Yesterday, the 7th International Conference on Alzheimer’s and Parkinson’s disease—AD/PD 2005 for short—drew to a close in beautiful Sorrento, Italy. Nearly 1,500 scientists from 52 countries met to brush up on news and trends in research and treatment for neurodegenerative diseases. Abraham Fisher from the Israel Institute for Biological Research in Ness Ziona, and Israel Hanin at Loyola University, Chicago, have organized this biannual conference for the past 20 years. They include co-organizers from a different host country each time, this year they were Maurizio Memo from University of Breschia, and Fabrizio Stocchi from La Sapienza University in Rome. AD/PD 2005 took place in a hotel built into a Sorrento mountainside amid lemon and orange orchards. The setting offered sweeping views over an ancient town that is built spectacularly into the rocks of a peninsula rounding off the southern tip of the gulf of Naples.

The Alzforum news team will post news reports of this conference for the next two weeks. This first installment of our hot-off-the-podium dispatches summarizes the status of AD immunotherapy research. The setback in the AN 1792 trial cast a temporary pall over the field (see Alzforum Live Discussion), but soon after, researchers began trying again and the approach remains one of the most closely watched in the field. The Alzforum has covered this science in detail (for background, see Can a Shrinking Brain Be Good for You?, Vaccine, Microglia, NGF News Fill in Neuroimmunology Picture, Immunotherapy—The Game Is Still in Town); therefore, this story focuses on new human data and experimental approaches.

Dale Schenk of Elan led off an immunotherapy symposium by noting second-generation approaches his company is currently exploring. They include passive vaccination—Elan is completing a phase 1 trial of a humanized monoclonal antibody to Aβ and is preparing a phase 2 at several sites in the U.S.—and preclinical work on active immunogens made by conjugating small Aβ fragments to an external carrier. In parallel, investigators in many other laboratories are developing their own approaches.

An area that requires extensive investigation concerns the mechanisms that drive the distribution of Aβ between its different pools in plasma, CSF, and the brain. Adding antibodies to this already complicated process creates a level of complexity that research has not yet fully understood, Schenk said. For example, in the AN1792 trial, changes in Aβ levels followed no consistent pattern, which may be due to different antibodies being produced. Another major question in AD immunotherapy revolves around how amyloid gets cleared; different groups have proposed different routes including microglia devouring amyloid and the peripheral sink hypothesis. “This may be one of those questions where everyone is right,” Schenk said.

Those who perched on the edge of their seat expecting new clinical data eased back in disappointment when Schenk did not present any. He merely noted that further analysis of the phase 2 patients showed that the patients who had the strongest antibody response also tended to perform a tad better on some of the functional assessments, and that the tau levels in their cerebrospinal fluid had come down the farthest. This adds preliminary human data to an emerging body of animal data suggesting that tau pathogenesis occurs downstream of amyloid pathology and that amyloid removal can hold its early stages at bay. It is also the first demonstration of any therapeutic intervention that can lower tau in people with AD, Schenk noted.

Schenk added that, looking back, mice have had a mixed record of predicting what would happen in humans. The mouse data predicted plaque removal correctly. Their prediction of functional improvement remains largely untested, though immunotherapy buffs are putting hope into preliminary hints at a slight functional improvement in a few AN-1792 participants. On safety, the mouse models failed to predict the meningoencephalitis side effect while, on the other hand, they exhibit a bleeding side effect in the brain that has, to date, not been detected in any of the treated people (see Racke et al., 2005; Pfeifer et al., 2002; scroll down for Morgan’s presentation).

Nick Fox, of London’s University College, who had presented the MRI imaging results of the AN-1792 cohort last July in Philadelphia (see conference news story,) noted that they are slated to appear in the journal Neurology in May. He repeated that analyzing the surprising shrinkage in brain volume against other measures in the trial suggests that people whose brains had shrunk the most were also the ones who had the greatest drop in CSF tau levels. They also performed a tad better on some of the functional measures. The next challenge, besides avoiding encephalitis in future trials, is to understand what underlying processes in the brain reduce its volume after immunization, as amyloid removal alone can account at best for a fraction of the observed shrinkage. Fox suggested that fluid shifts may be part of the answer, while Roger Nitsch of University of Zurich suggested in his talk that the calming down of astrogliosis might free up measurable space.

Nitsch’s group perform clinical MRI on the Swiss cohort themselves. Christoph Hock, a longtime colleague of Nitsch’s, said that data from the year three imaging time point are not available yet. The year three clinical follow-up is being completed. Hock noted that some patients still have high antibody titers at this point, and that these “responders” still outperform the “non-responders” slightly on functional assessments. Addressing the question of which Aβ measurements are best suited to such trials, Hock said that the Zurich researchers observed patients with high ELISA titers decline but not those with high TAPIR values, suggesting that using ELISA alone might scramble the trial result and understate how well the responders did. (The TAPIR is an assay developed in Roger Nitsch’s group, where patient serum is used to stain AD tissue. Its developers claim that it is better at detecting pathogenic neoepitopes than standard ELISA tests.) Some researchers had deplored that the whole AN-1792 trial population was not being followed properly, but Hock said that TAPIR analysis of the whole cohort is under way.

