The methylation profile of cytosines in CpG-dinucleotides in promoter regions and other, mainly regulatory, sequences is an important determining factor of cell differentiation. The methylation profile of a tumor cell depends on the cell-of-origin from which the tumor cell originates. A second important factor that determines the methylation profile of a specific tumor cell is the tumor type. For brain tumors, a complete classification based on methylation profiling already exists. Methylation profiling has become a very useful tool for the diagnosis of difficult brain lesions.

In addition to the classification of brain tumors, there are several other interesting and promising applications for methylation profiling, such as Cancer of Unknown Primary (CUP), aggressive lymphoma, and early detection of liver cancer occurring in patients with liver cirrhosis. At the moment, methylation profiling is mainly based on array technology, which is only possible on biopsies from tumor tissue. The development of methylation profiling based on sequencing (e.g., cell-free reduced representation bisulfite sequencing, cfRRBS, developed in the lab of Prof. Nico Callewaert) enables methylation profiling on liquid biopsies, such as blood samples.

Cancer of unknown primary forms an important medical problem. A frequent diagnostic issue for practicing pathologists is the determination of the primary tumor from which a metastasis has originated. In about one-third of cases, it remains unclear from which primary tumor the metastasis originates, even after a complete immunohistochemical work-up. On the other hand, an exact typing of a metastasis is of utmost importance for further treatment of the patient. Our aim is the development and clinical validation of a method for methylation profiling based on sequencing for the specific diagnosis of CUP on formalin-fixed, paraffin-embedded tissue biopsy material. The same methodology will be applied to cell-free DNA in blood samples from patients with a CUP. The methylation pattern of DNA that is released from dying (apoptotic or necrotic) tumor cells is tumor entity-specific and is therefore diagnostically very useful. Our research supports the development of methylation profiling technology, with a focus on innovations that increase sensitivity and specificity. This is crucial for adequate tumor typing on tissue biopsies, as well as liquid biopsies. For tissue biopsies, our research group has access to a large tumor biobank. The biobank for liquid biopsies from patients with a CUP is set up prospectively by a dedicated data manager.

The same methylation profiling technology can be applied to liquid biopsies from patients with aggressive lymphoma. For the diagnosis of lymphoma, a tissue biopsy remains the gold standard, but Hodgkin and diffuse large B-cell lymphoma often affect organs that are difficult to reach. In addition, the differential diagnosis usually includes a wide range of malignant and non-malignant conditions, which makes a non-invasive test highly beneficial for patients. In a pilot study, using shallow whole-genome sequencing on cfDNA in blood, we demonstrated that abnormal copy number variant (CNV) profiles are detected in 84.2% and 74.1% of Hodgkin and diffuse large B-cell lymphoma patients, respectively, clearly proving the feasibility of a liquid biopsy. Moreover, CNV profiles are specific enough to differentiate Hodgkin from diffuse large B-cell lymphoma with high accuracy. To improve sensitivity and specificity, we will additionally apply a genome-wide methylation profiling technique (cfRRBS, see above) that works well on cfDNA. Methylation profiles are highly tumor-specific and thus may greatly enhance the specificity of our test. We hypothesize that combining CNV and methylation profiling will lead to a practical non-invasive test for lymphoma. For this research, we have prospectively collected liquid biopsies from more than 500 patients.

Other clinically important applications are the development of a liquid biopsy for early detection of liver cancer in liver cirrhosis patients and molecular profiling of soft tissue sarcomas. This research is a collective effort of numerous researchers and research teams from UGent and UZ Gent.