Abstract:
Bloom Syndrome (BS) is a recessive genetic disorder characterized by hyper- recombination and genome instability. It is caused by mutations in the conserved RecQ helicase gene, BLM, which is essential in maintaining genome integrity given its ability to unwind various aberrant DNA structures arising in DNA metabolic pathways, including a four-stranded helical structure called G-quadruplexes (G4). Clinically, BS individuals manifest developmental delay, immune deficiencies, a shorter lifespan and elevated cancer risks. Molecularly, BS is characterized by excessive sister chromatid exchange (SCE) events and genome instability. Recent studies hinted at the relevance of G4 in BS by the enrichment of putative G4-forming sequences in the differentially expressed genes and SCE events. However, having G4-forming sequences does not predict the formation of G4 in the cells, and it remains unclear if endogenous G4 structures play a pivotal role in BS. Given G4's regulatory roles in DNA replication, transcription, and chromatin organization, I hypothesized that BLM deficiency leads to perturbed G4 formation and resolution, causing the downstream molecular changes in BS.
To characterize the molecular changes in BS, we collected cell lines derived from sex- matched and roughly age-matched BS and healthy donors (WT). I assayed the molecular changes via ATAC-seq and RNA-seq. Importantly, instead of inferring the G4 formation by in silico predicted G4 motifs or in vitro validated G4-seq hits, I utilized G4 ChIP-seq and an antibody specific to G4 structures to map the endogenous G4. I found that in BS, increased G4 formation correlated with both increased chromatin accessibility and gene expression, and vice versa. Next, by applying a G4 stabilizing molecule to WT to mimic the defective G4-resolving abilities in BS, I showed that G4 stabilization partially phenocopies BS across both cell types, suggesting a partially causal role of G4. To dissect the molecular mechanism, I observed that regions/genes with the presence of G4 displayed elevated chromatin accessibility and gene expression. My results suggest a plausible molecular mechanism wherein changes in G4 formation directly influence local chromatin accessibility, thereby regulating gene expression. As a validation for this hypothesis, I collected ATAC-seq data from a BS family and showed that more accessible chromatin regions in BS individuals exhibited stronger enrichment for G4-forming sequences.
During data analysis, I noticed the impacts of copy number variation (CNV) in differential analyses. In chapter two, I demonstrated that CNV between contrasted samples can and does drive much of the observed differential signals. However, the impact of CNV is often overlooked despite it being common in diseases like cancer. I therefore developed a differential analysis pipeline featuring copy number (CN) normalization. I showcased its application and advantages using two examples in biomedical studies. Firstly, I applied it to ATAC-seq and ChIP-seq data generated from cell lines with complex chromosomal aberrations derived from a BS individual and a healthy donor. Using a conventional copy-number blind pipeline, differential signals were heavily skewed toward the sample with relatively higher copy numbers in the corresponding regions. Notably, applying our pipeline with CN normalization efficiently distinguishes differential signals driven by CNV and those due to the disease. In the second case, I applied it to ATAC-seq data generated from trisomy 21 and euploid cell lines. By combining the results from our pipeline with the common workflow, I was able to distinguish among open chromatin regions on chromosome 21 with dosage effects, compensatory effects and copy-number-independent regulatory changes.
Overall, this thesis presents the first study with direct evidence that G4 structures could emerge as a central factor in the molecular etiology of BS. Upon the loss of function of BLM, defective G4-resolving abilities lead to G4 formation changes, which subsequently trigger downstream molecular alterations in BS, thereby contributing to the clinical phenotypes observed in affected individuals. My findings enrich the understanding of Bloom Syndrome at a molecular level as well as the regulatory function of G4 in vivo and highlight the broad regulatory function of BLM, which extends beyond its well-established role as a helicase. Additionally, this thesis delivers the concept of CN normalization for differential analyses of count-based functional genomic assays.
BLM