Chromatin topology defines cell identity and phenotypic transition in human cancer and fungal pathogen

• Main Author: Chen, Yao
• Other Authors: Amartya Sanyal
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Science::Biological sciences
• Online Access: https://hdl.handle.net/10356/140893
• Institution: Nanyang Technological University

Higher-order chromatin organization plays an important role in epigenetic regulation of gene expression. Chromatin topology has been increasingly recognized to define cellular identity. Alterations in chromatin conformation are often associated with phenotypic changes and can even lead to cellular state transitions. The developments of chromosome conformation capture (3C) and its derived techniques have provided molecular approaches to study chromatin interactions based on proximity ligation. Using genome-wide derivatives of 3C such as Hi-C or 3C-seq, three-dimensional chromatin architecture inside the nucleus could be characterized. Comparison of interphase chromatin structures at different cellular states could therefore contribute to the understanding of underlying gene regulatory mechanisms during phenotypic transitions in development, differentiation and disease. Furthermore, the frequency of intra-chromosomal contacts between two chromatin regions decreases with increasing linear genomic distance. This fact can be employed for scaffolding of contigs to facilitate a rapid construction of the chromosome-level genome assembly. In the multiple studies included in this thesis, we explored the role of three-dimensional chromatin structure in defining cellular identity and in achieving phenotypic changes. Chapter 1 begins with a concise account about chromatin biology from historical perspective followed by a review of a few models of DNA packaging to understand the structural basis of chromatin function. We also discussed the impact of genomics and post-genomics era, which witnessed a huge increment in the genome assembly and annotation of genetic elements and chromatin features in various organisms. All these genome-wide datasets, as well as their large-scale analysis and integration due to tremendous advancement in computational and bioinformatics approaches, provide the necessary foundation for the modern understanding of chromatin function and gene regulatory system. Next, recent researches on higher-order chromatin organization using 3C-based techniques have been elucidated. This chapter concludes with an overview of the studies carried out in the thesis. In Chapter 2, we studied drug-induced changes of chromatin topology in cancer chemoresistance, one of the major challenges in cancer treatment. Although many genes were found to be involved in cancer drug resistance, the underlying gene-regulatory mechanism remains poorly characterized. Since chromatin organization is known to regulate gene expression during cellular state transition, we hypothesized that chromatin interaction network changes as cancer cells develop resistance to anticancer drugs. To understand the role of disease-causing chromatin alterations in drug resistance, we performed 3C-seq using doxorubicin-resistance cell line model derived from human small cell lung cancer (SCLC) cells. These cell lines include drug‐sensitive NCI‐H69, its in vitro multidrug resistant variant H69AR as well as the revertant H69PR that is resensitized to doxorubicin after a drug holiday. The genome-wide chromatin contact maps showed different genomic rearrangements in the three cell lines, especially the widespread translocations in H69AR, suggesting karyotype alterations during development of drug resistance. Further analysis revealed the switching of active and repressed compartments between sensitive and resistant cells, which is associated with concomitant changes in gene expression and enhancer profiles. Moreover, the comparison of topologically associating domain (TAD) boundaries suggests a relatively low similarity of TAD structures between cell lines. We found that changes of TAD boundaries are associated with alterations of gene expression patterns and enhancer activities, possibly through rewiring of enhancer-gene interactions. Furthermore, to study the time-resolved changes in chromatin interactions during the development of drug resistance, we have generated in-house drug-tolerant cells by culturing NCI-H69 cells in increasing doses of doxorubicin. The genome-wide chromatin interaction analysis revealed changes in higher-order chromatin organization at different levels between doxorubicin-tolerant H69DT cells and their age-matched control H69AMC cells, suggesting that genomic rearrangements, A/B compartment switching and TAD alterations are frequent in cancer cells like NCI-H69 under drug-selection pressure. Additionally, we found that some new genes are differentially expressed between in-house drug-tolerant and control cells. These genes are not differentially expressed between NCI-H69 and H69AR, suggesting that cancer cells can utilize different sets of genes to develop drug tolerance in response to the same drug challenge. In Chapter 3, we proposed a rapid and low-input method for genome-wide chromatin interaction study. The current genome-wide variants of the 3C technique have provided the mainstream strategy for characterization of the three-dimensional chromatin organization. However, the extended timeline and complex experimental procedures largely restrict their applicability in clinical settings, where the priority lies in the ease of protocol using a small sample in a limited time-scale. Therefore, we developed FANT-C, a 10-hour simple and straight-forward experimental protocol to prepare 3C library from a low-input sample (as low as 50,000 cells). This library can then be sonicated into shorter