After years of fits and starts, anti-amyloid immunotherapies are finally hitting their target effectively. At least four drugs have now demonstrated the ability to clear plaques from the brain: aducanumab, gantenerumab, Lilly’s LY3002813, and BAN2401 (Jul 2018 conference news). At the Alzheimer’s Association International Conference, held July 22–26 in Chicago, researchers presented new data from gantenerumab and LY3002813, aka N3pG/donanemab. It clinched the case that these antibodies can mop up brain amyloid, bringing many people with early symptomatic Alzheimer’s disease below the threshold for amyloid positivity. At one to two years, this clearance took a long time. But still: Gregory Klein at Roche claimed that two years of treatment with high-dose gantenerumab essentially resets a person’s trajectory of amyloid accumulation. “We are setting back the clock by 15 years,” Klein told the audience.

  • After two years, half of AD patients taking gantenerumab become amyloid-negative.
  • One-third develop ARIA-E, usually without symptoms.
  • Lilly’s N3pG, directed against pyroglutamate Aβ, clears plaques in six months.

To achieve these rates of clearance, researchers have had to greatly boost antibody dose, in many cases quadrupling the amounts used in earlier, unsuccessful trials. The same was done for two additional anti-amyloid antibodies that did not report on amyloid removal at AAIC. Crenezumab’s dose was ramped up midstream in the Colombian ADAD trial, and is high in the ongoing CREAD Phase 3 program in LOAD; solanezumab’s dose was pushed up in DIAN and A4, two ongoing secondary prevention trials (Jun 2017 news). 

These high doses bring a greater risk of infusion site reactions and ARIA-E, the occurrence of leaky blood vessels causing edema in the brain. In Chicago, scientists argued that these side effects are manageable with careful monitoring of patients. Moreover, ARIA-E can be lessened by gradually titrating up the antibody dose, they said.

Other researchers found the data encouraging. “It’s impressive there are now four drugs that can remove amyloid plaques from the brain,” Randall Bateman of Washington University, St. Louis, wrote to Alzforum. However, clinicians noted that the jury is still out on how much this will help AD patients. “Several of the antibodies are looking good at removing amyloid, but the clinical efficacy still needs to be demonstrated,” said Ron Petersen of the Mayo Clinic in Rochester, Minnesota.

Brain Drain.

Amyloid scans from three patients on gantenerumab at baseline (top row), one year (middle), and two years (bottom). Reduction shown in centiloids. [Courtesy of Greg Klein.]

Previously, at the 2017 CTAD conference, Roche scientists presented six- to nine-month gantenerumab data from ongoing open-label extensions of two Phase 3 studies. Participants were titrated up from 105 or 225 mg subcutaneous gantenerumab to 1,200 mg, and after nine months, their amyloid had receded by 15 percent, with one-third of participants dipping below the threshold of brain-wide amyloid positivity (Dec 2017 conference news). 

In Chicago, Klein added data from the one- and two-year timepoints. The researchers transformed SUVR PET data to the centiloid scale, following a new convention that eases comparisons between data obtained with different tracers (Feb 2013 conference news; Nov 2014 news). At the one-year mark, 61 participants had amyloid PET scans; 43 of them had received at least six doses of 900 or 1,200 mg. Overall, their amyloid reduction was still similar to that at nine months, Klein showed. The 15 participants from the Scarlet RoAD extension notched drops of 24 centiloid at nine months and 29 at one year. The 46 participants from Marguerite RoAD stayed steady at about 45 centiloid reduction at both timepoints. (Why the difference? Scarlet RoAD enrolled prodromal AD at baseline compared with mild to moderate AD for Marguerite RoAD. The former had less amyloid at baseline, and had received fewer high doses of gantenerumab by one year, Klein explained.)

By two years, however, things changed dramatically. The 28 participants who underwent PET at this timepoint on average had received 18 to 20 doses of 900 mg or higher. Scarlet RoAD participants lost 46 centiloids worth of amyloid, while Marguerite RoAD participants who had started out on placebo and switched to gantenerumab lost a whopping 78 centiloids. Participants previously in the active arm of Marguerite RoAD saw their scans shrink by 48, because this group had little amyloid left to lose, Klein noted. Lilly’s Adam Fleisher confirmed that with his company’s new antibody, too, the higher a person’s baseline SUVR, the more amyloid reduction they tended to post (see below).

