Toward Bridging Unmet Medical Need in Early Alzheimer's Disease: An Evaluation of Beta-Amyloid (Aβ) Plaque Burden as a Potential Drug Development Tool.

Abstract

This piece expresses our views on the use of reduction in beta-amyloid (Aβ) plaque as a potential tool to facilitate new drug development and patient access to promising therapies for Alzheimer's disease. By summarizing literature from seven anti-Aβ antibodies investigated in late-phase trials, we demonstrate a potential threshold of Aβ plaque reduction for clinical effect. The reduction in Aβ plaque of sufficient extent shows a relationship with clinical improvements as measured by a standard end point. Alzheimer’s disease (AD) is a progressive neurodegenerative disorder affecting ~ 6 million Americans 65 years and older, resulting in tremendous adverse clinical and economic impacts on patients and their families, caregivers, and health systems. Despite significant investments to identify the mechanisms of disease pathogenesis and prognosis, there has been a relative paucity of novel therapeutics for AD over the past several decades. Drug development tools such as biomarkers and animal models can help facilitate the development of safe and effective treatments for AD by increasing our understanding of (i) the role of the pharmacological target in modulating disease symptoms or progression, (ii) what would constitute adequate drug exposure at the target site, (iii) patient-specific factors associated with variable disease progression and/or drug response (e.g., to enable appropriate patient enrollment in clinical trials), and (iv) any off-target effects of a novel therapeutic agent. By using these critical drug development tools, we can bridge our knowledge gaps, facilitate drug development for this debilitating disease, and help expedite patient access to promising new therapies. Accumulation of brain Aβ with the subsequent formation of extraneuronal Aβ plaques and intraneuronal neurofibrillary tangles composed of hyperphosphorylated tau protein is hypothesized to be the major pathogenic mechanism for AD. Interrupting Aβ accumulation and removing existing aggregated Aβ plaque to prevent further amyloid cascade have been considered plausible pharmacological mechanisms for new compounds that aim to slow disease progression. Over the past few decades, a total of seven anti-Aβ immunoglobulin G (IgG) antibodies developed to remove Aβ have been investigated in late-phase controlled clinical trials (i.e., large phase Ib/II trials and/or phase III trials) in patients with early stages of disease (i.e., mild cognitive impairment, prodromal AD, or mild to moderate AD dementia): solanezumab, crenezumab, bapineuzumab, gantenerumab, donanemab, lecanemab, and aducanumab. Given the critical unmet medical need in AD, we examined whether available evidence supports the use of change in Aβ plaque burden as a potential biomarker that would be reasonably likely to predict a clinical benefit in patients with early AD across the development programs of these seven anti-Aβ IgG antibodies. Specifically, we evaluated results from publicly available late-phase trials that were randomized, placebo-controlled, double-blind studies investigating various biomarkers, including the change of Aβ plaque burden, and clinical outcomes following a long period of treatment (i.e., ≥12 months). Additionally, given the historical late-phase clinical trial failures of several anti-Aβ IgG antibodies, we sought to evaluate whether these failures can be explained by the nature of their effects on Aβ plaque burden. All trials used the Clinical Dementia Rating Scale–Sum of Boxes (CDR-SB) as a clinical outcome measure; CDR-SB is a quantitative index derived from assessment in six domains of patient functioning: memory, orientation, judgment / problem solving, community affairs, home and hobbies, and personal care. It is currently considered a validated clinical outcome measure to assess efficacy of new drugs being developed for AD. To ensure cross-program consistency in the analysis, placebo-adjusted, baseline-corrected CDR-SB was used as the clinical outcome measure for correlative analysis with changes in Aβ. Across the drug development programs of solanezumab, crenezumab, bapineuzumab, gantenerumab, lecanemab, and aducanumab, brain Aβ plaque burden was quantified as a composite standardized uptake value ratio (SUVR), which describes the composite ratios of radiotracer uptake in brain regions of interest (with Aβ pathology) to that in a reference region (regions without or with only minimal Aβ pathology). In the aducanumab program, Aβ plaque burden was also reported using Centiloid scaling, which generates standardized quantitative Aβ plaque levels independent of methodological approach to image acquisition. Only Centiloid scaling was reported for the donanemab program. The association between placebo-subtracted change in Aβ plaque burden and CDR-SB was explored using observations in patients receiving treatment for at least 1 year with most data collected following 1.5 years of treatment. This was intended to account for the potential delay between changes in Aβ plaque burden and effects on clinical outcomes. As shown in Figure 1, there is a relationship between change in Aβ plaque burden and change in CDR-SB.1-8 At the tested doses, solanezumab, crenezumab, bapineuzumab, and gantenerumab, which all showed minimal changes in Aβ plaque level, consistently failed to demonstrate superiority over placebo in slowing the disease. However, drugs that showed greater changes in Aβ plaque burden were associated with improvement in CDR-SB measures compared with placebo as seen in the aducanumab and lecanemab programs. Only the high-dose group in Study 301 of aducanumab seems to show an inconsistent pattern. Observations from a seventh compound, donanemab, appear to suggest a similar finding, i.e., that adequate reduction of Aβ plaque burden, as measured by Centiloid scaling, is associated with improvement in CDR-SB. Although our findings show a relationship between change in Aβ plaque level and change in CDR-SB, there are limitations to the current analysis. Our analysis was limited to clinical trials of patients with early AD, and it is unclear if such a relationship exists for other stages of AD. Importantly, there is expected to be a threshold effect for Aβ plaque burden reduction that leads to a clinical effect, so that a minimum level of reduction is needed to translate into improvement in CDR-SB. The quantitative value is yet to be robustly defined. One important distinction to make is that we assessed the relationship between Aβ plaque burden and CDR-SB at the randomized group level, not the individual patient level. Establishing biomarker and clinical outcome relationships based on group level observations is a standard approach. Because of randomization, using observations at group level avoids confounders (e.g., baseline disease severity) that attenuate the underlying relationship. In addition, this approach was intended to address the key question of whether a group-level change in the imaging parameter (e.g., in the context of a clinical trial) is likely to result in a group-level change in the clinical end point measure. A significant relationship might allow for regulatory actions intended to consider population-level effects. Our analysis does not establish a relationship between these measures at the individual patient level and does not allow for predictions of patient-level responses (e.g., for treatment decisions in clinical practice). Finally, our analyses were limited to anti-Aβ IgG antibodies and should not be extrapolated to therapeutic modalities with other mechanisms of action. In summary, our analyses suggest that the reduction in amyloid plaque of sufficient extent shows a relationship with improvement in AD, as measured by a standard trial end point. Biomarkers, such as Aβ plaque burden, offer potential to bridge our knowledge gaps in hopes of facilitating new drug development and patient access to promising new therapies for AD. The authors would like to thank Kevin Krudys, Kimberly Bergman, Lauren Milligan, Qi Liu, and Issam Zineh for their helpful input in the preparation of the manuscript. No funding was received for this work. The authors declared no competing interests for this work. This article reflects the views of the authors and should not be construed to represent official FDA policy.