fragments and sequenced using next-generation sequencing (NGS) to detect genome-wide chromatin interactions. In addition, a large proportion of FANT-C reads contain genomic DNA sequences that can provide information about genetic aberrations apart from the intended chromatin interaction information. Currently in clinics, detection of genetic aberrations by NGS is emerging as a primary method for molecular characterization of tumors and clinical decision-making of cancer patients. Hence, FANT-C can be applied to cancer biopsy samples to get additional information about chromatin interactions together with profiles of genetic abnormalities while taking care of patient comfort by avoiding multiple sample collection. In a proof-of-concept study, we performed FANT-C using 50,000 K562 cells and showed that, at different levels of chromatin organization, the FANT-C dataset is comparable with publicly-available conventional Hi-C and in situ Hi-C K562 datasets. We next implemented FANT-C in a longitudinal follow-up study using acute myeloid leukemia (AML) patient samples before induction therapy and after complete remission. Switching of chromatin compartments and alterations in TAD structures were detected between pre- and post-therapy samples. In addition, we detected an inversion in chromosome 16 of the patient with AML subtype M4Eo before and after therapy. Using the genomic DNA sequences from FANT-C dataset, we found chromosome 22 trisomy in the same patient after therapy. The ability of FANT-C to simultaneously provide information about chromatin conformation as well as genomic abnormalities of clinical samples within diagnostic timeline empowers its potential application in NGS-based routine cancer test. In Chapter 4, we incorporated chromatin interaction data for scaffolding and characterizing the genome of Candida tropicalis. As a member of the CUG-Ser1 clade of the fungal phylum of Ascomycota, C. tropicalis is a common and virulent fungal pathogen in human. The increasing cases of drug resistance in C. tropicalis have severely constrained the therapeutic strategy to combat this disease. Previous studies using Candida albicans, the well-studied organism from the same clade, revealed large-scale genomic changes during the development of drug resistance. However, the genomic features of C. tropicalis remain largely unknown due to the lack of complete genome assembly. Meanwhile, comparison of genomes in C. albicans and C. tropicalis, the two closely-related species, could potentially indicate the driving force of speciation in their common ancestor. By combining single-molecule real-time (SMRT) sequencing and 3C-seq data, we successfully constructed the complete chromosome-level genome assembly of C. tropicalis type strain MYA-3404 based on previously-available contig information. Similar to C. albicans, the genome-wide chromatin contact map of C. tropicalis revealed strong centromere-centromere interactions as well as telomere-telomere interactions. We also identified a balanced translocation between one homolog of chromosome 1 and 4 in the diploid genome. Furthermore, the read depth signal generated from the genomic DNA reads in the C. tropicalis 3C-seq dataset showed segmental amplifications on chromosomes 4 and R respectively. Orthologs of a few genes in these regions are known to be involved in pathobiology and antifungal drug resistance in C. albicans. Finally, genome-wide synteny analysis was performed to study the karyotype changes between seven chromosomes of C. tropicalis and eight chromosomes of C. albicans, which identified centromeres as possible hotspots for ancient chromosomal rearrangements during karyotype evolution in their common ancestor. Chapter 4 also includes part of our results in another study, where we analyzed publicly-available Hi-C dataset of C. albicans to examine the cis and trans chromatin interactions in centromere-proximal regions. We identified ~25 kb specialized pericentric domains surrounding the centromeres which form a compact chromatin neighborhood. Segregated from the bulk chromatin, this centromere-flanking compact chromatin (CFCC) was shown to restrict centromere activities, such as neocentromere formation, to the pericentric domains upon native centromere deletion. Inspired by this study, we examined the chromatin organization in pericentric regions of C. tropicalis. However, our analysis suggests that CFCC may not be present in C. tropicalis. Chapter 5 is the concluding chapter of the thesis which discusses the underlying impact of the research carried out in this thesis in the context of the results produced by earlier studies in the field. We also explored the future research undertakings that may provide deeper insights into some of the conclusions that remain unclear or unanswered in the present study. Overall our study on cellular state transitions, viz., drug sensitive vs. resistant SCLC cells and AML patient samples before treatment and after remission, unveiled the important role of chromatin organization in gene regulatory mechanisms. Next, we have put forward a low-input rapid method of generating 3C library which can be adapted in clinical settings to get additional information about genome-wide chromatin interactions along with the intended information about genetic aberrations in clinical samples. Finally, we applied 3C-seq data in constructing chromosome-level genome assembly of pathogenic yeast C. tropicalis and deciphering the chromatin organization and genetic alterations in this organism. Addtionally, the comparison between two closely-related Candida species, C. tropicalis and C. albicans, suggests interesting differences in their pericentric chromatin organization as well as the potential mechanism of karyotype evolution in their common ancestor.