News Synthesis | 22 Oct 2013

Neuromyelitis Optica, Part 3: Treatment and Prevention

After discussing NMO pathogenesis, we’ll talk about current treatments, as well as efforts to find better treatments and even prevent NMO attacks

Andreas von Bubnoff

Toward better treatment and prevention

Because NMO patients essentially suffer from inflammation, doctors first try to stop an acute NMO attack with intravenous steroids. But because long-term use of steroids has serious side effects and is not very good at preventing attacks, they then switch patients to drugs that suppress immune function. Lea Long, the NMO patient introduced in Part 1, for example, received an intravenous infusion of rituximab (Rituxan). Approved to treat lymphoma, it depletes B cells. Other patients receive azathioprine (Imuran) or mycophenolate mofetil (CellCept), drugs that decrease the numbers of B and T cells.

But drugs like Rituxan aren’t very specific, because they suppress all B cells, and azathioprine and mycophenolate mofetil suppress T cells as well. Therefore, they have side effects, such as a higher susceptibility to infection. And even though retrospective studies show that they can reduce relapse rates, they don’t always work: Only about 60% of NMO patients remain attack-free for 1 or 2 years after taking azathioprine or rituximab, Sean Pittock, M.D., an autoimmune neurologist at the Mayo Clinic in Rochester, Minnesota, told MSDF.

It’s also possible to remove antibodies from the blood by plasmapheresis, in which doctors draw a patient’s blood, separate the plasma from the blood cells, and then place the blood cells with antibody-free plasma back in the patient. Doctors usually use plasmapheresis to treat acute attacks that are especially severe and don’t respond well to steroids. But Dean Wingerchuk, M.D., at the Mayo Clinic in Arizona and his colleagues are now also studying whether regular plasmapheresis can prevent attacks.

In addition to these established treatments, two drugs are currently in clinical trials that consist of antibodies that inhibit elements of the immune system involved in NMO pathogenesis.

The first, eculizumab, contains an antibody to the protein C5, a part of the complement system. Produced by the company Alexion, eculizumab is approved to treat paroxysmal nocturnal hemoglobinuria, a rare disease in which red blood cells are destroyed.

Because the complement system is involved in the death of astrocytes in NMO patients, Pittock reasoned that it might also prevent NMO attacks. To study that idea in an early clinical trial, he chose 14 patients who were the “sickest of the sick,” he said: All had had at least two attacks in the 6 months before the trial, or at least three attacks in the preceding year, and previous treatments had failed for half of them. Pittock found that eculizumab treatment inhibited severe attacks for 1 year in all patients (Pittock et al., 2013). Only two, he said, complained about back pain and blurry vision, which could mean that they had mild attacks.

“It went very well,” Pittock said. “I was very impressed and actually quite surprised by the outcome.” One issue, however, is that the drug increases the risk of meningitis. Another is cost: The treatment’s price tag is $300,000 to $400,000 a year.

Jeffrey Bennett, M.D., Ph.D., a neurologist at the University of Colorado School of Medicine in Aurora, who was not involved in the study, calls these results very encouraging but adds that eculizumab inhibits complement activation but not antibody-dependent cellular cytotoxicity, which means that over time there might still be ADCC-related damage of the aquaporin antibody.

The second drug, tocilizumab, is approved to treat rheumatoid arthritis and contains an antibody that inhibits a sensor on the surface of certain immune cells that allows them to respond to a signaling molecule called IL-6. Two years ago, Japanese researchers, led by Takashi Yamamura, M.D., Ph.D., a neurologist and immunologist at the National Center of Neurology and Psychiatry in Tokyo, found that inhibiting this IL-6 sensor reduced the survival of plasmablasts—the cells that make the aquaporin-4 antibodies—that had been taken from NMO patients (Chihara et al., 2011).

Because tocilizumab inhibits the IL-6 sensor, Yamamura reasoned that it might reduce aquaporin-4 antibody secretion in NMO patients and therefore prevent relapses. To test that hypothesis, the Japanese team selected seven NMO patients, who had experienced between two and five relapses in the year prior to the start of the study and had not responded well to conventional treatments, for the first clinical tocilizumab trial in NMO patients. The trial is ongoing, and so far the researchers have found that a monthly intravenous tocilizumab injection could reduce the concentration of aquaporin-4 antibodies in the patients’ blood. Now that all seven patients have been treated for more than 12 months, only two of them have had a minor relapse.