For all groups, baseline centiloid averaged 80, and the reduction after two years was 59. A centiloid of 24, corresponding to an SUVR of 1.4, is the threshold for amyloid positivity on these florbetapir scans, Klein said. By the two-year mark, half the cohort had fallen below it, and the remainder were on a trajectory toward the amyloid floor of negative 50 centiloids, Klein said. Amyloid declined fastest in people with heavy loads, and more slowly as people approached the floor, he added.

But will this help people with their cognition and function? “We are quite optimistic that these large reductions will correspond to a clinical benefit,” Klein said.

And how safe are these high doses? In Chicago, Roche’s Danielle Abi-Saab presented safety findings from the Marguerite RoAD extension of 219 participants, and Mirjana Andjelkovic from the Scarlet RoAD extension, with 154 participants. In both, the higher doses brought more ARIA-E. A third of participants in the Marguerite RoAD extension developed it, compared with 12 percent in the original study. In the Scarlet RoAD extension, it was 28 percent. APOE4 carriers had the highest rates, topping out at 38 percent in the Marguerite RoAD extension. Without titration, 58 percent of APOE4 carriers would have been expected to develop ARIA-E at these doses, Abi-Saab noted, suggesting that titration worked to lower incidence.

The majority of ARIA-E instances were detectable only on MRI scans. That said, one-quarter of participants with ARIA-E in either cohort did report symptoms, typically headache. Four participants in the Marguerite RoAD extension, and three in Scarlet RoAD, had more serious reactions, including seizures, confusion, ischemic stroke, and neurological symptoms on one side of their body. After discontinuing drug, all these conditions cleared up, the speakers said. Everyone with ARIA-E had an MRI scan per month, and on average, ARIA-E resolved fully over three months. The majority of participants with ARIA-E were able to continue in the study.

The researchers saw fewer instances of ARIA-H, or microhemorrhages. These occurred in 9 percent of Marguerite RoAD, and 6.5 percent of Scarlet RoAD extension participants. The protocol forced an end to treatment in the majority of these participants.

In both studies, about one-third of participants developed rashes at the infusion site. These were typically mild, though one person dropped out due to this reaction.

At high doses of gantenerumab, safety remains favorable, and ARIA events are manageable, the speakers concluded. Data from these extension studies were used to design the Phase 3 GRADUATE 1 and 2 studies of subcutaneous gantenerumab titrated up to high doses. Both trials have started enrolling.

Lilly’s Anti-Pyroglutamate Aβ Antibody Removes Plaque Quickly

Fast Removal.

Amyloid scans from three patients taking N3pG. Baseline (top row), three months (middle), and six months (bottom). Readings given in SUVR/centiloids; red numbers are below cutoff for positivity. [Courtesy of Adam Fleisher.]

The findings for Eli Lilly’s N3pG are similar. This newer antibody targets a pyroglutamate form of Aβ that is particularly prone to aggregate and is found mostly in plaques. As with other antibodies, Lilly’s initial dosing turned out to be too timid. In Phase 1, only the highest dose used, 10 mg/kg, produced meaningful brain exposure. At this dose, brain amyloid dropped by about 44 centiloid, with almost no ARIA. But there was a catch. Most participants made anti-drug antibodies, and six of 37 had an infusion reaction (Aug 2016 conference news). 

This prompted Lilly to explore safety and immunogenicity of higher doses, and in Chicago, Fleisher presented interim data from this ongoing Phase 1b study. The 58 participants have prodromal to moderate AD, and range around 73 years old with an MMSE score of 21. About three-fourths carry an APOE4 allele, and their average baseline centiloid was 104.

The study tests three dosing regimens. In one, participants receive a single dose of 10, 20, or 40 mg/kg. In the second, they get 10 mg/kg every other week for 24 weeks. The third is a chronic dosing schedule of 10 or 20 mg/kg every month for 16 months, or until the participant’s amyloid scan turns negative. Each dosing scheme includes nine to 12 participants randomized 3:1 to drug or placebo, except for the 40 mg/kg single dose, which has only two participants. Besides regular assessment for anti-drug antibodies, amyloid and MRI scans happen quarterly, and cognitive testing every six months.

