Functional Genomics of Germ Cell Tumors: from balls to bytes and back again

Abstract

The work discussed in this thesis explains the role of the functional genome in germ cell tumor (GCT) pathogenesis by applying newly developed and existing computational methods to (genome-wide) functional genomic datasets. Specifically, epigenetic and (post-)transcriptional regulation in GCTs was studied to gain a deeper understanding of disease pathogenesis, also aiming at clinical application of the findings. GCTs are a unique class of neoplasms originating from (fetal) developing germ cells. Five subtypes can be distinguished (I-V), which are related to physiological germ cell development. Type I GCTs, also called infantile or pediatric GCTs, are rare and generally benign. Type II GCTs, also called germ cell cancer (GCC), include a heterogeneous set of histological subtypes with clearly defined totipotent stem cell components. GCC accounts for 60% of all cancers in Caucasian males between the ages of 20 and 40. Type III, IV and V GCTs are generally benign. They originate from more mature germ cells. Generally speaking, the spectrum of GCT subtypes represents a unique class of neoplastic entities in relatively young patients. Moreover, GCTs show strong and traceable onco-fetal roots advantageous for clinical applications. The functional genome includes a broad range of interacting regulatory features in a cell that determine its identity and behavior. These features are therefore highly relevant in the investigation of GCT pathogenesis and underlying biological processes. Functional genomics comprises the field of study targeted at integrating multiple (genome-wide) datasets, aiming to better understand these interactions. A general introduction into the field of GCTs and functional genomics is provided in chapter 1 of this thesis. Part one / chapter 2 gives a review of the onco-fetal roots of GCC. It illustrates that GCC is a fascinating model for studying onco-fetal processes like cell cycle control, pluripotency maintenance and KIT-KITLG signaling. In the second part of this thesis onco-fetal proteins and micro-RNAs (miR) are investigated that can serve as diagnostic markers and help understand GCT pathogenesis. For example, the OCT3/4 protein (transcription factor) is both a mainstream pluripotency regulator and a specific marker for GCCs. Chapter 3 underlines the specificity of the OCT3/4-A isoform for GCC showing at the same time that the other isoforms (B, B1) are not GCC specific. The findings illustrate the importance of using isoform specific primers and monoclonal antibodies against the N terminus of the OCT3/4 protein to specifically detect the pluripotency related isoform OCT3/4-A. Besides proteins like OCT3/4, specific miRs are important in GCC development/identity. miRs are short, non-coding RNAs that target mRNAs, primarily promoting their degradation and inhibiting translation. The detection of potentially functional miR-mRNA interactions is often hampered by the large number of predicted targets for each miR. Therefore, a tool (miMsg) is presented in chapter 4 to identify promising patterns of matched miR and mRNA expression from genome wide profiles. In this chapter, miMsg is also successfully applied to GCT data. Functionally, miRs are core players in GCC, facilitating the crossover between pluripotency, cell cycle regulation and therapy sensitivity of as reviewed in chapter 2. Previous studies have shown that the miR expression profile of GCC tissues reflects their pluripotent embryonic origin. Chapters 5 & 6 explore the potential clinical applicability of (embryonic) miRs as serum biomarkers in GCC. Chapter 5 describes the Targeted Serum miR (TSmiR) test, which uses the levels of the embryonic miR-371-2-3 and 302abc/367 clusters in serum as biomarkers for GCC. TSmiR achieved 98% sensitivity in detecting GCC in serum of patients versus controls, clearly outperforming classical serum markers like AFP and hCG. Prospective studies to validate and extend these findings are already initiated. Chapter 6 builds on the targeted observations in chapter 5 and validates miR-371 and 372 as GCC specific in a high throughput serum miR profile covering ≈750 miRs. In addition, potential novel serum miR biomarkers were identified, including miR-511 and, less prominent, miR-26b, 769, 23a, 106b, 365, 598, 340 and let-7a. The work presented in the third part of this thesis investigates components of the epigenetic blueprint of GCTs, and subsequently illustrates their close onco-fetal relation to (fetal) developing germ cells as reviewed in chapter 2. During embryonic development, the progenitors of mature germ cells migrate from the yolk sac, via the hindgut to the genital ridge where they further mature. During this migration and at the genital ridge, these PGCs / gonocytes undergo a characteristic epigenetic “reset,” including a reorganization of their methylation state. These changes are reflected in their malignant counterparts: GCTs. This process is extensively reviewed in chapter 9. Tools to compare the methylation status between groups of samples are widely available. Chapter 7 presents a new tool (DMRforPairs) which allows comparison of unique samples in cases where large groups of samples are not available. DMRforPairs was successfully applied to GCT cell line data in chapters 8 and 9. Chapter 8 illustrates the similarity of GCC cell lines to their PGC / gonocyte ancestor by integrating genome-wide histone modification, methylation and gene expression data. Subtle differences in the (integrated) epigenetic and expression profiles indicated the seminoma-like cell line TCam-2 to exhibit a more germ cell-like profile. The mediastinal embryonal carcinoma cell line NCCIT showed a more pluripotent phenotype. Chapter 9 investigates the genome wide methylation profiles of 91 GCTs and four representative cell lines. Different GCT subtypes could be clearly distinguished based on their global methylation profile. The methylation status of specific functional regions like imprinting control regions was also highly informative in identifying specific GCT subtypes. The observed methylation profiles matched with specific stages of germ cell development, providing insight into the developmental timing and biology of the various GCT subtypes.   To conclude, the studies presented in this thesis illuminate onco-fetal properties of GCTs from a developmental point of view, mainly by investigating (integrated) functional genomic datasets using existing and newly developed computational tools. In all studies, close collaboration with other disciplines including biologists, clinicians and engineers from various backgrounds contributed greatly to the research process and the results. All data and results are freely available online. The results presented provide insight into epigenetic and (post-)transcriptional regulation in GCT pathogenesis, focus on clinical application of the findings and contribute to the emergence of new hypothesis that can be experimentally and clinically validated in the future.