February 15, 2011 (Vol. 31, No. 4)

As Researchers Correct Past Shortcomings, Promising New Pathways Begin to Appear

Cries of “reports of gene therapy’s death have been greatly exaggerated” tend to be in a never-ending cycle with “gene therapy is dead.” Judging by the recent Congress of the European Society for Gene and Cellular Therapy, it looks like the former is in the ascendant: there is no shortage of activity in both laboratory and clinic.

Many researchers continue down promising pathways or attempt to correct shortcomings of the past, while others explore previously un- and undercharted territories. From the types of genes being delivered and the vectors they arrive in, to where they are headed and what they are destined to accomplish, to the means of ensuring that they heed Hippocrates’ admonition to above all do no harm, gene-delivery research seems to be alive and well.

Much gene therapy uses viral vectors to deliver corrective genetic material, allowing the target to produce a protein that it normally should have made. Sometimes the virus will be engineered to encode a protein designed to directly kill a wayward cancer cell, or to make it recognizable by the immune system. And sometimes the viruses themselves do the killing.

Mutations that allow cells to become malignant—such as the loss of growth control or the loss of key components of the interferon pathways—often make them much more susceptible to viral infection.

Len Seymour, Ph.D.’s group at the University of Oxford takes advantage of that, using adenovirus as a cytotoxic agent against colorectal cancer. On a molar basis, adenovirus is about 109 more potent than chemotherapy agents, and about 105 more potent on the basis of weight, “because the agents can amplify themselves within the target cells, lyse the target cells, and then spread to adjacent cells. It’s a very powerful approach if you can use it properly.”

One problem is that wild-type adenovirus is toxic to the liver, where colorectal cancer tends to metastasize. Attempts to attenuate a wild-type virus so that it is not active in liver cells generally end up attenuating its potency in cancer cells as well, he explains. So why not make a virus that is inactivated by liver cells?

miR-122 is a hepatocyte-specific microRNA that recognizes and inactivates mRNAs with the appropriate binding site. Dr. Seymour’s lab engineered an adenovirus so it contains miR-122 binding sites in its genome and found that “the virus is completely neutralized when it gets into hepatocytes, but still retains the activity of the wild-type virus when it gets into tumor cells.”

Dr. Seymour has also been using forced evolution to develop adenoviruses specific for colorectal cancer through a spin-out company he founded, Hybrid Biosystems (now PsiOxus Therapeutics) in collaboration with Schering (now Bayer Schering Pharma). “It has lost the components it needs for normal cells, but in tumors the mutations tend to complement those deletions,” he explained.

The Class B adenovirus—for which humans do not have natural antibodies—is highly active in tumors, but “has no obvious activity in normal cells.” Clinical trials for hepatic metastatic colorectal cancer will begin later this year, with the virus being injected directly into the hepatic artery of patients.

Sometimes Smaller Is Better

The immune system can inactivate a virus before it ever gets a chance to reach its target. But even more worrisome for the gene-delivery field is an adverse immune reaction to the target itself, generated against proteins encoded by the virus.

There’s a growing body of evidence that different adeno-associated viruses (AAVs), when delivered through the bloodstream, can introduce genetic material to a range of tissues, depending on the particular capsid of the AAV and the expression characteristics of the gene.

The University of London’s George Dickson, Ph.D., works with a serotype that infects skeletal muscle, liver, and other tissues—the object being to correct the function of, or replace, a Duchenne muscular dystrophy (DMD) patient’s faulty dystrophin gene. He carefully chose “a muscle tissue-specific expression system rather than a ubiquitous CMV-based promoter,” he explained, because the latter “would result in the transgene being expressed in tissue antigen-presenting cells—dendritic cells or macrophages and so on—and provoking an immune response.”

Because dystrophin is a large gene and AAV has a somewhat diminutive capacity, Dr. Dickson uses microdystrophins, which are “much shorter, engineered cDNAs that encode smaller dystrophin forms, which are nevertheless highly biologically active.” He also makes sure that the gene sequence is optimized for codon usage, translatability, and mRNA stability.

Another arm of Dr. Dickson’s DMD work—done as part of the MDEX Consortium—tackles stability by utilizing morpholinos. These essentially inert oligonucleotides, delivered through the bloodstream, act as antisense mRNA to inhibit the expression of particular X-linked dystrophin exons.

“If those exons have permutations, or are involved in induction of frameshifts in mutated genes, then in many cases the product of the mutated genes can be re-expressed.” Exon skipping, as it’s called, can result in an albeit slightly mutated or a shortened, but significantly biologically active, protein.

Given the range of different mutations in the gene that gives rise to DMD, only about half of affected patients can even theoretically be corrected by exon skipping. “For the remaining 40–50% it seems more than likely they will need a conventional gene-replacement strategy using something like an AAV vector,” Dr. Dickson said.


