CTAD provided excitement, important insights, and new questions, as would be expected for a first-rate meeting focused on the latest word in Alzheimer's trials. In my view, the key take-away was that giving a therapeutic agent which solely targets Aβ was found to significantly slow cognitive decline in AD patients. This result for Lilly’s solanezumab in an 80-week trial was statistically significant at p <0.05 in a prespecified secondary analysis of the total pooled subjects from the EXPEDITION1 and 2 trials. When just the pooled mild subjects (MMSE 20-26 at entry) were analyzed, treatment differences favoring solanezumab over placebo were highly significant on the ADAS-cog11, ADAS-cog14, and even the MMSE (all p   Read more

CTAD provided excitement, important insights, and new questions, as would be expected for a first-rate meeting focused on the latest word in Alzheimer's trials. In my view, the key take-away was that giving a therapeutic agent which solely targets Aβ was found to significantly slow cognitive decline in AD patients. This result for Lilly’s solanezumab in an 80-week trial was statistically significant at p <0.05 in a prespecified secondary analysis of the total pooled subjects from the EXPEDITION1 and 2 trials. When just the pooled mild subjects (MMSE 20-26 at entry) were analyzed, treatment differences favoring solanezumab over placebo were highly significant on the ADAS-cog11, ADAS-cog14, and even the MMSE (all p <0.001). (Importantly, no significant treatment effects occurred in the pooled moderate subjects.) The effect size in the pooled mild subjects may be viewed as modest: a ~34 percent slowing in the degree of cognitive decline by week 80, with the placebo and solanezumab curves starting to separate after week 28. While we all hoped for a greater therapeutic effect, the fact that one-third less decline in widely validated cognitive tests was accompanied by a small (~18 percent; p = 0.045) slowing on a more difficult-to-achieve functional test (ADCS-instrumental Activities of Daily Living) strengthens the conclusion that a modest but real clinical benefit had occurred. And the safety profile of solanezumab appeared good.

As regards biomarker results, why was there no significant decrease in CSF phospho-tau levels with solanezumab, whereas this biomarker showed a small but significant decrease in the bapineuzumab trial despite a lack of cognitive benefit in the analyses reported so far? This apparent paradox might be explained by considering the mechanisms of action of the two antibodies. Solanezumab (and its mouse precursor 266) principally binds and sequesters soluble Aβ monomers, although it remains possible that it also binds some portion of soluble oligomers in the AD brain. By chronically sequestering monomers, solanezumab may cause a gradual shift in Aβ equilibrium in the brain so that soluble oligomers and plaque-associated fibrils slowly decrease in concentration in the cortex, presumably contributing to solanezumab’s clinical benefit over 18 months.

Nevertheless, the levels of some or many neurotoxic oligomers may not go down much with solanezumab. Because soluble oligomers isolated from AD cortex have been shown to induce selective tau phosphorylation and neuritic degeneration (Jin et al., 2012), these changes may not have been sufficiently blunted by the principally monomer-directed solanezumab. On the other hand, bapineuzumab (and its mouse precursor 3D6) binds and neutralizes soluble Aβ oligomers from AD brain quite effectively (Shankar et al., 2008), and it can also affect amyloid plaque burden modestly in AD patients (Rinne et al., 2010). Therefore, bapineuzumab may have bound and neutralized phospho-tau-elevating Aβ oligomers better than solanezumab, but still not enough to slow clinical decline within 18 months. Other reasons for the apparent clinical-biomarker disconnect of the two antibodies could be trial related, for example, that not enough CSF samples from mild subjects were available to detect any small phospho-tau lowering effect that solanezumab might confer.

Many of us who attended CTAD came away with the impression that one reason for the different clinical outcomes of the solanezumab and bapineuzumab trials was the large difference in doses. Although the two antibodies have quite distinct properties that limit the meaningfulness of comparing doses, solanezumab was dosed at 400 mg per month, whereas bapineuzumab was dosed at 0.5 or 1.0 mg/kg (~35-70 mg) every three months due to its considerable risk of inducing ARIA-E. Thus, the cumulative three-month solanezumab dose was ~15-30-fold higher than that of bapineuzumab. Even higher solanezumab doses than this could potentially prove beneficial, whereas bapineuzumab cannot be dosed at such high levels, at least by the iv route. Other reasons for the different clinical outcomes of the solanezumab and bapineuzumab trials may well be revealed by deeper analyses of the data, particularly the correlations of patients’ total antibody exposures to their cognitive and biomarker outcomes.

