Speaker: Dr. Marcel Kool
Department: Paediatric neuro-oncology
Subject: Molecular classification of paediatric brain tumours reveals new molecular entities
Location: Erasmus MC, Rotterdam
Author: Romano van Genderen
The talk about Dr. Kool started with a small piece of background information on brain tumours. Some of the key characteristics of these are that there are extremely many types, compared to for example breast cancers. These types differ in “grade”, which is a number varying from 1 to 3, which experts use to determine the intensity of therapeutic treatment. Another thing they differ in is morphological structure. But Dr. Kool claimed that these distinctions are not enough. Even tumours that appear morphologically similar, differ a lot on the molecular scale.
To investigate the tumours on this scale, multiple techniques were used in the past, but each one of them has a specific shortcoming. Identifying a molecular subgroup based on mRNA expression is accurate, but it needs very fresh and undamaged samples of mRNA. For immunohistochemistry, very specific and accurate antibodies are needed, making this method useless in the case of rare sorts of tumours. Nanostring methods are accurate and can even use broken strands of mRNA, but the morphological type of the cell must be pinpointed perfectly, and you wouldn’t risk treating a patient with the wrong medicine based on an oversight by a pathologist. But now the scientific community has finally identified a method which seems to have no drawbacks. This method is based on identifying the tumours based on the DNA methylation pattern. Methylation is the process of adding -CH3 tags to the DNA in order to change the intensity of non-covalent interactions with the DNA.
Afterwards, he dealt with the results of this method. The first was medulloblastoma, a form of cancer located in the lower back of the brain, which appeared to have not 1, but 4 types. The type 1 was usually over-treated, leading to unnecessary brain damage, while type 3 was usually undertreated. Afterwards, he showed that pilocytic astrocytoma’s, which are a cancer in the astrocytes (star-shaped cells responsible for transferring nutrients from capillaries to neurons), have 2 pathways associated with them instead of one. Then he showed that glioblastomas (cancers in the glial cells, a cell group that supports and assists neurons in the central nervous system) can be caused by 2 different mutations in the same gene. Then he showed evidence that ependymal tumours (tumours in ependymal cells, the cells that produce the cerebro-spinal fluid that fills the cavities in the brain) have 9 subtypes and that in this case the usual type II and type III tumours are identical. Then he mentioned that all different classes of atypical teratoid rhabdoid tumours are in fact identical and that the differences were caused by varying degrees of expression.
Radiographics. 2005 Mar-Apr;25(2):486-90.
Best cases from the AFIP: supratentorial ependymoma.
Mermuys K, Jeuris W, Vanhoenacker PK, Van Hoe L, D’Haenens P.
The final type of tumour he mentioned was the CNS-primitive neuroectodermal tumour, which is caused by a mutation in the common progenitor of neurons, glial cells and ependymal cells (neuroectodermal cells). These cells look like poorly differentiated mature cells. After investigating their genome, he came to a very surprising result, namely that these tumours, based on their methylation pattern, almost all belong to other classes of tumours.
Finally he talked about his future plans, which were to make methylation databases to identify these tumours much more easily, as he mentioned, in a span of weeks.
This talk really showed very well how the collaboration between pathologists, oncologists and molecular biologists can help us learn more about cancer and work to come closer to better and faster treatments.