For more than a decade, industry and academia together have sunk many millions of dollars into the prospect that antibodies glomming onto Aβ peptides could help shrink amyloid plaques in the brains of Alzheimer disease patients. At last, live brain imaging confirms this does happen—in people receiving an investigational drug at doses similar to those used in ongoing Phase 3 AD trials. Using positron emission tomography (PET), “this study demonstrates for the first time the effect of an [immunotherapy] AD drug on pathological changes of the disease in living patients,” said Juha Rinne, University of Turku, Finland. Rinne led the study published online March 1 in Lancet Neurology. It remains to be seen whether amyloid clearance by this treatment, a humanized monoclonal anti-Aβ antibody known as bapineuzumab, brings measurable clinical improvement. Still, the demonstration that bapineuzumab works as expected biologically is a key proof of concept for the field and should expand possibilities for earlier treatment and prevention trials. (One much smaller study had previously shown a PIB and CSF biomarker change in response to treatment with phenserine; see Kadir et al., 2008).

In an 18-month Phase 2 trial of 234 mild to moderate AD patients, bapineuzumab did not offer clear clinical benefit, though a post-hoc analysis revealed that a subgroup of ApoE4 non-carriers fared better on some cognitive measures and experienced fewer adverse effects (ARF conference report and Salloway et al., 2009; see also Wilcock, 2010 commentary for more on ApoE4 angle). This trial did not include live brain imaging, which meant the researchers were giving antibody to participants without knowing how much amyloid existed in their brains at baseline, or how much was cleared by the treatment. The current study was designed as a small pilot trial to see whether PET imaging using the amyloid tracer Pittsburgh Compound B (PIB) could detect reduction of brain amyloid in AD patients treated with the compound, said senior investigator Michael Grundman of Elan Pharmaceuticals, South San Francisco. “The objective of this study was really very limited,” he said. “We wanted to see whether bapineuzumab was doing what it was supposed to do.

In the pilot study, 28 people with mild to moderate AD received intravenous bapineuzumab or placebo at doses of 0.5, 1, or 2 mg/kg every 13 weeks for up to 18 months (78 weeks). Beyond the PET-PIB imaging, participants underwent a range of other assessments, including clinical lab tests, fluorodeoxyglucose (FDG)-PET to measure glucose metabolism in the brain, cognitive/behavioral tests, volumetric and safety MRI, and, in those who agreed to spinal taps, cerebrospinal fluid (CSF) measurements of tau, phospho-tau, and Aβ42. As the primary endpoint, the researchers looked for changes in brain amyloid levels, measured by PIB-PET in six cortical areas, at weeks 20, 45, and 78 relative to baseline.

In participants receiving bapineuzumab, brain amyloid load declines reached statistical significance in all six brain regions, with overall amyloid load dropping on average 8.5 percent over 78 weeks compared to baseline. In the placebo group, PIB-PET detected a 16.9 percent rise in average brain amyloid over the study’s duration, adding up to a 25 percent difference between the groups. “For the first time, a study gives us a realistic indication of the amount of amyloid reduction over a 1- to 1.5-year period with these doses of immunotherapy,” Murali Doraiswamy, Duke University Medical Center, Durham, North Carolina, told ARF. “Until now, I don't think anyone knew.” Doraiswamy added, though, that the magnitude of amyloid clearance is “modest and far less than expected from prior autopsy studies of immunized patients and animal studies.” Immunization with Aβ peptides sharply reduced pathology in AD transgenic mice (Schenk et al., 1999 and ARF related news story), and brain autopsy of a trial participant who had received Elan’s ill-fated Aβ peptide vaccine (AN1792) showed that plaques had virtually vanished from large swaths of the brain (Nicoll et al., 2003 and ARF related news story).

The ultimate test lies in whether treatment with Aβ antibodies offers functional benefit. “At the end of the day, one can clear all the amyloid in the brain and if the patient does not improve, it still is not a useful therapy,” noted Doraiswamy, who has received research grants from, and served as a scientific adviser to, companies developing amyloid-based therapies or PET imaging methods. “I would like to see whether patients with the most amyloid cleared from their brains also had the biggest clinical improvement. That would be very important for the field,” he said. “That's a huge missing link.”

Based on follow-up analysis of participants from Elan’s AN1792 active immunotherapy program, which halted at Phase 2a because 6 percent of subjects developed encephalitis, it is not entirely clear whether those who responded to the Aβ vaccine fared better or worse clinically. Several studies of the Phase 2a cohort suggest that responders held their own in activities of daily living better than did placebo-treated patients (Vellas et al., 2009; Gilman et al., 2005 and ARF related news story). However, an analysis of eight people from the much smaller Phase 1 study concluded that clearance of plaques by full-length Aβ42 vaccination did not correlate with slowing of disease (Holmes et al., 2008 and ARF related news story).