Regarding other biomarkers used in these trials, vMRI has once again shown the “paradoxical” effect of slightly more hippocampal or whole brain volume loss in these trials. The explanation is unknown, but my speculation is that it may relate to changes in fluid balance in the AD cortex with the partial clearance of Aβ and consequent improvements in the robust inflammatory cell changes that occur in AD. PET imaging of fibrillar amyloid burden showed a trend towards reduction with solanezumab and with bapineuzumab. Thus, each antibody engaged its intended target, and this was further supported by how solanezumab shifted CSF Aβ40 levels: more bound and fewer free monomers.

The central point is that solanezumab’s clinical benefit—while modest at 18 months—confirms that an agent which could only be targeting Aβ can slow cognitive decline and move certain biomarkers in mild AD. This appears to represent the first proof from human trials of the Aβ hypothesis, and strongly encourages new, carefully designed trials of anti-Aβ monoclonals in mild, very mild, and/or presymptomatic AD. The results also recommend other Aβ-lowering approaches such as Aβ vaccination and β- or γ-secretase modulation. The next trials should rely on composite cognitive endpoints that are especially sensitive for the episodic memory deficits of early AD. And biomarkers of neuronal health should continue to be included, both CSF phospho-tau and new CSF indicators of synaptic structure that must be found. Finally, the results presented at CTAD, while suggesting that better Aβ-directed trials are ahead, take nothing away from the importance of accelerating strategies for tau and for cellular inflammation, as multimodal treatments for AD would be highly desirable.

References:
Jin M, Shepardson N, Yang T, Chen G, Walsh D, Selkoe DJ. Soluble amyloid beta-protein dimers isolated from Alzheimer cortex directly induce Tau hyperphosphorylation and neuritic degeneration. Proc Natl Acad Sci U S A. 2011 Apr 5;108(14):5819-24. Abstract

Shankar GM, Li S, Mehta TH, Garcia-Munoz A, Shepardson NE, Smith I, Brett FM, Farrell MA, Rowan MJ, Lemere CA, Regan CM, Walsh DM, Sabatini BL, Selkoe DJ. Amyloid-beta protein dimers isolated directly from Alzheimer's brains impair synaptic plasticity and memory. Nat Med. 2008 Aug;14(8):837-42. Abstract

Rinne JO, Brooks DJ, Rossor MN, Fox NC, Bullock R, Klunk WE, Mathis CA, Blennow K, Barakos J, Okello AA, Rodriguez Martinez de Liano S, Liu E, Koller M, Gregg KM, Schenk D, Black R, Grundman M. 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. Abstract

View all comments by Dennis Selkoe

Soluble β amyloid has a beneficial physiological role for the normal functioning of neurons. In the AD brain, the increase in soluble β amyloid concentrations, mainly via a decrease in its clearance, could simply represent the brain's attempt to medicate the neuronal suffering and death that occur for still-unknown reasons. The decrease in β amyloid clearance would lead to β amyloid accumulation with formation of neurotoxic oligomers and plaque deposition.

I would expect that any drug that lowers soluble β amyloid concentrations in the AD brain will worsen the situation. In many cases of familial AD, point mutations of presenilins lead to a loss of function of the γ-secretase complex, with less ability to produce β amyloid from APP (Chavez-Gutierrez et al., 2012). Consequently, this would lead to a compensatory decrease in β amyloid clearance and then β amyloid accumulation with formation of neurotoxic oligomers and plaque deposition. Again, in familial AD any drug that decreases the brain concentration of β amyloid would worsen cognition. This hypothesis could explain the...  Read more

Soluble β amyloid has a beneficial physiological role for the normal functioning of neurons. In the AD brain, the increase in soluble β amyloid concentrations, mainly via a decrease in its clearance, could simply represent the brain's attempt to medicate the neuronal suffering and death that occur for still-unknown reasons. The decrease in β amyloid clearance would lead to β amyloid accumulation with formation of neurotoxic oligomers and plaque deposition.

I would expect that any drug that lowers soluble β amyloid concentrations in the AD brain will worsen the situation. In many cases of familial AD, point mutations of presenilins lead to a loss of function of the γ-secretase complex, with less ability to produce β amyloid from APP (Chavez-Gutierrez et al., 2012). Consequently, this would lead to a compensatory decrease in β amyloid clearance and then β amyloid accumulation with formation of neurotoxic oligomers and plaque deposition. Again, in familial AD any drug that decreases the brain concentration of β amyloid would worsen cognition. This hypothesis could explain the detrimental cognitive effects observed with semagacestat and avagacestat (both decrease β amyloid levels in CSF), the neutral effects of bapineuzumab (no effects on β amyloid levels in CSF), and the relatively beneficial effects of solanezumab (small increases in β amyloid levels in CSF).