James Nicoll presented his analysis of a fourth postmortem case in addition to the three that have to date been published (Nicoll et al., 2003; Ferrer et al., 2004; Masliah et al. 2005). Case 4 was a British patient from the phase 1 trial of the AN1792 vaccine, who died four months after receiving it of an abdominal aortic aneurysm deemed unrelated to immunization. Unlike the other three cases, whose cortex was devoid of plaques in some areas and whose Aβ burden had dropped, case 4 had still had abundant amyloid plaques and the amyloid load was like that of unimmunized patients. However, the plaques looked different, Nicoll said. They appeared looser, moth-eaten, and nearby microglia carried lots of Aβ inside their lysosomes. Overall, it looked as if the patient had died during an active phase of plaque removal, Nicoll said.

Nicoll also talked about tau pathology, and noted that in cases 1, 2, and 3, tangles and neuropil threads stayed in place but dystrophic neurites had straightened out, as was reported in mice, as well (Brendza et al., 2005). This was yet another of many hints reverberating throughout the meeting that amyloid immunotherapy may resolve early stages in tau’s pathogenic pathway. (Other work in mice, most prominently some presented by Mike Hutton of the Mayo Clinic in Jacksonville, Florida, goes in the same direction. Hutton reported that shutting off a tau transgene in an inducible mouse model enabled the mice to recover from their memory deficit even while tangles stayed in place but dystrophic neurites resolved.)

Finally, Nicoll addressed the current worry that Aβ immunotherapy may remove parenchymal Aβ but cause trouble at the brain’s blood vessels. As a group, cases 1 through 4 had fairly severe cerebral amyloid angiopathy compared to what is typically seen in AD, Nicoll said. He is currently studying three additional autopsy cases from the phase 1 trial and gathering informed consent from the UK participants in these studies. “I want to reassure you that we are following up the UK patients,” he said.

Dave Morgan’s current work with mice ties into this concern about CAA, and its possible consequence of bleeding from damaged blood vessels. Morgan, who is at University of South Florida in Tampa, said that when his coworkers injected an anti-Aβ antibody into APP Tg2576 mice, this passive vaccine not only elicited a strong antibody response but also led to an increase in CAA. “This surprised us,” Morgan said. Further study of the different types of deposit in the mice showed that, after both three and five months of vaccination, parenchymal amyloid predictably decreased but vascular amyloid increased to where it became the predominant deposit, where previously most was parenchymal. Morgan speculated that microglia absorb amyloid in the parenchyma and dump it out at the blood vessels, where its amounts overwhelm local transport and drainage mechanisms so that it promptly deposits again. Dosing the vaccine to a slower rate of amyloid removal might address this problem, he said (see also Weller talk in St. Moritz meeting report).

Intrigued by Mathias Jucker’s research of micro-hemorrhages (Pfeifer et al., 2002) and Dave Holtzman’s work on ApoE and CAA (Fryer et al., 2003), Morgan decided to look for leaking blood vessels in the immunized mice with the dye Prussian Blue. Indeed, he found that, over time, increasing amounts of blood seeped out of blood vessels into the brain parenchyma. The mice seemed strangely unfazed by these micro-hemorrhages, as they continued to show a vaccine-induced improvement in the water maze test even as the bleeding increased. This may mean that the bleeds are harmless, or it might remind us how poorly water maze performance models human brain function. (Morgan used an antibody to the C-terminal end of Aβ; the current Elan passive vaccination trial uses an N-terminal antibody.)

One of the more promising new vaccination approaches lies in fashioning cleverer vaccines from small snippets of Aβ. That became necessary because scientists now agree that the side effect in the phase 2 trial arose from autoimmune Th1 lymphocytes. Egged on further by an adjuvant known to elicit Th-1 responses, these cells reacted to Aβ and infiltrated the brain. Broadly speaking, Th-1 immune responses tend toward cellular responses and predominate in autoimmune conditions, whereas Th2 responses drive humoral, i.e. antibody, immunity and tend to be seen in allergies. To an immunologist, the workup of the AN 1792 patients’ brains was simple, said Michael Agadjanyan of the Institute for Molecular Medicine in Huntington Beach, California: “The B cells are the good guys, the T cells are the bad guys.” To avoid rousing Th1 cells, several groups have buffed up their immunology skills and are now testing vaccines based on B cell epitopes. (Aβ houses B cell epitopes between its amino acids 1-15, and T cell epitopes in the middle and toward the C-terminal end.)

