Why lecanemab’s success spells the revival of small-molecule Alzheimer’s treatments

In 1999, a woman visited the office of neuroscientist Rudolph Tanzi of Massachusetts General Hospital and Harvard Medical School. She was in her early 30s, seemingly in good health. But she had learned that her mother carried a genetic mutation that predisposed her to Alzheimer’s disease. By the time the mother had reached 45, the disease had ravaged her brain. Tanzi explained to his visitor that she likely inherited the same genetic marker.

“She said, ‘I have two kids, 5 and 6 years old. You’re telling me that when those two kids are in college, I’m not going to know who they are anymore?’ ” Tanzi recalls. “That hit me pretty hard.”

That night, Tanzi called his friend, the late neurologist Steven Wagner, who had been working on an Alzheimer’s treatment for Bristol Myers Squibb. The treatment candidate had been discontinued because of its side effects. As the friends chatted, Tanzi’s recounting of the woman’s visit renewed Wagner’s determination to try again and spurred Tanzi to join the effort to devise a viable treatment. Tanzi recalls Wagner saying, “We’ve got to do something.”

The pair were well aware of the challenges they faced. Alzheimer’s was incurable. It still is today. But back then, the medical community had no successful drug precedent to go off—no validated disease pathway that guaranteed a druggable target.

Alzheimer’s affects over 6 million people in the US, according to the National Institute on Aging, and more than an estimated 32 million people worldwide (Alzheimer’s Dementia 2022, DOI: 10.1002/alz.12694). To treat the disease, drugmakers need a candidate that is widely accessible and, more importantly, safe. Small molecules, experts say, could check these boxes. In the decades since Tanzi’s meeting, researchers have experimented with several compounds, many of which stumbled in the clinic. But the field has been reenergized by the success of a different sort of compound: the biologic drug lecanemab.

Lecanemab’s legacy

While the causes of Alzheimer’s disease are still not fully understood, one hallmark of the condition is the accumulation of protein deposits called amyloid-β (Aβ) in the brain. These form plaques that are detected in 70–80% of people diagnosed with Alzheimer’s. Plaques’ presence in the brain has led to the so-called amyloid hypothesis, which suggests that Aβ is a driver of the disease. If the hypothesis holds, removing or preventing the formation of these plaques may ward off Alzheimer’s or, at the least, slow cognitive decline.

For decades, the scientific community was skeptical. But last year, vindication for the amyloid hypothesis arrived in the form of clinical success for the Alzheimer’s drug lecanemab, a monoclonal antibody developed by Eisai and Biogen that works by dissolving brain plaques. Lecanemab slowed cognitive decline in people with early-stage Alzheimer’s and was the first treatment for this disease to unequivocally pass its clinical trials.

“It’s both a breakthrough and a disappointment,” says Samuel Gandy, a neurologist at the Mount Sinai Alzheimer’s Disease Research Center.

A breakthrough because it’s the first Alzheimer’s drug to show any measurable cognitive improvement among trial participants; a disappointment because the effects were only modest.

One explanation for lecanemab’s low efficacy is that anti-Aβ treatments need to be deployed early to be effective. To illustrate the importance of early intervention, Tanzi often draws parallels between Alzheimer’s and heart disease, from which his long-time collaborator, Wagner, died last year. “Amyloid is to Alzheimer’s as cholesterol is to heart disease,” Tanzi says. Aβ may accumulate decades before Alzheimer’s symptoms show. Treating full-blown Alzheimer’s with antiamyloid medication after Aβ has wreaked havoc in the brain is like throwing cholesterol-lowering statins at heart failure when a coronary artery bypass is in order.

But monoclonal antibodies aren’t practical for preventive, long-term treatments. Like all biologic drugs, they’re expensive to manufacture and store. Lecanemab requires repeated intravenous infusion over at least 18 months and costs $26,500 per year, a figure analysts say is an overcharge.

Amyloid-removing antibodies also come with another caveat: people face a risk of potentially fatal bleeding and swelling in the brain. When monoclonal antibodies bind to plaques in the brain, they can inflame and damage the underlying cerebral blood vessels that are awash in amyloid.

