Speaker: Marcel Kool
Department: JNI Oncology
Subject: Molecular classification of pediatric brain tumors
reveals new entities
Location: Erasmus MC
Author: Hielke Walinga
Times that only surgery was a treatment for cancer are long over. Very different treatments have been developed and knowing which one to use for a patient can be crucial for the survival of the patient or can prevent severe damage to other tissues near the tumour. Especially very vulnerable tissues like the brain must be treated with the exact right treatment. Therefore some researchers focus on the classification of these tumours. Just like Marcel Kool whose seminar I listened to at 2 September. Marcel Kool tries to classify paediatric (children’s) brain tumours using molecular methods. He gave an overview of the latest discoveries of the field.
Cancer types always used to be classified by their histology (their appearance when viewing it under a microscope). This makes sense, since the way a cancer looks is usually a result from the specific mutations it has. However molecular research has shown that this classification isn’t always correct for guessing the type of cancer.
For example oligo astrocytoma, which looks like a bunch of star like structures, doesn’t exist on the molecular level. It is actually at the molecular level either astrocytoma (star like structure) or oligodendroglia (a bunch of tree like structures).
Another way to distinct different cancer types is to look at the clinical outcome. The survival rate should be a good sign on the different mutations of the tumour, and might be later be linked to something that good help to set a good diagnosis. This classification is done using grades. Higher grades (from one to three) stand for a worse outcome. However molecular research has shown that there is actually no difference between two and three of most types.
To reveal the gene expression the molecular methods that could be used for this research are for example Northern or Southern blotting, however it’s quite hard to obtain enough sample from the tumour, especially because it’s in the brain, to do these kind of research. A better method might be to use NanoString to reveal certain fragments RNA. A method, that’s better to be done, and also studies by Marcel Kool, is actually to reveal DNA methylation. However, it’s left to discussion on how this actually relates to the nature of the tumour.
An example of a methylation heat map created in a research in which Marcel Kool has cooperated. (Source: Hendrik Sturm, Hotspot Mutations in H3F3A and IDH1 Define Distinct Epigenetic and Biological Subgroups of Glioblastoma, Cancer Cell: Volume 22, Issue 4, 16 October 2012, Pages 425–437)
The research shows a lot of new entities of brain tumours, and the researchers also have been able to link the results to different causes of the cancer. Their distinction almost always came from different pathways that are affected by some kind of mutations. Sometimes they were able to link these together to one certain factor, like the MITF TF of the AT RT tumours. But it also revealed that glioblastoma is a cancer caused by a histone gene mutation. They also were able to link their results to the clinical outcome of the disease. For example, medulastoma has a molecular subgroup (Wnt) which almost always meant a survival for the patient.
To use all these result for the benefit of the patient, the next step would be to create a worldwide database of some sort, which doctors could use by their diagnosis of the patients. Therefore this will be the next step in molecular neuropathology. Actually this has partly already begun and it is called INFORM (INdividualized therapy FOr Relapsed Malignancies in childhood). Still this only focusses on the relapse cases, because they usually have a dismal prognosis.