In conclusion, I propose that the accumulation of β amyloid in the brain of sporadic AD patients is a compensatory reaction to neuronal death. In the brain of familial AD patients, it would be an accumulation of β amyloid as a compensatory reaction to loss of presenilin function. API and DIAN prevention trials in presymptomatic subjects could test this unorthodox view.

References:
Chávez-Gutiérrez L, Bammens L, Benilova I, Vandersteen A, Benurwar M, Borgers M, Lismont S, Zhou L, Van Cleynenbreugel S, Esselmann H, Wiltfang J, Serneels L, Karran E, Gijsen H, Schymkowitz J, Rousseau F, Broersen K, De Strooper B. The mechanism of γ-Secretase dysfunction in familial Alzheimer disease. EMBO J. 2012 May 16;31(10):2261-74. Abstract

View all comments by Bruno Pietro Imbimbo

Dr. Imbimbo’s perspective has been voiced before by a minority of AD investigators who believe that lowering brain Aβ levels in various ways would perforce be detrimental to the patient’s cognitive status. There is certainly nothing wrong with revisiting this line of reasoning, as it remains important to openly debate all possible explanations for the complex pathogenic process of AD. Even so, several statements in Dr. Imbimbo’s commentary do not fit with published data.

First, the straightforward declaration that “soluble β amyloid has a beneficial physiological role for the normal functioning of neurons” is more unsettled than he implies. Only a few published studies have addressed the normal function of soluble Aβ at physiological levels, and more than one molecular function has been suggested without widespread confirmation in multiple labs. Importantly, some such papers have examined Aβ42 peptides, and yet studies of normal Aβ should be focusing on Aβ40, since this is ~10-fold more abundant physiologically and since Aβ42 has a confounding tendency to aggregate into...  Read more

Dr. Imbimbo’s perspective has been voiced before by a minority of AD investigators who believe that lowering brain Aβ levels in various ways would perforce be detrimental to the patient’s cognitive status. There is certainly nothing wrong with revisiting this line of reasoning, as it remains important to openly debate all possible explanations for the complex pathogenic process of AD. Even so, several statements in Dr. Imbimbo’s commentary do not fit with published data.

First, the straightforward declaration that “soluble β amyloid has a beneficial physiological role for the normal functioning of neurons” is more unsettled than he implies. Only a few published studies have addressed the normal function of soluble Aβ at physiological levels, and more than one molecular function has been suggested without widespread confirmation in multiple labs. Importantly, some such papers have examined Aβ42 peptides, and yet studies of normal Aβ should be focusing on Aβ40, since this is ~10-fold more abundant physiologically and since Aβ42 has a confounding tendency to aggregate into oligomers (even at nanomolar concentrations) that have quite different biological properties than do Aβ40 monomers. So we need more detailed studies of the normal activities in neurons (and other cells) of endogenous Aβ40 (not synthetic Aβ42) at entirely physiological levels. Pending these, it remains possible that Aβ peptides are simply byproducts of the physiological processing of APP by β- and γ-secretase that, respectively, release the bioactive APPs-β and AICD derivatives, and that the Aβ fragment per se did not acquire an important salutary function during evolution, particularly as regards cognition. More evidence for such a normal function is needed.

Second, there is abundant evidence from numerous labs that soluble Aβ concentrations in the brain are substantially elevated in preclinical and clinical AD subjects versus age-matched normal subjects, and are far above the physiological levels found in young brains (e.g., Naslund et al., 2000). Therefore, the lowering of soluble Aβ by current experimental therapeutics would be very unlikely to bring these high levels fully down to or below those found in the normal cognitive state. Indeed, in the interrupted semagacestat study that Dr. Imbimbo references, no evidence has been published that brain Aβ levels fell precipitously into the normal range, and any adverse cognitive side effects are more likely to be attributable to interference with the normal functions of one or more of the many other substrates (including Notch) that semagacestat could affect.

Third, it is not correct to simply say that “point mutations of presenilins lead to a loss of function of the γ-secretase complex, with less ability to produce β amyloid from APP.” Many presenilin mutations have been shown to increase Aβ42 production with relatively little or no change in total Aβ or Aβ40 levels. The elevated Aβ42/40 ratio can then allow the formation of Aβ42 oligomers, which can induce synaptotoxicity. This sequence is supported by numerous studies of mouse models bearing AD-causing presenilin mutations, as well as by studies of human brain tissue and CSF. Indeed, recent biomarker analyses of young presenilin mutation carriers showed that an early elevation of CSF Aβ42 levels began to decline to subnormal levels some 20-25 years prior to symptom onset, as fibrillar Aβ42 began accumulating in brain parenchyma (Bateman et al., 2012).

