| THE
STUDY OF FIBROBLAST GROWTH FACTOR (FGF) SIGNALLING MOLECULES
IN DIAGNOSTIC PROSTATE CANCER BIOPSIES AS PROGNOSTIC
MARKERS FOR DISEASE PROGRESSION.
Principal Investigator: Dr Vincent Gnanapragasam Cancer Research UK Clinician Scientist
and Clinical Lecturer in Urology |
 |
The development of Laser Capture Microdissection of diagnostic prostate core biopsies for molecular prognostic studies in prostate cancer.
Final report (October 2006-September 2007)
In these 12 months we have successfully completed key aims of the study. Much of the time has been spent in developing and optimising the use of biopsies for expression analysis as this has never been reported before in the literature. An important consideration for us was that the technique described could be applicable to formalin fixed paraffin embedded material which is the standard for prostate biopsy. To achieve this we also carried out careful experiments to investigate the effect of formalin fixation on transcript profiles and how this might be corrected for. The result of this work is a sound foundation on which to develop real time gene profiling studies at the point of diagnosis in prostate cancer. Because of time required to develop and optimise the technique we did not collect as many biopsies as we had intended. Nevertheless, we were able to profile a number of transcripts and FGF signalling molecules in the cohort we had and demonstrate clear application of the technique. We are now preparing to extend the study with further grant applications to develop a molecular archive of prostate biopsies and to trial real time transcript prognostic studies using diagnostic prostate biopsies.
Key achievements
1. First use of laser capture microdissection in archival standard prostate biopsies.
2. Establishment of a protocol for gene transcript profiling in formalin fixed paraffin embedded prostate needle biopsies.
3. Assessment of the effect of formalin fixation on RNA transcript expression and validation of a technique to allow gene quantification in RNA from prostate biopsies.
5. Expression profiling of FGF signalling molecules in mRNA from prostate biopsies.
6. Presentation of results at national meetings, submission to international
meetings as well as the preparation of a manuscript for submission.
Laser capture microdissection (LCM) in prostate biopsies
At the completion of this study we have established the protocol for
LCM in standard archival prostate biopsies. Through multiple optimisation
steps we have defined the optimal method to mark and isolate different
tumour populations. Key basic requirements have also been defined
including the use of a fresh diagnostic slide to map malignant and
benign epithelium, the minimum number of sections required and close
involvement of a consultant histopathologist. At the completion of
this study a cohort of 10 benign and 40 cancer biopsies have been
identified, marked and microdissected. We have also isolated pure
stromal tissue from 5 biopsies. Material from all these biopsies
have been used for transcript profiling studies by quantitative real
time PCR.
Gene transcript profiling in formalin fixed paraffin embedded (FFPE)
prostate needle biopsies
In this study we established that at least 6 serial sections from each
biopsy core are required to produce sufficient RNA for analysis. We
were successful in obtaining sufficient RNA and cDNA from all samples
used for transcription analysis. In a pilot subset of samples we confirmed
transcript expression of a range of genes including GAPDH, RPL13, PSA
and Vimentin. Despite this success, one important issue which arose
was that the small quantities of cells derived from needle biopsies
produce small amounts of RNA for analysis. A future aim of our work
however, is to expand the application of standard biopsy material for
large scale transcript expression and genome wide profiling. To achieve
this, we trialled the use of RNA amplification. In the time of this
study we established that this technique could be reliably used to amplify
RNA from microdissected biopsies. This step however needs more validation
and optimisation in future work. In particular how amplification might
affect global gene expression signatures and the efficiency of microarray
platforms. To comparatively assess the relative abundance of RNA we
also successfully performed LCM on a small cohort of frozen prostate
biopsies.
