The Paradox of Daclizumab

Three might be the lucky number for daclizumab, an investigational drug for relapsing-remitting multiple sclerosis (RRMS) with broad and unexpected immune effects.

Under the brand name Zinbryta (Biogen Idec/AbbVie), a newly patented version of the drug, called daclizumab high-yield process, or DAC HYP, recently captured headlines after researchers presented more efficacy and safety results from the phase 3 clinical trial (Kappos et al., 2014).

The unpublished results seem to sustain expectations that daclizumab may be as effective as the best MS drugs now approved but with lower risk (Pfender and Martin, 2014). Meanwhile, researchers have identified at least three ways the drug works to counter MS, a finding that is illuminating new details in disease pathology and may lead to novel therapeutic strategies (Bielekova, 2013; Pfender and Martin, 2014).

The phase 3 study, called DECIDE, is “a pivotal trial,” said Ellen  Mowry, M.D., M.C.R., a neurologist at Johns Hopkins University in Baltimore, Maryland. She cited the notable drop in disease activity as measured by relapses, MRI, and disability progression, compared to a popular older drug, interferon β-1a (Avonex, Biogen). She also noted daclizumab’s signature skin side effects as well as a small increase in serious infections.

Daclizumab “deserves to be approved and to be one of the options” available to people with MS and their physicians, said Ludwig Kappos, M.D., chair of neurology and head of the MS Research Group at University Hospital Basel, Switzerland, and lead investigator for DECIDE. The relative effectiveness of the drug could be “at least” at the level of fingolimod (Gilenya, Novartis) or dimethyl fumarate (Tecfidera, Biogen), Kappos speculated in an interview with MSDF. He cautioned that results of this and other drug studies cannot be compared directly, due to differences in design, selection criteria, and other details.

Kappos presented the DECIDE findings earlier this month at MSBoston2014, shorthand for the joint meeting of the Americas and European committees for treatment and research in multiple sclerosis (ACTRIMS-ECTRIMS), held September 10-13 in Boston. In the closing session, Mowry cited the findings as a clinical research highlight of the meeting.

Other meeting reports confirmed and extended some of the surprising ways daclizumab pulls the strings of the immune system. One of the new mechanisms of action has shown promise as an early biomarker of how the drug is working.

In another finding from the meeting, an analysis of samples from blood and cerebrospinal fluid suggests that daclizumab nudges the abnormal numbers of innate and adaptive immune cells in RRMS back to more normal physiological levels found in people without MS (Lin et al., 2014).

Top-level clinical findings

The double-blind phase 3 DECIDE trial pitted monthly injections of daclizumab against weekly interferon injections in people with RRMS for a minimum of 2 years. Each group received placebo injections of the other drug. A little more than 1,800 people started the 2-year DECIDE study. About two-thirds were women. On average, participants had had RRMS for about 4 years and mild disability of about 2.5 on the Expanded Disability Status Scale. About 30% dropped out for various reasons, mostly because of adverse events and lack of efficacy.

In the main finding, daclizumab reduced the annualized relapse rate by nearly half compared to interferon. Translated to an evidence-based medicine figure called “number needed to treat,” each patient on daclizumab had a 1-in-5.6 chance of having fewer relapses. In secondary endpoints, several brain imaging techniques also showed about half the new and enlarging lesions in the daclizumab treatment arm compared to interferon (Arnold et al., 2014).

Many of the top findings were reported in a press release in June. A new 6-month analysis presented at the meeting showed a slightly lowered risk of disability progression. Another new graph showed a reduction, although not statistically significant, in patient-reported physical decline (Kappos et al., 2014).

Skin and other side effects

The opposite trend was true with daclizumab’s distinctive side effect: skin adverse events, mostly rash and inflammation. More than one-third of people on daclizumab reported cutaneous issues, twice as many as in the interferon group, including some cases severe enough to discontinue the drug. The skin effects were independent of the injection site and were mostly manageable by corticosteroid creams, Kappos said.

Both groups had similar rates of injection site pain, about 10%, and daclizumab came out ahead on influenzalike illness, with 10% compared to 37% of people in the interferon group (Selmaj et al., 2014).

