Abstract
Cancer, a formidable global health challenge, witnessed a staggering toll in 2020, with over 19 million new cases and approximately 10 million cancer-related deaths reported worldwide. Notably, cancer is a highly heterogeneous disease, exhibiting profound variations across patients in terms of diverse molecular factors, including genetic mutations, epigenetic modifications, and the intricacies of the tumor microenvironment. Characterizing cancer heterogeneity is essential for enhancing cancer prevention, early detection, and treatment strategies. Technological advancements, exemplified by next-generation sequencing and single-cell analysis, have yielded numerous multi-omics profiles and provided valuable insights into the complexity and heterogeneity that underpin tumors.Prior investigations into the dissection of cancer heterogeneity primarily relied on omics data. Given the profound association between aging and the gradual deterioration of biological functions and molecular and cellular degenerations, the process of cancer could be considered a disease of aging. Nevertheless, our current understanding regarding the interplay between aging and the diverse biological properties exhibited within tumors remains limited. From the perspective of aging, we could explore the tumor progression and clinical features of young-onset patients from a unique angle.
Distinct molecular characteristics among different subtypes of cancers contribute to divergent clinical outcomes. Thus, the dissection of subtype-specific regulatory mechanisms could identify potential therapeutic target designing. Long non-coding RNAs (lncRNAs) have also emerged as crucial regulators in tumorigenesis and tumor progression. However, little is known about how lncRNAs function in specific cancer subtypes. Consequently, there exists an imperative requirement to comprehensively unravel cancer heterogeneity from an aging perspective, elucidate the underlying regulatory mechanisms specific to molecular subtypes, and devise dependable tools for the accurate prediction of cancer subtypes.
The work in this thesis focuses on utilizing genome-wide data analysis to dissect cancer heterogeneity from an aging perspective and elucidate cancer subtype-specific lncRNA regulatory mechanisms underlying the poor prognosis subtypes. The contents of each chapter are summarized below:
Chapter 1: An introduction to molecular subtyping and the insights of the aging process into cancer progression. We would discuss the current molecular subtyping systems and clinical characterizations as well as epidemical features of subtypes in breast cancer (BC), gastric cancer (GC), and head and neck cancer (HNSC). Subsequently, the aging process, its hallmarks, and its association with cancer will be discussed. We then summarize the advantages, commonly used data repositories, and current limitations of molecular subtype applications.
Chapter 2: In this chapter, we implemented intrinsic subtyping on BC and dissected the multi-omics characterizations in different age groups, showing that the age-dependent prevalence of molecular subtypes primarily influenced the poor prognosis and aggressive biology of young-onset BC patients. We then constructed a DNA methylation aging clock to quantitively measure the aging process, which indicated that young-onset BC patients had faster aging acceleration. Furthermore, a classifier using spatial organization features from histological images was developed and could accurately predict the aging acceleration-based BC subtypes.
Chapter 3: In this chapter, we, in the first of its kind, developed an R package GCclassifier to predict the molecular subtypes of GC using gene expression profiles in three different systems. The independent validation confirmed the reliability of GCclassifier, and a Shiny web tool was developed to facilitate its usage for clinicians and biologists. Moreover, an integrative network approach was implemented to identify potential subtype-specific master regulator lncRNAs and the possible mechanism of lncRNAs functioning as competing endogenous (ceRNA) to sponge miRNAs, thereby repressing the expression of target genes in GC.
Chapter 4: In this chapter, we proposed an integrative network method to dissect the subtype-specific lncRNA regulatory mechanism underlying the mesenchymal subtype of head and neck cancer (HNSC), followed by the identification of master regulatory lncRNAs mediating the epithelial-mesenchymal transition (EMT) pathway. Finally, using comprehensive drug sensitivity analysis, we found that the five putative targets of identified master regulator lncRNA were also predictive of the sensitivities of multiple drugs, some of which are FDA-approved.
Chapter 5: In this chapter, we provide a comprehensive summary of the discoveries presented in this thesis while delineating the prospects that emerge from our research.
| Date of Award | 10 Sept 2024 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Rebecca Y M CHIN (Supervisor) & Xin WANG (Co-supervisor) |
Cite this
- Standard