EACR 2016
Glioblastoma heterogeneity in circulating tumour DNA
Dr Joan Seoane - Vall d'Hebron Institute of Oncology (VHIO), Barcelona, Spain
What I do in the Vall d’Hebron Institute of Oncology I am the Director of the translational research programme and I have a group that is devoted to studying glioblastoma which is the most common brain tumour. So what we do is we try to understand the biology of glioblastoma, trying to look for new therapeutic targets, biomarkers to stratify patients to be treated, mechanisms of resistance, but also we look for circulating biomarkers to try to characterise the tumour in a non-invasive manner.
What have you found so far?
We found that, for example, it’s very important to detect circulating tumour DNA in the CSF of patients because by sequencing the DNA that is in the CSF you can characterise, you can know, what are the genomic alterations found in the brain tumour. Also we have, in a separate study, looked at the characteristics of the cancer initiating cells that are present in glioblastoma and how these cells are regulated by a factor called LEAF.
What techniques do you use to extract the circulating tumour DNA?
This is free DNA so what happens is the tumour, when it grows, some of the cells die and then they release their DNA to the microenvironment and then this DNA is washed out by the cerebrospinal fluid that is bathing the whole parenchyma. This CSF, cerebrospinal fluid, goes from the parenchyma down to the spinal cord. So if we have a lumbar puncture and we obtain a sample of this CSF and then we capture the DNA that is present there we can sequence it and so we can see which are the genomic alterations of the tumour without having to do a biopsy which, as you can imagine, is very invasive because you have to go through the skull and get into the brain.
How successful does this tend to be?
This is much more successful than assessing ctDNA, circulating tumour DNA, in the blood. So for some reason brain tumours do not release so much ctDNA into the blood, they normally get it into the CSF and then the concentration of ctDNA is higher because the amount of normal DNA is very low so the ctDNA that you find is normally all coming from the brain tumour. Second, the volume is smaller than the blood so that’s why it’s not so diluted.
Do you see this happening more frequently in clinics now?
Yes, we are making a future to try to translate these discoveries into the clinic because we think that it’s very important to really characterise the tumour in a non-invasive, or relatively non-invasive, manner, especially brain tumours which are difficult to access, in order to improve the treatment but also to try to know which is the prognosis of these tumours. We want to know whether the tumour is benign or aggressive because this will dictate the type of treatment that we have to do.
Would this be useful for biomarker work?
Of course, this is something that we think is very beneficial for diagnosis, so to know exactly which is the tumour that you have in the brain. Second, prognosis – whether it is benign or aggressive. Third, monitor the tumour growth – whether it is responding or not. We know that the levels of ctDNA correlate with the burden, with the tumour burden, so if the tumour is reducing you have less levels of ctDNA and if the tumour is progressing you have more levels. Then obviously if we are lucky and we see some of the genomic alterations that are druggable, that there is a drug that can be used, then we can also impact on the therapy.
What else is your lab looking into?
Our department is, as I said, the translational department so what we want to do is translate as fast as possible the discoveries in basic research to the patient. We have a very nice multidisciplinary group that includes many different disciplines – oncologists, pathologists, surgeons and scientists – all together trying to help in understanding and trying to help in the treatment of patients. One of the models that we are using is what we call PDX models, patient derived xenograft models, so basically we obtain the sample of the tumours of the patients that are undergoing surgery in our hospital and with this sample we inoculate them into mice to recapitulate the tumour of a patient in a mouse. Then we can study the tumour in the mouse and also we can use that model as a preclinical model to understand the mechanism of action of some of the compounds that are in development.