Working with David Cribbs at University of California, Irvine, Agadjanyan took a B cell epitope and fused it with a generic T cell epitope called PADRE, which serves to activate Th2 helper cells. They, in turn, help fully activate Aβ-specific B cells into antibody-producing cells, as a proper immune response is thought to require. The trick in this scheme is that the T cells are directed against PADRE, not Aβ, and that the B cells make antibody against Aβ, not PADRE. An added technical twist boosts the strength of the immune response by planting the vaccine on a "lysine tree". The performance of this vaccine in Balb/C is published (Agadjanyan et al., 2005). In Sorrento, Agadjanyan said that in the APP Tg2576 and triple transgenic AD mouse models, too, the vaccine generates a strong antibody response but no T cell response. The cytokine profile in the vaccinated mice was consistent with a Th2 response, not an autoimmune reaction, and the antibodies bound specifically to plaques from AD brain.

Cindy Lemere, of Brigham and Women’s Hospital in Boston, has developed different new vaccines that do not evoke an Aβ-specific T cell response. She uses the Aβ1-15 peptide, either multiple copies attached to a lysine tree (the "dendrimeric" vaccine) or two copies connected by a lysine spacer and attached to an RGD motif to boost immunogenicity. An added appeal of her approach lies in the fact that she delivers her vaccines through the nose, raising the distant prospect of nose drops against AD.

Lemere’s team first described the performance of these new vaccines in wild-type mice at last year’s Society for Neuroscience conference in San Diego. In Sorrento, Lemere added data from tests in APP-transgenic mice. Both vaccines elicited a humoral immune response that was much stronger than that generated by prior vaccines using full-length Aβ. The antibodies labeled plaques and vascular amyloid in a TAPIR assay in AD tissue. When immunizing Lennart Mucke’s J20 APP-transgenic mice with the new vaccines for 6 months, Lemere found that the plasma levels of total Aβ shot up midway through the study and then came down toward the end. Plaque deposition decreased strongly, and cerebral Aβ levels dropped, as well. The mice exhibited no side effects. Lemere said that she hopes to test the vaccine in triple transgenic mice to look for effects on tau pathology, and then test it in monkeys (see also Lemere et al., 2004.)

Only the mouth can rival the nose as the simplest entry point into the human body. Dangle the prospect of a little white pill, and drug companies will listen with more interest to an academic’s budding experimental therapy than if it requires injection. Enter Takeshi Tabira’s from the National Institute of Longevity Sciences in Obu, Japan. Tabira described his group’s development of an oral vaccine that exploits the intestinal immune system’s ability to generate responses that follow a Th2 pattern.

The immune system in the gut is often overlooked as the dull, puny cousin of the more formidable system in the spleen and lymph nodes. However, it is quite extensive, spreading through the 300 square meters of intestinal mucosa an adult carries coiled up in the belly. To avoid peptide digestion in the stomach, Tabira’s group fashioned gene therapy vaccines out of Aβ1-43 and Aβ1-21 packaged into adeno-associated virus (AAV). This virus is used in experimental gene therapy regimens but comes with its own safety concerns. (Various gene therapy approaches are in preclinical development, but in the clinic all gene therapy remains problematic.)

Tabira said that antibodies against the Aβ peptides appeared in the lamina propria of the upper part of the small intestine of young Tg2576 mice treated with the vaccine. The mice developed mostly IgG antibodies, and their serum stained plaques from AD tissue in a TAPIR assay. The vaccine inhibited amyloid aggregation in vitro and reduced Aβ deposition and burden in vivo. It also appeared to reduce levels of phospho-tau. T lymphocytes from these mice did not react to full-length Aβ; the mice showed no evidence of inflammation save for microglial markers, Tabira reported. Older mice receiving the vaccine produced fewer antibodies, but also exhibited a reduction of their amyloid burden and inflammatory markers such as TGF-beta1 and certain cytokines. Tabira noted that these effects followed a single oral dose and then persisted for weeks.

Anticipating safety concerns regarding the use of AAV, Tabira noted that the cells at the top of the intestinal mucosa turn over quickly. In mouse and rat, these dead cells exfoliate into the gut lumen and are excreted, whereas monkeys internalize the dead cells first and then degrade them. Tabira was unable to detect the AAV vector in other tissues of the mouse, but even so, this issue will require careful study before moving this approach into humans. The vaccine has entered testing in monkeys, Tabira noted.

Overall, the talks brought into focus these common themes and questions:

• Safer antigens seem to be effective in mice,
• questions around vascular amyloid and hemorrhages need attention,
• detecting reactivity in tissue may be a better way of quantifying the antibody response than ELISA,
• clearance mechanisms remain murky but microglia are part of it,
• amyloid immunotherapy might stem tau pathology,
• all eyes will be on the passive vaccine once it has entered phase 2.

The AD/PD meeting featured other talks on vaccination, such as an update by Beka Solomon on her approach using filamentous bacteriophages to deliver Aβ peptides through the nose (Solomon, 2005), but this writer was unable to attend them all. As always, meeting attendees are cordially invited to complete this update with their own notes and corrections.—Gabrielle Strobel.