Blood-based circRNA signatures may help identify early Alzheimer’s disease biology and progression risk, providing a promising new layer beyond amyloid and tau testing.
Research: Blood-based circular RNA for early diagnosis of Alzheimer’s disease. Image credit: Andrii Vodolazhskyi / Shutterstock
In a recent study published in the journal natural medicineResearchers have identified cyclic ribonucleic acid (circRNA) in the blood that has a high diagnostic accuracy for early Alzheimer’s disease (AD) confirmed by biomarkers. Combining these circRNAs with established markers such as phosphorylated tau-217 (pTau217) provides the highest predictive power. These findings suggest that circRNA investigations may ultimately complement blood-based AD biomarker panels to identify those with early AD biology or at increased risk of progression. However, this finding needs to be validated in a larger and more diverse prospective clinical cohort.
AD is the main cause of dementia. Because pathological changes in this condition appear before cognitive decline occurs, scientists are developing new strategies to detect Alzheimer’s disease early and support timely interventions aimed at slowing disease progression. Early identification of the disease, before clinical symptoms appear, allows for prompt treatment and better clinical planning, which, when combined with effective interventions, may improve outcomes while reducing mortality associated with severe disease.
About research
In the current study, researchers used RNA sequencing (RNA-seq) to analyze blood samples from 1,221 participants, including 405 people with Alzheimer’s disease and 816 adults without cognitive impairment. They aimed to identify and validate blood-based circRNAs that may be useful in Alzheimer’s disease diagnosis and monitoring disease progression. They examined circRNA expression across 33 tissues using the CircAtlas 3.0 database and assessed selected circRNA expression in these tissues using quantitative polymerase chain reaction (qPCR).
The research team calculated area under the curve (AUC) values to determine the diagnostic utility of the blood-based circRNA-based model. They compared their results to blood pTau217 levels to classify biomarker-confirmed AD conditions. The researchers also replicated their results among 551 participants at the Knight Alzheimer’s Disease Research Center (Knight ADRC), including 76 with Alzheimer’s disease and 475 without cognitive impairment. They further tested the model with preclinical anti-amyloid therapy in an asymptomatic Alzheimer’s disease cohort (A4, 1,767 participants). In this model, nearly all participants were cognitively intact at baseline. They used a logistic regression model including the top differentially expressed circRNAs for statistical analysis.
Among Knight ADRC participants, the team evaluated the ability of blood circRNA and pTau217 biomarkers, as well as amyloid PET status, to predict symptom progression. They used a Cox regression model to estimate the hazard ratio (HR) for this analysis.
The research team evaluated the specificity of blood-based circRNAs in disease detection by comparing findings with other neurodegenerative diseases, including Parkinson’s disease (PD), dementia with Lewy bodies (DLB), and frontotemporal dementia (FTD). They also assessed whether the overall 34-circRNA model could predict the progression of dementia severity using Clinical Dementia Rating (CDR) scores. They also conducted sensitivity analyzes stratified by gender, ancestry, and apolipoprotein E4 (APOE4) status. They performed principal component analysis (PCA) to generate genetic ancestry covariates.
result
The research team identified 34 circRNAs associated with AD clinical conditions. The overall 34-circRNA predictive signal increased linearly and consistently from the presymptomatic stage, approximately 2–4 years before symptom onset, to symptomatic AD. Although most of the identified AD-associated circRNAs were highly expressed and showed preferential expression in the brain, this study could not prove that blood circRNAs are brain-derived, and an association with clinical AD was observed regardless of their cognate linear messenger RNA counterparts. The overall circRNA model score was associated with dementia severity and was able to capture dynamic signals of AD progression that may be missed by other pathology-focused biomarkers.
The results were comparable to blood pTau217 levels and were replicated in the A4 and Knight ADRC study groups. The circRNA-based model outperformed blood pTau217 alone for biomarker-confirmed AT- without cognitive impairment AD classification and A+T+ AD classification, achieving an AUC of 0.95 compared to 0.88 for blood pTau217 alone. The team achieved the highest AUC (0.97-0.98) by integrating both biomarkers. The combined model of circRNA and pTau217 helped distinguish between non-progressors and high-risk progressors. This could potentially be useful for monitoring AD progression in the era of new AD therapeutics, especially those targeting amyloid plaques, as circRNAs can signal a wide range of biological changes and symptom progression beyond amyloid pathology.
Blood-based circRNA models also specifically detected AD-related changes and showed poor predictive performance for conditions such as PD, DLB, and FTD. Therefore, these markers may be useful for stratifying progression risk and may be studied for monitoring disease biology beyond amyloid pathology. Among Knight ADRC participants, CircRNA (HR, 2.9) was superior to pTau217 (HR, 1.8) and amyloid PET in predicting progression to the symptomatic stage of AD. Sensitivity analyzes yielded similar results, highlighting the robustness of our main results. The results were broadly similar for European, African, and mixed populations, supporting potential robustness across ancestry, although some ancestry subgroups were small.
conclusion
This finding highlights that blood-based circRNAs are promising, non-invasive, scalable, and highly accurate investigational biomarkers for predicting biomarker-confirmed AD status and symptom progression risk. Based on these findings, circRNA detection in blood may one day be used as an adjunct for early detection of AD, if the findings are validated in large prospective clinical studies. In the future, researchers should also investigate the effects of Alzheimer’s disease-related comorbidities on blood-based circRNA levels.
This finding is particularly relevant because circRNAs are highly stable, tissue-specific, and can be measured in blood. This approach may be clinically useful, as traditional AD biomarker assessment often relied on cerebrospinal fluid (CSF) obtained by lumbar puncture or expensive amyloid PET scans.
Want to read it later? Click here to download a PDF copy.