In response to a question from the audience, Kappos said there was no sign of secondary autoimmunity arising from daclizumab use, although “we might interpret the cutaneous events as autoimmune,” he said. “Every intervention with immune-modifying therapy has this risk, but we did not observe additional risk in these cases.” Other adverse events included a higher rate of infections (65% in daclizumab and 57% in interferon), with twice as many serious infections (4% to 2%), and elevated liver enzymes.

Two deaths in earlier daclizumab trials had raised concerns about secondary autoimmune disease, which is a serious side effect of another therapeutic monoclonal antibody, alemtuzumab. In a 1-year phase 2 extension study, SELECTION, a person on daclizumab died of autoimmune hepatitis. In the phase 2 study SELECT, one person died of complications of psoas abscess after recovering from a serious rash. In both cases, the experimental drug could not be ruled out as a contributing factor (Barkhof and Ciccarelli, 2014; Pfender and Martin, 2014).

In the phase 3 DECIDE study, there were four deaths in the interferon group and one death in the daclizumab group, none of which was considered treatment related. The death in the daclizumab group was from an MS attack in the brainstem 4 months after the patient stopped the experimental treatment and withdrew from the study. DECIDE researchers reported no cases of the life-threatening side effect, progressive multifocal leukencephalopathy, which has been associated with natalizumab, another potent therapeutic monoclonal antibody.

“A rigorous safety monitoring program will need to be implemented should daclizumab be prescribed in clinical practice,” advised an editorial (Barkhof and Ciccarelli, 2014) that accompanied the 1-year phase 2 SELECTION extension findings in March (Giovannoni et al., 2014). Longer-term experience with more people is needed to evaluate the risks for rare adverse events, agreed the authors of another review, who also termed daclizumab a “well tolerated therapy option in MS that is safe in the vast majority of patients” (Pfender and Martin, 2014).

Biogen Idec in Cambridge, Massachusetts, and AbbVie in North Chicago, Illinois, are jointly developing the drug. They plan to file marketing applications for daclizumab (Zinbryta) with regulatory authorities in the United States, Europe, and other jurisdictions during the first half of 2015, said Gilmore O’Neill, vice president for MS research and development at Biogen.

Namesake natural killer cells

One analysis was notable for its absence at the ACTRIMS-ECTRIMS meeting. In phase 2 trials, the therapeutic efficacy of daclizumab paralleled the expansion of a key target cell population, called CD56bright natural killer cells (CD56bright NK). That raised hopes for a biomarker that could measure who was responding and predict who would do best.

“In this study, the [CD56bright NK expansion] response was seen in nearly all patients,” dashing hopes for its utility as a biomarker, Kappos told MSDF in an interview. But investigators collected other potential biomarkers during DECIDE, he said, and further analysis of them may help discern which patients would benefit.

The phase 2b studies provided a “comprehensive view of how the drug’s levels affected the size of regulatory T cells and CD56bright NK cells,” O’Neill said in an interview with MSDF. O’Neill cited several posters at ACTRIMS-ECTRIMS detailing the immunological responses in studies of daclizumab’s safety and effectiveness (Mehta et al., 2014; Othman et al., 2014; Huss, Fragoso et al., 2014; Huss, Mehta et al., 2014).

The next big thing for Biogen and other companies is “understanding how you can use biomarkers to predict how aggressive someone’s MS is going to be,” O’Neill told MSDF. “Layered on that, we’re interested in how you can predict which drug is better for which patient. We believe there are different mechanisms driving disease in different patients.”

Daclizumab’s unexpected actions

Daclizumab is one of a long line of drugs to be repurposed in MS. The list includes natalizumab (Tysabri, Biogen/Elan), alemtuzumab (Lemtrada, Genzyme), dimethyl fumarate (Tecfidera, Biogen), fingolimod (Gilenya, Novartis), and teriflunomide (Aubagio, Genzyme) (Schmierer, 2014).

The credit for developing daclizumab in all its uses goes to researchers in several labs at the U.S. National Institutes of Health (NIH). Daclizumab began 30 years ago as a promising mouse molecule for T cell leukemia for its ability to block activated inflammatory T cells (Bielekova, 2013; Pfender and Martin, 2014). By the same presumed mechanism, the humanized antibody was approved as an add-on therapy for preventing rejection of solid organ transplants and marketed as Zenapax by Hoffmann-La Roche until the company withdrew the product for commercial reasons in 2008 in the European Union and in 2009 in the United States.

