miR 371-373 and miR 302 cluster expression in body fluids as potential biomarkers of malignant germ cell tumours

Share :
Published: 29 May 2013
Views: 5277
Rating:
Save
Dr Matthew Murray - Cambridge Cancer Centre, UK

Dr Murray presents data on potential biomarkers for maliganant GCTs at the 3rd International CNS Germ Cell Tumour Symposium in Cambridge, UK.

The second abstract which I’ll present today which is the use of these microRNAs in serum body fluids as potential biomarkers of this disease and then will lend itself to Dr Terashima’s talk after that.

So the case I’m going to describe first is a four year old boy who presented to our institution with a large extra-cranial malignant germ cell tumour that histologically was a yolk sac tumour and had raised AFP. So from a serum sample that was left over after AFP and HCG estimation we extracted RNA and then we performed a PCR looking at the eight microRNAs that I’ve previously identified to you as being overexpressed in these tumours, so microRNAs from the miR-371-373 on the left and the miR-302 family shown here on the right. We did a PCR involving the pre-amplification step and when we compared the levels to normal pooled serum from healthy individuals shown here on this log scale we saw that the levels were highly elevated at the time of diagnosis, in particular miR-371-373 microRNAs with miR-372 being over 700-fold higher than in pooled normal serum.

So that was at diagnosis, we then obtained longitudinal follow-up samples in this patient and that data is shown here in this graph. So we have time from initial diagnosis out here. This patient received a biopsy and then six cycles of chemotherapy and surgery around day 100 and what we note is the miR-372 levels where the serum sample was available fell to normal and remained normal in uneventful clinical follow-up which mirrored the fall in the AFP level which was shown in the green circles.

So this is a marker positive case. What we then went on to do was look in a slightly larger cohort and that data is shown here. This now includes a small number of cases of both paediatric and adult disease; it includes gonadal and extra-gonadal cases and a range of histological subtypes we’ve previously looked at. The graph on the left is for miR-372 which is from the miR-371-373 cluster and miR-367 shown here on the right is from the miR-302 cluster. Importantly although the three yolk sac tumour cases were positive for AFP the five cases shown here which are for embryonal carcinoma and for seminoma were marker negative.

Subsequent to this an independent German group using the same technology demonstrated in adult patients with testicular disease that these three microRNAs are high at diagnosis, shown by the black bars and that after orchidectomy the levels fall to the same levels as seen in healthy individuals which is shown in the third column. They then looked at those samples individually and looked at the three microRNAs and what they showed was that following treatment with orchidectomy for stage 1 disease the levels fell with miR-371 in this case having some of the… all falling and having significant falls following treatment.

So where we are with this work now is that we have combined our samples with the German group, this is work being led by Linda at Erasmus in the Netherlands. This is now looking at a total of 128 patients and this is a receiver operating characteristic curve: 79 malignant germ cell tumours and 49 control samples of patients who have other testicular pathology including testicular lymphoma, epididymitis, to use a negative control. What we see with this is that each of the individual microRNAs has a very high area under the curve, 0.88 and up to 0.96 individually, showing that they’re highly sensitive and specific. If we use a cut-off of 90% for specificity then all of these microRNAs individually have a sensitivity of over 80%. When these four microRNAs were combined with AFP and HCG then all malignant germ cell tumours were correctly identified compared to non-malignant controls.

So the conclusions from this work are that these microRNA clusters are over-expressed in serum in extracranial malignant GCT diagnosis and the levels fall during treatment. It therefore has a potential application for non-invasive diagnosis and monitoring. Some of the benefits, therefore, of looking both in the serum and CSF for samples in intracranial germ cell tumours will be important. That leads very nicely onto Dr Terashima’s talk. Thank you.