14 April 2011. The only approved treatments for Alzheimer’s disease are symptomatic, doing nothing to stem the disease’s inexorable progression. Better treatments are urgently needed, and a decade of setbacks in the clinic has left researchers searching for new ideas. The focus is broadening from amyloid plaques to blocking Aβ oligomers, now widely believed to be the most toxic form of the peptide. At the 10th International Conference on Alzheimer’s and Parkinson’s Diseases, held 9-13 March 2011 in Barcelona, Spain, scientists discussed numerous approaches for tackling the disease, from the preclinical to Phase 1. Many talks centered on ways to target oligomeric or protofibrillar species, from antibodies to small molecules. Other strategies roamed further afield, looking at the role of metals and the potential of a recently discovered anti-aging gene.

Immunotherapy strategies for AD are at all stages of clinical trials, though success has been mixed so far. Past trials showed that antibodies cleared amyloid plaques, but have not so far slowed mild to moderate disease (see ARF related news story). Some participants suffered encephalitis, while others develop a poorly understood side effect called vasogenic brain edema (see ARF related AD/PD story). Even so, researchers remain optimistic about the potential of antibodies to mop up Aβ and promote its clearance from the brain (see ARF related news story and ARF related SfN story). Current clinical trials employ heightened vigilance in hopes to avoid and learn about side effects, said Norman Relkin at Weill Cornell Medical College, New York City, and new trials are starting up.

Most current antibodies target sequence-based antigens of Aβ; hence, they bind peripheral Aβ. In contrast, conformation-specific antibodies against oligomers/protofibrils would be more like “guided missiles,” Relkin said, targeting what is thought to be the most toxic form of Aβ. At AD/PD, Kaj Blennow of Sahlgrenska Academy at Göteborg University, Molndal, Sweden, noted that this strategy draws its most recent support from assays measuring oligomeric Aβ in human cerebrospinal fluid (CSF). These were originally reported by Japanese scientists (see Fukumoto et al., 2010), but have now been developed in several other labs independently, including Blennow’s. All show an increase of large-sized oligomers in human CSF from people with AD compared to controls, Blennow noted.

Lars Lannfelt at Uppsala University, Sweden, with Hans Basun and other colleagues, developed an antibody, mAb158, specifically directed against such large-sized oligomers, aka protofibrils. A biotech company Lannfelt co-founded, BioArctic Neuroscience AB in Stockholm, collaborated with Eisai Co., Ltd., to develop a humanized version, dubbed BAN2401, which Eisai is now testing clinically. Andrew Satlin at Eisai told the AD/PD audience about an ongoing Phase 1 trial and another imminent one. The antibody has a thousand times greater affinity for protofibrils than for monomers, Satlin said. In cell culture, BAN2401 lowered the binding of Aβ protofibrils to hippocampal neurons and neutralized their damaging effect on those cells.

The human trials will assess safety, tolerability, and pharmacokinetics, Satlin said. The trials include a single-ascending dose phase, in which the researchers will test six doses from 0.1 up to 15 mg/kg given intravenously, followed by a multiple-ascending dose phase in which four doses will be given one month apart. The four doses in the multiple phase range from 0.3 to 10 mg/kg, Satlin said. The multiple ascending dose phase will only begin with its lowest dose after a higher dose has passed the safety muster in the single dose phase, Satlin said. This design shows that the scientists are watching out for potential safety concerns, which, according to Satlin, include vasogenic edema, microhemorrhage, hypersensitivity reactions, hypotension, formation of anti-BAN2401 antibodies, or inflammation. All this will need to be monitored closely, Satlin said.

In response to a question from immunologist Michael Agadjanyan at the Institute for Molecular Medicine in Huntington Beach, California, Satlin acknowledged that the doses at the upper end of the spectrum are high, hence, the stepwise design. Satlin’s talk prompted discussion about how to find the right dose for immunotherapy trials. Satlin said he expects the 0.3 mg/kg dose to provide roughly the amount of drug that had effects in experiments in AD transgenic mice. He also said that he expects some 0.5 to 1 percent of the antibody to penetrate the brain and work there, not through a peripheral mechanism. He hopes one of the lower doses will achieve effective exposure in the brain that is nonetheless safe in plasma.