It made perfect sense to try the drug in MS, given all the evidence that autoreactive T cells drive the disease. “It appealed to us for all the wrong reasons,” said Bibiana Bielekova, M.D., who started working on the drug as a fellow and has continued to work out the mechanisms in her own lab as an NIH investigator.

Bielekova sat down with MSDF at ACTRIMS-ECTRIMS to tell her 15-year story of tracking down daclizumab’s unexpected effects on the immune system and its implications for understanding MS and future therapeutics.

“It’s a typical story of discovery,” she said. “We think one thing. We are wrong. By being really good observers, we notice something and get it right. But we are wrong most of the time.”

The way it works

In 1999, Bielekova and her colleagues opened a pilot study of daclizumab in their patients for whom interferon was not working. By 6 months, contrast-enhancing MRI showed a dramatic 80% reduction in new lesions, and clinical symptoms stabilized. In an extension of the study in which interferon was withdrawn, they observed that daclizumab alone could sustain the benefits, although some people experienced a greater synergistic effect from taking both drugs together. A subsequent open-label phase 2 study showed comparable results with daclizumab as the first-line treatment.

In sharp contrast to the robust clinical demonstrations, the laboratory work was one negative finding after another. At first, “we were trying to prove the hypothesis that the drug was inhibiting activated T cells,” Bielekova said. “We could not demonstrate any effect on T cell functions.”

The original idea, still included in reviews and posters on the drug, was that daclizumab directly shut down inflammatory T cells. It grabs on to part of a high-affinity receptor for the inflammatory cytokine interleukin-2 (IL-2), effectively getting in the way without turning it on. The targeted receptor piece is called CD25 and is plentiful on effector T cells. So, the reasoning went, daclizumab must bind to autoreactive T cells and block their activity.

It made sense to Bielekova too. After all, she and her colleagues fed IL-2 to grow T cells in culture dishes. If it worked in vitro, why wouldn’t it work in vivo? In the process of trying to prove this, Bielekova found that therapeutic levels of daclizumab in people seemed to do all the wrong things for T cells. Effector T cells could still activate, proliferate, and secrete inflammatory cytokines. Worse, regulatory T cells (Tregs), which suppress inflammation, were significantly inhibited by the drug. Yet, daclizumab obviously was working in people with a T-cell-mediated autoimmune disease. It didn’t make sense to her.

At the end of three frustrating years in the lab, she noticed another population of cells with an intermediate-affinity IL-2 receptor. An Internet search suggested they were a small subset of NK cells called CD56bright, previously called immunoregulatory. Interestingly, Bielekova said, they are known to increase during pregnancy and are thought to have something to do with maintaining tolerance to the foreign body that is the fetus. The rare CD56bright subset normally makes up about 5% to 10% of NK cells, which in turn represent only about 1% of lymphocytes in the blood. In daclizumab-treated patients, the tiny subset of cells expanded up to 500%.

Bielekova formally phenotyped those cells and then tried to figure out what they were doing. Normally, when she added therapeutic amounts of daclizumab to culture dishes of patient blood samples, the T cells could activate and produce cytokines, but they did not survive. In a key experiment, when she extracted the NK cells, the T cells could not only function in the presence of the drug, but they also survived.

“We demonstrated that the natural killer cells were killing autoreactive T cells,” Bielekova told MSDF. “It was difficult to publish. There was a dogma that natural killer cells should not kill autologous T cells.” Later, she and her colleagues showed that the CD56bright NK subset wielded a specialized weapon, granzyme K, not available to their other NK cousins.

The team also showed why and how daclizumab expands and activates the subset. It turns out that CD56bright NK cells have huge amounts of the intermediate-affinity IL-2 receptor, which is missing the CD25 component of the high-affinity receptor blocked by daclizumab. That enables the NK subset to consume the excess IL-2.

The NIH patented daclizumab for MS, based on the new finding that daclizumab deployed innate cells, the expanded CD56bright NK subset, as contract killers of activated T cells. “We were really excited to see that mechanism of action,” Bielekova said. “We felt it was an interesting compound that works very differently than drugs that we have. It was inducing regulation that normally happens in humans in pregnancy and other circumstances.”

Even more interesting was the strong correlation between the expansion of CD56bright cells in a person and the reduction of new contrast-enhancing lesions in the brain. Combined with the relative decrease in T cells, Bielekova and her colleagues found that the cell ratios in the blood could differentiate full responders from partial responders and could be a useful biomarker.