Fleisher reported to the AAIC audience that plaque load drops after even a single dose of the antibody, and stabilizes at the new level during follow-up. Higher doses effect a greater reduction, with the highest so far being an average drop of 70 centiloid with six months of 20 mg/kg. In this highest dose group, three of six patients have become amyloid-negative so far, Fleisher said. He added that not everyone responded to 10 mg/kg chronic dosing, but all responded to 20 mg/kg.

Alas, Lilly also saw more ARIA at higher doses. About one-quarter of dosed participants developed ARIA-E. Two people reported symptoms such as headaches, confusion, and sleepiness, which went away after dosing stopped. In contrast to the first trial, just one participant experienced a rash at the infusion site, and only on the first dose. As before, however, nearly all patients made anti-drug antibodies. These have caused no apparent side effects thus far. They did not affect Np3G’s efficacy, either, except in one patient whose B cells pumped out massive amounts of antibody, more than 100,000 per drug molecule. Researchers stopped treatment in this participant, as well as in another with a microhemorrhage and in 11 with ARIA-E. Lilly is closely monitoring all participants for 18 months.

While this study is ongoing, Lilly has advanced N3pG to Phase 2, where it is being tested in combination with Lilly’s BACE inhibitor LY3202626. Dubbed TRAILBLAZER, this trial is the first to test a combination of disease-modifying therapies for AD, Lilly’s Michael Irizarry said at AAIC. The trial is currently enrolling, targeting 375 participants with early symptomatic AD. They will be divided evenly among three arms. One arm will receive placebo, another N3pG, and the third both N3pG and the BACE inhibitor. The primary endpoint will be the integrated Alzheimer’s Disease Rating Scale (iADRS) developed by Lilly, with ADAS-Cog, CDR, MMSE, and ADCS iADL as secondary outcomes. The researchers will also look at volumetric MRI, amyloid PET, and tau PET, as well as safety and tolerability.

Notably, the trial will enroll only people whose neurofibrillary tangle buildup falls within a low-to-medium range defined as an SUVR of 1.15 to 1.36 by flortaucipir PET. This is because people with less tau than the lower cutoff show little cognitive decline over one to two years, while people with more tau than the upper cutoff might be at too advanced a stage to still benefit from an anti-amyloid treatment. Stratifying participants in previous trials by their tau PET signal has shown a close correspondence between tau pathology at baseline and subsequent rate of cognitive decline, Irizarry noted.—Madolyn Bowman Rogers.


  1. Getting rid of deleterious "junk" aids recovery from many diseases. Surgeons lance and drain pus-filled furuncles. Pulmonologists suction obstructive mucus out of airways of pneumonia patients. But these helpful steps do not address the root cause of the observed problem. Hopefully, this dramatic removal of amyloid will help cognition and other functions in Alzheimer's patients. But even then, we must press for what triggers the amyloid to appear in the first place.

  2. That antibodies remove plaque can be explained by the fact that IgGs, by binding to aggregated Aβ, disrupt the hydrophobic interactions holding the aggregates together. Because of the hydrophilic character of amyloid-IgG complexes, their formation would result in soluble complexes. 

    Because of the non-specificity of this process, practically any anti-amyloid antibody should be able to remove plaque, which apparently is the case. But if the antibodies cannot neutralize the cytotoxic-diffusable amyloid aggregates, then we should not expect any improvement. Indeed, the finding that many elder but cognitively normal people have high levels of plaque argues against plaque being the etiological agent of Alzheimer. Most of the failed vaccines and antibodies had successfully removed plaque without improving the patients' condition. Finally, there is a large body of evidence showing that Aβ oligomers and protofibrils cause cell damages associated with Alzheimer in the absence of plaque; which questions the role of plaque in this disease.

Make a Comment

To make a comment you must login or register.


News Citations

  1. BAN2401 Removes Brain Amyloid, Possibly Slows Cognitive Decline
  2. A4 Researchers Raise Solanezumab Dosage, Lengthen the Trial
  3. High-Dose Gantenerumab Lowers Plaque Load
  4. HAI—Standardizing Amyloid PET: The Centiloid Project
  5. Paper Alert: Centiloid Scale Aims to Unify Amyloid PET
  6. New Ways to Target Aβ and BACE Show Promising Phase 1 Data

Therapeutics Citations

  1. Gantenerumab
  2. Donanemab
  3. LY3202626

Other Citations

  1. N3pG

External Citations

  2. 2

Further Reading