The monthly injections also improved symptoms from previous attacks: A woman in her late 30s, who before the study couldn’t walk for more than 100 meters because of serious pain in her legs, can now walk for more than 1 kilometer, Yamamura told MSDF, adding that she is also less tired, an effect of the drug that’s already known from rheumatoid arthritis patients. And a 28-year-old man who went blind several years ago and needed assistance to walk around in the hospital before the treatment started can now slowly walk without any help.

For the most part, the drug is safe, Yamamura said. Still, as with most other approaches, there is a risk of infections, again because the effects of tocilizumab are unspecific: Tocilizumab reduces the formation of all types of plasmablasts, not just the ones that make aquaporin-specific antibodies.

This lack of specificity is a major drawback of approaches like Rituxan, eculizumab, or tocilizumab, said Lawrence Steinman, M.D., a neurologist and immunologist at Stanford University. He calls them “sledgehammers,” because they inhibit the immune system in general and not just aquaporin-4 antibody function or production. This is why they often have side effects such as a higher risk of infections. “It is with a great deal of frustration [to] see that more and more energy is put on sledgehammers,” he said, adding that instead, researchers should develop drugs that specifically inhibit the effect of aquaporin-4 antibodies. “This is staring us in the face,” he said. “Why don’t we try something much more reasonable [rather] than merely getting rid of all the complement cascade or all of the B cells?”

For his part, Steinman is studying whether vaccination to aquaporin 4 can prevent NMO attacks, by training the immune system to tolerate aquaporin 4. The vaccination will involve weekly intramuscular injections of aquaporin-4-coding DNA. Some of the nucleotides (the building blocks of this DNA) are modified: Short six-nucleotide-long stretches with the nucleotide pair CG—which usually induce a pathogenic immune response—are changed so that they contain GG instead. Such an immunization, Steinman said, should train CD4+ T cells to ignore aquaporin 4 and to stop attacking astrocytes that express it. (Steinman has recently used this approach to reduce proinsulin-specific immune responses in people with type 1 diabetes, in which proinsulin immune responses are thought to contribute to the disease.) Compared with other treatments researchers are currently exploring, Steinman said, such a vaccine would have the main advantage that it is more specific and therefore has fewer side effects.

Bennett and Alan Verkman, M.D., Ph.D., a clinician-scientist at the University of California, San Francisco, are working on another aquaporin-specific approach: aquaporumab, a modified monoclonal aquaporin-4 antibody that still binds to aquaporin 4 but is unable to activate the complement system or induce ADCC. Binding of aquaporumab to aquaporin 4, the researchers hope, can therefore keep the harmful, pathogenic aquaporin-4 antibodies away from their target. And indeed, a recent study suggests that the approach works, at least in animals: When the researchers injected serum from an NMO patient—which contains aquaporin-4 antibody—together with human complement into the brain of mice, they were able to induce NMO-like lesions, but coinjection of aquaporumab almost completely kept the lesions from forming (Tradtrantip et al., 2012).

Because aquaporumab is specific, it’s less likely to have side effects, Bennett told MSDF, and could therefore be used to treat acute attacks and be given continually to prevent them. “This is the only nonimmunosuppressive medication that’s currently out there for potential treatment of this disease,” Bennett said. “You are competing against just those antibodies that are causing the damage and have a therapy that’s competitive but not immunosuppressive and very specific for this disease.”

If everything goes well, Bennett and Verkman hope to move aquaporumab into phase 1 clinical trials within the next few years.

Steinman, who is not involved in the project, is enthusiastic. “It gives a fantastic opportunity to try antigen-specific therapies for this disease,” he said.

But others are skeptical. It’s unclear, Lennon and Pittock both told MSDF, how aquaporumab would cross the BBB. But Bennett doesn’t think that’s an issue. “If whatever it is that initiates an attack opens up the blood-brain barrier for the peripheral antibody that’s causing the disease, it’s also going to be opening up the blood-brain barrier for the inhibitory antibody that’s circulating with it to block it,” he said.

Even if aquaporumab can enter the brain, Lennon said, it could be harmful, because her observations (which Bennett and Verkman say they can’t confirm) suggest that, because the aquaporumab antibody can bind aquaporin 4, it might block the water channel or cause it to internalize, which would cause swelling.

And Vincent is somewhat skeptical for another reason: Similar attempts to develop therapies for other antibody-mediated diseases such as myasthenia gravis were unsuccessful. “About 20 years ago,” she said, “my colleagues tried very hard to find specific ways of treating myasthenia, and it never came to anything.” One likely problem, she said, is that different NMO patients have antibodies that target different parts of aquaporin 4, and aquaporumab might not be able to inhibit the binding of all of them.