Small-molecule alternatives show promise at circumventing the challenges of monoclonal antibodies. Structurally less complex, generally cheaper, and compatible with at-home oral consumption, small molecules make it feasible for people to start their drug regimen early and stay on it longer. They also tamper with Aβ synthesis instead of removing established plaques later, sidestepping the events that directly cause swelling and bleeding.

Previous small-molecule candidates have been foiled by their own safety issues and low efficacy.

But thanks to lecanemab’s role in clarifying a possible druggable pathway, researchers are revisiting small molecules that target this same pathway. Lecanemab “has raised the energy and the enthusiasm for the work that maybe we can actually get somewhere,” Gandy says. He calls the drug “the rising tide that raises all boats.”

The secretase strategy

Alzheimer’s telltale plaques form in the brain from the amyloid-β precursor protein (APP), which dangles from the cell membrane. APP itself is harmless, but in individuals with Alzheimer’s, it’s snipped into fragments by the enzymes β-secretase and γ-secretase. One fragment that’s partly embedded in the cell membrane, the pesky Aβ itself, escapes from the surface and clumps with other similarly liberated stubs. These monomers assemble into small clusters called oligomers, which then form fibrils and eventually aggregate into what scientists call Aβ plaque.

In the first decade of the 2000s, molecules that inhibit the secretases were seen as ways to nip Aβ plaques in the bud. A flurry of small-molecule β- and γ-secretase inhibitors entered the clinic. But while these inhibitors successfully reduced Aβ, the drug candidates also gave rise to severe complications. Scientists realized that the multitasking secretase enzymes cleave many substrates in the cell beyond APP.

γ-Secretase inhibitors were particularly problematic. The enzyme acts on over 150 targets besides APP, including a crucial receptor in the Notch signaling pathway, which regulates intercellular communication and the cell’s life cycle. Clinical trials proved that γ-secretase inhibitors wreaked more mayhem than they tamed. Scientists have since closed the chapter on γ-secretase inhibitor development.

A string of failures also marked the development of β-secretase inhibitors. Blanket deactivation of the secretases, it seemed, lacked the finesse and specificity to treat Alzheimer’s while sparing other important biological processes in the body. Researchers needed a more nuanced approach.

Modify, don’t block

After these failures, companies looked for compounds that modulated the activity of target enzymes instead of fully blocking it. γ-Secretase drew particular interest as a target because of its role in generating the most deleterious species of Aβ peptides, Aβ42, which is 42 amino acid residues long.

γ-Secretase is a sloppy cutter. The enzyme produces Aβ peptides of various lengths. Aβ42 is the predominant species found in the brain plaques of people with Alzheimer’s. Some researchers think that this fragment is more toxic and prone to aggregation than shorter products.

The hope was that modulators could bind to an allosteric site, or side pocket, on γ-secretase and trigger a conformational change that would make the enzyme less likely to produce Aβ42. But like other early secretase inhibitors, many modulator drug candidates faltered under clinical settings.

The first-generation modulators were nonsteroidal anti-inflammatory drugs such as ibuprofen. These modulator candidates curbed the levels of Aβ42 in favor of the shorter Aβ38. But their bulky molecular sizes prevented them from crossing the blood-brain barrier efficiently, and they recorded low efficacy in the clinic.

Later, the field pivoted to heterocyclic compounds with the goal of increasing brain penetration and limiting off-target effects. Bristol Myers Squibb designed a bicyclic pyrimidine that increased the proportion of sub-40-length Aβ relative to the longer variants. But liver toxicity issues forced the developers to abandon their compound. Pfizer’s bicyclic pyridinone derivative similarly showed a reduction of Aβ42 levels among patients during Phase 1 trials, but the pharma giant has since paused the development of its program.

Today, a molecule that Wagner and Tanzi discovered is one of the few remaining γ-secretase modulators in the works and the closest to clinical debut. Experts such as Gandy say this candidate is the most promising one yet. The small molecule forces γ-secretase to open its Pac Man–like mouth a tad wider so the enzyme can chomp off more APP to produce a stump shorter than 42 amino acid residues. So far, this molecule has shown to be effective at eliminating Aβ42 production in rodents—creating Aβ38 and Aβ37 in its place—at low dosages with no side effects (J. Exp. Med. 2021, DOI: 10.1084/jem.20202560). The research team is gearing up for Phase 1 clinical trials this year.