Three strikes and you’re in

Since then, Bielekova and her colleagues have shown two additional ways daclizumab works. One lesson came from a patient who responded well to the drug despite the lack of a CD56bright NK expansion in her blood. To solve this mystery, the investigators turned to other cells that expressed CD25: dendritic cells and monocytes.

After a few false starts, they realized CD25 on dendritic cells is not part of a high-affinity receptor for IL-2. Instead, the component has another job deep within the synapse dendritic cells form with T cells when activating them to fight a pathogen, in the case of infection, or to react to an antigen, in the case of autoimmune response. In a twist, CD25 on dendritic cells doesn't consume IL-2; it spoons the cytokine into the T cell.

“To make a long story short, dendritic cells secrete IL-2, use CD25 to hold IL-2, and feed it to T cells, like a good mother,” Bielekova said. The increased risk for serious infections in people on daclizumab in the phase 3 trial may be due to daclizumab blocking this action. “We’re still chasing the dendritic cell story,” she said.

In one of the wrong turns in studying how dendritic cells behaved in people, Bielekova expected that the annual flu vaccine would not activate those innate immune cells in patients treated with daclizumab. “Unfortunately,” research fellow Yen-Chih Lin, Ph.D., told her after looking at the samples, “you were wrong.”

But Lin pointed out another unknown group of cells that expanded in controls after vaccination but decreased in people on daclizumab. They look like T cells and can produce cytokines, but they are part of the innate immune system and can respond early. They turned out to be lymphoid tissue inducer (LTi) cells.

“In the fetus, they’re important for making lymph nodes,” Bielekova told MSDF. “It’s not clear what [LTi cells] do in adulthood. In animal models, it’s been shown they can still mediate formation of the lymphoid follicles in the gut. People with progressive MS are known to have lymphoid follicles in the meninges. We were wondering if these cells had something to do with MS disease process.”

Indeed, untreated MS patients had significantly higher levels of circulating LTi cells in comparison with healthy controls, they reported. Daclizumab seemed to normalize the LTi cell numbers in the blood. “We think this is another mechanism of action,” she said. She and her team showed that LTi and CD56bright NK cells come from the same precursor cells whose fate is determined by the IL-2 signaling. With the high-affinity receptor blocked by daclizumab, the balance tips toward more of the NK subset.

At ACTRIMS-ECTRIMS, Lin presented a poster showing that LTi cells were also expanded in the cerebrospinal fluid of untreated people with RRMS and that daclizumab nudges the numbers and proportions of LTi and other immune cells back closer to that of people without MS (Lin et al., 2014).

A new treatment paradigm

“Science is not simple,” Bielekova told MSDF. “Nature always has these back loops and controls. You have to step back and see bigger pictures. IL-2 is so important in orchestrating immune response.”

To sum up, daclizumab may end up being a useful therapy for MS, a T-cell mediated disease, for reasons completely different than originally thought. It knocks down activated T cells, but not in the way anyone had assumed when they envisioned the possibilities for people with MS. Instead, it harnesses the innate immune system and the IL-2 signaling pathway. The emerging science has revealed new roles for innate immune cells in MS, as well as the biology of daclizumab’s chief nemesis, the cytokine IL-2.

“There are some earlier studies that described direct inhibitory effect on T cells, but these used non-physiologically high levels of daclizumab,” Bielekova clarified in an email. “Using concentrations of daclizumab achievable in-vivo, we have shown absolutely no direct inhibitory effect on T cell proliferation or cytokine secretion. Daclizumab does inhibit high affinity IL-2 signaling in activated T cells, but this does not lead to their functional inhibition. Instead, it enhances their survival. There is a direct effect on Treg cells. Because these cells are dependent on high affinity IL-2 signaling for their survival and function, we and multiple other groups have shown that they are directly inhibited by daclizumab therapy. Both inhibition of Tregs and enhanced survival of effector T cells should have detrimental consequences for MS disease process, if current paradigms are right. And as you have heard, instead, MS disease process gets better on daclizumab therapy.”

Additional reporting from Dan Keller.

Key open questions

• What new efficacy, safety, and mechanistic information will emerge from extensions of the phase 2 and phase 3 studies of daclizumab in MS?
• What biomarkers will be useful to predict which patients will respond best to daclizumab and other MS drugs?
• How will daclizumab compare in efficacy with other highly effective MS drugs, if tested together in one study?
• What new treatment strategies will emerge from the new understanding of the role of the innate immune system and IL-2 signaling in MS?

