Can toxic genes be blocked to treat disease?
Since the 1990s, gene therapy - the insertion of functional genes to compensate for nonfunctional ones - has been the goal of researchers working in several muscular dystrophies, spinal muscular atrophy, Friedreich's ataxia, metabolic muscle diseases and myotubular myopathy.
These diseases have something in common: The genetic mutations that cause them do so by leading to a lack of a functioning muscle protein. The solution, though technically fraught with complications, is conceptually simple: Add genes that can make the normal muscle protein and compensate for its loss in muscle cells.
But there are other diseases in which the solution isn't a matter of adding something. Rather, it's blocking or removing a gene or protein that has a toxic effect. In disorders such as myotonic dystrophy, oculopharyngeal muscular dystrophy, one form of Emery-Dreifuss muscular dystrophy, one form of amyotrophic lateral sclerosis, spinal-bulbar muscular atrophy, some forms of Charcot-Marie-Tooth disease, and several myopathies, the problem is clearly what scientists call a "toxic gain of function." In these disorders, a gene takes on a toxic function that can't be overcome merely by adding more of its normal counterpart.
These diseases are generally inherited in a dominant pattern, meaning a mutation inherited from only one parent is enough to cause symptoms.
In these and other conditions, scientists are hoping to use new strategies, known as RNA interference (RNAi) and antisense, to block or destroy genetic material that causes disease.
"The beauty and power of RNAi is that it gives us the opportunity to design a drug that can turn off a specific gene, any specific gene," says Henry Paulson, an associate professor of neurology at the University of Iowa College of Medicine.
In at least two neuromuscular diseases, clinical trials of drugs that block RNA are slated for this year or are under way.
In 1993, MDA-supported researchers identified the gene for superoxide dismutase 1 (SOD1) as a cause of an inherited form of amyotrophic lateral sclerosis (ALS), a disease that ravages the muscle-controlling nerve cells of the brain and spinal cord and usually results in paralysis and death within a few years. Although only some 2 percent of ALS cases have been traced to a mutation in the SOD1 gene, toxic SOD1 RNA provides an obvious target.
|If triggering RNAi isn’t the goal, scientists can introduce a single strand of RNA that forms the antisense for the gene they want to block.|
|Single-stranded antisense that sticks to an existing piece of RNA doesn’t attract the RNAi silencing complex. It can either stay on the targeted sense RNA and block it to silence it; or, the whole sense-antisense molecule can be destroyed by an enzyme. The process doesn’t appear to be self-perpetuating.|
"If you're going to think about silencing a gene, it makes the most sense to silence SOD1," says Timothy Miller, a neurologist and neuroscientist at the University of California-San Diego. "It's abundantly clear from rat and mouse models that if you have less SOD1, the disease is less severe, or they're less likely to get the disease."
Miller, with colleagues Richard Smith and Don Cleveland at UC-San Diego, and Isis Pharmaceuticals in Carlsbad, Calif., has been blocking SOD1 in rodents and is planning to try it in patients as soon as possible. People with ALS have one normal and one abnormal copy of the SOD1 gene, so one obvious question is, will the antisense target both? And if it does, is that dangerous?
"I don't think there's a danger in reducing SOD1," Miller says, "in part because knocking out SOD1 in mice resulted in surprisingly little phenotype [symptoms].
"With antisense therapy, and with most gene downregulation therapy, you're not going to reduce the protein level to zero. It will more likely be about a 50 percent knockdown, and that may be a very appropriate level to decrease toxicity of SOD1 but still maintain enough of function of the enzyme."
Miller, whose application for MDA funding is being reviewed, says he's not planning to use RNAi, at least initially, in people with ALS, because it could have unwanted effects that might be hard to control or shut off.
Instead, he'll use small, laboratory-designed compounds that resemble antisense RNA and can block SOD1 RNA, but aren't expected to trigger the more wide-reaching RNAi mechanism or a potentially deadly antiviral defensive action.
These antisense pieces will have only a transient effect and will constantly have to be replenished, but Miller considers the gain in safety to be a good tradeoff for the loss of treatment duration.
In a trial planned for later this year, he plans to infuse the synthetic antisense molecules into the fluid surrounding the spinal cord, using electronic pumps that he says can be "turned down, turned up or turned off." Other researchers are working on making RNAi easy to shut off, too, he says, "but the technology is not there yet."
The story of myotonic dystrophy, a multisystem disease whose hallmarks are weakness and inability to relax muscles at will, is a little like the SOD1 ALS story but with some interesting twists.
When the mutation underlying type 1 MMD was identified in 1992, and found to be in a gene called DMPK, scientists started experimenting to see whether the DMPK protein was overproduced, underproduced or toxic. But in this disease, none of those things proved to be true. Instead, the toxicity appears to be in the RNA itself.
