A biomarker signature for Alzheimer’s disease, independent of clinical diagnosis, accurately classified patients with the disease and was unexpectedly present in more than one-third of cognitively normal controls, researchers reported in the August Archives of Neurology.
“Classical biomarker studies in Alzheimer’s disease always considered that the healthy controls—elderly people without disease symptoms—did not have the disease,” Geert De Meyer, PhD, of the Department of Applied Mathematics and Computer Science, Ghent University in Belgium, told Neurology Reviews. “Using an alternative approach, our paper provides (indirect) evidence that this simple assumption does not hold. Alzheimer’s disease might well be present years before the first symptoms appear. We also provide an approach to study this presymptomatic period.”
Investigators tested CSF beta-amyloid protein 1-42, total tau protein, and phosphorylated tau181Pprotein (P-Tau181P) concentrations in a subsample of adults (ages 55 to 90) enrolled in the US Alzheimer’s Disease Neuroimaging Initiative (US-ADNI). The study included 114 cognitively normal persons, 200 patients with mild cognitive impairment (MCI), and 102 patients with Alzheimer’s disease who agreed to undergo a lumbar puncture at baseline.
Dr. De Meyer and colleagues used a novel mixture model, incorporating two components based on an expectation-maximization algorithm. “The mixture model technique takes the assumption that the biomarker distribution is a mixture of two normal distributions—one of the healthy people and one of people with Alzheimer’s disease,” Dr. De Meyer explained. “It uses no prior (diagnosis) information but an iterative probabilistic procedure to establish membership of either distribution.”
Using this technique, the investigators identified two distinct signatures—one related to an Alzheimer’s disease group and the other to a non-Alzheimer’s disease group. “The Alzheimer’s disease signature appears to be ‘naturally’ present in the data and is expressed as a homogeneous group, consistent with a single pathologic process underlying Alzheimer’s disease,” the researchers noted.
Early Evidence of Disease Pathology
Reduced levels of beta-amyloid 1-42 and increased P-Tau181P concentrations were present in 90% of subjects with Alzheimer’s disease, in 72% of subjects with MCI, and in 36% of cognitively normal subjects. The results were validated on two other datasets, including a 94% sensitivity in a set of autopsy-confirmed cases and a 100% sensitivity in predicting disease progression in a group of patients with MCI.
“The unexpected presence of the Alzheimer’s disease signature in more than one-third of cognitively normal subjects suggests that Alzheimer’s disease pathology is active and detectable earlier than has heretofore been envisioned,” Dr. De Meyer and colleagues stated.
“From a medical perspective, it was quite striking to find the Alzheimer’s biomarker signature in about 30% of the cognitively normal controls,” he added. “In these people, the signature was already quite comparable to that in Alzheimer’s patients, suggesting that the disease is already well developed before the first symptoms appear.
“Our paper opens the presymptomatic window of Alzheimer’s disease for therapeutic interventions,” he continued. “To date, Alzheimer’s drugs were not able to reverse the disease when applied in the later stages of the disease. These or other drugs have a higher chance of being effective when given earlier.”
Determining Disease Progression
Of the cognitively normal individuals who displayed the biomarker signature, it was “significantly enriched” in apolipoprotein (APOE) ε4 carriers, the researchers noted. “There was a clear association between mixture model feature and APOE genotype in cognitively normal subjects, with the odds of having an APOE ε4 allele in the Alzheimer’s cluster being 6.88 times the odds for having [the allele] in the non-Alzheimer’s cluster,” Dr. De Meyer’s group wrote.
When comparing biomarker levels between cognitively normal subjects, those with MCI, and those with Alzheimer’s disease, the researchers found no statistically significant differences in beta-amyloid 1-42 concentrations between the diagnosis groups. However, a clear difference was found in P-Tau181P levels. Patients with advanced Alzheimer’s disease had even higher levels of P-Tau181P than those in earlier stages of the disease, a finding that the study authors noted was consistent with a previously observed increase in P-Tau181P during cognitive decline and dementia.
“It also suggests an intrinsically different role for the two biomarkers in the mixture model, with CSF beta-amyloid 1-42 as an initial marker and CSF P-Tau181P as a subsequent stage marker related to dementia symptoms and disease progression,” Dr. De Meyer and colleagues wrote. “This could also imply that CSF beta-amyloid 1-42 is the initial driver of Alzheimer’s disease pathology and that changes in CSF total tau and/or CSF P-Tau181P concentrations are a secondary effect.”
A Reflection of Neuropathology
One important feature of the Alzheimer’s signature biomarker is that it accurately reflects the biochemical processes of the disease—senile plaques composed of amyloid fragments and intraneuronal fibrillary tangles with neuropil threads composed of hyperphosphorylated tau protein, wrote A. Zara Herskovits, MD, PhD, and John H. Growdon, MD, in an accompanying editorial.