Effect of Pathway-Specific Polygenic Risk Scores for Alzheimer’s Disease (AD) on Rate of Change in Cognitive Function and AD-Related Biomarkers Among Asymptomatic Individuals

Yuexuan Xu, Eva Vasiljevic, Yuetiva Deming, Erin M. Jonaitis, Rebecca L. Koscik, Carol A. Van Hulle, Qiongshi Lu, Margherita Carboni, Gwendlyn Kollmorgen, Norbert Wild, Cynthia M. Carlsson, Sterling C. Johnson, Henrik Zetterberg, Kaj Blennow, Corinne D. Engelman

Abstract

In addition to APOE, pathway-specific polygenic risk scores (p-PRSs) can predict age-dependent changes in beta-amyloid, tau, and cognition. Once validated, they could be used to identify individuals with an elevated genetic risk of accumulating beta-amyloid and tau, long before the onset of clinical symptoms.

Introduction

Late-onset Alzheimer’s Disease (LOAD) is a highly heritable, complex condition characterized by a "slow burn" of pathological changes occurring decades before clinical symptoms emerge. While the APOE gene is the most significant single genetic risk factor, it does not fully explain the variance in disease onset or progression. Traditional Polygenic Risk Scores (PRS) aggregate hundreds of small-effect genetic variants into a single metric, which can obscure the specific biological mechanisms—such as lipid metabolism or immune response—driving an individual's risk.

This study investigates whether breaking down genetic risk into pathway-specific PRSs (p-PRSs) provides a more nuanced understanding of disease progression. By focusing on specific biological "neighborhoods," researchers aimed to predict the rate of change in cognitive function and cerebrospinal fluid (CSF) biomarkers (beta-amyloid and tau) in healthy, asymptomatic individuals as they age.

Methods

The researchers mapped 23 significant genetic variants to six distinct biological pathways identified in previous genome-wide association studies (GWAS):

  1. APP metabolism (Amyloid Precursor Protein processing)
  2. Cholesterol metabolism
  3. Endocytosis (Cellular absorption processes)
  4. Tau pathology
  5. Immune response
  6. Axonal development

Statistical Modeling

The study utilized linear mixed-effects models to analyze longitudinal data. This approach accounted for familial relationships and repeated measurements over time. A key feature of the methodology was the inclusion of interaction terms between the p-PRSs and age. This allowed the team to test the hypothesis that genetic risk is not a static burden but one that accelerates pathological changes as individuals grow older. Models were analyzed both with and without the APOE gene to isolate the contribution of other genetic pathways.

Datasets

  • Discovery Cohort: The Wisconsin Registry for Alzheimer’s Prevention (WRAP), consisting of 1,175 initially healthy individuals (mean age ~59) followed for up to 10 years. A subset of 197 participants provided longitudinal CSF samples.
  • Replication Cohort: The Wisconsin Alzheimer’s Disease Research Center (ADRC), used to validate findings in an independent group (427 for cognition; 259 for biomarkers).

Key Insights and Results

The 15-Year Predictive Advantage

One of the most striking findings was that when APOE was excluded, p-PRSs for the APP and cholesterol pathways predicted changes in beta-amyloid levels approximately 15 years earlier than a standard, genome-wide PRS. This suggests that these specific pathways are the "early responders" in the Alzheimer's pathological cascade.

Age-Dependent Genetic Impact

The study confirmed that genetic risk is highly dependent on age. The impact of these genetic variants on the rate of cognitive and biomarker decline was found to accelerate significantly after age 65–70. This underscores the importance of longitudinal monitoring in the preclinical phase.

Sequential Pathological Changes

The data revealed a clear chronological order in how genetic risk manifests:

  • Ages 55–60: Significant genetic effects on beta-amyloid (Aβ42) levels emerge.
  • Age 65: Genetic influence on tau pathology becomes detectable.
  • Ages 65–70: Genetic risk begins to significantly impact cognitive decline.

APOE vs. Pathway Nuance

While APOE remains the dominant driver for amyloid accumulation, it has a notably smaller impact on tau pathology (the neurofibrillary tangles associated with cell death). Non-APOE genetic factors, particularly those involved in endocytosis, were found to be more critical for predicting tau levels and cognitive decline across multiple domains.

Discussion and Limitations

The ability of p-PRSs to predict biomarker changes decades before clinical symptoms suggests they could serve as a powerful "genetic early warning system." Identifying which biological pathway is most compromised in an individual could eventually allow for personalized preventative strategies or more targeted enrollment in clinical trials.

However, several limitations remain:

  • Ancestral Bias: The study was limited to non-Hispanic white participants. Because genetic risk scores are often population-specific, these results may not generalize to more diverse cohorts.
  • Conservative Mapping: The strict criteria used to assign genes to pathways (requiring >50% literature consensus) may have excluded relevant but less-studied variants.
  • Sample Size: While the cognitive data was robust, the CSF biomarker sub-samples were relatively small, necessitating further validation in larger longitudinal studies.

Conclusion

Pathway-specific polygenic risk scores offer a more granular view of Alzheimer’s risk than global scores. By identifying the specific biological drivers of amyloid and tau accumulation, p-PRSs provide a window into the preclinical phase of the disease, offering the potential for intervention long before the first signs of cognitive impairment appear.

Summary updated

Abstract

In addition to APOE, the p-PRSs can predict age-dependent changes in amyloid-β, tau, and cognition. Once validated, they could be used to identify individuals with an elevated genetic risk of accumulating amyloid-β and tau, long before the onset of clinical symptoms.
BibTeX
@article{xu2023effect,
  author = {Yuexuan Xu and Eva Vasiljevic and Yuetiva Deming and Erin M. Jonaitis and Rebecca L. Koscik and Carol A. Van Hulle and Qiongshi Lu and Margherita Carboni and Gwendlyn Kollmorgen and Norbert Wild and Cynthia M. Carlsson and Sterling C. Johnson and Henrik Zetterberg and Kaj Blennow and Corinne D. Engelman},
  title = {{Effect of Pathway-Specific Polygenic Risk Scores for Alzheimer’s Disease (AD) on Rate of Change in Cognitive Function and AD-Related Biomarkers Among Asymptomatic Individuals}},
  journal = {Journal of Alzheimer s Disease},
  year = {2023},
  doi = {10.3233/jad-230097},
}
Figure