Transcriptomic Profiling of Senescence- Induced Patient Derived Fibroblasts Reveals Alzheimer’s Disease-Specific Gene Signatures
- 주제(키워드) Alzheimer's disease , Patient-derived fibroblasts , Replicative senescence , RNA sequencing , Machine Learning
- 주제(DDC) 570
- 발행기관 아주대학교 일반대학원
- 지도교수 Hyun Woong Roh
- 발행년도 2026
- 학위수여년월 2026. 2
- 학위명 석사
- 학과 및 전공 일반대학원 의생명과학과
- 실제URI http://www.dcollection.net/handler/ajou/000000035447
- 본문언어 영어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
초록/요약
Transcriptomic Profiling of Senescence-Induced Patient Derived Fibroblasts Reveals Alzheimer’s Disease-Specific Gene Signatures Alzheimer’s disease (AD) is the most common cause of dementia worldwide, and its global burden continues to rise with population aging. Although aging is the strongest risk factor for AD, the molecular and cellular mechanisms linking aging to disease onset remain poorly understood. Existing research models have limitations, as they inadequately capture human aging processes. To address these limitations, we utilized patient-derived dermal fibroblasts, which preserve donor-specific characteristics and undergo replicative senescence through serial passaging in vitro. Fibroblasts from 13 individuals with AD dementia and 13 controls were cultured to model three senescence stages: young (passage 7), mid-old (passage 18), and old (passage 25 to 28). Senescence progression was validated using integrated staining, molecular, and transcriptome-based methodologies. Differential expression analysis of ribonucleic acid (RNA) sequencing data revealed stage-specific transcriptomic shifts that substantially differed between the control and AD dementia groups, informing the construction of a 605-gene senescence-associated gene set for downstream disease prediction model. Machine-learning classifiers trained on the expression profiles of the 605 senescence-associated genes showed that models based on mid-old stage data achieved the highest AD dementia classification accuracy (>0.9), outperforming those derived from the young and old stages, suggesting that senescence characteristics at the mid-old stage may most effectively capture the molecular differences between control and AD dementia fibroblasts. Key mid-old signature genes, including H2AC18, H1-2, and LTBP1, demonstrated strong correlations with established AD biomarkers such as amyloid burden and plasma phosphorylated tau- 217 (pTau217). Application of these transcriptomic signatures to the iLINCS (Integrated Library of Network-based Cellular Signatures) platform further identified putative compounds inversely correlated with the senescence-associated expression patterns, suggesting potential agents capable of counteracting age-related molecular changes. Collectively, this study suggests that patient-derived fibroblasts may serve as a promising ex vivo platform for modeling senescence-associated molecular changes relevant to AD. The mid-old stage emerges as a critical window where AD dementia specific transcriptomic divergence is most prominent, underscoring the importance of senescence-stage stratification for biomarker discovery, early-stage disease characterization, and drug-repurposing strategies targeting aging pathways. Keywords: Alzheimer’s disease; Patient-derived fibroblasts; Replicative senescence; RNA sequencing; Machine Learning
more목차
I. INTRODUCTION 1
A. Alzheimer's disease & aging 1
B. Cellular senescence 2
C. Rationale for using patient-derived fibroblasts 3
D. Study aims 5
II. MATERIALS AND METHODS 6
A. Study participants 6
B. Isolation of patient-derived fibroblasts and cell culture 7
C. Three experimental strategies for inducing cellular senescence 8
D. RNA extraction and sequencing 12
E. Validation of senescence 13
F. Identification of senescence-associated gene signatures 14
G. Functional enrichment and network analysis 15
H. Machine learning-based disease prediction modeling 15
I. Correlation analysis between gene expression and biomarkers 16
J. Transcriptome-based drug repurposing via iLINCS 17
K. Statistical analysis 18
III. RESULTS 20
A. Overall study design 20
B. Demographic characteristics of study participants 22
C. Validation of cellular senescence across senescence stages 24
D. Delayed senescence-associated transcriptional response in AD dementia fibroblasts 26
E. Extraction of a senescence-associated gene set for disease prediction 28
F. Machine learning–based AD prediction across senescence stages 30
G. Mid-old stage as a window of differential gene expression and biomarker correlation 32
H. Perturbagen analysis using iLINCS platform 34
IV. DISCUSSION 37
V. CONCLUSSION 46
REFERENCES 47
국문요약 55