Fourth, compelling evidence appeared recently that a lifelong decrease in total Aβ production from inheritance of a novel mutation in APP (A673T) confers strong protection against the development of AD, and it was even associated with significantly less cognitive decline in non-AD elders (Jonsson et al., 2012). The mutation was shown in cell culture to lower Aβ some 40 percent by decreasing the β-secretase cleavage of APP. This “natural human experiment” is perhaps as clear-cut an indication as we can get that chronically subnormal Aβ levels in the brain are not associated with a cognitive detriment, either during brain development or with age.

Fifth, Dr. Imbimbo’s interpretation seems the opposite of what the new solanezumab data actually suggest. An antibody that lowers essentially all forms of soluble Aβ very avidly by targeting its mid-region has been found to significantly slow the rate of cognitive decline on well-validated tests by ~34 percent over 80 weeks, without significant side effects. If Dr. Imbimbo’s concepts were correct that “in the AD brain, an increase in soluble β amyloid concentrations … could simply represent the brain's attempt to medicate the neuronal suffering and death that occur for still-unknown reasons” and that “any drug that lowers soluble β amyloid concentrations in the AD brain will worsen the situation,” then solanezumab’s presumed lowering of soluble Aβ levels in brain (and free Aβ40 levels in CSF) should have had a negative effect on cognition.

Finally, Dr. Imbimbo’s reprise of the longstanding speculation of a “still-unknown reason” for neuronal death in sporadic AD has not yet led to identifying an antecedent precipitant of neuronal death unrelated to Aβ, despite decades of research by hundreds of laboratories. And the notion that “the accumulation of β amyloid in the brain of sporadic AD patients is a compensatory reaction to neuronal death” seems counterintuitive (neurons are the principal source of Aβ) and is not supported by human or animal model data of which I am aware. Indeed, the general hypothesis that Aβ undergoes “compensatory” rises or falls whenever levels of the peptide change in one direction or another requires experimental evidence.

References:
Näslund J, Haroutunian V, Mohs R, Davis KL, Davies P, Greengard P, Buxbaum JD. Correlation between elevated levels of amyloid beta-peptide in the brain and cognitive decline. JAMA. 2000 Mar 22-29;283(12):1571-7. Abstract

Bateman RJ, Xiong C, Benzinger TL, Fagan AM, Goate A, Fox NC, Marcus DS, Cairns NJ, Xie X, Blazey TM, Holtzman DM, Santacruz A, Buckles V, Oliver A, Moulder K, Aisen PS, Ghetti B, Klunk WE, McDade E, Martins RN, Masters CL, Mayeux R, Ringman JM, Rossor MN, Schofield PR, Sperling RA, Salloway S, Morris JC, Dominantly Inherited Alzheimer Network. Clinical and biomarker changes in dominantly inherited Alzheimer's disease. N Engl J Med. 2012 Aug 30;367(9):795-804. Abstract

Jonsson T, Atwal JK, Steinberg S, Snædal J, Jonsson PV, Bjornsson S, Stefansson H, Sulem P, Gudbjartsson D, Maloney J, Hoyte K, Gustafson A, Liu Y, Lu Y, Bhangale T, Graham RR, Huttenlocher J, Bjornsdottir G, Andreassen OA, Jönsson EG, Palotie A, Behrens TW, Magnusson OT, Kong A, Thorsteinsdottir U, Watts RJ, Stefansson K. A mutation in APP protects against Alzheimer's disease and age-related cognitive decline. Nature. 2012 Aug 2;488(7409):96-9. Abstract

View all comments by Dennis Selkoe

Regarding the question of how presenilin FAD mutations affect Aβ production that is being debated here, current research suggests a complex answer. In brief, mutations causing FAD affect γ-secretase processing at three levels to different extents. The most consistent effect is, however, that the mutations interfere with the fourth cycle in the progressive cutting of the Aβ processing, which causes the cleavage of Aβ43 and Aβ42 to Aβ40 and Aβ38, respectively. This probably happens by destabilizing the retention of the peptide in the protease. (For a full explanation, see Chávez-Gutiérrez et al., 2012.)