The duo took 20 years and over $20 million in research funding to arrive at the current state, Tanzi says. “We kept going and going,” he says. “In essence, we kind of became the last [ones] standing for a small molecule that can hopefully safely hit Aβ production.”

Arresting assembly

The antiamyloid approach may not be just about prohibiting plaque production from the get-go. Some experts think it’s more effective to target Aβ midstream—once the monomers misfold and cluster into oligomers but before they harden into plaques. The oligomeric form of Aβ is often thought to be the instigator of neuronal death. A key to plaque prevention could lie in shutting down these oligomers.

The Massachusetts-based company Alzheon claims it’s the closest to the first oral treatment for Alzheimer’s, with an oligomer-ousting molecule called ALZ-801 making its way through the clinic. According to Chief Scientific Officer John Hey, the company’s drug candidate contorts the shape of Aβ monomers to bar them from stacking into longer oligomers and plaques.

ALZ-801 is derived from tramiprosate, a small molecule that flunked clinical trials in the mid-aughts. Its then developer, Neurochem, reported low efficacy among people with mild to moderate Alzheimer’s. Nine years later, Alzheon researchers reanalyzed the data and found that the drug slowed cognitive decline in the subgroup of participants whose genes predisposed them to Aβ accumulation in the brain.

Alzheon thinks tramiprosate may have failed the first time around because it was offered to the wrong group of trial participants—those showing Alzheimer’s symptoms but not confirmed to have high levels of amyloid in the brain, which constitute a small fraction of people with the disease. More crucially, tramiprosate is metabolized in the gut, which leads to low uptake of the compound into the bloodstream. Alzheon’s work-around is to formulate a prodrug in which tramiprosate is conjugated to a valine amino acid group. This valine attachment breaks off in the gut, allowing tramiprosate to enter the bloodstream intact.

Phase 3 clinical trials are underway for people with the genetic markers for Aβ production. Alzheon is aiming for ALZ-801 to hit the market in 2025. The company also hopes to carry out a preventive trial, the ultimate goal of small-molecule antiamyloid treatments. “The earlier you start the treatment, the better the outcome will be,” Hey says.

Like many other disease treatments that experience safety setbacks, Alzheimer’s drug candidates suffer from community-wide clinical inertia that’s challenging to break away from. Some researchers carry cautious hope for Alzheon’s candidate despite one strike against it previously. “The history of the drug is not promising,” Gandy says. “I tend to be skeptical, though I see some evidence to the contrary.”

Alzheon is not the only company targeting Aβ oligomers. Dieter Willbold’s team at Heinrich Heine University Düsseldorf and the Jülich Research Center aims to reverse the formation of oligomers using a short peptide around 1.6 kDa. It’s not a small molecule, but it can be taken orally and has many of the same advantages. On top of disassembling oligomers, the peptide also binds to Aβ monomers and stabilizes them, driving equilibrium away from the aggregates (ACS Chem. Neurosci. 2019, DOI: 10.1021/acschemneuro.9b00458).

“We fold [oligomers] back into monomers,” the biophysicist says. “The monomers are not really the problem. They go their own way to be excreted.”

Being larger and structurally more intricate than a small molecule, the peptide has lower chances of off-target binding. Moreover, targeting the errant oligomers rather than an enzyme would in theory eliminate any side reactions.

Early preclinical and clinical results for the drug, PRI-002, showed improvements in outward symptoms in mice and Phase 1 trial participants with Alzheimer’s. Willbold’s group has since sold the peptide’s intellectual property rights to the company Priavoid, which Willbold also cofounded and currently supervises as a board member. Priavoid will carry out Phase 2 clinical trials spanning 18 months later this year.