In MMD, overly long, repeated sequences of DNA lead to the production of very long strands of RNA. In MMD1, the repeats are in the DMPK gene, whose function has remained uncertain. In MMD2, the repeats are in the ZNF9 gene, but the diseases are very similar.
This similarity between the two kinds of MMD, despite their being caused by mutations affecting two very different genes, has led scientists to believe that it's the overly long pieces of RNA, which get hung up in cell nuclei and interfere with their operations, that cause the disease.
The RNAi mechanism would theoretically destroy the extra RNA and treat the disease, but there's one major problem: It's only been studied outside the nucleus, and there's no certainty that it can occur inside it.
Molecular biologist Mani Mahadevan, an MDA grantee at the University of Virginia in Charlottesville, says he isn't deterred. He's conducted experiments with cells in a lab dish that suggest that RNAi or a similar process does in fact occur inside the nucleus. When he used an RNAi strategy against RNA that was retained in the nucleus and contained the recipe for a blue protein, "most of the cells were not blue anymore. That means it got into the nucleus.
"If you're going to treat a disease that's caused by a poison, the simplest way to treat it is to get rid of the poison," Mahadevan says of the MMD-causing RNA.
He's considering using "short hairpin" RNA pieces that are like the double-stranded RNAs that cells naturally make to start the RNAi process, and letting the cells continue the process from there.
Short hairpin RNAs are slightly longer (about 80 nucleotides) than the double-stranded RNAs that actually lead to RNAi (which have nucleotide lengths in the 20s), but shorter than those that lead to an antiviral response, Mahadevan says. These, he says, "are supposed to be more efficient, because they're using the cell's own machinery."
For Nan Sook Lee at City of Hope National Medical Center in Duarte, Calif., the existence of RNAi in the nucleus also seems likely, although she says this hypothesis "goes against conventional theory."
Lee has had MDA funding to study RNA in myotonic dystrophy and has worked closely with MDA grantee Jack Puymirat of Laval University in Quebec on this disease. She wants to study nuclear RNAi and says myotonic dystrophy cells are the perfect place to do it.
If the normal DMPK protein proves essential, she says, you could theoretically give some DMPK protein back to the patient. "RNA interference cannot attack the protein," she says, "so you can do these simultaneously."
Although it isn't clear that people need DMPK, Mahadevan would like to develop hairpin RNA sections that target only the abnormal DMPK RNA. Since this RNA is concentrated in the nucleus, while the normal RNA is only there transiently before it moves to the cytoplasm, a hairpin RNA that stays in the nucleus would mostly hit the abnormal RNA. That kind of strategy, he says, "would be ideal."
Using antisense or RNAi against toxic genes isn't limited to fighting dominantly inherited diseases, although they're certainly good candidates for it.
One creative way to use antisense has been developed by Ester Neurosciences of Herzliya, Israel, with Hermona (Mona) Soreq at the Hebrew University of Jerusalem.
This strategy targets the RNA for acetylcholinesterase (AChE), an enzyme that normally breaks down acetylcholine, a chemical carrier of signals from nerve cell endings to muscle fibers.
In people with various forms of myasthenia, the acetylcholine signal is disrupted in some way, and debilitating weakness results. Sometimes the immune system attacks the landing site on muscle fibers that's supposed to receive it, resulting in myasthenia gravis. Sometimes it attacks the nerve ending that's supposed to release it, resulting in Lambert-Eaton syndrome. And in other cases, a genetic flaw disrupts either the sending or reception of acetylcholine.
What most myasthenias have in common is that they can be treated to some extent by making a patient's acetylcholine last longer. You can do that by slowing its breakdown by AChE; in fact, the drug pyridostigmine (Mestinon) does just that, by targeting the AChE protein.
The drawback is that the body doesn't like having a protein it considers useful destroyed. When medications are given that do this, the body fights back by making more of the protein, creating an escalating war between the medication and the body.
The Ester Neurosciences drug, Monarsen, is a short, synthetic and chemically protected antisense compound that's designed to match up perfectly with its target, the RNA for AChE, and keep the enzyme from being made. It's now being tested in a clinical trial in people with myasthenia gravis in the United Kingdom and Israel.
|To trigger the RNA interference (RNAi) mechanism, small pieces of double-stranded RNA (a sense strand stuck to an antisense strand for a specific gene) are introduced into a cell||The double-stranded RNA molecule attracts the RNA interference silencing complex (RISC).|
|The RISC causes the double-stranded RNA to unwind. Only the antisense strand stays attached
to the RISC.
|The RISC and its antisense cargo seek and land on the target gene, which is a sense strand RNA.|
|The RISC and its antisense cargo cut the target RNA at a precise point.||An enzyme destroys the targeted sense RNA, thereby blocking the gene. The RNAi process can be self-perpetuating.|
In 1991, Albert La Spada was a predoctoral student working in the laboratory of MDA grantee Kenneth Fischbeck at the University of Pennsylvania studying spinal-bulbar muscular atrophy (SBMA), also known as Kennedy's disease, a disorder that resembles ALS but is more slowly progressive and less severe. In addition, it has an X-linked inheritance pattern and only affects males.