In the current analysis, though cognitive and behavioral tests were included, the authors emphasize that it was designed as a biomarker study and was not powered to assess clinical outcomes. “Clinical outcomes are so noisy that we really couldn't evaluate correlations between the PIB and the clinical outcomes,” Grundman said. “Hopefully the Phase 3 trials will be able to speak to that issue.” This reporter pressed both Grundman and Rinne on this point, but both insisted on waiting for more data.

Lon Schneider, University of Southern California, Los Angeles, agrees it would be imprudent to measure clinical outcomes in a trial of this size. “I would distinguish the very small pharmacologic study from the clinical efficacy trial,” said Schneider, who was not involved in this study but has consulted for its sponsors. “If we could measure clinical efficacy with 28 patients, we wouldn't need to do trials with a thousand patients.”

Rather, this pilot study highlights “the sensitivity of this kind of imaging technique,” Rinne said. “Even with a small number of patients, you can really see if the drug has an effect on brain amyloid.” Demonstrating the proposed biologic activity of a treatment at its proposed site of action in the body is a technical step in the drug development process that is sometimes neglected in AD drug development.

This should help efforts to target people at earlier stages of disease, eventually even those with pathological brain changes who have yet to develop cognitive symptoms, said Grundman. “Now we can give a drug to milder patients and monitor for amyloid reduction and for what effects that has clinically over time.”

Even in the current study, PIB-PET imaging likely affected the outcome by enabling the investigators to scrap potential participants on the basis of their baseline amyloid load. About 15 percent of the patients screened for this study did not meet inclusion criteria because their PIB-PET scan showed insufficient brain amyloid. Rinne noted that these people may not necessarily have been PIB-negative, but had to be excluded for not meeting pre-specified cortex-to-cerebellum ratios of PIB retention in all three required brain areas. It may be that some met the criteria in one or two areas and would have later developed high amyloid in the third, he suggested. “In the future, it would be very interesting to follow up on patients who for some reason have low PIB uptake—to see whether they have pathologically proven AD or something else, and whether they will develop amyloid later during the course of disease,” Rinne told ARF.

In Doraiswamy’s view, the study’s positive result likely hinged on the ability to exclude people based on brain amyloid load. “This shows the power of PET amyloid imaging to select people who have pathology in order to maximize your chance of a drug effect,” he wrote in an e-mail to ARF. “Prior to this, we were treating AD patients blindly, without knowing how much amyloid they had in their brains—a bit like treating people with a statin without knowing their cholesterol level.” (See full comment below.)

Several issues may yet complicate the use of PET amyloid imaging in future early intervention and prevention studies. In an accompanying commentary, Sam Gandy, Mount Sinai School of Medicine, New York, wrote that PIB binds fibrillar but not oligomeric Aβ, though growing evidence suggests the latter may be the most neurotoxic in AD. Furthermore, Doraiswamy cautioned, large-scale imaging studies are likely to use 18F-labeled amyloid tracers with longer half-life than the current 11C-labeled PIB, and varying methods for computing amyloid load with these tracers may cause determinations of “normal” and “abnormal” to differ between studies. He presented Phase 2 multicenter data on Avid Radiopharmaceuticals’ F18 amyloid agent, florbetapir (formerly known as AV-45), at the 2009 International Conference on Alzheimer’s Disease in Vienna (ARF conference report), and expects that Phase 3 validation data on this tracer will be reported later this year. “Once the validation is complete, it will really jumpstart the use of the PET amyloid imaging in secondary and primary prevention trials of both drugs and lifestyle interventions. It will also give us more insight into the role of amyloid in aging and dementia, and allow us to test mechanistic hypotheses,” he wrote.—Esther Landhuis

Comments

  1. This is a very impressive study. It is the kind of pilot biomarker study that every top investigator dreams of doing, and kudos to the team that did it.

    I noticed some 15 percent of AD patients were dropped from entering the trial because the scan showed they did not have sufficient amyloid in the brain. Without dropping these people, the study would likely have had no chance of showing a positive result and might have also exposed more people to risks. This shows the power of PET amyloid imaging to select people who have pathology in order to maximize your chance of a drug effect. Prior to this, we were treating AD patients blindly without knowing how much amyloid they had in their brains, a bit like treating people with a statin without knowing their cholesterol level.

    With regard to the bapineuzumab therapy, the magnitude of amyloid clearance seems consistent and real, but at around 20 percent is modest. That is far less than was expected from prior autopsy studies of immunized patients or animal studies which suggested the vaccines might have a much bigger effect. This is a new insight and we might need to lower our expectations. The potential promise with this technology is that we might be able to test how different doses of therapy affect amyloid clearance at an early stage, allowing companies to select the most optimal dose for definitive trials.

    Going beyond amyloid, at the end of the day, one can clear all the amyloid in the brain and if the patient does not improve, it still is not a useful therapy. So what's missing is for the field to now show that clearing amyloid eventually leads to a meaningful cognitive and functional benefit for the individual.