Back to BACEs

A small camp of researchers is also holding out for inhibitors of β-secretase, also known as BACE, to make a comeback. Without the Notch-interfering ability of γ-secretase inhibitors, BACE inhibitors are considered by some researchers to be salvageable. Where previous developers went wrong was likely in the high doses, says Robert Vassar, a neurologist at Northwestern University and one of the leading researchers of these compounds. He says that a lower-dosing “sweet spot” of BACE inhibitors could allow physicians to tiptoe past the toxicities and into the narrow therapeutic window for blocking Aβ.

“It’s kind of walking a tightrope,” Vassar says, but “it’s probably a safe bet.”

And there is evidence to back up that optimism. In one study, mice with one BACE gene deletion had up to 50% lower β-secretase activity than those with two intact copies, had much less Aβ accumulation in the brain, and were protected from the symptoms of Alzheimer’s (J. Biol. Chem. 2007, DOI: 10.1074/jbc.M611687200). Similarly, people with a so-called Icelandic mutation (most common in those of Icelandic or Scandinavian heritage) have a version of APP that is less susceptible to β-secretase cleavage. Middle-aged to elderly individuals with this mutation have 30% lower Aβ levels and have better cognitive function on average than those without (Nature 2012, DOI: 10.1038/nature11283).

These observations suggest that attenuating BACE activity just enough could be a viable strategy to treat Alzheimer’s while bypassing the off-target effects observed with previous attempts. Vassar advocates for revisiting the data in BACE inhibitors’ failed clinical trials to look for subpopulations in which they offered a benefit. He suspects that the BACE inhibitors were unknowingly given to participants who had Alzheimer’s but little amyloid, or perhaps the trials weren’t long enough to reveal the full effects of this slow-acting drug.

Hitting the right dose of BACE inhibitors will be tricky because people will have highly variable inhibition responses. “You cannot achieve that exact level that you need with a BACE inhibitor,” Alzheon’s Hey says. “It’s very difficult from a drug development standpoint to thread the needle.”

Parallel pursuits

Extending the reach of antiamyloid medications requires developing accessible Aβ detection methods in presymptomatic individuals. Without presymptomatic detection, many people who could benefit from the drugs would be missed. While some individuals, such as Tanzi’s 1999 visitor, are aware of their family history and carry the gene for Aβ production in the brain, there are others who aren’t genetic carriers—yet Aβ already lurks in their brains.

“If we can detect who has amyloid building in their brains early enough, 10–20 years before symptoms, we can prevent Alzheimer’s,” Tanzi says. “Think about the godsend it is for these families.”

Aβ can be measured in the brain via positron-emission tomography or in the cerebrospinal fluid collected via a lumbar puncture—a needle inserted into the lower back. These procedures’ high costs and complexities mean neither is feasible for widespread screening.

Several companies have emerged in recent years promising blood tests for Aβ that could make its detection more widespread. C₂N Diagnostics and Shimadzu are rolling out mass spectrometry–based services that can detect the fraction of deleterious Aβ42 among other peptides of its ilk, as diagnostic aids or to identify amyloid-positive candidates for clinical trials. Cheap and needing no special equipment, blood-based Alzheimer’s tests can be folded into routine checkups.

“We’d like a screening test for everybody 40 years of age and up,” says Valerie Daggett, a biophysicist at the University of Washington and CEO of the diagnostics company AltPep. The start-up is working toward a diagnostic assay specifically for Aβ oligomers in the blood (Proc. Natl. Acad. Sci. U.S.A. 2022, DOI: 10.1073/pnas.2213157119).

Reliable blood tests are still a work in progress because scientists haven’t mapped all the nuances of Alzheimer’s. The levels of any biomarker may vary with comorbidities and across demographic factors such as sex and ethnicity.

As far as researchers have advanced in their understanding of Alzheimer’s in recent years, they still have far to go to devise an effective treatment. In addition to Aβ, researchers advocate pursuing other targets to realize combination and personalized therapy for people with the disease. It’s early days yet, but lecanemab’s success, an outlier in the history of Alzheimer’s treatment, has given researchers dawning hope for thwarting Alzheimer’s one day. “It shows that you can do something,” Gandy says. “If you can do something, maybe you can do more.”