"I worked pretty independently on the SBMA project," he says, "and I made a rather unexpected and exciting discovery."
The discovery was that SBMA is caused by what was at the time an unknown type of genetic mutation - repeated groups of three nucleotides (triplet repeats), leading to the presence of extra molecules of the amino acid glutamine in the final protein. In this case, the final protein is the androgen receptor, a cellular docking site for male hormones (androgens). The extra glutamine molecules cause it to lose some of its normal functions and exhibit a toxic gain of function.
SBMA would be the first triplet repeat and the first polyglutamine (extra glutamine) disorder to be identified. (A year later, MMD1 would become the third identified triplet repeat disorder, although these triplets don't result in excess glutamine. MMD2 is caused by a quadruplet repeat mutation identified in 2001.)
La Spada, now an independent researcher at the University of Washington-Seattle who has had five years of MDA funding to study SBMA, says he became interested in RNAi starting about 2002.
"Once it became clear that you could use the technology in higher organisms, mammalian organisms, then that's when we realized that it could be applicable to disorders that involved dominant mutations in neurological diseases; and the polyglutamine repeat diseases certainly fall into that category."
A turning point came in 2004, when Beverly Davidson at the University of Iowa and colleagues showed that RNAi could be used to improve cellular health and disease symptoms in a mouse with type 1 spinocerebellar ataxia (SCA1), another triplet repeat, polyglutamine disorder that causes neurodegeneration. The paper was "an important proof of principle," La Spada says, that caused him to start thinking about using RNAi in SBMA.
His enthusiasm, however, is tempered by a major stumbling block specific to this disease: Unlike DMPK and SOD1, the androgen receptor gene is necessary for normal functioning in males. And, to make matters worse, it's on the X chromosome, and men only have one of those. Anything that knocked out the mutated androgen receptor RNA would, by definition, also remove any remaining normal function the receptor might have.
Androgen receptors, even in men with SBMA, seem to retain some of their normal, necessary properties, so getting rid of them altogether doesn't seem like a good idea.
"With SBMA," La Spada says, "you're stuck with the fact that when you knock down the disease-causing [gene], you're also at the same time knocking down [remaining] normal function."
To overcome that problem, he's considered two possible scenarios. One is "some sort of knock-down strategy that would be supplemented by delivering stem cells with a normal androgen receptor gene." Another would be to deliver RNAi only to the cells affected by androgen receptor toxicity in this disease, leaving the partially functioning androgen receptors in other tissues, where they apparently aren't terribly toxic.
"I think that even though there's a certain degree of uncertainty, it deserves to be considered and pursued. It's something that we haven't decided to immediately pursue, but something that we're interested in, and we have a wonderful mouse model to do it in."
"Just because it works in mice doesn't mean it's going to work in humans," La Spada says of RNAi. "There has to be a study that says we can develop it, refine it and make it work in humans."
Step one, he notes, is making sure you're clear on what cells are affected in a particular disease and what RNA you need to target.
Step two is making sure your weapon (RNAi or antisense) isn't going to have "off-target" effects, that is, landing on RNA that's exactly or almost exactly like the sequence you're actually targeting.
"If you do the math," La Spada says, "the odds are limited" that you would hit a stretch of RNA that's exactly like the target. But, he cautions, "the sequence is sufficiently small that you could get partial matches, and that's where the concern about off-target effects comes in."
A third question is how gene silencing compounds can be safely and effectively delivered to human cells. "Whenever you use the RNAi approach, you have to proceed with considerable caution," La Spada says.
For now, the best way to administer RNAi is to package DNA instructions for it into an altered virus that will then keep churning out RNAi molecules. The adeno-associated virus, which is being used to insert therapeutic genes in two muscular dystrophy trials, is appealing as a delivery vehicle, he says, "because it doesn't elicit as much of an immune response as other types of viruses, it's easy to manipulate, and it doesn't seem to be associated with toxicity in humans."
But he suspects there are other safer, effective ways to administer RNAi molecules, such as packaging them into a synthetic structure.