The researchers will evaluate changes in Aβ and tau CSF and plasma biomarkers, and will look for interaction with ApoE genotype on those exploratory measures. Effects on neuroimaging markers and cognition are also part of the trial’s exploratory readouts, and will hopefully help in the dose setting and design of further trials, Satlin said. People 50 or older with clinically mild AD will be eligible (i.e., those who have a score between 16 and 28 on the Mini-Mental State Exam); the trial will exclude anyone with a history of microhemorrhages (microhemorrhages are not the same as vasogenic edema, but some scientists fear they may be connected, and more research is needed). Single ascending dose trials are ongoing at eight sites in the U.S., and a multiple ascending dose trial will start in May 2011 in the U.S. and later expand to sites in Sweden, including one led by Martin Ingelsson at Uppsala University.

The clinical trial program follows results of animal tests of mAb158, which Lannfelt discussed at AD/PD. Lannfelt developed the antibody after his discovery of the Arctic mutation in APP. AD caused by this mutation is particularly aggressive, and is driven by Aβ protofibrils, not plaques. Lannfelt said that mAb158 selectively binds Aβ protofibrils, not normal monomeric Aβ or other amyloids such as those made by α-synuclein (see Englund et al., 2007). When mAb158 was given to 10-month-old transgenic AD mice, levels of soluble Aβ dropped by three-quarters, but existing plaques remained undisturbed, Lannfelt said. In young mice, the treatment prevented plaque formation and again lowered soluble Aβ (see Lord et al., 2009). Lower levels of Aβ protofibrils correlated with better cognition, Lannfelt said.

On the heels of this clinical anti-protofibril antibody comes a different, preclinical one that supposedly also recognizes Aβ “aggregates” much more strongly than monomers, though no structural information about the particular antigen is available. At AD/PD, Christoph Hock of Neurimmune, a biotech company based in Schlieren, Switzerland, introduced BIIB037, a monoclonal antibody that Neurimmune’s larger U.S. partner Biogen Idec of Cambridge, Massachusetts, will push into human trials this year, according to Hock. This antibody is unusual in that it is not a humanized form of a mouse antibody but was isolated directly from humans in what Hock called “reverse translational medicine.” This means that the Swiss scientists started with living older people who appear resistant to AD, and tried to exploit their immune response. “We speculate that natural generation of antibodies protects some people,” Hock said.

How would this work? Using a donor cohort of 265 research volunteers in their seventies, the scientists screened the B cells of those people who were cognitively stable over a period of three years, recovered from mild cognitive impairment, or whose AD barely progressed. The scientists generated recombinant versions of naturally occurring antibodies from sequence information, and selected for development an IgG1 that binds to aggregated Aβ with an affinity below one nanomolar. With repeated intraperitoneal injection into Tg2576 mice, the antibody crossed the blood-brain barrier, accumulated on amyloid plaques, and persisted in brain over the measurement period of 2 weeks, Hock said. Repeated injection to mimic chronic treatment of chimeric versions of the human-derived antibody dose-dependently reduced soluble and insoluble Aβ40 and 42, as well as plaque load in hippocampus and cortex. “This human antibody dramatically lowered all Aβ species where deposits can be measured by biochemistry and immunohistochemistry,” Hock said.

At the same doses (3 to 30 mg/kg), microglial activation went up around plaques; amyloid angiopathy or frequency of microhemorrhages stayed unchanged, Hock added. He presented data suggesting that the antibody appears to protect the shape and survival of adult-born neurons in APP/PS1 transgenic mice as well as dendritic spines in cultured neurons (Biscaro et al., 2009). Fielding a question about memory tests in those mice, Hock noted that the antibody improves their performance in contextual fear conditioning and other tests. In response to a second question about exactly what BIIB037 binds to, Hock demurred, saying only that binding to monomer is manifold weaker than to “aggregates.” “We are not disclosing the epitope.”

As the field at large develops the basic science and translational knowledge base for further anti-oligomer or anti-protofibril approaches, investigators would be well advised to become much more precise in how they prepare, study, and name those potential targets, Dominic Walsh of Brigham and Women’s Hospital in Boston told the audience in Barcelona. “I don’t like the term ‘oligomers’ because it is nebulous. The chemical definition of oligomer as a ‘low-N mer’ tells us nothing about structure,” Walsh said. There are studies about Aβ56*, about ADDLs, about dimers extracted from human brain, the globulomer, protofibrils, and various other synthetic species. “We cannot all dump them into the same category, call them ‘oligomers’ and essentially consider them the same. We need to be much more specific,” Walsh said.

