Date of Award

Fall 1-1-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Biomedical Engineering (ENAS)

First Advisor

Fan, Rong

Abstract

Glioblastoma (GBM) is a grade IV astrocytoma, the most common malignant brain tumor in adults, which has a poor prognosis and high recurrence. To better understand the disease's fundamental mechanism, tremendous effort has been made to investigate the molecular and cellular interactions between malignant cells and the tumor microenvironment using single-cell technology. However, the spatial intra-tumoral information of GBM, including the disease progression regulatory network and cellular dynamics, has yet to be widely constructed. Our investigation undertakes a novel approach to address the challenges. Specifically, we employ different spatial omics technologies at all levels from epigenetics, transcriptomics, post-transcriptomics, and proteomics, utilizing the microfluidic-based Deterministic Barcoding in Tissue (DBiT) technology for epigenetics and transcriptomics as well as the subcellular-resolution high-plex fluorescent imaging for proteomics. The spatial landscape of specimens derived from different patients with regions representing both magnetic resonance imaging (MRI) contrast-enhanced and non-contrast-enhanced domains was investigated. With the integration of single-nucleus ATAC-seq (Assay for Transposase-Accessible Chromatin), multi-ome results, and public single-cell RNA-seq databases, we not only assembled the complexity and inherent intra-tumoral heterogeneity of the GBM microenvironment but also unraveled the dynamic cellular interactions and the cell state transitions. The bioinformatics analysis began by distinguishing marker genes, delineating the phenotype identifications of GBM subtypes, and other cell identities. Then the spatial ligand-receptor analysis dives into neighborhood-to-cell interactions that show the cellular regulatory network and interactome. Using the maximal entropy approximation, we also performed the surprisal analysis, which deconvolutes the hidden pattern that suggests the disease progression. Specifically, the neuronal-cancer interaction was investigated, and the spatial correlation between hijacked neuron behaviors and cancer cells was found in detail. Beyond that, we developed the in situ poly-adenylation integrated platform, Patho-DBiT, which provides the potential of capturing the whole spectrum of RNA in the clinically archived samples. Our preliminary data demonstrated the potential of employing the FFPE samples to demonstrate the spatial prevalence of onco-miRNA in GBM. Collectively, these datasets provide the first fine-scale, multi-layered atlas of GBM epigenome, transcriptome, and interactome across spatially distinct niches. The work clarifies gene-regulatory programs that underlie cellular plasticity and progression and establishes computational pipelines, including multi-modal pseudo-time reconstruction and spatial interactome, for dissecting disease dynamics. These findings enhance our comprehension of GBM biology and offer promising pathways for future research and therapeutic interventions in the battle against this formidable disease.

Download

Share

COinS