Even with all the caveats, it's hard not to get excited about gene silencing strategies. Neuroscientist Henry Paulson, whose University of Iowa lab has studied mechanisms of dominant diseases for several years, puts it this way: "Once it became clear that there was an elegant and powerful way to turn off genes, it was a no-brainer to try to turn off dominantly active genes, instead of jumping through lots of hoops to target their downstream effects."
Since 1953, thanks to the work of American scientist James Watson and British scientist Francis Crick (for which they would win a Nobel Prize in 1962), biologists have known that the code of life in all plants and animals is contained in DNA, a molecule in the shape of a twisted ladder - the now famous double helix.
Watson and Crick described the chemical rungs of the DNA ladder as being composed of pairs of four chemicals called nucleotides, each one precisely bound to its matching counterpart.
The next few decades would reveal that the classic DNA double helix, with its pairs of nucleotides, is actually the reference library for the genetic code. A very similar compound, RNA, would be identified as the molecule that contains a moveable copy of each DNA gene. It's these copies, which can be compared to photocopies made from a library reference book, from which a cell gets instructions for making each protein.
Watson and Crick would also learn that the DNA ladder, with its precisely matched pairs of nucleotide rungs, is a relatively inactive structure, with the rungs shielded from their environment. When DNA is replicated during the creation of new cells, or when an RNA copy of a stretch of DNA is made, the vertical pieces of the ladder open, leaving the rungs exposed.
Paired nucleotides, they learned, usually mean a compound is inactive (at least temporarily); unpaired nucleotides signal that a compound is ready to interact with its environment.
If scientists want to keep a gene from making a protein, they can block its RNA by pairing its nucleotides with their partners, thereby closing it off from the cell's environment. A strand of RNA or DNA that sticks to the original strand in this way is called an antisense strand, because it keeps the cell from following (making sense of) genetic instructions.
Today, some researchers are using laboratory-engineered antisense compounds that resist destruction to block specific genes or portions of genes, and some are investigating other strategies for gene silencing.
In the late 1980s, when molecular biology was still in its early years, scientist Richard Jorgensen (now at the University of Arizona in Tucson) and co-workers at DNA Plant Technology Corp. in Oakland, Calif., were trying to create a deep purple petunia flower by introducing extra genes for an enzyme called chalcone synthase into the plant during its development.
What they got, however, wasn't what they expected. Instead of darkening the petunias, their efforts yielded flowers that were either completely or partly white. The researchers thought they'd made some technical error, but they failed to find it.
More than a decade later, in 1998, Andrew Fire at Carnegie Institute of Washington in Baltimore, Craig Mello at the University of Massachusetts in Worcester, and colleagues, purposely set out to block the gene for a muscle filament protein called unc-22 in worms, testing a variety of molecular constructions to see which one would work best.
Much to their surprise, the scientists found that single strands of antisense to unc-22 RNA had only a modest effect in diminishing production of the unc-22 protein. But when they introduced double-stranded unc-22 RNA, it had a profound effect in turning off the gene. The worms treated with the double-stranded RNA had so little of the muscle filament protein that they developed severe twitching.
Fire and Mello speculated that they had identified a potent, natural process of "physiological gene silencing" that could cross cell boundaries and spread throughout the body and that cells must have a biological purpose for such a mechanism. They called it RNA interference, and in 2006, they would be awarded a Nobel Prize in Medicine for this work.
Although short antisense strands can be used to block specific genes or even parts of genes, long strands can trigger a potent antiviral defense mechanism in the cells of mammals. When these cells see a long, ladder-shaped portion of RNA, it looks like a virus, because many viruses use double-stranded RNA to reproduce themselves.
To save the whole organism from what they perceive as a spreading viral infection, the cells of higher organisms can shut down almost all protein synthesis and even commit cell suicide.
It's possible that some organisms developed gene-specific RNA interference (RNAi) in response to short, double-stranded RNA as a less drastic alternative than a total cell shutdown for defending themselves in certain circumstances against invasion by viruses.
Using this more specific defense system, a mammalian cell could theoretically target double-stranded, viral-shaped RNA and any existing RNA that looks like it, while saving the operation of the rest of the cell. It's this mechanism that scientists hope to manipulate to silence toxic genes in human disease.
Jorgensen's petunias turned pink and white instead of deep purple because the newly introduced chalcone synthase gene, for some reason, formed a double-stranded molecule. That caused the plants' cells to destroy not only the new color genes but some or all of the existing color genes. Likewise, Fire and Mello's worms developed twitches when the scientists introduced double-stranded pieces of specifically targeted RNA, because these pieces activated the RNA interference mechanism against the worms' own muscle filament RNA genes.
Once the RNAi mechanism was understood, and it looked possible to avoid triggering a global antiviral defense, scientists be-came tantalized by the idea of harnessing its power to stop disease-causing RNA.