    Some minor methodologic issues: differences at baseline in cognition and amyloid burden (not unexpected in small studies) between treatment groups add a bit of uncertainty as to interpretation. The method for computing standardized uptake values relative to cerebellum (i.e., the amyloid ratios) varies slightly from one study to another, and one sees different ratios being called normal or abnormal. This makes it hard to compare findings across studies and across tracers. So I think we need head-to-head comparisons and also some standardization of the way one determines a positive from a negative scan.

    At HAI and AAN in Toronto, and ICAD in Honolulu, we will see lots of new data on these tracers in terms of cognitive correlates in normals and MCI subjects. I also expect AVID's florbetapir (formerly known as AV-45) will be the first to present validation data from a multicenter autopsy study. Once the validation is complete, this will really jumpstart the use of PET amyloid imaging in secondary and primary prevention trials of both drugs and lifestyle interventions. It will also give us more insight into the role of amyloid in aging and dementia, and allow us to test mechanistic hypotheses.

  2. This is more a question than a comment. Both PIB and β amyloid antibodies bind to the same relatively small peptide, β amyloid. Is there any work testing whether they displace each other?

    View all comments by Chris Carter

Make a Comment

To make a comment you must login or register.

References

Therapeutics Citations

  1. Bapineuzumab

News Citations

  1. Chicago: Bapineuzumab’s Phase 2—Was the Data Better Than the Spin?
  2. Vaccinating Against Plaques
  3. Trials and Tribulations—Autopsy Reveals Pros and Cons of AD Vaccine
  4. Atorvastatin, Vaccine Trial Data Published
  5. AD Clinical Pipeline: Immunotherapy Woes, Dimebon Boons
  6. Vienna: New Genes, Anyone? ICAD Saves Best for Last

Paper Citations

  1. . Effect of phenserine treatment on brain functional activity and amyloid in Alzheimer's disease. Ann Neurol. 2008 May;63(5):621-31. PubMed.
  2. . A phase 2 multiple ascending dose trial of bapineuzumab in mild to moderate Alzheimer disease. Neurology. 2009 Dec 15;73(24):2061-70. PubMed.
  3. . Bapineuzumab in Alzheimer's disease: where now?. Lancet Neurol. 2010 Feb;9(2):134-6. PubMed.
  4. . Immunization with amyloid-beta attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature. 1999 Jul 8;400(6740):173-7. PubMed.
  5. . Neuropathology of human Alzheimer disease after immunization with amyloid-beta peptide: a case report. Nat Med. 2003 Apr;9(4):448-52. PubMed.
  6. . Long-term follow-up of patients immunized with AN1792: reduced functional decline in antibody responders. Curr Alzheimer Res. 2009 Apr;6(2):144-51. PubMed.
  7. . Clinical effects of A{beta} immunization (AN1792) in patients with AD in an interrupted trial. Neurology. 2005 Apr 7; PubMed.
  8. . Long-term effects of Abeta42 immunisation in Alzheimer's disease: follow-up of a randomised, placebo-controlled phase I trial. Lancet. 2008 Jul 19;372(9634):216-23. PubMed.

Other Citations

  1. AN1792

External Citations

  1. Phase 3 trials
  2. Phase 3 validation data

Further Reading

Papers

  1. . Clinical effects of A{beta} immunization (AN1792) in patients with AD in an interrupted trial. Neurology. 2005 Apr 7; PubMed.
  2. . A phase 2 multiple ascending dose trial of bapineuzumab in mild to moderate Alzheimer disease. Neurology. 2009 Dec 15;73(24):2061-70. PubMed.
  3. . Immunization with amyloid-beta attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature. 1999 Jul 8;400(6740):173-7. PubMed.
  4. . Long-term follow-up of patients immunized with AN1792: reduced functional decline in antibody responders. Curr Alzheimer Res. 2009 Apr;6(2):144-51. PubMed.
  5. . Long-term effects of Abeta42 immunisation in Alzheimer's disease: follow-up of a randomised, placebo-controlled phase I trial. Lancet. 2008 Jul 19;372(9634):216-23. PubMed.

Primary Papers

  1. . Testing the amyloid hypothesis of Alzheimer's disease in vivo. Lancet Neurol. 2010 Apr;9(4):333-5. PubMed.
  2. . Bapineuzumab in Alzheimer's disease: where now?. Lancet Neurol. 2010 Feb;9(2):134-6. PubMed.
  3. . 11C-PiB PET assessment of change in fibrillar amyloid-beta load in patients with Alzheimer's disease treated with bapineuzumab: a phase 2, double-blind, placebo-controlled, ascending-dose study. Lancet Neurol. 2010 Apr;9(4):363-72. Epub 2010 Feb 26 PubMed.