Get documents about "A New Sensitive and Accurate Diagnostic Test for Prostate
Document Sample
28th February 2007
Prostate Research Campaign : Project Report
A New Sensitive and Accurate Diagnostic Test for Prostate Cancer
Dr Christiane Fenske
Christodoulos Pipinikas
Medical Genetics
Clinical Developmental Sciences
St George’s University of London
1
CONTENTS:
THE DIAGNOSTIC TEST ……………………………………..……..……………………...….. 3
TOWARDS TEST VALIDATION………….…………………………………………………….. 3
Statistical Analysis ………………………………..………………………..…. 3
Patient recruitment……………..…………………………………..………….... 3
PAX (Bioanalytix) Tube………………………………..……….………………..3
FACS: Fluorescence-Activated Cell Sorting ……………….…..................... 4
Inflammatory markers…………………………………………………………… 4
Commercialisation and development……….. …..…………………………… 4
INVESTIGATION OF ADDITIONAL MARKERS …………………………………….……...... 4
Therapeutic role/mechanisms…………………………………....……………. 4
Early screening role…………………………………………….……………….. 5
STATISTICAL APPROACHES……………………………………………..……………..……. 6
PROJECT CONCLUSIONS……………………………………………………...……………… 6
GLOSSARY- abbreviations………………………………………………….…………………... 6
APPENDIX 1:
Discriminant diagnosis between cancer and non-cancer……..…………….………………. 7
APPENDIX 2:
Differential diagnosis between stages of prostate cancer development…………………… 7
APPENDIX 3:
FACS analysis of prostate specific cells ………………………………………………………. 8
APPENDIX 4:
Analysis of marker expression in patients with inflammatory disease……………………….9
APPENDIX 5:
Interactions between genes involved in CaP initiation, progression and aggressiveness... 10
2
The interim report (August 2006) from this research group has demonstrated increased
accuracy, specificity and sensitivity of our blood-based quantitative RT-PCR diagnostic test,
by adding to our patient numbers and applying a number of robust approaches in statistical
analysis, identifying key marker combinations.
I. THE DIAGNOSTIC TEST
We have shown that 4 markers, when used in combination in qRT-PCR using patient blood
samples, result in over 95% accurate discriminant diagnosis between cancer and non-
cancer (appendix 1).
α
Two of these, hypoxia-inducible factor 1α (HIF-1α) and E2F3, a transcription factor, are up-
regulated early on in disease development.
Matrix metalloproteinase 9 (MMP9), is activated in the late metastatic stages.
RECK, an inhibitor and down-regulator of MMPs, suppresses the invasive and metastatic
potential in many tumors. The down-regulated levels of RECK (resulting in tumour
development), when taken together with increased levels of E2F3, HIF-1α and MMP9, as a
ratio gives highly differentiated values, resulting in the potential of increased discriminant
diagnosis.
Differential diagnosis of developmental stages
Specific combinations of these markers are able to differentially diagnose the stages of CaP
with E2F3, HIF-1α and MMP9 being excellent for diagnosing early stages of CaP (NEOM vs
LocCap) HIF-1α differentially diagnosing the later stages of disease development (LocCap vs
MetCap).
Response to therapy
E2F3, MMP9 and RECK are all excellent markers for differentiating between NEOM and
MetCap, and hence may be useful in monitoring response to therapy (appendix 2 shows %
correct calling and ROC curve analysis where high values for area under curve, AUC,
indicates excellent discriminatory power of marker).
The inclusion of serum PSA levels with quantified expression levels of these markers
increases diagnostic accuracy to an overall sensitivity and specificity of 96%.
II. TOWARDS TEST VALIDATION
i. Statistical Analysis: New collaborations with statisticians at the University of Cranfield
have focused on in depth analysis of our data, examination and close scrutiny of our sampling
methods and procedures, and statistical approaches.
Results have been excellent, confirming our findings, thereby increasing confidence in the
accuracy of diagnosis our test provides, and offering an extremely optimistic view of further
validation procedures.
ii. Patient recruitment: A new patient recruitment system has been established (in
collaboration with Dr Tom Swallow and Dr Chris Anderson of the Urology Department, St
George’s Hospital) with increased control regarding the time scales involved in blood sampling
for qRT-PCR, serum PSA and biopsy.
Results: This allows greater alignment of qRT-PCR results with serum PSA and diagnosis
These tightly controlled parameters allow for increased confidence in correlation of our results
with diagnosis.
iii. PAX (Bioanalytix) tube
This blood tube allows the collection of blood samples and storage/transportation at ambient
temperature, with guaranteed stability of mRNA for up to 10 days.
3
Results: Having validated these claims, we are now routinely using the tube, enabling
maximisation of patient recruitment (approximately doubling numbers of patient recruited each
week). The incorporation of the use of the tube in further test validation and standardisation
will allow patient blood sampling with subsequent transportation to a routine testing laboratory
iii. FACS: Fluorescence-Activated Cell Sorting
The technique, constituting a control for the potential diagnostic test, is essential for increased
confidence in the accuracy of the test, demonstrating whether variation in marker expression
levels is due to increased numbers of cells or due to differences in expression levels.
