particles infused during the dose-escalation study. The X-SCID...


Gene therapists typically distinguish between the pre- and post-Gelsinger eras. Many, like the University of Pennsylvania's Carl June, subdivide even further, referring to the present day as the "post-Gelsinger, post-SCID" era. These represent not just ticks on a timeline, but also a fundamental shift in the field.

A defining characteristic of the present era is the markedly higher level of oversight to which clinical trials are subjected. In the wake of Gelsinger, disclosure and reporting requirements, and patient inclusion/exclusion criteria in the United States and abroad were strengthened. The X-SCID leukemias resulted in a temporary halt of about one-third of all gene-therapy trials, and almost all retroviral trials, most of which were subsequently allowed to resume but with increased patient monitoring.

Regulatory hurdles have gotten so time- and cost-prohibitive that smaller companies and academic institutions now shy away from these kinds of trials, says Stanford University's Mark Kay. Because of all the requirements, it's become difficult to fund these kinds of trials. Kay sees gene therapy as "being examined under a microscope," and that creates a problem for the field. Safety must remain the primary concern, he says, yet there needs to be a balance to prevent research from being stifled.

June, echoing these thoughts, refers to the HIV trial he is involved with, run by VIRxSYS of Gaithersburg, Md. "You wouldn't believe the amount of data we have to gather from one patient, now, in order to treat another." He explains that they are required to follow-up patients for 15 years to monitor for insertional mutagenesis-caused leukemias.



Courtesy of Carl June

The vicissitudes of fervor for gene therapy since its conception in 1988.

But not every trial is subject to quite so much scrutiny. The VIRxSYS trial is unique, in that this five-patient, Phase I safety trial is the first in which a lentiviral vector has been used to modify cells that are subsequently returned to a human subject. Four patients have been treated with autologous T cells transfected with an antisense code directed against the HIV envelope RNA, says VIRxSYS founder Boro Dropulic. The company plans to treat the fifth patient next month.

Unlike other retroviruses (such as the Moloney leukemia virus, MLV) currently in use or under investigation for gene therapy, lentiviruses can transfect a high percentage of nondividing cells, making them a promising vector for terminally differentiated tissues. "I can pretty much alter a T cell at will now," something that was nearly impossible just a few years ago, says June, an immunologist. "It's a huge breakthrough for immunotherapy that way."

These vectors are derived from pathogens such as HIV and its kin, and so the field proceeds with utmost caution. In theory, the vectors have the potential to recombine with endogenous host or viral sequences to produce replication-competent virus. "Our target patient population is already laden with wild-type HIV," Dropulic told the Food and Drug Administration at a meeting of the Center for Biologics Evaluation and Research's Biological Response Modifiers Advisory Committee (BRMAC). He noted that the patients were chosen because they have "no good treatment options left."1

Other lentiviral trials wait in the wings for the results of this pioneering trial. "There are all sorts of things where the preclinical data exists; now it's just [a matter of] overcoming the inertia of the regulatory and financial barriers," June notes. He predicts that 50 lentivirus trials will be in the pipeline within two years.


Retroviruses (including lenti) and adeno-associated viruses (AAV) integrate into the target's genome, allowing the transgene to be retained throughout the life of the cell and reliably passed on to its daughters. Yet this ability to integrate makes them subject to inactivation caused by chromatin remodeling, and at the same time is a factor that makes these vectors potentially dangerous.

The genomic insertion sites that presumably led to leukemia in two patients treated for X-SCID were examined and found to be related to each other, leading BRMAC to conclude that it "cannot be considered a random event."2 Indeed, neither MLV (used in the X-SCID trials) nor HIV integrates randomly. But, according to Shawn Burgess, whose National Institutes of Health laboratory has developed methods for high-throughput mapping of viral integration, these viruses use different sets of cellular cofactors and thus "are actually integrating in different kinds of nonrandom fashion."

MLV has a preference for landing very close to the start site of a gene, whereas HIV preferentially lands within the gene, explains Burgess. Thus, the potential problems of vectors constructed are likely to be different. MLV-ferried transgenes, because they contain promoters, could inappropriately turn on a gene; HIV vectors are more likely to inactivate a gene by disrupting it.

Like HIV, AAV is likely to land within a gene when it integrates, which it does at a fairly low frequency. Yet unlike retroviruses, AAV also has the capacity to exist as a freely replicating episome.

The other major classes of viral vector, adenovirus (AV) and poxvirus, exist and propagate only as episomes. Because they do not integrate, these vectors can be used to transfect both dividing and nondividing cells (often in high copy number), albeit with no assurance that the transgene will be passed down to daughter cells. And because the vectors require exogenous proteins, they tend to be more immunogenic than their integrating counterparts.


Viruses, because they enter cells by means of receptor-mediated uptake, deliver their cargo in a cell-specific manner. But genetic material can be introduced in other ways into cells. Of the nearly 1000 clinical trials listed on the Journal of Gene Medicine's online "Gene Therapy Clinical Trials Worldwide" database, more than one-fourth use nonviral vectors.3

This broad category encompasses a variety of technologies, essentially all of which involve complexing (and sometimes condensing) DNA with an agent that allows it to nonspecifically enter a cell, either by membrane fusion, endocytosis, or in connection with some membrane-disrupting event such as electric, chemical, or pressure shock. Research also is underway to target the DNA to specific cell types by complexing it with surface-receptor ligands, and according to Perry Hackett, chief scientific officer at Discovery Genomics in Minneapolis, the approach may soon yield success.