“I do think the protofibrils are one of the mediators of toxicity. The stable dimer boosts protofibril formation and may be the fundamental building block of synaptotoxic protofibrils. Going forward, I prefer that we talk about biophysically defined species,” Walsh went on. Indeed, a strongly worded editorial published in the current Nature Neuroscience makes the same point. Charging that a multitude of different methods and vague language “muddle the field,” the journal calls for scientists to state exactly where the particular form of Aβ used in their experiment comes from, to characterize its aggregation state rigorously, and to discuss its physiological relevance.

Notwithstanding the precise antigen, a slew of vaccine-based approaches are wending their way through trials. Relkin surveyed these studies, in particular noting three passive vaccines that are now in Phase 3: Eli Lilly and Company, Indianapolis, has solanezumab in trials (EXPEDITION and EXPEDITION2); Janssen Alzheimer Immunotherapy, South San Francisco, California, has bapineuzumab in trials (ApoE4 Carrier and ApoE4 Non-Carrier); and Baxter Healthcare Corporation, Deerfield, Illinois, has intravenous immunoglobulin in trials. Results from these studies should come out in 2012 and 2013, Relkin said, at which point the field can reassess what works and what does not. In addition, Merck, Genentech/Roche, Novartis, Affiris, and other companies have antibodies in Phases 2 and 1, and many labs are working on yet more preclinical ones.

With all these immunotherapies, does the field really need another? Indeed it does, Agadjanyan told Alzforum. First, multiple approved vaccines will be needed, because some people will respond to one vaccine but not another. Secondly, the field of AD immunotherapy is so young—and in Agadjanyan’s view, still a bit thinly staffed with card-carrying immunologists—that many of the vaccines that are currently in development have not been deliberately optimized for an elderly population. Vaccines for older people should be specifically designed to tap and reactivate the patient’s pre-existing memory T helper cells. That’s because in older people, those constitute by far the larger pool of T helper cells than naïve ones, which predominate in younger people. In other words, an AD vaccine could rouse an older patient’s memory T helper cells if it contained as its carrier protein a foreign T helper epitope from a conventional public health vaccine against which large numbers of older people had already mounted a T cell response earlier in life. Those could be epitopes that proved their mettle in childhood vaccines, such as from the diphtheria or tetanus toxins, or in influenza vaccines widely used in the elderly. When coupled with multiple copies of the right self-epitope from Aβ’s N-terminus, such heterologous vaccines would rally memory T helper cells to ignite a robust immune response, whereas the actual antibodies the AD patient’s B cells churn out would target Aβ. “If these vaccines prove safe, they may help especially older people overcome immune senescence,” Agadjanyan told the audience.

His group experiments with various peptide-based epitope vaccines along these lines. In Barcelona, he told the audience that some of these generate a strong response in mouse models and rabbits, whereby the T helper cell reaction was specific to the vaccine’s T cell epitope and the humoral response to Aβ. One vaccine is ready for tests in rhesus monkeys; another one, which boosts pre-existing memory T helper cells specific to the tetanus toxin epitope P30, reduced cored and diffuse amyloid plaques in Tg2576 mice. Another category of vaccine use DNA, not peptides. For those, the challenge lies in strengthening the body’s immune response, which tends to be feeble. Commercial vaccine research is developing ways to do that, for example, adjuvant patches that stimulate Langerhans cells in the skin or hand-held electroporation devices, Agadjanyan said. In addition, his group is working out prime-boost regimens that enhance both humoral and T helper cell responses with successive injections of DNA and protein (Davtyan et al., 2010).—Madolyn Bowman Rogers and Gabrielle Strobel.

This is Part 1 of a two-part summary of experimental therapies. For a discussion of small-molecule, metal-targeting, and other AD therapeutic approaches, see Part 2.

Reference:
[No authors listed]. State of aggregation. Nat Neurosci. 2011 Apr;14(4):399. Abstract