The technique uses fluorescently tagged antibodies to identify and count prostate-, cancer-
and prostate cancer- specific cells in patient blood samples.
We have optimised the technique using HER-2 as a cancer-specific marker.
This lengthy and expensive technique has proven to be complicated.
Prostate cancer cells have been cultured, harvested, purified and counted in the optimisation
of the technique
Results have been excellent, with distinct and accurate identification of cancer cells, allowing
their enumeration (appendix 3)
Further optimization of the technique is necessary, using prostate- (PSMA) and additional
cancer-specific (EPCAM) antibodies, and different prostate cancer cell lines, reflecting
different metastatic potential (low→moderate→high). Once we are able to identify and count
specifically prostate cancer cells in control samples, the technique will be applied to patient
blood samples, and cell numbers correlated with marker expression levels, and disease
diagnosis, resulting in a full picture identifying causes of marker expression level variation.
FACS is not envisaged to be part of the final diagnostic test.
iv. Inflammatory markers
Results:
Our interim report has shown correct diagnosis of 92% patients with inflammatory disease,
using the 4 key markers outlined above, confirming their prostate cancer-specificity, with some
possible up-regulation in patients with inflammatory disease, but at much lower levels than
that seen in prostate cancer patients (appendix 4).
Therefore, inflammation (seen in some patients with prostate disease) will not be a potential
factor of error in the diagnostic test.
v. Commercialisation and development
Our excellent results, together with the confirmatory research, have resulted in a great deal of
external interest (IPSO Ventures), with a view to standardisation, implementation of standard
operating procedures, validation, formatting, and bringing the test into routine use.
III. INVESTIGATION OF ADDITIONAL MARKERS
Several new markers, investigated regarding their potential of increasing the strength of
diagnosis of the different stages of CaP, have also shown promising potential role in therapy
i. Therapeutic role/mechanism
The human epidermal growth factor receptors have demonstrated both therapeutic and
prognostic significance in breast cancer.
HER2, a proto-oncogene and member of the epidermal growth factor receptor (EGFR) family,
is notable for its role in the pathogenesis of breast cancer and as a target of treatment (using
the monoclonal antibody trastuzumab, marketed as Herceptin). However, trastuzumab is only
effective in breast cancer where the HER2/neu receptor is over expressed.
HER2 is part of the pathway, including MMP9 and RECK, respectively up-regulating and
down-regulating these markers. Therefore, identification of HER2 over-expression in CaP
4
patient blood samples and possible interaction with MMP9 and RECK has been the focus of
our research.
Further research will determine the value of including HER2 in our diagnostic test. In
particular, due to its proven effectiveness in antibody therapy of breast cancer, HER2 may
also be of value in CaP therapy
HER2 is up-regulated and HER4 down-regulated with tumour progression. Therefore,
combining results as a ratio may increase diagnostic accuracy.
Results:
Primers for HER2 and HER4 have been designed and manufactured.
qRT-PCR has been optimised using prostate cancer cell lines (LnCAP, Du145, and PC3)
HIF-1α affects the expression of vascular epithelial growth factor (VEG-F), whose over-
expression is known to be associated with various tumours. The inclusion of VEG-F in our
diagnostic test may provide a further opportunity for therapy via anti-sense technology, which,
by blocking VEG-F expression, will suppress tumour progression. In addition, its interaction
with HIF-1α and, therefore, also the carbonic anhydrases, may offer clues regarding the
mechanism of CaP development.
Results: Primers are being designed and synthesised for optimisation.
Appendix 5 shows interaction of these markers, demonstrating their synergistic action. This
may well prove to be a basis for a diagnostic test involving their combination, and resulting in
high degrees of accuracy.
In addition, therapy involving the specific targeting of one marker (e.g HIF-1α or HER2) may
well have a knock on effect, affecting transcriptional regulation and expression of markers
further down the metabolic pathway (e.g. VEG-F, MMP9 and carbonic anhydrases, or MMP9
respectively).
ii. Early screening role
Alternative splicing
Results: Fatty acid synthase (FAS), carbonic anhydrase IX (CAIX), EZH2, and e-cad initially
suggested the possibility of alternative splicing. However, further experiments, using
increased patient numbers have not confirmed this phenomenon.
Recent publications have identified two markers EPCA and IGF-1, up-regulated before
prostate cancer is even diagnosed and, therefore, early risk indicators for prostate cancer.
These are currently being optimised for qRT-PCR.
Early prostate cancer antigen (EPCA), a nuclear matrix protein, identified as a novel
prostate cancer marker, has shown expression in negative biopsies of men who, as much as
5 years later, were diagnosed with prostate cancer.