Effect of formalin fixation on transcript expression and validation
of a technique to normalise gene expression
In current clinical practice, preservation of the glandular architecture
in needle biopsies is optimally achieved by formalin fixation. This
as well as archival storage inevitably results in RNA degradation. In
this study we identified that the extent of degradation might alter
transcript quantification in our samples. This would have an important
impact on the practical usability of our technique for routine prognostic
application. To investigate how transcript quantification might be affected
by RNA degradation, we used two different models which respectively
tested the effect of varying levels of degradation and the effect of
formalin fixation and archival storage. In the first instance we induced
artificial degradation in prostate cell lines. In this model we observed
that increasing degradation resulted in a progressive loss in housekeeping
and test transcript expression. Correction of test gene quantity with
a housekeeping gene however, consistently normalised expression to within
a 50% margin of the intact control samples. In the second model we compared
transcript expression in identical matched FFPE and frozen xenograft
tumours (CWR22) previously harvested in our laboratory. As expected
total expression for all genes tested was significantly lower in FFPE
samples. Normalisation of the test transcript for a housekeeping gene
however again resulted in a corrected expression to within a 50% range
of the frozen controls. These results suggest that gene expression can
be reasonably quantified in a FFPE archival biopsy provided normalisation
is made for a constant housekeeping gene and all samples are treated
in the same way. Housekeeping gene correction therefore allows quantification
to within an acceptable and consistent margin of error provided all
samples are treated and processed in an identical manner.
Transcript expression profiling in a clinical cohort of biopsies
Our results have shown that RNA suitable for quantitative real time
PCR can be derived from archival FFPE microdissected prostate biopsies.
To test the validity of this protocol we investigated if this method
could identify expected differences in gene expression in malignant
compared to benign prostate epithelium. To do this we tested for
transcript expression of the polycomb group protein EZH2 which is
known to be over-expressed in prostate cancer. In these experiments
EZH2 (normalised for a housekeeping gene) was expressed in 79% of
malignant glands but in only 27% of benign glands studied. No expression
was seen in stromal controls. Among the cancers, expression was seen
in 85% of grade 4 and 5 tumours and in only 55% of grade 3 tumours.
These results are entirely consistent with the published literature
and confirm that distinct differences in expression patterns between
benign and malignant glands as well as stromal tissue can be detected
using our technique. We also investigated the expression of epithelial
and stromal markers within dissected sections to test the accuracy
of microdissection. Using primers for the PSA transcript we were
able to demonstrate strong expression in epithelial cells but no
expression in stromal cells. In contrast, expression of the filament
protein vimentin was up to 10 fold higher in stromal cells compared
to microdissected glandular epithelium.
Expression of FGF signalling molecules in prostate biopsies
Work in our laboratory has defined important changes in components of
the FGF signalling system in prostate cancer. Ligands and receptors
are increased in aggressive cancers while endogenous inhibitors are
down-regulated. Recent studies have demonstrated key roles for the
FGF receptors 1 and 4 and for the tyrosine kinase inhibitor Sef.
In contrast the signalling inhibitor Sef has been shown to be down-regulated
in prostate cancer. These molecules were studied in prostate biopsies
with similar trends observed as in the published literature. The
numbers in our series at the stage of completion however was not
sufficient to definitively identify if these molecules were strong
prognostic markers. A larger study is proposed now that the basic
methodology has been established. Our work however supports the use
of prostate biopsies for transcript investigation as prognostic markers.
Presentation of results and preparation of manuscript
The preliminary results of this study were presented in the British
Prostate Group meeting in March of this year. More recently we have
presented our updated work at the NCRI meeting in Birmingham. A paper
is currently being prepared for submission describing the technical
details we have established. An abstract has also been submitted
for next year’s European Association of Urology meeting.
Finally
We hope that the medical committee will agree that we have made good use of the research money granted by Prostate UK (formally the Prostate Research Campaign UK). This work will now form the basis of a much larger programme that will seek to establish the application of molecular profiling at the same time as routine histopathological diagnosis and grading. We full intend to approach Prostate UK again for further funding and hope that an application will be received favourably.
Research Final report dated 12 November 2007
Project 2005/06