In general, nonvirally delivered DNA is only transiently expressed, because it very rarely integrates into the genome. Several companies and academic researchers are using transposon systems, however, to direct integration of naked DNA. With Discovery Genomics' system, for example, a plasmid containing a gene of interest flanked by transposable elements is cotransduced with a plasmid coding for the Sleeping Beauty (SB) transposase, explains Hackett. SB excises the gene and randomly pastes it into a chromosome, at rates at least 100-fold greater than that at which naked DNA integrates. Hackett and his colleagues are preparing an investigational new drug application to the FDA for the use of an SB-delivered gene to treat Fanconi anemia.

Integration of plasmids and transposon-mediated sequences, like integrating viruses, can have deleterious consequences if they land in the wrong place. Poetic Genetics of Burlingame, Calif., has been developing a system based on a phage integrase which, says company founder Michele Calos, can direct integration into specific sequences of mammalian chromatin.

Another approach to avoiding insertional mutagenesis involves using an entirely separate chromosome. Several artificial chromosome systems exist, each having their own telomeric and centromeric sequences. "The cell treats it just like its own chromosome," notes Harry Ledebur, Jr., vice president of research and development at Chromos Molecular Systems in British Columbia. In Chromos' ACE system, genes of interest are incorporated into the artificial chromosome (there are about 70 acceptor sites), which is then introduced into cells by standard transfection techniques. The company is currently prototyping delivery methods, Ledebur says, estimating that they are probably two to four years away from any human gene therapy.



Courtesy of Shawn Burgess

Different host factors may affect a retroviral vector's pre-integration complex (PIC). Moloney Leukemia virus appears to be influenced by cellular factors surrounding the transcriptional start site, while HIV-1 likely interacts with factors that target it to the transcribed region of a gene. The factors influencing the integrations are unknown.

"Each vector has inherent properties that make it better or worse for a particular application," points out Kay. "There are inherent properties of the vector, and inherent pathophysiological events in the disease, and the two have to basically match up." For infectious diseases such as hepatitis, "you probably don't need lifelong gene transfer if you can rid the virus from the body, but you need efficient gene transfer," says Kay, a liver specialist.

On the other end of the spectrum, even 5% of normal expression might be sufficient to cure the individual of a congenital metabolic disorder. But, Kay adds, "you need something that is going to last for the life of the individual, or something that can be readministered in a safe manner." The latter is another reason gene therapists are so interested in nonviral vectors: " [Viruses] all cause immune responses," notes Hackett. Once an immune response is engendered, future administrations will likely be rejected, "so you can only apply them once, at best."

Yet sometimes immunogenicity is just what is desired, explains High. "You have to know when it's going to help you and when it's going to hurt you." Many of the presentations at ASGT's 7th annual meeting, held in Minneapolis in June, focused on ways to boost the immune response to cancer and infectious disease following gene-based vaccination.


Tissue specificity also looms as a challenge. "All of these viral vectors have specific tissue tropisms," explains High. Viral vectors can in many cases be pseudotyped to bind different cell types, by modifying the capsid to include tissue-specific ligands, she continues. Yet even a virus with exquisite specificity for, say, pancreatic β cells may have a difficult time finding its target.

But much has changed since the early focus on cells that could be harvested, manipulated in vitro, and returned to the patient. The 1000+ ASGT talks and posters featured in vivo work on virtually every organ and organ system, sometimes with stunning results.

Paul Gregorevic from the University of Washington, Seattle, announced the transfection of extensive sections of cardiac and skeletal musculature following a single intravenous injection of AAV vector. At that same meeting, Italian and American researchers detailed the establishment of nonhuman primate models for disorders such as diabetic retinopathy and choroidal neovascularization, by injecting an AAV vector encoding the gene for vascular endothelial growth factor (VEGF) into the eye. Others have used genes for zinc finger transcription factors to initiate transcription of VEGF in models of peripheral arterial disease, and to manipulate a variety of other genes in other models. A group from Case Western Reserve University showed that condensed DNA, complexed with a peptide ligand for the serpin enzyme complex receptor, can be targeted to various organs depending upon the route of administration. Numerous presentations also detailed advances in gene silencing with RNA interference and short hairpin RNA gene therapy.

Introgen Therapeutics in Houston plans to file the first-ever application to the FDA to sell a gene therapy product later this year, the adenovirally delivered tumor suppressor p53 (brand named Advexin). Yet even with such exciting results, gene therapy for the most part remains years away from the clinic. For now, AV, lentivirus, and AAV are the most popular vectors, but Kay says that "nonviral systems are going to become much more prominent over time."

Regardless of what vector is used, "I think most people are hoping for intravenous infusion for most applications ... with the ultimate goal to have a tissue-specific vector," adds Kay, who is the ASGT's president-elect.

In trying to create the perfect balance, researchers hope the course of gene therapy will level out, but more dips and surges inevitably lie past the next curve in the track. Says Kay, "It's a relatively new field, and you can't expect miracles overnight."

Josh P. Roberts jroberts@the-scientist.com

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?