Results:
Publication merely of the protein sequence has meant that primer design has been a more
than usually laborious task, involving the identification of amino acid sequences that would
result in a minimum number of base ’wobbles’, the manufacture of degenerate primers and
intense optimisation.
Work is ongoing
Higher levels of insulin-like growth factor 1 (IGF-1), linked to increased risk of CaP in
younger men, precede clinical diagnosis as early as 5 years.
Results:
5
Primers have been designed for IGF-1 and qRT-PCR is currently being optimised using CaP
cell line mRNA. Following optimisation, qRT-PCR will be carried out on patient samples,
correlating results with diagnosis as before.
Our ultimate research aims are to incorporate these markers into our diagnostic test, allowing
its use in a potential prostate cancer screening programme.
Validation of these markers in early screening will be the basis on a longitudinal study, to be
designed and established separately form this research.
IV. STATISTICAL APPROACHES
Further statistical approaches are currently being investigated, including artificial neural
networks (ANN) and support vector machine analysis (SVM) with assisted algorithms.
These will enable the identification of any interaction between risk factors for CaP e.g. marker
expression levels, age, Gleason score, ethnicity, and the extent of their effect. The ultimate
application will be that by taking all factors into consideration, increased discrimination
between all 3 patient groups (BPH, localised, and metastatic cancer) may be achieved.
V. PROJECT CONCLUSIONS
We have established a non-invasive blood-based sensitive molecular prostate cancer
diagnostic test, consisting of 4 markers and with 95% accuracy of differential diagnosis
between cancer and non-cancer. Following diagnosis of cancer, subsequent testing will
involve determination of stage of cancer development. Robust control experiments and
statistical analysis have confirmed and justified confidence in its accuracy.
The test is now ready for validation, standardisation and formatting, prior to introduction for
routine diagnosis. At this point high throughput automation should be a crucial consideration.
Further markers are being investigated to determine whether their inclusion in the test will
increase diagnostic accuracy of developmental stages. Specific markers have additional
value in potential therapy, enabling not only diagnosis, but also specific targeted therapy.
Others, due to very early changes offer the potential of prostate cancer screening- identifying
risk up to 5 years before disease onset. These markers are still at the early stages of our
analysis.
GLOSSARY
Abbreviations
RT-PCR reverse transcriptase polymerase chain reaction
qRT-PCR quantitative reverse transcriptase polymerase chain reaction
NEOM no evidence of malignancy
BPH benign prostatic hyperplasia
LocCap localised cancer
MetCap metastatic cancer
6
APPENDICES
1. Scatter/Dot – CaP discriminant diagnosis (cancer vs non-cancer)
α
Using combination of HIF-1α, E2F3, MMP9 as ratio with RECK
7
2. Differential diagnosis between stages of CaP development
i. % correct calling
NEOM LocCap MetCap % correct
NEOM 19 2 0 90%
LocCap 2 33 9 75%
MetCap 1 1 9 82%
22 36 18 Over all 80%
ii. ROC analysis for differential diagnosis (higher area under curve , AUC, values
indicate increased marker discriminatory power
marker NEOM vs LocCap vs NEOM vs
LocCap MetCap MetCap
E2F3 0.879 0.691 0.803
HIF-1a 0.883 0.87 0.546
MMP9 0.812 0.788 0.944
RECK 0.86 0.857 0.981
3. FACS analysis
Mouse anti – HER-2/neu conjugated to phycoerythrin (PE) to stain DU145 prostate
cancer cells. Positive staining of DU145 prostate cancer cells for HER-2/neu (top) and
use of the appropriate isotype control (mouse IgG1 conjugated to PE) to correct for
any background, non-specific binding/staining (bottom).
8
4. Analysis of marker expression in patients with inflammatory disease
i. % correct calling
Discriminant analysis classification matrix including all 4 test markers
α
(E2F3, HIF-1α, MMP9, RECK)
Benign CRP Malignant %correct
Benign (n=23) 21 0 2 91
CRP (n=10) 0 9 1 90
Malignant 1 3 51 93
(n=55)
Total 22 12 54 92
ii. Discriminant analysis between prostatic (benign and malignant) and CRP
cases.:
A canonical scores logarithmic plot showing discrimination using HIF-1α, E2F3, MMP9
and RECK combination.
Canonical Scores Plot
6
4
FACTOR(2)
2
0
CLASSIFICAT
-2
Benign
CRP
-4 Malignant
-4 -2 0 2 4 6
FACTOR(1)
9
5. Interactions between genes involved in CaP initiation, progression and
aggressiveness.
Red and green arrows indicate increased and reduced transcriptional activity, respectively.
ECM : extra cellular matrix
EZH2 E-cad (EZH2/E-cad)
E2F3
c-MET/HGF
MTSP-1
AR?
IGF-1 uPAR
α
HIF-1α
MMPs
(ECM degradation)
IGFBP-3
VEGF Carbonic anhydrases
HER2 RECK
10
Related docs