References:
Chávez-Gutiérrez L, Bammens L, Benilova I, Vandersteen A, Benurwar M, Borgers M, Lismont S, Zhou L, Van Cleynenbreugel S, Esselmann H, Wiltfang J, Serneels L, Karran E, Gijsen H, Schymkowitz J, Rousseau F, Broersen K, De Strooper B. The mechanism of γ-Secretase dysfunction in familial Alzheimer disease. EMBO J. 2012 May 16;31(10):2261-74. Abstract

View all comments by Bart De Strooper

Lilly is pleased that the solanezumab data have stimulated such interesting mechanistic discussions. Whether solanezumab has cognitive effects due to binding to Aβ monomers, equilibria shifts affecting Aβ oligomers, or other species of Aβ (including intracellular), cannot be answered definitively at this time. We will also continue to investigate these mechanistic possibilities, but would like to emphasize that, regardless of the mechanism, we are pleased that the rate of cognitive decline in the pooled Phase 3 studies was slowed by 34 percent.

View all comments by Eric Siemers

The fact that there is a silver lining to the solanezumab trial is welcome news, even though it is tempered by the study’s failure to meet the primary endpoint and that the data were pooled from different trials to gain statistical significance. Whether such results represent a clinically meaningful therapy for patients and their families remains to be seen.

If one accepts the optimistic interpretation and considers the solanezumab trial a success, then the study raises a non-trivial problem for the field; these results invalidate the theragnostic value of the biomarkers and raise questions about the dynamic biomarker trajectory proposed by Jack et al. (1). In one case, the biomarkers moved in the right direction without cognitive benefits (bapineuzumab), and in the other case, cognitive benefits appeared without the biomarker response (solanezumab). These findings pose a significant problem for the current thought process regarding AD pathogenesis, as well as for the planned presymptomatic clinical trials.

The solanezumab study will indeed represent the first proof from...  Read more

The fact that there is a silver lining to the solanezumab trial is welcome news, even though it is tempered by the study’s failure to meet the primary endpoint and that the data were pooled from different trials to gain statistical significance. Whether such results represent a clinically meaningful therapy for patients and their families remains to be seen.

If one accepts the optimistic interpretation and considers the solanezumab trial a success, then the study raises a non-trivial problem for the field; these results invalidate the theragnostic value of the biomarkers and raise questions about the dynamic biomarker trajectory proposed by Jack et al. (1). In one case, the biomarkers moved in the right direction without cognitive benefits (bapineuzumab), and in the other case, cognitive benefits appeared without the biomarker response (solanezumab). These findings pose a significant problem for the current thought process regarding AD pathogenesis, as well as for the planned presymptomatic clinical trials.

The solanezumab study will indeed represent the first proof from human trials of the Aβ hypothesis, if these results are replicated. However, if there is one lesson learned from all the failed trials of the past decade, it is that an amyloid-focused strategy is not effective in patients who already show neurological deficits. Thus, the Aβ hypothesis may have become irrelevant in practical terms for yielding a therapeutic treatment for the patient population that needs it the most.

Starting the therapy in presymptomatic patients is no panacea either. Only a fraction of patients will be willing to undergo the treatment in the absence of neurological symptoms (What, me? I feel fine!). Since there is no way to know with certainty who will develop AD, a substantial fraction of the population will receive unnecessary treatment or overtreatment (e.g., prostate cancer patients). The financial burden of such a potentially lifelong treatment will be unsustainable, especially if started at presymptomatic stages. Finally, there is no guarantee that results from a trial that was successful in a specialized, early-onset, younger AD patient population will be applicable to the much older, heterogeneous, late-onset sporadic AD patients (see, e.g., 2, and comments by Daniel Davis on the paper.

Therefore, practically speaking, we need a therapeutic strategy that will be effective in symptomatic late-onset AD patients, the 99 percent. This goal may be difficult, but not impossible if we allow ourselves to think innovatively and act boldly by exploring alternative therapeutic strategies.

It is gratifying to hear amyloid proponents emphasize the importance of accelerating strategies for alternative pathological mechanisms (such as tau and cellular inflammation). However, in the environment of extremely limited resources, every dollar spent on the amyloid strategy is a dollar taken away from pursuing these alternative strategies. There is a reason why most major pharmaceuticals have moved away from the amyloid approach; will the academic world follow suit?

References:
1. Jack CR Jr, Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, Petersen RC, Trojanowski JQ. Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol. 2010 Jan;9(1):119-28. Abstract

2. Cerreta F, Eichler HG, Rasi G. Drug policy for an aging population—the European Medicines Agency's geriatric medicines strategy. N Engl J Med. 2012 Nov 22;367(21):1972-4. Abstract

View all comments by Sanjay W. Pimplikar