A recombinant microdystrophin (red) is double stained for intracellular Ig (green), which identifies muscle fibers damaged by dystrophin deficiency. [University of London]

Cut It Out

Gene therapy can be used against viruses as well. Cellectis Therapeutics (a subsidiary of Cellectis) uses meganucleases engineered to precisely disrupt integrated viruses such as HIV, HBV, and HSV.

Meganucleases (also known as homing endonucleases) target a specific, statistically unique, sequence—generally at least 14 base pairs—creating a double-strand break in the DNA. The cell then attempts to repair the break, often with some minor mutation or loss of sequence.

Cellectis’ most advanced meganuclease project targets an essential gene in the herpes simplex virus. “We were able to show that if we take cells and introduce a plasmid that expresses a meganuclease that targets the viral genome, and then we infect those cells with an HSV virus, we can decrease the infectivity of that virus,” related Julianne Smith, Ph.D., head of the meganuclease recombination system group. “The number of infected cells significantly decreases when the meganuclease is present.”

If there is a homologous sequence present, the cell may attempt to repair the break by homologous recombination. In a therapeutic context, “you could go in and actually create a double-stranded break and introduce a plasmid that contains wild-type information, and go on to correct the mutated allele at its genomic location.”

To test this idea, Cellectis developed a series of meganucleases targeting the RAG-1 gene (involved in a form of severe combined immunodeficiency), and used them to demonstrate homologous recombination between the chromosomal locus and a repair plasmid.

The efficiency by which meganucleases get into cells and do their job is about the same as other means of gene therapy. The important difference is the safety aspect, said Dr. Smith. “If we can specifically target a genomic site, we could avoid some of the serious adverse effects that have been observed concerning random insertions and activation of adjacent genes.” Similarly, random insertions are also subject to gene silencing through gene extinction, but with precise targeting you can select your locus.

In certain cases, it’s even possible to use meganucleases to generate an event in a cell that couldn’t be done any other way, she added, such as cleaving the genome of a virus or perhaps even knocking out a gene. Most antivirals are based on blocking replication, not on eliminating the genome.


Cellectis is developing various genome engineering applications using meganucleases. The process involves introducing into the cell a meganuclease that is specific to the targeted site and inserting a gene with the attributes required to stimulate homologous recombination. The meganuclease breaks the DNA molecule and the homologous recombination system corrects this break by taking as a model the gene introduced at the same time as the meganuclease. [© Frank Beloncle]

Image-Guided, Convection-Enhanced

Precise delivery also is of paramount concern to the University of California, San Francisco’s Krystof Bankiewicz, M.D., Ph.D. He too is worried about causing toxicity as well as engendering adverse immune reactions. Specifically, his work focuses on delivering therapies—whether genetic or pharmaceutical—to specific regions of the brain for the treatment of cancers and neurological disorders.

Typically the location of an intracerebral injection was imprecise and, because of the pressure in the brain, there was a propensity for therapeutics to travel along the needle and back to the surface. Dr. Bankiewicz’ group developed a delivery system that utilizes MRI to guide, in real-time, a small-diameter silica and ceramic cannula to exactly where the therapy is needed.

By gently increasing the pressure at the tip of the cannula, the extracellular fluid is pushed away, “allowing us to infuse pretty large volumes of the fluid that find a space now because it displaces fluid that is already in the brain,” he explained. A tracer in the injection allows it to be tracked. “Basically any structure within the brain”—even large volumes of the brain—can be covered at will by viral particles in a “phenomenally” precise and predictable manner.

It’s not as simple as just watching as a needle goes into the brain, squirting, and watching as the injection magically reaches its target, though. The delivery system itself is like one leg of a three-legged stool. Dr. Bankiewicz has spent years studying things like the movement of cerebrospinal fluid. “It’s being driven by blood vessels causing peristaltic action for the perivascular space.”

“It’s highly predictable and well described how the fluid circulates within the brain.” It’s also of vital importance to know what the viral vector itself is doing: AAV has multiple serotypes, utilizing different trafficking patterns and with different propensities to transduce different cell types. “We have developed a full understanding of how specific serotypes behave in the brain, and how much transgene—gene product—is being delivered, and where it goes.”

The main theme is that if you don’t know how to deliver the drug, it is never going to work, Dr. Bankiewicz summarized. “It will be all over the map—it is never going to be consistent. You have to have a technology that is very precise, very predictable, and reproducible.”

Previous articlePerkinElmer Takes Over Chemagen Biopolymer-Technologie
Next articleResearch Implicates Hair Follicle Stem Cells in Wound Healing-Related Basal Cell Carcinoma