Disclosures and sources of funding

The DECIDE study was supported by Biogen Idec and AbbVie Biotherapeutics. Co-authors:

Ludwig Kappos’ institution (University Hospital Basel) received in the last 3 years and used exclusively for research support: steering committee, advisory board and consultancy fees from Actelion, Addex, Bayer Health Care, Biogen, Biotica, Genzyme, Lilly, Merck, Mitsubishi, Novartis, Ono Pharma, Pfizer, Receptos, Sanofi-Aventis, Santhera, Siemens, Teva, UCB, Xenoport; speaker fees from Bayer Health Care, Biogen, Merck, Novartis, Sanofi-Aventis, Teva; support of educational activities from Bayer Health Care, Biogen, CSL Behring, Genzyme, Merck, Novartis, Sanofi, Teva; royalties from Neurostatus Systems GmbH; grants from Bayer Health Care, Biogen, Merck, Novartis, Roche, Swiss MS Society, the Swiss National Research Foundation, the European Union, and Roche Research Foundations.

Krzysztof Selmaj: compensation for consulting services from Genzyme, Novartis, Ono, Roche, Synthon, and Teva; compensation for speaking from Biogen Idec.

Douglas Arnold: honoraria from Bayer Health Care, Biogen Idec Inc., EMD Serono, Genentech, Genzyme, GlaxoSmithKline, Novartis, Roche, Merck Serono, Teva; salary from NeuroRx Research, and owns stock in NeuroRx Research.

Eva Havrdova: speakers’ honoraria and research grant support from Bayer Schering Healthcare, Biogen Idec, Genzyme, Merck Serono, Novartis, and Teva, and compensation for advisory board activities from Biogen Idec, Genzyme, Merck Serono, Novartis, and Teva.

Alexey Boyko: consulting fees and speaker’s honoraria from Bayer Schering, Merck Serono, Teva, Novartis, Biogen Idec, Nycomed, Genzyme and Sanofi-Aventis.

Michael Kaufman: honoraria and research support from Biogen Idec, financial support from Bayer, EMD Serono, Novartis, and Teva, and consultant for Department of Defense and Dechert.

Heinz Wiendl: honoraria and consultation fees from Bayer Health Care, Biogen Idec, Fresenius Medical Care, GlaxoSmithKline, GW Pharmaceuticals, Merck Serono, Novartis, Sanofi-Genzyme BioVentures, and Teva Pharmaceutical Industries; grants and contracts with Bayer Health Care, Biogen Idec, the German Ministry for Education and Research, Deutsche Forschungsgesellschaft, the Else Kröner Fresenius Foundation, the Fresenius Foundation, the Hertie Foundation, Merck Serono, Novartis, the NRW Ministry of Education and Research, the Interdisciplinary Center for Clinical Studies in Münster, Germany, the RE Children’s Foundation, Sanofi-Aventis/Genzyme, and Teva Pharmaceutical Industries.

John Rose: research support from Biogen Idec, AbbVie Biotherapeutics, Teva, Cumming Foundation, National Multiple Sclerosis Society, VA, and NIH.

Steven Greenberg: full-time employee of AbbVie Biotherapeutics.

Marianne Sweetser, Katherine Riester, and Jacob Elkins: full-time employees of Biogen Idec.

Alexandra Silveira of Biogen Idec provided editorial support in the development of this presentation.

Bibiana Bielekova is a co-inventor of the National Institutes of Health patents related to daclizumab therapy and as such has received patent royalty payments.

Frederik Barkhof is a paid consultant for Bayer-Schering Pharma, Sanofi-Aventis, Biogen Idec, Teva, Merck-Serono, Novartis, Roche, Synthon BV, Janssen Research, Genzyme, and receives research support from the Dutch MS Society, Serono Symposia Foundation, and MedScape.

Olga Ciccarelli is a paid consultant for Novartis, Biogen Idec, General Electric, and Genzyme, and receives research support from the UK Multiple Sclerosis Society, National Institute for Health Research University College London Hospitals Biomedical Research Centre, and Engineering and Physical Sciences Research Council.

Does not include disclosures from Pfender and Martin nor additional poster authors.